LLVM Design:

15 Dec 2003 · Misha Brukman

longjmp
253.perlbmk
configure
configure.ac
configure
configure
llvm/projects
-licm
-basicaa
-basicaa
format
format_arg
non_null
longjmp
longjmp
setjmp
llvm/
llvm/
@@ -388,9 +386,19 @@
       foo(X);
     }
 
unwind
longjmp
cd where-you-want-the-front-end-to-live
gunzip --stdout cfrontend.platform.tar.gz | tar -xvf
+  
gunzip --stdout cfrontend-version.platform.tar.gz | tar -xvf
       -

 

 
-If you are on a Sparc/Solaris machine, you will need to fix the header
-files:
+If you are using Solaris/Sparc or MacOS X/PPC, you will need to fix the
+header files:
 
 cd cfrontend/sparc

    ./fixheaders
@@ -443,7 +496,8 @@
 Once checked out from the CVS repository, the LLVM suite source code must be
 configured via the configure script.  This script sets variables in
 llvm/Makefile.config and llvm/include/Config/config.h.  It
-also populates OBJ_ROOT with the Makefiles needed to build LLVM.
+also populates OBJ_ROOT with the Makefiles needed to begin building
+LLVM.
 
 The following environment variables are used by the configure
 script to configure the build system:
@@ -477,7 +531,7 @@
   --with-llvmgccdir=LLVMGCCDIR
   

     Path to the location where the LLVM C front end binaries and
-    associated libraries will be installed.  This must be specified as an
+    associated libraries were installed.  This must be specified as an
     absolute pathname.
     
   
--enable-optimized
@@ -488,7 +542,8 @@
     
   
--enable-jit
   

-    Compile the Just In Time (JIT) functionality.  This is not available
+    Compile the Just In Time (JIT) compiler functionality.  This is not
+    available
     on all platforms.  The default is dependent on platform, so it is best
     to explicitly enable it if you want it.
     
@@ -521,10 +576,10 @@
 LLVM_LIB_SEARCH_PATH environment variable in your startup scripts.
 This environment variable is used to locate "system" libraries like
 "-lc" and "-lm" when linking.  This variable should be set to
-the absolute path for the bytecode-libs subdirectory of the GCC front end
-install, or LLVMGCCDIR/bytecode-libs.  For example, one might set
+the absolute path of the bytecode-libs subdirectory of the GCC front
+end, or LLVMGCCDIR/bytecode-libs.  For example, one might set
 LLVM_LIB_SEARCH_PATH to
-/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the X86
+/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the x86
 version of the GCC front end on our research machines.
 
 
@@ -1091,7 +1146,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/13 19:45:28 $
+  Last modified: $Date: 2003/12/16 16:26:26 $
 
 
 




From criswell at cs.uiuc.edu  Tue Dec 16 10:52:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 10:52:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.i686-redhat-linux-gnu.tar.gz cfrontend-1.1.sparc-sun-solaris2.8.tar.gz
Message-ID: <200312161651.KAA27586@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.i686-redhat-linux-gnu.tar.gz updated: 1.1 -> 1.2
cfrontend-1.1.sparc-sun-solaris2.8.tar.gz updated: 1.1 -> 1.2

---
Log message:

Checkin of newer, faster GCC frontend.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.i686-redhat-linux-gnu.tar.gz


Index: llvm-www/releases/1.1/cfrontend-1.1.sparc-sun-solaris2.8.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 10:54:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 10:54:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.source.tar.gz
Message-ID: <200312161653.KAA27692@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.source.tar.gz updated: 1.1 -> 1.2

---
Log message:

Source code for newer GCC frontend.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.source.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 10:57:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 10:57:02 2003
Subject: [llvm-commits] [release_11] CVS: llvm/docs/LLVMVsTheWorld.html
Message-ID: <200312161656.KAA27805@choi.cs.uiuc.edu>


Changes in directory llvm/docs:

LLVMVsTheWorld.html (r1.5) removed

---
Log message:

Not ready for prime-time.



---
Diffs of the changes:  (+0 -0)



From criswell at cs.uiuc.edu  Tue Dec 16 11:32:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 11:32:01 2003
Subject: [llvm-commits] [release_11] CVS: llvm/docs/Stacker.html
Message-ID: <200312161731.LAA28000@choi.cs.uiuc.edu>


Changes in directory llvm/docs:

Stacker.html updated: 1.5 -> 1.5.2.1

---
Log message:

Fixed some punctuation and grammar.



---
Diffs of the changes:  (+73 -73)

Index: llvm/docs/Stacker.html
diff -u llvm/docs/Stacker.html:1.5 llvm/docs/Stacker.html:1.5.2.1

--- llvm/docs/Stacker.html:1.5	Sun Nov 30 23:12:09 2003
+++ llvm/docs/Stacker.html	Tue Dec 16 11:31:42 2003
@@ -61,7 +61,7 @@
 about LLVM through the experience of creating a simple programming language
 named Stacker.  Stacker was invented specifically as a demonstration of
 LLVM. The emphasis in this document is not on describing the
-intricacies of LLVM itself, but on how to use it to build your own
+intricacies of LLVM itself but on how to use it to build your own
 compiler system.
 
 
@@ -70,7 +70,7 @@
 Amongst other things, LLVM is a platform for compiler writers.
 Because of its exceptionally clean and small IR (intermediate
 representation), compiler writing with LLVM is much easier than with
-other system. As proof, the author of Stacker wrote the entire
+other systems. As proof, the author of Stacker wrote the entire
 compiler (language definition, lexer, parser, code generator, etc.) in
 about four days! That's important to know because it shows 
 how quickly you can get a new
@@ -78,11 +78,11 @@
 language the author ever created using LLVM. The learning curve is 
 included in that four days.
 The language described here, Stacker, is Forth-like. Programs
-are simple collections of word definitions and the only thing definitions
+are simple collections of word definitions, and the only thing definitions
 can do is manipulate a stack or generate I/O.  Stacker is not a "real" 
-programming language; its very simple.  Although it is computationally 
+programming language; it's very simple.  Although it is computationally 
 complete, you wouldn't use it for your next big project. However, 
-the fact that it is complete, its simple, and it doesn't have 
+the fact that it is complete, it's simple, and it doesn't have 
 a C-like syntax make it useful for demonstration purposes. It shows
 that LLVM could be applied to a wide variety of languages.
 The basic notions behind stacker is very simple. There's a stack of 
@@ -96,11 +96,11 @@
 : MAIN hello_world ;

 This has two "definitions" (Stacker manipulates words, not
 functions and words have definitions): MAIN and 
-hello_world. The MAIN definition is standard, it
+hello_world. The MAIN definition is standard; it
 tells Stacker where to start. Here, MAIN is defined to 
 simply invoke the word hello_world. The
 hello_world definition tells stacker to push the 
-"Hello, World!" string onto the stack, print it out 
+"Hello, World!" string on to the stack, print it out 
 (>s), pop it off the stack (DROP), and
 finally print a carriage return (CR). Although 
 hello_world uses the stack, its net effect is null. Well
@@ -124,7 +124,7 @@
 
Although I knew that LLVM uses a Single Static Assignment (SSA) format, 
 it wasn't obvious to me how prevalent this idea was in LLVM until I really
 started using it.  Reading the 
-Programmer's Manual and Language Reference
+Programmer's Manual and Language Reference,
 I noted that most of the important LLVM IR (Intermediate Representation) C++ 
 classes were derived from the Value class. The full power of that simple
 design only became fully understood once I started constructing executable
@@ -200,7 +200,7 @@
 

     Create your blocks early. While writing your compiler, you 
     will encounter several situations where you know apriori that you will
-    need several blocks. For example, if-then-else, switch, while and for
+    need several blocks. For example, if-then-else, switch, while, and for
     statements in C/C++ all need multiple blocks for expression in LVVM. 
     The rule is, create them early.
     Terminate your blocks early. This just reduces the chances 
@@ -261,15 +261,15 @@
 the instructions for the "then" and "else" parts. They would use the third part
 of the idiom almost exclusively (inserting new instructions before the 
 terminator). Furthermore, they could even recurse back to handle_if 
-should they encounter another if/then/else statement and it will just work.
+should they encounter another if/then/else statement, and it will just work.
 Note how cleanly this all works out. In particular, the push_back methods on
 the BasicBlock's instruction list. These are lists of type 
 Instruction which also happen to be Values. To create 
-the "if" branch we merely instantiate a BranchInst that takes as 
+the "if" branch, we merely instantiate a BranchInst that takes as 
 arguments the blocks to branch to and the condition to branch on. The blocks
 act like branch labels! This new BranchInst terminates
 the BasicBlock provided as an argument. To give the caller a way
-to keep inserting after calling handle_if we create an "exit" block
+to keep inserting after calling handle_if, we create an "exit" block
 which is returned to the caller.  Note that the "exit" block is used as the 
 terminator for both the "then" and the "else" blocks. This guarantees that no
 matter what else "handle_if" or "fill_in" does, they end up at the "exit" block.
@@ -283,7 +283,7 @@
 method on the various lists. This is so common that it is worth mentioning.
 The "push_back" inserts a value into an STL list, vector, array, etc. at the
 end. The method might have also been named "insert_tail" or "append".
-Althought I've used STL quite frequently, my use of push_back wasn't very
+Although I've used STL quite frequently, my use of push_back wasn't very
 high in other programs. In LLVM, you'll use it all the time.
 
 
@@ -292,17 +292,17 @@
 
 
 It took a little getting used to and several rounds of postings to the LLVM
-mail list to wrap my head around this instruction correctly. Even though I had
+mailing list to wrap my head around this instruction correctly. Even though I had
 read the Language Reference and Programmer's Manual a couple times each, I still
 missed a few very key points:
 
 
-    GetElementPtrInst gives you back a Value for the last thing indexed
+    
GetElementPtrInst gives you back a Value for the last thing indexed.
     
All global variables in LLVM  are pointers.
     
Pointers must also be dereferenced with the GetElementPtrInst instruction.
 
 This means that when you look up an element in the global variable (assuming
-its a struct or array), you must deference the pointer first! For many
+it's a struct or array), you must deference the pointer first! For many
 things, this leads to the idiom:
 
 
@@ -319,13 +319,13 @@
 variable and the address of its first element as the same. That tripped me up
 for a while until I realized that they really do differ .. by type.
 Remember that LLVM is a strongly typed language itself. Everything
-has a type.  The "type" of the global variable is [24 x int]*. That is, its
+has a type.  The "type" of the global variable is [24 x int]*. That is, it's
 a pointer to an array of 24 ints.  When you dereference that global variable with
 a single (0) index, you now have a "[24 x int]" type.  Although
 the pointer value of the dereferenced global and the address of the zero'th element
 in the array will be the same, they differ in their type. The zero'th element has
 type "int" while the pointer value has type "[24 x int]".
-Get this one aspect of LLVM right in your head and you'll save yourself
+
Get this one aspect of LLVM right in your head, and you'll save yourself
 a lot of compiler writing headaches down the road.
 

 
@@ -334,7 +334,7 @@
 Linkage types in LLVM can be a little confusing, especially if your compiler
 writing mind has affixed very hard concepts to particular words like "weak",
 "external", "global", "linkonce", etc. LLVM does not use the precise
-definitions of say ELF or GCC even though they share common terms. To be fair,
+definitions of, say, ELF or GCC, even though they share common terms. To be fair,
 the concepts are related and similar but not precisely the same. This can lead
 you to think you know what a linkage type represents but in fact it is slightly
 different. I recommend you read the 
@@ -342,10 +342,10 @@
 carefully. Then, read it again.

 
Here are some handy tips that I discovered along the way:
 
-    Unitialized means external. That is, the symbol is declared in the current
+    
Uninitialized means external. That is, the symbol is declared in the current
     module and can be used by that module but it is not defined by that module.
     Setting an initializer changes a global's linkage type from whatever it was
-    to a normal, defind global (not external). You'll need to call the setLinkage()
+    to a normal, defined global (not external). You'll need to call the setLinkage()
     method to reset it if you specify the initializer after the GlobalValue has been
     constructed. This is important for LinkOnce and Weak linkage types. 
     Appending linkage can be used to keep track of compilation information at
@@ -362,7 +362,7 @@
 functions in the LLVM IR that make things easier. Here's what I learned: 
 
  Constants are Values like anything else and can be operands of instructions
- Integer constants, frequently needed can be created using the static "get"
+ 
Integer constants, frequently needed, can be created using the static "get"
  methods of the ConstantInt, ConstantSInt, and ConstantUInt classes. The nice thing
  about these is that you can "get" any kind of integer quickly.
  There's a special method on Constant class which allows you to get the null 
@@ -379,14 +379,14 @@
 proceeding, a few words about the stack are in order. The stack is simply
 a global array of 32-bit integers or pointers. A global index keeps track
 of the location of the top of the stack. All of this is hidden from the 
-programmer but it needs to be noted because it is the foundation of the 
+programmer, but it needs to be noted because it is the foundation of the 
 conceptual programming model for Stacker. When you write a definition,
 you are, essentially, saying how you want that definition to manipulate
 the global stack.
 Manipulating the stack can be quite hazardous. There is no distinction
 given and no checking for the various types of values that can be placed
 on the stack. Automatic coercion between types is performed. In many 
-cases this is useful. For example, a boolean value placed on the stack
+cases, this is useful. For example, a boolean value placed on the stack
 can be interpreted as an integer with good results. However, using a
 word that interprets that boolean value as a pointer to a string to
 print out will almost always yield a crash. Stacker simply leaves it
@@ -406,9 +406,9 @@
 
So, your typical definition will have the form:
 : name ... ;
 The name is up to you but it must start with a letter and contain
-only letters numbers and underscore. Names are case sensitive and must not be
+only letters, numbers, and underscore. Names are case sensitive and must not be
 the same as the name of a built-in word. The ... is replaced by
-the stack manipulting words that you wish define name as. 

+the stack manipulating words that you wish to define name as. 

 
 
 
Comments
@@ -429,12 +429,12 @@
 
 Literals
 
-    There are three kinds of literal values in Stacker. Integer, Strings,
+    
There are three kinds of literal values in Stacker: Integers, Strings,
     and Booleans. In each case, the stack operation is to simply push the
-    value onto the stack. So, for example:

+    value on to the stack. So, for example:

      42 " is the answer." TRUE 

-    will push three values onto the stack: the integer 42, the
-    string " is the answer." and the boolean TRUE.
+    will push three values on to the stack: the integer 42, the
+    string " is the answer.", and the boolean TRUE.
 
 
 Words
@@ -464,20 +464,20 @@
 The built-in words of the Stacker language are put in several groups 
 depending on what they do. The groups are as follows:
  
-    LogicalThese words provide the logical operations for
+    
Logical: These words provide the logical operations for
     comparing stack operands.
The words are: < > <= >= 
     = <> true false.
-    BitwiseThese words perform bitwise computations on 
+    
Bitwise: These words perform bitwise computations on 
     their operands. 
 The words are: << >> XOR AND NOT
-    ArithmeticThese words perform arithmetic computations on
+    
Arithmetic: These words perform arithmetic computations on
     their operands. 
 The words are: ABS NEG + - * / MOD */ ++ -- MIN MAX
-    StackThese words manipulate the stack directly by moving
+    
Stack: These words manipulate the stack directly by moving
     its elements around.
 The words are: DROP DUP SWAP OVER ROT DUP2 DROP2 PICK TUCK
-    MemoryThese words allocate, free and manipulate memory
+    
Memory: These words allocate, free, and manipulate memory
     areas outside the stack.
The words are: MALLOC FREE GET PUT
-    ControlThese words alter the normal left to right flow
+    
Control: These words alter the normal left to right flow
     of execution.
The words are: IF ELSE ENDIF WHILE END RETURN EXIT RECURSE
-    I/O These words perform output on the standard output
+    
I/O: These words perform output on the standard output
     and input on the standard input. No other I/O is possible in Stacker.
     
The words are: SPACE TAB CR >s >d >c <s <d <c.
 
@@ -554,12 +554,12 @@
 

-    

+    

-    

+    

@@ -604,75 +604,75 @@
     

+    on to the stack
 

+    on to the stack
 

+    on to the stack
 

+    on to the stack
 

+    of w1 by w2 is pushed back on to the stack
 

+    divided by w3. The result is pushed back on to the stack.
 

+    pushed back on to the stack.
 

+    pushed back on to the stack.
 

+    on to the stack.
 

+	on to the stack.
 

@@ -703,13 +703,13 @@
 

-    

+    

-    

@@ -717,7 +717,7 @@
     

@@ -725,27 +725,27 @@
     

-    

@@ -822,7 +822,7 @@
     

+	block is pushed on to the stack.
 

@@ -911,7 +911,7 @@
     

+	resulting string is pushed on to the stack.
 

+	resulting value is pushed on to the stack
 

+	" %c". The value is converted to an integer and pushed on to the stack.
 

@@ -989,9 +989,9 @@
 The following fully documented program highlights many features of both
 the Stacker language and what is possible with LLVM. The program has two modes
 of operations. If you provide numeric arguments to the program, it checks to see
-if those arguments are prime numbers, prints out the results. Without any 
-aruments, the program prints out any prime numbers it finds between 1 and one 
-million (there's a log of them!). The source code comments below tell the 
+if those arguments are prime numbers and prints out the results. Without any 
+arguments, the program prints out any prime numbers it finds between 1 and one 
+million (there's a lot of them!). The source code comments below tell the 
 remainder of the story.
 
 
@@ -1015,7 +1015,7 @@
 : exit_loop FALSE;
     
 ################################################################################
-# This definition tryies an actual division of a candidate prime number. It
+# This definition tries an actual division of a candidate prime number. It
 # determines whether the division loop on this candidate should continue or
 # not.
 # STACK<:
@@ -1075,7 +1075,7 @@
 # STACK<:
 #   p - the prime number to check
 # STACK>:
-#   yn - boolean indiating if its a prime or not
+#   yn - boolean indicating if its a prime or not
 #   p - the prime number checked
 ################################################################################
 : try_harder
@@ -1248,7 +1248,7 @@
 under the LLVM "projects" directory. You will need to obtain the LLVM sources
 to find it (either via anonymous CVS or a tarball. See the 
 Getting Started document).
-Under the "projects" directory there is a directory named "stacker". That
+
Under the "projects" directory there is a directory named "Stacker". That
 directory contains everything, as follows:
 
     lib - contains most of the source code
@@ -1301,7 +1301,7 @@
 definitions, the ROLL word is not implemented. This word was left out of 
 Stacker on purpose so that it can be an exercise for the student.  The exercise 
 is to implement the ROLL functionality (in your own workspace) and build a test 
-program for it.  If you can implement ROLL you understand Stacker and probably 
+program for it.  If you can implement ROLL, you understand Stacker and probably 
 a fair amount about LLVM since this is one of the more complicated Stacker 
 operations. The work will almost be completely limited to the 
 compiler.  
@@ -1326,7 +1326,7 @@
     emitted currently is somewhat wasteful. It gets cleaned up a lot by existing
     passes but more could be done.
     Add -O -O1 -O2 and -O3 optimization switches to the compiler driver to
-    allow LLVM optimization without using "opt"
+    allow LLVM optimization without using "opt."
     Make the compiler driver use the LLVM linking facilities (with IPO) before 
     depending on GCC to do the final link.
     Clean up parsing. It doesn't handle errors very well.
@@ -1346,6 +1346,6 @@
 
 Reid Spencer
 The LLVM Compiler Infrastructure 
-
Last modified: $Date: 2003/12/01 05:12:09 $ 
+
Last modified: $Date: 2003/12/16 17:31:42 $ 
 
 




From criswell at cs.uiuc.edu  Tue Dec 16 11:36:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 11:36:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/llvm-1.1.tar.gz
Message-ID: <200312161735.LAA28131@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

llvm-1.1.tar.gz updated: 1.3 -> 1.4

---
Log message:

Minor documentation and configure script updates.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/llvm-1.1.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 11:42:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 11:42:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.sparc-sun-solaris2.8.tar.gz
Message-ID: <200312161741.LAA28255@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.sparc-sun-solaris2.8.tar.gz updated: 1.2 -> 1.3

---
Log message:

Removed Solaris header files again.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.sparc-sun-solaris2.8.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 14:19:03 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 14:19:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz cfrontend-1.1.i386-unknown-freebsd5.1.tar.gz
Message-ID: <200312162018.OAA21808@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz updated: 1.1 -> 1.2
cfrontend-1.1.i386-unknown-freebsd5.1.tar.gz updated: 1.1 -> 1.2

---
Log message:

Updated to the newer and faster GCC frontend.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz


Index: llvm-www/releases/1.1/cfrontend-1.1.i386-unknown-freebsd5.1.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 14:37:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 14:37:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz
Message-ID: <200312162036.OAA21953@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz updated: 1.2 -> 1.3

---
Log message:

Updated README with information on how to fix the header files.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz




From criswell at cs.uiuc.edu  Tue Dec 16 15:05:04 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:05:04 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312162104.PAA22107@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.5 -> 1.6

---
Log message:

Added the ability to set a cookie.



---
Diffs of the changes:  (+10 -2)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.5 llvm-www/releases/register.cgi:1.6
--- llvm-www/releases/register.cgi:1.5	Mon Dec 15 13:15:11 2003
+++ llvm-www/releases/register.cgi	Tue Dec 16 15:04:27 2003
@@ -5,6 +5,7 @@
 import smtplib
 import os
 import sys
+import Cookie
 
 # List of email addresses that want to know when people download
 notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
@@ -159,7 +160,7 @@
 	print ('Status: 400 Bad Request')
 	print ('')
 	print ('')
-	print ('Error in form:' + error)
+	print ('Error in form: ' + error)
 	print ('')
 	sys.exit (0)
 
@@ -174,7 +175,14 @@
 Subscribe (form)
 
 #
-# Everything so far has worked.  Send the user to the download page.
+# Create a cookie so that the user does not need to register again.
+#
+usercookie = Cookie.SimpleCookie()
+usercookie['LLVM Download'] = 'Yes'
+usercookie.output()
+
+#
+# Send the user to the download page.
 #
 print ('Location: /releases/download.html')
 print ('')




From criswell at cs.uiuc.edu  Tue Dec 16 15:07:06 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:07:06 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312162106.PAA22218@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.6 -> 1.7

---
Log message:

Remove new functionality until it is tested.



---
Diffs of the changes:  (+3 -3)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.6 llvm-www/releases/register.cgi:1.7
--- llvm-www/releases/register.cgi:1.6	Tue Dec 16 15:04:27 2003
+++ llvm-www/releases/register.cgi	Tue Dec 16 15:06:04 2003
@@ -177,9 +177,9 @@
 #
 # Create a cookie so that the user does not need to register again.
 #
-usercookie = Cookie.SimpleCookie()
-usercookie['LLVM Download'] = 'Yes'
-usercookie.output()
+#usercookie = Cookie.SimpleCookie()
+#usercookie['LLVM Download'] = 'Yes'
+#usercookie.output()
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Tue Dec 16 15:07:13 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:07:13 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi testregister.html
Message-ID: <200312162107.PAA22333@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi added (r1.1)
testregister.html added (r1.1)

---
Log message:

Test files for testing new functionality without breaking the old.



---
Diffs of the changes:  (+276 -0)

Index: llvm-www/releases/testregister.cgi
diff -c /dev/null llvm-www/releases/testregister.cgi:1.1
*** /dev/null	Tue Dec 16 15:07:02 2003
--- llvm-www/releases/testregister.cgi	Tue Dec 16 15:06:52 2003
***************
*** 0 ****
--- 1,191 ----
+ #!/usr/dcs/software/supported/bin/python
+ 
+ import cgi
+ import urllib
+ import smtplib
+ import os
+ import sys
+ import Cookie
+ 
+ # List of email addresses that want to know when people download
+ notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
+ 
+ #
+ # Function: Subscribe()
+ #
+ # Description:
+ #	This function subscribes the specified user to the LLVM announce mailing
+ #	list.
+ #
+ def Subscribe (form):
+ 	# Name of the LLVM subscription CGI script
+ 	scriptURL='http://mail.cs.uiuc.edu/mailman/subscribe/llvm-announce'
+ 
+ 	#
+ 	# Extract from the form any information that we need
+ 	#
+ 	email     = form.getfirst ('email', '')
+ 	pw1       = email
+ 	pw2       = email
+ 	announce  = form.getfirst ('announce','yes')
+ 
+ 	#
+ 	# Exit now if we do not need to subscribe.
+ 	#
+ 	if (announce != 'yes'):
+ 		return
+ 
+ 	#
+ 	# Create a POST request with all of the information that the CGI
+ 	# script will require.
+ 	#
+ 	options = {'email': email, 'pw': pw1, 'pw-conf': pw2, 'digest': '0'}
+ 
+ 	#
+ 	# Convert the POST options into a string that we can send to the
+ 	# subscription CGI script.
+ 	#
+ 	postdata = urllib.urlencode (options)
+ 
+ 	#
+ 	# Subscribe the user.
+ 	#
+ 	urllib.urlopen (scriptURL, postdata)
+ 
+ 	return
+ 
+ #
+ # Function: ValidateForm ()
+ #
+ # Description:
+ #	Make sure that the input to the CGI script is valid.
+ #
+ def ValidateForm (form):
+ 	#
+ 	# Verify if the required fields have been supplied.
+ 	#
+ 	firstname = form.getfirst ('firstname', '')
+ 	lastname  = form.getfirst ('lastname', '')
+ 	email     = form.getfirst ('email', '')
+ 	announce  = form.getfirst ('announce','yes')
+ 
+ 	#
+ 	# Verify that the name and email fields have been filled in.
+ 	#
+ 	if (firstname == ''):
+ 		return 'First name is empty.'
+ 
+ 	if (lastname == ''):
+ 		return 'Last name is empty.'
+ 
+ 	if (email == ''):
+ 		return 'Email address is empty.'
+ 
+ 	#
+ 	# Verify that the email address has an at sign and some periods.
+ 	#
+ 	length = len (email) - 1
+ 	while (length != 0):
+ 		if (email[length] == '@'):
+ 			break
+ 		length=length - 1
+ 	else:
+ 		return 'Email address has no at sign'
+ 
+ 	length = len (email) - 1
+ 	while (length != 0):
+ 		if (email[length] == '.'):
+ 			break
+ 		length=length - 1
+ 	else:
+ 		return 'Email address has no periods'
+ 
+ 	return ''
+ 
+ #
+ # Function: LogForm ()
+ #
+ # Description:
+ #	Append a log record to the logfile that another users has registered to
+ #	download LLVM.
+ #
+ def LogForm (form):
+ 	#
+ 	# Extract the information from the form that we want.
+ 	#
+ 	firstname = form.getfirst ('firstname', '')
+ 	lastname = form.getfirst ('lastname', '')
+ 	title = form.getfirst ('title','')
+ 	organization = form.getfirst ('organization', '')
+ 	email = form.getfirst ('email','')
+ 	plans  = form.getfirst ('plans','')
+ 	if (plans == ''):
+ 		plans = 'No plans.'
+ 
+ 	#
+ 	# Construct an email message describing the user who is downloading
+ 	# LLVM.
+ 	#
+ 	msg = 'Subject: [LLVM Download]\r\n\r\n'
+ 	msg = msg + 'Name: ' + firstname + ' ' + lastname + '\n'
+ 	msg = msg + 'Email: ' + email + '\n'
+ 	msg = msg + 'Title: ' + title + '\n'
+ 	msg = msg + 'Organization: ' + organization + '\n'
+ 	msg = msg + 'Plans with LLVM:\n' + plans + '\n'
+ 
+ 	#
+ 	# Send email to notify that someone has downloaded LLVM yet again!
+ 	#
+ 	mailer = smtplib.SMTP ('localhost')
+ 	for receiver in notifylist:
+ 		header = 'From: ' + email + '\r\nTo: ' + receiver + '\r\n'
+ 		try:
+ 			mailer.sendmail (email, receiver, msg)
+ 		except:
+ 			pass
+ 	mailer.quit ()
+ 	return
+ 
+ #
+ # Parse the CGI input
+ #
+ form = cgi.FieldStorage ()
+ 
+ #
+ # Verify that everything in the form is correct.
+ #
+ error = ValidateForm (form)
+ if (error != ''):
+ 	print ('Content-type: text/html')
+ 	print ('Status: 400 Bad Request')
+ 	print ('')
+ 	print ('')
+ 	print ('Error in form: ' + error)
+ 	print ('')
+ 	sys.exit (0)
+ 
+ #
+ # Log the information provided by the form.
+ #
+ LogForm (form)
+ 
+ #
+ # Subscribe the user the LLVM Announcements list (if they so desire)
+ #
+ Subscribe (form)
+ 
+ #
+ # Create a cookie so that the user does not need to register again.
+ #
+ usercookie = Cookie.SimpleCookie()
+ usercookie['LLVM Download'] = 'Yes'
+ usercookie.output()
+ 
+ #
+ # Send the user to the download page.
+ #
+ print ('Location: /releases/download.html')
+ print ('')
+ 
+ sys.exit (0)
+ 


Index: llvm-www/releases/testregister.html
diff -c /dev/null llvm-www/releases/testregister.html:1.1
*** /dev/null	Tue Dec 16 15:07:02 2003
--- llvm-www/releases/testregister.html	Tue Dec 16 15:06:52 2003
***************
*** 0 ****
--- 1,85 ----
+ 
+ 
+   LLVM Registration Page
+   
+ 
+ 
+ 
+ 
+   LLVM Registration Page
+ 
+ 
+ 
+ 
+ 
+ Welcome to the LLVM registration page!
+ 


+ We'd like to collect some information about people and organizations who are
+ interested in LLVM.  Please complete the following form below to download a copy
+ of the LLVM software.
+ 
+ 
+ 

+ 

+ 

+ 
+ 
+ 
+ 
+ 
+ 
+ 
+   
+   
+ 
+ 
+   
+   
+ 
+ 
+   
+   
+ 
+ 
+   
+   
+ 
+ 
+   
+   
+ 
+ First Name (required): 
Last Name (required): 
Email Address (required): 
Organizational Affiliation: 
Position/Title: 
+ 
+ 
+ The LLVM Announcements List is a low volume mailing list that provides LLVM
+ users with important information on new LLVM releases.  If you would like to
+ subscribe to this list, please select "yes" and respond to the confirmation
+ email that you will receive (your email address will be your password):
+ 
+ 
+ Subscribe to 
+ LLVM
+ Announcements List (default is yes):
+ 

+  Yes

+  No

+ 
+ 
+ We would be interested to know how you plan to use LLVM.
+ 

+ This information is strictly optional and will be kept completely confidential:

+ 
+ 
+ 
+ Finally, click "Download" to go to the download page:

+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 




From criswell at cs.uiuc.edu  Tue Dec 16 15:10:05 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:10:05 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312162109.PAA22445@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.1 -> 1.2

---
Log message:

Don't bother logging when we're just testing.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.1 llvm-www/releases/testregister.cgi:1.2
--- llvm-www/releases/testregister.cgi:1.1	Tue Dec 16 15:06:52 2003
+++ llvm-www/releases/testregister.cgi	Tue Dec 16 15:08:56 2003
@@ -167,7 +167,7 @@
 #
 # Log the information provided by the form.
 #
-LogForm (form)
+#LogForm (form)
 
 #
 # Subscribe the user the LLVM Announcements list (if they so desire)




From criswell at cs.uiuc.edu  Tue Dec 16 15:12:05 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:12:05 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312162112.PAA22567@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.2 -> 1.3

---
Log message:

Working cookie support that doesn't do anything.



---
Diffs of the changes:  (+6 -5)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.2 llvm-www/releases/testregister.cgi:1.3
--- llvm-www/releases/testregister.cgi:1.2	Tue Dec 16 15:08:56 2003
+++ llvm-www/releases/testregister.cgi	Tue Dec 16 15:11:53 2003
@@ -154,7 +154,8 @@
 #
 # Verify that everything in the form is correct.
 #
-error = ValidateForm (form)
+#error = ValidateForm (form)
+error = ''
 if (error != ''):
 	print ('Content-type: text/html')
 	print ('Status: 400 Bad Request')
@@ -172,14 +173,14 @@
 #
 # Subscribe the user the LLVM Announcements list (if they so desire)
 #
-Subscribe (form)
+#Subscribe (form)
 
 #
 # Create a cookie so that the user does not need to register again.
 #
-usercookie = Cookie.SimpleCookie()
-usercookie['LLVM Download'] = 'Yes'
-usercookie.output()
+usercookie = Cookie.SmartCookie()
+usercookie["LLVM"] = "Yes"
+print usercookie
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Tue Dec 16 15:40:04 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:40:04 2003
Subject: [llvm-commits] CVS: llvm-www/releases/checkreg.cgi
Message-ID: <200312162139.PAA22753@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

checkreg.cgi added (r1.1)

---
Log message:

Script to check for previous registration.



---
Diffs of the changes:  (+27 -0)

Index: llvm-www/releases/checkreg.cgi
diff -c /dev/null llvm-www/releases/checkreg.cgi:1.1
*** /dev/null	Tue Dec 16 15:39:06 2003
--- llvm-www/releases/checkreg.cgi	Tue Dec 16 15:38:56 2003
***************
*** 0 ****
--- 1,27 ----
+ #!/usr/dcs/software/supported/bin/python
+ 
+ import cgi
+ import urllib
+ import smtplib
+ import os
+ import sys
+ import Cookie
+ 
+ #
+ # Create a cookie so that the user does not need to register again.
+ #
+ usercookie = Cookie.SmartCookie()
+ usercookie.load (environ['HTTP_COOKIE'])
+ 
+ #
+ # Send the user to the download page.
+ #
+ if (usercookie['LLVM'] == 'Yes'):
+   print ('Location: /releases/download.html')
+ else:
+   print ('Location: /releases/register.html')
+ 
+ print ('')
+ 
+ sys.exit (0)
+ 




From criswell at cs.uiuc.edu  Tue Dec 16 15:42:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:42:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/checkreg.cgi
Message-ID: <200312162141.PAA22907@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

checkreg.cgi updated: 1.1 -> 1.2

---
Log message:

Fixed some name resolution problems.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/checkreg.cgi
diff -u llvm-www/releases/checkreg.cgi:1.1 llvm-www/releases/checkreg.cgi:1.2
--- llvm-www/releases/checkreg.cgi:1.1	Tue Dec 16 15:38:56 2003
+++ llvm-www/releases/checkreg.cgi	Tue Dec 16 15:41:08 2003
@@ -11,7 +11,7 @@
 # Create a cookie so that the user does not need to register again.
 #
 usercookie = Cookie.SmartCookie()
-usercookie.load (environ['HTTP_COOKIE'])
+usercookie.load (os.environ['HTTP_COOKIE'])
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Tue Dec 16 15:50:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:50:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/checkreg.cgi
Message-ID: <200312162149.PAA22952@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

checkreg.cgi updated: 1.2 -> 1.3

---
Log message:

Added code to check for the existance of cookie information.



---
Diffs of the changes:  (+7 -2)

Index: llvm-www/releases/checkreg.cgi
diff -u llvm-www/releases/checkreg.cgi:1.2 llvm-www/releases/checkreg.cgi:1.3
--- llvm-www/releases/checkreg.cgi:1.2	Tue Dec 16 15:41:08 2003
+++ llvm-www/releases/checkreg.cgi	Tue Dec 16 15:49:13 2003
@@ -10,8 +10,13 @@
 #
 # Create a cookie so that the user does not need to register again.
 #
-usercookie = Cookie.SmartCookie()
-usercookie.load (os.environ['HTTP_COOKIE'])
+if (os.environ.has_key ('HTTP_COOKIE')):
+  usercookie = Cookie.SmartCookie()
+  usercookie.load (os.environ['HTTP_COOKIE'])
+else:
+  print ('Location: /releases/register.html')
+  print ('')
+  sys.exit (0)
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Tue Dec 16 15:54:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 15:54:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/checkreg.cgi
Message-ID: <200312162153.PAA23156@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

checkreg.cgi updated: 1.3 -> 1.4

---
Log message:

Move the output into the actual stuff that the browser prints.
This should make it a little easier to debug.



---
Diffs of the changes:  (+3 -2)

Index: llvm-www/releases/checkreg.cgi
diff -u llvm-www/releases/checkreg.cgi:1.3 llvm-www/releases/checkreg.cgi:1.4
--- llvm-www/releases/checkreg.cgi:1.3	Tue Dec 16 15:49:13 2003
+++ llvm-www/releases/checkreg.cgi	Tue Dec 16 15:52:53 2003
@@ -1,12 +1,13 @@
 #!/usr/dcs/software/supported/bin/python
 
 import cgi
-import urllib
-import smtplib
 import os
 import sys
 import Cookie
 
+print ('Content-type: text/plain')
+print ('')
+print ('')
 #
 # Create a cookie so that the user does not need to register again.
 #




From gaeke at cs.uiuc.edu  Tue Dec 16 15:56:01 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Tue Dec 16 15:56:01 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/Transforms/IPO.h
Message-ID: <200312162155.PAA18325@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/Transforms:

IPO.h updated: 1.25 -> 1.26

---
Log message:

Fix typo in comment.

---
Diffs of the changes:  (+1 -1)

Index: llvm/include/llvm/Transforms/IPO.h
diff -u llvm/include/llvm/Transforms/IPO.h:1.25 llvm/include/llvm/Transforms/IPO.h:1.26
--- llvm/include/llvm/Transforms/IPO.h:1.25	Tue Nov 11 16:41:31 2003
+++ llvm/include/llvm/Transforms/IPO.h	Tue Dec 16 15:55:45 2003
@@ -104,7 +104,7 @@
 Pass *createDeadArgEliminationPass();
 
 // DeadArgHacking pass - Same as DAE, but delete arguments of external functions
-// as well.  This is definately not safe, and should only be used by bugpoint.
+// as well.  This is definitely not safe, and should only be used by bugpoint.
 Pass *createDeadArgHackingPass();
 
 //===----------------------------------------------------------------------===//




From lattner at cs.uiuc.edu  Tue Dec 16 16:35:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Tue Dec 16 16:35:02 2003
Subject: [llvm-commits] CVS: llvm/docs/FAQ.html
Message-ID: <200312162234.QAA19112@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

FAQ.html updated: 1.15 -> 1.16

---
Log message:

Add a faq entry for the demo page


---
Diffs of the changes:  (+24 -6)

Index: llvm/docs/FAQ.html
diff -u llvm/docs/FAQ.html:1.15 llvm/docs/FAQ.html:1.16
--- llvm/docs/FAQ.html:1.15	Mon Dec  8 00:22:05 2003
+++ llvm/docs/FAQ.html	Tue Dec 16 16:33:55 2003
@@ -70,6 +70,7 @@
   
      What is this __main() call that gets inserted into
          main()?
+     Where did all of my code go??
   
   
 

@@ -406,11 +407,9 @@
   Questions about code generated by the GCC front-end
 
 
-
-
+

 What is this __main() call that gets inserted into main()?
-
-
+
 
 
 
@@ -426,10 +425,29 @@
 llvm/runtime/GCCLibraries/crtend/ directory in the source-base, and is
 linked in automatically when you link the program.
 
-
 
 
+
+
+
+Where did all of my code go??
+
 
+
+
+If you are using the LLVM demo page, you may often wonder what happened to all
+of the code that you typed in.  Remember that the demo script is running the
+code through the LLVM optimizers, so if you code doesn't actually do anything
+useful, it might all be deleted.
+
+
+
+To prevent this, make sure that the code is actually needed.  For example, if
+you are computing some expression, return the value from the function instead of
+leaving it in a local variable.  If you really want to constrain the optimizer,
+you can read from and assign to volatile global variables.
+
+
 
 
 
@@ -438,7 +456,7 @@
 
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/08 06:22:05 $
+  Last modified: $Date: 2003/12/16 22:33:55 $
 
 
 




From brukman at cs.uiuc.edu  Tue Dec 16 16:58:02 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Tue Dec 16 16:58:02 2003
Subject: [llvm-commits] CVS: llvm/runtime/GCCLibraries/libc/memory.c
Message-ID: <200312162257.QAA19398@zion.cs.uiuc.edu>


Changes in directory llvm/runtime/GCCLibraries/libc:

memory.c updated: 1.3 -> 1.4

---
Log message:

Disable __attribute__((weak)) on Mac OS X and other lame platforms.


---
Diffs of the changes:  (+20 -5)

Index: llvm/runtime/GCCLibraries/libc/memory.c
diff -u llvm/runtime/GCCLibraries/libc/memory.c:1.3 llvm/runtime/GCCLibraries/libc/memory.c:1.4
--- llvm/runtime/GCCLibraries/libc/memory.c:1.3	Mon Dec 15 16:32:50 2003
+++ llvm/runtime/GCCLibraries/libc/memory.c	Tue Dec 16 16:57:30 2003
@@ -2,14 +2,29 @@
 // 
 // A lot of this code is ripped gratuitously from glibc and libiberty.
 //
-//===----------------------------------------------------------------------===//
+//===---------------------------------------------------------------------===//
 
 #include 
 
-void *malloc(size_t) __attribute__((weak));
-void free(void *) __attribute__((weak));
-void *memset(void *, int, size_t) __attribute__((weak));
-void *calloc(size_t nelem, size_t elsize) __attribute__((weak));
+// If we're not being compiled with GCC, turn off attributes. Question is how
+// to handle overriding of memory allocation functions in that case.
+#ifndef __GNUC__
+#define __attribute__(X)
+#endif
+    
+// For now, turn off the weak linkage attribute on Mac OS X.
+#if defined(__GNUC__) && defined(__APPLE_CC__)
+#define __ATTRIBUTE_WEAK__
+#elif defined(__GNUC__)
+#define __ATTRIBUTE_WEAK__ __attribute__((weak))
+#else
+#define __ATTRIBUTE_WEAK__
+#endif
+
+void *malloc(size_t) __ATTRIBUTE_WEAK__;
+void free(void *) __ATTRIBUTE_WEAK__;
+void *memset(void *, int, size_t) __ATTRIBUTE_WEAK__;
+void *calloc(size_t nelem, size_t elsize) __ATTRIBUTE_WEAK__;
 
 void *calloc(size_t nelem, size_t elsize) {
   void *Result = malloc(nelem*elsize);




From gaeke at cs.uiuc.edu  Tue Dec 16 17:15:02 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Tue Dec 16 17:15:02 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Applications/siod/slibu.c
Message-ID: <200312162314.RAA12710@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Applications/siod:

slibu.c updated: 1.1 -> 1.2

---
Log message:

Take out reference to putpwent().
Provide a reasonable default definition for PATH_MAX.


---
Diffs of the changes:  (+5 -11)

Index: llvm/test/Programs/MultiSource/Applications/siod/slibu.c
diff -u llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.1 llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.2
--- llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.1	Fri Oct 17 13:48:45 2003
+++ llvm/test/Programs/MultiSource/Applications/siod/slibu.c	Tue Dec 16 17:14:10 2003
@@ -70,6 +70,10 @@
 #include 
 #endif
 
+#ifndef PATH_MAX
+#define PATH_MAX 1024
+#endif
+
 #if defined(hpux)
 #define PATH_MAX MAXPATHLEN
 #endif
@@ -100,7 +104,7 @@
 
 static void init_slibu_version(void)
 {setvar(cintern("*slibu-version*"),
-	cintern("$Id: slibu.c,v 1.1 2003/10/17 18:48:45 gaeke Exp $"),
+	cintern("$Id: slibu.c,v 1.2 2003/12/16 23:14:10 gaeke Exp $"),
 	NIL);}
 
 
@@ -287,15 +291,6 @@
  return(NIL);}
 #endif
 
-LISP lputpwent(LISP alist,LISP file)
-{int iflag = no_interrupt(1);
- int status;
- struct passwd p;
- lencode_pwent(alist,&p);
- status = putpwent(&p,get_c_file(file,NULL));
- no_interrupt(iflag);
- return(NIL);}
-
 LISP laccess_problem(LISP lfname,LISP lacc)
 {char *fname = get_c_string(lfname);
  char *acc = get_c_string(lacc),*p;
@@ -2171,7 +2166,6 @@
 #if defined(__osf__)
  init_subr_1("setpwfile",lsetpwfile);
 #endif
- init_subr_2("putpwent",lputpwent);
  init_subr_2("access-problem?",laccess_problem);
  init_subr_3("utime",lutime);
  init_subr_2("fchmod",lfchmod);




From criswell at cs.uiuc.edu  Tue Dec 16 17:30:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Tue Dec 16 17:30:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz
Message-ID: <200312162329.RAA23399@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz updated: 1.3 -> 1.4

---
Log message:

Renamed util to utils so that the fixheaders script will work.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz




From gaeke at cs.uiuc.edu  Tue Dec 16 18:07:01 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Tue Dec 16 18:07:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Applications/siod/Makefile
Message-ID: <200312170006.SAA32629@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Applications/siod:

Makefile updated: 1.1 -> 1.2

---
Log message:

Make sure 'unix' is always defined


---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/MultiSource/Applications/siod/Makefile
diff -u llvm/test/Programs/MultiSource/Applications/siod/Makefile:1.1 llvm/test/Programs/MultiSource/Applications/siod/Makefile:1.2
--- llvm/test/Programs/MultiSource/Applications/siod/Makefile:1.1	Fri Oct 17 13:48:45 2003
+++ llvm/test/Programs/MultiSource/Applications/siod/Makefile	Tue Dec 16 18:06:07 2003
@@ -1,7 +1,7 @@
 
 LEVEL = ../../../../..
 PROG     = siod
-CPPFLAGS = -D__USE_MISC -D__USE_GNU -D__USE_SVID -D__USE_XOPEN_EXTENDED -D__USE_XOPEN
+CPPFLAGS = -D__USE_MISC -D__USE_GNU -D__USE_SVID -D__USE_XOPEN_EXTENDED -D__USE_XOPEN -Dunix
 LDFLAGS  = -lm -ldl -lcrypt
 
 RUN_OPTIONS = $(BUILD_SRC_DIR)/test.scm




From gaeke at cs.uiuc.edu  Tue Dec 16 18:10:06 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Tue Dec 16 18:10:06 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Applications/siod/slibu.c
Message-ID: <200312170010.SAA00487@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Applications/siod:

slibu.c updated: 1.2 -> 1.3

---
Log message:

Take out lchown()


---
Diffs of the changes:  (+1 -15)

Index: llvm/test/Programs/MultiSource/Applications/siod/slibu.c
diff -u llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.2 llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.3
--- llvm/test/Programs/MultiSource/Applications/siod/slibu.c:1.2	Tue Dec 16 17:14:10 2003
+++ llvm/test/Programs/MultiSource/Applications/siod/slibu.c	Tue Dec 16 18:09:51 2003
@@ -104,7 +104,7 @@
 
 static void init_slibu_version(void)
 {setvar(cintern("*slibu-version*"),
-	cintern("$Id: slibu.c,v 1.2 2003/12/16 23:14:10 gaeke Exp $"),
+	cintern("$Id: slibu.c,v 1.3 2003/12/17 00:09:51 gaeke Exp $"),
 	NIL);}
 
 
@@ -1448,17 +1448,6 @@
 
 #endif
 
-#if defined(unix) && !defined(linux)
-LISP l_lchown(LISP path,LISP uid,LISP gid)
-{long iflag;
- iflag = no_interrupt(1);
- if (lchown(get_c_string(path),get_c_long(uid),get_c_long(gid)))
-   err("lchown",cons(path,llast_c_errmsg(-1)));
- no_interrupt(iflag);
- return(NIL);}
-#endif
-
-
 #ifdef unix
 
 LISP popen_l(LISP name,LISP how)
@@ -2242,9 +2231,6 @@
 #endif
 #ifndef WIN32
  init_subr_3("chown",l_chown);
-#endif
-#if defined(unix) && !defined(linux)
- init_subr_3("lchown",l_lchown);
 #endif
  init_subr_1("http-date",http_date);
 #if defined(__osf__)




From gaeke at cs.uiuc.edu  Tue Dec 16 18:19:01 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Tue Dec 16 18:19:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Bytecode/Reader/ArchiveReader.cpp
Message-ID: <200312170018.SAA01319@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Bytecode/Reader:

ArchiveReader.cpp updated: 1.10 -> 1.11

---
Log message:

Make getObjectType() smarter about ranlibbed Mac OS X archives.

---
Diffs of the changes:  (+2 -0)

Index: llvm/lib/Bytecode/Reader/ArchiveReader.cpp
diff -u llvm/lib/Bytecode/Reader/ArchiveReader.cpp:1.10 llvm/lib/Bytecode/Reader/ArchiveReader.cpp:1.11
--- llvm/lib/Bytecode/Reader/ArchiveReader.cpp:1.10	Wed Dec 10 18:38:04 2003
+++ llvm/lib/Bytecode/Reader/ArchiveReader.cpp	Tue Dec 16 18:18:18 2003
@@ -53,6 +53,8 @@
   // Check for sections with special names...
   if (!memcmp(H->name, "__.SYMDEF       ", 16))
     return ArchiveSymbolTable;
+  if (!memcmp(H->name, "__.SYMDEF SORTED", 16))
+    return ArchiveSymbolTable;
   if (!memcmp(H->name, "//              ", 16))
     return SVR4LongFilename;
 




From criswell at cs.uiuc.edu  Wed Dec 17 12:16:09 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 12:16:09 2003
Subject: [llvm-commits] CVS: llvm/CREDITS.TXT
Message-ID: <200312171815.MAA25447@choi.cs.uiuc.edu>


Changes in directory llvm:

CREDITS.TXT updated: 1.7 -> 1.8

---
Log message:

Added Stacker to Reid Spencer's list of contributions.



---
Diffs of the changes:  (+1 -1)

Index: llvm/CREDITS.TXT
diff -u llvm/CREDITS.TXT:1.7 llvm/CREDITS.TXT:1.8
--- llvm/CREDITS.TXT:1.7	Thu Nov 13 23:37:10 2003
+++ llvm/CREDITS.TXT	Wed Dec 17 12:15:04 2003
@@ -56,7 +56,7 @@
 N: Reid Spencer
 E: rspencer at x10sys.com
 W: http://extprosys.sourceforge.net/
-D: Complete 'llvm' namespacification, bug fixes and improvements
+D: Complete 'llvm' namespacification, Stacker, bug fixes, and improvements
 
 N: Bill Wendling
 E: wendling at isanbard.org




From brukman at cs.uiuc.edu  Wed Dec 17 12:48:02 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 12:48:02 2003
Subject: [llvm-commits] CVS: llvm/docs/AliasAnalysis.html
Message-ID: <200312171846.MAA24568@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

AliasAnalysis.html updated: 1.12 -> 1.13

---
Log message:

Since we are now in the 'llvm' namespace, the mangled C++ name of structs
and hence the links, all change.


---
Diffs of the changes:  (+2 -2)

Index: llvm/docs/AliasAnalysis.html
diff -u llvm/docs/AliasAnalysis.html:1.12 llvm/docs/AliasAnalysis.html:1.13
--- llvm/docs/AliasAnalysis.html:1.12	Fri Nov 21 19:26:21 2003
+++ llvm/docs/AliasAnalysis.html	Wed Dec 17 12:46:49 2003
@@ -199,7 +199,7 @@
 straight-forward.  There are already several implementations that you can use
 for examples, and the following information should help fill in any details.
 For a minimal example, take a look at the no-aa implementation.
+href="/doxygen/structllvm_1_1NoAA.html">no-aa implementation.
 
 
 
@@ -486,7 +486,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/11/22 01:26:21 $
+  Last modified: $Date: 2003/12/17 18:46:49 $
 
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 14:16:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:16:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/cfrontend-1.1.sparc-sun-solaris2.8.tar.gz
Message-ID: <200312172015.OAA25424@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

cfrontend-1.1.sparc-sun-solaris2.8.tar.gz updated: 1.3 -> 1.4

---
Log message:

New fixheaders script that fixes more headers.



---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/cfrontend-1.1.sparc-sun-solaris2.8.tar.gz




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:02 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/External/Makefile
Message-ID: <200312172035.OAA29661@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/External:

Makefile updated: 1.3 -> 1.3.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/External/Makefile
diff -u llvm/test/Programs/External/Makefile:1.3 llvm/test/Programs/External/Makefile:1.3.4.1
--- llvm/test/Programs/External/Makefile:1.3	Tue Aug 19 16:30:28 2003
+++ llvm/test/Programs/External/Makefile	Wed Dec 17 14:35:03 2003
@@ -8,7 +8,7 @@
 #
 # Create the list of directories to compile
 #
-DIRS := $(sort $(filter-out CVS/, $(wildcard */)))
+DIRS := SPEC
 ifndef USE_SPEC
 DIRS := $(filter-out SPEC/, $(DIRS))
 endif




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:06 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:06 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/External/SPEC/Makefile
Message-ID: <200312172035.OAA29687@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/External/SPEC:

Makefile updated: 1.4 -> 1.4.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/External/SPEC/Makefile
diff -u llvm/test/Programs/External/SPEC/Makefile:1.4 llvm/test/Programs/External/SPEC/Makefile:1.4.4.1
--- llvm/test/Programs/External/SPEC/Makefile:1.4	Tue Aug 19 16:30:30 2003
+++ llvm/test/Programs/External/SPEC/Makefile	Wed Dec 17 14:35:06 2003
@@ -1,3 +1,3 @@
 LEVEL = ../../../..
-DIRS := $(sort $(filter-out CVS/, $(wildcard */)))
+DIRS := CFP2000 CINT2000
 include ${LEVEL}/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:10 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:10 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile
Message-ID: <200312172035.OAA29694@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/FreeBench:

Makefile updated: 1.1 -> 1.1.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.1 llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.1.4.1
--- llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.1	Sat Oct 11 16:18:41 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile	Wed Dec 17 14:35:07 2003
@@ -1,6 +1,6 @@
 # MultiSource/Olden Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := distray mason pcompress2 analyzer fourinarow neural pifft
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:14 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:14 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile
Message-ID: <200312172035.OAA29701@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/McCat:

Makefile updated: 1.3 -> 1.3.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.3.4.1
--- llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.3	Fri Sep 12 11:01:59 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile	Wed Dec 17 14:35:08 2003
@@ -1,6 +1,7 @@
 # MultiSource/McCat Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := 01-qbsort 04-bisect 08-main 12-IOtest 17-bintr 03-testtrie 05-eks 09-vor 15-trie 18-imp
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:18 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:18 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile
Message-ID: <200312172035.OAA29708@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/Olden:

Makefile updated: 1.3 -> 1.3.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.3.4.1
--- llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.3	Fri Sep 12 11:02:21 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile	Wed Dec 17 14:35:09 2003
@@ -1,6 +1,7 @@
 # MultiSource/Olden Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := bh em3d mst power tsp bisort health perimeter treeadd voronoi
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Wed Dec 17 14:36:22 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:36:22 2003
Subject: [llvm-commits] [release_11] CVS: llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile
Message-ID: <200312172035.OAA29715@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/Ptrdist:

Makefile updated: 1.3 -> 1.3.4.1

---
Log message:

Hard coded subdirectories to build to ensure that they build.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.3.4.1
--- llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.3	Fri Sep 12 11:02:35 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile	Wed Dec 17 14:35:10 2003
@@ -1,6 +1,7 @@
 # MultiSource/Ptrdist Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := anagram bc ft ks yacr2
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Wed Dec 17 14:38:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 14:38:01 2003
Subject: [llvm-commits] [release_11] CVS: llvm/CREDITS.TXT
Message-ID: <200312172037.OAA29754@choi.cs.uiuc.edu>


Changes in directory llvm:

CREDITS.TXT updated: 1.7 -> 1.7.2.1

---
Log message:

Added Stacker Reid Spencer's contributions.



---
Diffs of the changes:  (+1 -1)

Index: llvm/CREDITS.TXT
diff -u llvm/CREDITS.TXT:1.7 llvm/CREDITS.TXT:1.7.2.1
--- llvm/CREDITS.TXT:1.7	Thu Nov 13 23:37:10 2003
+++ llvm/CREDITS.TXT	Wed Dec 17 14:37:38 2003
@@ -56,7 +56,7 @@
 N: Reid Spencer
 E: rspencer at x10sys.com
 W: http://extprosys.sourceforge.net/
-D: Complete 'llvm' namespacification, bug fixes and improvements
+D: Complete 'llvm' namespacification, Stacker, bug fixes, and improvements
 
 N: Bill Wendling
 E: wendling at isanbard.org




From criswell at cs.uiuc.edu  Wed Dec 17 15:05:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 15:05:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/announcement.txt
Message-ID: <200312172104.PAA21837@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

announcement.txt updated: 1.3 -> 1.4

---
Log message:

Fixed a little punctuation and grammar.



---
Diffs of the changes:  (+2 -2)

Index: llvm-www/releases/1.1/announcement.txt
diff -u llvm-www/releases/1.1/announcement.txt:1.3 llvm-www/releases/1.1/announcement.txt:1.4
--- llvm-www/releases/1.1/announcement.txt:1.3	Sun Dec 14 10:27:54 2003
+++ llvm-www/releases/1.1/announcement.txt	Wed Dec 17 15:04:43 2003
@@ -11,7 +11,7 @@
   and "idle-time" optimization of programs from arbitrary programming
   languages.  LLVM is written in C++ and has been developed over the past 3
   years at the University of Illinois.  It currently supports compilation of
-  C and C++ programs, using front-ends derived from GCC 3.4.  New front-ends
+  C and C++ programs using front-ends derived from GCC 3.4.  New front-ends
   are being written for Java bytecode and CAML.
 
   The LLVM infrastructure is publicly available under a non-restrictive open
@@ -30,7 +30,7 @@
   A full list of new features and bug-fixes are listed in the Release Notes:
     http://llvm.cs.uiuc.edu/releases/1.1/docs/ReleaseNotes.html#whatsnew
 
-  For an easier to read set of changes, please see the status updates, see:
+  For an easier to read set of changes, please see the status updates:
    http://mail.cs.uiuc.edu/pipermail/llvm-announce/2003-November/000003.html
    ***** Second 1.1 status update *****
 




From criswell at cs.uiuc.edu  Wed Dec 17 15:23:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 15:23:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/llvm-1.1.tar.gz
Message-ID: <200312172122.PAA27193@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

llvm-1.1.tar.gz updated: 1.4 -> 1.5

---
Log message:

Fixed some test suite Makefiles to ensure that they always run the tests.
Updated the CREDITS file since we've added Stacker while I was at it.




---
Diffs of the changes:  (+0 -0)

Index: llvm-www/releases/1.1/llvm-1.1.tar.gz




From brukman at cs.uiuc.edu  Wed Dec 17 16:05:02 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 16:05:02 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/SparcFrameInfo.cpp SparcFrameInfo.h SparcInstrInfo.h SparcRegInfo.h SparcTargetMachine.h SparcInstrInfo.cpp SparcInstrSelection.cpp SparcInternals.h SparcRegClassInfo.cpp SparcRegInfo.cpp SparcTargetMachine.cpp SparcV9CodeEmitter.cpp UltraSparcSchedInfo.cpp
Message-ID: <200312172204.QAA25321@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

SparcFrameInfo.cpp added (r1.1)
SparcFrameInfo.h added (r1.1)
SparcInstrInfo.h added (r1.1)
SparcRegInfo.h updated: 1.6 -> 1.7
SparcTargetMachine.h added (r1.1)
SparcInstrInfo.cpp updated: 1.58 -> 1.59
SparcInstrSelection.cpp updated: 1.127 -> 1.128
SparcInternals.h updated: 1.108 -> 1.109
SparcRegClassInfo.cpp updated: 1.32 -> 1.33
SparcRegInfo.cpp updated: 1.114 -> 1.115
SparcTargetMachine.cpp updated: 1.91 -> 1.92
SparcV9CodeEmitter.cpp updated: 1.46 -> 1.47
UltraSparcSchedInfo.cpp updated: 1.8 -> 1.9

---
Log message:

Reorganized the Sparc backend to be more modular -- each different
implementation of a Target{RegInfo, InstrInfo, Machine, etc} now has a separate
header and a separate implementation file.

This means that instead of a massive SparcInternals.h that forces a
recompilation of the whole target whenever a minor detail is changed, you should
only recompile a few files.

Note that SparcInternals.h is still around; its contents should be minimized.


---
Diffs of the changes:  (+845 -773)

Index: llvm/lib/Target/Sparc/SparcFrameInfo.cpp
diff -c /dev/null llvm/lib/Target/Sparc/SparcFrameInfo.cpp:1.1
*** /dev/null	Wed Dec 17 16:04:11 2003
--- llvm/lib/Target/Sparc/SparcFrameInfo.cpp	Wed Dec 17 16:04:00 2003
***************
*** 0 ****
--- 1,65 ----
+ //===-- Sparc.cpp - General implementation file for the Sparc Target ------===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ // 
+ // Interface to stack frame layout info for the UltraSPARC.  Starting offsets
+ // for each area of the stack frame are aligned at a multiple of
+ // getStackFrameSizeAlignment().
+ // 
+ //===----------------------------------------------------------------------===//
+ 
+ #include "llvm/CodeGen/MachineFunction.h"
+ #include "llvm/CodeGen/MachineFunctionInfo.h"
+ #include "llvm/Target/TargetFrameInfo.h"
+ #include "SparcFrameInfo.h"
+ 
+ using namespace llvm;
+ 
+ int
+ SparcFrameInfo::getFirstAutomaticVarOffset(MachineFunction&, bool& pos) const {
+   pos = false;                          // static stack area grows downwards
+   return StaticAreaOffsetFromFP;
+ }
+ 
+ int
+ SparcFrameInfo::getRegSpillAreaOffset(MachineFunction& mcInfo, bool& pos) const 
+ {
+   // ensure no more auto vars are added
+   mcInfo.getInfo()->freezeAutomaticVarsArea();
+   
+   pos = false;                          // static stack area grows downwards
+   unsigned autoVarsSize = mcInfo.getInfo()->getAutomaticVarsSize();
+   return StaticAreaOffsetFromFP - autoVarsSize; 
+ }
+ 
+ int SparcFrameInfo::getTmpAreaOffset(MachineFunction& mcInfo, bool& pos) const {
+   MachineFunctionInfo *MFI = mcInfo.getInfo();
+   MFI->freezeAutomaticVarsArea();     // ensure no more auto vars are added
+   MFI->freezeSpillsArea();            // ensure no more spill slots are added
+   
+   pos = false;                          // static stack area grows downwards
+   unsigned autoVarsSize = MFI->getAutomaticVarsSize();
+   unsigned spillAreaSize = MFI->getRegSpillsSize();
+   int offset = autoVarsSize + spillAreaSize;
+   return StaticAreaOffsetFromFP - offset;
+ }
+ 
+ int
+ SparcFrameInfo::getDynamicAreaOffset(MachineFunction& mcInfo, bool& pos) const {
+   // Dynamic stack area grows downwards starting at top of opt-args area.
+   // The opt-args, required-args, and register-save areas are empty except
+   // during calls and traps, so they are shifted downwards on each
+   // dynamic-size alloca.
+   pos = false;
+   unsigned optArgsSize = mcInfo.getInfo()->getMaxOptionalArgsSize();
+   if (int extra = optArgsSize % getStackFrameSizeAlignment())
+     optArgsSize += (getStackFrameSizeAlignment() - extra);
+   int offset = optArgsSize + FirstOptionalOutgoingArgOffsetFromSP;
+   assert((offset - OFFSET) % getStackFrameSizeAlignment() == 0);
+   return offset;
+ }


Index: llvm/lib/Target/Sparc/SparcFrameInfo.h
diff -c /dev/null llvm/lib/Target/Sparc/SparcFrameInfo.h:1.1
*** /dev/null	Wed Dec 17 16:04:11 2003
--- llvm/lib/Target/Sparc/SparcFrameInfo.h	Wed Dec 17 16:04:00 2003
***************
*** 0 ****
--- 1,174 ----
+ //===-- SparcFrameInfo.h - Define TargetFrameInfo for Sparc -----*- C++ -*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // Interface to stack frame layout info for the UltraSPARC.
+ // Starting offsets for each area of the stack frame are aligned at
+ // a multiple of getStackFrameSizeAlignment().
+ //
+ //----------------------------------------------------------------------------
+ 
+ #ifndef SPARC_FRAMEINFO_H
+ #define SPARC_FRAMEINFO_H
+ 
+ #include "llvm/Target/TargetFrameInfo.h"
+ #include "llvm/Target/TargetMachine.h"
+ #include "llvm/Target/TargetRegInfo.h"
+ 
+ namespace llvm {
+ 
+ class SparcFrameInfo: public TargetFrameInfo {
+   const TargetMachine ⌖
+ public:
+   SparcFrameInfo(const TargetMachine &TM)
+     : TargetFrameInfo(StackGrowsDown, StackFrameSizeAlignment, 0), target(TM) {}
+   
+ public:
+   // These methods provide constant parameters of the frame layout.
+   // 
+   int  getStackFrameSizeAlignment() const { return StackFrameSizeAlignment;}
+   int  getMinStackFrameSize()       const { return MinStackFrameSize; }
+   int  getNumFixedOutgoingArgs()    const { return NumFixedOutgoingArgs; }
+   int  getSizeOfEachArgOnStack()    const { return SizeOfEachArgOnStack; }
+   bool argsOnStackHaveFixedSize()   const { return true; }
+ 
+   // This method adjusts a stack offset to meet alignment rules of target.
+   // The fixed OFFSET (0x7ff) must be subtracted and the result aligned.
+   virtual int  adjustAlignment(int unalignedOffset, bool growUp,
+                                unsigned int align) const {
+     return unalignedOffset + (growUp? +1:-1)*((unalignedOffset-OFFSET) % align);
+   }
+ 
+   // These methods compute offsets using the frame contents for a
+   // particular function.  The frame contents are obtained from the
+   // MachineCodeInfoForMethod object for the given function.
+   // 
+   int getFirstIncomingArgOffset(MachineFunction& mcInfo, bool& growUp) const {
+     growUp = true;                         // arguments area grows upwards
+     return FirstIncomingArgOffsetFromFP;
+   }
+   int getFirstOutgoingArgOffset(MachineFunction& mcInfo, bool& growUp) const {
+     growUp = true;                         // arguments area grows upwards
+     return FirstOutgoingArgOffsetFromSP;
+   }
+   int getFirstOptionalOutgoingArgOffset(MachineFunction& mcInfo,
+                                         bool& growUp) const {
+     growUp = true;                         // arguments area grows upwards
+     return FirstOptionalOutgoingArgOffsetFromSP;
+   }
+   
+   int getFirstAutomaticVarOffset(MachineFunction& mcInfo, bool& growUp) const;
+   int getRegSpillAreaOffset(MachineFunction& mcInfo, bool& growUp) const;
+   int getTmpAreaOffset(MachineFunction& mcInfo, bool& growUp) const;
+   int getDynamicAreaOffset(MachineFunction& mcInfo, bool& growUp) const;
+ 
+   //
+   // These methods specify the base register used for each stack area
+   // (generally FP or SP)
+   // 
+   virtual int getIncomingArgBaseRegNum() const {
+     return (int) target.getRegInfo().getFramePointer();
+   }
+   virtual int getOutgoingArgBaseRegNum() const {
+     return (int) target.getRegInfo().getStackPointer();
+   }
+   virtual int getOptionalOutgoingArgBaseRegNum() const {
+     return (int) target.getRegInfo().getStackPointer();
+   }
+   virtual int getAutomaticVarBaseRegNum() const {
+     return (int) target.getRegInfo().getFramePointer();
+   }
+   virtual int getRegSpillAreaBaseRegNum() const {
+     return (int) target.getRegInfo().getFramePointer();
+   }
+   virtual int getDynamicAreaBaseRegNum() const {
+     return (int) target.getRegInfo().getStackPointer();
+   }
+ 
+   virtual int getIncomingArgOffset(MachineFunction& mcInfo, 
+                                    unsigned argNum) const {
+     assert(argsOnStackHaveFixedSize()); 
+   
+     unsigned relativeOffset = argNum * getSizeOfEachArgOnStack();
+     bool growUp;                          // do args grow up or down
+     int firstArg = getFirstIncomingArgOffset(mcInfo, growUp);
+     return growUp ? firstArg + relativeOffset : firstArg - relativeOffset; 
+   }
+ 
+   virtual int getOutgoingArgOffset(MachineFunction& mcInfo,
+ 				   unsigned argNum) const {
+     assert(argsOnStackHaveFixedSize()); 
+     //assert(((int) argNum - this->getNumFixedOutgoingArgs())
+     //     <= (int) mcInfo.getInfo()->getMaxOptionalNumArgs());
+     
+     unsigned relativeOffset = argNum * getSizeOfEachArgOnStack();
+     bool growUp;                          // do args grow up or down
+     int firstArg = getFirstOutgoingArgOffset(mcInfo, growUp);
+     return growUp ? firstArg + relativeOffset : firstArg - relativeOffset; 
+   }
+   
+ private:
+   /*----------------------------------------------------------------------
+     This diagram shows the stack frame layout used by llc on Sparc V9.
+     Note that only the location of automatic variables, spill area,
+     temporary storage, and dynamically allocated stack area are chosen
+     by us.  The rest conform to the Sparc V9 ABI.
+     All stack addresses are offset by OFFSET = 0x7ff (2047).
+ 
+     Alignment assumptions and other invariants:
+     (1) %sp+OFFSET and %fp+OFFSET are always aligned on 16-byte boundary
+     (2) Variables in automatic, spill, temporary, or dynamic regions
+         are aligned according to their size as in all memory accesses.
+     (3) Everything below the dynamically allocated stack area is only used
+         during a call to another function, so it is never needed when
+         the current function is active.  This is why space can be allocated
+         dynamically by incrementing %sp any time within the function.
+     
+     STACK FRAME LAYOUT:
+ 
+        ...
+        %fp+OFFSET+176      Optional extra incoming arguments# 1..N
+        %fp+OFFSET+168      Incoming argument #6
+        ...                 ...
+        %fp+OFFSET+128      Incoming argument #1
+        ...                 ...
+     ---%fp+OFFSET-0--------Bottom of caller's stack frame--------------------
+        %fp+OFFSET-8        Automatic variables <-- ****TOP OF STACK FRAME****
+                            Spill area
+                            Temporary storage
+        ...
+ 
+        %sp+OFFSET+176+8N   Bottom of dynamically allocated stack area
+        %sp+OFFSET+168+8N   Optional extra outgoing argument# N
+        ...                 ...
+        %sp+OFFSET+176      Optional extra outgoing argument# 1
+        %sp+OFFSET+168      Outgoing argument #6
+        ...                 ...
+        %sp+OFFSET+128      Outgoing argument #1
+        %sp+OFFSET+120      Save area for %i7
+        ...                 ...
+        %sp+OFFSET+0        Save area for %l0 <-- ****BOTTOM OF STACK FRAME****
+ 
+    *----------------------------------------------------------------------*/
+ 
+   // All stack addresses must be offset by 0x7ff (2047) on Sparc V9.
+   static const int OFFSET                                  = (int) 0x7ff;
+   static const int StackFrameSizeAlignment                 =  16;
+   static const int MinStackFrameSize                       = 176;
+   static const int NumFixedOutgoingArgs                    =   6;
+   static const int SizeOfEachArgOnStack                    =   8;
+   static const int FirstIncomingArgOffsetFromFP            = 128 + OFFSET;
+   static const int FirstOptionalIncomingArgOffsetFromFP    = 176 + OFFSET;
+   static const int StaticAreaOffsetFromFP                  =   0 + OFFSET;
+   static const int FirstOutgoingArgOffsetFromSP            = 128 + OFFSET;
+   static const int FirstOptionalOutgoingArgOffsetFromSP    = 176 + OFFSET;
+ };
+ 
+ } // End llvm namespace
+ 
+ #endif


Index: llvm/lib/Target/Sparc/SparcInstrInfo.h
diff -c /dev/null llvm/lib/Target/Sparc/SparcInstrInfo.h:1.1
*** /dev/null	Wed Dec 17 16:04:11 2003
--- llvm/lib/Target/Sparc/SparcInstrInfo.h	Wed Dec 17 16:04:00 2003
***************
*** 0 ****
--- 1,201 ----
+ //===-- SparcInstrInfo.h - Define TargetInstrInfo for Sparc -----*- C++ -*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // This class contains information about individual instructions.
+ // Most information is stored in the SparcMachineInstrDesc array above.
+ // Other information is computed on demand, and most such functions
+ // default to member functions in base class TargetInstrInfo. 
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef SPARC_INSTRINFO_H
+ #define SPARC_INSTRINFO_H
+ 
+ #include "llvm/Target/TargetInstrInfo.h"
+ #include "llvm/CodeGen/MachineInstr.h"
+ #include "SparcInternals.h"
+ 
+ namespace llvm {
+ 
+ struct SparcInstrInfo : public TargetInstrInfo {
+   SparcInstrInfo();
+ 
+   // All immediate constants are in position 1 except the
+   // store instructions and SETxx.
+   // 
+   virtual int getImmedConstantPos(MachineOpCode opCode) const {
+     bool ignore;
+     if (this->maxImmedConstant(opCode, ignore) != 0) {
+       // 1st store opcode
+       assert(! this->isStore((MachineOpCode) V9::STBr - 1));
+       // last store opcode
+       assert(! this->isStore((MachineOpCode) V9::STXFSRi + 1));
+ 
+       if (opCode == V9::SETSW || opCode == V9::SETUW ||
+           opCode == V9::SETX  || opCode == V9::SETHI)
+         return 0;
+       if (opCode >= V9::STBr && opCode <= V9::STXFSRi)
+         return 2;
+       return 1;
+     }
+     else
+       return -1;
+   }
+ 
+   /// createNOPinstr - returns the target's implementation of NOP, which is
+   /// usually a pseudo-instruction, implemented by a degenerate version of
+   /// another instruction, e.g. X86: xchg ax, ax; SparcV9: sethi 0, g0
+   ///
+   MachineInstr* createNOPinstr() const {
+     return BuildMI(V9::SETHI, 2).addZImm(0).addReg(SparcIntRegClass::g0);
+   }
+ 
+   /// isNOPinstr - not having a special NOP opcode, we need to know if a given
+   /// instruction is interpreted as an `official' NOP instr, i.e., there may be
+   /// more than one way to `do nothing' but only one canonical way to slack off.
+   ///
+   bool isNOPinstr(const MachineInstr &MI) const {
+     // Make sure the instruction is EXACTLY `sethi g0, 0'
+     if (MI.getOpcode() == V9::SETHI && MI.getNumOperands() == 2) {
+       const MachineOperand &op0 = MI.getOperand(0), &op1 = MI.getOperand(1);
+       if (op0.isImmediate() && op0.getImmedValue() == 0 &&
+           op1.isMachineRegister() &&
+           op1.getMachineRegNum() == SparcIntRegClass::g0)
+       {
+         return true;
+       }
+     }
+     return false;
+   }
+   
+   virtual bool hasResultInterlock(MachineOpCode opCode) const
+   {
+     // All UltraSPARC instructions have interlocks (note that delay slots
+     // are not considered here).
+     // However, instructions that use the result of an FCMP produce a
+     // 9-cycle stall if they are issued less than 3 cycles after the FCMP.
+     // Force the compiler to insert a software interlock (i.e., gap of
+     // 2 other groups, including NOPs if necessary).
+     return (opCode == V9::FCMPS || opCode == V9::FCMPD || opCode == V9::FCMPQ);
+   }
+ 
+   //-------------------------------------------------------------------------
+   // Queries about representation of LLVM quantities (e.g., constants)
+   //-------------------------------------------------------------------------
+ 
+   virtual bool ConstantMayNotFitInImmedField(const Constant* CV,
+                                              const Instruction* I) const;
+ 
+   //-------------------------------------------------------------------------
+   // Code generation support for creating individual machine instructions
+   //-------------------------------------------------------------------------
+ 
+   // Get certain common op codes for the current target.  This and all the
+   // Create* methods below should be moved to a machine code generation class
+   // 
+   virtual MachineOpCode getNOPOpCode() const { return V9::NOP; }
+ 
+   // Get the value of an integral constant in the form that must
+   // be put into the machine register.  The specified constant is interpreted
+   // as (i.e., converted if necessary to) the specified destination type.  The
+   // result is always returned as an uint64_t, since the representation of
+   // int64_t and uint64_t are identical.  The argument can be any known const.
+   // 
+   // isValidConstant is set to true if a valid constant was found.
+   // 
+   virtual uint64_t ConvertConstantToIntType(const TargetMachine &target,
+                                             const Value *V,
+                                             const Type *destType,
+                                             bool  &isValidConstant) const;
+ 
+   // Create an instruction sequence to put the constant `val' into
+   // the virtual register `dest'.  `val' may be a Constant or a
+   // GlobalValue, viz., the constant address of a global variable or function.
+   // The generated instructions are returned in `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void  CreateCodeToLoadConst(const TargetMachine& target,
+                                       Function* F,
+                                       Value* val,
+                                       Instruction* dest,
+                                       std::vector& mvec,
+                                       MachineCodeForInstruction& mcfi) const;
+ 
+   // Create an instruction sequence to copy an integer value `val'
+   // to a floating point value `dest' by copying to memory and back.
+   // val must be an integral type.  dest must be a Float or Double.
+   // The generated instructions are returned in `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void  CreateCodeToCopyIntToFloat(const TargetMachine& target,
+                                        Function* F,
+                                        Value* val,
+                                        Instruction* dest,
+                                        std::vector& mvec,
+                                        MachineCodeForInstruction& mcfi) const;
+ 
+   // Similarly, create an instruction sequence to copy an FP value
+   // `val' to an integer value `dest' by copying to memory and back.
+   // The generated instructions are returned in `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void  CreateCodeToCopyFloatToInt(const TargetMachine& target,
+                                        Function* F,
+                                        Value* val,
+                                        Instruction* dest,
+                                        std::vector& mvec,
+                                        MachineCodeForInstruction& mcfi) const;
+   
+   // Create instruction(s) to copy src to dest, for arbitrary types
+   // The generated instructions are returned in `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void CreateCopyInstructionsByType(const TargetMachine& target,
+                                        Function* F,
+                                        Value* src,
+                                        Instruction* dest,
+                                        std::vector& mvec,
+                                        MachineCodeForInstruction& mcfi) const;
+ 
+   // Create instruction sequence to produce a sign-extended register value
+   // from an arbitrary sized value (sized in bits, not bytes).
+   // The generated instructions are appended to `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void CreateSignExtensionInstructions(const TargetMachine& target,
+                                        Function* F,
+                                        Value* srcVal,
+                                        Value* destVal,
+                                        unsigned int numLowBits,
+                                        std::vector& mvec,
+                                        MachineCodeForInstruction& mcfi) const;
+ 
+   // Create instruction sequence to produce a zero-extended register value
+   // from an arbitrary sized value (sized in bits, not bytes).
+   // The generated instructions are appended to `mvec'.
+   // Any temp. registers (TmpInstruction) created are recorded in mcfi.
+   // Any stack space required is allocated via mcff.
+   // 
+   virtual void CreateZeroExtensionInstructions(const TargetMachine& target,
+                                        Function* F,
+                                        Value* srcVal,
+                                        Value* destVal,
+                                        unsigned int numLowBits,
+                                        std::vector& mvec,
+                                        MachineCodeForInstruction& mcfi) const;
+ };
+ 
+ } // End llvm namespace
+ 
+ #endif


Index: llvm/lib/Target/Sparc/SparcRegInfo.h
diff -u /dev/null llvm/lib/Target/Sparc/SparcRegInfo.h:1.7
--- /dev/null	Wed Dec 17 16:04:11 2003
+++ llvm/lib/Target/Sparc/SparcRegInfo.h	Wed Dec 17 16:04:00 2003
@@ -0,0 +1,190 @@
+//===-- SparcRegInfo.h - Define TargetRegInfo for Sparc ---------*- C++ -*-===//
+// 
+//                     The LLVM Compiler Infrastructure
+//
+// This file was developed by the LLVM research group and is distributed under
+// the University of Illinois Open Source License. See LICENSE.TXT for details.
+// 
+//===----------------------------------------------------------------------===//
+//
+// This class implements the virtual class TargetRegInfo for Sparc.
+//
+//----------------------------------------------------------------------------
+
+#ifndef SPARC_REGINFO_H
+#define SPARC_REGINFO_H
+
+namespace llvm {
+
+class SparcRegInfo : public TargetRegInfo {
+
+private:
+
+  // Number of registers used for passing int args (usually 6: %o0 - %o5)
+  //
+  unsigned const NumOfIntArgRegs;
+
+  // Number of registers used for passing float args (usually 32: %f0 - %f31)
+  //
+  unsigned const NumOfFloatArgRegs;
+
+  // The following methods are used to color special live ranges (e.g.
+  // function args and return values etc.) with specific hardware registers
+  // as required. See SparcRegInfo.cpp for the implementation.
+  //
+  void suggestReg4RetAddr(MachineInstr *RetMI, 
+			  LiveRangeInfo &LRI) const;
+
+  void suggestReg4CallAddr(MachineInstr *CallMI, LiveRangeInfo &LRI) const;
+  
+  // Helper used by the all the getRegType() functions.
+  int getRegTypeForClassAndType(unsigned regClassID, const Type* type) const;
+
+public:
+  // Type of registers available in Sparc. There can be several reg types
+  // in the same class. For instace, the float reg class has Single/Double
+  // types
+  //
+  enum RegTypes {
+    IntRegType,
+    FPSingleRegType,
+    FPDoubleRegType,
+    IntCCRegType,
+    FloatCCRegType,
+    SpecialRegType
+  };
+
+  // The actual register classes in the Sparc
+  //
+  // **** WARNING: If this enum order is changed, also modify 
+  // getRegisterClassOfValue method below since it assumes this particular 
+  // order for efficiency.
+  // 
+  enum RegClassIDs { 
+    IntRegClassID,                      // Integer
+    FloatRegClassID,                    // Float (both single/double)
+    IntCCRegClassID,                    // Int Condition Code
+    FloatCCRegClassID,                  // Float Condition code
+    SpecialRegClassID                   // Special (unallocated) registers
+  };
+
+  SparcRegInfo(const SparcTargetMachine &tgt);
+
+  // To find the register class used for a specified Type
+  //
+  unsigned getRegClassIDOfType(const Type *type,
+                               bool isCCReg = false) const;
+
+  // To find the register class to which a specified register belongs
+  //
+  unsigned getRegClassIDOfRegType(int regType) const;
+  
+  // getZeroRegNum - returns the register that contains always zero this is the
+  // unified register number
+  //
+  virtual int getZeroRegNum() const;
+
+  // getCallAddressReg - returns the reg used for pushing the address when a
+  // function is called. This can be used for other purposes between calls
+  //
+  unsigned getCallAddressReg() const;
+
+  // Returns the register containing the return address.
+  // It should be made sure that this  register contains the return 
+  // value when a return instruction is reached.
+  //
+  unsigned getReturnAddressReg() const;
+
+  // Number of registers used for passing int args (usually 6: %o0 - %o5)
+  // and float args (usually 32: %f0 - %f31)
+  //
+  unsigned const getNumOfIntArgRegs() const   { return NumOfIntArgRegs; }
+  unsigned const getNumOfFloatArgRegs() const { return NumOfFloatArgRegs; }
+  
+  // Compute which register can be used for an argument, if any
+  // 
+  int regNumForIntArg(bool inCallee, bool isVarArgsCall,
+                      unsigned argNo, unsigned& regClassId) const;
+
+  int regNumForFPArg(unsigned RegType, bool inCallee, bool isVarArgsCall,
+                     unsigned argNo, unsigned& regClassId) const;
+  
+  // The following methods are used to color special live ranges (e.g.
+  // function args and return values etc.) with specific hardware registers
+  // as required. See SparcRegInfo.cpp for the implementation for Sparc.
+  //
+  void suggestRegs4MethodArgs(const Function *Meth, 
+			      LiveRangeInfo& LRI) const;
+
+  void suggestRegs4CallArgs(MachineInstr *CallMI, 
+			    LiveRangeInfo& LRI) const; 
+
+  void suggestReg4RetValue(MachineInstr *RetMI, 
+                           LiveRangeInfo& LRI) const;
+  
+  void colorMethodArgs(const Function *Meth,  LiveRangeInfo& LRI,
+                       std::vector& InstrnsBefore,
+                       std::vector& InstrnsAfter) const;
+
+  // method used for printing a register for debugging purposes
+  //
+  void printReg(const LiveRange *LR) const;
+  
+  // returns the # of bytes of stack space allocated for each register
+  // type. For Sparc, currently we allocate 8 bytes on stack for all 
+  // register types. We can optimize this later if necessary to save stack
+  // space (However, should make sure that stack alignment is correct)
+  //
+  inline int getSpilledRegSize(int RegType) const {
+    return 8;
+  }
+
+
+  // To obtain the return value and the indirect call address (if any)
+  // contained in a CALL machine instruction
+  //
+  const Value * getCallInstRetVal(const MachineInstr *CallMI) const;
+  const Value * getCallInstIndirectAddrVal(const MachineInstr *CallMI) const;
+
+  // The following methods are used to generate "copy" machine instructions
+  // for an architecture.
+  //
+  // The function regTypeNeedsScratchReg() can be used to check whether a
+  // scratch register is needed to copy a register of type `regType' to
+  // or from memory.  If so, such a scratch register can be provided by
+  // the caller (e.g., if it knows which regsiters are free); otherwise
+  // an arbitrary one will be chosen and spilled by the copy instructions.
+  //
+  bool regTypeNeedsScratchReg(int RegType,
+                              int& scratchRegClassId) const;
+
+  void cpReg2RegMI(std::vector& mvec,
+                   unsigned SrcReg, unsigned DestReg,
+                   int RegType) const;
+
+  void cpReg2MemMI(std::vector& mvec,
+                   unsigned SrcReg, unsigned DestPtrReg,
+                   int Offset, int RegType, int scratchReg = -1) const;
+
+  void cpMem2RegMI(std::vector& mvec,
+                   unsigned SrcPtrReg, int Offset, unsigned DestReg,
+                   int RegType, int scratchReg = -1) const;
+
+  void cpValue2Value(Value *Src, Value *Dest,
+                     std::vector& mvec) const;
+
+  // Get the register type for a register identified different ways.
+  // Note that getRegTypeForLR(LR) != getRegTypeForDataType(LR->getType())!
+  // The reg class of a LR depends both on the Value types in it and whether
+  // they are CC registers or not (for example).
+  int getRegTypeForDataType(const Type* type) const;
+  int getRegTypeForLR(const LiveRange *LR) const;
+  int getRegType(int unifiedRegNum) const;
+
+  virtual unsigned getFramePointer() const;
+  virtual unsigned getStackPointer() const;
+};
+
+} // End llvm namespace
+
+#endif


Index: llvm/lib/Target/Sparc/SparcTargetMachine.h
diff -c /dev/null llvm/lib/Target/Sparc/SparcTargetMachine.h:1.1
*** /dev/null	Wed Dec 17 16:04:11 2003
--- llvm/lib/Target/Sparc/SparcTargetMachine.h	Wed Dec 17 16:04:00 2003
***************
*** 0 ****
--- 1,59 ----
+ //===-- SparcTargetMachine.h - Define TargetMachine for Sparc ---*- C++ -*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ // 
+ // This file declares the primary interface to machine description for the
+ // UltraSPARC.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef SPARC_TARGETMACHINE_H
+ #define SPARC_TARGETMACHINE_H
+ 
+ #include "llvm/PassManager.h"
+ #include "llvm/Target/TargetFrameInfo.h"
+ #include "llvm/Target/TargetMachine.h"
+ #include "SparcInstrInfo.h"
+ #include "SparcInternals.h"
+ #include "SparcRegInfo.h"
+ #include "SparcFrameInfo.h"
+ 
+ namespace llvm {
+ 
+ class SparcTargetMachine : public TargetMachine {
+   SparcInstrInfo instrInfo;
+   SparcSchedInfo schedInfo;
+   SparcRegInfo   regInfo;
+   SparcFrameInfo frameInfo;
+   SparcCacheInfo cacheInfo;
+ public:
+   SparcTargetMachine();
+ 
+   virtual const TargetInstrInfo  &getInstrInfo() const { return instrInfo; }
+   virtual const TargetSchedInfo  &getSchedInfo() const { return schedInfo; }
+   virtual const TargetRegInfo    &getRegInfo()   const { return regInfo; }
+   virtual const TargetFrameInfo  &getFrameInfo() const { return frameInfo; }
+   virtual const TargetCacheInfo  &getCacheInfo() const { return cacheInfo; }
+ 
+   virtual bool addPassesToEmitAssembly(PassManager &PM, std::ostream &Out);
+   virtual bool addPassesToJITCompile(FunctionPassManager &PM);
+   virtual bool addPassesToEmitMachineCode(FunctionPassManager &PM,
+                                           MachineCodeEmitter &MCE);
+   virtual void replaceMachineCodeForFunction(void *Old, void *New);
+ 
+   /// getJITStubForFunction - Create or return a stub for the specified
+   /// function.  This stub acts just like the specified function, except that it
+   /// allows the "address" of the function to be taken without having to
+   /// generate code for it.
+   ///
+   ///virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE);
+ };
+ 
+ } // End llvm namespace
+ 
+ #endif


Index: llvm/lib/Target/Sparc/SparcInstrInfo.cpp
diff -u llvm/lib/Target/Sparc/SparcInstrInfo.cpp:1.58 llvm/lib/Target/Sparc/SparcInstrInfo.cpp:1.59
--- llvm/lib/Target/Sparc/SparcInstrInfo.cpp:1.58	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/SparcInstrInfo.cpp	Wed Dec 17 16:04:00 2003
@@ -9,8 +9,6 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "SparcInternals.h"
-#include "SparcInstrSelectionSupport.h"
 #include "llvm/Constants.h"
 #include "llvm/DerivedTypes.h"
 #include "llvm/Function.h"
@@ -22,6 +20,9 @@
 #include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/CodeGen/MachineCodeForInstruction.h"
 #include "llvm/CodeGen/MachineInstrBuilder.h"
+#include "SparcInternals.h"
+#include "SparcInstrSelectionSupport.h"
+#include "SparcInstrInfo.h"
 
 namespace llvm {
 
@@ -41,7 +42,7 @@
 //---------------------------------------------------------------------------
 
 uint64_t
-UltraSparcInstrInfo::ConvertConstantToIntType(const TargetMachine &target,
+SparcInstrInfo::ConvertConstantToIntType(const TargetMachine &target,
                                               const Value *V,
                                               const Type *destType,
                                               bool  &isValidConstant) const
@@ -406,7 +407,7 @@
 
 
 //---------------------------------------------------------------------------
-// class UltraSparcInstrInfo 
+// class SparcInstrInfo 
 // 
 // Purpose:
 //   Information about individual instructions.
@@ -416,7 +417,7 @@
 //---------------------------------------------------------------------------
 
 /*ctor*/
-UltraSparcInstrInfo::UltraSparcInstrInfo()
+SparcInstrInfo::SparcInstrInfo()
   : TargetInstrInfo(SparcMachineInstrDesc,
                     /*descSize = */ V9::NUM_TOTAL_OPCODES,
                     /*numRealOpCodes = */ V9::NUM_REAL_OPCODES)
@@ -425,7 +426,7 @@
 }
 
 bool
-UltraSparcInstrInfo::ConstantMayNotFitInImmedField(const Constant* CV,
+SparcInstrInfo::ConstantMayNotFitInImmedField(const Constant* CV,
                                                    const Instruction* I) const
 {
   if (I->getOpcode() >= MaxConstantsTable.size()) // user-defined op (or bug!)
@@ -455,12 +456,12 @@
 // Any stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateCodeToLoadConst(const TargetMachine& target,
-                                           Function* F,
-                                           Value* val,
-                                           Instruction* dest,
-                                           std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const
+SparcInstrInfo::CreateCodeToLoadConst(const TargetMachine& target,
+                                      Function* F,
+                                      Value* val,
+                                      Instruction* dest,
+                                      std::vector& mvec,
+                                      MachineCodeForInstruction& mcfi) const
 {
   assert(isa(val) || isa(val) &&
          "I only know about constant values and global addresses");
@@ -552,7 +553,7 @@
 // Any stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateCodeToCopyIntToFloat(const TargetMachine& target,
+SparcInstrInfo::CreateCodeToCopyIntToFloat(const TargetMachine& target,
                                         Function* F,
                                         Value* val,
                                         Instruction* dest,
@@ -613,7 +614,7 @@
 // Temporary stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateCodeToCopyFloatToInt(const TargetMachine& target,
+SparcInstrInfo::CreateCodeToCopyFloatToInt(const TargetMachine& target,
                                         Function* F,
                                         Value* val,
                                         Instruction* dest,
@@ -664,11 +665,11 @@
 // Any stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateCopyInstructionsByType(const TargetMachine& target,
-                                                  Function *F,
-                                                  Value* src,
-                                                  Instruction* dest,
-                                               std::vector& mvec,
+SparcInstrInfo::CreateCopyInstructionsByType(const TargetMachine& target,
+                                             Function *F,
+                                             Value* src,
+                                             Instruction* dest,
+                                             std::vector& mvec,
                                           MachineCodeForInstruction& mcfi) const
 {
   bool loadConstantToReg = false;
@@ -760,7 +761,7 @@
 // Any stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateSignExtensionInstructions(
+SparcInstrInfo::CreateSignExtensionInstructions(
                                         const TargetMachine& target,
                                         Function* F,
                                         Value* srcVal,
@@ -782,7 +783,7 @@
 // Any stack space required is allocated via MachineFunction.
 // 
 void
-UltraSparcInstrInfo::CreateZeroExtensionInstructions(
+SparcInstrInfo::CreateZeroExtensionInstructions(
                                         const TargetMachine& target,
                                         Function* F,
                                         Value* srcVal,


Index: llvm/lib/Target/Sparc/SparcInstrSelection.cpp
diff -u llvm/lib/Target/Sparc/SparcInstrSelection.cpp:1.127 llvm/lib/Target/Sparc/SparcInstrSelection.cpp:1.128
--- llvm/lib/Target/Sparc/SparcInstrSelection.cpp:1.127	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/SparcInstrSelection.cpp	Wed Dec 17 16:04:00 2003
@@ -11,9 +11,6 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "SparcInstrSelectionSupport.h"
-#include "SparcInternals.h"
-#include "SparcRegClassInfo.h"
 #include "llvm/Constants.h"
 #include "llvm/ConstantHandling.h"
 #include "llvm/DerivedTypes.h"
@@ -28,6 +25,10 @@
 #include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/CodeGen/MachineInstrBuilder.h"
 #include "llvm/CodeGen/MachineInstrAnnot.h"
+#include "SparcInstrSelectionSupport.h"
+#include "SparcInternals.h"
+#include "SparcRegClassInfo.h"
+#include "SparcRegInfo.h"
 #include "Support/MathExtras.h"
 #include 
 #include 
@@ -1583,8 +1584,8 @@
         //     -- For non-FP values, create an add-with-0 instruction
         // 
         if (retVal != NULL) {
-          const UltraSparcRegInfo& regInfo =
-            (UltraSparcRegInfo&) target.getRegInfo();
+          const SparcRegInfo& regInfo =
+            (SparcRegInfo&) target.getRegInfo();
           const Type* retType = retVal->getType();
           unsigned regClassID = regInfo.getRegClassIDOfType(retType);
           unsigned retRegNum = (retType->isFloatingPoint()
@@ -2488,8 +2489,8 @@
           MachineFunction& MF = MachineFunction::get(currentFunc);
           MachineCodeForInstruction& mcfi =
             MachineCodeForInstruction::get(callInstr); 
-          const UltraSparcRegInfo& regInfo =
-            (UltraSparcRegInfo&) target.getRegInfo();
+          const SparcRegInfo& regInfo =
+            (SparcRegInfo&) target.getRegInfo();
           const TargetFrameInfo& frameInfo = target.getFrameInfo();
 
           // Create hidden virtual register for return address with type void*


Index: llvm/lib/Target/Sparc/SparcInternals.h
diff -u llvm/lib/Target/Sparc/SparcInternals.h:1.108 llvm/lib/Target/Sparc/SparcInternals.h:1.109
--- llvm/lib/Target/Sparc/SparcInternals.h:1.108	Wed Nov 12 18:17:20 2003
+++ llvm/lib/Target/Sparc/SparcInternals.h	Wed Dec 17 16:04:00 2003
@@ -28,7 +28,7 @@
 namespace llvm {
 
 class LiveRange;
-class UltraSparc;
+class SparcTargetMachine;
 class Pass;
 
 enum SparcInstrSchedClass {
@@ -71,571 +71,34 @@
   };
 }
 
-
 // Array of machine instruction descriptions...
 extern const TargetInstrDescriptor SparcMachineInstrDesc[];
 
-
 //---------------------------------------------------------------------------
-// class UltraSparcInstrInfo 
-// 
-// Purpose:
-//   Information about individual instructions.
-//   Most information is stored in the SparcMachineInstrDesc array above.
-//   Other information is computed on demand, and most such functions
-//   default to member functions in base class TargetInstrInfo. 
-//---------------------------------------------------------------------------
-
-struct UltraSparcInstrInfo : public TargetInstrInfo {
-  UltraSparcInstrInfo();
-
-  //
-  // All immediate constants are in position 1 except the
-  // store instructions and SETxx.
-  // 
-  virtual int getImmedConstantPos(MachineOpCode opCode) const {
-    bool ignore;
-    if (this->maxImmedConstant(opCode, ignore) != 0) {
-      // 1st store opcode
-      assert(! this->isStore((MachineOpCode) V9::STBr - 1));
-      // last store opcode
-      assert(! this->isStore((MachineOpCode) V9::STXFSRi + 1));
-
-      if (opCode == V9::SETSW || opCode == V9::SETUW ||
-          opCode == V9::SETX  || opCode == V9::SETHI)
-        return 0;
-      if (opCode >= V9::STBr && opCode <= V9::STXFSRi)
-        return 2;
-      return 1;
-    }
-    else
-      return -1;
-  }
-
-  /// createNOPinstr - returns the target's implementation of NOP, which is
-  /// usually a pseudo-instruction, implemented by a degenerate version of
-  /// another instruction, e.g. X86: xchg ax, ax; SparcV9: sethi 0, g0
-  ///
-  MachineInstr* createNOPinstr() const {
-    return BuildMI(V9::SETHI, 2).addZImm(0).addReg(SparcIntRegClass::g0);
-  }
-
-  /// isNOPinstr - not having a special NOP opcode, we need to know if a given
-  /// instruction is interpreted as an `official' NOP instr, i.e., there may be
-  /// more than one way to `do nothing' but only one canonical way to slack off.
-  ///
-  bool isNOPinstr(const MachineInstr &MI) const {
-    // Make sure the instruction is EXACTLY `sethi g0, 0'
-    if (MI.getOpcode() == V9::SETHI && MI.getNumOperands() == 2) {
-      const MachineOperand &op0 = MI.getOperand(0), &op1 = MI.getOperand(1);
-      if (op0.isImmediate() && op0.getImmedValue() == 0 &&
-          op1.isMachineRegister() &&
-          op1.getMachineRegNum() == SparcIntRegClass::g0)
-      {
-        return true;
-      }
-    }
-    return false;
-  }
-  
-  virtual bool hasResultInterlock(MachineOpCode opCode) const
-  {
-    // All UltraSPARC instructions have interlocks (note that delay slots
-    // are not considered here).
-    // However, instructions that use the result of an FCMP produce a
-    // 9-cycle stall if they are issued less than 3 cycles after the FCMP.
-    // Force the compiler to insert a software interlock (i.e., gap of
-    // 2 other groups, including NOPs if necessary).
-    return (opCode == V9::FCMPS || opCode == V9::FCMPD || opCode == V9::FCMPQ);
-  }
-
-  //-------------------------------------------------------------------------
-  // Queries about representation of LLVM quantities (e.g., constants)
-  //-------------------------------------------------------------------------
-
-  virtual bool ConstantMayNotFitInImmedField(const Constant* CV,
-                                             const Instruction* I) const;
-
-  //-------------------------------------------------------------------------
-  // Code generation support for creating individual machine instructions
-  //-------------------------------------------------------------------------
-
-  // Get certain common op codes for the current target.  This and all the
-  // Create* methods below should be moved to a machine code generation class
-  // 
-  virtual MachineOpCode getNOPOpCode() const { return V9::NOP; }
-
-  // Get the value of an integral constant in the form that must
-  // be put into the machine register.  The specified constant is interpreted
-  // as (i.e., converted if necessary to) the specified destination type.  The
-  // result is always returned as an uint64_t, since the representation of
-  // int64_t and uint64_t are identical.  The argument can be any known const.
-  // 
-  // isValidConstant is set to true if a valid constant was found.
-  // 
-  virtual uint64_t ConvertConstantToIntType(const TargetMachine &target,
-                                            const Value *V,
-                                            const Type *destType,
-                                            bool  &isValidConstant) const;
-
-  // Create an instruction sequence to put the constant `val' into
-  // the virtual register `dest'.  `val' may be a Constant or a
-  // GlobalValue, viz., the constant address of a global variable or function.
-  // The generated instructions are returned in `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void  CreateCodeToLoadConst(const TargetMachine& target,
-                                      Function* F,
-                                      Value* val,
-                                      Instruction* dest,
-                                      std::vector& mvec,
-                                      MachineCodeForInstruction& mcfi) const;
-
-  // Create an instruction sequence to copy an integer value `val'
-  // to a floating point value `dest' by copying to memory and back.
-  // val must be an integral type.  dest must be a Float or Double.
-  // The generated instructions are returned in `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void  CreateCodeToCopyIntToFloat(const TargetMachine& target,
-                                       Function* F,
-                                       Value* val,
-                                       Instruction* dest,
-                                       std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const;
-
-  // Similarly, create an instruction sequence to copy an FP value
-  // `val' to an integer value `dest' by copying to memory and back.
-  // The generated instructions are returned in `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void  CreateCodeToCopyFloatToInt(const TargetMachine& target,
-                                       Function* F,
-                                       Value* val,
-                                       Instruction* dest,
-                                       std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const;
-  
-  // Create instruction(s) to copy src to dest, for arbitrary types
-  // The generated instructions are returned in `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void CreateCopyInstructionsByType(const TargetMachine& target,
-                                       Function* F,
-                                       Value* src,
-                                       Instruction* dest,
-                                       std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const;
-
-  // Create instruction sequence to produce a sign-extended register value
-  // from an arbitrary sized value (sized in bits, not bytes).
-  // The generated instructions are appended to `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void CreateSignExtensionInstructions(const TargetMachine& target,
-                                       Function* F,
-                                       Value* srcVal,
-                                       Value* destVal,
-                                       unsigned int numLowBits,
-                                       std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const;
-
-  // Create instruction sequence to produce a zero-extended register value
-  // from an arbitrary sized value (sized in bits, not bytes).
-  // The generated instructions are appended to `mvec'.
-  // Any temp. registers (TmpInstruction) created are recorded in mcfi.
-  // Any stack space required is allocated via mcff.
-  // 
-  virtual void CreateZeroExtensionInstructions(const TargetMachine& target,
-                                       Function* F,
-                                       Value* srcVal,
-                                       Value* destVal,
-                                       unsigned int numLowBits,
-                                       std::vector& mvec,
-                                       MachineCodeForInstruction& mcfi) const;
-};
-
-
-//----------------------------------------------------------------------------
-// class UltraSparcRegInfo
-//
-// This class implements the virtual class TargetRegInfo for Sparc.
-//
-//----------------------------------------------------------------------------
-
-class UltraSparcRegInfo : public TargetRegInfo {
-
-private:
-
-  // Number of registers used for passing int args (usually 6: %o0 - %o5)
-  //
-  unsigned const NumOfIntArgRegs;
-
-  // Number of registers used for passing float args (usually 32: %f0 - %f31)
-  //
-  unsigned const NumOfFloatArgRegs;
-
-  // ========================  Private Methods =============================
-
-  // The following methods are used to color special live ranges (e.g.
-  // function args and return values etc.) with specific hardware registers
-  // as required. See SparcRegInfo.cpp for the implementation.
-  //
-  void suggestReg4RetAddr(MachineInstr *RetMI, 
-			  LiveRangeInfo &LRI) const;
-
-  void suggestReg4CallAddr(MachineInstr *CallMI, LiveRangeInfo &LRI) const;
-  
-  // Helper used by the all the getRegType() functions.
-  int getRegTypeForClassAndType(unsigned regClassID, const Type* type) const;
-
-public:
-  // Type of registers available in Sparc. There can be several reg types
-  // in the same class. For instace, the float reg class has Single/Double
-  // types
-  //
-  enum RegTypes {
-    IntRegType,
-    FPSingleRegType,
-    FPDoubleRegType,
-    IntCCRegType,
-    FloatCCRegType,
-    SpecialRegType
-  };
-
-  // The actual register classes in the Sparc
-  //
-  // **** WARNING: If this enum order is changed, also modify 
-  // getRegisterClassOfValue method below since it assumes this particular 
-  // order for efficiency.
-  // 
-  enum RegClassIDs { 
-    IntRegClassID,                      // Integer
-    FloatRegClassID,                    // Float (both single/double)
-    IntCCRegClassID,                    // Int Condition Code
-    FloatCCRegClassID,                  // Float Condition code
-    SpecialRegClassID                   // Special (unallocated) registers
-  };
-
-  UltraSparcRegInfo(const UltraSparc &tgt);
-
-  // To find the register class used for a specified Type
-  //
-  unsigned getRegClassIDOfType(const Type *type,
-                               bool isCCReg = false) const;
-
-  // To find the register class to which a specified register belongs
-  //
-  unsigned getRegClassIDOfRegType(int regType) const;
-  
-  // getZeroRegNum - returns the register that contains always zero this is the
-  // unified register number
-  //
-  virtual int getZeroRegNum() const;
-
-  // getCallAddressReg - returns the reg used for pushing the address when a
-  // function is called. This can be used for other purposes between calls
-  //
-  unsigned getCallAddressReg() const;
-
-  // Returns the register containing the return address.
-  // It should be made sure that this  register contains the return 
-  // value when a return instruction is reached.
-  //
-  unsigned getReturnAddressReg() const;
-
-  // Number of registers used for passing int args (usually 6: %o0 - %o5)
-  // and float args (usually 32: %f0 - %f31)
-  //
-  unsigned const getNumOfIntArgRegs() const   { return NumOfIntArgRegs; }
-  unsigned const getNumOfFloatArgRegs() const { return NumOfFloatArgRegs; }
-  
-  // Compute which register can be used for an argument, if any
-  // 
-  int regNumForIntArg(bool inCallee, bool isVarArgsCall,
-                      unsigned argNo, unsigned& regClassId) const;
-
-  int regNumForFPArg(unsigned RegType, bool inCallee, bool isVarArgsCall,
-                     unsigned argNo, unsigned& regClassId) const;
-  
-  // The following methods are used to color special live ranges (e.g.
-  // function args and return values etc.) with specific hardware registers
-  // as required. See SparcRegInfo.cpp for the implementation for Sparc.
-  //
-  void suggestRegs4MethodArgs(const Function *Meth, 
-			      LiveRangeInfo& LRI) const;
-
-  void suggestRegs4CallArgs(MachineInstr *CallMI, 
-			    LiveRangeInfo& LRI) const; 
-
-  void suggestReg4RetValue(MachineInstr *RetMI, 
-                           LiveRangeInfo& LRI) const;
-  
-  void colorMethodArgs(const Function *Meth,  LiveRangeInfo& LRI,
-                       std::vector& InstrnsBefore,
-                       std::vector& InstrnsAfter) const;
-
-  // method used for printing a register for debugging purposes
-  //
-  void printReg(const LiveRange *LR) const;
-  
-  // returns the # of bytes of stack space allocated for each register
-  // type. For Sparc, currently we allocate 8 bytes on stack for all 
-  // register types. We can optimize this later if necessary to save stack
-  // space (However, should make sure that stack alignment is correct)
-  //
-  inline int getSpilledRegSize(int RegType) const {
-    return 8;
-  }
-
-
-  // To obtain the return value and the indirect call address (if any)
-  // contained in a CALL machine instruction
-  //
-  const Value * getCallInstRetVal(const MachineInstr *CallMI) const;
-  const Value * getCallInstIndirectAddrVal(const MachineInstr *CallMI) const;
-
-  // The following methods are used to generate "copy" machine instructions
-  // for an architecture.
-  //
-  // The function regTypeNeedsScratchReg() can be used to check whether a
-  // scratch register is needed to copy a register of type `regType' to
-  // or from memory.  If so, such a scratch register can be provided by
-  // the caller (e.g., if it knows which regsiters are free); otherwise
-  // an arbitrary one will be chosen and spilled by the copy instructions.
-  //
-  bool regTypeNeedsScratchReg(int RegType,
-                              int& scratchRegClassId) const;
-
-  void cpReg2RegMI(std::vector& mvec,
-                   unsigned SrcReg, unsigned DestReg,
-                   int RegType) const;
-
-  void cpReg2MemMI(std::vector& mvec,
-                   unsigned SrcReg, unsigned DestPtrReg,
-                   int Offset, int RegType, int scratchReg = -1) const;
-
-  void cpMem2RegMI(std::vector& mvec,
-                   unsigned SrcPtrReg, int Offset, unsigned DestReg,
-                   int RegType, int scratchReg = -1) const;
-
-  void cpValue2Value(Value *Src, Value *Dest,
-                     std::vector& mvec) const;
-
-  // Get the register type for a register identified different ways.
-  // Note that getRegTypeForLR(LR) != getRegTypeForDataType(LR->getType())!
-  // The reg class of a LR depends both on the Value types in it and whether
-  // they are CC registers or not (for example).
-  int getRegTypeForDataType(const Type* type) const;
-  int getRegTypeForLR(const LiveRange *LR) const;
-  int getRegType(int unifiedRegNum) const;
-
-  virtual unsigned getFramePointer() const;
-  virtual unsigned getStackPointer() const;
-};
-
-
-
-
-//---------------------------------------------------------------------------
-// class UltraSparcSchedInfo
+// class SparcSchedInfo
 // 
 // Purpose:
 //   Interface to instruction scheduling information for UltraSPARC.
 //   The parameter values above are based on UltraSPARC IIi.
 //---------------------------------------------------------------------------
 
-
-class UltraSparcSchedInfo: public TargetSchedInfo {
+class SparcSchedInfo: public TargetSchedInfo {
 public:
-  UltraSparcSchedInfo(const TargetMachine &tgt);
+  SparcSchedInfo(const TargetMachine &tgt);
 protected:
   virtual void initializeResources();
 };
 
-
-//---------------------------------------------------------------------------
-// class UltraSparcFrameInfo 
-// 
-// Purpose:
-//   Interface to stack frame layout info for the UltraSPARC.
-//   Starting offsets for each area of the stack frame are aligned at
-//   a multiple of getStackFrameSizeAlignment().
-//---------------------------------------------------------------------------
-
-class UltraSparcFrameInfo: public TargetFrameInfo {
-  const TargetMachine ⌖
-public:
-  UltraSparcFrameInfo(const TargetMachine &TM)
-    : TargetFrameInfo(StackGrowsDown, StackFrameSizeAlignment, 0), target(TM) {}
-  
-public:
-  // These methods provide constant parameters of the frame layout.
-  // 
-  int  getStackFrameSizeAlignment() const { return StackFrameSizeAlignment;}
-  int  getMinStackFrameSize()       const { return MinStackFrameSize; }
-  int  getNumFixedOutgoingArgs()    const { return NumFixedOutgoingArgs; }
-  int  getSizeOfEachArgOnStack()    const { return SizeOfEachArgOnStack; }
-  bool argsOnStackHaveFixedSize()   const { return true; }
-
-  // This method adjusts a stack offset to meet alignment rules of target.
-  // The fixed OFFSET (0x7ff) must be subtracted and the result aligned.
-  virtual int  adjustAlignment                  (int unalignedOffset,
-                                                 bool growUp,
-                                                 unsigned int align) const {
-    return unalignedOffset + (growUp? +1:-1)*((unalignedOffset-OFFSET) % align);
-  }
-
-  // These methods compute offsets using the frame contents for a
-  // particular function.  The frame contents are obtained from the
-  // MachineCodeInfoForMethod object for the given function.
-  // 
-  int getFirstIncomingArgOffset  (MachineFunction& mcInfo,
-                                  bool& growUp) const
-  {
-    growUp = true;                         // arguments area grows upwards
-    return FirstIncomingArgOffsetFromFP;
-  }
-  int getFirstOutgoingArgOffset  (MachineFunction& mcInfo,
-                                  bool& growUp) const
-  {
-    growUp = true;                         // arguments area grows upwards
-    return FirstOutgoingArgOffsetFromSP;
-  }
-  int getFirstOptionalOutgoingArgOffset(MachineFunction& mcInfo,
-                                        bool& growUp)const
-  {
-    growUp = true;                         // arguments area grows upwards
-    return FirstOptionalOutgoingArgOffsetFromSP;
-  }
-  
-  int getFirstAutomaticVarOffset (MachineFunction& mcInfo,
-                                  bool& growUp) const;
-  int getRegSpillAreaOffset      (MachineFunction& mcInfo,
-                                  bool& growUp) const;
-  int getTmpAreaOffset           (MachineFunction& mcInfo,
-                                  bool& growUp) const;
-  int getDynamicAreaOffset       (MachineFunction& mcInfo,
-                                  bool& growUp) const;
-
-  //
-  // These methods specify the base register used for each stack area
-  // (generally FP or SP)
-  // 
-  virtual int getIncomingArgBaseRegNum()               const {
-    return (int) target.getRegInfo().getFramePointer();
-  }
-  virtual int getOutgoingArgBaseRegNum()               const {
-    return (int) target.getRegInfo().getStackPointer();
-  }
-  virtual int getOptionalOutgoingArgBaseRegNum()       const {
-    return (int) target.getRegInfo().getStackPointer();
-  }
-  virtual int getAutomaticVarBaseRegNum()              const {
-    return (int) target.getRegInfo().getFramePointer();
-  }
-  virtual int getRegSpillAreaBaseRegNum()              const {
-    return (int) target.getRegInfo().getFramePointer();
-  }
-  virtual int getDynamicAreaBaseRegNum()               const {
-    return (int) target.getRegInfo().getStackPointer();
-  }
-
-  virtual int getIncomingArgOffset(MachineFunction& mcInfo,
-				   unsigned argNum) const {
-    assert(argsOnStackHaveFixedSize()); 
-  
-    unsigned relativeOffset = argNum * getSizeOfEachArgOnStack();
-    bool growUp;                          // do args grow up or down
-    int firstArg = getFirstIncomingArgOffset(mcInfo, growUp);
-    return growUp ? firstArg + relativeOffset : firstArg - relativeOffset; 
-  }
-
-  virtual int getOutgoingArgOffset(MachineFunction& mcInfo,
-				   unsigned argNum) const {
-    assert(argsOnStackHaveFixedSize()); 
-    //assert(((int) argNum - this->getNumFixedOutgoingArgs())
-    //     <= (int) mcInfo.getInfo()->getMaxOptionalNumArgs());
-    
-    unsigned relativeOffset = argNum * getSizeOfEachArgOnStack();
-    bool growUp;                          // do args grow up or down
-    int firstArg = getFirstOutgoingArgOffset(mcInfo, growUp);
-    return growUp ? firstArg + relativeOffset : firstArg - relativeOffset; 
-  }
-  
-private:
-  /*----------------------------------------------------------------------
-    This diagram shows the stack frame layout used by llc on Sparc V9.
-    Note that only the location of automatic variables, spill area,
-    temporary storage, and dynamically allocated stack area are chosen
-    by us.  The rest conform to the Sparc V9 ABI.
-    All stack addresses are offset by OFFSET = 0x7ff (2047).
-
-    Alignment assumptions and other invariants:
-    (1) %sp+OFFSET and %fp+OFFSET are always aligned on 16-byte boundary
-    (2) Variables in automatic, spill, temporary, or dynamic regions
-        are aligned according to their size as in all memory accesses.
-    (3) Everything below the dynamically allocated stack area is only used
-        during a call to another function, so it is never needed when
-        the current function is active.  This is why space can be allocated
-        dynamically by incrementing %sp any time within the function.
-    
-    STACK FRAME LAYOUT:
-
-       ...
-       %fp+OFFSET+176      Optional extra incoming arguments# 1..N
-       %fp+OFFSET+168      Incoming argument #6
-       ...                 ...
-       %fp+OFFSET+128      Incoming argument #1
-       ...                 ...
-    ---%fp+OFFSET-0--------Bottom of caller's stack frame--------------------
-       %fp+OFFSET-8        Automatic variables <-- ****TOP OF STACK FRAME****
-                           Spill area
-                           Temporary storage
-       ...
-
-       %sp+OFFSET+176+8N   Bottom of dynamically allocated stack area
-       %sp+OFFSET+168+8N   Optional extra outgoing argument# N
-       ...                 ...
-       %sp+OFFSET+176      Optional extra outgoing argument# 1
-       %sp+OFFSET+168      Outgoing argument #6
-       ...                 ...
-       %sp+OFFSET+128      Outgoing argument #1
-       %sp+OFFSET+120      Save area for %i7
-       ...                 ...
-       %sp+OFFSET+0        Save area for %l0 <-- ****BOTTOM OF STACK FRAME****
-
-   *----------------------------------------------------------------------*/
-
-  // All stack addresses must be offset by 0x7ff (2047) on Sparc V9.
-  static const int OFFSET                                  = (int) 0x7ff;
-  static const int StackFrameSizeAlignment                 =  16;
-  static const int MinStackFrameSize                       = 176;
-  static const int NumFixedOutgoingArgs                    =   6;
-  static const int SizeOfEachArgOnStack                    =   8;
-  static const int FirstIncomingArgOffsetFromFP            = 128 + OFFSET;
-  static const int FirstOptionalIncomingArgOffsetFromFP    = 176 + OFFSET;
-  static const int StaticAreaOffsetFromFP                  =   0 + OFFSET;
-  static const int FirstOutgoingArgOffsetFromSP            = 128 + OFFSET;
-  static const int FirstOptionalOutgoingArgOffsetFromSP    = 176 + OFFSET;
-};
-
-
 //---------------------------------------------------------------------------
-// class UltraSparcCacheInfo 
+// class SparcCacheInfo 
 // 
 // Purpose:
 //   Interface to cache parameters for the UltraSPARC.
 //   Just use defaults for now.
 //---------------------------------------------------------------------------
 
-struct UltraSparcCacheInfo: public TargetCacheInfo {
-  UltraSparcCacheInfo(const TargetMachine &T) : TargetCacheInfo(T) {} 
+struct SparcCacheInfo: public TargetCacheInfo {
+  SparcCacheInfo(const TargetMachine &T) : TargetCacheInfo(T) {} 
 };
 
 
@@ -663,38 +126,6 @@
 /// getBytecodeAsmPrinterPass - Emits final LLVM bytecode to assembly file.
 ///
 Pass* createBytecodeAsmPrinterPass(std::ostream &Out);
-
-//---------------------------------------------------------------------------
-// class UltraSparc 
-// 
-// Purpose:
-//   Primary interface to machine description for the UltraSPARC.
-//   Primarily just initializes machine-dependent parameters in
-//   class TargetMachine, and creates machine-dependent subclasses
-//   for classes such as InstrInfo, SchedInfo and RegInfo. 
-//---------------------------------------------------------------------------
-
-class UltraSparc : public TargetMachine {
-  UltraSparcInstrInfo instrInfo;
-  UltraSparcSchedInfo schedInfo;
-  UltraSparcRegInfo   regInfo;
-  UltraSparcFrameInfo frameInfo;
-  UltraSparcCacheInfo cacheInfo;
-public:
-  UltraSparc();
-
-  virtual const TargetInstrInfo  &getInstrInfo() const { return instrInfo; }
-  virtual const TargetSchedInfo  &getSchedInfo() const { return schedInfo; }
-  virtual const TargetRegInfo    &getRegInfo()   const { return regInfo; }
-  virtual const TargetFrameInfo  &getFrameInfo() const { return frameInfo; }
-  virtual const TargetCacheInfo  &getCacheInfo() const { return cacheInfo; }
-
-  virtual bool addPassesToEmitAssembly(PassManager &PM, std::ostream &Out);
-  virtual bool addPassesToJITCompile(FunctionPassManager &PM);
-  virtual bool addPassesToEmitMachineCode(FunctionPassManager &PM,
-                                          MachineCodeEmitter &MCE);
-  virtual void replaceMachineCodeForFunction(void *Old, void *New);
-};
 
 } // End llvm namespace
 


Index: llvm/lib/Target/Sparc/SparcRegClassInfo.cpp
diff -u llvm/lib/Target/Sparc/SparcRegClassInfo.cpp:1.32 llvm/lib/Target/Sparc/SparcRegClassInfo.cpp:1.33
--- llvm/lib/Target/Sparc/SparcRegClassInfo.cpp:1.32	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/SparcRegClassInfo.cpp	Wed Dec 17 16:04:00 2003
@@ -11,9 +11,10 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "llvm/Type.h"
 #include "SparcRegClassInfo.h"
 #include "SparcInternals.h"
-#include "llvm/Type.h"
+#include "SparcRegInfo.h"
 #include "../../CodeGen/RegAlloc/RegAllocCommon.h"   // FIXME!
 #include "../../CodeGen/RegAlloc/IGNode.h"           // FIXME!
 
@@ -320,8 +321,8 @@
                                         int RegTypeWanted,
                                     std::vector &IsColorUsedArr) const
 {
-  if (UserRegType == UltraSparcRegInfo::FPDoubleRegType ||
-      RegTypeWanted == UltraSparcRegInfo::FPDoubleRegType) {
+  if (UserRegType == SparcRegInfo::FPDoubleRegType ||
+      RegTypeWanted == SparcRegInfo::FPDoubleRegType) {
     // This register is used as or is needed as a double-precision reg.
     // We need to mark the [even,odd] pair corresponding to this reg.
     // Get the even numbered register corresponding to this reg.
@@ -348,7 +349,7 @@
 int SparcFloatRegClass::findUnusedColor(int RegTypeWanted,
                                 const std::vector &IsColorUsedArr) const
 {
-  if (RegTypeWanted == UltraSparcRegInfo::FPDoubleRegType) {
+  if (RegTypeWanted == SparcRegInfo::FPDoubleRegType) {
     unsigned NC = 2 * this->getNumOfAvailRegs();
     assert(IsColorUsedArr.size() == NC && "Invalid colors-used array");
     for (unsigned c = 0; c < NC; c+=2)


Index: llvm/lib/Target/Sparc/SparcRegInfo.cpp
diff -u llvm/lib/Target/Sparc/SparcRegInfo.cpp:1.114 llvm/lib/Target/Sparc/SparcRegInfo.cpp:1.115
--- llvm/lib/Target/Sparc/SparcRegInfo.cpp:1.114	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/SparcRegInfo.cpp	Wed Dec 17 16:04:00 2003
@@ -12,8 +12,6 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "SparcInternals.h"
-#include "SparcRegClassInfo.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/CodeGen/InstrSelection.h"
@@ -22,10 +20,14 @@
 #include "llvm/CodeGen/MachineInstrAnnot.h"
 #include "../../CodeGen/RegAlloc/LiveRangeInfo.h"   // FIXME!!
 #include "../../CodeGen/RegAlloc/LiveRange.h"       // FIXME!!
+#include "llvm/DerivedTypes.h"
+#include "llvm/Function.h"
 #include "llvm/iTerminators.h"
 #include "llvm/iOther.h"
-#include "llvm/Function.h"
-#include "llvm/DerivedTypes.h"
+#include "SparcInternals.h"
+#include "SparcRegClassInfo.h"
+#include "SparcRegInfo.h"
+#include "SparcTargetMachine.h"
 
 namespace llvm {
 
@@ -33,7 +35,7 @@
   BadRegClass = ~0
 };
 
-UltraSparcRegInfo::UltraSparcRegInfo(const UltraSparc &tgt)
+SparcRegInfo::SparcRegInfo(const SparcTargetMachine &tgt)
   : TargetRegInfo(tgt), NumOfIntArgRegs(6), NumOfFloatArgRegs(32)
 {
   MachineRegClassArr.push_back(new SparcIntRegClass(IntRegClassID));
@@ -50,16 +52,16 @@
 // getZeroRegNum - returns the register that contains always zero.
 // this is the unified register number
 //
-int UltraSparcRegInfo::getZeroRegNum() const {
-  return getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+int SparcRegInfo::getZeroRegNum() const {
+  return getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                           SparcIntRegClass::g0);
 }
 
 // getCallAddressReg - returns the reg used for pushing the address when a
 // method is called. This can be used for other purposes between calls
 //
-unsigned UltraSparcRegInfo::getCallAddressReg() const {
-  return getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+unsigned SparcRegInfo::getCallAddressReg() const {
+  return getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                           SparcIntRegClass::o7);
 }
 
@@ -67,8 +69,8 @@
 // It should be made sure that this  register contains the return 
 // value when a return instruction is reached.
 //
-unsigned UltraSparcRegInfo::getReturnAddressReg() const {
-  return getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+unsigned SparcRegInfo::getReturnAddressReg() const {
+  return getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                           SparcIntRegClass::i7);
 }
 
@@ -133,14 +135,14 @@
 }
 
 // Get unified reg number for frame pointer
-unsigned UltraSparcRegInfo::getFramePointer() const {
-  return getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+unsigned SparcRegInfo::getFramePointer() const {
+  return getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                           SparcIntRegClass::i6);
 }
 
 // Get unified reg number for stack pointer
-unsigned UltraSparcRegInfo::getStackPointer() const {
-  return getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+unsigned SparcRegInfo::getStackPointer() const {
+  return getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                           SparcIntRegClass::o6);
 }
 
@@ -173,7 +175,7 @@
 //                           regClassId is set to the register class ID.
 // 
 int
-UltraSparcRegInfo::regNumForIntArg(bool inCallee, bool isVarArgsCall,
+SparcRegInfo::regNumForIntArg(bool inCallee, bool isVarArgsCall,
                                    unsigned argNo, unsigned& regClassId) const
 {
   regClassId = IntRegClassID;
@@ -192,7 +194,7 @@
 //                           regClassId is set to the register class ID.
 // 
 int
-UltraSparcRegInfo::regNumForFPArg(unsigned regType,
+SparcRegInfo::regNumForFPArg(unsigned regType,
                                   bool inCallee, bool isVarArgsCall,
                                   unsigned argNo, unsigned& regClassId) const
 {
@@ -221,7 +223,7 @@
 // The following 4  methods are used to find the RegType (SparcInternals.h)
 // of a LiveRange, a Value, and for a given register unified reg number.
 //
-int UltraSparcRegInfo::getRegTypeForClassAndType(unsigned regClassID,
+int SparcRegInfo::getRegTypeForClassAndType(unsigned regClassID,
                                                  const Type* type) const
 {
   switch (regClassID) {
@@ -237,17 +239,17 @@
   }
 }
 
-int UltraSparcRegInfo::getRegTypeForDataType(const Type* type) const
+int SparcRegInfo::getRegTypeForDataType(const Type* type) const
 {
   return getRegTypeForClassAndType(getRegClassIDOfType(type), type);
 }
 
-int UltraSparcRegInfo::getRegTypeForLR(const LiveRange *LR) const
+int SparcRegInfo::getRegTypeForLR(const LiveRange *LR) const
 {
   return getRegTypeForClassAndType(LR->getRegClassID(), LR->getType());
 }
 
-int UltraSparcRegInfo::getRegType(int unifiedRegNum) const
+int SparcRegInfo::getRegType(int unifiedRegNum) const
 {
   if (unifiedRegNum < 32) 
     return IntRegType;
@@ -267,7 +269,7 @@
 
 // To find the register class used for a specified Type
 //
-unsigned UltraSparcRegInfo::getRegClassIDOfType(const Type *type,
+unsigned SparcRegInfo::getRegClassIDOfType(const Type *type,
                                                 bool isCCReg) const {
   Type::PrimitiveID ty = type->getPrimitiveID();
   unsigned res;
@@ -290,7 +292,7 @@
     return res;
 }
 
-unsigned UltraSparcRegInfo::getRegClassIDOfRegType(int regType) const {
+unsigned SparcRegInfo::getRegClassIDOfRegType(int regType) const {
   switch(regType) {
   case IntRegType:      return IntRegClassID;
   case FPSingleRegType:
@@ -307,7 +309,7 @@
 // Suggests a register for the ret address in the RET machine instruction.
 // We always suggest %i7 by convention.
 //---------------------------------------------------------------------------
-void UltraSparcRegInfo::suggestReg4RetAddr(MachineInstr *RetMI, 
+void SparcRegInfo::suggestReg4RetAddr(MachineInstr *RetMI, 
 					   LiveRangeInfo& LRI) const {
 
   assert(target.getInstrInfo().isReturn(RetMI->getOpCode()));
@@ -336,7 +338,7 @@
 // Sparc ABI dictates that %o7 be used for this purpose.
 //---------------------------------------------------------------------------
 void
-UltraSparcRegInfo::suggestReg4CallAddr(MachineInstr * CallMI,
+SparcRegInfo::suggestReg4CallAddr(MachineInstr * CallMI,
                                        LiveRangeInfo& LRI) const
 {
   CallArgsDescriptor* argDesc = CallArgsDescriptor::get(CallMI); 
@@ -360,7 +362,7 @@
 //  If the arg is passed on stack due to the lack of regs, NOTHING will be
 //  done - it will be colored (or spilled) as a normal live range.
 //---------------------------------------------------------------------------
-void UltraSparcRegInfo::suggestRegs4MethodArgs(const Function *Meth, 
+void SparcRegInfo::suggestRegs4MethodArgs(const Function *Meth, 
 					       LiveRangeInfo& LRI) const 
 {
   // Check if this is a varArgs function. needed for choosing regs.
@@ -393,7 +395,7 @@
 // the correct hardware registers if they did not receive the correct
 // (suggested) color through graph coloring.
 //---------------------------------------------------------------------------
-void UltraSparcRegInfo::colorMethodArgs(const Function *Meth, 
+void SparcRegInfo::colorMethodArgs(const Function *Meth, 
                             LiveRangeInfo &LRI,
                             std::vector& InstrnsBefore,
                             std::vector& InstrnsAfter) const {
@@ -566,7 +568,7 @@
 // This method is called before graph coloring to suggest colors to the
 // outgoing call args and the return value of the call.
 //---------------------------------------------------------------------------
-void UltraSparcRegInfo::suggestRegs4CallArgs(MachineInstr *CallMI, 
+void SparcRegInfo::suggestRegs4CallArgs(MachineInstr *CallMI, 
 					     LiveRangeInfo& LRI) const {
   assert ( (target.getInstrInfo()).isCall(CallMI->getOpCode()) );
 
@@ -634,7 +636,7 @@
 // this method is called for an LLVM return instruction to identify which
 // values will be returned from this method and to suggest colors.
 //---------------------------------------------------------------------------
-void UltraSparcRegInfo::suggestReg4RetValue(MachineInstr *RetMI, 
+void SparcRegInfo::suggestReg4RetValue(MachineInstr *RetMI, 
                                             LiveRangeInfo& LRI) const {
 
   assert( (target.getInstrInfo()).isReturn( RetMI->getOpCode() ) );
@@ -662,7 +664,7 @@
 //---------------------------------------------------------------------------
 
 bool
-UltraSparcRegInfo::regTypeNeedsScratchReg(int RegType,
+SparcRegInfo::regTypeNeedsScratchReg(int RegType,
                                           int& scratchRegType) const
 {
   if (RegType == IntCCRegType)
@@ -679,7 +681,7 @@
 //---------------------------------------------------------------------------
 
 void
-UltraSparcRegInfo::cpReg2RegMI(std::vector& mvec,
+SparcRegInfo::cpReg2RegMI(std::vector& mvec,
                                unsigned SrcReg,
                                unsigned DestReg,
                                int RegType) const {
@@ -695,7 +697,7 @@
     if (getRegType(DestReg) == IntRegType) {
       // copy intCC reg to int reg
       MI = (BuildMI(V9::RDCCR, 2)
-            .addMReg(getUnifiedRegNum(UltraSparcRegInfo::IntCCRegClassID,
+            .addMReg(getUnifiedRegNum(SparcRegInfo::IntCCRegClassID,
                                       SparcIntCCRegClass::ccr))
             .addMReg(DestReg,MOTy::Def));
     } else {
@@ -705,7 +707,7 @@
       MI = (BuildMI(V9::WRCCRr, 3)
             .addMReg(SrcReg)
             .addMReg(SparcIntRegClass::g0)
-            .addMReg(getUnifiedRegNum(UltraSparcRegInfo::IntCCRegClassID,
+            .addMReg(getUnifiedRegNum(SparcRegInfo::IntCCRegClassID,
                                       SparcIntCCRegClass::ccr), MOTy::Def));
     }
     break;
@@ -743,7 +745,7 @@
 
 
 void
-UltraSparcRegInfo::cpReg2MemMI(std::vector& mvec,
+SparcRegInfo::cpReg2MemMI(std::vector& mvec,
                                unsigned SrcReg, 
                                unsigned PtrReg,
                                int Offset, int RegType,
@@ -763,7 +765,7 @@
       OffReg = PRA.getUnusedUniRegAtMI(RC, RegType, MInst, LVSetBef);
 #else
       // Default to using register g4 for holding large offsets
-      OffReg = getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+      OffReg = getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                                 SparcIntRegClass::g4);
 #endif
       assert(OffReg >= 0 && "FIXME: cpReg2MemMI cannot find an unused reg.");
@@ -796,7 +798,7 @@
     assert(scratchReg >= 0 && "Need scratch reg to store %ccr to memory");
     assert(getRegType(scratchReg) ==IntRegType && "Invalid scratch reg");
     MI = (BuildMI(V9::RDCCR, 2)
-          .addMReg(getUnifiedRegNum(UltraSparcRegInfo::IntCCRegClassID,
+          .addMReg(getUnifiedRegNum(SparcRegInfo::IntCCRegClassID,
                                     SparcIntCCRegClass::ccr))
           .addMReg(scratchReg, MOTy::Def));
     mvec.push_back(MI);
@@ -805,7 +807,7 @@
     return;
 
   case FloatCCRegType: {
-    unsigned fsrReg =  getUnifiedRegNum(UltraSparcRegInfo::SpecialRegClassID,
+    unsigned fsrReg =  getUnifiedRegNum(SparcRegInfo::SpecialRegClassID,
                                            SparcSpecialRegClass::fsr);
     if (target.getInstrInfo().constantFitsInImmedField(V9::STXFSRi, Offset))
       MI=BuildMI(V9::STXFSRi,3).addMReg(fsrReg).addMReg(PtrReg).addSImm(Offset);
@@ -827,7 +829,7 @@
 
 
 void
-UltraSparcRegInfo::cpMem2RegMI(std::vector& mvec,
+SparcRegInfo::cpMem2RegMI(std::vector& mvec,
                                unsigned PtrReg,	
                                int Offset,
                                unsigned DestReg,
@@ -848,7 +850,7 @@
       OffReg = PRA.getUnusedUniRegAtMI(RC, RegType, MInst, LVSetBef);
 #else
       // Default to using register g4 for holding large offsets
-      OffReg = getUnifiedRegNum(UltraSparcRegInfo::IntRegClassID,
+      OffReg = getUnifiedRegNum(SparcRegInfo::IntRegClassID,
                                 SparcIntRegClass::g4);
 #endif
       assert(OffReg >= 0 && "FIXME: cpReg2MemMI cannot find an unused reg.");
@@ -890,12 +892,12 @@
     MI = (BuildMI(V9::WRCCRr, 3)
           .addMReg(scratchReg)
           .addMReg(SparcIntRegClass::g0)
-          .addMReg(getUnifiedRegNum(UltraSparcRegInfo::IntCCRegClassID,
+          .addMReg(getUnifiedRegNum(SparcRegInfo::IntCCRegClassID,
                                     SparcIntCCRegClass::ccr), MOTy::Def));
     break;
     
   case FloatCCRegType: {
-    unsigned fsrRegNum =  getUnifiedRegNum(UltraSparcRegInfo::SpecialRegClassID,
+    unsigned fsrRegNum =  getUnifiedRegNum(SparcRegInfo::SpecialRegClassID,
                                            SparcSpecialRegClass::fsr);
     if (target.getInstrInfo().constantFitsInImmedField(V9::LDXFSRi, Offset))
       MI = BuildMI(V9::LDXFSRi, 3).addMReg(PtrReg).addSImm(Offset)
@@ -919,7 +921,7 @@
 
 
 void
-UltraSparcRegInfo::cpValue2Value(Value *Src, Value *Dest,
+SparcRegInfo::cpValue2Value(Value *Src, Value *Dest,
                                  std::vector& mvec) const {
   int RegType = getRegTypeForDataType(Src->getType());
   MachineInstr * MI = NULL;
@@ -948,7 +950,7 @@
 // Print the register assigned to a LR
 //---------------------------------------------------------------------------
 
-void UltraSparcRegInfo::printReg(const LiveRange *LR) const {
+void SparcRegInfo::printReg(const LiveRange *LR) const {
   unsigned RegClassID = LR->getRegClassID();
   std::cerr << " Node ";
 


Index: llvm/lib/Target/Sparc/SparcTargetMachine.cpp
diff -u llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.91 llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.92
--- llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.91	Wed Nov 12 18:16:28 2003
+++ llvm/lib/Target/Sparc/SparcTargetMachine.cpp	Wed Dec 17 16:04:00 2003
@@ -6,27 +6,31 @@
 // the University of Illinois Open Source License. See LICENSE.TXT for details.
 // 
 //===----------------------------------------------------------------------===//
-//
-// This file contains the code for the Sparc Target that does not fit in any of
-// the other files in this directory.
-//
+// 
+// Primary interface to machine description for the UltraSPARC.  Primarily just
+// initializes machine-dependent parameters in class TargetMachine, and creates
+// machine-dependent subclasses for classes such as TargetInstrInfo.
+// 
 //===----------------------------------------------------------------------===//
 
-#include "SparcInternals.h"
-#include "MappingInfo.h" 
 #include "llvm/Function.h"
 #include "llvm/PassManager.h"
 #include "llvm/Assembly/PrintModulePass.h"
-#include "llvm/Transforms/Scalar.h"
-#include "llvm/CodeGen/MachineFunction.h"
-#include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/CodeGen/InstrSelection.h"
 #include "llvm/CodeGen/InstrScheduling.h"
+#include "llvm/CodeGen/MachineFunction.h"
+#include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/CodeGen/MachineCodeForInstruction.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/Target/TargetMachineImpls.h"
+#include "llvm/Transforms/Scalar.h"
+#include "MappingInfo.h" 
+#include "SparcInternals.h"
+#include "SparcTargetMachine.h"
 #include "Support/CommandLine.h"
 
+using namespace llvm;
+
 namespace llvm {
 
 static const unsigned ImplicitRegUseList[] = { 0 }; /* not used yet */
@@ -62,87 +66,9 @@
                           cl::Hidden);
 }
 
-//----------------------------------------------------------------------------
-// allocateSparcTargetMachine - Allocate and return a subclass of TargetMachine
-// that implements the Sparc backend. (the llvm/CodeGen/Sparc.h interface)
-//----------------------------------------------------------------------------
-
-TargetMachine *allocateSparcTargetMachine(const Module &M) {
-  return new UltraSparc();
-}
-
-//---------------------------------------------------------------------------
-// class UltraSparcFrameInfo 
-// 
-//   Interface to stack frame layout info for the UltraSPARC.
-//   Starting offsets for each area of the stack frame are aligned at
-//   a multiple of getStackFrameSizeAlignment().
-//---------------------------------------------------------------------------
-
-int
-UltraSparcFrameInfo::getFirstAutomaticVarOffset(MachineFunction& ,
-                                                bool& pos) const
-{
-  pos = false;                          // static stack area grows downwards
-  return StaticAreaOffsetFromFP;
-}
-
-int
-UltraSparcFrameInfo::getRegSpillAreaOffset(MachineFunction& mcInfo,
-                                           bool& pos) const
-{
-  // ensure no more auto vars are added
-  mcInfo.getInfo()->freezeAutomaticVarsArea();
-  
-  pos = false;                          // static stack area grows downwards
-  unsigned autoVarsSize = mcInfo.getInfo()->getAutomaticVarsSize();
-  return StaticAreaOffsetFromFP - autoVarsSize; 
-}
-
-int
-UltraSparcFrameInfo::getTmpAreaOffset(MachineFunction& mcInfo,
-                                      bool& pos) const
-{
-  MachineFunctionInfo *MFI = mcInfo.getInfo();
-  MFI->freezeAutomaticVarsArea();     // ensure no more auto vars are added
-  MFI->freezeSpillsArea();            // ensure no more spill slots are added
-  
-  pos = false;                          // static stack area grows downwards
-  unsigned autoVarsSize = MFI->getAutomaticVarsSize();
-  unsigned spillAreaSize = MFI->getRegSpillsSize();
-  int offset = autoVarsSize + spillAreaSize;
-  return StaticAreaOffsetFromFP - offset;
-}
-
-int
-UltraSparcFrameInfo::getDynamicAreaOffset(MachineFunction& mcInfo,
-                                          bool& pos) const
-{
-  // Dynamic stack area grows downwards starting at top of opt-args area.
-  // The opt-args, required-args, and register-save areas are empty except
-  // during calls and traps, so they are shifted downwards on each
-  // dynamic-size alloca.
-  pos = false;
-  unsigned optArgsSize = mcInfo.getInfo()->getMaxOptionalArgsSize();
-  if (int extra = optArgsSize % getStackFrameSizeAlignment())
-    optArgsSize += (getStackFrameSizeAlignment() - extra);
-  int offset = optArgsSize + FirstOptionalOutgoingArgOffsetFromSP;
-  assert((offset - OFFSET) % getStackFrameSizeAlignment() == 0);
-  return offset;
-}
-
-//---------------------------------------------------------------------------
-// class UltraSparcMachine 
-// 
-// Purpose:
-//   Primary interface to machine description for the UltraSPARC.
-//   Primarily just initializes machine-dependent parameters in
-//   class TargetMachine, and creates machine-dependent subclasses
-//   for classes such as TargetInstrInfo. 
-// 
-//---------------------------------------------------------------------------
+} // End llvm namespace
 
-UltraSparc::UltraSparc()
+SparcTargetMachine::SparcTargetMachine()
   : TargetMachine("UltraSparc-Native", false),
     schedInfo(*this),
     regInfo(*this),
@@ -153,7 +79,8 @@
 // addPassesToEmitAssembly - This method controls the entire code generation
 // process for the ultra sparc.
 //
-bool UltraSparc::addPassesToEmitAssembly(PassManager &PM, std::ostream &Out)
+bool
+SparcTargetMachine::addPassesToEmitAssembly(PassManager &PM, std::ostream &Out)
 {
   // The following 3 passes used to be inserted specially by llc.
   // Replace malloc and free instructions with library calls.
@@ -225,7 +152,7 @@
 // addPassesToJITCompile - This method controls the JIT method of code
 // generation for the UltraSparc.
 //
-bool UltraSparc::addPassesToJITCompile(FunctionPassManager &PM) {
+bool SparcTargetMachine::addPassesToJITCompile(FunctionPassManager &PM) {
   const TargetData &TD = getTargetData();
 
   PM.add(new TargetData("lli", TD.isLittleEndian(), TD.getPointerSize(),
@@ -269,4 +196,15 @@
   return false; // success!
 }
 
-} // End llvm namespace
+//----------------------------------------------------------------------------
+// allocateSparcTargetMachine - Allocate and return a subclass of TargetMachine
+// that implements the Sparc backend. (the llvm/CodeGen/Sparc.h interface)
+//----------------------------------------------------------------------------
+
+namespace llvm {
+
+TargetMachine *allocateSparcTargetMachine(const Module &M) {
+  return new SparcTargetMachine();
+}
+
+}


Index: llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp
diff -u llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.46 llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.47
--- llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.46	Sun Dec 14 07:24:15 2003
+++ llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp	Wed Dec 17 16:04:00 2003
@@ -35,6 +35,8 @@
 #include "Support/hash_set"
 #include "Support/Statistic.h"
 #include "SparcInternals.h"
+#include "SparcTargetMachine.h"
+#include "SparcRegInfo.h"
 #include "SparcV9CodeEmitter.h"
 #include "Config/alloca.h"
 
@@ -46,8 +48,8 @@
   Statistic<> CallbackCalls("callback", "Number CompilationCallback() calls");
 }
 
-bool UltraSparc::addPassesToEmitMachineCode(FunctionPassManager &PM,
-                                            MachineCodeEmitter &MCE) {
+bool SparcTargetMachine::addPassesToEmitMachineCode(FunctionPassManager &PM,
+                                                    MachineCodeEmitter &MCE) {
   MachineCodeEmitter *M = &MCE;
   DEBUG(M = MachineCodeEmitter::createFilePrinterEmitter(MCE));
   PM.add(new SparcV9CodeEmitter(*this, *M));
@@ -102,10 +104,10 @@
 
   private:
     uint64_t emitStubForFunction(Function *F);
-    static void SaveRegisters(uint64_t DoubleFP[], uint64_t &FSR,
-                              uint64_t &FPRS, uint64_t &CCR);
-    static void RestoreRegisters(uint64_t DoubleFP[], uint64_t &FSR,
-                                 uint64_t &FPRS, uint64_t &CCR);
+    static void SaveRegisters(uint64_t DoubleFP[], uint64_t CC[],
+                              uint64_t Globals[]);
+    static void RestoreRegisters(uint64_t DoubleFP[], uint64_t CC[],
+                                 uint64_t Globals[]);
     static void CompilationCallback();
     uint64_t resolveFunctionReference(uint64_t RetAddr);
 
@@ -209,17 +211,20 @@
   }
 }
 
-void JITResolver::SaveRegisters(uint64_t DoubleFP[], uint64_t &FSR, 
-                                uint64_t &FPRS, uint64_t &CCR) {
+void JITResolver::SaveRegisters(uint64_t DoubleFP[], uint64_t CC[], 
+                                uint64_t Globals[]) {
 #if defined(sparc) || defined(__sparc__) || defined(__sparcv9)
 
-#if 0
   __asm__ __volatile__ (// Save condition-code registers
                         "stx %%fsr, %0;\n\t" 
                         "rd %%fprs, %1;\n\t" 
                         "rd %%ccr,  %2;\n\t"
-                        : "=m"(FSR), "=r"(FPRS), "=r"(CCR));
-#endif
+                        : "=m"(CC[0]), "=r"(CC[1]), "=r"(CC[2]));
+
+  __asm__ __volatile__ (// Save globals g1 and g5
+                        "stx %%g1, %0;\n\t"
+                        "stx %%g5, %0;\n\t"
+                        : "=m"(Globals[0]), "=m"(Globals[1]));
 
   // GCC says: `asm' only allows up to thirty parameters!
   __asm__ __volatile__ (// Save Single/Double FP registers, part 1
@@ -261,18 +266,21 @@
 }
 
 
-void JITResolver::RestoreRegisters(uint64_t DoubleFP[], uint64_t &FSR, 
-                                   uint64_t &FPRS, uint64_t &CCR)
+void JITResolver::RestoreRegisters(uint64_t DoubleFP[], uint64_t CC[], 
+                                   uint64_t Globals[])
 {
 #if defined(sparc) || defined(__sparc__) || defined(__sparcv9)
 
-#if 0
   __asm__ __volatile__ (// Restore condition-code registers
                         "ldx %0,    %%fsr;\n\t" 
                         "wr  %1, 0, %%fprs;\n\t"
                         "wr  %2, 0, %%ccr;\n\t" 
-                        :: "m"(FSR), "r"(FPRS), "r"(CCR));
-#endif
+                        :: "m"(CC[0]), "r"(CC[1]), "r"(CC[2]));
+
+  __asm__ __volatile__ (// Restore globals g1 and g5
+                        "ldx %0, %%g1;\n\t"
+                        "ldx %0, %%g5;\n\t"
+                        :: "m"(Globals[0]), "m"(Globals[1]));
 
   // GCC says: `asm' only allows up to thirty parameters!
   __asm__ __volatile__ (// Restore Single/Double FP registers, part 1
@@ -314,11 +322,12 @@
 }
 
 void JITResolver::CompilationCallback() {
-  // Local space to save double registers
+  // Local space to save the registers
   uint64_t DoubleFP[32];
-  uint64_t FSR, FPRS, CCR;
+  uint64_t CC[3];
+  uint64_t Globals[2];
 
-  SaveRegisters(DoubleFP, FSR, FPRS, CCR);
+  SaveRegisters(DoubleFP, CC, Globals);
   ++CallbackCalls;
 
   uint64_t CameFrom = (uint64_t)(intptr_t)__builtin_return_address(0);
@@ -394,7 +403,7 @@
   __asm__ __volatile__ ("sub %%i7, %0, %%i7" : : "r" (Offset+12));
 #endif
 
-  RestoreRegisters(DoubleFP, FSR, FPRS, CCR);
+  RestoreRegisters(DoubleFP, CC, Globals);
 }
 
 /// emitStubForFunction - This method is used by the JIT when it needs to emit
@@ -476,7 +485,7 @@
   fakeReg = RI.getClassRegNum(fakeReg, regClass);
 
   switch (regClass) {
-  case UltraSparcRegInfo::IntRegClassID: {
+  case SparcRegInfo::IntRegClassID: {
     // Sparc manual, p31
     static const unsigned IntRegMap[] = {
       // "o0", "o1", "o2", "o3", "o4", "o5",       "o7",
@@ -494,12 +503,12 @@
     return IntRegMap[fakeReg];
     break;
   }
-  case UltraSparcRegInfo::FloatRegClassID: {
+  case SparcRegInfo::FloatRegClassID: {
     DEBUG(std::cerr << "FP reg: " << fakeReg << "\n");
-    if (regType == UltraSparcRegInfo::FPSingleRegType) {
+    if (regType == SparcRegInfo::FPSingleRegType) {
       // only numbered 0-31, hence can already fit into 5 bits (and 6)
       DEBUG(std::cerr << "FP single reg, returning: " << fakeReg << "\n");
-    } else if (regType == UltraSparcRegInfo::FPDoubleRegType) {
+    } else if (regType == SparcRegInfo::FPDoubleRegType) {
       // FIXME: This assumes that we only have 5-bit register fields!
       // From Sparc Manual, page 40.
       // The bit layout becomes: b[4], b[3], b[2], b[1], b[5]
@@ -509,7 +518,7 @@
     }
     return fakeReg;
   }
-  case UltraSparcRegInfo::IntCCRegClassID: {
+  case SparcRegInfo::IntCCRegClassID: {
     /*                                   xcc, icc, ccr */
     static const unsigned IntCCReg[] = {  6,   4,   2 };
     
@@ -518,7 +527,7 @@
     DEBUG(std::cerr << "IntCC reg: " << IntCCReg[fakeReg] << "\n");
     return IntCCReg[fakeReg];
   }
-  case UltraSparcRegInfo::FloatCCRegClassID: {
+  case SparcRegInfo::FloatCCRegClassID: {
     /* These are laid out %fcc0 - %fcc3 => 0 - 3, so are correct */
     DEBUG(std::cerr << "FP CC reg: " << fakeReg << "\n");
     return fakeReg;
@@ -573,7 +582,7 @@
   }
 }
 
-void UltraSparc::replaceMachineCodeForFunction (void *Old, void *New) {
+void SparcTargetMachine::replaceMachineCodeForFunction (void *Old, void *New) {
   assert (TheJITResolver &&
 	"Can only call replaceMachineCodeForFunction from within JIT");
   uint64_t Target = (uint64_t)(intptr_t)New;


Index: llvm/lib/Target/Sparc/UltraSparcSchedInfo.cpp
diff -u llvm/lib/Target/Sparc/UltraSparcSchedInfo.cpp:1.8 llvm/lib/Target/Sparc/UltraSparcSchedInfo.cpp:1.9
--- llvm/lib/Target/Sparc/UltraSparcSchedInfo.cpp:1.8	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/UltraSparcSchedInfo.cpp	Wed Dec 17 16:04:00 2003
@@ -728,7 +728,7 @@
 
 
 //---------------------------------------------------------------------------
-// class UltraSparcSchedInfo 
+// class SparcSchedInfo 
 // 
 // Purpose:
 //   Scheduling information for the UltraSPARC.
@@ -737,7 +737,7 @@
 //---------------------------------------------------------------------------
 
 /*ctor*/
-UltraSparcSchedInfo::UltraSparcSchedInfo(const TargetMachine& tgt)
+SparcSchedInfo::SparcSchedInfo(const TargetMachine& tgt)
   : TargetSchedInfo(tgt,
                      (unsigned int) SPARC_NUM_SCHED_CLASSES,
 		     SparcRUsageDesc,
@@ -764,7 +764,7 @@
 }
 
 void
-UltraSparcSchedInfo::initializeResources()
+SparcSchedInfo::initializeResources()
 {
   // Compute TargetSchedInfo::instrRUsages and TargetSchedInfo::issueGaps
   TargetSchedInfo::initializeResources();




From brukman at cs.uiuc.edu  Wed Dec 17 16:06:06 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 16:06:06 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/Sparc.cpp
Message-ID: <200312172205.QAA25343@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

Sparc.cpp (r1.91) removed

---
Log message:

The contents of this file have been absorbed into SparcTargetMachine, et al.


---
Diffs of the changes:  (+0 -0)



From brukman at cs.uiuc.edu  Wed Dec 17 16:07:01 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 16:07:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/PreSelection.cpp
Message-ID: <200312172206.QAA25377@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

PreSelection.cpp updated: 1.26 -> 1.27

---
Log message:

Doxygenified some comments, reduced extraneous space.


---
Diffs of the changes:  (+12 -19)

Index: llvm/lib/Target/Sparc/PreSelection.cpp
diff -u llvm/lib/Target/Sparc/PreSelection.cpp:1.26 llvm/lib/Target/Sparc/PreSelection.cpp:1.27
--- llvm/lib/Target/Sparc/PreSelection.cpp:1.26	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/PreSelection.cpp	Wed Dec 17 16:06:08 2003
@@ -94,13 +94,11 @@
     : NULL;
 }
 
-
-// Wrapper on Constant::classof to use in find_if :-(
+// Wrapper on Constant::classof to use in find_if
 inline static bool nonConstant(const Use& U) {
   return ! isa(U);
 }
 
-
 static Instruction* DecomposeConstantExpr(ConstantExpr* CE,
                                           Instruction& insertBefore)
 {
@@ -177,15 +175,15 @@
   }
 }
 
-// visitOperands() transforms individual operands of all instructions:
-// -- Load "large" int constants into a virtual register.  What is large
-//    depends on the type of instruction and on the target architecture.
-// -- For any constants that cannot be put in an immediate field,
-//    load address into virtual register first, and then load the constant.
-// 
-// firstOp and lastOp can be used to skip leading and trailing operands.
-// If lastOp is 0, it defaults to #operands or #incoming Phi values.
-//  
+/// visitOperands - transform individual operands of all instructions:
+/// -- Load "large" int constants into a virtual register.  What is large
+///    depends on the type of instruction and on the target architecture.
+/// -- For any constants that cannot be put in an immediate field,
+///    load address into virtual register first, and then load the constant.
+/// 
+/// firstOp and lastOp can be used to skip leading and trailing operands.
+/// If lastOp is 0, it defaults to #operands or #incoming Phi values.
+///  
 inline void PreSelection::visitOperands(Instruction &I, int firstOp) {
   // For any instruction other than PHI, copies go just before the instr.
   for (unsigned i = firstOp, e = I.getNumOperands(); i != e; ++i)
@@ -205,15 +203,12 @@
   // do not call visitOperands!
 }
 
-
-
 // Common work for *all* instructions.  This needs to be called explicitly
 // by other visit functions.
 inline void PreSelection::visitInstruction(Instruction &I) { 
   visitOperands(I);              // Perform operand transformations
 }
 
-
 // GetElementPtr instructions: check if pointer is a global
 void PreSelection::visitGetElementPtrInst(GetElementPtrInst &I) { 
   Instruction* curI = &I;
@@ -239,10 +234,8 @@
   visitOperands(I, (/*firstOp=*/ I.getCalledFunction()? 1 : 0));
 }
 
-//===----------------------------------------------------------------------===//
-// createPreSelectionPass - Public entrypoint for pre-selection pass
-// and this file as a whole...
-//
+/// createPreSelectionPass - Public entry point for the PreSelection pass
+///
 FunctionPass* createPreSelectionPass(const TargetMachine &TM) {
   return new PreSelection(TM);
 }




From brukman at cs.uiuc.edu  Wed Dec 17 16:07:07 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 16:07:07 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/EmitAssembly.cpp
Message-ID: <200312172206.QAA25401@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

EmitAssembly.cpp updated: 1.97 -> 1.98

---
Log message:

Reordered #includes.


---
Diffs of the changes:  (+5 -5)

Index: llvm/lib/Target/Sparc/EmitAssembly.cpp
diff -u llvm/lib/Target/Sparc/EmitAssembly.cpp:1.97 llvm/lib/Target/Sparc/EmitAssembly.cpp:1.98
--- llvm/lib/Target/Sparc/EmitAssembly.cpp:1.97	Sun Dec 14 07:24:15 2003
+++ llvm/lib/Target/Sparc/EmitAssembly.cpp	Wed Dec 17 16:06:28 2003
@@ -18,16 +18,16 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "llvm/CodeGen/MachineInstr.h"
-#include "llvm/CodeGen/MachineConstantPool.h"
-#include "llvm/CodeGen/MachineFunction.h"
-#include "llvm/CodeGen/MachineFunctionInfo.h"
 #include "llvm/Constants.h"
 #include "llvm/DerivedTypes.h"
 #include "llvm/Module.h"
-#include "llvm/SlotCalculator.h"
 #include "llvm/Pass.h"
+#include "llvm/SlotCalculator.h"
 #include "llvm/Assembly/Writer.h"
+#include "llvm/CodeGen/MachineConstantPool.h"
+#include "llvm/CodeGen/MachineFunction.h"
+#include "llvm/CodeGen/MachineFunctionInfo.h"
+#include "llvm/CodeGen/MachineInstr.h"
 #include "Support/StringExtras.h"
 #include "Support/Statistic.h"
 #include "SparcInternals.h"




From brukman at cs.uiuc.edu  Wed Dec 17 16:09:01 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 16:09:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/PeepholeOpts.cpp
Message-ID: <200312172208.QAA25448@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

PeepholeOpts.cpp updated: 1.16 -> 1.17

---
Log message:

* Converted C-style comments to C++
* Doxygenified comments
* Reordered #includes


---
Diffs of the changes:  (+15 -13)

Index: llvm/lib/Target/Sparc/PeepholeOpts.cpp
diff -u llvm/lib/Target/Sparc/PeepholeOpts.cpp:1.16 llvm/lib/Target/Sparc/PeepholeOpts.cpp:1.17
--- llvm/lib/Target/Sparc/PeepholeOpts.cpp:1.16	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/Sparc/PeepholeOpts.cpp	Wed Dec 17 16:08:20 2003
@@ -13,12 +13,12 @@
 //===----------------------------------------------------------------------===//
 
 #include "SparcInternals.h"
+#include "llvm/BasicBlock.h"
+#include "llvm/Pass.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineInstr.h"
-#include "llvm/Target/TargetMachine.h"
 #include "llvm/Target/TargetInstrInfo.h"
-#include "llvm/BasicBlock.h"
-#include "llvm/Pass.h"
+#include "llvm/Target/TargetMachine.h"
 
 namespace llvm {
 
@@ -62,7 +62,7 @@
 
 static bool IsUselessCopy(const TargetMachine &target, const MachineInstr* MI) {
   if (MI->getOpCode() == V9::FMOVS || MI->getOpCode() == V9::FMOVD) {
-    return (/* both operands are allocated to the same register */
+    return (// both operands are allocated to the same register
             MI->getOperand(0).getAllocatedRegNum() == 
             MI->getOperand(1).getAllocatedRegNum());
   } else if (MI->getOpCode() == V9::ADDr || MI->getOpCode() == V9::ORr ||
@@ -78,14 +78,14 @@
     if (srcWithDestReg == 2)
       return false;
     else {
-      /* else source and dest are allocated to the same register */
+      // else source and dest are allocated to the same register
       unsigned otherOp = 1 - srcWithDestReg;
-      return (/* either operand otherOp is register %g0 */
+      return (// either operand otherOp is register %g0
               (MI->getOperand(otherOp).hasAllocatedReg() &&
                MI->getOperand(otherOp).getAllocatedRegNum() ==
                target.getRegInfo().getZeroRegNum()) ||
               
-              /* or operand otherOp == 0 */
+              // or operand otherOp == 0
               (MI->getOperand(otherOp).getType()
                == MachineOperand::MO_SignExtendedImmed &&
                MI->getOperand(otherOp).getImmedValue() == 0));
@@ -117,6 +117,11 @@
   PeepholeOpts(const TargetMachine &TM): target(TM) { }
   bool runOnBasicBlock(BasicBlock &BB); // apply this pass to each BB
   virtual const char *getPassName() const { return "Peephole Optimization"; }
+
+  // getAnalysisUsage - this pass preserves the CFG
+  void getAnalysisUsage(AnalysisUsage &AU) const {
+    AU.setPreservesCFG();
+  }
 };
 
 // Apply a list of peephole optimizations to this machine instruction
@@ -125,7 +130,7 @@
 //
 bool PeepholeOpts::visit(MachineBasicBlock& mvec,
                          MachineBasicBlock::iterator BBI) const {
-  /* Remove redundant copy instructions */
+  // Remove redundant copy instructions
   return RemoveUselessCopies(mvec, BBI, target);
 }
 
@@ -157,11 +162,8 @@
   return true;
 }
 
-
-//===----------------------------------------------------------------------===//
-// createPeepholeOptsPass - Public entrypoint for peephole optimization
-// and this file as a whole...
-//
+/// createPeepholeOptsPass - Public entry point for peephole optimization
+///
 FunctionPass* createPeepholeOptsPass(const TargetMachine &TM) {
   return new PeepholeOpts(TM);
 }




From criswell at cs.uiuc.edu  Wed Dec 17 16:48:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 16:48:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/docs/AliasAnalysis.html CFEBuildInstrs.html CodingStandards.html CommandLine.html FAQ.html GettingStarted.html HowToSubmitABug.html LLVMVsTheWorld.html LangRef.html OpenProjects.html Projects.html ReleaseNotes.html Stacker.html TestingGuide.html
Message-ID: <200312172247.QAA01383@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1/docs:

AliasAnalysis.html updated: 1.1 -> 1.2
CFEBuildInstrs.html updated: 1.1 -> 1.2
CodingStandards.html updated: 1.1 -> 1.2
CommandLine.html updated: 1.1 -> 1.2
FAQ.html updated: 1.1 -> 1.2
GettingStarted.html updated: 1.1 -> 1.2
HowToSubmitABug.html updated: 1.1 -> 1.2
LLVMVsTheWorld.html updated: 1.1 -> 1.2
LangRef.html updated: 1.1 -> 1.2
OpenProjects.html updated: 1.1 -> 1.2
Projects.html updated: 1.1 -> 1.2
ReleaseNotes.html updated: 1.3 -> 1.4
Stacker.html updated: 1.1 -> 1.2
TestingGuide.html updated: 1.1 -> 1.2

---
Log message:

Updating to new docs in LLVM tarball.



---
Diffs of the changes:  (+254 -165)

Index: llvm-www/releases/1.1/docs/AliasAnalysis.html
diff -u llvm-www/releases/1.1/docs/AliasAnalysis.html:1.1 llvm-www/releases/1.1/docs/AliasAnalysis.html:1.2
--- llvm-www/releases/1.1/docs/AliasAnalysis.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/AliasAnalysis.html	Wed Dec 17 16:47:16 2003
@@ -486,7 +486,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/CFEBuildInstrs.html
diff -u llvm-www/releases/1.1/docs/CFEBuildInstrs.html:1.1 llvm-www/releases/1.1/docs/CFEBuildInstrs.html:1.2
--- llvm-www/releases/1.1/docs/CFEBuildInstrs.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/CFEBuildInstrs.html	Wed Dec 17 16:47:16 2003
@@ -262,7 +262,7 @@
   Brian Gaeke
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/CodingStandards.html
diff -u llvm-www/releases/1.1/docs/CodingStandards.html:1.1 llvm-www/releases/1.1/docs/CodingStandards.html:1.2
--- llvm-www/releases/1.1/docs/CodingStandards.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/CodingStandards.html	Wed Dec 17 16:47:16 2003
@@ -952,7 +952,7 @@
   Chris Lattner
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/CommandLine.html
diff -u llvm-www/releases/1.1/docs/CommandLine.html:1.1 llvm-www/releases/1.1/docs/CommandLine.html:1.2
--- llvm-www/releases/1.1/docs/CommandLine.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/CommandLine.html	Wed Dec 17 16:47:16 2003
@@ -1692,7 +1692,7 @@
   Chris Lattner
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/FAQ.html
diff -u llvm-www/releases/1.1/docs/FAQ.html:1.1 llvm-www/releases/1.1/docs/FAQ.html:1.2
--- llvm-www/releases/1.1/docs/FAQ.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/FAQ.html	Wed Dec 17 16:47:16 2003
@@ -438,7 +438,7 @@
 
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/GettingStarted.html
diff -u llvm-www/releases/1.1/docs/GettingStarted.html:1.1 llvm-www/releases/1.1/docs/GettingStarted.html:1.2
--- llvm-www/releases/1.1/docs/GettingStarted.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/GettingStarted.html	Wed Dec 17 16:47:16 2003
@@ -182,24 +182,55 @@
 
   Linux on x86 (Pentium and above)
   
-    Approximately 760 MB of Free Disk Space
+    
Approximately 918 MB of Free Disk Space
     
-      Source code: 30 MB
-      Object code: 670 MB
-      GCC front end: 60 MB
+      Source code: 28 MB
+      Object code: 850 MB
+      GCC front end: 40 MB
     
-  
+  
+  
+
+  
 
   Solaris on SparcV9 (Ultrasparc)
   
-    Approximately 1.24 GB of Free Disk Space
+    
Approximately 1.52 GB of Free Disk Space
+      
+        Source code: 28 MB
+        Object code: 1470 MB
+        GCC front end: 50 MB
+      
+  
+  
+
+  
+
+  FreeBSD on x86 (Pentium and above)
+  
+    Approximately 918 MB of Free Disk Space
+    
+      Source code: 28 MB
+      Object code: 850 MB
+      GCC front end: 40 MB
+    
+  
+  
+
+  
+
+  MacOS X on PowerPC
+  
+    No native code generation
+    
Approximately 1.20 GB of Free Disk Space
       
-        Source code: 30 MB
-        Object code: 1000 MB
-        GCC front end: 210 MB
+        Source code: 28 MB
+        Object code: 1160 MB
+        GCC front end: 40 MB
       
-  
+  
 
+  
 
 
 The LLVM suite may compile on other platforms, but it is not
@@ -253,7 +284,6 @@
 
 
 
-
 
The remainder of this guide is meant to get you up and running with
 LLVM and to give you some basic information about the LLVM environment.
 A complete guide to installation is provided in the
@@ -348,22 +378,31 @@
 
 

 If you have the LLVM distribution, you will need to unpack it before you
-can begin to compile it.  LLVM is distributed as a set of three files.  Each
+can begin to compile it.  LLVM is distributed as a set of two files: the LLVM
+suite and the LLVM GCC front end compiled for your platform.  Each
 file is a TAR archive that is compressed with the gzip program.
 
 
- The three files are as follows:
+
 The files are as follows:
 

-    llvm.tar.gz
+    
llvm-1.1.tar.gz
     
This is the source code to the LLVM suite.
     
 
-    
cfrontend.sparc.tar.gz
+    
cfrontend-1.1.sparc-sun-solaris2.8.tar.gz
     
This is the binary release of the GCC front end for Solaris/Sparc.
     
 
-    
cfrontend.x86.tar.gz
+    
cfrontend-1.1.i686-redhat-linux-gnu.tar.gz
     
This is the binary release of the GCC front end for Linux/x86.
+    
+
+    
cfrontend-1.1.i386-unknown-freebsd5.1.tar.gz
+    
This is the binary release of the GCC front end for FreeBSD/x86.
+    
+
+    
cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz
+    
This is the binary release of the GCC front end for MacOS X/PPC.
 
 
 
@@ -391,6 +430,20 @@
 directory and fully populate it with the LLVM source code, Makefiles,
 test directories, and local copies of documentation files.
 
+
+If you want to get a specific release (as opposed to the most recent revision),
+you can specify a label.  The following releases have the following label:
+

+  
+  Release 1.1: RELEASE_11
+  
+
+  
+  Release 1.0: RELEASE_1
+  
+
+
+
 Note that the GCC front end is not included in the CVS repository. You
 should have downloaded the binary distribution for your platform.
 
@@ -412,12 +465,12 @@
 
 
   cd where-you-want-the-front-end-to-live
-  gunzip --stdout cfrontend.platform.tar.gz | tar -xvf
+  
gunzip --stdout cfrontend-version.platform.tar.gz | tar -xvf
       -

 

 
-If you are on a Sparc/Solaris machine, you will need to fix the header
-files:
+If you are using Solaris/Sparc or MacOS X/PPC, you will need to fix the
+header files:
 
 cd cfrontend/sparc

    ./fixheaders
@@ -443,7 +496,8 @@
 Once checked out from the CVS repository, the LLVM suite source code must be
 configured via the configure script.  This script sets variables in
 llvm/Makefile.config and llvm/include/Config/config.h.  It
-also populates OBJ_ROOT with the Makefiles needed to build LLVM.
+also populates OBJ_ROOT with the Makefiles needed to begin building
+LLVM.
 
 The following environment variables are used by the configure
 script to configure the build system:
@@ -477,7 +531,7 @@
   --with-llvmgccdir=LLVMGCCDIR
   

     Path to the location where the LLVM C front end binaries and
-    associated libraries will be installed.  This must be specified as an
+    associated libraries were installed.  This must be specified as an
     absolute pathname.
     
   
--enable-optimized
@@ -488,7 +542,8 @@
     
   
--enable-jit
   
LLVM_LIB_SEARCH_PATH
-lc
-lm
bytecode-libs
bytecode-libs
LLVM_LIB_SEARCH_PATH
/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs
/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs
longjmp
253.perlbmk
configure
configure.ac
configure
configure
llvm/projects
-licm
-basicaa
-basicaa
@@ -395,8 +399,27 @@
     }
 
format
format_arg
non_null
longjmp
longjmp
setjmp
The LLVM C++ front-end follows the unwind
longjmp

@@ -319,13 +319,13 @@
 variable and the address of its first element as the same. That tripped me up
 for a while until I realized that they really do differ .. by type.
 Remember that LLVM is a strongly typed language itself. Everything
-has a type.  The "type" of the global variable is [24 x int]*. That is, its
+has a type.  The "type" of the global variable is [24 x int]*. That is, it's
 a pointer to an array of 24 ints.  When you dereference that global variable with
 a single (0) index, you now have a "[24 x int]" type.  Although
 the pointer value of the dereferenced global and the address of the zero'th element
 in the array will be the same, they differ in their type. The zero'th element has
 type "int" while the pointer value has type "[24 x int]".
-Get this one aspect of LLVM right in your head and you'll save yourself
+
Get this one aspect of LLVM right in your head, and you'll save yourself
 a lot of compiler writing headaches down the road.
 

-    Compile the Just In Time (JIT) functionality.  This is not available
+    Compile the Just In Time (JIT) compiler functionality.  This is not
+    available
     on all platforms.  The default is dependent on platform, so it is best
     to explicitly enable it if you want it.
     
@@ -521,10 +576,10 @@
 LLVM_LIB_SEARCH_PATH environment variable in your startup scripts.
 This environment variable is used to locate "system" libraries like
 "-lc" and "-lm" when linking.  This variable should be set to
-the absolute path for the bytecode-libs subdirectory of the GCC front end
-install, or LLVMGCCDIR/bytecode-libs.  For example, one might set
+the absolute path of the bytecode-libs subdirectory of the GCC front
+end, or LLVMGCCDIR/bytecode-libs.  For example, one might set
 LLVM_LIB_SEARCH_PATH to
-/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the X86
+/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the x86
 version of the GCC front end on our research machines.
 
 
@@ -1091,7 +1146,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/HowToSubmitABug.html
diff -u llvm-www/releases/1.1/docs/HowToSubmitABug.html:1.1 llvm-www/releases/1.1/docs/HowToSubmitABug.html:1.2
--- llvm-www/releases/1.1/docs/HowToSubmitABug.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/HowToSubmitABug.html	Wed Dec 17 16:47:16 2003
@@ -311,7 +311,7 @@
   Chris Lattner
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/LLVMVsTheWorld.html
diff -u llvm-www/releases/1.1/docs/LLVMVsTheWorld.html:1.1 llvm-www/releases/1.1/docs/LLVMVsTheWorld.html:1.2
--- llvm-www/releases/1.1/docs/LLVMVsTheWorld.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/LLVMVsTheWorld.html	Wed Dec 17 16:47:16 2003
@@ -173,7 +173,7 @@
   Brian R. Gaeke
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/LangRef.html
diff -u llvm-www/releases/1.1/docs/LangRef.html:1.1 llvm-www/releases/1.1/docs/LangRef.html:1.2
--- llvm-www/releases/1.1/docs/LangRef.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/LangRef.html	Wed Dec 17 16:47:16 2003
@@ -1692,6 +1692,6 @@
 
 Chris Lattner
 The LLVM Compiler Infrastructure 

-Last modified: $Date: 2003/12/13 22:03:42 $ 
+Last modified: $Date: 2003/12/17 22:47:16 $ 
 
 


Index: llvm-www/releases/1.1/docs/OpenProjects.html
diff -u llvm-www/releases/1.1/docs/OpenProjects.html:1.1 llvm-www/releases/1.1/docs/OpenProjects.html:1.2
--- llvm-www/releases/1.1/docs/OpenProjects.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/OpenProjects.html	Wed Dec 17 16:47:16 2003
@@ -313,7 +313,7 @@
   Chris Lattner
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/13 22:03:42 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/Projects.html
diff -u llvm-www/releases/1.1/docs/Projects.html:1.1 llvm-www/releases/1.1/docs/Projects.html:1.2
--- llvm-www/releases/1.1/docs/Projects.html:1.1	Sat Dec 13 16:03:42 2003
+++ llvm-www/releases/1.1/docs/Projects.html	Wed Dec 17 16:47:16 2003
@@ -385,7 +385,7 @@
 
 John Criswell


 The LLVM Compiler Infrastructure

-Last modified: $Date: 2003/12/13 22:03:42 $
+Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/ReleaseNotes.html
diff -u llvm-www/releases/1.1/docs/ReleaseNotes.html:1.3 llvm-www/releases/1.1/docs/ReleaseNotes.html:1.4
--- llvm-www/releases/1.1/docs/ReleaseNotes.html:1.3	Sun Dec 14 16:36:45 2003
+++ llvm-www/releases/1.1/docs/ReleaseNotes.html	Wed Dec 17 16:47:16 2003
@@ -7,9 +7,7 @@
 
 
 
-
-  LLVM 1.1 Release Notes
-
+LLVM 1.1 Release Notes
  
 
   Introduction
@@ -18,7 +16,6 @@
   Installation Instructions
   Known Problems
   
-
     Known problems with the LLVM Core
     
Known problems with the C Front-end
     
Known problems with the C++ Front-end
@@ -30,7 +27,7 @@
 
 
 
-  Written by Chris Lattner

+  
Written by the LLVM team

 
 
 
@@ -43,10 +40,10 @@
 
 This document contains the release notes for the LLVM compiler
 infrastructure, release 1.1.  Here we describe the status of LLVM, including any
-known problems, and bug fixes from the previous release.  The most up-to-date
+known problems and bug fixes from the previous release.  The most up-to-date
 version of this document can be found on the LLVM 1.1 web site.  If you are
-not reading this on the LLVM web pages, you should probably go there, because
+not reading this on the LLVM web pages, you should probably go there because
 this document may be updated after the release.
 
 For more information about LLVM, including information about potentially more
@@ -72,10 +69,12 @@
 
 
This is the second public release of the LLVM compiler infrastructure.  This
 release is primarily a bugfix release, dramatically improving the C/C++
-front-end, and improving support for C++ in the LLVM core.  This release also
-includes a few new features, such as a simple profiler, support for Mac OS/X,
+front-end and improving support for C++ in the LLVM core.  This release also
+includes a few new features, such as a simple profiler, support for Mac OS X,
 better interoperability with external source bases, a new example language
-front-end, and improvements in a few optimizations.
+front-end, and improvements in a few optimizations.  The performance of several
+LLVM components has been improved, and several gratuitous type-safety issues in
+the C front-end have been fixed.
 
 At this time, LLVM is known to correctly compile and run all non-unwinding C
 & C++ SPEC CPU2000 benchmarks, the Olden benchmarks, and the Ptrdist
@@ -85,10 +84,10 @@
 
 
 
-The LLVM native code generators are very stable, but do not currently support
+The LLVM native code generators are very stable but do not currently support
 unwinding (exception throwing or longjmping), which prevent them from
 working with programs like the 253.perlbmk in SPEC CPU2000.  The C
-backend and the rest of LLVM does support these programs however, so you can
+backend and the rest of LLVM supports these programs, so you can
 still use LLVM with them.  Support for unwinding will be added in a future
 release.
 
@@ -102,15 +101,15 @@
 

 A new
-LLVM profiler, similar to gprof is available
+LLVM profiler, similar to gprof, is available.
 
-LLVM and the C/C++ front-end now compile on Mac OS/X!  Mac OS/X users can
+
LLVM and the C/C++ front-end now compile on Mac OS X!  Mac OS X users can
 now explore the LLVM optimizer with the C backend and interpreter.  Note that
-LLVM requires GCC 3.3 on Mac OS/X.
+LLVM requires GCC 3.3 on Mac OS X.
 
 LLVM has been moved
-into an 'llvm' C++ namespace, for easier integration with third-party
+into an 'llvm' C++ namespace for easier integration with third-party
 code. Note that due to lack of namespace support in GDB 5.x, you will probably
 want to upgrade to GDB 6 or better to debug LLVM code.
 
@@ -120,7 +119,8 @@
   
   
   New directories can be added to the source tree, and the build will
-  automatically pick them up (i.e. no need to re-run configure).
+  automatically pick them up (i.e. no need to edit configure.ac and
+  re-run configure).
   
 
   
@@ -132,7 +132,9 @@
 
 
A front-end for "Stacker" (a simple Forth-like language) is now
     included in the main LLVM tree.
-    Additionally, Reid Spencer, the author, contributed a document describing his experiences writing Stacker, and the language itself.  This document is invaluable for others writing front-ends targetting LLVM.
+    Additionally, Reid Spencer, the author, contributed a document
+    describing his experiences writing Stacker and the language itself.
+    This document is invaluable for others writing front-ends targetting LLVM.
 
 The configure script will now configure all projects placed in the
 llvm/projects directory.
@@ -143,12 +145,12 @@
 The -licm pass can now sink instructions out the bottom of loops
 in addition to being able to hoist them out the top.
 
-The -basicaa pass (the default alias analysis) has been upgraded
-to be significantly more 
+
The -basicaa pass (the default alias analysis pass) has been
+upgraded to be significantly more 
 precise.
 
 LLVM 1.1 implements a simple size optimization for LLVM bytecode files.  
-    This means that the 1.1 files are smaller than 1.0, but that 1.0 won't 
+    This means that the 1.1 files are smaller than 1.0, but LLVM 1.0 won't 
     read 1.1 bytecode files.
 
 The gccld program produces a runner script that includes command-line options to load the necessary shared objects.
@@ -185,15 +187,15 @@
 
 The C++ front-end now compiles functions to 
     use the linkonce linkage type 
-    more, giving the optimizer more freedom.
+    more, giving the optimizer more freedom.
 
 The C front-end now generates 
-    type-safe code in several cases that it did not before, which prevented
-    some important optimizations.
+    type-safe code in several cases that it did not before, allowing
+    optimization of code that could not be optimized previously.
 
 The LLVM build system has been taught to catch some common configuration 
     problems that caused it to get 
-    horribly confused before.
+    horribly confused.
 
 The LLVM header files are now 
     -Wold-style-cast clean.
@@ -210,15 +212,11 @@
 generated N^2 amounts of duplicated cleanup code in some cases.
 
 The JIT used to generate code for 
-    all functions pointed to by globals immediately, before the program 
-    started execution, but now it waits until the first time they are called to
+    all functions pointed to by globals before the program 
+    started execution.  Now, it waits until the first time the functions are
+    called to
     compile them.  This dramatically speeds up short runs of large C++ programs,
     which often have large numbers of functions pointed to by vtables.
-
-The C front-end had an inefficient
-    implementation of the LLVM type-system, which caused the C++ front-end
-    to be very slow in some cases.
-
 
 
 
@@ -238,7 +236,7 @@
 llvm-as crashes when labels are used in phi nodes
 [build problem] Callgraph.cpp not pulled in from libipa.a
 Variables in scope of output setjmp
-calls should be volatile (Note that this does not effect correctness on
+calls should be volatile (Note that this does not affect correctness on
 many platforms, such as X86).
 [X86] Emission of global bool initializers broken
 [gccld] The -r (relinking) option does not work correctly
@@ -274,7 +272,7 @@
 CFrontend crashes when compiling C99 compound expressions
 llvm-gcc infinite loops on "case MAXINT:"
 [C++] Catch blocks make unparsable labels
-[C++] Initializing array with constructable objects fail
+[C++] Initializing array with constructible objects fail
 llvm-gcc tries to add bools
 [c++] C++ Frontend lays out superclasses like anonymous bitfields!
 C front-end miscompiles unsigned enums whose LLVM types are signed
@@ -307,13 +305,13 @@
 
 
 LLVM has been extensively tested on Intel and AMD machines running Red
-Hat Linux and FreeBSD.  It has also been tested on Sun UltraSPARC workstations running Solaris 8.
+Hat Linux and has been tested on Sun UltraSPARC workstations running Solaris 8.
 Additionally, 
-LLVM works on Mac OS/X 10.3 and above, but only with the C backend or 
+LLVM works on Mac OS X 10.3 and above, but only with the C backend or 
 interpreter (no native backend for the PowerPC is available yet).
 The core LLVM infrastructure uses "autoconf" for portability, so hopefully we
 work on more platforms than that.  However, it is likely that we
-missed something, and that minor porting is required to get LLVM to work on 
+missed something and that minor porting is required to get LLVM to work on 
 new platforms.  We welcome portability patches and error messages.
 
 
@@ -329,7 +327,7 @@
 This section contains all known problems with the LLVM system, listed by
 component.  As new problems are discovered, they will be added to these
 sections.  If you run into a problem, please check the LLVM bug database, and submit a bug if
+href="http://llvm.cs.uiuc.edu/bugs/">LLVM bug database and submit a bug if
 there isn't already one.
 
 
@@ -364,10 +362,18 @@
 LLVM cannot handle structures with
 more than 256 elements.
 
-Bugs in 1.1 fixed in 1.2:
-
-
[licm] LICM promotes volatile memory locations to registers
+
+The gccld program 
+
+does not link objects/archives in the order specified on the command line.
+
+
 
+
+
+Tail duplication does not update SSA form correctly.
+
+
 
 
 
@@ -383,9 +389,7 @@
 
 
 
-
 
-
 C99 Variable sized arrays do not release stack memory when they go out of 
     scope.  Thus, the following program may run out of stack space:
 @@ -395,8 +399,27 @@
     }
 
 
-
+
+Initialization of global union variables can only be done
+with the largest
+union member.
+
 
+
+
+Functions marked "extern inline" are not compiled into LLVM with linkonce
+linkage.
+
+
+
+
+
+The memory management functions in the libc runtime
+need weak linkage so that they can be
+overridden.
+
+
+
 
 
 
@@ -428,11 +451,11 @@
     the following extensions are known to not be supported:
   

   Local Labels: Labels local to a block.
-  Labels as Values: Getting pointers to labels, and computed gotos.
+  Labels as Values: Getting pointers to labels and computed gotos.
   Nested Functions: As in Algol and Pascal, lexical scoping of functions.
   Constructing Calls: Dispatching a call to another function.
   Extended Asm: Assembler instructions with C expressions as operands.
-  Constraints: Constraints for asm operands
+  Constraints: Constraints for asm operands.
   Asm Labels: Specifying the assembler name to use for a C symbol.
   Explicit Reg Vars: Defining variables residing in specified registers.
   Return Address: Getting the return or frame address of a function.
@@ -445,7 +468,7 @@
   The following GCC extensions are partially supported.  An ignored
   attribute means that the LLVM compiler ignores the presence of the attribute,
   but the code should still work.  An unsupported attribute is one which is
-  ignored by the LLVM compiler, which will cause a different interpretation of
+  ignored by the LLVM compiler and will cause a different interpretation of
   the program.
 
   

@@ -455,7 +478,7 @@
 
   Function Attributes:
 
-      Declaring that functions have no side effects, or that they can never
+      Declaring that functions have no side effects or that they can never
       return.

 
       Supported: format, format_arg, non_null,
@@ -514,7 +537,8 @@
   
Subscripting: Any array can be subscripted, even if not an lvalue.
   Pointer Arith: Arithmetic on void-pointers and function pointers.
   Initializers: Non-constant initializers.
-  Compound Literals: Compound literals give structures, unions or arrays as values.
+  Compound Literals: Compound literals give structures, unions,
+or arrays as values.
   Designated Inits: Labeling elements of initializers.
   Cast to Union: Casting to union type from any member of the union.
   Case Ranges: `case 1 ... 9' and such.
@@ -546,7 +570,7 @@
 
 
 
-For this release, the C++ front-end is considered to be fully functional, but
+
For this release, the C++ front-end is considered to be fully functional but
 has not been tested as thoroughly as the C front-end.  It has been tested and 
 works for a number of non-trivial programs, but there may be lurking bugs.
 Please report any bugs or problems.
@@ -562,9 +586,14 @@
 
 
 The C++ front-end inherits all problems afflicting the C
-    front-end
-
+    front-end.
 
+
+
+Code is generated for empty classes.
+
+
+
 
 
 
@@ -584,7 +613,7 @@
 Destructors for local objects are not always run when a longjmp is
     performed. In particular, destructors for objects in the longjmping
     function and in the setjmp receiver function may not be run.
-    Objects in intervening stack frames will be destroyed however (which is
+    Objects in intervening stack frames will be destroyed, however (which is
     better than most compilers).
 
 The LLVM C++ front-end follows the unwind instruction, so code that throws a C++ exception
 or calls the C longjmp function will abort.
 
+
+
+The llc program can crash on legal code.
+
+
 
 
 
@@ -673,7 +707,7 @@
 
 
 A wide variety of additional information is available on the LLVM web page,
-including mailing lists publications describing algorithms and components
+including mailing lists and publications describing algorithms and components
 implemented in LLVM.  The web page also contains versions of the API
 documentation which is up-to-date with the CVS version of the source code.  You
 can access versions of these documents specific to this release by going into
@@ -695,7 +729,7 @@
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" />
 
   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/14 22:36:45 $
+  Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 


Index: llvm-www/releases/1.1/docs/Stacker.html
diff -u llvm-www/releases/1.1/docs/Stacker.html:1.1 llvm-www/releases/1.1/docs/Stacker.html:1.2
--- llvm-www/releases/1.1/docs/Stacker.html:1.1	Sat Dec 13 16:03:43 2003
+++ llvm-www/releases/1.1/docs/Stacker.html	Wed Dec 17 16:47:16 2003
@@ -61,7 +61,7 @@
 about LLVM through the experience of creating a simple programming language
 named Stacker.  Stacker was invented specifically as a demonstration of
 LLVM. The emphasis in this document is not on describing the
-intricacies of LLVM itself, but on how to use it to build your own
+intricacies of LLVM itself but on how to use it to build your own
 compiler system.
 
 
@@ -70,7 +70,7 @@
 Amongst other things, LLVM is a platform for compiler writers.
 Because of its exceptionally clean and small IR (intermediate
 representation), compiler writing with LLVM is much easier than with
-other system. As proof, the author of Stacker wrote the entire
+other systems. As proof, the author of Stacker wrote the entire
 compiler (language definition, lexer, parser, code generator, etc.) in
 about four days! That's important to know because it shows 
 how quickly you can get a new
@@ -78,11 +78,11 @@
 language the author ever created using LLVM. The learning curve is 
 included in that four days.
 The language described here, Stacker, is Forth-like. Programs
-are simple collections of word definitions and the only thing definitions
+are simple collections of word definitions, and the only thing definitions
 can do is manipulate a stack or generate I/O.  Stacker is not a "real" 
-programming language; its very simple.  Although it is computationally 
+programming language; it's very simple.  Although it is computationally 
 complete, you wouldn't use it for your next big project. However, 
-the fact that it is complete, its simple, and it doesn't have 
+the fact that it is complete, it's simple, and it doesn't have 
 a C-like syntax make it useful for demonstration purposes. It shows
 that LLVM could be applied to a wide variety of languages.
 The basic notions behind stacker is very simple. There's a stack of 
@@ -96,11 +96,11 @@
 : MAIN hello_world ;

 This has two "definitions" (Stacker manipulates words, not
 functions and words have definitions): MAIN and 
-hello_world. The MAIN definition is standard, it
+hello_world. The MAIN definition is standard; it
 tells Stacker where to start. Here, MAIN is defined to 
 simply invoke the word hello_world. The
 hello_world definition tells stacker to push the 
-"Hello, World!" string onto the stack, print it out 
+"Hello, World!" string on to the stack, print it out 
 (>s), pop it off the stack (DROP), and
 finally print a carriage return (CR). Although 
 hello_world uses the stack, its net effect is null. Well
@@ -124,7 +124,7 @@
 
Although I knew that LLVM uses a Single Static Assignment (SSA) format, 
 it wasn't obvious to me how prevalent this idea was in LLVM until I really
 started using it.  Reading the 
-Programmer's Manual and Language Reference
+Programmer's Manual and Language Reference,
 I noted that most of the important LLVM IR (Intermediate Representation) C++ 
 classes were derived from the Value class. The full power of that simple
 design only became fully understood once I started constructing executable
@@ -200,7 +200,7 @@
 

     Create your blocks early. While writing your compiler, you 
     will encounter several situations where you know apriori that you will
-    need several blocks. For example, if-then-else, switch, while and for
+    need several blocks. For example, if-then-else, switch, while, and for
     statements in C/C++ all need multiple blocks for expression in LVVM. 
     The rule is, create them early.
     
Terminate your blocks early. This just reduces the chances 
@@ -261,15 +261,15 @@
 the instructions for the "then" and "else" parts. They would use the third part
 of the idiom almost exclusively (inserting new instructions before the 
 terminator). Furthermore, they could even recurse back to handle_if 
-should they encounter another if/then/else statement and it will just work.
+should they encounter another if/then/else statement, and it will just work.
 Note how cleanly this all works out. In particular, the push_back methods on
 the BasicBlock's instruction list. These are lists of type 
 Instruction which also happen to be Values. To create 
-the "if" branch we merely instantiate a BranchInst that takes as 
+the "if" branch, we merely instantiate a BranchInst that takes as 
 arguments the blocks to branch to and the condition to branch on. The blocks
 act like branch labels! This new BranchInst terminates
 the BasicBlock provided as an argument. To give the caller a way
-to keep inserting after calling handle_if we create an "exit" block
+to keep inserting after calling handle_if, we create an "exit" block
 which is returned to the caller.  Note that the "exit" block is used as the 
 terminator for both the "then" and the "else" blocks. This guarantees that no
 matter what else "handle_if" or "fill_in" does, they end up at the "exit" block.
@@ -283,7 +283,7 @@
 method on the various lists. This is so common that it is worth mentioning.
 The "push_back" inserts a value into an STL list, vector, array, etc. at the
 end. The method might have also been named "insert_tail" or "append".
-Althought I've used STL quite frequently, my use of push_back wasn't very
+Although I've used STL quite frequently, my use of push_back wasn't very
 high in other programs. In LLVM, you'll use it all the time.
 
 
@@ -292,17 +292,17 @@
 
 
 It took a little getting used to and several rounds of postings to the LLVM
-mail list to wrap my head around this instruction correctly. Even though I had
+mailing list to wrap my head around this instruction correctly. Even though I had
 read the Language Reference and Programmer's Manual a couple times each, I still
 missed a few very key points:
 
 
-    GetElementPtrInst gives you back a Value for the last thing indexed
+    
GetElementPtrInst gives you back a Value for the last thing indexed.
     
All global variables in LLVM  are pointers.
     
Pointers must also be dereferenced with the GetElementPtrInst instruction.
 
 This means that when you look up an element in the global variable (assuming
-its a struct or array), you must deference the pointer first! For many
+it's a struct or array), you must deference the pointer first! For many
 things, this leads to the idiom:
 
 
@@ -319,13 +319,13 @@
 variable and the address of its first element as the same. That tripped me up
 for a while until I realized that they really do differ .. by type.
 Remember that LLVM is a strongly typed language itself. Everything
-has a type.  The "type" of the global variable is [24 x int]*. That is, its
+has a type.  The "type" of the global variable is [24 x int]*. That is, it's
 a pointer to an array of 24 ints.  When you dereference that global variable with
 a single (0) index, you now have a "[24 x int]" type.  Although
 the pointer value of the dereferenced global and the address of the zero'th element
 in the array will be the same, they differ in their type. The zero'th element has
 type "int" while the pointer value has type "[24 x int]".
-Get this one aspect of LLVM right in your head and you'll save yourself
+
Get this one aspect of LLVM right in your head, and you'll save yourself
 a lot of compiler writing headaches down the road.
 

 
@@ -334,7 +334,7 @@
 Linkage types in LLVM can be a little confusing, especially if your compiler
 writing mind has affixed very hard concepts to particular words like "weak",
 "external", "global", "linkonce", etc. LLVM does not use the precise
-definitions of say ELF or GCC even though they share common terms. To be fair,
+definitions of, say, ELF or GCC, even though they share common terms. To be fair,
 the concepts are related and similar but not precisely the same. This can lead
 you to think you know what a linkage type represents but in fact it is slightly
 different. I recommend you read the 
@@ -342,10 +342,10 @@
 carefully. Then, read it again.

 
Here are some handy tips that I discovered along the way:
 

-    Unitialized means external. That is, the symbol is declared in the current
+    
Uninitialized means external. That is, the symbol is declared in the current
     module and can be used by that module but it is not defined by that module.
     Setting an initializer changes a global's linkage type from whatever it was
-    to a normal, defind global (not external). You'll need to call the setLinkage()
+    to a normal, defined global (not external). You'll need to call the setLinkage()
     method to reset it if you specify the initializer after the GlobalValue has been
     constructed. This is important for LinkOnce and Weak linkage types. 
     
Appending linkage can be used to keep track of compilation information at
@@ -362,7 +362,7 @@
 functions in the LLVM IR that make things easier. Here's what I learned: 
 

  Constants are Values like anything else and can be operands of instructions
- Integer constants, frequently needed can be created using the static "get"
+ 
Integer constants, frequently needed, can be created using the static "get"
  methods of the ConstantInt, ConstantSInt, and ConstantUInt classes. The nice thing
  about these is that you can "get" any kind of integer quickly.
  There's a special method on Constant class which allows you to get the null 
@@ -379,14 +379,14 @@
 proceeding, a few words about the stack are in order. The stack is simply
 a global array of 32-bit integers or pointers. A global index keeps track
 of the location of the top of the stack. All of this is hidden from the 
-programmer but it needs to be noted because it is the foundation of the 
+programmer, but it needs to be noted because it is the foundation of the 
 conceptual programming model for Stacker. When you write a definition,
 you are, essentially, saying how you want that definition to manipulate
 the global stack.
 Manipulating the stack can be quite hazardous. There is no distinction
 given and no checking for the various types of values that can be placed
 on the stack. Automatic coercion between types is performed. In many 
-cases this is useful. For example, a boolean value placed on the stack
+cases, this is useful. For example, a boolean value placed on the stack
 can be interpreted as an integer with good results. However, using a
 word that interprets that boolean value as a pointer to a string to
 print out will almost always yield a crash. Stacker simply leaves it
@@ -406,9 +406,9 @@
 
So, your typical definition will have the form:
 : name ... ;
 The name is up to you but it must start with a letter and contain
-only letters numbers and underscore. Names are case sensitive and must not be
+only letters, numbers, and underscore. Names are case sensitive and must not be
 the same as the name of a built-in word. The ... is replaced by
-the stack manipulting words that you wish define name as. 

+the stack manipulating words that you wish to define name as. 

 
 
 
Comments
@@ -429,12 +429,12 @@
 
 Literals
 
-    There are three kinds of literal values in Stacker. Integer, Strings,
+    
There are three kinds of literal values in Stacker: Integers, Strings,
     and Booleans. In each case, the stack operation is to simply push the
-    value onto the stack. So, for example:

+    value on to the stack. So, for example:

      42 " is the answer." TRUE 

-    will push three values onto the stack: the integer 42, the
-    string " is the answer." and the boolean TRUE.
+    will push three values on to the stack: the integer 42, the
+    string " is the answer.", and the boolean TRUE.
 
 
 Words
@@ -464,20 +464,20 @@
 The built-in words of the Stacker language are put in several groups 
 depending on what they do. The groups are as follows:
  
-    LogicalThese words provide the logical operations for
+    
Logical: These words provide the logical operations for
     comparing stack operands.
The words are: < > <= >= 
     = <> true false.
-    BitwiseThese words perform bitwise computations on 
+    
Bitwise: These words perform bitwise computations on 
     their operands. 
 The words are: << >> XOR AND NOT
-    ArithmeticThese words perform arithmetic computations on
+    
Arithmetic: These words perform arithmetic computations on
     their operands. 
 The words are: ABS NEG + - * / MOD */ ++ -- MIN MAX
-    StackThese words manipulate the stack directly by moving
+    
Stack: These words manipulate the stack directly by moving
     its elements around.
 The words are: DROP DUP SWAP OVER ROT DUP2 DROP2 PICK TUCK
-    MemoryThese words allocate, free and manipulate memory
+    
Memory: These words allocate, free, and manipulate memory
     areas outside the stack.
The words are: MALLOC FREE GET PUT
-    ControlThese words alter the normal left to right flow
+    
Control: These words alter the normal left to right flow
     of execution.
The words are: IF ELSE ENDIF WHILE END RETURN EXIT RECURSE
-    I/O These words perform output on the standard output
+    
I/O: These words perform output on the standard output
     and input on the standard input. No other I/O is possible in Stacker.
     
The words are: SPACE TAB CR >s >d >c <s <d <c.
 
@@ -554,12 +554,12 @@
 FALSE
     FALSE
      -- b
-    The boolean value FALSE (0) is pushed onto the stack.
+    The boolean value FALSE (0) is pushed on to the stack.
 
 TRUE
     TRUE
      -- b
-    The boolean value TRUE (-1) is pushed onto the stack.
+    The boolean value TRUE (-1) is pushed on to the stack.
 
 BITWISE OPERATIONS
 WordNameOperationDescription
@@ -604,75 +604,75 @@
     ABS
     w -- |w|
     One value s popped off the stack; its absolute value is computed
-    and then pushed onto the stack. If w1 is -1 then w2 is 1. If w1 is
+    and then pushed on to the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is also 1.
 
 NEG
     NEG
     w -- -w
     One value is popped off the stack which is negated and then
-    pushed back onto the stack. If w1 is -1 then w2 is 1. If w1 is
+    pushed back on to the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is -1.
 
  + 
     ADD
     w1 w2 -- w2+w1
     Two values are popped off the stack. Their sum is pushed back
-    onto the stack
+    on to the stack
 
  - 
     SUB
     w1 w2 -- w2-w1
     Two values are popped off the stack. Their difference is pushed back
-    onto the stack
+    on to the stack
 
  * 
     MUL
     w1 w2 -- w2*w1
     Two values are popped off the stack. Their product is pushed back
-    onto the stack
+    on to the stack
 
  / 
     DIV
     w1 w2 -- w2/w1
     Two values are popped off the stack. Their quotient is pushed back
-    onto the stack
+    on to the stack
 
 MOD
     MOD
     w1 w2 -- w2%w1
     Two values are popped off the stack. Their remainder after division
-    of w1 by w2 is pushed back onto the stack
+    of w1 by w2 is pushed back on to the stack
 
  */ 
     STAR_SLAH
     w1 w2 w3 -- (w3*w2)/w1
     Three values are popped off the stack. The product of w1 and w2 is
-    divided by w3. The result is pushed back onto the stack.
+    divided by w3. The result is pushed back on to the stack.
 
  ++ 
     INCR
     w -- w+1
     One value is popped off the stack. It is incremented by one and then
-    pushed back onto the stack.
+    pushed back on to the stack.
 
  -- 
     DECR
     w -- w-1
     One value is popped off the stack. It is decremented by one and then
-    pushed back onto the stack.
+    pushed back on to the stack.
 
 MIN
     MIN
     w1 w2 -- (w2<w1?w2:w1)
     Two values are popped off the stack. The larger one is pushed back
-    onto the stack.
+    on to the stack.
 
 MAX
     MAX
     w1 w2 -- (w2>w1?w2:w1)
     Two values are popped off the stack. The larger value is pushed back
-	onto the stack.
+	on to the stack.
 
 STACK MANIPULATION OPERATIONS
 WordNameOperationDescription
@@ -703,13 +703,13 @@
 DUP
     DUP
     w1 -- w1 w1
-    One value is popped off the stack. That value is then pushed onto
-	the stack twice to duplicate the top stack vaue.
+    One value is popped off the stack. That value is then pushed on to
+	the stack twice to duplicate the top stack value.
 
 DUP2
     DUP2
     w1 w2 -- w1 w2 w1 w2
-    The top two values on the stack are duplicated. That is, two vaues
+    The top two values on the stack are duplicated. That is, two values
 	are popped off the stack. They are alternately pushed back on the
 	stack twice each.
 
@@ -717,7 +717,7 @@
     SWAP
     w1 w2 -- w2 w1
     The top two stack items are reversed in their order. That is, two
-	values are popped off the stack and pushed back onto the stack in
+	values are popped off the stack and pushed back on to the stack in
 	the opposite order they were popped.
 
 SWAP2
@@ -725,27 +725,27 @@
     w1 w2 w3 w4 -- w3 w4 w2 w1
     The top four stack items are swapped in pairs. That is, two values
 	are popped and retained. Then, two more values are popped and retained.
-	The values are pushed back onto the stack in the reverse order but
+	The values are pushed back on to the stack in the reverse order but
 	in pairs.
 
 OVER
     OVER
     w1 w2-- w1 w2 w1
     Two values are popped from the stack. They are pushed back
-	onto the stack in the order w1 w2 w1. This seems to cause the
+	on to the stack in the order w1 w2 w1. This seems to cause the
 	top stack element to be duplicated "over" the next value.
 
 OVER2
     OVER2
     w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2
-    The third and fourth values on the stack are replicated onto the
+    The third and fourth values on the stack are replicated on to the
 	top of the stack
 
 ROT
     ROT
     w1 w2 w3 -- w2 w3 w1
     The top three values are rotated. That is, three value are popped
-	off the stack. They are pushed back onto the stack in the order
+	off the stack. They are pushed back on to the stack in the order
 	w1 w3 w2.
 
 ROT2
@@ -822,7 +822,7 @@
     One value is popped off the stack. The value is used as the size
 	of a memory block to allocate. The size is in bytes, not words.
         The memory allocation is completed and the address of the memory
-	block is pushed onto the stack.
+	block is pushed on to the stack.
 
 FREE
     FREE
@@ -911,7 +911,7 @@
     The boolean value on the top of the stack is examined. If it is non-zero then the 
 	"words..." between WHILE and END are executed. Execution then begins again at the WHILE where another
 	boolean is popped off the stack. To prevent this operation from eating up the entire
-	stack, you should push onto the stack (just before the END) a boolean value that indicates
+	stack, you should push on to the stack (just before the END) a boolean value that indicates
 	whether to terminate. Note that since booleans and integers can be coerced you can
 	use the following "for loop" idiom:

 	(push count) WHILE (words...) -- END

@@ -960,19 +960,19 @@
     IN_STR
      -- s 
     A string is read from the input via the scanf(3) format string " %as". The
-	resulting string is pushed onto the stack.
+	resulting string is pushed on to the stack.
 
 <d
     IN_STR
      -- w 
     An integer is read from the input via the scanf(3) format string " %d". The
-	resulting value is pushed onto the stack
+	resulting value is pushed on to the stack
 
 <c
     IN_CHR
      -- w 
     A single character is read from the input via the scanf(3) format string 
-	" %c". The value is converted to an integer and pushed onto the stack.
+	" %c". The value is converted to an integer and pushed on to the stack.
 
 DUMP
     DUMP
@@ -989,9 +989,9 @@
 The following fully documented program highlights many features of both
 the Stacker language and what is possible with LLVM. The program has two modes
 of operations. If you provide numeric arguments to the program, it checks to see
-if those arguments are prime numbers, prints out the results. Without any 
-aruments, the program prints out any prime numbers it finds between 1 and one 
-million (there's a log of them!). The source code comments below tell the 
+if those arguments are prime numbers and prints out the results. Without any 
+arguments, the program prints out any prime numbers it finds between 1 and one 
+million (there's a lot of them!). The source code comments below tell the 
 remainder of the story.
 
 
@@ -1015,7 +1015,7 @@
 : exit_loop FALSE;
     
 ################################################################################
-# This definition tryies an actual division of a candidate prime number. It
+# This definition tries an actual division of a candidate prime number. It
 # determines whether the division loop on this candidate should continue or
 # not.
 # STACK<:
@@ -1075,7 +1075,7 @@
 # STACK<:
 #   p - the prime number to check
 # STACK>:
-#   yn - boolean indiating if its a prime or not
+#   yn - boolean indicating if its a prime or not
 #   p - the prime number checked
 ################################################################################
 : try_harder
@@ -1248,7 +1248,7 @@
 under the LLVM "projects" directory. You will need to obtain the LLVM sources
 to find it (either via anonymous CVS or a tarball. See the 
 Getting Started document).
-Under the "projects" directory there is a directory named "stacker". That
+
Under the "projects" directory there is a directory named "Stacker". That
 directory contains everything, as follows:
 
     lib - contains most of the source code
@@ -1301,7 +1301,7 @@
 definitions, the ROLL word is not implemented. This word was left out of 
 Stacker on purpose so that it can be an exercise for the student.  The exercise 
 is to implement the ROLL functionality (in your own workspace) and build a test 
-program for it.  If you can implement ROLL you understand Stacker and probably 
+program for it.  If you can implement ROLL, you understand Stacker and probably 
 a fair amount about LLVM since this is one of the more complicated Stacker 
 operations. The work will almost be completely limited to the 
 compiler.  
@@ -1326,7 +1326,7 @@
     emitted currently is somewhat wasteful. It gets cleaned up a lot by existing
     passes but more could be done.
     Add -O -O1 -O2 and -O3 optimization switches to the compiler driver to
-    allow LLVM optimization without using "opt"
+    allow LLVM optimization without using "opt."
     Make the compiler driver use the LLVM linking facilities (with IPO) before 
     depending on GCC to do the final link.
     Clean up parsing. It doesn't handle errors very well.
@@ -1346,6 +1346,6 @@
 
 Reid Spencer
 The LLVM Compiler Infrastructure 
-
Last modified: $Date: 2003/12/13 22:03:43 $ 
+
Last modified: $Date: 2003/12/17 22:47:16 $ 
 
 


Index: llvm-www/releases/1.1/docs/TestingGuide.html
diff -u llvm-www/releases/1.1/docs/TestingGuide.html:1.1 llvm-www/releases/1.1/docs/TestingGuide.html:1.2
--- llvm-www/releases/1.1/docs/TestingGuide.html:1.1	Sat Dec 13 16:03:43 2003
+++ llvm-www/releases/1.1/docs/TestingGuide.html	Wed Dec 17 16:47:16 2003
@@ -422,7 +422,7 @@
 John T. Criswell
 The LLVM Compiler Infrastructure
 

-Last modified: $Date: 2003/12/13 22:03:43 $
+Last modified: $Date: 2003/12/17 22:47:16 $
 
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 16:50:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 16:50:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/index.html
Message-ID: <200312172249.QAA01517@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

index.html updated: 1.10 -> 1.11

---
Log message:

Blessing LLVM 1.1.



---
Diffs of the changes:  (+6 -1)

Index: llvm-www/releases/index.html
diff -u llvm-www/releases/index.html:1.10 llvm-www/releases/index.html:1.11
--- llvm-www/releases/index.html:1.10	Fri Dec 12 16:09:41 2003
+++ llvm-www/releases/index.html	Wed Dec 17 16:48:50 2003
@@ -6,7 +6,7 @@
 
 
 
-  LLVM Release 1.0
+  LLVM Public Release
 
 
 
@@ -39,6 +39,10 @@
 
 
 
+The 1.1 release (17 Dec 2003) is available for
+download. You can read the Release Notes
+here.
+
 The 1.0 release (24 Oct 2003) is available for
 download. You can read the Release Notes
 here.
@@ -54,6 +58,7 @@
 the documentation here:
 
 
+  Complete documentation for release 1.1.
   Complete documentation for release 1.0.
   Up-to-date LLVM documentation
       on the LLVM Web page.




From criswell at cs.uiuc.edu  Wed Dec 17 16:52:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 16:52:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/download.html
Message-ID: <200312172251.QAA01637@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

download.html updated: 1.7 -> 1.8

---
Log message:

Blessing LLVM 1.1.



---
Diffs of the changes:  (+29 -4)

Index: llvm-www/releases/download.html
diff -u llvm-www/releases/download.html:1.7 llvm-www/releases/download.html:1.8
--- llvm-www/releases/download.html:1.7	Fri Dec 12 16:16:10 2003
+++ llvm-www/releases/download.html	Wed Dec 17 16:50:50 2003
@@ -1,12 +1,12 @@
 
 
   
-  LLVM 1.0 Download Page
+  LLVM Download Page
 
 
 
 
-  LLVM 1.0 Download Page
+  LLVM Download Page
 
 
 
@@ -20,11 +20,36 @@
 
 
 
-
+Download
Download LLVM 1.1
 
 
 
-Please read the Release Notes before
+Please read the Release Notes before
+downloading:
+
+
+	LLVM source code (6.2M)
+
+  GCC Front End
+      Binaries for Linux/x86 (15.3M)
+
+  GCC Front End
+      Binaries for FreeBSD/x86 (14.4M)
+
+  GCC Front End Binaries
+      for Solaris/Sparc (43.3M)
+
+	GCC Front End Source Code (36.1M)
+
+
+
+
+
+
+Download LLVM 1.0
+
+
+Please read the Release Notes before
 downloading:
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 16:57:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 16:57:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/download.html
Message-ID: <200312172256.QAA01795@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

download.html updated: 1.8 -> 1.9

---
Log message:

Bless MacOS X too.
:)



---
Diffs of the changes:  (+3 -0)

Index: llvm-www/releases/download.html
diff -u llvm-www/releases/download.html:1.8 llvm-www/releases/download.html:1.9
--- llvm-www/releases/download.html:1.8	Wed Dec 17 16:50:50 2003
+++ llvm-www/releases/download.html	Wed Dec 17 16:56:04 2003
@@ -39,6 +39,9 @@
   GCC Front End Binaries
       for Solaris/Sparc (43.3M)
 
+  GCC Front End Binaries
+      for MacOS X/PowerPC (43.3M)
+
 	GCC Front End Source Code (36.1M)
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 17:03:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 17:03:01 2003
Subject: [llvm-commits] CVS: llvm-www/header.incl www-index.html
Message-ID: <200312172302.RAA01940@choi.cs.uiuc.edu>


Changes in directory llvm-www:

header.incl updated: 1.6 -> 1.7
www-index.html updated: 1.92 -> 1.93

---
Log message:

Update to 1.1.



---
Diffs of the changes:  (+4 -4)

Index: llvm-www/header.incl
diff -u llvm-www/header.incl:1.6 llvm-www/header.incl:1.7
--- llvm-www/header.incl:1.6	Mon Dec 15 14:14:58 2003
+++ llvm-www/header.incl	Wed Dec 17 17:02:40 2003
@@ -32,7 +32,7 @@
 Download!
 
 Download now:
-LLVM 1.0

+LLVM 1.1

 
 

 Try the:

@@ -40,7 +40,7 @@
 

 
 View the open-source

-license
+license
 
 
 



Index: llvm-www/www-index.html
diff -u llvm-www/www-index.html:1.92 llvm-www/www-index.html:1.93
--- llvm-www/www-index.html:1.92	Fri Dec 12 12:23:51 2003
+++ llvm-www/www-index.html	Wed Dec 17 17:02:40 2003
@@ -101,9 +101,9 @@
 Public LLVM Release!
 
   
-  The LLVM 1.0 is now available for
+  The LLVM 1.1 is now available for
   download!  LLVM is publicly available under the OSI-certified University of Illinois Open-Source
+  href="releases/1.1/LICENSE.TXT">University of Illinois Open-Source
   License.  See answers to common
   licensing questions.

 




From brukman at cs.uiuc.edu  Wed Dec 17 17:12:09 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Wed Dec 17 17:12:09 2003
Subject: [llvm-commits] CVS: llvm/docs/AliasAnalysis.html
Message-ID: <200312172310.RAA30705@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

AliasAnalysis.html updated: 1.13 -> 1.14

---
Log message:

Fix the links to match our 'llvm' namespace in doxygen-generated docs.


---
Diffs of the changes:  (+5 -5)

Index: llvm/docs/AliasAnalysis.html
diff -u llvm/docs/AliasAnalysis.html:1.13 llvm/docs/AliasAnalysis.html:1.14
--- llvm/docs/AliasAnalysis.html:1.13	Wed Dec 17 12:46:49 2003
+++ llvm/docs/AliasAnalysis.html	Wed Dec 17 17:10:49 2003
@@ -65,7 +65,7 @@
 pointers do point to the same object, might point to the same object, or are
 known not to point to the same object.
 
-The AliasAnalysis class is the
+
The AliasAnalysis class is the
 centerpiece of the LLVM Alias Analysis related infrastructure.  This class is
 the common interface between clients of alias analysis information and the
 implementations providing it.  In addition to simple alias analysis information,
@@ -89,7 +89,7 @@
 
 

 
-The AliasAnalysis class
+
The AliasAnalysis class
 defines the interface that Alias Analysis implementations should support.  This
 class exports two important enums: AliasResult and
 ModRefResult which represent the result of an alias query or a mod/ref
@@ -277,7 +277,7 @@
 
 

 
-All of the AliasAnalysis
+
All of the AliasAnalysis
 virtual methods default to providing conservatively correct information
 (returning "May" Alias and "Mod/Ref" for alias and mod/ref queries
 respectively).  Depending on the capabilities of the analysis you are
@@ -355,7 +355,7 @@
 
 
Many transformations need information about alias sets that are active
 in some scope, rather than information about pairwise aliasing.  The AliasSetTracker class is used
+href="/doxygen/classllvm_1_1AliasSetTracker.html">AliasSetTracker class is used
 to efficiently build these Alias Sets from the pairwise alias analysis
 information provided by the AliasAnalysis interface.
 
@@ -486,7 +486,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/17 18:46:49 $
+  Last modified: $Date: 2003/12/17 23:10:49 $
 
 
 




From lattner at cs.uiuc.edu  Wed Dec 17 17:48:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Wed Dec 17 17:48:01 2003
Subject: [llvm-commits] CVS: llvm-www/header.incl
Message-ID: <200312172347.RAA31397@zion.cs.uiuc.edu>


Changes in directory llvm-www:

header.incl updated: 1.7 -> 1.8

---
Log message:

Add link to new status update


---
Diffs of the changes:  (+1 -0)

Index: llvm-www/header.incl
diff -u llvm-www/header.incl:1.7 llvm-www/header.incl:1.8
--- llvm-www/header.incl:1.7	Wed Dec 17 17:02:40 2003
+++ llvm-www/header.incl	Wed Dec 17 17:46:57 2003
@@ -48,6 +48,7 @@
 Status Updates
 
   
+  December 17, 2003

   November 18, 2003

   October 7, 2003

   September 10, 2003





From lattner at cs.uiuc.edu  Wed Dec 17 17:51:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Wed Dec 17 17:51:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/announcement.txt
Message-ID: <200312172350.RAA31671@zion.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1:

announcement.txt updated: 1.4 -> 1.5

---
Log message:

Add link


---
Diffs of the changes:  (+2 -2)

Index: llvm-www/releases/1.1/announcement.txt
diff -u llvm-www/releases/1.1/announcement.txt:1.4 llvm-www/releases/1.1/announcement.txt:1.5
--- llvm-www/releases/1.1/announcement.txt:1.4	Wed Dec 17 15:04:43 2003
+++ llvm-www/releases/1.1/announcement.txt	Wed Dec 17 17:50:16 2003
@@ -22,7 +22,7 @@
 
   This release is primarily a bugfix release, dramatically improving the C/C++
   front-end, and improving support for C++ in the LLVM core. This release also
-  includes a few new features, such as a simple profiler, support for Mac OS/X,
+  includes a few new features, such as a simple profiler, support for Mac OS X,
   better interoperability with external source bases, a new example language
   front-end, and improvements in a few optimizations.  This release also
   features several substantial speedups and implementation of missing features.
@@ -32,7 +32,7 @@
 
   For an easier to read set of changes, please see the status updates:
    http://mail.cs.uiuc.edu/pipermail/llvm-announce/2003-November/000003.html
-   ***** Second 1.1 status update *****
+   http://mail.cs.uiuc.edu/pipermail/llvm-announce/2003-December/000004.html
 
 HOW DO I GET IT?
 




From criswell at cs.uiuc.edu  Wed Dec 17 19:22:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 19:22:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/download.html
Message-ID: <200312180121.TAA02337@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

download.html updated: 1.9 -> 1.10

---
Log message:

Update the sizes to be more accurate.



---
Diffs of the changes:  (+5 -5)

Index: llvm-www/releases/download.html
diff -u llvm-www/releases/download.html:1.9 llvm-www/releases/download.html:1.10
--- llvm-www/releases/download.html:1.9	Wed Dec 17 16:56:04 2003
+++ llvm-www/releases/download.html	Wed Dec 17 19:21:03 2003
@@ -31,18 +31,18 @@
 	LLVM source code (6.2M)
 
   GCC Front End
-      Binaries for Linux/x86 (15.3M)
+      Binaries for Linux/x86 (14.5M)
 
   GCC Front End
-      Binaries for FreeBSD/x86 (14.4M)
+      Binaries for FreeBSD/x86 (13.9M)
 
   GCC Front End Binaries
-      for Solaris/Sparc (43.3M)
+      for Solaris/Sparc (16.1M)
 
   GCC Front End Binaries
-      for MacOS X/PowerPC (43.3M)
+      for MacOS X/PowerPC (12.2M)
 
-	GCC Front End Source Code (36.1M)
+	GCC Front End Source Code (32.4M)
 

 
 





From criswell at cs.uiuc.edu  Wed Dec 17 19:38:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 19:38:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/docs/CommandGuide/index.html
Message-ID: <200312180137.TAA02490@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1/docs/CommandGuide:

index.html updated: 1.1 -> 1.2

---
Log message:

Update command guide.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/1.1/docs/CommandGuide/index.html
diff -u llvm-www/releases/1.1/docs/CommandGuide/index.html:1.1 llvm-www/releases/1.1/docs/CommandGuide/index.html:1.2
--- llvm-www/releases/1.1/docs/CommandGuide/index.html:1.1	Sat Dec 13 16:26:47 2003
+++ llvm-www/releases/1.1/docs/CommandGuide/index.html	Wed Dec 17 19:37:38 2003
@@ -116,7 +116,7 @@
 

 Maintained by the 
 LLVM Team.

-Last modified: $Date: 2003/12/13 22:26:47 $
+Last modified: $Date: 2003/12/18 01:37:38 $
 
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 19:41:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 19:41:01 2003
Subject: [llvm-commits] CVS: llvm-www/Features.html
Message-ID: <200312180140.TAA02614@choi.cs.uiuc.edu>


Changes in directory llvm-www:

Features.html added (r1.1)

---
Log message:

List of LLVM Features.



---
Diffs of the changes:  (+172 -0)

Index: llvm-www/Features.html
diff -c /dev/null llvm-www/Features.html:1.1
*** /dev/null	Wed Dec 17 19:40:42 2003
--- llvm-www/Features.html	Wed Dec 17 19:40:32 2003
***************
*** 0 ****
--- 1,172 ----
+ 
+ LLVM Features
+ 
+ 
+ The public release of LLVM is intended to be a fully functional release of our
+ compiler system for C and C++.  As such, it includes the following:
+ 
+ 
+ 
+   
+   Front-ends for C and C++ based on GCC 3.4, supporting the full ANSI-standard
+   C and C++ languages, plus many GCC extensions.
+   
+ 
+   
+   A wide range of global scalar optimizations.
+   
+ 
+   
+   A link-time interprocedural optimization framework with a rich set of
+   analyses and transformations, including sophisticated whole-program pointer
+   analysis and call graph construction.
+   
+ 
+   
+   Native code generators for x86 and Sparc.
+   
+ 
+   
+   A JIT code generation system for x86 and Sparc.
+   
+ 
+   
+   A C back-end, useful for testing and to support other targets.
+   
+ 
+   
+   A test framework with a number of benchmark codes and some applications.
+   
+ 
+   
+   APIs and debugging tools to simplify rapid development.
+   
+ 
+ 
+ LLVM Features:
+ 
+                 
+                 
+                         LLVM: A Compilation Framework for Lifelong Program
+                         Analysis & Transformation:
+                     - Describes the LLVM instruction set and compilation
+                     strategy.  This should be the first document you read to
+                     get an overview of LLVM. 
+                 LLVM Reference Manual
+                     - Defines the LLVM
+                         intermediate representation, the assembly
+                         form of the different nodes, and provides
+                         reference information about the different
+                         tools in LLVM. 
+ 
+                 
+ 
+                 LLVM User Guides:
+ 
+                 
+                 
+                     The LLVM Getting Started Guide
+                     
+                     - Discusses how to get up and running
+                         quickly with the LLVM infrastructure.
+                         Everything from unpacking and
+                         compilation of the distribution to
+                         execution of some tools. 
+                     
+                 LLVM Command Guide
+                     - A reference manual for the LLVM command line
+                     utilities ("man" pages for LLVM tools). 
+ 
+                 LLVM Project Guide
+ 		    - How-to guide and templates for new
+ 		      projects that use the LLVM infrastructure.
+ 		      The templates (directory organization, Makefiles,
+ 		      and test tree) allow the project code to be located
+ 		      outside (or inside) the llvm/ tree,
+ 		      while using LLVM header files and libraries. 
+ 
+                 How to Submit A Bug
+                     Report - Instructions for properly submitting information
+                     about any bugs you run into in the LLVM system.
+ 
+                 LLVM Test Suite Guide
+                     - A reference manual for using the LLVM test suite.
+ 
+                 How to build the C/C++
+                     front-end - Instructions for building the front-end
+                     from source.
+                 
+ 
+                 Programming Documentation:
+ 
+                 
+                     The LLVM
+                         Programmers Manual - Introduction to the general
+                         layout of the LLVM sourcebase, important classes and
+                         APIs, and some tips & tricks. 
+ 
+                     Writing
+                         an LLVM Pass - Information on how to
+                         write LLVM transformations and analyses. 
+ 
+                     Alias Analysis in
+                         LLVM - Information on how to write a new alias
+                         analysis implementation or how to use existing
+                         analyses.
+ 
+                     CommandLine
+                         library Reference Manual
+                         - Provides information on using the command line
+                           parsing library.
+ 
+                     Recommended
+                         LLVM Coding standards
+                         - Details the LLVM coding standards and provides useful
+                           information on writing efficient C++ code.
+ 
+                     Open Projects
+                     - Look here if you are interested in doing something with
+                     LLVM but aren't sure what needs to be done.
+ 
+                     The Stacker Cronicles
+                     - This document describes both the Stacker language and
+                     LLVM frontend, but also some details about LLVM useful for
+                     those writing front-ends.
+ 
+                     Doxygen generated documentation
+                     (classes)
+                     
+                     CVSWeb CVS Tree Browser
+ 
+                 
+ 
+ 
+ 
+ 
+ LLVM Support
+ 
+                 
+                     Frequently Asked Questions
+                     - A list of common questions and problems and their
+                       solutions.
+ 
+                     
+                     How-To-Submit-A-Bug
+                     - Instructions for the proper way to submit information
+                     about a bug you ran into in the LLVM system.
+ 
+                     LLVM Mailing Lists:
+                     
+                     The
+                     
+                     LLVM Announcements List
+                     The Developer's List (LLVM user & developer support)
+                     The Bugs & Patches Archive
+                     The CVS Commits Archive
+ 
+                     
+                     
Results from the LLVM Automated Nightly Tester
+                 

+ 
+ 
+ 




From criswell at cs.uiuc.edu  Wed Dec 17 19:44:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 19:44:01 2003
Subject: [llvm-commits] CVS: llvm-www/Features.html
Message-ID: <200312180143.TAA02731@choi.cs.uiuc.edu>


Changes in directory llvm-www:

Features.html updated: 1.1 -> 1.2

---
Log message:

Get rid of old contents of original file.



---
Diffs of the changes:  (+2 -126)

Index: llvm-www/Features.html
diff -u llvm-www/Features.html:1.1 llvm-www/Features.html:1.2
--- llvm-www/Features.html:1.1	Wed Dec 17 19:40:32 2003
+++ llvm-www/Features.html	Wed Dec 17 19:43:01 2003
@@ -1,6 +1,8 @@
 
 
LLVM Features
 
+LLVM Features:
+
 
 The public release of LLVM is intended to be a fully functional release of our
 compiler system for C and C++.  As such, it includes the following:
@@ -42,131 +44,5 @@
   APIs and debugging tools to simplify rapid development.
   
 

-
-LLVM Features:
-
-                
-                
-                        LLVM: A Compilation Framework for Lifelong Program
-                        Analysis & Transformation:
-                    - Describes the LLVM instruction set and compilation
-                    strategy.  This should be the first document you read to
-                    get an overview of LLVM. 
-                LLVM Reference Manual
-                    - Defines the LLVM
-                        intermediate representation, the assembly
-                        form of the different nodes, and provides
-                        reference information about the different
-                        tools in LLVM. 
-
-                
-
-                LLVM User Guides:
-
-                
-                
-                    The LLVM Getting Started Guide
-                    
-                    - Discusses how to get up and running
-                        quickly with the LLVM infrastructure.
-                        Everything from unpacking and
-                        compilation of the distribution to
-                        execution of some tools. 
-                    
-                LLVM Command Guide
-                    - A reference manual for the LLVM command line
-                    utilities ("man" pages for LLVM tools). 
-
-                LLVM Project Guide
-		    - How-to guide and templates for new
-		      projects that use the LLVM infrastructure.
-		      The templates (directory organization, Makefiles,
-		      and test tree) allow the project code to be located
-		      outside (or inside) the llvm/ tree,
-		      while using LLVM header files and libraries. 
-
-                How to Submit A Bug
-                    Report - Instructions for properly submitting information
-                    about any bugs you run into in the LLVM system.
-
-                LLVM Test Suite Guide
-                    - A reference manual for using the LLVM test suite.
-
-                How to build the C/C++
-                    front-end - Instructions for building the front-end
-                    from source.
-                
-
-                Programming Documentation:
-
-                
-                    The LLVM
-                        Programmers Manual - Introduction to the general
-                        layout of the LLVM sourcebase, important classes and
-                        APIs, and some tips & tricks. 
-
-                    Writing
-                        an LLVM Pass - Information on how to
-                        write LLVM transformations and analyses. 
-
-                    Alias Analysis in
-                        LLVM - Information on how to write a new alias
-                        analysis implementation or how to use existing
-                        analyses.
-
-                    CommandLine
-                        library Reference Manual
-                        - Provides information on using the command line
-                          parsing library.
-
-                    Recommended
-                        LLVM Coding standards
-                        - Details the LLVM coding standards and provides useful
-                          information on writing efficient C++ code.
-
-                    Open Projects
-                    - Look here if you are interested in doing something with
-                    LLVM but aren't sure what needs to be done.
-
-                    The Stacker Cronicles
-                    - This document describes both the Stacker language and
-                    LLVM frontend, but also some details about LLVM useful for
-                    those writing front-ends.
-
-                    Doxygen generated documentation
-                    (classes)
-                    
-                    CVSWeb CVS Tree Browser
-
-                
-
-
-
-
-LLVM Support
-
-                
-                    Frequently Asked Questions
-                    - A list of common questions and problems and their
-                      solutions.
-
-                    
-                    How-To-Submit-A-Bug
-                    - Instructions for the proper way to submit information
-                    about a bug you ran into in the LLVM system.
-
-                    LLVM Mailing Lists:
-                    
-                    The
-                    
-                    LLVM Announcements List
-                    The Developer's List (LLVM user & developer support)
-                    The Bugs & Patches Archive
-                    The CVS Commits Archive
-
-                    
-                    
Results from the LLVM Automated Nightly Tester
-                

-
 
 




From criswell at cs.uiuc.edu  Wed Dec 17 20:25:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Wed Dec 17 20:25:01 2003
Subject: [llvm-commits] CVS: llvm-www/Features.html
Message-ID: <200312180224.UAA03106@choi.cs.uiuc.edu>


Changes in directory llvm-www:

Features.html updated: 1.2 -> 1.3

---
Log message:

Added information on how might find it useful and links to where interested
parties can find more information.



---
Diffs of the changes:  (+43 -5)

Index: llvm-www/Features.html
diff -u llvm-www/Features.html:1.2 llvm-www/Features.html:1.3
--- llvm-www/Features.html:1.2	Wed Dec 17 19:43:01 2003
+++ llvm-www/Features.html	Wed Dec 17 20:24:31 2003
@@ -1,8 +1,6 @@
 
 
LLVM Features
 
-LLVM Features:
-
 
 The public release of LLVM is intended to be a fully functional release of our
 compiler system for C and C++.  As such, it includes the following:
@@ -10,8 +8,8 @@
 
 

   
-  Front-ends for C and C++ based on GCC 3.4, supporting the full ANSI-standard
-  C and C++ languages, plus many GCC extensions.
+  Front-ends for C and C++ based on GCC 3.4.  They support the full
+  ANSI-standard C and C++ languages and many GCC extensions.
   
 
   
@@ -33,7 +31,7 @@
   
 
   
-  A C back-end, useful for testing and to support other targets.
+  A C back-end useful for testing and for supporting other targets.
   
 
   
@@ -44,5 +42,45 @@
   APIs and debugging tools to simplify rapid development.
   
 
+
+LLVM Audience
+
+
+LLVM can be used in many different kinds of projects.  You might be interested
+in LLVM if you are:
+
+
+
+  
+  A compiler researcher interested in compile-time, link-time
+  (interprocedural), and runtime transformations for C and C++ programs.
+  
+
+  
+  A virtual machine researcher/developer interested in a portable,
+  language-independent instruction set and compilation framework.
+  
+
+  
+  An architecture researcher interested in compiler/hardware techniques.
+  
+
+  
+  A security researcher interested in static analysis or instrumentation.
+  
+
+  
+  An instructor (or other person) interested in a system for quick
+  prototyping of compiler transformations.
+  
+
+
+Want to Know More?
+
+
+You can browse the documentation online,
+try LLVM in your web browser,  or
+download the source code.
+
 
 




From lattner at cs.uiuc.edu  Wed Dec 17 21:10:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Wed Dec 17 21:10:01 2003
Subject: [llvm-commits] CVS: llvm-www/Features.html
Message-ID: <200312180309.VAA04044@zion.cs.uiuc.edu>


Changes in directory llvm-www:

Features.html updated: 1.3 -> 1.4

---
Log message:

minor cleanups


---
Diffs of the changes:  (+11 -6)

Index: llvm-www/Features.html
diff -u llvm-www/Features.html:1.3 llvm-www/Features.html:1.4
--- llvm-www/Features.html:1.3	Wed Dec 17 20:24:31 2003
+++ llvm-www/Features.html	Wed Dec 17 21:08:50 2003
@@ -9,7 +9,7 @@
 
   
   Front-ends for C and C++ based on GCC 3.4.  They support the full
-  ANSI-standard C and C++ languages and many GCC extensions.
+  ANSI-standard C and C++ languages, plus many GCC extensions.
   
 
   
@@ -27,11 +27,16 @@
   
 
   
-  A JIT code generation system for x86 and Sparc.
+  A Just-In-Time (JIT) code generation system for x86 and Sparc.
   
 
   
-  A C back-end useful for testing and for supporting other targets.
+  A C back-end useful for testing, for supporting targets other than X86 and 
+  Sparc, and allowing LLVM to use a native code generator.
+  
+
+  
+  A simple profiling system, similar to gprof.
   
 
   
@@ -39,9 +44,9 @@
   
 
   
-  APIs and debugging tools to simplify rapid development.
+  APIs and debugging tools to simplify rapid development of LLVM components.
   
-
+
 
 LLVM Audience
 
@@ -78,7 +83,7 @@
 Want to Know More?
 
 
-You can browse the documentation online,
+You can browse the documentation online,
 try LLVM in your web browser,  or
 download the source code.
 




From lattner at cs.uiuc.edu  Wed Dec 17 21:11:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Wed Dec 17 21:11:01 2003
Subject: [llvm-commits] CVS: llvm-www/header.incl
Message-ID: <200312180310.VAA04368@zion.cs.uiuc.edu>


Changes in directory llvm-www:

header.incl updated: 1.8 -> 1.9

---
Log message:

Add features link to the sidebar


---
Diffs of the changes:  (+1 -0)

Index: llvm-www/header.incl
diff -u llvm-www/header.incl:1.8 llvm-www/header.incl:1.9
--- llvm-www/header.incl:1.8	Wed Dec 17 17:46:57 2003
+++ llvm-www/header.incl	Wed Dec 17 21:10:11 2003
@@ -18,6 +18,7 @@
 Site Map:
 
 Overview

+Features

 Documentation

 Support

 FAQ





From lattner at cs.uiuc.edu  Thu Dec 18 00:41:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 00:41:01 2003
Subject: [llvm-commits] CVS: llvm/docs/Stacker.html
Message-ID: <200312180640.AAA11913@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

Stacker.html updated: 1.5 -> 1.6

---
Log message:

Check in patch that Reid submitted


---
Diffs of the changes:  (+384 -336)

Index: llvm/docs/Stacker.html
diff -u llvm/docs/Stacker.html:1.5 llvm/docs/Stacker.html:1.6
--- llvm/docs/Stacker.html:1.5	Sun Nov 30 23:12:09 2003
+++ llvm/docs/Stacker.html	Thu Dec 18 00:40:22 2003
@@ -57,7 +57,7 @@
 
 This document is another way to learn about LLVM. Unlike the 
 LLVM Reference Manual or 
-LLVM Programmer's Manual, we learn
+LLVM Programmer's Manual, here we learn
 about LLVM through the experience of creating a simple programming language
 named Stacker.  Stacker was invented specifically as a demonstration of
 LLVM. The emphasis in this document is not on describing the
@@ -70,11 +70,10 @@
 
Amongst other things, LLVM is a platform for compiler writers.
 Because of its exceptionally clean and small IR (intermediate
 representation), compiler writing with LLVM is much easier than with
-other system. As proof, the author of Stacker wrote the entire
-compiler (language definition, lexer, parser, code generator, etc.) in
-about four days! That's important to know because it shows 
-how quickly you can get a new
-language up when using LLVM. Furthermore, this was the first 
+other system. As proof, I wrote the entire compiler (language definition, 
+lexer, parser, code generator, etc.) in about four days! 
+That's important to know because it shows how quickly you can get a new
+language running when using LLVM. Furthermore, this was the first 
 language the author ever created using LLVM. The learning curve is 
 included in that four days.
 The language described here, Stacker, is Forth-like. Programs
@@ -136,7 +135,7 @@
 You could write a function using LLVM that computes this expression like this: 
 
 Value* 
-expression(BasicBlock*bb, Value* a, Value* b, Value* x, Value* y )
+expression(BasicBlock* bb, Value* a, Value* b, Value* x, Value* y )
 {
     Instruction* tail = bb->getTerminator();
     ConstantSInt* one = ConstantSInt::get( Type::IntTy, 1);
@@ -154,14 +153,16 @@
     return mult1;
 }
 
-"Okay, big deal," you say.  It is a big deal. Here's why. Note that I didn't
+
"Okay, big deal," you say?  It is a big deal. Here's why. Note that I didn't
 have to tell this function which kinds of Values are being passed in. They could be
-Instructions, Constants, GlobalVariables, 
-etc. Furthermore, if you specify Values that are incorrect for this sequence of 
+Instructions, Constants, GlobalVariables, or
+any of the other subclasses of Value that LLVM supports.
+Furthermore, if you specify Values that are incorrect for this sequence of 
 operations, LLVM will either notice right away (at compilation time) or the LLVM 
-Verifier will pick up the inconsistency when the compiler runs. In no case will 
-you make a type error that gets passed through to the generated program. 
-This really helps you write a compiler that always generates correct code!

+Verifier will pick up the inconsistency when the compiler runs. In either case 
+LLVM prevents you from making a type error that gets passed through to the 
+generated program.  This really helps you write a compiler that 
+always generates correct code!

 
The second point is that we don't have to worry about branching, registers,
 stack variables, saving partial results, etc. The instructions we create 
 are the values we use. Note that all that was created in the above
@@ -235,26 +236,26 @@
 MyCompiler::handle_if( BasicBlock* bb, SetCondInst* condition )
 {
     // Create the blocks to contain code in the structure of if/then/else
-    BasicBlock* then = new BasicBlock(); 
-    BasicBlock* else = new BasicBlock();
-    BasicBlock* exit = new BasicBlock();
+    BasicBlock* then_bb = new BasicBlock(); 
+    BasicBlock* else_bb = new BasicBlock();
+    BasicBlock* exit_bb = new BasicBlock();
 
     // Insert the branch instruction for the "if"
-    bb->getInstList().push_back( new BranchInst( then, else, condition ) );
+    bb->getInstList().push_back( new BranchInst( then_bb, else_bb, condition ) );
 
     // Set up the terminating instructions
-    then->getInstList().push_back( new BranchInst( exit ) );
-    else->getInstList().push_back( new BranchInst( exit ) );
+    then->getInstList().push_back( new BranchInst( exit_bb ) );
+    else->getInstList().push_back( new BranchInst( exit_bb ) );
 
     // Fill in the then part .. details excised for brevity
-    this->fill_in( then );
+    this->fill_in( then_bb );
 
     // Fill in the else part .. details excised for brevity
-    this->fill_in( else );
+    this->fill_in( else_bb );
 
     // Return a block to the caller that can be filled in with the code
     // that follows the if/then/else construct.
-    return exit;
+    return exit_bb;
 }
 
 
Presumably in the foregoing, the calls to the "fill_in" method would add 
@@ -264,15 +265,17 @@
 should they encounter another if/then/else statement and it will just work.
 Note how cleanly this all works out. In particular, the push_back methods on
 the BasicBlock's instruction list. These are lists of type 
-Instruction which also happen to be Values. To create 
+Instruction (which is also of type Value). To create 
 the "if" branch we merely instantiate a BranchInst that takes as 
-arguments the blocks to branch to and the condition to branch on. The blocks
-act like branch labels! This new BranchInst terminates
-the BasicBlock provided as an argument. To give the caller a way
-to keep inserting after calling handle_if we create an "exit" block
-which is returned to the caller.  Note that the "exit" block is used as the 
-terminator for both the "then" and the "else" blocks. This guarantees that no
-matter what else "handle_if" or "fill_in" does, they end up at the "exit" block.
+arguments the blocks to branch to and the condition to branch on. The 
+BasicBlock objects act like branch labels! This new 
+BranchInst terminates the BasicBlock provided 
+as an argument. To give the caller a way to keep inserting after calling 
+handle_if we create an exit_bb block which is returned 
+to the caller.  Note that the exit_bb block is used as the 
+terminator for both the then_bb and the else_bb
+blocks. This guarantees that no matter what else handle_if
+or fill_in does, they end up at the exit_bb block.
 
 
 
@@ -318,8 +321,8 @@
 will run against your grain because you'll naturally think of the global array
 variable and the address of its first element as the same. That tripped me up
 for a while until I realized that they really do differ .. by type.
-Remember that LLVM is a strongly typed language itself. Everything
-has a type.  The "type" of the global variable is [24 x int]*. That is, its
+Remember that LLVM is strongly typed. Everything has a type.  
+The "type" of the global variable is [24 x int]*. That is, its
 a pointer to an array of 24 ints.  When you dereference that global variable with
 a single (0) index, you now have a "[24 x int]" type.  Although
 the pointer value of the dereferenced global and the address of the zero'th element
@@ -332,7 +335,7 @@
 Getting Linkage Types Right
 
 Linkage types in LLVM can be a little confusing, especially if your compiler
-writing mind has affixed very hard concepts to particular words like "weak",
+writing mind has affixed firm concepts to particular words like "weak",
 "external", "global", "linkonce", etc. LLVM does not use the precise
 definitions of say ELF or GCC even though they share common terms. To be fair,
 the concepts are related and similar but not precisely the same. This can lead
@@ -342,16 +345,19 @@
 carefully. Then, read it again.

 
Here are some handy tips that I discovered along the way:
 
-    Unitialized means external. That is, the symbol is declared in the current
+    
Unitialized means external. That is, the symbol is declared in the current
     module and can be used by that module but it is not defined by that module.
-    Setting an initializer changes a global's linkage type from whatever it was
-    to a normal, defind global (not external). You'll need to call the setLinkage()
-    method to reset it if you specify the initializer after the GlobalValue has been
-    constructed. This is important for LinkOnce and Weak linkage types. 
-    Appending linkage can be used to keep track of compilation information at
-    runtime. It could be used, for example, to build a full table of all the C++
-    virtual tables or hold the C++ RTTI data, or whatever. Appending linkage can 
-    only be applied to arrays. The arrays are concatenated together at link time.
+    Setting an initializer changes a global' linkage type. Setting an 
+    initializer changes a global's linkage type from whatever it was to a normal, 
+    defind global (not external). You'll need to call the setLinkage() method to 
+    reset it if you specify the initializer after the GlobalValue has been constructed. 
+    This is important for LinkOnce and Weak linkage types. 
+    Appending linkage can keep track of things. Appending linkage can 
+    be used to keep track of compilation information at runtime. It could be used, 
+    for example, to build a full table of all the C++ virtual tables or hold the 
+    C++ RTTI data, or whatever. Appending linkage can only be applied to arrays. 
+    All arrays with the same name in each module are concatenated together at link 
+    time.
 
 
 
@@ -423,7 +429,7 @@
 # This is a comment to end of line
 ( This is an enclosed comment )
 
-See the example program to see how this works in 
+
See the example program to see comments in use in 
 a real program.
 
 
@@ -446,9 +452,9 @@
 transformation he applies will affect the program.
 Words in a definition come in two flavors: built-in and programmer
 defined. Simply mentioning the name of a previously defined or declared
-programmer-defined word causes that word's definition to be invoked. It
+programmer-defined word causes that word's stack actions to be invoked. It
 is somewhat like a function call in other languages. The built-in
-words have various effects, described below.
+words have various effects, described below.
 Sometimes you need to call a word before it is defined. For this, you can
 use the FORWARD declaration. It looks like this:
 FORWARD name ;
@@ -472,7 +478,8 @@
     ArithmeticThese words perform arithmetic computations on
     their operands. 
 The words are: ABS NEG + - * / MOD */ ++ -- MIN MAX
     StackThese words manipulate the stack directly by moving
-    its elements around.
 The words are: DROP DUP SWAP OVER ROT DUP2 DROP2 PICK TUCK
+    its elements around.
 The words are: DROP DROP2 NIP NIP2 DUP DUP2 
+    SWAP SWAP2 OVER OVER2 ROT ROT2 RROT RROT2 TUCK TUCK2 PICK SELECT ROLL
     MemoryThese words allocate, free and manipulate memory
     areas outside the stack.
The words are: MALLOC FREE GET PUT
     ControlThese words alter the normal left to right flow
@@ -500,311 +507,331 @@
     
p - a pointer to a malloc'd memory block
 

 
-
-
-
-
-
-
-    
-    
-    Definition Of Operation Of Built In Words
LOGICAL OPERATIONS
Word Name Operation Description
< LT w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
+    
+
+
+
+    
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-
-
-    
-    
-    
+    
+    
+    
+
+
+    
+    
+    
+
+
+
+    
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-
-
-    
-    
-    
+
+    
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
+
+    
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-
-
-    
-    
-    
+
+    
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-
-
-    
-    
-    
+
+    
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
-
-
-    
-    
-    
+
+    
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
+
+
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
-
-    
-    
-    
+    
+    
+    
@@ -989,9 +1031,9 @@
 The following fully documented program highlights many features of both
 the Stacker language and what is possible with LLVM. The program has two modes
 of operations. If you provide numeric arguments to the program, it checks to see
-if those arguments are prime numbers, prints out the results. Without any 
+if those arguments are prime numbers and prints out the results. Without any 
 aruments, the program prints out any prime numbers it finds between 1 and one 
-million (there's a log of them!). The source code comments below tell the 
+million (there's a lot of them!). The source code comments below tell the 
 remainder of the story.
 
@@ -1321,7 +1363,13 @@
 interested, here are some things that could be implemented better:

     Write an LLVM pass to compute the correct stack depth needed by the
-    program.
+    program. Currently the stack is set to a fixed number which means programs
+    with large numbers of definitions might fail.
+    Enhance to run on 64-bit platforms like SPARC. Right now the size of a
+    pointer on 64-bit machines will cause incorrect results because of the 32-bit 
+    size of a stack element currently supported. This feature was not implemented
+    because LLVM needs a union type to be able to support the different sizes
+    correctly (portably and efficiently).
     Write an LLVM pass to optimize the use of the global stack. The code
     emitted currently is somewhat wasteful. It gets cleaned up a lot by existing
     passes but more could be done.
@@ -1335,10 +1383,10 @@
     technique out until I was nearly done with LLVM. As it is, its a bad example 
     of how to insert instructions!
     Provide for I/O to arbitrary files instead of just stdin/stdout.
-    Write additional built-in words.
+    Write additional built-in words; with inspiration from FORTH
     Write additional sample Stacker programs.
-    Add your own compiler writing experiences and tips in the 
-    Lessons I Learned About LLVM section.
+    Add your own compiler writing experiences and tips in the 
+    Lessons I Learned About LLVM section.
 
@@ -1346,6 +1394,6 @@
 
 Reid Spencer
 The LLVM Compiler Infrastructure 
-
Last modified: $Date: 2003/12/01 05:12:09 $ 
+
Last modified: $Date: 2003/12/18 06:40:22 $ 




From lattner at cs.uiuc.edu  Thu Dec 18 02:12:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 02:12:02 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/Analysis/AliasSetTracker.h
Message-ID: <200312180811.CAA28368@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/Analysis:

AliasSetTracker.h updated: 1.10 -> 1.11

---
Log message:

Add a new AliassetTracker::remove method.  Because we need to be able to remove
a pointer from an AliasSet, maintain the pointer values on a doubly linked
list instead of a singly linked list, to permit efficient removal from the 
middle of the list.



---
Diffs of the changes:  (+33 -9)

Index: llvm/include/llvm/Analysis/AliasSetTracker.h
diff -u llvm/include/llvm/Analysis/AliasSetTracker.h:1.10 llvm/include/llvm/Analysis/AliasSetTracker.h:1.11
--- llvm/include/llvm/Analysis/AliasSetTracker.h:1.10	Sat Dec 13 22:51:34 2003
+++ llvm/include/llvm/Analysis/AliasSetTracker.h	Thu Dec 18 02:11:11 2003
@@ -37,15 +37,20 @@
   typedef std::pair HashNodePair;
 
   class PointerRec {
-    HashNodePair *NextInList;
+    HashNodePair **PrevInList, *NextInList;
     AliasSet *AS;
     unsigned Size;
   public:
-    PointerRec() : NextInList(0), AS(0), Size(0) {}
+    PointerRec() : PrevInList(0), NextInList(0), AS(0), Size(0) {}
 
     HashNodePair *getNext() const { return NextInList; }
     bool hasAliasSet() const { return AS != 0; }
 
+    HashNodePair** setPrevInList(HashNodePair **PIL) {
+      PrevInList = PIL;
+      return &NextInList;
+    }
+
     void updateSize(unsigned NewSize) {
       if (NewSize > Size) Size = NewSize;
     }
@@ -68,13 +73,14 @@
       assert(AS == 0 && "Already have an alias set!");
       AS = as;
     }
-    void setTail(HashNodePair *T) {
-      assert(NextInList == 0 && "Already have tail!");
-      NextInList = T;
+
+    void removeFromList() {
+      if (NextInList) NextInList->second.PrevInList = PrevInList;
+      *PrevInList = NextInList;
     }
   };
 
-  HashNodePair *PtrListHead, *PtrListTail; // Singly linked list of nodes
+  HashNodePair *PtrList, **PtrListEnd;  // Doubly linked list of nodes
   AliasSet *Forward;             // Forwarding pointer
   AliasSet *Next, *Prev;         // Doubly linked list of AliasSets
 
@@ -135,7 +141,7 @@
   // Alias Set iteration - Allow access to all of the pointer which are part of
   // this alias set...
   class iterator;
-  iterator begin() const { return iterator(PtrListHead); }
+  iterator begin() const { return iterator(PtrList); }
   iterator end()   const { return iterator(); }
 
   void print(std::ostream &OS) const;
@@ -175,11 +181,22 @@
 
 private:
   // Can only be created by AliasSetTracker
-  AliasSet() : PtrListHead(0), PtrListTail(0), Forward(0), RefCount(0),
+  AliasSet() : PtrList(0), PtrListEnd(&PtrList), Forward(0), RefCount(0),
                AccessTy(NoModRef), AliasTy(MustAlias), Volatile(false) {
   }
+  AliasSet(const AliasSet &AS) {
+    // AliasSet's only get copy constructed in simple circumstances.  In
+    // particular, they cannot have any pointers in their list.  Despite this,
+    // we have to be sure to update the PtrListEnd to not point to the source
+    // AliasSet's list.
+    assert(AS.PtrList == 0 && "AliasSet has pointers in it!");
+    PtrList = 0; PtrListEnd = &PtrList;
+    Forward = AS.Forward; RefCount = AS.RefCount;
+    AccessTy = AS.AccessTy; AliasTy = AS.AliasTy; Volatile = AS.Volatile;
+  }
+
   HashNodePair *getSomePointer() const {
-    return PtrListHead ? PtrListHead : 0;
+    return PtrList;
   }
 
   /// getForwardedTarget - Return the real alias set this represents.  If this
@@ -246,6 +263,13 @@
   void add(Instruction *I);       // Dispatch to one of the other add methods...
   void add(BasicBlock &BB);       // Add all instructions in basic block
   void add(const AliasSetTracker &AST); // Add alias relations from another AST
+
+  /// remove method - This method is used to remove a pointer value from the
+  /// AliasSetTracker entirely.  It should be used when an instruction is
+  /// deleted from the program to update the AST.  If you don't use this, you
+  /// would have dangling pointers to deleted instructions.
+  ///
+  void remove(Value *PtrVal);
 
   /// getAliasSets - Return the alias sets that are active.
   const ilist &getAliasSets() const { return AliasSets; }




From lattner at cs.uiuc.edu  Thu Dec 18 02:13:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 02:13:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Analysis/AliasSetTracker.cpp
Message-ID: <200312180812.CAA29360@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Analysis:

AliasSetTracker.cpp updated: 1.9 -> 1.10

---
Log message:

Add a new AliassetTracker::remove method.  Because we need to be able to remove
a pointer from an AliasSet, maintain the pointer values on a doubly linked
list instead of a singly linked list, to permit efficient removal from the
middle of the list.


---
Diffs of the changes:  (+36 -8)

Index: llvm/lib/Analysis/AliasSetTracker.cpp
diff -u llvm/lib/Analysis/AliasSetTracker.cpp:1.9 llvm/lib/Analysis/AliasSetTracker.cpp:1.10
--- llvm/lib/Analysis/AliasSetTracker.cpp:1.9	Sat Dec 13 22:52:11 2003
+++ llvm/lib/Analysis/AliasSetTracker.cpp	Thu Dec 18 02:11:56 2003
@@ -47,9 +47,14 @@
   RefCount++;         // AS is now pointing to us...
 
   // Merge the list of constituent pointers...
-  PtrListTail->second.setTail(AS.PtrListHead);
-  PtrListTail = AS.PtrListTail;
-  AS.PtrListHead = AS.PtrListTail = 0;
+  if (AS.PtrList) {
+    *PtrListEnd = AS.PtrList;
+    AS.PtrList->second.setPrevInList(PtrListEnd);
+    PtrListEnd = AS.PtrListEnd;
+
+    AS.PtrList = 0;
+    AS.PtrListEnd = &AS.PtrList;
+  }
 }
 
 void AliasSetTracker::removeAliasSet(AliasSet *AS) {
@@ -82,11 +87,9 @@
   Entry.second.updateSize(Size);
 
   // Add it to the end of the list...
-  if (PtrListTail)
-    PtrListTail->second.setTail(&Entry);
-  else
-    PtrListHead = &Entry;
-  PtrListTail = &Entry;
+  assert(*PtrListEnd == 0 && "End of list is not null?");
+  *PtrListEnd = &Entry;
+  PtrListEnd = Entry.second.setPrevInList(PtrListEnd);
   RefCount++;               // Entry points to alias set...
 }
 
@@ -252,6 +255,31 @@
                    (AliasSet::AccessType)AS.AccessTy);
     }
 }
+
+
+// remove method - This method is used to remove a pointer value from the
+// AliasSetTracker entirely.  It should be used when an instruction is deleted
+// from the program to update the AST.  If you don't use this, you would have
+// dangling pointers to deleted instructions.
+//
+void AliasSetTracker::remove(Value *PtrVal) {
+  // First, look up the PointerRec for this pointer...
+  hash_map::iterator I = PointerMap.find(PtrVal);
+  if (I == PointerMap.end()) return;  // Noop
+
+  // If we found one, remove the pointer from the alias set it is in.
+  AliasSet::HashNodePair &PtrValEnt = *I;
+  AliasSet *AS = PtrValEnt.second.getAliasSet(*this);
+
+  // Unlink from the list of values...
+  PtrValEnt.second.removeFromList();
+  // Stop using the alias set
+  if (--AS->RefCount == 0)
+    AS->removeFromTracker(*this);
+
+  PointerMap.erase(I);
+}
+
 
 //===----------------------------------------------------------------------===//
 //               AliasSet/AliasSetTracker Printing Support




From lattner at cs.uiuc.edu  Thu Dec 18 02:13:08 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 02:13:08 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/LICM.cpp
Message-ID: <200312180812.CAA30197@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

LICM.cpp updated: 1.51 -> 1.52

---
Log message:

When we delete instructions from the loop, make sure to remove them from the
AliasSetTracker as well.


---
Diffs of the changes:  (+2 -0)

Index: llvm/lib/Transforms/Scalar/LICM.cpp
diff -u llvm/lib/Transforms/Scalar/LICM.cpp:1.51 llvm/lib/Transforms/Scalar/LICM.cpp:1.52
--- llvm/lib/Transforms/Scalar/LICM.cpp:1.51	Sat Dec 13 22:52:31 2003
+++ llvm/lib/Transforms/Scalar/LICM.cpp	Thu Dec 18 02:12:32 2003
@@ -387,6 +387,7 @@
   if (ExitBlocks.size() == 1) {
     if (!isExitBlockDominatedByBlockInLoop(ExitBlocks[0], I.getParent())) {
       // Instruction is not used, just delete it.
+      CurAST->remove(&I);
       I.getParent()->getInstList().erase(&I);
     } else {
       // Move the instruction to the start of the exit block, after any PHI
@@ -399,6 +400,7 @@
     }
   } else if (ExitBlocks.size() == 0) {
     // The instruction is actually dead if there ARE NO exit blocks.
+    CurAST->remove(&I);
     I.getParent()->getInstList().erase(&I);
   } else {
     // Otherwise, if we have multiple exits, use the PromoteMem2Reg function to




From lattner at cs.uiuc.edu  Thu Dec 18 02:17:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 02:17:01 2003
Subject: [llvm-commits] CVS: llvm/docs/ReleaseNotes.html
Message-ID: <200312180816.CAA32722@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

ReleaseNotes.html updated: 1.86 -> 1.87

---
Log message:

add boog


---
Diffs of the changes:  (+3 -1)

Index: llvm/docs/ReleaseNotes.html
diff -u llvm/docs/ReleaseNotes.html:1.86 llvm/docs/ReleaseNotes.html:1.87
--- llvm/docs/ReleaseNotes.html:1.86	Mon Dec 15 16:48:12 2003
+++ llvm/docs/ReleaseNotes.html	Thu Dec 18 02:16:25 2003
@@ -137,6 +137,8 @@
 
 [licm] LICM promotes volatile memory
 locations to registers
+[licm] Memory read after free causes 
+infrequent crash
 
 
 
@@ -542,7 +544,7 @@
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" />
 
   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/15 22:48:12 $
+  Last modified: $Date: 2003/12/18 08:16:25 $
 
 
 




From lattner at cs.uiuc.edu  Thu Dec 18 02:21:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 02:21:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/1.1/docs/ReleaseNotes.html
Message-ID: <200312180820.CAA00793@zion.cs.uiuc.edu>


Changes in directory llvm-www/releases/1.1/docs:

ReleaseNotes.html updated: 1.4 -> 1.5

---
Log message:

Add boogs, fixed in 1.2


---
Diffs of the changes:  (+11 -1)

Index: llvm-www/releases/1.1/docs/ReleaseNotes.html
diff -u llvm-www/releases/1.1/docs/ReleaseNotes.html:1.4 llvm-www/releases/1.1/docs/ReleaseNotes.html:1.5
--- llvm-www/releases/1.1/docs/ReleaseNotes.html:1.4	Wed Dec 17 16:47:16 2003
+++ llvm-www/releases/1.1/docs/ReleaseNotes.html	Thu Dec 18 02:20:31 2003
@@ -374,10 +374,20 @@
 Tail duplication does not update SSA form correctly.
 
 
+
+Bugs in 1.0 fixed in 1.1:
+
+[licm] LICM promotes volatile memory
+locations to registers
+[licm] Memory read after free causes 
+infrequent crash
+
 
 
 
 
+
+
 
 
   Known problems with the C front-end
@@ -729,7 +739,7 @@
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" />
 
   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/17 22:47:16 $
+  Last modified: $Date: 2003/12/18 08:20:31 $
 
 
 




From alkis at cs.uiuc.edu  Thu Dec 18 02:49:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 02:49:01 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/CodeGen/LiveIntervals.h
Message-ID: <200312180848.CAA28098@morpheus.cs.uiuc.edu>


Changes in directory llvm/include/llvm/CodeGen:

LiveIntervals.h updated: 1.3 -> 1.4

---
Log message:

Handle multiple virtual register definitions gracefully.

Move some of the longer LiveIntervals::Interval method out of the
header and add debug information to them. Fix bug and simplify range
merging code.


---
Diffs of the changes:  (+4 -47)

Index: llvm/include/llvm/CodeGen/LiveIntervals.h
diff -u llvm/include/llvm/CodeGen/LiveIntervals.h:1.3 llvm/include/llvm/CodeGen/LiveIntervals.h:1.4
--- llvm/include/llvm/CodeGen/LiveIntervals.h:1.3	Fri Dec  5 04:38:28 2003
+++ llvm/include/llvm/CodeGen/LiveIntervals.h	Thu Dec 18 02:48:48 2003
@@ -59,55 +59,12 @@
                 return end() <= index;
             }
 
-            bool overlaps(unsigned index) const {
-                for (Ranges::const_iterator
-                         i = ranges.begin(), e = ranges.end(); i != e; ++i) {
-                    if (index >= i->first && index < i->second) {
-                        return true;
-                    }
-                }
-                return false;
-            }
-
-            void addRange(unsigned start, unsigned end) {
-                Range range = std::make_pair(start, end);
-                Ranges::iterator it =
-                    std::lower_bound(ranges.begin(), ranges.end(), range);
-
-                if (it == ranges.end()) {
-                    it = ranges.insert(it, range);
-                    goto exit;
-                }
-
-                assert(range.first <= it->first && "got wrong iterator?");
-                // merge ranges if necesary
-                if (range.first < it->first) {
-                    if (range.second >= it->first) {
-                        it->first = range.first;
-                    }
-                    else {
-                        it = ranges.insert(it, range);
-                        assert(it != ranges.end() && "wtf?");
-                        goto exit;
-                    }
-                }
-
-            exit:
-                mergeRangesIfNecessary(it);
-            }
+            void addRange(unsigned start, unsigned end);
 
         private:
-            void mergeRangesIfNecessary(Ranges::iterator it) {
-                while (it != ranges.begin()) {
-                    Ranges::iterator prev = it - 1;
-                    if (prev->second < it->first) {
-                        break;
-                    }
-                    prev->second = it->second;
-                    ranges.erase(it);
-                    it = prev;
-                }
-            }
+            void mergeRangesForward(Ranges::iterator it);
+
+            void mergeRangesBackward(Ranges::iterator it);
         };
 
         struct StartPointComp {




From alkis at cs.uiuc.edu  Thu Dec 18 02:49:08 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 02:49:08 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/LiveIntervals.cpp
Message-ID: <200312180848.CAA28090@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

LiveIntervals.cpp updated: 1.8 -> 1.9

---
Log message:

Handle multiple virtual register definitions gracefully.

Move some of the longer LiveIntervals::Interval method out of the
header and add debug information to them. Fix bug and simplify range
merging code.


---
Diffs of the changes:  (+67 -51)

Index: llvm/lib/CodeGen/LiveIntervals.cpp
diff -u llvm/lib/CodeGen/LiveIntervals.cpp:1.8 llvm/lib/CodeGen/LiveIntervals.cpp:1.9
--- llvm/lib/CodeGen/LiveIntervals.cpp:1.8	Sun Dec 14 22:55:38 2003
+++ llvm/lib/CodeGen/LiveIntervals.cpp	Thu Dec 18 02:48:47 2003
@@ -122,60 +122,46 @@
 
     LiveVariables::VarInfo& vi = lv_->getVarInfo(reg);
 
+    Interval* interval = 0;
     Reg2IntervalMap::iterator r2iit = r2iMap_.find(reg);
-    // handle multiple definition case (machine instructions violating
-    // ssa after phi-elimination
-    if (r2iit != r2iMap_.end()) {
-        unsigned ii = r2iit->second;
-        Interval& interval = intervals_[ii];
-        unsigned end = getInstructionIndex(mbb->back()) + 1;
-        DEBUG(std::cerr << "\t\t\t\tadding range: ["
-              << instrIndex << ',' << end << "]\n");
-        interval.addRange(instrIndex, end);
-        DEBUG(std::cerr << "\t\t\t\t" << interval << '\n');
-    }
-    else {
+    if (r2iit == r2iMap_.end()) {
         // add new interval
         intervals_.push_back(Interval(reg));
-        Interval& interval = intervals_.back();
         // update interval index for this register
         r2iMap_[reg] = intervals_.size() - 1;
+        interval = &intervals_.back();
+    }
+    else {
+        interval = &intervals_[r2iit->second];
+    }
 
-        for (MbbIndex2MbbMap::iterator
-                 it = mbbi2mbbMap_.begin(), itEnd = mbbi2mbbMap_.end();
-             it != itEnd; ++it) {
-            unsigned liveBlockIndex = it->first;
-            MachineBasicBlock* liveBlock = it->second;
-            if (liveBlockIndex < vi.AliveBlocks.size() &&
-                vi.AliveBlocks[liveBlockIndex]) {
-                unsigned start =  getInstructionIndex(liveBlock->front());
-                unsigned end = getInstructionIndex(liveBlock->back()) + 1;
-                DEBUG(std::cerr << "\t\t\t\tadding range: ["
-                      << start << ',' << end << "]\n");
-                interval.addRange(start, end);
-            }
-        }
-
-        bool killedInDefiningBasicBlock = false;
-        for (int i = 0, e = vi.Kills.size(); i != e; ++i) {
-            MachineBasicBlock* killerBlock = vi.Kills[i].first;
-            MachineInstr* killerInstr = vi.Kills[i].second;
-            killedInDefiningBasicBlock |= mbb == killerBlock;
-            unsigned start = (mbb == killerBlock ?
-                              instrIndex :
-                              getInstructionIndex(killerBlock->front()));
-            unsigned end = getInstructionIndex(killerInstr) + 1;
-            DEBUG(std::cerr << "\t\t\t\tadding range: ["
-                  << start << ',' << end << "]\n");
-            interval.addRange(start, end);
+    for (MbbIndex2MbbMap::iterator
+             it = mbbi2mbbMap_.begin(), itEnd = mbbi2mbbMap_.end();
+         it != itEnd; ++it) {
+        unsigned liveBlockIndex = it->first;
+        MachineBasicBlock* liveBlock = it->second;
+        if (liveBlockIndex < vi.AliveBlocks.size() &&
+            vi.AliveBlocks[liveBlockIndex]) {
+            unsigned start =  getInstructionIndex(liveBlock->front());
+            unsigned end = getInstructionIndex(liveBlock->back()) + 1;
+            interval->addRange(start, end);
         }
+    }
 
-        if (!killedInDefiningBasicBlock) {
-            unsigned end = getInstructionIndex(mbb->back()) + 1;
-            interval.addRange(instrIndex, end);
-        }
+    bool killedInDefiningBasicBlock = false;
+    for (int i = 0, e = vi.Kills.size(); i != e; ++i) {
+        MachineBasicBlock* killerBlock = vi.Kills[i].first;
+        MachineInstr* killerInstr = vi.Kills[i].second;
+        killedInDefiningBasicBlock |= mbb == killerBlock;
+        unsigned start = (mbb == killerBlock ?
+                          instrIndex :
+                          getInstructionIndex(killerBlock->front()));
+        unsigned end = getInstructionIndex(killerInstr) + 1;
+    }
 
-        DEBUG(std::cerr << "\t\t\t\t" << interval << '\n');
+    if (!killedInDefiningBasicBlock) {
+        unsigned end = getInstructionIndex(mbb->back()) + 1;
+        interval->addRange(instrIndex, end);
     }
 }
 
@@ -220,20 +206,14 @@
     if (r2iit != r2iMap_.end()) {
         unsigned ii = r2iit->second;
         Interval& interval = intervals_[ii];
-        DEBUG(std::cerr << "\t\t\t\tadding range: ["
-              << start << ',' << end << "]\n");
         interval.addRange(start, end);
-        DEBUG(std::cerr << "\t\t\t\t" << interval << '\n');
     }
     else {
         intervals_.push_back(Interval(reg));
         Interval& interval = intervals_.back();
         // update interval index for this register
         r2iMap_[reg] = intervals_.size() - 1;
-        DEBUG(std::cerr << "\t\t\t\tadding range: ["
-              << start << ',' << end << "]\n");
         interval.addRange(start, end);
-        DEBUG(std::cerr << "\t\t\t\t" << interval << '\n');
     }
 }
 
@@ -308,6 +288,42 @@
     std::sort(intervals_.begin(), intervals_.end(), StartPointComp());
     DEBUG(std::copy(intervals_.begin(), intervals_.end(),
                     std::ostream_iterator(std::cerr, "\n")));
+}
+
+void LiveIntervals::Interval::addRange(unsigned start, unsigned end)
+{
+    DEBUG(std::cerr << "\t\t\t\tadding range: [" << start <<','<< end << "]\n");
+    //assert(start < end && "invalid range?");
+    Range range = std::make_pair(start, end);
+    Ranges::iterator it =
+        ranges.insert(std::upper_bound(ranges.begin(), ranges.end(), range),
+                      range);
+
+    DEBUG(std::cerr << "\t\t\t\tbefore merge forward: " << *this << '\n');
+    mergeRangesForward(it);
+    DEBUG(std::cerr << "\t\t\t\tbefore merge backward: " << *this << '\n');
+    mergeRangesBackward(it);
+    DEBUG(std::cerr << "\t\t\t\tafter merging: " << *this << '\n');
+}
+
+void LiveIntervals::Interval::mergeRangesForward(Ranges::iterator it)
+{
+    for (Ranges::iterator next = it + 1;
+         next != ranges.end() && it->second >= next->first; ) {
+        it->second = std::max(it->second, next->second);
+        next = ranges.erase(next);
+    }
+}
+
+void LiveIntervals::Interval::mergeRangesBackward(Ranges::iterator it)
+{
+    for (Ranges::iterator prev = it - 1;
+         it != ranges.begin() && it->first <= prev->second; ) {
+        it->first = std::min(it->first, prev->first);
+        it->second = std::max(it->second, prev->second);
+        it = ranges.erase(prev);
+        prev = it - 1;
+    }
 }
 
 std::ostream& llvm::operator<<(std::ostream& os,




From alkis at cs.uiuc.edu  Thu Dec 18 02:54:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 02:54:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/LiveIntervals.cpp
Message-ID: <200312180853.CAA28154@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

LiveIntervals.cpp updated: 1.9 -> 1.10

---
Log message:

When a variable is killed and redifined in a basic block only one
killing instruction is tracked. This causes the LiveIntervals to
create bogus intervals. The workaound is to add a range to the
interval from the redefinition to the end of the basic block.


---
Diffs of the changes:  (+4 -1)

Index: llvm/lib/CodeGen/LiveIntervals.cpp
diff -u llvm/lib/CodeGen/LiveIntervals.cpp:1.9 llvm/lib/CodeGen/LiveIntervals.cpp:1.10
--- llvm/lib/CodeGen/LiveIntervals.cpp:1.9	Thu Dec 18 02:48:47 2003
+++ llvm/lib/CodeGen/LiveIntervals.cpp	Thu Dec 18 02:53:43 2003
@@ -152,11 +152,14 @@
     for (int i = 0, e = vi.Kills.size(); i != e; ++i) {
         MachineBasicBlock* killerBlock = vi.Kills[i].first;
         MachineInstr* killerInstr = vi.Kills[i].second;
-        killedInDefiningBasicBlock |= mbb == killerBlock;
         unsigned start = (mbb == killerBlock ?
                           instrIndex :
                           getInstructionIndex(killerBlock->front()));
         unsigned end = getInstructionIndex(killerInstr) + 1;
+        if (start < end) {
+            killedInDefiningBasicBlock |= mbb == killerBlock;
+            interval->addRange(start, end);
+        }
     }
 
     if (!killedInDefiningBasicBlock) {




From alkis at cs.uiuc.edu  Thu Dec 18 02:57:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 02:57:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/RegAllocLinearScan.cpp
Message-ID: <200312180856.CAA28202@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

RegAllocLinearScan.cpp updated: 1.10 -> 1.11

---
Log message:

Rename LiveIntervals::expired() to LiveIntervals::expiredAt().


---
Diffs of the changes:  (+1 -1)

Index: llvm/lib/CodeGen/RegAllocLinearScan.cpp
diff -u llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.10 llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.11
--- llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.10	Sun Dec 14 07:24:16 2003
+++ llvm/lib/CodeGen/RegAllocLinearScan.cpp	Thu Dec 18 02:56:11 2003
@@ -464,7 +464,7 @@
         // an interval expires this is going to be the last use. in
         // this case we can reuse the register for a def in the same
         // instruction
-        if ((*i)->expired(cur->start() + 1)) {
+        if ((*i)->expiredAt(cur->start() + 1)) {
             DEBUG(std::cerr << "\t\tinterval " << **i << " expired\n");
             if (reg < MRegisterInfo::FirstVirtualRegister) {
                 clearReservedPhysReg(reg);




From alkis at cs.uiuc.edu  Thu Dec 18 02:57:08 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 02:57:08 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/CodeGen/LiveIntervals.h
Message-ID: <200312180856.CAA28211@morpheus.cs.uiuc.edu>


Changes in directory llvm/include/llvm/CodeGen:

LiveIntervals.h updated: 1.4 -> 1.5

---
Log message:

Rename LiveIntervals::expired() to LiveIntervals::expiredAt().


---
Diffs of the changes:  (+1 -1)

Index: llvm/include/llvm/CodeGen/LiveIntervals.h
diff -u llvm/include/llvm/CodeGen/LiveIntervals.h:1.4 llvm/include/llvm/CodeGen/LiveIntervals.h:1.5
--- llvm/include/llvm/CodeGen/LiveIntervals.h:1.4	Thu Dec 18 02:48:48 2003
+++ llvm/include/llvm/CodeGen/LiveIntervals.h	Thu Dec 18 02:56:11 2003
@@ -55,7 +55,7 @@
                 return ranges.back().second;
             }
 
-            bool expired(unsigned index) const {
+            bool expiredAt(unsigned index) const {
                 return end() <= index;
             }
 




From alkis at cs.uiuc.edu  Thu Dec 18 07:06:03 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:06:03 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/CodeGen/LiveVariables.h
Message-ID: <200312181302.HAA29036@morpheus.cs.uiuc.edu>


Changes in directory llvm/include/llvm/CodeGen:

LiveVariables.h updated: 1.9 -> 1.10

---
Log message:

Add TwoAddressInstructionPass to handle instructions that have two or
more operands and the two first operands are constrained to be the
same. The pass takes an instruction of the form:

        a = b op c

and transforms it into:

        a = b
        a = a op c

and also preserves live variables.


---
Diffs of the changes:  (+56 -12)

Index: llvm/include/llvm/CodeGen/LiveVariables.h
diff -u llvm/include/llvm/CodeGen/LiveVariables.h:1.9 llvm/include/llvm/CodeGen/LiveVariables.h:1.10
--- llvm/include/llvm/CodeGen/LiveVariables.h:1.9	Sat Dec 13 05:11:02 2003
+++ llvm/include/llvm/CodeGen/LiveVariables.h	Thu Dec 18 07:02:36 2003
@@ -57,15 +57,18 @@
 
     VarInfo() : DefBlock(0), DefInst(0) {}
 
-    /// removeKill - Delete a kill corresponding to the specified machine instr
-    void removeKill(MachineInstr *MI) {
-      for (unsigned i = 0; ; ++i) {
-        assert(i < Kills.size() && "Machine instr is not a kill!");
-        if (Kills[i].second == MI) {
-          Kills.erase(Kills.begin()+i);
-          return;
+    /// removeKill - Delete a kill corresponding to the specified
+    /// machine instruction. Returns true if there was a kill
+    /// corresponding to this instruction, false otherwise.
+    bool removeKill(MachineInstr *MI) {
+      for (std::vector >::iterator
+             i = Kills.begin(); i != Kills.end(); ++i) {
+        if (i->second == MI) {
+          Kills.erase(i);
+          return true;
         }
       }
+      return false;
     }
   };
 
@@ -156,30 +159,71 @@
   /// specified register is killed after being used by the specified
   /// instruction.
   ///
-  void addVirtualRegisterKilled(unsigned IncomingReg, MachineBasicBlock *MBB,
+  void addVirtualRegisterKilled(unsigned IncomingReg,
+                                MachineBasicBlock *MBB,
                                 MachineInstr *MI) {
     RegistersKilled.insert(std::make_pair(MI, IncomingReg));
     getVarInfo(IncomingReg).Kills.push_back(std::make_pair(MBB, MI));
   }
 
+  /// removeVirtualRegisterKilled - Remove the specified virtual
+  /// register from the live variable information. Returns true if the
+  /// variable was marked as killed by the specified instruction,
+  /// false otherwise.
+  bool removeVirtualRegisterKilled(unsigned reg,
+                                   MachineBasicBlock *MBB,
+                                   MachineInstr *MI) {
+    if (!getVarInfo(reg).removeKill(MI))
+      return false;
+    for (killed_iterator i = killed_begin(MI), e = killed_end(MI); i != e; ) {
+      if (i->second == reg)
+        RegistersKilled.erase(i++);
+      else
+        ++i;
+    }
+    return true;
+  }
+
   /// removeVirtualRegistersKilled - Remove all of the specified killed
   /// registers from the live variable information.
   void removeVirtualRegistersKilled(killed_iterator B, killed_iterator E) {
-    for (killed_iterator I = B; I != E; ++I)  // Remove VarInfo entries...
-      getVarInfo(I->second).removeKill(I->first);
+    for (killed_iterator I = B; I != E; ++I) { // Remove VarInfo entries...
+      bool removed = getVarInfo(I->second).removeKill(I->first);
+      assert(removed && "kill not in register's VarInfo?");
+    }
     RegistersKilled.erase(B, E);
   }
 
   /// addVirtualRegisterDead - Add information about the fact that the specified
   /// register is dead after being used by the specified instruction.
   ///
-  void addVirtualRegisterDead(unsigned IncomingReg, MachineBasicBlock *MBB,
+  void addVirtualRegisterDead(unsigned IncomingReg,
+                              MachineBasicBlock *MBB,
                               MachineInstr *MI) {
     RegistersDead.insert(std::make_pair(MI, IncomingReg));
     getVarInfo(IncomingReg).Kills.push_back(std::make_pair(MBB, MI));
   }
 
-  /// removeVirtualRegistersKilled - Remove all of the specified killed
+  /// removeVirtualRegisterDead - Remove the specified virtual
+  /// register from the live variable information. Returns true if the
+  /// variable was marked dead at the specified instruction, false
+  /// otherwise.
+  bool removeVirtualRegisterDead(unsigned reg,
+                                 MachineBasicBlock *MBB,
+                                 MachineInstr *MI) {
+    if (!getVarInfo(reg).removeKill(MI))
+      return false;
+
+    for (killed_iterator i = killed_begin(MI), e = killed_end(MI); i != e; ) {
+      if (i->second == reg)
+        RegistersKilled.erase(i++);
+      else
+        ++i;
+    }
+    return true;
+  }
+
+  /// removeVirtualRegistersDead - Remove all of the specified dead
   /// registers from the live variable information.
   void removeVirtualRegistersDead(killed_iterator B, killed_iterator E) {
     for (killed_iterator I = B; I != E; ++I)  // Remove VarInfo entries...




From alkis at cs.uiuc.edu  Thu Dec 18 07:07:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:07:01 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/CodeGen/TwoAddressInstructionPass.h
Message-ID: <200312181306.HAA29088@morpheus.cs.uiuc.edu>


Changes in directory llvm/include/llvm/CodeGen:

TwoAddressInstructionPass.h added (r1.1)

---
Log message:

Add TwoAddressInstructionPass to handle instructions that have two or
more operands and the two first operands are constrained to be the
same. The pass takes an instruction of the form:

        a = b op c

and transforms it into:

        a = b
        a = a op c

and also preserves live variables.


---
Diffs of the changes:  (+51 -0)

Index: llvm/include/llvm/CodeGen/TwoAddressInstructionPass.h
diff -c /dev/null llvm/include/llvm/CodeGen/TwoAddressInstructionPass.h:1.1
*** /dev/null	Thu Dec 18 07:06:14 2003
--- llvm/include/llvm/CodeGen/TwoAddressInstructionPass.h	Thu Dec 18 07:06:04 2003
***************
*** 0 ****
--- 1,51 ----
+ //===-- llvm/CodeGen/TwoAddressInstructionPass.h - Two-Address instruction pass  -*- C++ -*-===//
+ //
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ //
+ //===----------------------------------------------------------------------===//
+ //
+ // This file implements the Two-Address instruction rewriter pass. In
+ // some architectures instructions have a combined source/destination
+ // operand. In those cases the instruction cannot have three operands
+ // as the destination is implicit (for example ADD %EAX, %EBX on the
+ // IA-32). After code generation this restrictions are not handled and
+ // instructions may have three operands. This pass remedies this and
+ // reduces all two-address instructions to two operands.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef LLVM_CODEGEN_TWOADDRESSINSTRUCTIONPASS_H
+ #define LLVM_CODEGEN_TWOADDRESSINSTRUCTIONPASS_H
+ 
+ #include "llvm/CodeGen/MachineFunctionPass.h"
+ #include "llvm/CodeGen/MachineBasicBlock.h"
+ #include 
+ #include 
+ 
+ namespace llvm {
+ 
+     class LiveVariables;
+     class MRegisterInfo;
+ 
+     class TwoAddressInstructionPass : public MachineFunctionPass
+     {
+     private:
+         MachineFunction* mf_;
+         const TargetMachine* tm_;
+         const MRegisterInfo* mri_;
+         LiveVariables* lv_;
+ 
+     public:
+         virtual void getAnalysisUsage(AnalysisUsage &AU) const;
+ 
+     private:
+         /// runOnMachineFunction - pass entry point
+         bool runOnMachineFunction(MachineFunction&);
+     };
+ 
+ } // End llvm namespace
+ 
+ #endif




From alkis at cs.uiuc.edu  Thu Dec 18 07:07:10 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:07:10 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/TwoAddressInstructionPass.cpp
Message-ID: <200312181306.HAA29083@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

TwoAddressInstructionPass.cpp added (r1.1)

---
Log message:

Add TwoAddressInstructionPass to handle instructions that have two or
more operands and the two first operands are constrained to be the
same. The pass takes an instruction of the form:

        a = b op c

and transforms it into:

        a = b
        a = a op c

and also preserves live variables.


---
Diffs of the changes:  (+149 -0)

Index: llvm/lib/CodeGen/TwoAddressInstructionPass.cpp
diff -c /dev/null llvm/lib/CodeGen/TwoAddressInstructionPass.cpp:1.1
*** /dev/null	Thu Dec 18 07:06:14 2003
--- llvm/lib/CodeGen/TwoAddressInstructionPass.cpp	Thu Dec 18 07:06:03 2003
***************
*** 0 ****
--- 1,149 ----
+ //===-- TwoAddressInstructionPass.cpp - Two-Address instruction pass ------===//
+ //
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ //
+ //===----------------------------------------------------------------------===//
+ //
+ // This file implements the LiveInterval analysis pass which is used
+ // by the Linear Scan Register allocator. This pass linearizes the
+ // basic blocks of the function in DFS order and uses the
+ // LiveVariables pass to conservatively compute live intervals for
+ // each virtual and physical register.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #define DEBUG_TYPE "twoaddrinstr"
+ #include "llvm/CodeGen/TwoAddressInstructionPass.h"
+ #include "llvm/Function.h"
+ #include "llvm/CodeGen/LiveVariables.h"
+ #include "llvm/CodeGen/MachineFrameInfo.h"
+ #include "llvm/CodeGen/MachineFunctionPass.h"
+ #include "llvm/CodeGen/MachineInstr.h"
+ #include "llvm/CodeGen/Passes.h"
+ #include "llvm/CodeGen/SSARegMap.h"
+ #include "llvm/Target/MRegisterInfo.h"
+ #include "llvm/Target/TargetInstrInfo.h"
+ #include "llvm/Target/TargetMachine.h"
+ #include "llvm/Target/TargetRegInfo.h"
+ #include "Support/Debug.h"
+ #include "Support/Statistic.h"
+ #include "Support/STLExtras.h"
+ #include 
+ 
+ using namespace llvm;
+ 
+ namespace {
+     RegisterAnalysis X(
+         "twoaddressinstruction", "Two-Address instruction pass");
+ 
+     Statistic<> numTwoAddressInstrs("twoaddressinstruction",
+                                     "Number of two-address instructions");
+     Statistic<> numInstrsAdded("twoaddressinstruction",
+                                "Number of instructions added");
+ };
+ 
+ void TwoAddressInstructionPass::getAnalysisUsage(AnalysisUsage &AU) const
+ {
+     AU.addPreserved();
+     AU.addRequired();
+     AU.addPreservedID(PHIEliminationID);
+     AU.addRequiredID(PHIEliminationID);
+     MachineFunctionPass::getAnalysisUsage(AU);
+ }
+ 
+ /// runOnMachineFunction - Reduce two-address instructions to two
+ /// operands
+ ///
+ bool TwoAddressInstructionPass::runOnMachineFunction(MachineFunction &fn) {
+     DEBUG(std::cerr << "Machine Function\n");
+     mf_ = &fn;
+     tm_ = &fn.getTarget();
+     mri_ = tm_->getRegisterInfo();
+     lv_ = &getAnalysis();
+ 
+     const TargetInstrInfo& tii = tm_->getInstrInfo();
+ 
+     for (MachineFunction::iterator mbbi = mf_->begin(), mbbe = mf_->end();
+          mbbi != mbbe; ++mbbi) {
+         for (MachineBasicBlock::iterator mii = mbbi->begin();
+              mii != mbbi->end(); ++mii) {
+             MachineInstr* mi = *mii;
+ 
+             unsigned opcode = mi->getOpcode();
+             // ignore if it is not a two-address instruction
+             if (!tii.isTwoAddrInstr(opcode))
+                 continue;
+ 
+             ++numTwoAddressInstrs;
+ 
+             DEBUG(std::cerr << "\tinstruction: "; mi->print(std::cerr, *tm_));
+ 
+             // we have nothing to do if the two operands are the same
+             if (mi->getOperand(0).getAllocatedRegNum() ==
+                 mi->getOperand(1).getAllocatedRegNum())
+                 continue;
+ 
+             assert(mi->getOperand(1).isRegister() &&
+                    mi->getOperand(1).getAllocatedRegNum() &&
+                    mi->getOperand(1).isUse() &&
+                    "two address instruction invalid");
+ 
+             // rewrite:
+             //     a = b op c
+             // to:
+             //     a = b
+             //     a = a op c
+             unsigned regA = mi->getOperand(0).getAllocatedRegNum();
+             unsigned regB = mi->getOperand(1).getAllocatedRegNum();
+             bool regAisPhysical = regA < MRegisterInfo::FirstVirtualRegister;
+             bool regBisPhysical = regB < MRegisterInfo::FirstVirtualRegister;
+ 
+             const TargetRegisterClass* rc = regAisPhysical ?
+                 mri_->getRegClass(regA) :
+                 mf_->getSSARegMap()->getRegClass(regA);
+ 
+             numInstrsAdded += mri_->copyRegToReg(*mbbi, mii, regA, regB, rc);
+ 
+             MachineInstr* prevMi = *(mii - 1);
+             DEBUG(std::cerr << "\t\tadded instruction: ";
+                   prevMi->print(std::cerr, *tm_));
+ 
+             // update live variables for regA
+             if (regAisPhysical) {
+                 lv_->HandlePhysRegDef(regA, prevMi);
+             }
+             else {
+                 LiveVariables::VarInfo& varInfo = lv_->getVarInfo(regA);
+                 varInfo.DefInst = prevMi;
+             }
+ 
+             // update live variables for regB
+             if (regBisPhysical) {
+                 lv_->HandlePhysRegUse(regB, prevMi);
+             }
+             else {
+                 if (lv_->removeVirtualRegisterKilled(regB, &*mbbi, mi))
+                     lv_->addVirtualRegisterKilled(regB, &*mbbi, prevMi);
+ 
+                 if (lv_->removeVirtualRegisterDead(regB, &*mbbi, mi))
+                     lv_->addVirtualRegisterDead(regB, &*mbbi, prevMi);
+             }
+ 
+             // replace all occurences of regB with regA
+             for (unsigned i = 1; i < mi->getNumOperands(); ++i) {
+                 if (mi->getOperand(i).isRegister() &&
+                     mi->getOperand(i).getReg() == regB)
+                     mi->SetMachineOperandReg(i, regA);
+             }
+             DEBUG(std::cerr << "\t\tmodified original to: ";
+                   mi->print(std::cerr, *tm_));
+             assert(mi->getOperand(0).getAllocatedRegNum() ==
+                    mi->getOperand(1).getAllocatedRegNum());
+         }
+     }
+ 
+     return numInstrsAdded != 0;
+ }




From alkis at cs.uiuc.edu  Thu Dec 18 07:10:05 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:10:05 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/RegAllocLocal.cpp
Message-ID: <200312181309.HAA29127@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

RegAllocLocal.cpp updated: 1.34 -> 1.35

---
Log message:

Modify local register allocator to use the two-address instruction pass.


---
Diffs of the changes:  (+5 -18)

Index: llvm/lib/CodeGen/RegAllocLocal.cpp
diff -u llvm/lib/CodeGen/RegAllocLocal.cpp:1.34 llvm/lib/CodeGen/RegAllocLocal.cpp:1.35
--- llvm/lib/CodeGen/RegAllocLocal.cpp:1.34	Sun Dec 14 07:24:16 2003
+++ llvm/lib/CodeGen/RegAllocLocal.cpp	Thu Dec 18 07:08:52 2003
@@ -19,6 +19,7 @@
 #include "llvm/CodeGen/SSARegMap.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/LiveVariables.h"
+#include "llvm/CodeGen/TwoAddressInstructionPass.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
 #include "Support/CommandLine.h"
@@ -112,6 +113,7 @@
       if (!DisableKill)
 	AU.addRequired();
       AU.addRequiredID(PHIEliminationID);
+      AU.addRequired();
       MachineFunctionPass::getAnalysisUsage(AU);
     }
 
@@ -571,28 +573,13 @@
         unsigned DestVirtReg = MI->getOperand(i).getAllocatedRegNum();
         unsigned DestPhysReg;
 
-        // If DestVirtReg already has a value, forget about it.  Why doesn't
-        // getReg do this right?
+        // If DestVirtReg already has a value, use it.
         std::map::iterator DestI =
           Virt2PhysRegMap.find(DestVirtReg);
         if (DestI != Virt2PhysRegMap.end()) {
-          unsigned PhysReg = DestI->second;
-          Virt2PhysRegMap.erase(DestI);
-          removePhysReg(PhysReg);
+          DestPhysReg = DestI->second;
         }
-
-        if (TM->getInstrInfo().isTwoAddrInstr(MI->getOpcode()) && i == 0) {
-          // must be same register number as the first operand
-          // This maps a = b + c into b += c, and saves b into a's spot
-          assert(MI->getOperand(1).isPhysicalRegister()  &&
-                 MI->getOperand(1).getAllocatedRegNum() &&
-                 MI->getOperand(1).isUse() &&
-                 "Two address instruction invalid!");
-          DestPhysReg = MI->getOperand(1).getAllocatedRegNum();
-
-          liberatePhysReg(MBB, I, DestPhysReg);
-          assignVirtToPhysReg(DestVirtReg, DestPhysReg);
-        } else {
+        else {
           DestPhysReg = getReg(MBB, I, DestVirtReg);
         }
         markVirtRegModified(DestVirtReg);




From alkis at cs.uiuc.edu  Thu Dec 18 07:13:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:13:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/RegAllocLinearScan.cpp
Message-ID: <200312181312.HAA29179@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

RegAllocLinearScan.cpp updated: 1.11 -> 1.12

---
Log message:

Fix bug in reserved registers. DH actually aliases DX and EDX which
are not reserved registers.


---
Diffs of the changes:  (+2 -2)

Index: llvm/lib/CodeGen/RegAllocLinearScan.cpp
diff -u llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.11 llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.12
--- llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.11	Thu Dec 18 02:56:11 2003
+++ llvm/lib/CodeGen/RegAllocLinearScan.cpp	Thu Dec 18 07:12:18 2003
@@ -227,14 +227,14 @@
 
     // reserve R32: EDI, EBX,
     //         R16:  DI,  BX,
-    //         R8:   DH,  BH,
+    //         R8:   BH,  BL
     //         RFP: FP5, FP6
     reserved_.push_back(19); /* EDI */
     reserved_.push_back(17); /* EBX */
     reserved_.push_back(12); /*  DI */
     reserved_.push_back( 7); /*  BX */
-    reserved_.push_back(11); /*  DH */
     reserved_.push_back( 4); /*  BH */
+    reserved_.push_back( 5); /*  BL */
     reserved_.push_back(28); /* FP5 */
     reserved_.push_back(29); /* FP6 */
 




From alkis at cs.uiuc.edu  Thu Dec 18 07:16:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 07:16:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/RegAllocLinearScan.cpp LiveIntervals.cpp
Message-ID: <200312181315.HAA29233@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

RegAllocLinearScan.cpp updated: 1.12 -> 1.13
LiveIntervals.cpp updated: 1.10 -> 1.11

---
Log message:

Modify linear scan register allocator to use the two-address
instruction pass. This also fixes all remaining bugs for this new
allocator to pass all tests under test/Programs.


---
Diffs of the changes:  (+39 -76)

Index: llvm/lib/CodeGen/RegAllocLinearScan.cpp
diff -u llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.12 llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.13
--- llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.12	Thu Dec 18 07:12:18 2003
+++ llvm/lib/CodeGen/RegAllocLinearScan.cpp	Thu Dec 18 07:15:02 2003
@@ -84,6 +84,11 @@
         /// runOnMachineFunction - register allocate the whole function
         bool runOnMachineFunction(MachineFunction&);
 
+        /// verifyIntervals - verify that we have no inconsistencies
+        /// in the register assignments we have in active and inactive
+        /// lists
+        bool verifyIntervals();
+
         /// processActiveIntervals - expire old intervals and move
         /// non-overlapping ones to the incative list
         void processActiveIntervals(Intervals::const_iterator cur);
@@ -245,6 +250,8 @@
 
         DEBUG(printIntervals("\tactive", active_.begin(), active_.end()));
         DEBUG(printIntervals("\tinactive", inactive_.begin(), inactive_.end()));
+        assert(verifyIntervals());
+
         processActiveIntervals(i);
         // processInactiveIntervals(i);
 
@@ -321,7 +328,7 @@
             for (unsigned i = 0, e = (*currentInstr_)->getNumOperands();
                  i != e; ++i) {
                 MachineOperand& op = (*currentInstr_)->getOperand(i);
-                if (op.isVirtualRegister() && op.isUse()) {
+                if (op.isVirtualRegister() && op.isUse() && !op.isDef()) {
                     unsigned virtReg = op.getAllocatedRegNum();
                     unsigned physReg = v2pMap_[virtReg];
                     if (!physReg) {
@@ -361,81 +368,6 @@
                 }
             }
 
-
-            // if the instruction is a two address instruction and the
-            // source operands are not identical we need to insert
-            // extra instructions.
-
-            unsigned opcode = (*currentInstr_)->getOpcode();
-            if (tm_->getInstrInfo().isTwoAddrInstr(opcode) &&
-                (*currentInstr_)->getOperand(0).getAllocatedRegNum() !=
-                (*currentInstr_)->getOperand(1).getAllocatedRegNum()) {
-                assert((*currentInstr_)->getOperand(1).isRegister() &&
-                       (*currentInstr_)->getOperand(1).getAllocatedRegNum() &&
-                       (*currentInstr_)->getOperand(1).isUse() &&
-                       "Two address instruction invalid");
-
-                unsigned regA =
-                    (*currentInstr_)->getOperand(0).getAllocatedRegNum();
-                unsigned regB =
-                    (*currentInstr_)->getOperand(1).getAllocatedRegNum();
-                unsigned regC =
-                    ((*currentInstr_)->getNumOperands() > 2 &&
-                     (*currentInstr_)->getOperand(2).isRegister()) ?
-                    (*currentInstr_)->getOperand(2).getAllocatedRegNum() :
-                    0;
-
-                const TargetRegisterClass* rc = mri_->getRegClass(regA);
-
-                // special case: "a = b op a". If b is a temporary
-                // reserved register rewrite as: "b = b op a; a = b"
-                // otherwise use a temporary reserved register t and
-                // rewrite as: "t = b; t = t op a; a = t"
-                if (regC && regA == regC) {
-                    // b is a temp reserved register
-                    if (find(reserved_.begin(), reserved_.end(),
-                             regB) != reserved_.end()) {
-                        (*currentInstr_)->SetMachineOperandReg(0, regB);
-                        ++currentInstr_;
-                        instrAdded_ += mri_->copyRegToReg(*currentMbb_,
-                                                          currentInstr_,
-                                                          regA,
-                                                          regB,
-                                                          rc);
-                        --currentInstr_;
-                    }
-                    // b is just a normal register
-                    else {
-                        unsigned tempReg = getFreeTempPhysReg(rc);
-                        assert (tempReg &&
-                                "no free temp reserved physical register?");
-                        instrAdded_ += mri_->copyRegToReg(*currentMbb_,
-                                                          currentInstr_,
-                                                          tempReg,
-                                                          regB,
-                                                          rc);
-                        (*currentInstr_)->SetMachineOperandReg(0, tempReg);
-                        (*currentInstr_)->SetMachineOperandReg(1, tempReg);
-                        ++currentInstr_;
-                        instrAdded_ += mri_->copyRegToReg(*currentMbb_,
-                                                          currentInstr_,
-                                                          regA,
-                                                          tempReg,
-                                                          rc);
-                        --currentInstr_;
-                    }
-                }
-                // "a = b op c" gets rewritten to "a = b; a = a op c"
-                else {
-                    instrAdded_ += mri_->copyRegToReg(*currentMbb_,
-                                                      currentInstr_,
-                                                      regA,
-                                                      regB,
-                                                      rc);
-                    (*currentInstr_)->SetMachineOperandReg(1, regA);
-                }
-            }
-
             DEBUG(std::cerr << "\t\tspilling temporarily defined operands "
                   "of this instruction:\n");
             ++currentInstr_; // we want to insert after this instruction
@@ -452,6 +384,35 @@
         }
     }
 
+    return true;
+}
+
+bool RA::verifyIntervals()
+{
+    std::set assignedRegisters;
+    for (IntervalPtrs::iterator i = active_.begin(); i != active_.end(); ++i) {
+        if ((*i)->reg >= MRegisterInfo::FirstVirtualRegister) {
+            unsigned reg = v2pMap_.find((*i)->reg)->second;
+
+            bool inserted = assignedRegisters.insert(reg).second;
+            assert(inserted && "registers in active list conflict");
+        }
+    }
+
+    for (IntervalPtrs::iterator i = active_.begin(); i != active_.end(); ++i) {
+        unsigned reg = (*i)->reg;
+        if (reg >= MRegisterInfo::FirstVirtualRegister) {
+            reg = v2pMap_.find((*i)->reg)->second;
+        }
+
+        for (const unsigned* as = mri_->getAliasSet(reg); *as; ++as) {
+            assert(assignedRegisters.find(*as) == assignedRegisters.end() &&
+                   "registers in active list alias each other");
+        }
+    }
+
+    // TODO: add checks between active and inactive and make sure we
+    // do not overlap anywhere
     return true;
 }
 


Index: llvm/lib/CodeGen/LiveIntervals.cpp
diff -u llvm/lib/CodeGen/LiveIntervals.cpp:1.10 llvm/lib/CodeGen/LiveIntervals.cpp:1.11
--- llvm/lib/CodeGen/LiveIntervals.cpp:1.10	Thu Dec 18 02:53:43 2003
+++ llvm/lib/CodeGen/LiveIntervals.cpp	Thu Dec 18 07:15:02 2003
@@ -24,6 +24,7 @@
 #include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/CodeGen/SSARegMap.h"
+#include "llvm/CodeGen/TwoAddressInstructionPass.h"
 #include "llvm/Target/MRegisterInfo.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
@@ -49,6 +50,7 @@
     AU.addRequired();
     AU.addPreservedID(PHIEliminationID);
     AU.addRequiredID(PHIEliminationID);
+    AU.addRequired();
     MachineFunctionPass::getAnalysisUsage(AU);
 }
 




From criswell at cs.uiuc.edu  Thu Dec 18 10:42:04 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:42:04 2003
Subject: [llvm-commits] CVS: llvm/autoconf/configure.ac
Message-ID: <200312181641.KAA05027@choi.cs.uiuc.edu>


Changes in directory llvm/autoconf:

configure.ac updated: 1.61 -> 1.62

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+1 -1)

Index: llvm/autoconf/configure.ac
diff -u llvm/autoconf/configure.ac:1.61 llvm/autoconf/configure.ac:1.62
--- llvm/autoconf/configure.ac:1.61	Fri Dec 12 14:09:33 2003
+++ llvm/autoconf/configure.ac	Thu Dec 18 10:41:17 2003
@@ -15,7 +15,7 @@
 dnl **************************************************************************
 dnl * Initialize
 dnl **************************************************************************
-AC_INIT([[[LLVM]]],[[[1.0]]],[llvmbugs at cs.uiuc.edu])
+AC_INIT([[[LLVM]]],[[[1.1]]],[llvmbugs at cs.uiuc.edu])
 
 dnl Place all of the extra autoconf files into the config subdirectory
 AC_CONFIG_AUX_DIR([autoconf])




From criswell at cs.uiuc.edu  Thu Dec 18 10:42:12 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:42:12 2003
Subject: [llvm-commits] CVS: llvm/configure
Message-ID: <200312181641.KAA05020@choi.cs.uiuc.edu>


Changes in directory llvm:

configure updated: 1.61 -> 1.62

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+9 -9)

Index: llvm/configure
diff -u llvm/configure:1.61 llvm/configure:1.62
--- llvm/configure:1.61	Fri Dec 12 14:15:23 2003
+++ llvm/configure	Thu Dec 18 10:41:15 2003
@@ -1,6 +1,6 @@
 #! /bin/sh
 # Guess values for system-dependent variables and create Makefiles.
-# Generated by GNU Autoconf 2.57 for [LLVM] [1.0].
+# Generated by GNU Autoconf 2.57 for [LLVM] [1.1].
 #
 # Report bugs to .
 #
@@ -422,8 +422,8 @@
 # Identity of this package.
 PACKAGE_NAME='[LLVM]'
 PACKAGE_TARNAME='--llvm--'
-PACKAGE_VERSION='[1.0]'
-PACKAGE_STRING='[LLVM] [1.0]'
+PACKAGE_VERSION='[1.1]'
+PACKAGE_STRING='[LLVM] [1.1]'
 PACKAGE_BUGREPORT='llvmbugs at cs.uiuc.edu'
 
 ac_subdirs_all="$ac_subdirs_all projects/${i}"
@@ -954,7 +954,7 @@
   # Omit some internal or obsolete options to make the list less imposing.
   # This message is too long to be a string in the A/UX 3.1 sh.
   cat <<_ACEOF
-\`configure' configures [LLVM] [1.0] to adapt to many kinds of systems.
+\`configure' configures [LLVM] [1.1] to adapt to many kinds of systems.
 
 Usage: $0 [OPTION]... [VAR=VALUE]...
 
@@ -1016,7 +1016,7 @@
 
 if test -n "$ac_init_help"; then
   case $ac_init_help in
-     short | recursive ) echo "Configuration of [LLVM] [1.0]:";;
+     short | recursive ) echo "Configuration of [LLVM] [1.1]:";;
    esac
   cat <<\_ACEOF
 
@@ -1131,7 +1131,7 @@
 test -n "$ac_init_help" && exit 0
 if $ac_init_version; then
   cat <<\_ACEOF
-[LLVM] configure [1.0]
+[LLVM] configure [1.1]
 generated by GNU Autoconf 2.57
 
 Copyright 1992, 1993, 1994, 1995, 1996, 1998, 1999, 2000, 2001, 2002
@@ -1146,7 +1146,7 @@
 This file contains any messages produced by compilers while
 running configure, to aid debugging if configure makes a mistake.
 
-It was created by [LLVM] $as_me [1.0], which was
+It was created by [LLVM] $as_me [1.1], which was
 generated by GNU Autoconf 2.57.  Invocation command line was
 
   $ $0 $@
@@ -23279,7 +23279,7 @@
 } >&5
 cat >&5 <<_CSEOF
 
-This file was extended by [LLVM] $as_me [1.0], which was
+This file was extended by [LLVM] $as_me [1.1], which was
 generated by GNU Autoconf 2.57.  Invocation command line was
 
   CONFIG_FILES    = $CONFIG_FILES
@@ -23342,7 +23342,7 @@
 
 cat >>$CONFIG_STATUS <<_ACEOF
 ac_cs_version="\\
-[LLVM] config.status [1.0]
+[LLVM] config.status [1.1]
 configured by $0, generated by GNU Autoconf 2.57,
   with options \\"`echo "$ac_configure_args" | sed 's/[\\""\`\$]/\\\\&/g'`\\"
 




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:00 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:00 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/External/SPEC/Makefile
Message-ID: <200312181642.KAA05074@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/External/SPEC:

Makefile updated: 1.4 -> 1.5

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/External/SPEC/Makefile
diff -u llvm/test/Programs/External/SPEC/Makefile:1.4 llvm/test/Programs/External/SPEC/Makefile:1.5
--- llvm/test/Programs/External/SPEC/Makefile:1.4	Tue Aug 19 16:30:30 2003
+++ llvm/test/Programs/External/SPEC/Makefile	Thu Dec 18 10:42:05 2003
@@ -1,3 +1,3 @@
 LEVEL = ../../../..
-DIRS := $(sort $(filter-out CVS/, $(wildcard */)))
+DIRS := CFP2000 CINT2000
 include ${LEVEL}/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:07 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:07 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile
Message-ID: <200312181642.KAA05094@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/Ptrdist:

Makefile updated: 1.3 -> 1.4

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.4
--- llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile:1.3	Fri Sep 12 11:02:35 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/Ptrdist/Makefile	Thu Dec 18 10:42:06 2003
@@ -1,6 +1,7 @@
 # MultiSource/Ptrdist Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := anagram bc ft ks yacr2
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:14 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:14 2003
Subject: [llvm-commits] CVS: llvm/test/QMTest/expectations.linux.qmr expectations.sunos.qmr
Message-ID: <200312181642.KAA05101@choi.cs.uiuc.edu>


Changes in directory llvm/test/QMTest:

expectations.linux.qmr updated: 1.5 -> 1.6
expectations.sunos.qmr updated: 1.5 -> 1.6

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+2 -2)

Index: llvm/test/QMTest/expectations.linux.qmr
diff -u llvm/test/QMTest/expectations.linux.qmr:1.5 llvm/test/QMTest/expectations.linux.qmr:1.6
--- llvm/test/QMTest/expectations.linux.qmr:1.5	Thu Dec 11 17:18:33 2003
+++ llvm/test/QMTest/expectations.linux.qmr	Thu Dec 18 10:42:06 2003
@@ -3,13 +3,13 @@
 qoq}q(U
_Result__kindqUtestqU_Result__outcomeqUPASSqU_Result__annotationsq}U_Result__idq	U0Regression.Assembler.2002-08-16-ConstExprInlinedq
 U_Result__contextq(cqm.test.context
 Context
-qo}q
(U_Context__propertiesq}U_Context__temporariesq}ubub.(hoq}q(hhhhh}h	U(Regression.Transforms.PruneEH.simpletestqh(ho}q(h}h}ubub.(hoq}q(hhhhh}h	U/Regression.CBackend.2002-08-19-HardConstantExprqh(ho}q(h}h}ubub.(hoq}q(hhhhh}h	URegression.Jello.test-shiftqh(ho}q(h}h}ubub.(hoq}q(hhhUFAILqh}h	U.Regression.Transforms.CorrelatedExprs.looptestqh(ho}q (h}h}ubub.(hoq!}q"(hhhhh}h	U-Regression.Linker.2003-08-23-GlobalVarLinkingq#h(ho}q$(h}h}ubub.(hoq%}q&(hhhhh}h	U?Regression.Transforms.Inline.2003-09-22-PHINodesInExceptionDestq'h(ho}q((h}h}ubub.(hoq)}q*(hhhhh}h	U,Regression.CFrontend.2003-02-12-NonlocalGotoq+h(ho}q,(h}h}ubub.(hoq-}q.(hhhhh}h	U2Regression.Transforms.FunctionResolve.retmismatch1q/h(ho}q0(h}h}ubub.(hoq1}q2(hhhhh}h	U8Regression.Transforms.ADCE.2003-01-22-PredecessorProblemq3h(ho}q4(h}h}ubub.(hoq5}q6(hhhhh}h	U=Regression.Transforms.LevelRaise.2002-10-08-!
 VarArgCallInfLoopq7h(ho}q8(h}h}ubub.(hoq9}q:(hhhhh}h	U5Regression.CFrontend.2003-07-22-ArrayAccessTypeSafetyq;h(ho}q<(h}h}ubub.(hoq=}q>(hhhhh}h	U8Regression.C++Frontend.2003-09-29-ArgumentNumberMismatchq?h(ho}q@(h}h}ubub.(hoqA}qB(hhhhh}h	U+Regression.CFrontend.2002-09-19-StarInLabelqCh(ho}qD(h}h}ubub.(hoqE}qF(hhhhh}h	U.Regression.CFrontend.2003-08-23-LocalUnionTestqGh(ho}qH(h}h}ubub.(hoqI}qJ(hhhhh}h	U-Regression.C++Frontend.2003-08-28-SaveExprBugqKh(ho}qL(h}h}ubub.(hoqM}qN(hhhhh}h	U"Regression.Jello.2003-06-05-PHIBugqOh(ho}qP(h}h}ubub.(hoqQ}qR(hhhhh}h	U/Regression.Linker.2003-04-26-NullPtrLinkProblemqSh(ho}qT(h}h}ubub.(hoqU}qV(hhhhh}h	U)Regression.Transforms.Reassociate.subtestqWh(ho}qX(h}h}ubub.(hoqY}qZ(hhhhh}h	U)Regression.Linker.2002-08-20-ConstantExprq[h(ho}q\(h}h}ubub.(hoq]}q^(hhhhh}h	U=Regression.Transforms.Reassociate.2002-05-15-AgressiveSubMoveq_h(ho}q`(h}h}ubub.(!
 hoqa}qb(hhhhh}h	UARegression.Transforms.CorrelatedExprs.2002-10-
07-DominatorProblemqch(ho}qd(h}h}ubub.(hoqe}qf(hhhhh}h	U>Regression.Transforms.InstCombine.2003-06-22-ConstantExprCrashqgh(ho}qh(h}h}ubub.(hoqi}qj(hhhhh}h	U3Regression.Transforms.PiNodeInserter.substitutetestqkh(ho}ql(h}h}ubub.(hoqm}qn(hhhhh}h	U4Regression.Transforms.SCCP.2003-08-26-InvokeHandlingqoh(ho}qp(h}h}ubub.(hoqq}qr(hhhhh}h	U$Regression.Jello.2003-01-04-LoopTestqsh(ho}qt(h}h}ubub.(hoqu}qv(hhhhh}h	U1Regression.Assembler.2002-04-04-PureVirtMethCall2qwh(ho}qx(h}h}ubub.(hoqy}qz(hhhhh}h	U)Regression.CFrontend.2002-07-14-MiscTestsq{h(ho}q|(h}h}ubub.(hoq}}q~(hhhhh}h	U(hhhhh}h	U8Regression.C++Frontend.2003-09-29-ArgumentNumberMismatchq?h(ho}q@(h}h}ubub.(hoqA}qB(hhhhh}h	U+Regression.CFrontend.2002-09-19-StarInLabelqCh(ho}qD(h}h}ubub.(hoqE}qF(hhhhh}h	U.Regression.CFrontend.2003-08-23-LocalUnionTestqGh(ho}qH(h}h}ubub.(hoqI}qJ(hhhhh}h	U-Regression.C++Frontend.2003-08-28-SaveExprBugqKh(ho}qL(h}h}ubub.(hoqM}qN(hhhhh}h	U"Regression.Jello.2003-06-05-PHIBugqOh(ho}qP(h}h}ubub.(hoqQ}qR(hhhhh}h	U/Regression.Linker.2003-04-26-NullPtrLinkProblemqSh(ho}qT(h}h}ubub.(hoqU}qV(hhhhh}h	U)Regression.Transforms.Reassociate.subtestqWh(ho}qX(h}h}ubub.(hoqY}qZ(hhhhh}h	U)Regression.Linker.2002-08-20-ConstantExprq[h(ho}q\(h}h}ubub.(hoq]}q^(hhhhh}h	U=Regression.Transforms.Reassociate.2002-05-15-AgressiveSubMoveq_h(ho}q`(h}h}ubub.(hoqa}qb(hhhhh}h!
 	UARegression.Transforms.CorrelatedExprs.2002-10-07-DominatorProblemqc
h(ho}qd(h}h}ubub.(hoqe}qf(hhhhh}h	U3Regression.Transforms.PiNodeInserter.substitutetestqgh(ho}qh(h}h}ubub.(hoqi}qj(hhhhh}h	U4Regression.Transforms.SCCP.2003-08-26-InvokeHandlingqkh(ho}ql(h}h}ubub.(hoqm}qn(hhhhh}h	U$Regression.Jello.2003-01-04-LoopTestqoh(ho}qp(h}h}ubub.(hoqq}qr(hhhhh}h	U1Regression.Assembler.2002-04-04-PureVirtMethCall2qsh(ho}qt(h}h}ubub.(hoqu}qv(hhhhh}h	U)Regression.CFrontend.2002-07-14-MiscTestsqwh(ho}qx(h}h}ubub.(hoqy}qz(hhhhh}h	U(hhhhh}h	U Regression.Reoptimizer.ticm.ticmq?h(ho}q@(h}h}ubub.(hoqA}qB(hhhhh}h	U8Regression.C++Frontend.2003-09-29-ArgumentNumberMismatchqCh(ho}qD(h}h}ubub.(hoqE}qF(hhhhh}h	U+Regression.CFrontend.2002-09-19-StarInLabelqGh(ho}qH(h}h}ubub.(hoqI}qJ(hhhhh}h	U.Regression.CFrontend.2003-08-23-LocalUnionTestqKh(ho}qL(h}h}ubub.(hoqM}qN(hhhhh}h	U/Regression.Linker.2003-04-26-NullPtrLinkProblemqOh(ho}qP(h}h}ubub.(hoqQ}qR(hhhhh}h	U)Regression.Transforms.Reassociate.subtestqSh(ho}qT(h}h}ubub.(hoqU}qV(hhhhh}h	U)Regression.Linker.2002-08-20-ConstantExprqWh(ho}qX(h}h}ubub.(hoqY}qZ(hhhhh}h	U=Regression.Transforms.Reassociate.2002-05-15-AgressiveSubMoveq[h(ho}q\(h}h}ubub.(hoq]}q^(hhhhh}h	UARegression.Transforms.CorrelatedExprs.2002-10-07-DominatorProblemq_!
 h(ho}q`(h}h}ubub.(hoqa}qb(hhhhh}h	U3Regression.Transforms.Pi
NodeInserter.substitutetestqch(ho}qd(h}h}ubub.(hoqe}qf(hhhhh}h	U4Regression.Transforms.SCCP.2003-08-26-InvokeHandlingqgh(ho}qh(h}h}ubub.(hoqi}qj(hhhhh}h	U-Regression.CFrontend.2002-02-18-64bitConstantqkh(ho}ql(h}h}ubub.(hoqm}qn(hhhhh}h	U1Regression.Assembler.2002-04-04-PureVirtMethCall2qoh(ho}qp(h}h}ubub.(hoqq}qr(hhhhh}h	U(Regression.Reoptimizer.BinInterface.testqsh(ho}qt(h}h}ubub.(hoqu}qv(hhhhh}h	U(hhhhh}h	U8Regression.C++Frontend.2003-09-29-ArgumentNumberMismatchq?h(ho}q@(h}h}ubub.(hoqA}qB(hhhhh}h	U+Regression.CFrontend.2002-09-19-StarInLabelqCh(ho}qD(h}h}ubub.(hoqE}qF(hhhhh}h	U.Regression.CFrontend.2003-08-23-LocalUnionTestqGh(ho}qH(h}h}ubub.(hoqI}qJ(hhhhh}h	U-Regression.C++Frontend.2003-08-28-SaveExprBugqKh(ho}qL(h}h}ubub.(hoqM}qN(hhhhh}h	U"Regression.Jello.2003-06-05-PHIBugqOh(ho}qP(h}h}ubub.(hoqQ}qR(hhhhh}h	U/Regression.Linker.2003-04-26-NullPtrLinkProblemqSh(ho}qT(h}h}ubub.(hoqU}qV(hhhhh}h	U)Regression.Transforms.Reassociate.subtestqWh(ho}qX(h}h}ubub.(hoqY}qZ(hhhhh}h	U)Regression.Linker.2002-08-20-ConstantExprq[h(ho}q\(h}h}ubub.(hoq]}q^(hhhhh}h	U=Regression.Transforms.Reassociate.2002-05-15-AgressiveSubMoveq_h(ho}q`(h}h}ubub.(hoqa}qb(hhhhh}h!
 	UARegression.Transforms.CorrelatedExprs.2002-10-07-DominatorProblemqc
h(ho}qd(h}h}ubub.(hoqe}qf(hhhhh}h	U3Regression.Transforms.PiNodeInserter.substitutetestqgh(ho}qh(h}h}ubub.(hoqi}qj(hhhhh}h	U4Regression.Transforms.SCCP.2003-08-26-InvokeHandlingqkh(ho}ql(h}h}ubub.(hoqm}qn(hhhhh}h	U$Regression.Jello.2003-01-04-LoopTestqoh(ho}qp(h}h}ubub.(hoqq}qr(hhhhh}h	U1Regression.Assembler.2002-04-04-PureVirtMethCall2qsh(ho}qt(h}h}ubub.(hoqu}qv(hhhhh}h	U)Regression.CFrontend.2002-07-14-MiscTestsqwh(ho}qx(h}h}ubub.(hoqy}qz(hhhhh}h	U


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/Olden:

Makefile updated: 1.3 -> 1.4

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.4
--- llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile:1.3	Fri Sep 12 11:02:21 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/Olden/Makefile	Thu Dec 18 10:42:06 2003
@@ -1,6 +1,7 @@
 # MultiSource/Olden Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := bh em3d mst power tsp bisort health perimeter treeadd voronoi
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:27 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:27 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/External/Makefile
Message-ID: <200312181642.KAA05053@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/External:

Makefile updated: 1.3 -> 1.4

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/External/Makefile
diff -u llvm/test/Programs/External/Makefile:1.3 llvm/test/Programs/External/Makefile:1.4
--- llvm/test/Programs/External/Makefile:1.3	Tue Aug 19 16:30:28 2003
+++ llvm/test/Programs/External/Makefile	Thu Dec 18 10:41:58 2003
@@ -8,7 +8,7 @@
 #
 # Create the list of directories to compile
 #
-DIRS := $(sort $(filter-out CVS/, $(wildcard */)))
+DIRS := SPEC
 ifndef USE_SPEC
 DIRS := $(filter-out SPEC/, $(DIRS))
 endif




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:33 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:33 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile
Message-ID: <200312181642.KAA05084@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/McCat:

Makefile updated: 1.3 -> 1.4

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.3 llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.4
--- llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile:1.3	Fri Sep 12 11:01:59 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/McCat/Makefile	Thu Dec 18 10:42:06 2003
@@ -1,6 +1,7 @@
 # MultiSource/McCat Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := 01-qbsort 04-bisect 08-main 12-IOtest 17-bintr 03-testtrie 05-eks 09-vor 15-trie 18-imp
+
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Thu Dec 18 10:43:40 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:43:40 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile
Message-ID: <200312181642.KAA05079@choi.cs.uiuc.edu>


Changes in directory llvm/test/Programs/MultiSource/Benchmarks/FreeBench:

Makefile updated: 1.1 -> 1.2

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+1 -1)

Index: llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile
diff -u llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.1 llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.2
--- llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile:1.1	Sat Oct 11 16:18:41 2003
+++ llvm/test/Programs/MultiSource/Benchmarks/FreeBench/Makefile	Thu Dec 18 10:42:06 2003
@@ -1,6 +1,6 @@
 # MultiSource/Olden Makefile:  Build all subdirectories automatically
 
 LEVEL = ../../../../..
-PARALLEL_DIRS  := $(filter-out %-disabled/, $(sort $(filter-out CVS/, $(wildcard */))))
+PARALLEL_DIRS  := distray mason pcompress2 analyzer fourinarow neural pifft
 
 include $(LEVEL)/test/Programs/Makefile.programs




From criswell at cs.uiuc.edu  Thu Dec 18 10:44:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:44:01 2003
Subject: [llvm-commits] CVS: llvm/runtime/GCCLibraries/crtend/listend.ll
Message-ID: <200312181643.KAA05236@choi.cs.uiuc.edu>


Changes in directory llvm/runtime/GCCLibraries/crtend:

listend.ll updated: 1.3 -> 1.4

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+0 -3)

Index: llvm/runtime/GCCLibraries/crtend/listend.ll
diff -u llvm/runtime/GCCLibraries/crtend/listend.ll:1.3 llvm/runtime/GCCLibraries/crtend/listend.ll:1.4
--- llvm/runtime/GCCLibraries/crtend/listend.ll:1.3	Thu May 22 14:30:11 2003
+++ llvm/runtime/GCCLibraries/crtend/listend.ll	Thu Dec 18 10:43:17 2003
@@ -1,9 +1,6 @@
 ; global_ctors/global_dtors terminator: this is used to add a terminating null
 ; value to the initialization list.
 
-target endian = little
-target pointersize = 32
-
 %struct..TorRec = type { int, void ()* }
 
 %llvm.global_ctors = appending global [1 x %struct..TorRec] [




From criswell at cs.uiuc.edu  Thu Dec 18 10:44:07 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:44:07 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
Message-ID: <200312181643.KAA05243@choi.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

IndVarSimplify.cpp updated: 1.45 -> 1.46

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+8 -15)

Index: llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
diff -u llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.45 llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.46
--- llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.45	Mon Dec 15 11:34:02 2003
+++ llvm/lib/Transforms/Scalar/IndVarSimplify.cpp	Thu Dec 18 10:42:57 2003
@@ -15,19 +15,19 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Transforms/Scalar.h"
-#include "llvm/Constants.h"
-#include "llvm/Type.h"
-#include "llvm/iPHINode.h"
-#include "llvm/iOther.h"
 #include "llvm/Analysis/InductionVariable.h"
 #include "llvm/Analysis/LoopInfo.h"
+#include "llvm/iPHINode.h"
+#include "llvm/iOther.h"
+#include "llvm/Type.h"
+#include "llvm/Constants.h"
 #include "llvm/Support/CFG.h"
-#include "llvm/Transforms/Utils/Local.h"
 #include "Support/Debug.h"
 #include "Support/Statistic.h"
 #include "Support/STLExtras.h"
 #include 
-using namespace llvm;
+
+namespace llvm {
 
 namespace {
   Statistic<> NumRemoved ("indvars", "Number of aux indvars removed");
@@ -141,8 +141,6 @@
 
     DEBUG(IV->print(std::cerr));
 
-    while (isa(AfterPHIIt)) ++AfterPHIIt;
-
     // Don't do math with pointers...
     const Type *IVTy = IV->Phi->getType();
     if (isa(IVTy)) IVTy = Type::ULongTy;
@@ -188,12 +186,6 @@
       IV->Phi->setName("");
       Val->setName(OldName);
 
-      // Get the incoming values used by the PHI node
-      std::vector PHIOps;
-      PHIOps.reserve(IV->Phi->getNumIncomingValues());
-      for (unsigned i = 0, e = IV->Phi->getNumIncomingValues(); i != e; ++i)
-        PHIOps.push_back(IV->Phi->getIncomingValue(i));
-
       // Delete the old, now unused, phi node...
       Header->getInstList().erase(IV->Phi);
 
@@ -250,7 +242,8 @@
                                            "Canonicalize Induction Variables");
 }
 
-Pass *llvm::createIndVarSimplifyPass() {
+Pass *createIndVarSimplifyPass() {
   return new InductionVariableSimplify();
 }
 
+} // End llvm namespace




From criswell at cs.uiuc.edu  Thu Dec 18 10:44:14 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 10:44:14 2003
Subject: [llvm-commits] CVS: llvm/docs/CFEBuildInstrs.html GettingStarted.html ReleaseNotes.html Stacker.html
Message-ID: <200312181643.KAA05208@choi.cs.uiuc.edu>


Changes in directory llvm/docs:

CFEBuildInstrs.html updated: 1.10 -> 1.11
GettingStarted.html updated: 1.46 -> 1.47
ReleaseNotes.html updated: 1.87 -> 1.88
Stacker.html updated: 1.6 -> 1.7

---
Log message:

Merged in RELEASE_11.



---
Diffs of the changes:  (+276 -121)

Index: llvm/docs/CFEBuildInstrs.html
diff -u llvm/docs/CFEBuildInstrs.html:1.10 llvm/docs/CFEBuildInstrs.html:1.11
--- llvm/docs/CFEBuildInstrs.html:1.10	Wed Dec 10 14:04:13 2003
+++ llvm/docs/CFEBuildInstrs.html	Thu Dec 18 10:43:11 2003
@@ -14,6 +14,7 @@
 
   A Cautionary Note
   
Instructions
+  
License Information
 
 
     
@@ -114,7 +115,23 @@
    --enable-languages=c,c++ --host=sparcv9-sun-solaris2.8
  % gmake all-gcc
  % setenv LLVM_LIB_SEARCH_PATH `pwd`/gcc 
- % gmake all; gmake install
+ % gmake all
+
+
+
+At this point, libstdc++ may fail to build because of wchar errors (look for
+errors that reference vfwscanf or wcstof).  If that happens,
+edit sparcv9-sun-solaris2.8/libstdc++-v3/config.h and comment out the
+line that defines _GLIBCXX_USE_WCHAR_T.
+
+
+
+Then, continue as below:
+
+
++ % gmake all
+ % gmake install
 
 
  Common Problem: You may get error messages regarding the fact
@@ -197,13 +214,55 @@
 
 
 
+
+  License Information
+
+
+
+
+The LLVM GCC frontend is licensed to you under the GNU General Public License
+and the GNU Lesser General Public License.  Please see the files COPYING and
+COPYING.LIB for more details.
+
+
+
+The software also has the following additional copyrights:
+
+
++Copyright (c) 1994
+Hewlett-Packard Company
+
+Permission to use, copy, modify, distribute and sell this software
+and its documentation for any purpose is hereby granted without fee,
+provided that the above copyright notice appear in all copies and
+that both that copyright notice and this permission notice appear
+in supporting documentation.  Hewlett-Packard Company makes no
+representations about the suitability of this software for any
+purpose.  It is provided "as is" without express or implied warranty.
+
+Copyright (c) 1996, 1997, 1998, 1999
+Silicon Graphics Computer Systems, Inc.
+
+Permission to use, copy, modify, distribute and sell this software
+and its documentation for any purpose is hereby granted without fee,
+provided that the above copyright notice appear in all copies and
+that both that copyright notice and this permission notice appear
+in supporting documentation.  Silicon Graphics makes no
+representations about the suitability of this software for any
+purpose.  It is provided "as is" without express or implied warranty.
+
+
+
+
+
 
 
 
   Brian Gaeke
   The LLVM Compiler Infrastructure
   

-  Last modified: $Date: 2003/12/10 20:04:13 $
+  Last modified: $Date: 2003/12/18 16:43:11 $
 
 
 


Index: llvm/docs/GettingStarted.html
diff -u llvm/docs/GettingStarted.html:1.46 llvm/docs/GettingStarted.html:1.47
--- llvm/docs/GettingStarted.html:1.46	Fri Nov 21 19:30:01 2003
+++ llvm/docs/GettingStarted.html	Thu Dec 18 10:43:11 2003
@@ -129,7 +129,8 @@
         header files for the default platform. Useful options include:
       
         --with-llvmgccdir=directory
-            Specify where the LLVM GCC frontend is installed.
+            Specify the full pathname of where the LLVM GCC frontend is
+            installed.
         --enable-spec2000=directory
             Enable the SPEC2000 benchmarks for testing.  The SPEC2000
             benchmarks should be available in
@@ -181,24 +182,55 @@
 
   
Linux on x86 (Pentium and above)
   
-    Approximately 760 MB of Free Disk Space
+    
Approximately 918 MB of Free Disk Space
     
-      Source code: 30 MB
-      Object code: 670 MB
-      GCC front end: 60 MB
+      Source code: 28 MB
+      Object code: 850 MB
+      GCC front end: 40 MB
     
-  
+  
+  
+
+  
 
   Solaris on SparcV9 (Ultrasparc)
   
-    Approximately 1.24 GB of Free Disk Space
+    
Approximately 1.52 GB of Free Disk Space
+      
+        Source code: 28 MB
+        Object code: 1470 MB
+        GCC front end: 50 MB
+      
+  
+  
+
+  
+
+  FreeBSD on x86 (Pentium and above)
+  
+    Approximately 918 MB of Free Disk Space
+    
+      Source code: 28 MB
+      Object code: 850 MB
+      GCC front end: 40 MB
+    
+  
+  
+
+  
+
+  MacOS X on PowerPC
+  
+    No native code generation
+    
Approximately 1.20 GB of Free Disk Space
       
-        Source code: 30 MB
-        Object code: 1000 MB
-        GCC front end: 210 MB
+        Source code: 28 MB
+        Object code: 1160 MB
+        GCC front end: 40 MB
       
-  
+  
 
+  
 
 
 The LLVM suite may compile on other platforms, but it is not
@@ -252,7 +284,6 @@
 
 
 
-
 
The remainder of this guide is meant to get you up and running with
 LLVM and to give you some basic information about the LLVM environment.
 A complete guide to installation is provided in the
@@ -347,22 +378,31 @@
 
 

 If you have the LLVM distribution, you will need to unpack it before you
-can begin to compile it.  LLVM is distributed as a set of three files.  Each
+can begin to compile it.  LLVM is distributed as a set of two files: the LLVM
+suite and the LLVM GCC front end compiled for your platform.  Each
 file is a TAR archive that is compressed with the gzip program.
 
 
- The three files are as follows:
+
 The files are as follows:
 

-    llvm.tar.gz
+    
llvm-1.1.tar.gz
     
This is the source code to the LLVM suite.
     
 
-    
cfrontend.sparc.tar.gz
+    
cfrontend-1.1.sparc-sun-solaris2.8.tar.gz
     
This is the binary release of the GCC front end for Solaris/Sparc.
     
 
-    
cfrontend.x86.tar.gz
+    
cfrontend-1.1.i686-redhat-linux-gnu.tar.gz
     
This is the binary release of the GCC front end for Linux/x86.
+    
+
+    
cfrontend-1.1.i386-unknown-freebsd5.1.tar.gz
+    
This is the binary release of the GCC front end for FreeBSD/x86.
+    
+
+    
cfrontend-1.1.powerpc-apple-darwin7.0.0.tar.gz
+    
This is the binary release of the GCC front end for MacOS X/PPC.
 
 
 
@@ -390,6 +430,20 @@
 directory and fully populate it with the LLVM source code, Makefiles,
 test directories, and local copies of documentation files.
 
+
+If you want to get a specific release (as opposed to the most recent revision),
+you can specify a label.  The following releases have the following label:
+

+  
+  Release 1.1: RELEASE_11
+  
+
+  
+  Release 1.0: RELEASE_1
+  
+
+
+
 Note that the GCC front end is not included in the CVS repository. You
 should have downloaded the binary distribution for your platform.
 
@@ -411,12 +465,12 @@
 
 
   cd where-you-want-the-front-end-to-live
-  gunzip --stdout cfrontend.platform.tar.gz | tar -xvf
+  
gunzip --stdout cfrontend-version.platform.tar.gz | tar -xvf
       -

 

 
-If you are on a Sparc/Solaris machine, you will need to fix the header
-files:
+If you are using Solaris/Sparc or MacOS X/PPC, you will need to fix the
+header files:
 
 cd cfrontend/sparc

    ./fixheaders
@@ -442,7 +496,8 @@
 Once checked out from the CVS repository, the LLVM suite source code must be
 configured via the configure script.  This script sets variables in
 llvm/Makefile.config and llvm/include/Config/config.h.  It
-also populates OBJ_ROOT with the Makefiles needed to build LLVM.
+also populates OBJ_ROOT with the Makefiles needed to begin building
+LLVM.
 
 The following environment variables are used by the configure
 script to configure the build system:
@@ -476,7 +531,8 @@
   --with-llvmgccdir=LLVMGCCDIR
   

     Path to the location where the LLVM C front end binaries and
-    associated libraries will be installed.
+    associated libraries were installed.  This must be specified as an
+    absolute pathname.
     
   
--enable-optimized
   

@@ -486,7 +542,8 @@
     
   
--enable-jit
   
-    Compile the Just In Time (JIT) functionality.  This is not available
+    Compile the Just In Time (JIT) compiler functionality.  This is not
+    available
     on all platforms.  The default is dependent on platform, so it is best
     to explicitly enable it if you want it.
     
@@ -519,10 +576,10 @@
 LLVM_LIB_SEARCH_PATH environment variable in your startup scripts.
 This environment variable is used to locate "system" libraries like
 "-lc" and "-lm" when linking.  This variable should be set to
-the absolute path for the bytecode-libs subdirectory of the GCC front end
-install, or LLVMGCCDIR/bytecode-libs.  For example, one might set
+the absolute path of the bytecode-libs subdirectory of the GCC front
+end, or LLVMGCCDIR/bytecode-libs.  For example, one might set
 LLVM_LIB_SEARCH_PATH to
-/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the X86
+/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs for the x86
 version of the GCC front end on our research machines.
 
 
@@ -1089,7 +1146,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/11/22 01:30:01 $
+  Last modified: $Date: 2003/12/18 16:43:11 $
 
 
 


Index: llvm/docs/ReleaseNotes.html
diff -u llvm/docs/ReleaseNotes.html:1.87 llvm/docs/ReleaseNotes.html:1.88
--- llvm/docs/ReleaseNotes.html:1.87	Thu Dec 18 02:16:25 2003
+++ llvm/docs/ReleaseNotes.html	Thu Dec 18 10:43:14 2003
@@ -18,7 +18,6 @@
   Installation Instructions
   Known Problems
   
-
     Known problems with the LLVM Core
     
Known problems with the C Front-end
     
Known problems with the C++ Front-end
@@ -30,7 +29,7 @@
 
 
 
-  Written by Chris Lattner

+  
Written by the LLVM team

 
 
 
@@ -43,10 +42,10 @@
 
 This document contains the release notes for the LLVM compiler
 infrastructure, release 1.2.  Here we describe the status of LLVM, including any
-known problems, and bug fixes from the previous release.  The most up-to-date
+known problems and bug fixes from the previous release.  The most up-to-date
 version of this document can be found on the LLVM 1.2 web site.  If you are
-not reading this on the LLVM web pages, you should probably go there, because
+not reading this on the LLVM web pages, you should probably go there because
 this document may be updated after the release.
 
 For more information about LLVM, including information about potentially more
@@ -70,8 +69,7 @@
 
 

 
-This is the third public release of the LLVM compiler infrastructure.  OTHER
-OVERVIEW STUFF HERE.
+
This is the third public release of the LLVM compiler infrastructure.
 
 
 At this time, LLVM is known to correctly compile and run all non-unwinding C
@@ -82,10 +80,10 @@
 
 
 
-The LLVM native code generators are very stable, but do not currently support
+The LLVM native code generators are very stable but do not currently support
 unwinding (exception throwing or longjmping), which prevent them from
 working with programs like the 253.perlbmk in SPEC CPU2000.  The C
-backend and the rest of LLVM does support these programs however, so you can
+backend and the rest of LLVM supports these programs, so you can
 still use LLVM with them.  Support for unwinding will be added in a future
 release.
 
@@ -164,11 +162,11 @@
 LLVM has been extensively tested on Intel and AMD machines running Red
 Hat Linux and FreeBSD.  It has also been tested on Sun UltraSPARC workstations running Solaris 8.
 Additionally, 
-LLVM works on Mac OS/X 10.3 and above, but only with the C backend or 
+LLVM works on Mac OS X 10.3 and above, but only with the C backend or 
 interpreter (no native backend for the PowerPC is available yet).
 The core LLVM infrastructure uses "autoconf" for portability, so hopefully we
 work on more platforms than that.  However, it is likely that we
-missed something, and that minor porting is required to get LLVM to work on 
+missed something and that minor porting is required to get LLVM to work on 
 new platforms.  We welcome portability patches and error messages.
 
 
@@ -184,7 +182,7 @@
 This section contains all known problems with the LLVM system, listed by
 component.  As new problems are discovered, they will be added to these
 sections.  If you run into a problem, please check the LLVM bug database, and submit a bug if
+href="http://llvm.cs.uiuc.edu/bugs/">LLVM bug database and submit a bug if
 there isn't already one.
 
 
@@ -219,6 +217,18 @@
 
LLVM cannot handle structures with
 more than 256 elements.
 
+
+The gccld program 
+
+does not link objects/archives in the order specified on the command line.
+
+
+
+
+
+Tail duplication does not update SSA form correctly.
+
+
 
 
 
@@ -232,9 +242,7 @@
 Bugs
 
 
-
 
-
 C99 Variable sized arrays do not release stack memory when they go out of 
     scope.  Thus, the following program may run out of stack space:
 @@ -244,8 +252,27 @@
     }
 
 
+
+Initialization of global union variables can only be done
+with the largest
+union member.
+
+
+
+
+Functions marked "extern inline" are not compiled into LLVM with linkonce
+linkage.
+
+
+
+
+
+The memory management functions in the libc runtime
+need weak linkage so that they can be
+overridden.
+
+
 
-
 
 
 
@@ -277,11 +304,11 @@
     the following extensions are known to not be supported:
   
   Local Labels: Labels local to a block.
-  Labels as Values: Getting pointers to labels, and computed gotos.
+  Labels as Values: Getting pointers to labels and computed gotos.
   Nested Functions: As in Algol and Pascal, lexical scoping of functions.
   Constructing Calls: Dispatching a call to another function.
   Extended Asm: Assembler instructions with C expressions as operands.
-  Constraints: Constraints for asm operands
+  Constraints: Constraints for asm operands.
   Asm Labels: Specifying the assembler name to use for a C symbol.
   Explicit Reg Vars: Defining variables residing in specified registers.
   Return Address: Getting the return or frame address of a function.
@@ -294,7 +321,7 @@
   The following GCC extensions are partially supported.  An ignored
   attribute means that the LLVM compiler ignores the presence of the attribute,
   but the code should still work.  An unsupported attribute is one which is
-  ignored by the LLVM compiler, which will cause a different interpretation of
+  ignored by the LLVM compiler and will cause a different interpretation of
   the program.
 
   
@@ -304,7 +331,7 @@
 
   Function Attributes:
 
-      Declaring that functions have no side effects, or that they can never
+      Declaring that functions have no side effects or that they can never
       return.

 
       Supported: format, format_arg, non_null,
@@ -363,7 +390,8 @@
   
Subscripting: Any array can be subscripted, even if not an lvalue.
   Pointer Arith: Arithmetic on void-pointers and function pointers.
   Initializers: Non-constant initializers.
-  Compound Literals: Compound literals give structures, unions or arrays as values.
+  Compound Literals: Compound literals give structures, unions,
+or arrays as values.
   Designated Inits: Labeling elements of initializers.
   Cast to Union: Casting to union type from any member of the union.
   Case Ranges: `case 1 ... 9' and such.
@@ -395,7 +423,7 @@
 
 
 
-For this release, the C++ front-end is considered to be fully functional, but
+
For this release, the C++ front-end is considered to be fully functional but
 has not been tested as thoroughly as the C front-end.  It has been tested and 
 works for a number of non-trivial programs, but there may be lurking bugs.
 Please report any bugs or problems.
@@ -411,9 +439,14 @@
 
 
 The C++ front-end inherits all problems afflicting the C
-    front-end
-
+    front-end.
 
+
+
+Code is generated for empty classes.
+
+
+
 
 
 
@@ -433,7 +466,7 @@
 Destructors for local objects are not always run when a longjmp is
     performed. In particular, destructors for objects in the longjmping
     function and in the setjmp receiver function may not be run.
-    Objects in intervening stack frames will be destroyed however (which is
+    Objects in intervening stack frames will be destroyed, however (which is
     better than most compilers).
 
 The LLVM C++ front-end follows the unwind instruction, so code that throws a C++ exception
 or calls the C longjmp function will abort.
 
+
+
+The llc program can crash on legal code.
+
+
 
 
 
@@ -522,7 +560,7 @@
 
 
 A wide variety of additional information is available on the LLVM web page,
-including mailing lists publications describing algorithms and components
+including mailing lists and publications describing algorithms and components
 implemented in LLVM.  The web page also contains versions of the API
 documentation which is up-to-date with the CVS version of the source code.  You
 can access versions of these documents specific to this release by going into
@@ -544,7 +582,7 @@
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" />
 
   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/18 08:16:25 $
+  Last modified: $Date: 2003/12/18 16:43:14 $
 
 
 


Index: llvm/docs/Stacker.html
diff -u llvm/docs/Stacker.html:1.6 llvm/docs/Stacker.html:1.7
--- llvm/docs/Stacker.html:1.6	Thu Dec 18 00:40:22 2003
+++ llvm/docs/Stacker.html	Thu Dec 18 10:43:14 2003
@@ -61,7 +61,7 @@
 about LLVM through the experience of creating a simple programming language
 named Stacker.  Stacker was invented specifically as a demonstration of
 LLVM. The emphasis in this document is not on describing the
-intricacies of LLVM itself, but on how to use it to build your own
+intricacies of LLVM itself but on how to use it to build your own
 compiler system.
 
 
@@ -77,11 +77,11 @@
 language the author ever created using LLVM. The learning curve is 
 included in that four days.
 The language described here, Stacker, is Forth-like. Programs
-are simple collections of word definitions and the only thing definitions
+are simple collections of word definitions, and the only thing definitions
 can do is manipulate a stack or generate I/O.  Stacker is not a "real" 
-programming language; its very simple.  Although it is computationally 
+programming language; it's very simple.  Although it is computationally 
 complete, you wouldn't use it for your next big project. However, 
-the fact that it is complete, its simple, and it doesn't have 
+the fact that it is complete, it's simple, and it doesn't have 
 a C-like syntax make it useful for demonstration purposes. It shows
 that LLVM could be applied to a wide variety of languages.
 The basic notions behind stacker is very simple. There's a stack of 
@@ -95,11 +95,11 @@
 : MAIN hello_world ;

 This has two "definitions" (Stacker manipulates words, not
 functions and words have definitions): MAIN and 
-hello_world. The MAIN definition is standard, it
+hello_world. The MAIN definition is standard; it
 tells Stacker where to start. Here, MAIN is defined to 
 simply invoke the word hello_world. The
 hello_world definition tells stacker to push the 
-"Hello, World!" string onto the stack, print it out 
+"Hello, World!" string on to the stack, print it out 
 (>s), pop it off the stack (DROP), and
 finally print a carriage return (CR). Although 
 hello_world uses the stack, its net effect is null. Well
@@ -123,7 +123,7 @@
 
Although I knew that LLVM uses a Single Static Assignment (SSA) format, 
 it wasn't obvious to me how prevalent this idea was in LLVM until I really
 started using it.  Reading the 
-Programmer's Manual and Language Reference
+Programmer's Manual and Language Reference,
 I noted that most of the important LLVM IR (Intermediate Representation) C++ 
 classes were derived from the Value class. The full power of that simple
 design only became fully understood once I started constructing executable
@@ -201,7 +201,7 @@
 

     Create your blocks early. While writing your compiler, you 
     will encounter several situations where you know apriori that you will
-    need several blocks. For example, if-then-else, switch, while and for
+    need several blocks. For example, if-then-else, switch, while, and for
     statements in C/C++ all need multiple blocks for expression in LVVM. 
     The rule is, create them early.
     Terminate your blocks early. This just reduces the chances 
@@ -262,7 +262,7 @@
 the instructions for the "then" and "else" parts. They would use the third part
 of the idiom almost exclusively (inserting new instructions before the 
 terminator). Furthermore, they could even recurse back to handle_if 
-should they encounter another if/then/else statement and it will just work.
+should they encounter another if/then/else statement, and it will just work.
 Note how cleanly this all works out. In particular, the push_back methods on
 the BasicBlock's instruction list. These are lists of type 
 Instruction (which is also of type Value). To create 
@@ -271,7 +271,8 @@
 BasicBlock objects act like branch labels! This new 
 BranchInst terminates the BasicBlock provided 
 as an argument. To give the caller a way to keep inserting after calling 
-handle_if we create an exit_bb block which is returned 
+handle_if, we create an exit_bb block which is
+returned 
 to the caller.  Note that the exit_bb block is used as the 
 terminator for both the then_bb and the else_bb
 blocks. This guarantees that no matter what else handle_if
@@ -286,7 +287,7 @@
 method on the various lists. This is so common that it is worth mentioning.
 The "push_back" inserts a value into an STL list, vector, array, etc. at the
 end. The method might have also been named "insert_tail" or "append".
-Althought I've used STL quite frequently, my use of push_back wasn't very
+Although I've used STL quite frequently, my use of push_back wasn't very
 high in other programs. In LLVM, you'll use it all the time.
 
 
@@ -295,17 +296,17 @@
 
 
 It took a little getting used to and several rounds of postings to the LLVM
-mail list to wrap my head around this instruction correctly. Even though I had
+mailing list to wrap my head around this instruction correctly. Even though I had
 read the Language Reference and Programmer's Manual a couple times each, I still
 missed a few very key points:
 
 
-    GetElementPtrInst gives you back a Value for the last thing indexed
+    
GetElementPtrInst gives you back a Value for the last thing indexed.
     
All global variables in LLVM  are pointers.
     
Pointers must also be dereferenced with the GetElementPtrInst instruction.
 
 This means that when you look up an element in the global variable (assuming
-its a struct or array), you must deference the pointer first! For many
+it's a struct or array), you must deference the pointer first! For many
 things, this leads to the idiom:
 
 
@@ -322,13 +323,13 @@
 variable and the address of its first element as the same. That tripped me up
 for a while until I realized that they really do differ .. by type.
 Remember that LLVM is strongly typed. Everything has a type.  
-The "type" of the global variable is [24 x int]*. That is, its
+The "type" of the global variable is [24 x int]*. That is, it's
 a pointer to an array of 24 ints.  When you dereference that global variable with
 a single (0) index, you now have a "[24 x int]" type.  Although
 the pointer value of the dereferenced global and the address of the zero'th element
 in the array will be the same, they differ in their type. The zero'th element has
 type "int" while the pointer value has type "[24 x int]".
-Get this one aspect of LLVM right in your head and you'll save yourself
+
Get this one aspect of LLVM right in your head, and you'll save yourself
 a lot of compiler writing headaches down the road.
 

 
@@ -337,7 +338,7 @@
 Linkage types in LLVM can be a little confusing, especially if your compiler
 writing mind has affixed firm concepts to particular words like "weak",
 "external", "global", "linkonce", etc. LLVM does not use the precise
-definitions of say ELF or GCC even though they share common terms. To be fair,
+definitions of, say, ELF or GCC, even though they share common terms. To be fair,
 the concepts are related and similar but not precisely the same. This can lead
 you to think you know what a linkage type represents but in fact it is slightly
 different. I recommend you read the 
@@ -346,10 +347,10 @@
 
Here are some handy tips that I discovered along the way:
 
     Unitialized means external. That is, the symbol is declared in the current
-    module and can be used by that module but it is not defined by that module.
+    module and can be used by that module, but it is not defined by that module.
     Setting an initializer changes a global' linkage type. Setting an 
     initializer changes a global's linkage type from whatever it was to a normal, 
-    defind global (not external). You'll need to call the setLinkage() method to 
+    defined global (not external). You'll need to call the setLinkage() method to 
     reset it if you specify the initializer after the GlobalValue has been constructed. 
     This is important for LinkOnce and Weak linkage types. 
     Appending linkage can keep track of things. Appending linkage can 
@@ -368,7 +369,7 @@
 functions in the LLVM IR that make things easier. Here's what I learned: 
 
  Constants are Values like anything else and can be operands of instructions
- Integer constants, frequently needed can be created using the static "get"
+ 
Integer constants, frequently needed, can be created using the static "get"
  methods of the ConstantInt, ConstantSInt, and ConstantUInt classes. The nice thing
  about these is that you can "get" any kind of integer quickly.
  There's a special method on Constant class which allows you to get the null 
@@ -385,14 +386,14 @@
 proceeding, a few words about the stack are in order. The stack is simply
 a global array of 32-bit integers or pointers. A global index keeps track
 of the location of the top of the stack. All of this is hidden from the 
-programmer but it needs to be noted because it is the foundation of the 
+programmer, but it needs to be noted because it is the foundation of the 
 conceptual programming model for Stacker. When you write a definition,
 you are, essentially, saying how you want that definition to manipulate
 the global stack.
 Manipulating the stack can be quite hazardous. There is no distinction
 given and no checking for the various types of values that can be placed
 on the stack. Automatic coercion between types is performed. In many 
-cases this is useful. For example, a boolean value placed on the stack
+cases, this is useful. For example, a boolean value placed on the stack
 can be interpreted as an integer with good results. However, using a
 word that interprets that boolean value as a pointer to a string to
 print out will almost always yield a crash. Stacker simply leaves it
@@ -412,9 +413,9 @@
 
So, your typical definition will have the form:
 : name ... ;
 The name is up to you but it must start with a letter and contain
-only letters numbers and underscore. Names are case sensitive and must not be
+only letters, numbers, and underscore. Names are case sensitive and must not be
 the same as the name of a built-in word. The ... is replaced by
-the stack manipulting words that you wish define name as. 

+the stack manipulating words that you wish to define name as. 

 
 
 
Comments
@@ -435,12 +436,12 @@
 
 Literals
 
-    There are three kinds of literal values in Stacker. Integer, Strings,
+    
There are three kinds of literal values in Stacker: Integers, Strings,
     and Booleans. In each case, the stack operation is to simply push the
-    value onto the stack. So, for example:

+    value on to the stack. So, for example:

      42 " is the answer." TRUE 

-    will push three values onto the stack: the integer 42, the
-    string " is the answer." and the boolean TRUE.
+    will push three values on to the stack: the integer 42, the
+    string " is the answer.", and the boolean TRUE.
 
 
 Words
@@ -470,21 +471,21 @@
 The built-in words of the Stacker language are put in several groups 
 depending on what they do. The groups are as follows:
  
-    LogicalThese words provide the logical operations for
+    
Logical: These words provide the logical operations for
     comparing stack operands.
The words are: < > <= >= 
     = <> true false.
-    BitwiseThese words perform bitwise computations on 
+    
Bitwise: These words perform bitwise computations on 
     their operands. 
 The words are: << >> XOR AND NOT
-    ArithmeticThese words perform arithmetic computations on
+    
Arithmetic: These words perform arithmetic computations on
     their operands. 
 The words are: ABS NEG + - * / MOD */ ++ -- MIN MAX
     StackThese words manipulate the stack directly by moving
     its elements around.
 The words are: DROP DROP2 NIP NIP2 DUP DUP2 
     SWAP SWAP2 OVER OVER2 ROT ROT2 RROT RROT2 TUCK TUCK2 PICK SELECT ROLL
-    MemoryThese words allocate, free and manipulate memory
+    
MemoryThese words allocate, free, and manipulate memory
     areas outside the stack.
The words are: MALLOC FREE GET PUT
-    ControlThese words alter the normal left to right flow
+    
Control: These words alter the normal left to right flow
     of execution.
The words are: IF ELSE ENDIF WHILE END RETURN EXIT RECURSE
-    I/O These words perform output on the standard output
+    
I/O: These words perform output on the standard output
     and input on the standard input. No other I/O is possible in Stacker.
     
The words are: SPACE TAB CR >s >d >c <s <d <c.
 
@@ -566,12 +567,12 @@
 
-    
+    
-    
+    
@@ -626,75 +627,75 @@
     
+    on to the stack
 
+    on to the stack
 
+    on to the stack
 
+    on to the stack
 
+    of w1 by w2 is pushed back on to the stack
 
+    divided by w3. The result is pushed back on to the stack.
 
+    pushed back on to the stack.
 
+    pushed back on to the stack.
 
+    on to the stack.
 
+	on to the stack.
 
@@ -730,7 +731,7 @@
 
-    
@@ -744,7 +745,7 @@
     
@@ -752,27 +753,27 @@
     
-    
@@ -854,7 +855,7 @@
     
+	block is pushed on to the stack.
 
@@ -948,7 +949,7 @@
     
+	resulting string is pushed on to the stack.
 
+	resulting value is pushed on to the stack
 
+	" %c". The value is converted to an integer and pushed on to the stack.
 
@@ -1030,9 +1031,9 @@
 
 The following fully documented program highlights many features of both
 the Stacker language and what is possible with LLVM. The program has two modes
-of operations. If you provide numeric arguments to the program, it checks to see
+of operation. If you provide numeric arguments to the program, it checks to see
 if those arguments are prime numbers and prints out the results. Without any 
-aruments, the program prints out any prime numbers it finds between 1 and one 
+arguments, the program prints out any prime numbers it finds between 1 and one 
 million (there's a lot of them!). The source code comments below tell the 
 remainder of the story.
 
@@ -1057,7 +1058,7 @@
 : exit_loop FALSE;
     
 ################################################################################
-# This definition tryies an actual division of a candidate prime number. It
+# This definition tries an actual division of a candidate prime number. It
 # determines whether the division loop on this candidate should continue or
 # not.
 # STACK<:
@@ -1117,7 +1118,7 @@
 # STACK<:
 #   p - the prime number to check
 # STACK>:
-#   yn - boolean indiating if its a prime or not
+#   yn - boolean indicating if its a prime or not
 #   p - the prime number checked
 ################################################################################
 : try_harder
@@ -1290,7 +1291,7 @@
 under the LLVM "projects" directory. You will need to obtain the LLVM sources
 to find it (either via anonymous CVS or a tarball. See the 
 Getting Started document).
-Under the "projects" directory there is a directory named "stacker". That
+
Under the "projects" directory there is a directory named "Stacker". That
 directory contains everything, as follows:
 
     lib - contains most of the source code
@@ -1343,7 +1344,7 @@
 definitions, the ROLL word is not implemented. This word was left out of 
 Stacker on purpose so that it can be an exercise for the student.  The exercise 
 is to implement the ROLL functionality (in your own workspace) and build a test 
-program for it.  If you can implement ROLL you understand Stacker and probably 
+program for it.  If you can implement ROLL, you understand Stacker and probably 
 a fair amount about LLVM since this is one of the more complicated Stacker 
 operations. The work will almost be completely limited to the 
 compiler.  
@@ -1374,7 +1375,7 @@
     emitted currently is somewhat wasteful. It gets cleaned up a lot by existing
     passes but more could be done.
     Add -O -O1 -O2 and -O3 optimization switches to the compiler driver to
-    allow LLVM optimization without using "opt"
+    allow LLVM optimization without using "opt."
     Make the compiler driver use the LLVM linking facilities (with IPO) before 
     depending on GCC to do the final link.
     Clean up parsing. It doesn't handle errors very well.
@@ -1394,6 +1395,6 @@
 
 Reid Spencer
 The LLVM Compiler Infrastructure 
-
Last modified: $Date: 2003/12/18 06:40:22 $ 
+
Last modified: $Date: 2003/12/18 16:43:14 $ 
 
 




From criswell at cs.uiuc.edu  Thu Dec 18 11:20:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:20:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
Message-ID: <200312181719.LAA09598@choi.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

IndVarSimplify.cpp updated: 1.46 -> 1.47

---
Log message:

Reverted back to previous revision - this was previously merged
according to the CVS log messages.



---
Diffs of the changes:  (+15 -8)

Index: llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
diff -u llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.46 llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.47
--- llvm/lib/Transforms/Scalar/IndVarSimplify.cpp:1.46	Thu Dec 18 10:42:57 2003
+++ llvm/lib/Transforms/Scalar/IndVarSimplify.cpp	Thu Dec 18 11:19:19 2003
@@ -15,19 +15,19 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Transforms/Scalar.h"
-#include "llvm/Analysis/InductionVariable.h"
-#include "llvm/Analysis/LoopInfo.h"
+#include "llvm/Constants.h"
+#include "llvm/Type.h"
 #include "llvm/iPHINode.h"
 #include "llvm/iOther.h"
-#include "llvm/Type.h"
-#include "llvm/Constants.h"
+#include "llvm/Analysis/InductionVariable.h"
+#include "llvm/Analysis/LoopInfo.h"
 #include "llvm/Support/CFG.h"
+#include "llvm/Transforms/Utils/Local.h"
 #include "Support/Debug.h"
 #include "Support/Statistic.h"
 #include "Support/STLExtras.h"
 #include 
-
-namespace llvm {
+using namespace llvm;
 
 namespace {
   Statistic<> NumRemoved ("indvars", "Number of aux indvars removed");
@@ -141,6 +141,8 @@
 
     DEBUG(IV->print(std::cerr));
 
+    while (isa(AfterPHIIt)) ++AfterPHIIt;
+
     // Don't do math with pointers...
     const Type *IVTy = IV->Phi->getType();
     if (isa(IVTy)) IVTy = Type::ULongTy;
@@ -186,6 +188,12 @@
       IV->Phi->setName("");
       Val->setName(OldName);
 
+      // Get the incoming values used by the PHI node
+      std::vector PHIOps;
+      PHIOps.reserve(IV->Phi->getNumIncomingValues());
+      for (unsigned i = 0, e = IV->Phi->getNumIncomingValues(); i != e; ++i)
+        PHIOps.push_back(IV->Phi->getIncomingValue(i));
+
       // Delete the old, now unused, phi node...
       Header->getInstList().erase(IV->Phi);
 
@@ -242,8 +250,7 @@
                                            "Canonicalize Induction Variables");
 }
 
-Pass *createIndVarSimplifyPass() {
+Pass *llvm::createIndVarSimplifyPass() {
   return new InductionVariableSimplify();
 }
 
-} // End llvm namespace




From criswell at cs.uiuc.edu  Thu Dec 18 11:25:00 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:25:00 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181724.LAA21396@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.4 -> 1.5

---
Log message:

Re-enabling logging.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.4 llvm-www/releases/testregister.cgi:1.5
--- llvm-www/releases/testregister.cgi:1.4	Thu Dec 18 11:24:21 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:24:49 2003
@@ -181,7 +181,7 @@
 #
 # Log the information provided by the form.
 #
-#LogForm (form)
+LogForm (form)
 
 #
 # Subscribe the user the LLVM Announcements list (if they so desire)




From criswell at cs.uiuc.edu  Thu Dec 18 11:25:08 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:25:08 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181724.LAA21040@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.3 -> 1.4

---
Log message:

Testing logging functionality.



---
Diffs of the changes:  (+13 -6)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.3 llvm-www/releases/testregister.cgi:1.4
--- llvm-www/releases/testregister.cgi:1.3	Tue Dec 16 15:11:53 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:24:21 2003
@@ -10,6 +10,9 @@
 # List of email addresses that want to know when people download
 notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
 
+# Full pathname of the logfile
+logfilename='/var/tmp/LLVMDownloads'
+
 #
 # Function: Subscribe()
 #
@@ -144,6 +147,16 @@
 		except:
 			pass
 	mailer.quit ()
+
+	#
+	# Open the logfile, lock it, and write a new entry.
+	#
+	logfile = file (logfilename, 'ra')
+	os.lockf (logfile, os.LOCK_EX)
+	logfile.write (email + ':' + lastname + ':' + firstname + '\n')
+	logfile.flush ()
+	logfile.close ()
+
 	return
 
 #
@@ -175,12 +188,6 @@
 #
 #Subscribe (form)
 
-#
-# Create a cookie so that the user does not need to register again.
-#
-usercookie = Cookie.SmartCookie()
-usercookie["LLVM"] = "Yes"
-print usercookie
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Thu Dec 18 11:26:00 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:26:00 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181725.LAA22683@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.5 -> 1.6

---
Log message:

Don't bother Chris and Vikram while testing.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.5 llvm-www/releases/testregister.cgi:1.6
--- llvm-www/releases/testregister.cgi:1.5	Thu Dec 18 11:24:49 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:25:18 2003
@@ -8,7 +8,7 @@
 import Cookie
 
 # List of email addresses that want to know when people download
-notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
+notifylist=['criswell at uiuc.edu]
 
 # Full pathname of the logfile
 logfilename='/var/tmp/LLVMDownloads'




From criswell at cs.uiuc.edu  Thu Dec 18 11:27:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:27:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181726.LAA22877@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.6 -> 1.7

---
Log message:

Fixed syntax error.


---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.6 llvm-www/releases/testregister.cgi:1.7
--- llvm-www/releases/testregister.cgi:1.6	Thu Dec 18 11:25:18 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:26:37 2003
@@ -8,7 +8,7 @@
 import Cookie
 
 # List of email addresses that want to know when people download
-notifylist=['criswell at uiuc.edu]
+notifylist=['criswell at uiuc.edu']
 
 # Full pathname of the logfile
 logfilename='/var/tmp/LLVMDownloads'




From criswell at cs.uiuc.edu  Thu Dec 18 11:32:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:32:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181731.LAA23015@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.7 -> 1.8

---
Log message:

Fixed errors.



---
Diffs of the changes:  (+3 -3)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.7 llvm-www/releases/testregister.cgi:1.8
--- llvm-www/releases/testregister.cgi:1.7	Thu Dec 18 11:26:37 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:31:42 2003
@@ -4,8 +4,8 @@
 import urllib
 import smtplib
 import os
+import fcntl
 import sys
-import Cookie
 
 # List of email addresses that want to know when people download
 notifylist=['criswell at uiuc.edu']
@@ -151,8 +151,8 @@
 	#
 	# Open the logfile, lock it, and write a new entry.
 	#
-	logfile = file (logfilename, 'ra')
-	os.lockf (logfile, os.LOCK_EX)
+	logfile = file (logfilename, 'a')
+	fcntl.lockf (logfile, fcntl.LOCK_EX)
 	logfile.write (email + ':' + lastname + ':' + firstname + '\n')
 	logfile.flush ()
 	logfile.close ()




From criswell at cs.uiuc.edu  Thu Dec 18 11:36:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:36:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312181735.LAA23148@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.7 -> 1.8

---
Log message:

Added logging support.



---
Diffs of the changes:  (+13 -0)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.7 llvm-www/releases/register.cgi:1.8
--- llvm-www/releases/register.cgi:1.7	Tue Dec 16 15:06:04 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 11:35:09 2003
@@ -10,6 +10,9 @@
 # List of email addresses that want to know when people download
 notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
 
+# Full pathname of the logfile
+logfilename='/var/tmp/LLVMDownloads'
+
 #
 # Function: Subscribe()
 #
@@ -144,6 +147,16 @@
 		except:
 			pass
 	mailer.quit ()
+
+	#
+	# Open the logfile, lock it, and write a new entry.
+	#
+	logfile = file (logfilename, 'a')
+	fcntl.lockf (logfile, fcntl.LOCK_EX)
+	logfile.write (email + ':' + lastname + ':' + firstname + '\n')
+	logfile.flush ()
+	logfile.close ()
+
 	return
 
 #




From criswell at cs.uiuc.edu  Thu Dec 18 11:39:03 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:39:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312181738.LAA23270@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.8 -> 1.9

---
Log message:

Fix script for person trying to download as I speak.



---
Diffs of the changes:  (+5 -5)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.8 llvm-www/releases/register.cgi:1.9
--- llvm-www/releases/register.cgi:1.8	Thu Dec 18 11:35:09 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 11:38:05 2003
@@ -151,11 +151,11 @@
 	#
 	# Open the logfile, lock it, and write a new entry.
 	#
-	logfile = file (logfilename, 'a')
-	fcntl.lockf (logfile, fcntl.LOCK_EX)
-	logfile.write (email + ':' + lastname + ':' + firstname + '\n')
-	logfile.flush ()
-	logfile.close ()
+	#logfile = file (logfilename, 'a')
+	#fcntl.lockf (logfile, fcntl.LOCK_EX)
+	#logfile.write (email + ':' + lastname + ':' + firstname + '\n')
+	#logfile.flush ()
+	#logfile.close ()
 
 	return
 




From criswell at cs.uiuc.edu  Thu Dec 18 11:50:03 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 11:50:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181749.LAA23406@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.8 -> 1.9

---
Log message:

Add code to unlock the file when done.
Added code that helps to ensure that we return successfully even if
writing to the logfile fails.



---
Diffs of the changes:  (+11 -8)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.8 llvm-www/releases/testregister.cgi:1.9
--- llvm-www/releases/testregister.cgi:1.8	Thu Dec 18 11:31:42 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 11:48:56 2003
@@ -151,11 +151,15 @@
 	#
 	# Open the logfile, lock it, and write a new entry.
 	#
-	logfile = file (logfilename, 'a')
-	fcntl.lockf (logfile, fcntl.LOCK_EX)
-	logfile.write (email + ':' + lastname + ':' + firstname + '\n')
-	logfile.flush ()
-	logfile.close ()
+	try:
+		logfile = file (logfilename, 'a')
+		fcntl.lockf (logfile, fcntl.LOCK_EX)
+		logfile.write (email + ':' + lastname + ':' + firstname + '\n')
+		logfile.flush ()
+		fcntl.lockf (logfile, fcntl.LOCK_UN)
+		logfile.close ()
+	except:
+		pass
 
 	return
 
@@ -167,8 +171,7 @@
 #
 # Verify that everything in the form is correct.
 #
-#error = ValidateForm (form)
-error = ''
+error = ValidateForm (form)
 if (error != ''):
 	print ('Content-type: text/html')
 	print ('Status: 400 Bad Request')
@@ -186,7 +189,7 @@
 #
 # Subscribe the user the LLVM Announcements list (if they so desire)
 #
-#Subscribe (form)
+Subscribe (form)
 
 
 #




From criswell at cs.uiuc.edu  Thu Dec 18 12:02:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 12:02:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181801.MAA23542@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.9 -> 1.10

---
Log message:

Added extra fields to logfile.



---
Diffs of the changes:  (+2 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.9 llvm-www/releases/testregister.cgi:1.10
--- llvm-www/releases/testregister.cgi:1.9	Thu Dec 18 11:48:56 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 12:00:52 2003
@@ -8,6 +8,7 @@
 import sys
 
 # List of email addresses that want to know when people download
+#notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
 notifylist=['criswell at uiuc.edu']
 
 # Full pathname of the logfile
@@ -154,7 +155,7 @@
 	try:
 		logfile = file (logfilename, 'a')
 		fcntl.lockf (logfile, fcntl.LOCK_EX)
-		logfile.write (email + ':' + lastname + ':' + firstname + '\n')
+		logfile.write (email + ':' + firstname + ':' + lastname + ':' + organization + ':' + title + '\n')
 		logfile.flush ()
 		fcntl.lockf (logfile, fcntl.LOCK_UN)
 		logfile.close ()




From criswell at cs.uiuc.edu  Thu Dec 18 13:09:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:09:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181908.NAA23751@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.10 -> 1.11

---
Log message:

Check fields for colons.



---
Diffs of the changes:  (+37 -4)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.10 llvm-www/releases/testregister.cgi:1.11
--- llvm-www/releases/testregister.cgi:1.10	Thu Dec 18 12:00:52 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:08:28 2003
@@ -15,6 +15,20 @@
 logfilename='/var/tmp/LLVMDownloads'
 
 #
+# Function: checkForColon()
+#
+# Description:
+#	This function checks a string to see if it contains a colon.
+#
+def checkForColon (item):
+	length = len (item) - 1
+	while (length != 0):
+		if (item[length] == ':'):
+			return true
+		length=length - 1
+	return false
+
+#
 # Function: Subscribe()
 #
 # Description:
@@ -68,10 +82,11 @@
 	#
 	# Verify if the required fields have been supplied.
 	#
-	firstname = form.getfirst ('firstname', '')
-	lastname  = form.getfirst ('lastname', '')
-	email     = form.getfirst ('email', '')
-	announce  = form.getfirst ('announce','yes')
+	firstname    = form.getfirst ('firstname', '')
+	lastname     = form.getfirst ('lastname', '')
+	email        = form.getfirst ('email', '')
+	title        = form.getfirst ('title','')
+	organization = form.getfirst ('organization', '')
 
 	#
 	# Verify that the name and email fields have been filled in.
@@ -103,6 +118,24 @@
 		length=length - 1
 	else:
 		return 'Email address has no periods'
+
+	#
+	# Verify that there are no colons in fields placed in the logfile.
+	#
+	if (checkForColon (email)):
+		return 'Email address has a colon'
+
+	if (checkForColon (firstname)):
+		return 'First name has a colon'
+
+	if (checkForColon (lastname)):
+		return 'Last name has a colon'
+
+	if (checkForColon (organization)):
+		return 'Organization has a colon'
+
+	if (checkForColon (title)):
+		return 'Title has a colon'
 
 	return ''
 




From criswell at cs.uiuc.edu  Thu Dec 18 13:12:08 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:12:08 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181912.NAA23872@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.11 -> 1.12

---
Log message:

Fixed the check for colons for zero length strings.



---
Diffs of the changes:  (+2 -0)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.11 llvm-www/releases/testregister.cgi:1.12
--- llvm-www/releases/testregister.cgi:1.11	Thu Dec 18 13:08:28 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:11:55 2003
@@ -21,6 +21,8 @@
 #	This function checks a string to see if it contains a colon.
 #
 def checkForColon (item):
+  if (len (item) == 0):
+    return false
 	length = len (item) - 1
 	while (length != 0):
 		if (item[length] == ':'):




From criswell at cs.uiuc.edu  Thu Dec 18 13:14:03 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:14:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181913.NAA23986@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.12 -> 1.13

---
Log message:

Modified truth/false values.



---
Diffs of the changes:  (+3 -3)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.12 llvm-www/releases/testregister.cgi:1.13
--- llvm-www/releases/testregister.cgi:1.12	Thu Dec 18 13:11:55 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:13:14 2003
@@ -22,13 +22,13 @@
 #
 def checkForColon (item):
   if (len (item) == 0):
-    return false
+    return False
 	length = len (item) - 1
 	while (length != 0):
 		if (item[length] == ':'):
-			return true
+			return True
 		length=length - 1
-	return false
+	return False
 
 #
 # Function: Subscribe()




From criswell at cs.uiuc.edu  Thu Dec 18 13:16:00 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:16:00 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181915.NAA24104@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.13 -> 1.14

---
Log message:

Removing code segments for debugging.



---
Diffs of the changes:  (+2 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.13 llvm-www/releases/testregister.cgi:1.14
--- llvm-www/releases/testregister.cgi:1.13	Thu Dec 18 13:13:14 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:14:58 2003
@@ -21,7 +21,7 @@
 #	This function checks a string to see if it contains a colon.
 #
 def checkForColon (item):
-  if (len (item) == 0):
+  if ((len (item)) == 0):
     return False
 	length = len (item) - 1
 	while (length != 0):
@@ -121,6 +121,7 @@
 	else:
 		return 'Email address has no periods'
 
+  return ''
 	#
 	# Verify that there are no colons in fields placed in the logfile.
 	#




From criswell at cs.uiuc.edu  Thu Dec 18 13:19:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:19:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181918.NAA24226@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.14 -> 1.15

---
Log message:

Modified boolean values again.



---
Diffs of the changes:  (+3 -3)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.14 llvm-www/releases/testregister.cgi:1.15
--- llvm-www/releases/testregister.cgi:1.14	Thu Dec 18 13:14:58 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:18:24 2003
@@ -22,13 +22,13 @@
 #
 def checkForColon (item):
   if ((len (item)) == 0):
-    return False
+    return 0
 	length = len (item) - 1
 	while (length != 0):
 		if (item[length] == ':'):
-			return True
+			return 1
 		length=length - 1
-	return False
+	return 0
 
 #
 # Function: Subscribe()




From criswell at cs.uiuc.edu  Thu Dec 18 13:21:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:21:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181920.NAA24343@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.15 -> 1.16

---
Log message:

Added new CGI headers so that I can see the output.



---
Diffs of the changes:  (+3 -0)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.15 llvm-www/releases/testregister.cgi:1.16
--- llvm-www/releases/testregister.cgi:1.15	Thu Dec 18 13:18:24 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:20:05 2003
@@ -200,6 +200,9 @@
 
 	return
 
+print ('Content-type: text/plain\n\n')
+print ('')
+
 #
 # Parse the CGI input
 #




From criswell at cs.uiuc.edu  Thu Dec 18 13:22:06 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:22:06 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181921.NAA24457@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.16 -> 1.17

---
Log message:

Fixed indentation problem.



---
Diffs of the changes:  (+2 -2)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.16 llvm-www/releases/testregister.cgi:1.17
--- llvm-www/releases/testregister.cgi:1.16	Thu Dec 18 13:20:05 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:21:13 2003
@@ -21,8 +21,8 @@
 #	This function checks a string to see if it contains a colon.
 #
 def checkForColon (item):
-  if ((len (item)) == 0):
-    return 0
+	if ((len (item)) == 0):
+		return 0
 	length = len (item) - 1
 	while (length != 0):
 		if (item[length] == ':'):




From criswell at cs.uiuc.edu  Thu Dec 18 13:24:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:24:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181923.NAA24576@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.17 -> 1.18

---
Log message:

Fixed more indentation issues.



---
Diffs of the changes:  (+0 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.17 llvm-www/releases/testregister.cgi:1.18
--- llvm-www/releases/testregister.cgi:1.17	Thu Dec 18 13:21:13 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:23:08 2003
@@ -121,7 +121,6 @@
 	else:
 		return 'Email address has no periods'
 
-  return ''
 	#
 	# Verify that there are no colons in fields placed in the logfile.
 	#




From criswell at cs.uiuc.edu  Thu Dec 18 13:48:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:48:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312181947.NAA24712@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.18 -> 1.19

---
Log message:

Get the program to check for a colon in the first character of the string.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.18 llvm-www/releases/testregister.cgi:1.19
--- llvm-www/releases/testregister.cgi:1.18	Thu Dec 18 13:23:08 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 13:46:59 2003
@@ -24,7 +24,7 @@
 	if ((len (item)) == 0):
 		return 0
 	length = len (item) - 1
-	while (length != 0):
+	while (length != -1):
 		if (item[length] == ':'):
 			return 1
 		length=length - 1




From criswell at cs.uiuc.edu  Thu Dec 18 13:50:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:50:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312181949.NAA24742@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.9 -> 1.10

---
Log message:

Second attempt at adding logging and required validity checking of
form fields for logging.



---
Diffs of the changes:  (+54 -17)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.9 llvm-www/releases/register.cgi:1.10
--- llvm-www/releases/register.cgi:1.9	Thu Dec 18 11:38:05 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 13:49:34 2003
@@ -4,16 +4,33 @@
 import urllib
 import smtplib
 import os
+import fcntl
 import sys
-import Cookie
 
 # List of email addresses that want to know when people download
-notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
+#notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
+notifylist=['criswell at uiuc.edu']
 
 # Full pathname of the logfile
 logfilename='/var/tmp/LLVMDownloads'
 
 #
+# Function: checkForColon()
+#
+# Description:
+#	This function checks a string to see if it contains a colon.
+#
+def checkForColon (item):
+	if ((len (item)) == 0):
+		return 0
+	length = len (item) - 1
+	while (length != -1):
+		if (item[length] == ':'):
+			return 1
+		length=length - 1
+	return 0
+
+#
 # Function: Subscribe()
 #
 # Description:
@@ -67,10 +84,11 @@
 	#
 	# Verify if the required fields have been supplied.
 	#
-	firstname = form.getfirst ('firstname', '')
-	lastname  = form.getfirst ('lastname', '')
-	email     = form.getfirst ('email', '')
-	announce  = form.getfirst ('announce','yes')
+	firstname    = form.getfirst ('firstname', '')
+	lastname     = form.getfirst ('lastname', '')
+	email        = form.getfirst ('email', '')
+	title        = form.getfirst ('title','')
+	organization = form.getfirst ('organization', '')
 
 	#
 	# Verify that the name and email fields have been filled in.
@@ -103,6 +121,24 @@
 	else:
 		return 'Email address has no periods'
 
+	#
+	# Verify that there are no colons in fields placed in the logfile.
+	#
+	if (checkForColon (email)):
+		return 'Email address has a colon'
+
+	if (checkForColon (firstname)):
+		return 'First name has a colon'
+
+	if (checkForColon (lastname)):
+		return 'Last name has a colon'
+
+	if (checkForColon (organization)):
+		return 'Organization has a colon'
+
+	if (checkForColon (title)):
+		return 'Title has a colon'
+
 	return ''
 
 #
@@ -151,14 +187,21 @@
 	#
 	# Open the logfile, lock it, and write a new entry.
 	#
-	#logfile = file (logfilename, 'a')
-	#fcntl.lockf (logfile, fcntl.LOCK_EX)
-	#logfile.write (email + ':' + lastname + ':' + firstname + '\n')
-	#logfile.flush ()
-	#logfile.close ()
+	try:
+		logfile = file (logfilename, 'a')
+		fcntl.lockf (logfile, fcntl.LOCK_EX)
+		logfile.write (email + ':' + firstname + ':' + lastname + ':' + organization + ':' + title + '\n')
+		logfile.flush ()
+		fcntl.lockf (logfile, fcntl.LOCK_UN)
+		logfile.close ()
+	except:
+		pass
 
 	return
 
+#print ('Content-type: text/plain\n\n')
+#print ('')
+
 #
 # Parse the CGI input
 #
@@ -187,12 +230,6 @@
 #
 Subscribe (form)
 
-#
-# Create a cookie so that the user does not need to register again.
-#
-#usercookie = Cookie.SimpleCookie()
-#usercookie['LLVM Download'] = 'Yes'
-#usercookie.output()
 
 #
 # Send the user to the download page.




From criswell at cs.uiuc.edu  Thu Dec 18 13:52:03 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 13:52:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312181951.NAA24860@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.10 -> 1.11

---
Log message:

Put Vikram and Chris back into the loop.



---
Diffs of the changes:  (+1 -2)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.10 llvm-www/releases/register.cgi:1.11
--- llvm-www/releases/register.cgi:1.10	Thu Dec 18 13:49:34 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 13:50:55 2003
@@ -8,8 +8,7 @@
 import sys
 
 # List of email addresses that want to know when people download
-#notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
-notifylist=['criswell at uiuc.edu']
+notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
 
 # Full pathname of the logfile
 logfilename='/var/tmp/LLVMDownloads'




From lattner at cs.uiuc.edu  Thu Dec 18 14:26:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 14:26:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/RegAllocLinearScan.cpp
Message-ID: <200312182025.OAA04377@zion.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

RegAllocLinearScan.cpp updated: 1.13 -> 1.14

---
Log message:

Prune some #includes
Add a statistic for # reloads


---
Diffs of the changes:  (+2 -3)

Index: llvm/lib/CodeGen/RegAllocLinearScan.cpp
diff -u llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.13 llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.14
--- llvm/lib/CodeGen/RegAllocLinearScan.cpp:1.13	Thu Dec 18 07:15:02 2003
+++ llvm/lib/CodeGen/RegAllocLinearScan.cpp	Thu Dec 18 14:25:31 2003
@@ -22,18 +22,16 @@
 #include "llvm/Target/MRegisterInfo.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
-#include "llvm/Target/TargetRegInfo.h"
 #include "llvm/Support/CFG.h"
 #include "Support/Debug.h"
 #include "Support/DepthFirstIterator.h"
 #include "Support/Statistic.h"
 #include "Support/STLExtras.h"
-#include 
-
 using namespace llvm;
 
 namespace {
     Statistic<> numSpilled ("ra-linearscan", "Number of registers spilled");
+    Statistic<> numReloaded("ra-linearscan", "Number of registers reloaded");
 
     class RA : public MachineFunctionPass {
     public:
@@ -762,6 +760,7 @@
     const TargetRegisterClass* rc = mf_->getSSARegMap()->getRegClass(virtReg);
     int frameIndex = getStackSlot(virtReg);
     DEBUG(std::cerr << " from stack slot #" << frameIndex << '\n');
+    ++numReloaded;
     instrAdded_ += mri_->loadRegFromStackSlot(*currentMbb_, currentInstr_,
                                               physReg, frameIndex, rc);
     assignVirt2PhysReg(virtReg, physReg);




From gaeke at cs.uiuc.edu  Thu Dec 18 14:58:01 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Thu Dec 18 14:58:01 2003
Subject: [llvm-commits] CVS: llvm/Makefile.rules
Message-ID: <200312182057.OAA16084@gally.cs.uiuc.edu>


Changes in directory llvm:

Makefile.rules updated: 1.165 -> 1.166

---
Log message:

Add install target for libraries.


---
Diffs of the changes:  (+24 -6)

Index: llvm/Makefile.rules
diff -u llvm/Makefile.rules:1.165 llvm/Makefile.rules:1.166
--- llvm/Makefile.rules:1.165	Tue Dec  9 18:26:28 2003
+++ llvm/Makefile.rules	Thu Dec 18 14:57:48 2003
@@ -71,12 +71,13 @@
 ifdef SHARED_LIBRARY
 # if SHARED_LIBRARY is specified, the default is to build the dynamic lib
 all:: dynamic
+install:: install-dynamic
 endif
 
 ifdef BYTECODE_LIBRARY
 # if BYTECODE_LIBRARY is specified, the default is to build the bytecode lib
 all:: bytecodelib
-install:: bytecodelib-install
+install:: install-bytecode-library
 endif
 
 # Default Rule:  Make sure it's also a :: rule
@@ -116,7 +117,7 @@
 # slightly since GNU Make will not try to find implicit rules for targets
 # which are marked as Phony.
 #
-.PHONY: all dynamic bytecodelib bytecodelib-install
+.PHONY: all dynamic bytecodelib install-bytecode-library
 .PHONY: clean distclean install test bytecode prdirs
 
 ###########################################################################
@@ -462,23 +463,27 @@
 LIBNAME_O    := $(DESTLIBRELEASE)/lib$(LIBRARYNAME).so
 LIBNAME_P    := $(DESTLIBPROFILE)/lib$(LIBRARYNAME).so
 LIBNAME_G    := $(DESTLIBDEBUG)/lib$(LIBRARYNAME).so
+LIBNAME_CUR  := $(DESTLIBCURRENT)/lib$(LIBRARYNAME).so
 LIBNAME_AO   := $(DESTLIBRELEASE)/lib$(LIBRARYNAME).a
 LIBNAME_AP   := $(DESTLIBPROFILE)/lib$(LIBRARYNAME).a
 LIBNAME_AG   := $(DESTLIBDEBUG)/lib$(LIBRARYNAME).a
+LIBNAME_ACUR := $(DESTLIBCURRENT)/lib$(LIBRARYNAME).a
 LIBNAME_OBJO := $(DESTLIBRELEASE)/$(LIBRARYNAME).o
 LIBNAME_OBJP := $(DESTLIBPROFILE)/$(LIBRARYNAME).o
 LIBNAME_OBJG := $(DESTLIBDEBUG)/$(LIBRARYNAME).o
+LIBNAME_OBJCUR := $(DESTLIBCURRENT)/$(LIBRARYNAME).o
 LIBNAME_BC   := $(DESTLIBBYTECODE)/lib$(LIBRARYNAME).bc
 
+
 #--------------------------------------------------------------------
 # Library Targets
 #	Modify the top level targets to build the desired libraries.
 #--------------------------------------------------------------------
 
 # dynamic target builds a shared object version of the library...
-dynamic:: $(DESTLIBCURRENT)/lib$(LIBRARYNAME).so
+dynamic:: $(LIBNAME_CUR)
 bytecodelib:: $(LIBNAME_BC)
-bytecodelib-install:: $(LLVMGCCDIR)/bytecode-libs/lib$(LIBRARYNAME).bc
+install-bytecode-library:: $(LLVMGCCDIR)/bytecode-libs/lib$(LIBRARYNAME).bc
 
 $(LLVMGCCDIR)/bytecode-libs/lib$(LIBRARYNAME).bc: $(LIBNAME_BC)
 	@${ECHO} ======= Installing $(LIBRARYNAME) bytecode library =======
@@ -486,12 +491,14 @@
 
 # Does the library want a .o version built?
 ifndef DONT_BUILD_RELINKED
-all:: $(DESTLIBCURRENT)/$(LIBRARYNAME).o
+all:: $(LIBNAME_OBJCUR)
+install:: install-single-object-library
 endif
 
 # Does the library want an archive version built?
 ifdef BUILD_ARCHIVE
-all:: $(DESTLIBCURRENT)/lib$(LIBRARYNAME).a
+all:: $(LIBNAME_ACUR)
+install:: install-archive-library
 endif
 
 #--------------------------------------------------------------------
@@ -536,6 +543,10 @@
 	$(VERB) $(LIBTOOL) --mode=install $(INSTALL) $*.la $(DESTLIBCURRENT)
 	@${ECHO} ======= Finished building $(LIBRARYNAME) dynamic debug library =======
 
+install-dynamic-library: $(LIBNAME_CUR)
+	$(MKDIR) $(libdir)
+	$(VERB) $(LIBTOOL) --mode=install $(INSTALL) $(LIBNAME_CUR) $(libdir)/lib$(LIBRARYNAME).so
+
 #
 # Rules for building static archive libraries.
 #
@@ -557,6 +568,9 @@
 	$(VERB) $(Link) -g $(STRIP) -o $@ $(ObjectsG) $(LibSubDirs) -static
 	@${ECHO} ======= Finished building $(LIBRARYNAME) archive debug library =======
 
+install-archive-library: $(LIBNAME_ACUR)
+	$(MKDIR) $(libdir)
+	$(VERB) $(LIBTOOL) --mode=install $(INSTALL) $(LIBNAME_ACUR) $(libdir)/lib$(LIBRARYNAME).a
 
 #
 # Rules for building .o libraries.
@@ -583,6 +597,10 @@
 $(LIBNAME_OBJG): $(ObjectsG) $(LibSubDirs) $(DESTLIBDEBUG)/.dir
 	@${ECHO} "Linking `basename $@`"
 	$(VERB) $(Relink) -o $@ $(RObjectsG) $(LibSubDirs)
+
+install-single-object-library: $(LIBNAME_OBJCUR)
+	$(MKDIR) $(libdir)
+	$(VERB) $(LIBTOOL) --mode=install $(INSTALL) $(LIBNAME_OBJCUR) $(libdir)/$(LIBRARYNAME).o
 
 endif
 




From criswell at cs.uiuc.edu  Thu Dec 18 14:59:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 14:59:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312182058.OAA25270@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.19 -> 1.20

---
Log message:

Try using pickling.



---
Diffs of the changes:  (+5 -4)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.19 llvm-www/releases/testregister.cgi:1.20
--- llvm-www/releases/testregister.cgi:1.19	Thu Dec 18 13:46:59 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 14:58:08 2003
@@ -6,13 +6,14 @@
 import os
 import fcntl
 import sys
+import pickle
 
 # List of email addresses that want to know when people download
 #notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
 notifylist=['criswell at uiuc.edu']
 
 # Full pathname of the logfile
-logfilename='/var/tmp/LLVMDownloads'
+logfilename='/var/tmp/testLLVMDownloads'
 
 #
 # Function: checkForColon()
@@ -190,7 +191,7 @@
 	try:
 		logfile = file (logfilename, 'a')
 		fcntl.lockf (logfile, fcntl.LOCK_EX)
-		logfile.write (email + ':' + firstname + ':' + lastname + ':' + organization + ':' + title + '\n')
+		pickle.dump ((email, firstname, lastname, organization, title), logfile) 
 		logfile.flush ()
 		fcntl.lockf (logfile, fcntl.LOCK_UN)
 		logfile.close ()
@@ -199,8 +200,8 @@
 
 	return
 
-print ('Content-type: text/plain\n\n')
-print ('')
+#print ('Content-type: text/plain\n\n')
+#print ('')
 
 #
 # Parse the CGI input




From criswell at cs.uiuc.edu  Thu Dec 18 15:02:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:02:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312182101.PAA25396@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.11 -> 1.12

---
Log message:

Use Python's pickling abilities to save all of the registration information.


---
Diffs of the changes:  (+2 -1)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.11 llvm-www/releases/register.cgi:1.12
--- llvm-www/releases/register.cgi:1.11	Thu Dec 18 13:50:55 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 15:01:03 2003
@@ -6,6 +6,7 @@
 import os
 import fcntl
 import sys
+import pickle
 
 # List of email addresses that want to know when people download
 notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
@@ -189,7 +190,7 @@
 	try:
 		logfile = file (logfilename, 'a')
 		fcntl.lockf (logfile, fcntl.LOCK_EX)
-		logfile.write (email + ':' + firstname + ':' + lastname + ':' + organization + ':' + title + '\n')
+		pickle.dump ((email, firstname, lastname, organization, title, plans), logfile) 
 		logfile.flush ()
 		fcntl.lockf (logfile, fcntl.LOCK_UN)
 		logfile.close ()




From criswell at cs.uiuc.edu  Thu Dec 18 15:19:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:19:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312182118.PAA25533@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.20 -> 1.21

---
Log message:

Attempt at implementing skipping the registration process.



---
Diffs of the changes:  (+59 -7)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.20 llvm-www/releases/testregister.cgi:1.21
--- llvm-www/releases/testregister.cgi:1.20	Thu Dec 18 14:58:08 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 15:18:19 2003
@@ -76,6 +76,42 @@
 	return
 
 #
+# Function: checkEmail()
+#
+# Description:
+#	This function searches through the database to see if the email address
+# has already been registered.
+#pickle.dump ((email, firstname, lastname, organization, title), logfile) 
+#
+def checkEmail (email):
+
+	#
+	# Assume we haven't found a matching entry.
+	#
+	foundentry = ('','','','','')
+	try:
+		#
+		# Open the logfile read only.
+		#
+		logfile = file (logfilename, 'r')
+
+		found = 0
+		while (found == 0):
+			entry = pickle.load (logfile)
+			if (entry[0] == email):
+				entry = foundentry
+				break
+	except:
+		pass
+
+	try:
+		logfile.close ()
+	except:
+		pass
+
+	return foundentry
+
+#
 # Function: ValidateForm ()
 #
 # Description:
@@ -92,14 +128,8 @@
 	organization = form.getfirst ('organization', '')
 
 	#
-	# Verify that the name and email fields have been filled in.
+	# Verify the email address first.
 	#
-	if (firstname == ''):
-		return 'First name is empty.'
-
-	if (lastname == ''):
-		return 'Last name is empty.'
-
 	if (email == ''):
 		return 'Email address is empty.'
 
@@ -123,6 +153,23 @@
 		return 'Email address has no periods'
 
 	#
+	# Determine if the email address has been used in a previous
+	# registration.
+	#
+	entry = checkEmail (email)
+	if (entry[0] == email):
+		return 'registered'
+
+	#
+	# Verify that the name fields have been filled in.
+	#
+	if (firstname == ''):
+		return 'First name is empty.'
+
+	if (lastname == ''):
+		return 'Last name is empty.'
+
+	#
 	# Verify that there are no colons in fields placed in the logfile.
 	#
 	if (checkForColon (email)):
@@ -212,6 +259,11 @@
 # Verify that everything in the form is correct.
 #
 error = ValidateForm (form)
+if (error == 'registered'):
+	print ('Location: /releases/download.html')
+	print ('')
+	sys.exit (0)
+
 if (error != ''):
 	print ('Content-type: text/html')
 	print ('Status: 400 Bad Request')




From criswell at cs.uiuc.edu  Thu Dec 18 15:26:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:26:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/testregister.cgi
Message-ID: <200312182125.PAA25656@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

testregister.cgi updated: 1.21 -> 1.22

---
Log message:

The lvalue goes on the left side of the equals sign.  :)



---
Diffs of the changes:  (+2 -2)

Index: llvm-www/releases/testregister.cgi
diff -u llvm-www/releases/testregister.cgi:1.21 llvm-www/releases/testregister.cgi:1.22
--- llvm-www/releases/testregister.cgi:1.21	Thu Dec 18 15:18:19 2003
+++ llvm-www/releases/testregister.cgi	Thu Dec 18 15:25:29 2003
@@ -99,7 +99,7 @@
 		while (found == 0):
 			entry = pickle.load (logfile)
 			if (entry[0] == email):
-				entry = foundentry
+				foundentry = entry
 				break
 	except:
 		pass
@@ -157,7 +157,7 @@
 	# registration.
 	#
 	entry = checkEmail (email)
-	if (entry[0] == email):
+	if (entry[0] != ''):
 		return 'registered'
 
 	#




From criswell at cs.uiuc.edu  Thu Dec 18 15:28:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:28:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.cgi
Message-ID: <200312182127.PAA25780@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.cgi updated: 1.12 -> 1.13

---
Log message:

Attempt to accept only the email address if the user has already registered.



---
Diffs of the changes:  (+60 -8)

Index: llvm-www/releases/register.cgi
diff -u llvm-www/releases/register.cgi:1.12 llvm-www/releases/register.cgi:1.13
--- llvm-www/releases/register.cgi:1.12	Thu Dec 18 15:01:03 2003
+++ llvm-www/releases/register.cgi	Thu Dec 18 15:26:59 2003
@@ -75,6 +75,42 @@
 	return
 
 #
+# Function: checkEmail()
+#
+# Description:
+#	This function searches through the database to see if the email address
+# has already been registered.
+#pickle.dump ((email, firstname, lastname, organization, title), logfile) 
+#
+def checkEmail (email):
+
+	#
+	# Assume we haven't found a matching entry.
+	#
+	foundentry = ('','','','','')
+	try:
+		#
+		# Open the logfile read only.
+		#
+		logfile = file (logfilename, 'r')
+
+		found = 0
+		while (found == 0):
+			entry = pickle.load (logfile)
+			if (entry[0] == email):
+				foundentry = entry
+				break
+	except:
+		pass
+
+	try:
+		logfile.close ()
+	except:
+		pass
+
+	return foundentry
+
+#
 # Function: ValidateForm ()
 #
 # Description:
@@ -91,14 +127,8 @@
 	organization = form.getfirst ('organization', '')
 
 	#
-	# Verify that the name and email fields have been filled in.
+	# Verify the email address first.
 	#
-	if (firstname == ''):
-		return 'First name is empty.'
-
-	if (lastname == ''):
-		return 'Last name is empty.'
-
 	if (email == ''):
 		return 'Email address is empty.'
 
@@ -122,6 +152,23 @@
 		return 'Email address has no periods'
 
 	#
+	# Determine if the email address has been used in a previous
+	# registration.
+	#
+	entry = checkEmail (email)
+	if (entry[0] != ''):
+		return 'registered'
+
+	#
+	# Verify that the name fields have been filled in.
+	#
+	if (firstname == ''):
+		return 'First name is empty.'
+
+	if (lastname == ''):
+		return 'Last name is empty.'
+
+	#
 	# Verify that there are no colons in fields placed in the logfile.
 	#
 	if (checkForColon (email)):
@@ -190,7 +237,7 @@
 	try:
 		logfile = file (logfilename, 'a')
 		fcntl.lockf (logfile, fcntl.LOCK_EX)
-		pickle.dump ((email, firstname, lastname, organization, title, plans), logfile) 
+		pickle.dump ((email, firstname, lastname, organization, title), logfile) 
 		logfile.flush ()
 		fcntl.lockf (logfile, fcntl.LOCK_UN)
 		logfile.close ()
@@ -211,6 +258,11 @@
 # Verify that everything in the form is correct.
 #
 error = ValidateForm (form)
+if (error == 'registered'):
+	print ('Location: /releases/download.html')
+	print ('')
+	sys.exit (0)
+
 if (error != ''):
 	print ('Content-type: text/html')
 	print ('Status: 400 Bad Request')




From criswell at cs.uiuc.edu  Thu Dec 18 15:40:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:40:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.html
Message-ID: <200312182139.PAA25910@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.html updated: 1.1 -> 1.2

---
Log message:

Added a "Quick Download" link.



---
Diffs of the changes:  (+20 -1)

Index: llvm-www/releases/register.html
diff -u llvm-www/releases/register.html:1.1 llvm-www/releases/register.html:1.2
--- llvm-www/releases/register.html:1.1	Fri Dec 12 16:09:41 2003
+++ llvm-www/releases/register.html	Thu Dec 18 15:39:15 2003
@@ -16,7 +16,26 @@
 


 We'd like to collect some information about people and organizations who are
 interested in LLVM.  Please complete the following form below to download a copy
-of the LLVM software.
+of the LLVM software, or use the Quick Download if you have registered since
+December 19, 2003.
+
+
+Definition Of Operation Of Built In Words
LOGICAL OPERATIONS
Word Name Operation Description
< LT w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is less than w2, TRUE is pushed back on
     the stack, otherwise FALSE is pushed back on the stack.
 
> GT w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
> GT w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is greater than w2, TRUE is pushed back on
     the stack, otherwise FALSE is pushed back on the stack.
 
>= GE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
>= GE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is greater than or equal to w2, TRUE is 
     pushed back on the stack, otherwise FALSE is pushed back 
     on the stack.
 
<= LE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
<= LE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is less than or equal to w2, TRUE is 
     pushed back on the stack, otherwise FALSE is pushed back 
     on the stack.
 
= EQ w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
= EQ w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is equal to w2, TRUE is 
     pushed back on the stack, otherwise FALSE is pushed back 
     
 
<> NE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
+
<> NE w1 w2 -- b Two values (w1 and w2) are popped off the stack and
     compared. If w1 is equal to w2, TRUE is 
     pushed back on the stack, otherwise FALSE is pushed back 
     
 
FALSE FALSE  -- b The boolean value FALSE (0) is pushed onto the stack.
TRUE TRUE  -- b The boolean value TRUE (-1) is pushed onto the stack.
BITWISE OPERATIONS
Word Name Operation Description
<< SHL w1 w2 -- w1<<w2 Two values (w1 and w2) are popped off the stack. The w2
+
FALSE FALSE  -- b The boolean value FALSE (0) is pushed onto the stack.
TRUE TRUE  -- b The boolean value TRUE (-1) is pushed onto the stack.
BITWISE OPERATORS
Word Name Operation Description
<< SHL w1 w2 -- w1<<w2 Two values (w1 and w2) are popped off the stack. The w2
     operand is shifted left by the number of bits given by the
     w1 operand. The result is pushed back to the stack.
 
>> SHR w1 w2 -- w1>>w2 Two values (w1 and w2) are popped off the stack. The w2
+
>> SHR w1 w2 -- w1>>w2 Two values (w1 and w2) are popped off the stack. The w2
     operand is shifted right by the number of bits given by the
     w1 operand. The result is pushed back to the stack.
 
OR OR w1 w2 -- w2|w1 Two values (w1 and w2) are popped off the stack. The values
+
OR OR w1 w2 -- w2|w1 Two values (w1 and w2) are popped off the stack. The values
     are bitwise OR'd together and pushed back on the stack. This is 
     not a logical OR. The sequence 1 2 OR yields 3 not 1.
 
AND AND w1 w2 -- w2&w1 Two values (w1 and w2) are popped off the stack. The values
+
AND AND w1 w2 -- w2&w1 Two values (w1 and w2) are popped off the stack. The values
     are bitwise AND'd together and pushed back on the stack. This is 
     not a logical AND. The sequence 1 2 AND yields 0 not 1.
 
XOR XOR w1 w2 -- w2^w1 Two values (w1 and w2) are popped off the stack. The values
+
XOR XOR w1 w2 -- w2^w1 Two values (w1 and w2) are popped off the stack. The values
     are bitwise exclusive OR'd together and pushed back on the stack. 
     For example, The sequence 1 3 XOR yields 2.
 
ARITHMETIC OPERATIONS
Word Name Operation Description
ABS ABS w -- |w| One value s popped off the stack; its absolute value is computed
+
ARITHMETIC OPERATORS
Word Name Operation Description
ABS ABS w -- |w| One value s popped off the stack; its absolute value is computed
     and then pushed onto the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is also 1.
 
NEG NEG w -- -w One value is popped off the stack which is negated and then
+
NEG NEG w -- -w One value is popped off the stack which is negated and then
     pushed back onto the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is -1.
 
 + ADD w1 w2 -- w2+w1 Two values are popped off the stack. Their sum is pushed back
+
 + ADD w1 w2 -- w2+w1 Two values are popped off the stack. Their sum is pushed back
     onto the stack
 
 - SUB w1 w2 -- w2-w1 Two values are popped off the stack. Their difference is pushed back
+
 - SUB w1 w2 -- w2-w1 Two values are popped off the stack. Their difference is pushed back
     onto the stack
 
 * MUL w1 w2 -- w2*w1 Two values are popped off the stack. Their product is pushed back
+
 * MUL w1 w2 -- w2*w1 Two values are popped off the stack. Their product is pushed back
     onto the stack
 
 / DIV w1 w2 -- w2/w1 Two values are popped off the stack. Their quotient is pushed back
+
 / DIV w1 w2 -- w2/w1 Two values are popped off the stack. Their quotient is pushed back
     onto the stack
 
MOD MOD w1 w2 -- w2%w1 Two values are popped off the stack. Their remainder after division
+
MOD MOD w1 w2 -- w2%w1 Two values are popped off the stack. Their remainder after division
     of w1 by w2 is pushed back onto the stack
 
 */ STAR_SLAH w1 w2 w3 -- (w3*w2)/w1 Three values are popped off the stack. The product of w1 and w2 is
+
 */ STAR_SLAH w1 w2 w3 -- (w3*w2)/w1 Three values are popped off the stack. The product of w1 and w2 is
     divided by w3. The result is pushed back onto the stack.
 
 ++ INCR w -- w+1 One value is popped off the stack. It is incremented by one and then
+
 ++ INCR w -- w+1 One value is popped off the stack. It is incremented by one and then
     pushed back onto the stack.
 
 -- DECR w -- w-1 One value is popped off the stack. It is decremented by one and then
+
 -- DECR w -- w-1 One value is popped off the stack. It is decremented by one and then
     pushed back onto the stack.
 
MIN MIN w1 w2 -- (w2<w1?w2:w1) Two values are popped off the stack. The larger one is pushed back
+
MIN MIN w1 w2 -- (w2<w1?w2:w1) Two values are popped off the stack. The larger one is pushed back
     onto the stack.
 
MAX MAX w1 w2 -- (w2>w1?w2:w1) Two values are popped off the stack. The larger value is pushed back
+
MAX MAX w1 w2 -- (w2>w1?w2:w1) Two values are popped off the stack. The larger value is pushed back
 	onto the stack.
 
STACK MANIPULATION OPERATIONS
Word Name Operation Description
DROP DROP w -- One value is popped off the stack.
DROP2 DROP2 w1 w2 -- Two values are popped off the stack.
NIP NIP w1 w2 -- w2 The second value on the stack is removed from the stack. That is,
+
STACK MANIPULATION OPERATORS
Word Name Operation Description
DROP DROP w -- One value is popped off the stack.
DROP2 DROP2 w1 w2 -- Two values are popped off the stack.
NIP NIP w1 w2 -- w2 The second value on the stack is removed from the stack. That is,
 	a value is popped off the stack and retained. Then a second value is
 	popped and the retained value is pushed.
 
NIP2 NIP2 w1 w2 w3 w4 -- w3 w4 The third and fourth values on the stack are removed from it. That is,
+
NIP2 NIP2 w1 w2 w3 w4 -- w3 w4 The third and fourth values on the stack are removed from it. That is,
 	two values are popped and retained. Then two more values are popped and
 	the two retained values are pushed back on.
 
DUP DUP w1 -- w1 w1 One value is popped off the stack. That value is then pushed onto
+
DUP DUP w1 -- w1 w1 One value is popped off the stack. That value is then pushed onto
 	the stack twice to duplicate the top stack vaue.
 
DUP2 DUP2 w1 w2 -- w1 w2 w1 w2 The top two values on the stack are duplicated. That is, two vaues
+
DUP2 DUP2 w1 w2 -- w1 w2 w1 w2 The top two values on the stack are duplicated. That is, two vaues
 	are popped off the stack. They are alternately pushed back on the
 	stack twice each.
 
SWAP SWAP w1 w2 -- w2 w1 The top two stack items are reversed in their order. That is, two
+
SWAP SWAP w1 w2 -- w2 w1 The top two stack items are reversed in their order. That is, two
 	values are popped off the stack and pushed back onto the stack in
 	the opposite order they were popped.
 
SWAP2 SWAP2 w1 w2 w3 w4 -- w3 w4 w2 w1 The top four stack items are swapped in pairs. That is, two values
+
SWAP2 SWAP2 w1 w2 w3 w4 -- w3 w4 w2 w1 The top four stack items are swapped in pairs. That is, two values
 	are popped and retained. Then, two more values are popped and retained.
 	The values are pushed back onto the stack in the reverse order but
 	in pairs.
 
OVER OVER w1 w2-- w1 w2 w1 Two values are popped from the stack. They are pushed back
+
OVER OVER w1 w2-- w1 w2 w1 Two values are popped from the stack. They are pushed back
 	onto the stack in the order w1 w2 w1. This seems to cause the
 	top stack element to be duplicated "over" the next value.
 
OVER2 OVER2 w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2 The third and fourth values on the stack are replicated onto the
+
OVER2 OVER2 w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2 The third and fourth values on the stack are replicated onto the
 	top of the stack
 
ROT ROT w1 w2 w3 -- w2 w3 w1 The top three values are rotated. That is, three value are popped
+
ROT ROT w1 w2 w3 -- w2 w3 w1 The top three values are rotated. That is, three value are popped
 	off the stack. They are pushed back onto the stack in the order
 	w1 w3 w2.
 
ROT2 ROT2 w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2 Like ROT but the rotation is done using three pairs instead of
+
ROT2 ROT2 w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2 Like ROT but the rotation is done using three pairs instead of
 	three singles.
 
RROT RROT w1 w2 w3 -- w2 w3 w1 Reverse rotation. Like ROT, but it rotates the other way around.
+
RROT RROT w1 w2 w3 -- w2 w3 w1 Reverse rotation. Like ROT, but it rotates the other way around.
 	Essentially, the third element on the stack is moved to the top
 	of the stack.
 
RROT2 RROT2 w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2 Double reverse rotation. Like RROT but the rotation is done using 
+
RROT2 RROT2 w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2 Double reverse rotation. Like RROT but the rotation is done using 
 	three pairs instead of three singles. The fifth and sixth stack 
 	elements are moved to the first and second positions
 
TUCK TUCK w1 w2 -- w2 w1 w2 Similar to OVER except that the second operand is being 
+
TUCK TUCK w1 w2 -- w2 w1 w2 Similar to OVER except that the second operand is being 
 	replicated. Essentially, the first operand is being "tucked"
 	in between two instances of the second operand. Logically, two
 	values are popped off the stack. They are placed back on the
 	stack in the order w2 w1 w2.
 
TUCK2 TUCK2 w1 w2 w3 w4 -- w3 w4 w1 w2 w3 w4 Like TUCK but a pair of elements is tucked over two pairs.
+
TUCK2 TUCK2 w1 w2 w3 w4 -- w3 w4 w1 w2 w3 w4 Like TUCK but a pair of elements is tucked over two pairs.
 	That is, the top two elements of the stack are duplicated and
 	inserted into the stack at the fifth and positions.
 
PICK PICK x0 ... Xn n -- x0 ... Xn x0 The top of the stack is used as an index into the remainder of
+
PICK PICK x0 ... Xn n -- x0 ... Xn x0 The top of the stack is used as an index into the remainder of
 	the stack. The element at the nth position replaces the index 
 	(top of stack). This is useful for cycling through a set of 
 	values. Note that indexing is zero based. So, if n=0 then you
 	get the second item on the stack. If n=1 you get the third, etc.
 	Note also that the index is replaced by the n'th value. 
 
SELECT SELECT m n X0..Xm Xm+1 .. Xn -- Xm This is like PICK but the list is removed and you need to specify
+
SELECT SELECT m n X0..Xm Xm+1 .. Xn -- Xm This is like PICK but the list is removed and you need to specify
 	both the index and the size of the list. Careful with this one,
 	the wrong value for n can blow away a huge amount of the stack.
 
ROLL ROLL x0 x1 .. xn n -- x1 .. xn x0 Not Implemented. This one has been left as an exercise to
+
ROLL ROLL x0 x1 .. xn n -- x1 .. xn x0 Not Implemented. This one has been left as an exercise to
 	the student. See Exercise. ROLL requires 
     a value, "n", to be on the top of the stack. This value specifies how 
     far into the stack to "roll". The n'th value is moved (not
@@ -814,20 +841,25 @@
     how much to rotate. That is, ROLL with n=1 is the same as ROT and 
     ROLL with n=2 is the same as ROT2.
 
MEMORY OPERATIONS
Word Name Operation Description
MALLOC MALLOC w1 -- p One value is popped off the stack. The value is used as the size
+
MEMORY OPERATORS
Word Name Operation Description
MALLOC MALLOC w1 -- p One value is popped off the stack. The value is used as the size
 	of a memory block to allocate. The size is in bytes, not words.
         The memory allocation is completed and the address of the memory
 	block is pushed onto the stack.
 
FREE FREE p -- One pointer value is popped off the stack. The value should be
+
FREE FREE p -- One pointer value is popped off the stack. The value should be
 	the address of a memory block created by the MALLOC operation. The
 	associated memory block is freed. Nothing is pushed back on the
 	stack. Many bugs can be created by attempting to FREE something
@@ -839,20 +871,20 @@
 	the stack (for the FREE at the end) and that every use of the
 	pointer is preceded by a DUP to retain the copy for FREE.
 
GET GET w1 p -- w2 p An integer index and a pointer to a memory block are popped of
+
GET GET w1 p -- w2 p An integer index and a pointer to a memory block are popped of
 	the block. The index is used to index one byte from the memory
 	block. That byte value is retained, the pointer is pushed again
 	and the retained value is pushed. Note that the pointer value
 	s essentially retained in its position so this doesn't count
 	as a "use ptr" in the FREE idiom.
 
PUT PUT w1 w2 p -- p An integer value is popped of the stack. This is the value to
+
PUT PUT w1 w2 p -- p An integer value is popped of the stack. This is the value to
 	be put into a memory block. Another integer value is popped of
 	the stack. This is the indexed byte in the memory block. A
 	pointer to the memory block is popped off the stack. The
@@ -862,28 +894,33 @@
 	pushed back on the stack so this doesn't count as a "use ptr"
 	in the FREE idiom.
 
CONTROL FLOW OPERATIONS
Word Name Operation Description
RETURN RETURN  --  The currently executing definition returns immediately to its caller.
+
CONTROL FLOW OPERATORS
Word Name Operation Description
RETURN RETURN  --  The currently executing definition returns immediately to its caller.
 	Note that there is an implicit RETURN at the end of each
 	definition, logically located at the semi-colon. The sequence 
 	RETURN ;  is valid but redundant.
 
EXIT EXIT w1 -- A return value for the program is popped off the stack. The program is
+
EXIT EXIT w1 -- A return value for the program is popped off the stack. The program is
 	then immediately terminated. This is normally an abnormal exit from the
 	program. For a normal exit (when MAIN finishes), the exit
 	code will always be zero in accordance with UNIX conventions.
 
RECURSE RECURSE  -- The currently executed definition is called again. This operation is 
+
RECURSE RECURSE  -- The currently executed definition is called again. This operation is 
 	needed since the definition of a word doesn't exist until the semi colon
 	is reacher. Attempting something like:

 	 : recurser recurser ; 
 will yield and error saying that 
@@ -891,24 +928,24 @@
 	to:

 	 : recurser RECURSE ; 
 
IF (words...) ENDIF IF (words...) ENDIF b -- A boolean value is popped of the stack. If it is non-zero then the "words..." 
+
IF (words...) ENDIF IF (words...) ENDIF b -- A boolean value is popped of the stack. If it is non-zero then the "words..." 
 	are executed. Otherwise, execution continues immediately following the ENDIF.
 
IF (words...) ELSE (words...) ENDIF IF (words...) ELSE (words...) ENDIF b -- A boolean value is popped of the stack. If it is non-zero then the "words..."
+
IF (words...) ELSE (words...) ENDIF IF (words...) ELSE (words...) ENDIF b -- A boolean value is popped of the stack. If it is non-zero then the "words..."
 	between IF and ELSE are executed. Otherwise the words between ELSE and ENDIF are
 	executed. In either case, after the (words....) have executed, execution continues
         immediately following the ENDIF. 
 
WHILE (words...) END WHILE (words...) END b -- b The boolean value on the top of the stack is examined. If it is non-zero then the 
+
WHILE (words...) END WHILE (words...) END b -- b The boolean value on the top of the stack is examined. If it is non-zero then the 
 	"words..." between WHILE and END are executed. Execution then begins again at the WHILE where another
 	boolean is popped off the stack. To prevent this operation from eating up the entire
 	stack, you should push onto the stack (just before the END) a boolean value that indicates
@@ -924,60 +961,65 @@
 	the top of stack is decremented to 0 at which the WHILE test fails and control is
 	transfered to the word after the END.
 
INPUT & OUTPUT OPERATIONS
Word Name Operation Description
SPACE SPACE  --  A space character is put out. There is no stack effect.
TAB TAB  --  A tab character is put out. There is no stack effect.
CR CR  --  A carriage return character is put out. There is no stack effect.
>s OUT_STR  -- A string pointer is popped from the stack. It is put out.
>d OUT_STR  -- A value is popped from the stack. It is put out as a decimal integer.
>c OUT_CHR  -- A value is popped from the stack. It is put out as an ASCII character.
<s IN_STR  -- s A string is read from the input via the scanf(3) format string " %as". The
+
INPUT & OUTPUT OPERATORS
Word Name Operation Description
SPACE SPACE  --  A space character is put out. There is no stack effect.
TAB TAB  --  A tab character is put out. There is no stack effect.
CR CR  --  A carriage return character is put out. There is no stack effect.
>s OUT_STR  -- A string pointer is popped from the stack. It is put out.
>d OUT_STR  -- A value is popped from the stack. It is put out as a decimal integer.
>c OUT_CHR  -- A value is popped from the stack. It is put out as an ASCII character.
<s IN_STR  -- s A string is read from the input via the scanf(3) format string " %as". The
 	resulting string is pushed onto the stack.
 
<d IN_STR  -- w An integer is read from the input via the scanf(3) format string " %d". The
+
<d IN_STR  -- w An integer is read from the input via the scanf(3) format string " %d". The
 	resulting value is pushed onto the stack
 
<c IN_CHR  -- w A single character is read from the input via the scanf(3) format string 
+
<c IN_CHR  -- w A single character is read from the input via the scanf(3) format string 
 	" %c". The value is converted to an integer and pushed onto the stack.
 
DUMP DUMP  -- The stack contents are dumped to standard output. This is useful for
+
DUMP DUMP  -- The stack contents are dumped to standard output. This is useful for
 	debugging your definitions. Put DUMP at the beginning and end of a definition
 	to see instantly the net effect of the definition.
 
 
 
 
 
 
 FALSE
     FALSE
      -- b The boolean value FALSE (0) is pushed onto the stack. The boolean value FALSE (0) is pushed on to the stack.
 
 TRUE
     TRUE
      -- b The boolean value TRUE (-1) is pushed onto the stack. The boolean value TRUE (-1) is pushed on to the stack.
 
 BITWISE OPERATORS
 ABS
     w -- |w|
     One value s popped off the stack; its absolute value is computed
-    and then pushed onto the stack. If w1 is -1 then w2 is 1. If w1 is
+    and then pushed on to the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is also 1.
 
 NEG
     NEG
     w -- -w
     One value is popped off the stack which is negated and then
-    pushed back onto the stack. If w1 is -1 then w2 is 1. If w1 is
+    pushed back on to the stack. If w1 is -1 then w2 is 1. If w1 is
     1 then w2 is -1.
 
  + 
     ADD
     w1 w2 -- w2+w1
     Two values are popped off the stack. Their sum is pushed back
-    onto the stack
  - 
     SUB
     w1 w2 -- w2-w1
     Two values are popped off the stack. Their difference is pushed back
-    onto the stack
  * 
     MUL
     w1 w2 -- w2*w1
     Two values are popped off the stack. Their product is pushed back
-    onto the stack
  / 
     DIV
     w1 w2 -- w2/w1
     Two values are popped off the stack. Their quotient is pushed back
-    onto the stack
 MOD
     MOD
     w1 w2 -- w2%w1
     Two values are popped off the stack. Their remainder after division
-    of w1 by w2 is pushed back onto the stack
  */ 
     STAR_SLAH
     w1 w2 w3 -- (w3*w2)/w1
     Three values are popped off the stack. The product of w1 and w2 is
-    divided by w3. The result is pushed back onto the stack.
  ++ 
     INCR
     w -- w+1
     One value is popped off the stack. It is incremented by one and then
-    pushed back onto the stack.
  -- 
     DECR
     w -- w-1
     One value is popped off the stack. It is decremented by one and then
-    pushed back onto the stack.
 MIN
     MIN
     w1 w2 -- (w2<w1?w2:w1)
     Two values are popped off the stack. The larger one is pushed back
-    onto the stack.
 MAX
     MAX
     w1 w2 -- (w2>w1?w2:w1)
     Two values are popped off the stack. The larger value is pushed back
-	onto the stack.
 STACK MANIPULATION OPERATORS
 
DUP
     DUP
     w1 -- w1 w1 One value is popped off the stack. That value is then pushed onto
+    One value is popped off the stack. That value is then pushed on to
 	the stack twice to duplicate the top stack vaue.
 
 DUP2 SWAP
     w1 w2 -- w2 w1
     The top two stack items are reversed in their order. That is, two
-	values are popped off the stack and pushed back onto the stack in
+	values are popped off the stack and pushed back on to the stack in
 	the opposite order they were popped.
 
 SWAP2 w1 w2 w3 w4 -- w3 w4 w2 w1
     The top four stack items are swapped in pairs. That is, two values
 	are popped and retained. Then, two more values are popped and retained.
-	The values are pushed back onto the stack in the reverse order but
+	The values are pushed back on to the stack in the reverse order but
 	in pairs.
 
 OVER
     OVER
     w1 w2-- w1 w2 w1
     Two values are popped from the stack. They are pushed back
-	onto the stack in the order w1 w2 w1. This seems to cause the
+	on to the stack in the order w1 w2 w1. This seems to cause the
 	top stack element to be duplicated "over" the next value.
 
 OVER2
     OVER2
     w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2 The third and fourth values on the stack are replicated onto the
+    The third and fourth values on the stack are replicated on to the
 	top of the stack
 
 ROT
     ROT
     w1 w2 w3 -- w2 w3 w1
     The top three values are rotated. That is, three value are popped
-	off the stack. They are pushed back onto the stack in the order
+	off the stack. They are pushed back on to the stack in the order
 	w1 w3 w2.
 
 ROT2 One value is popped off the stack. The value is used as the size
 	of a memory block to allocate. The size is in bytes, not words.
         The memory allocation is completed and the address of the memory
-	block is pushed onto the stack.
 FREE
     FREE The boolean value on the top of the stack is examined. If it is non-zero then the 
 	"words..." between WHILE and END are executed. Execution then begins again at the WHILE where another
 	boolean is popped off the stack. To prevent this operation from eating up the entire
-	stack, you should push onto the stack (just before the END) a boolean value that indicates
+	stack, you should push on to the stack (just before the END) a boolean value that indicates
 	whether to terminate. Note that since booleans and integers can be coerced you can
 	use the following "for loop" idiom:

 	(push count) WHILE (words...) -- END

@@ -1002,19 +1003,19 @@
     IN_STR
      -- s 
     A string is read from the input via the scanf(3) format string " %as". The
-	resulting string is pushed onto the stack.
 <d
     IN_STR
      -- w 
     An integer is read from the input via the scanf(3) format string " %d". The
-	resulting value is pushed onto the stack
 <c
     IN_CHR
      -- w 
     A single character is read from the input via the scanf(3) format string 
-	" %c". The value is converted to an integer and pushed onto the stack.
 DUMP
     DUMP

+
+Quick Download

+
+
+
+
+
+
+
+
+  
+  
+
+
Email Address (required): 
+
 
 
 




From criswell at cs.uiuc.edu  Thu Dec 18 15:42:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:42:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.html
Message-ID: <200312182141.PAA26029@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.html updated: 1.2 -> 1.3

---
Log message:

Give the Quick Download section its own form.



---
Diffs of the changes:  (+2 -0)

Index: llvm-www/releases/register.html
diff -u llvm-www/releases/register.html:1.2 llvm-www/releases/register.html:1.3
--- llvm-www/releases/register.html:1.2	Thu Dec 18 15:39:15 2003
+++ llvm-www/releases/register.html	Thu Dec 18 15:41:18 2003
@@ -36,6 +36,8 @@
 
 
 
+
+
 

 
 




From criswell at cs.uiuc.edu  Thu Dec 18 15:52:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:52:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/quickreg.cgi
Message-ID: <200312182151.PAA26178@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

quickreg.cgi added (r1.1)

---
Log message:

Script used to do quick registration.



---
Diffs of the changes:  (+130 -0)

Index: llvm-www/releases/quickreg.cgi
diff -c /dev/null llvm-www/releases/quickreg.cgi:1.1
*** /dev/null	Thu Dec 18 15:51:43 2003
--- llvm-www/releases/quickreg.cgi	Thu Dec 18 15:51:33 2003
***************
*** 0 ****
--- 1,130 ----
+ #!/usr/dcs/software/supported/bin/python
+ 
+ import cgi
+ import urllib
+ import smtplib
+ import os
+ import fcntl
+ import sys
+ import pickle
+ 
+ # List of email addresses that want to know when people download
+ notifylist=['vadve at cs.uiuc.edu', 'criswell at uiuc.edu', 'sabre at nondot.org']
+ 
+ # Full pathname of the logfile
+ logfilename='/var/tmp/LLVMDownloads'
+ 
+ #
+ # Function: checkEmail()
+ #
+ # Description:
+ #	This function searches through the database to see if the email address
+ # has already been registered.
+ #pickle.dump ((email, firstname, lastname, organization, title), logfile) 
+ #
+ def checkEmail (email):
+ 
+ 	#
+ 	# Assume we haven't found a matching entry.
+ 	#
+ 	foundentry = ('','','','','')
+ 	try:
+ 		#
+ 		# Open the logfile read only.
+ 		#
+ 		logfile = file (logfilename, 'r')
+ 
+ 		found = 0
+ 		while (found == 0):
+ 			entry = pickle.load (logfile)
+ 			if (entry[0] == email):
+ 				foundentry = entry
+ 				break
+ 	except:
+ 		pass
+ 
+ 	try:
+ 		logfile.close ()
+ 	except:
+ 		pass
+ 
+ 	return foundentry
+ 
+ #
+ # Function: ValidateForm ()
+ #
+ # Description:
+ #	Make sure that the input to the CGI script is valid.
+ #
+ def ValidateForm (form):
+ 	#
+ 	# Verify if the required fields have been supplied.
+ 	#
+ 	firstname    = form.getfirst ('firstname', '')
+ 	lastname     = form.getfirst ('lastname', '')
+ 	email        = form.getfirst ('email', '')
+ 	title        = form.getfirst ('title','')
+ 	organization = form.getfirst ('organization', '')
+ 
+ 	#
+ 	# Verify the email address first.
+ 	#
+ 	if (email == ''):
+ 		return 'Email address is empty.'
+ 
+ 	#
+ 	# Verify that the email address has an at sign and some periods.
+ 	#
+ 	length = len (email) - 1
+ 	while (length != 0):
+ 		if (email[length] == '@'):
+ 			break
+ 		length=length - 1
+ 	else:
+ 		return 'Email address has no at sign'
+ 
+ 	length = len (email) - 1
+ 	while (length != 0):
+ 		if (email[length] == '.'):
+ 			break
+ 		length=length - 1
+ 	else:
+ 		return 'Email address has no periods'
+ 
+ 	#
+ 	# Determine if the email address has been used in a previous
+ 	# registration.
+ 	#
+ 	entry = checkEmail (email)
+ 	if (entry[0] != ''):
+ 		return 'registered'
+ 	return 'unregistered'
+ 
+ #
+ # Parse the CGI input
+ #
+ form = cgi.FieldStorage ()
+ 
+ #
+ # Verify that everything in the form is correct.
+ #
+ error = ValidateForm (form)
+ if (error == 'registered'):
+ 	print ('Location: /releases/download.html')
+ 	print ('')
+ 	sys.exit (0)
+ 
+ #
+ # Send the user to the download page.
+ #
+ print ('Content-type: text/html')
+ print ('')
+ 
+ print ('Sorry; not registered')
+ print ('')
+ print ('We are sorry, but we cannot find your email address in our database.')
+ print ('Please use the normal registration system.')
+ print ('')
+ 
+ sys.exit (0)
+ 




From criswell at cs.uiuc.edu  Thu Dec 18 15:53:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Thu Dec 18 15:53:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.html
Message-ID: <200312182152.PAA26295@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.html updated: 1.3 -> 1.4

---
Log message:

Use a different script for quick registration.



---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/register.html
diff -u llvm-www/releases/register.html:1.3 llvm-www/releases/register.html:1.4
--- llvm-www/releases/register.html:1.3	Thu Dec 18 15:41:18 2003
+++ llvm-www/releases/register.html	Thu Dec 18 15:52:44 2003
@@ -26,7 +26,7 @@
 
 
 
-
+
 
 





From gaeke at cs.uiuc.edu  Thu Dec 18 16:36:02 2003
From: gaeke at cs.uiuc.edu (Brian Gaeke)
Date: Thu Dec 18 16:36:02 2003
Subject: [llvm-commits] CVS: reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp
Message-ID: <200312182235.QAA18321@seraph.cs.uiuc.edu>


Changes in directory reopt/lib/LightWtProfiling:

UnpackTraceFunction.cpp updated: 1.22 -> 1.23

---
Log message:

Remove the bitrot induced by Misha's recent changes.

---
Diffs of the changes:  (+7 -6)

Index: reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp
diff -u reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp:1.22 reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp:1.23
--- reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp:1.22	Wed Nov 19 16:51:49 2003
+++ reopt/lib/LightWtProfiling/UnpackTraceFunction.cpp	Thu Dec 18 16:35:09 2003
@@ -8,7 +8,8 @@
 #include "TraceToFunction.h"
 #include "../../../../lib/CodeGen/RegAlloc/AllocInfo.h"
 #include "../../../../lib/CodeGen/RegAlloc/PhyRegAlloc.h"
-#include "../../../../lib/Target/Sparc/SparcInternals.h"
+#include "../../../../lib/Target/Sparc/SparcTargetMachine.h"
+#include "../../../../lib/Target/Sparc/SparcRegInfo.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/Support/InstIterator.h"
@@ -23,17 +24,17 @@
 ///
 void insertCopyMachineInstrs (AllocInfo &Source, AllocInfo &Target,
                               MachineBasicBlock &B, const Type *Ty) {
-  UltraSparc TM;
+  SparcTargetMachine TM;
   const TargetRegInfo &TRI = TM.getRegInfo ();
   std::vector mvec;
   int RegType;
   // Guess what kind of reg the register was allocated to.
   if (Ty == Type::FloatTy)
-    RegType = UltraSparcRegInfo::FPSingleRegType;
+    RegType = SparcRegInfo::FPSingleRegType;
   else if (Ty == Type::DoubleTy)
-    RegType = UltraSparcRegInfo::FPDoubleRegType;
+    RegType = SparcRegInfo::FPDoubleRegType;
   else
-    RegType = UltraSparcRegInfo::IntRegType;
+    RegType = SparcRegInfo::IntRegType;
   const unsigned FramePtrReg = TRI.getFramePointer ();
   const unsigned StackPtrReg = TRI.getStackPointer ();
   if (Source.AllocState == AllocInfo::Spilled
@@ -53,7 +54,7 @@
   } else if (Source.AllocState == AllocInfo::Spilled
              && Target.AllocState == AllocInfo::Spilled) {
     const unsigned TempReg = 
-      TRI.getUnifiedRegNum (UltraSparcRegInfo::IntRegClassID,
+      TRI.getUnifiedRegNum (SparcRegInfo::IntRegClassID,
                             SparcIntRegClass::g1); // just pick one?
     // Emit load instruction from stack loc. Source into register TempReg
     TRI.cpMem2RegMI (mvec, FramePtrReg, Source.Placement, TempReg, RegType);




From alkis at cs.uiuc.edu  Thu Dec 18 16:41:01 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 16:41:01 2003
Subject: [llvm-commits] CVS: llvm/lib/CodeGen/TwoAddressInstructionPass.cpp RegAllocLocal.cpp LiveIntervals.cpp
Message-ID: <200312182240.QAA00303@morpheus.cs.uiuc.edu>


Changes in directory llvm/lib/CodeGen:

TwoAddressInstructionPass.cpp updated: 1.1 -> 1.2
RegAllocLocal.cpp updated: 1.35 -> 1.36
LiveIntervals.cpp updated: 1.11 -> 1.12

---
Log message:

Remove TwoAddressInstruction from the public headers and add an ID
instead, since this pass doesn't expose any state to its users.


---
Diffs of the changes:  (+21 -6)

Index: llvm/lib/CodeGen/TwoAddressInstructionPass.cpp
diff -u llvm/lib/CodeGen/TwoAddressInstructionPass.cpp:1.1 llvm/lib/CodeGen/TwoAddressInstructionPass.cpp:1.2
--- llvm/lib/CodeGen/TwoAddressInstructionPass.cpp:1.1	Thu Dec 18 07:06:03 2003
+++ llvm/lib/CodeGen/TwoAddressInstructionPass.cpp	Thu Dec 18 16:40:23 2003
@@ -16,7 +16,6 @@
 //===----------------------------------------------------------------------===//
 
 #define DEBUG_TYPE "twoaddrinstr"
-#include "llvm/CodeGen/TwoAddressInstructionPass.h"
 #include "llvm/Function.h"
 #include "llvm/CodeGen/LiveVariables.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
@@ -36,7 +35,23 @@
 using namespace llvm;
 
 namespace {
-    RegisterAnalysis X(
+    class TwoAddressInstructionPass : public MachineFunctionPass
+    {
+    private:
+        MachineFunction* mf_;
+        const TargetMachine* tm_;
+        const MRegisterInfo* mri_;
+        LiveVariables* lv_;
+
+    public:
+        virtual void getAnalysisUsage(AnalysisUsage &AU) const;
+
+    private:
+        /// runOnMachineFunction - pass entry point
+        bool runOnMachineFunction(MachineFunction&);
+    };
+
+    RegisterPass X(
         "twoaddressinstruction", "Two-Address instruction pass");
 
     Statistic<> numTwoAddressInstrs("twoaddressinstruction",
@@ -44,6 +59,8 @@
     Statistic<> numInstrsAdded("twoaddressinstruction",
                                "Number of instructions added");
 };
+
+const PassInfo *llvm::TwoAddressInstructionPassID = X.getPassInfo();
 
 void TwoAddressInstructionPass::getAnalysisUsage(AnalysisUsage &AU) const
 {


Index: llvm/lib/CodeGen/RegAllocLocal.cpp
diff -u llvm/lib/CodeGen/RegAllocLocal.cpp:1.35 llvm/lib/CodeGen/RegAllocLocal.cpp:1.36
--- llvm/lib/CodeGen/RegAllocLocal.cpp:1.35	Thu Dec 18 07:08:52 2003
+++ llvm/lib/CodeGen/RegAllocLocal.cpp	Thu Dec 18 16:40:23 2003
@@ -19,7 +19,6 @@
 #include "llvm/CodeGen/SSARegMap.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/LiveVariables.h"
-#include "llvm/CodeGen/TwoAddressInstructionPass.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
 #include "Support/CommandLine.h"
@@ -113,7 +112,7 @@
       if (!DisableKill)
 	AU.addRequired();
       AU.addRequiredID(PHIEliminationID);
-      AU.addRequired();
+      AU.addRequiredID(TwoAddressInstructionPassID);
       MachineFunctionPass::getAnalysisUsage(AU);
     }
 


Index: llvm/lib/CodeGen/LiveIntervals.cpp
diff -u llvm/lib/CodeGen/LiveIntervals.cpp:1.11 llvm/lib/CodeGen/LiveIntervals.cpp:1.12
--- llvm/lib/CodeGen/LiveIntervals.cpp:1.11	Thu Dec 18 07:15:02 2003
+++ llvm/lib/CodeGen/LiveIntervals.cpp	Thu Dec 18 16:40:24 2003
@@ -24,7 +24,6 @@
 #include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/CodeGen/SSARegMap.h"
-#include "llvm/CodeGen/TwoAddressInstructionPass.h"
 #include "llvm/Target/MRegisterInfo.h"
 #include "llvm/Target/TargetInstrInfo.h"
 #include "llvm/Target/TargetMachine.h"
@@ -50,7 +49,7 @@
     AU.addRequired();
     AU.addPreservedID(PHIEliminationID);
     AU.addRequiredID(PHIEliminationID);
-    AU.addRequired();
+    AU.addRequiredID(TwoAddressInstructionPassID);
     MachineFunctionPass::getAnalysisUsage(AU);
 }
 




From alkis at cs.uiuc.edu  Thu Dec 18 16:41:09 2003
From: alkis at cs.uiuc.edu (Alkis Evlogimenos)
Date: Thu Dec 18 16:41:09 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/CodeGen/Passes.h TwoAddressInstructionPass.h
Message-ID: <200312182240.QAA00293@morpheus.cs.uiuc.edu>


Changes in directory llvm/include/llvm/CodeGen:

Passes.h updated: 1.10 -> 1.11
TwoAddressInstructionPass.h (r1.1) removed

---
Log message:

Remove TwoAddressInstruction from the public headers and add an ID
instead, since this pass doesn't expose any state to its users.


---
Diffs of the changes:  (+5 -0)

Index: llvm/include/llvm/CodeGen/Passes.h
diff -u llvm/include/llvm/CodeGen/Passes.h:1.10 llvm/include/llvm/CodeGen/Passes.h:1.11
--- llvm/include/llvm/CodeGen/Passes.h:1.10	Wed Nov 19 21:32:25 2003
+++ llvm/include/llvm/CodeGen/Passes.h	Thu Dec 18 16:40:24 2003
@@ -28,6 +28,11 @@
 //
 extern const PassInfo *PHIEliminationID;
 
+// TwoAddressInstruction pass - This pass reduces two-address
+// instructions to use two operands. This destroys SSA information but
+// it is desired by register allocators.
+extern const PassInfo *TwoAddressInstructionPassID;
+
 /// Creates a register allocator as the user specified on the command
 /// line.
 FunctionPass *createRegisterAllocator();




From lattner at cs.uiuc.edu  Thu Dec 18 21:07:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:07:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/GenerateReport.pl
Message-ID: <200312190306.VAA11307@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

GenerateReport.pl updated: 1.15 -> 1.16

---
Log message:

Hack off repeated prefixes from the first column if a variable is set.
This allows us to get rid of the External/SPEC/CINT2000 prefix, for example.


---
Diffs of the changes:  (+35 -6)

Index: llvm/test/Programs/GenerateReport.pl
diff -u llvm/test/Programs/GenerateReport.pl:1.15 llvm/test/Programs/GenerateReport.pl:1.16
--- llvm/test/Programs/GenerateReport.pl:1.15	Sun Sep 21 11:56:55 2003
+++ llvm/test/Programs/GenerateReport.pl	Thu Dec 18 21:06:37 2003
@@ -42,6 +42,10 @@
 my $SortReverse = 0;
 my $SortNumeric = 0;   # Sort numerically or textually?
 
+# If the report wants us to trim repeated path prefixes off of the start of the
+# strings in the first column of the report, we can do that.
+my $TrimRepeatedPrefix = 0;
+
 # Helper functions which may be called by the report description files...
 sub SumCols {
   my ($Cols, $Col, $NumRows) = @_;
@@ -92,26 +96,27 @@
   my @RowValues;
   my $Col = 0;
   for $Row (@Fields) {
+    my $Val = "*";
     if (scalar(@$Row)) {            # An actual value to read?
       if (ref ($Row->[1])) {        # Code to be executed?
-        push @RowValues, &{$Row->[1]}(\@RowValues, $Col);
+        $Val = &{$Row->[1]}(\@RowValues, $Col);
       } else {                      # Field to be read...
         $Record =~ m/$Row->[1]/;
         if (!defined($1)) {
-          push @RowValues, "*";
+          $Val = "*";
         } else {
           # If there is a formatting function, run it now...
-          my $Val = $1;
+          $Val = $1;
           if (scalar(@$Row) > 2) {
             $Val = &{$Row->[2]}($Val);
           }
-
-          push @RowValues, $Val;
         }
       }
     } else {                        # Just add a seperator...
-      push @RowValues, "|";
+      $Val = "|";
     }
+
+    push @RowValues, $Val;
     $Col++;
   }
 
@@ -122,6 +127,30 @@
   }
   push @RowValues, $Assert if (!$HTML);
   push @Values, [@RowValues];
+}
+
+
+# If the report wants it, we can trim excess cruft off of the beginning of the
+# first column (which is often a path).
+if ($TrimRepeatedPrefix and scalar(@Values)) {
+  OuterLoop: while (1) {
+    # Figure out what the first path prefix is:
+    $Values[0]->[0] =~ m|([^/]+/).|;
+    last OuterLoop if (!defined($1));
+
+    # Now that we have the prefix, check to see if all of the entries in the
+    # table start with this prefix.
+    foreach $Row (@Values) {
+      last OuterLoop if ((substr $Row->[0], 0, length $1) ne $1);
+    }
+
+    # If we get here, then all of the entries have the prefix.  Remove it now.
+    foreach $Row (@Values) {
+      $Row->[0] = substr $Row->[0], length $1;
+    }
+
+    print "FOUND PREFIX: $1\n";
+  }
 }
 
 #




From lattner at cs.uiuc.edu  Thu Dec 18 21:08:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:08:02 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/TEST.nightly.report
Message-ID: <200312190307.VAA11318@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

TEST.nightly.report updated: 1.18 -> 1.19

---
Log message:

Hack off repeated prefixes from the first column if a variable is set.
This allows us to get rid of the External/SPEC/CINT2000 prefix, for example.


---
Diffs of the changes:  (+1 -0)

Index: llvm/test/Programs/TEST.nightly.report
diff -u llvm/test/Programs/TEST.nightly.report:1.18 llvm/test/Programs/TEST.nightly.report:1.19
--- llvm/test/Programs/TEST.nightly.report:1.18	Thu Sep 18 10:27:11 2003
+++ llvm/test/Programs/TEST.nightly.report	Thu Dec 18 21:06:58 2003
@@ -6,6 +6,7 @@
 
 # Sort by program name
 $SortCol = 0;
+$TrimRepeatedPrefix = 1;
 
 my $WallTimeRE = "[A-Za-z0-9.: ]+\\(([0-9.]+) wall clock";
 




From lattner at cs.uiuc.edu  Thu Dec 18 21:21:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:21:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/GenerateReport.pl
Message-ID: <200312190320.VAA12487@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

GenerateReport.pl updated: 1.16 -> 1.17

---
Log message:

Err, yeah, I have a clue, really.


---
Diffs of the changes:  (+0 -2)

Index: llvm/test/Programs/GenerateReport.pl
diff -u llvm/test/Programs/GenerateReport.pl:1.16 llvm/test/Programs/GenerateReport.pl:1.17
--- llvm/test/Programs/GenerateReport.pl:1.16	Thu Dec 18 21:06:37 2003
+++ llvm/test/Programs/GenerateReport.pl	Thu Dec 18 21:20:04 2003
@@ -148,8 +148,6 @@
     foreach $Row (@Values) {
       $Row->[0] = substr $Row->[0], length $1;
     }
-
-    print "FOUND PREFIX: $1\n";
   }
 }
 




From lattner at cs.uiuc.edu  Thu Dec 18 21:33:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:33:02 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/External/SPEC/Makefile.spec
Message-ID: <200312190332.VAA13939@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs/External/SPEC:

Makefile.spec updated: 1.20 -> 1.21

---
Log message:

Add llc-ls target for testing the linear scan allocator, for use by alkis


---
Diffs of the changes:  (+8 -0)

Index: llvm/test/Programs/External/SPEC/Makefile.spec
diff -u llvm/test/Programs/External/SPEC/Makefile.spec:1.20 llvm/test/Programs/External/SPEC/Makefile.spec:1.21
--- llvm/test/Programs/External/SPEC/Makefile.spec:1.20	Wed Dec 10 09:16:33 2003
+++ llvm/test/Programs/External/SPEC/Makefile.spec	Thu Dec 18 21:32:09 2003
@@ -94,6 +94,14 @@
 	-(cd Output/llc-$(RUN_TYPE); cat $(LOCAL_OUTPUTS)) > $@
 	-cp Output/llc-$(RUN_TYPE)/$(STDOUT_FILENAME).time $@.time
 
+$(PROGRAMS_TO_TEST:%=Output/%.out-llc-ls): \
+Output/%.out-llc-ls: Output/%.llc-ls
+	$(SPEC_SANDBOX) llc-ls-$(RUN_TYPE) $@ $(REF_IN_DIR) \
+             $(RUNSAFELY) $(STDIN_FILENAME) $(STDOUT_FILENAME) \
+                  ../../$< $(RUN_OPTIONS)
+	-(cd Output/llc-ls-$(RUN_TYPE); cat $(LOCAL_OUTPUTS)) > $@
+	-cp Output/llc-ls-$(RUN_TYPE)/$(STDOUT_FILENAME).time $@.time
+
 $(PROGRAMS_TO_TEST:%=Output/%.out-cbe): \
 Output/%.out-cbe: Output/%.cbe
 	$(SPEC_SANDBOX) cbe-$(RUN_TYPE) $@ $(REF_IN_DIR) \




From lattner at cs.uiuc.edu  Thu Dec 18 21:34:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:34:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/TEST.nightly.report
Message-ID: <200312190333.VAA14182@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

TEST.nightly.report updated: 1.19 -> 1.20

---
Log message:

Add a llc-ls column, which will only do stuff on the X86 tester, but oh well.


---
Diffs of the changes:  (+2 -1)

Index: llvm/test/Programs/TEST.nightly.report
diff -u llvm/test/Programs/TEST.nightly.report:1.19 llvm/test/Programs/TEST.nightly.report:1.20
--- llvm/test/Programs/TEST.nightly.report:1.19	Thu Dec 18 21:06:58 2003
+++ llvm/test/Programs/TEST.nightly.report	Thu Dec 18 21:33:36 2003
@@ -59,5 +59,6 @@
  ["LLC"      , 'TEST-RESULT-llc-time: real\s*([.0-9m:]+)', \&FormatTime],
  ["JIT"      , 'TEST-RESULT-jit-time: real\s*([.0-9m:]+)', \&FormatTime],
  ["GCC/LLC"  , \&GCCLLCRatio],
- ["GCC/CBE"  , \&GCCCBERatio]
+ ["GCC/CBE"  , \&GCCCBERatio],
+ ["LLC-LS"   , 'TEST-RESULT-llc-ls-time: real\s*([.0-9m:]+)', \&FormatTime]
 );




From lattner at cs.uiuc.edu  Thu Dec 18 21:34:09 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:34:09 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/Makefile.programs
Message-ID: <200312190333.VAA14102@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

Makefile.programs updated: 1.104 -> 1.105

---
Log message:

Add llc-ls target for testing the linear scan allocator, for use by alkis


---
Diffs of the changes:  (+29 -3)

Index: llvm/test/Programs/Makefile.programs
diff -u llvm/test/Programs/Makefile.programs:1.104 llvm/test/Programs/Makefile.programs:1.105
--- llvm/test/Programs/Makefile.programs:1.104	Wed Dec 10 09:08:59 2003
+++ llvm/test/Programs/Makefile.programs	Thu Dec 18 21:32:59 2003
@@ -85,8 +85,8 @@
 # Completed bytecode for a program
 BYTECODE   := $(addsuffix .llvm.bc, $(PREFIXED_PROGRAMS_TO_TEST))
 
-# Generated code for llc (which does not require the target platform)
 LLCCODEGEN := $(addsuffix .llc.s,   $(PREFIXED_PROGRAMS_TO_TEST))
+LLCLSCODEGEN := $(addsuffix .llc-ls.s,   $(PREFIXED_PROGRAMS_TO_TEST))
 CBECODEGEN := $(addsuffix .cbe,     $(PREFIXED_PROGRAMS_TO_TEST))
 
 # Output produced by programs run
@@ -95,21 +95,23 @@
 LLIOUTPUT := $(addsuffix .out-lli,  $(PREFIXED_PROGRAMS_TO_TEST))
 JITOUTPUT := $(addsuffix .out-jit,  $(PREFIXED_PROGRAMS_TO_TEST))
 LLCOUTPUT := $(addsuffix .out-llc,  $(PREFIXED_PROGRAMS_TO_TEST))
+LLCLSOUTPUT := $(addsuffix .out-llc-ls,  $(PREFIXED_PROGRAMS_TO_TEST))
 CBEOUTPUT := $(addsuffix .out-cbe,  $(PREFIXED_PROGRAMS_TO_TEST))
 
 # Diffs of program runs vs the native program
 LLIDIFFS  := $(addsuffix .diff-lli, $(PREFIXED_PROGRAMS_TO_TEST))
 JITDIFFS  := $(addsuffix .diff-jit, $(PREFIXED_PROGRAMS_TO_TEST))
 LLCDIFFS  := $(addsuffix .diff-llc, $(PREFIXED_PROGRAMS_TO_TEST))
+LLCLSDIFFS  := $(addsuffix .diff-llc-ls, $(PREFIXED_PROGRAMS_TO_TEST))
 CBEDIFFS  := $(addsuffix .diff-cbe, $(PREFIXED_PROGRAMS_TO_TEST))
 
 # Build Program outputs:
-.PRECIOUS: Output/%.out-lli Output/%.out-jit Output/%.out-llc
+.PRECIOUS: Output/%.out-lli Output/%.out-jit Output/%.out-llc Output/%.out-llc-ls
 .PRECIOUS: Output/%.out-nat Output/%.out-cbe
 
 # Build diffs from the output...
 .PRECIOUS: Output/%.diff-lli Output/%.diff-jit
-.PRECIOUS: Output/%.diff-llc Output/%.diff-cbe
+.PRECIOUS: Output/%.diff-llc Output/%.diff-llc-ls Output/%.diff-cbe
 
 
 # Regardless of what other options are specified, build the program's bytecode
@@ -176,6 +178,10 @@
 endif
 endif
 
+ifdef ENABLE_LINEARSCAN
+all:: $(LLCLSDIFFS)
+endif
+
 ifdef ENABLE_LLI
 all:: $(LLIDIFFS)
 endif
@@ -240,6 +246,10 @@
 Output/%.llc.s: Output/%.llvm.bc $(LLC)
 	-$(LLC) $(LLCFLAGS) -f $< -o $@
 
+$(PROGRAMS_TO_TEST:%=Output/%.llc-ls.s): \
+Output/%.llc-ls.s: Output/%.llvm.bc $(LLC)
+	-$(LLC) $(LLCFLAGS) -f -regalloc=linearscan $< -o $@
+
 
 ## Alternative command to run llc remotely on another machine:
 ## This is very useful when debugging:
@@ -251,6 +261,10 @@
 Output/%.llc: Output/%.llc.s
 	-$(CC) $(CFLAGS) $< $(LLCLIBS) $(LDFLAGS) -o $@
 
+$(PROGRAMS_TO_TEST:%=Output/%.llc-ls): \
+Output/%.llc-ls: Output/%.llc-ls.s
+	-$(CC) $(CFLAGS) $< $(LLCLIBS) $(LDFLAGS) -o $@
+
 
 #
 # Rules to execute the program
@@ -290,6 +304,10 @@
 Output/%.out-llc: Output/%.llc
 	-$(RUNSAFELY) $(STDIN_FILENAME) $@ $< $(RUN_OPTIONS)
 
+$(PROGRAMS_TO_TEST:%=Output/%.out-llc-ls): \
+Output/%.out-llc-ls: Output/%.llc-ls
+	-$(RUNSAFELY) $(STDIN_FILENAME) $@ $< $(RUN_OPTIONS)
+
 $(PROGRAMS_TO_TEST:%=Output/%.out-cbe): \
 Output/%.out-cbe: Output/%.cbe
 	-$(RUNSAFELY) $(STDIN_FILENAME) $@ $< $(RUN_OPTIONS)
@@ -340,6 +358,10 @@
 Output/%.diff-llc: Output/%.out-nat Output/%.out-llc
 	-$(DIFFPROG) llc $* $(HIDEDIFF)
 
+$(PROGRAMS_TO_TEST:%=Output/%.diff-llc-ls): \
+Output/%.diff-llc-ls: Output/%.out-nat Output/%.out-llc-ls
+	-$(DIFFPROG) llc-ls $* $(HIDEDIFF)
+
 $(PROGRAMS_TO_TEST:%=Output/%.diff-cbe): \
 Output/%.diff-cbe: Output/%.out-nat Output/%.out-cbe
 	-$(DIFFPROG) cbe $* $(HIDEDIFF)
@@ -355,6 +377,10 @@
 
 $(PROGRAMS_TO_TEST:%=Output/%.exe-llc): \
 Output/%.exe-llc: Output/%.diff-llc
+	-cp $< $@
+
+$(PROGRAMS_TO_TEST:%=Output/%.exe-llc-ls): \
+Output/%.exe-llc-ls: Output/%.diff-llc-ls
 	-cp $< $@
 
 $(PROGRAMS_TO_TEST:%=Output/%.exe-cbe): \




From lattner at cs.uiuc.edu  Thu Dec 18 21:35:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:35:01 2003
Subject: [llvm-commits] CVS: llvm/utils/NightlyTestTemplate.html
Message-ID: <200312190334.VAA14383@zion.cs.uiuc.edu>


Changes in directory llvm/utils:

NightlyTestTemplate.html updated: 1.22 -> 1.23

---
Log message:

Update template


---
Diffs of the changes:  (+3 -0)

Index: llvm/utils/NightlyTestTemplate.html
diff -u llvm/utils/NightlyTestTemplate.html:1.22 llvm/utils/NightlyTestTemplate.html:1.23
--- llvm/utils/NightlyTestTemplate.html:1.22	Sat Nov  8 14:48:39 2003
+++ llvm/utils/NightlyTestTemplate.html	Thu Dec 18 21:34:46 2003
@@ -199,6 +199,9 @@
     GCC output: greater than 1 is a speedup, less than 1 is a slowdown.
 GCC/CBE - The speed-up of the CBE output vs the native 
     GCC output: greater than 1 is a speedup, less than 1 is a slowdown.
+LLC-LS - How long does the program generated by the static
+    backend LLC take to execute the program, when compiled with the linear scan
+    register allocator.  This is temporary, for tuning.
 
 
 A complete log of testing 




From lattner at cs.uiuc.edu  Thu Dec 18 21:36:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:36:02 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/TEST.nightly.Makefile
Message-ID: <200312190335.VAA14682@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

TEST.nightly.Makefile updated: 1.25 -> 1.26

---
Log message:

add linear scan tests if ENABLE_LINEARSCAN is defined


---
Diffs of the changes:  (+24 -1)

Index: llvm/test/Programs/TEST.nightly.Makefile
diff -u llvm/test/Programs/TEST.nightly.Makefile:1.25 llvm/test/Programs/TEST.nightly.Makefile:1.26
--- llvm/test/Programs/TEST.nightly.Makefile:1.25	Sun Aug 17 23:08:47 2003
+++ llvm/test/Programs/TEST.nightly.Makefile	Thu Dec 18 21:35:27 2003
@@ -1,4 +1,4 @@
-##===- test/Programs/TEST.nightly.Makefile --*- Makefile -*-=======- vim:ft=make
+##===- test/Programs/TEST.nightly.Makefile --*- Makefile -*- vim:ft=make --===##
 #
 # This test is used in conjunction with the llvm/utils/NightlyTest* stuff to
 # generate information about program status for the nightly report.
@@ -11,8 +11,12 @@
 CFLAGS  := -O3
 
 REPORTS_TO_GEN := compile nat llc cbe jit
+ifdef ENABLE_LINEARSCAN
+REPORTS_TO_GEN += llc-ls
+endif
 REPORTS_SUFFIX := $(addsuffix .report.txt, $(REPORTS_TO_GEN))
 
+
 TIMEOPT = -time-passes -stats -info-output-file=$(CURDIR)/$@.info
 EXTRA_LLI_OPTS = $(TIMEOPT)
 
@@ -61,6 +65,25 @@
 	else  \
 	  echo "TEST-FAIL: llc $(RELDIR)/$*" >> $@;\
 	fi
+
+# LLC-linearscan tests
+$(PROGRAMS_TO_TEST:%=Output/%.nightly.llc-ls.report.txt): \
+Output/%.nightly.llc-ls.report.txt: Output/%.llvm.bc Output/%.exe-llc-ls $(LLC)
+	@echo > $@
+	-head -n 100 Output/$*.exe-llc-ls >> $@
+	@-if test -f Output/$*.exe-llc-ls; then \
+	  echo "TEST-PASS: llc-ls $(RELDIR)/$*" >> $@;\
+	  $(LLC) $< -o /dev/null -f $(TIMEOPT) >> $@ 2>&1; \
+	  printf "TEST-RESULT-llc-ls: " >> $@;\
+	  grep "Total Execution Time" $@.info >> $@;\
+	  printf "TEST-RESULT-llc-ls-time: " >> $@;\
+	  grep "^real" Output/$*.out-llc-ls.time >> $@;\
+	  echo >> $@;\
+	else  \
+	  echo "TEST-FAIL: llc-ls $(RELDIR)/$*" >> $@;\
+	fi
+
+
 
 # CBE tests
 $(PROGRAMS_TO_TEST:%=Output/%.nightly.cbe.report.txt): \




From lattner at cs.uiuc.edu  Thu Dec 18 21:48:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:48:01 2003
Subject: [llvm-commits] CVS: llvm/utils/NightlyTest.pl
Message-ID: <200312190347.VAA15301@zion.cs.uiuc.edu>


Changes in directory llvm/utils:

NightlyTest.pl updated: 1.40 -> 1.41

---
Log message:

If -enable-linscan is specified, add ENABLE_LINEARSCAN=1 to the make commandline


---
Diffs of the changes:  (+6 -3)

Index: llvm/utils/NightlyTest.pl
diff -u llvm/utils/NightlyTest.pl:1.40 llvm/utils/NightlyTest.pl:1.41
--- llvm/utils/NightlyTest.pl:1.40	Sun Nov 30 23:31:12 2003
+++ llvm/utils/NightlyTest.pl	Thu Dec 18 21:47:31 2003
@@ -18,6 +18,8 @@
 #  -norunningtests  Do not run the Olden benchmark suite with
 #                   LARGE_PROBLEM_SIZE enabled.
 #  -parallel        Run two parallel jobs with GNU Make.
+#  -enable-linscan  Enable linearscan tests
+#
 # CVSROOT is the CVS repository from which the tree will be checked out,
 #  specified either in the full :method:user at host:/dir syntax, or
 #  just /dir if using a local repo.
@@ -117,7 +119,7 @@
 my $NOTEST     = 0;
 my $NORUNNINGTESTS = 0;
 my $MAKEOPTS   = "";
-
+my $ENABLELINEARSCAN = "";
 
 # Parse arguments...
 while (scalar(@ARGV) and ($_ = $ARGV[0], /^[-+]/)) {
@@ -130,6 +132,7 @@
   if (/^-notest$/)         { $NOTEST     = 1; $NORUNNINGTESTS = 1; next; }
   if (/^-norunningtests$/) { $NORUNNINGTESTS = 1; next; }
   if (/^-parallel$/)       { $MAKEOPTS   = "-j2 -l3.0"; next; }
+  if (/^-enable-linscan$/) { $ENABLELINEARSCAN = "ENABLE_LINEARSCAN=1"; }
 
   print "Unknown option: $_ : ignoring!\n";
 }
@@ -323,8 +326,8 @@
 
   # Run the programs tests... creating a report.nightly.html file
   if (!$NOTEST) {
-    system "gmake -k $MAKEOPTS report.nightly.html TEST=nightly "
-         . "> $Prefix-$SubDir-ProgramTest.txt 2>&1";
+    system "gmake -k $MAKEOPTS $ENABLELINEARSCAN report.nightly.html "
+         . "TEST=nightly > $Prefix-$SubDir-ProgramTest.txt 2>&1";
   } else {
     system "gunzip $Prefix-$SubDir-ProgramTest.txt.gz";
   }




From lattner at cs.uiuc.edu  Thu Dec 18 21:49:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 21:49:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/TEST.jit.report
Message-ID: <200312190348.VAA15322@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs:

TEST.jit.report updated: 1.2 -> 1.3

---
Log message:

Cleanup report a bit


---
Diffs of the changes:  (+1 -0)

Index: llvm/test/Programs/TEST.jit.report
diff -u llvm/test/Programs/TEST.jit.report:1.2 llvm/test/Programs/TEST.jit.report:1.3
--- llvm/test/Programs/TEST.jit.report:1.2	Mon Jun 16 10:52:02 2003
+++ llvm/test/Programs/TEST.jit.report	Thu Dec 18 21:48:48 2003
@@ -6,6 +6,7 @@
 
 # Sort by total execution time
 $SortCol = 2;
+$TrimRepeatedPrefix = 1;
 
 # Sort in descending order
 $SortReverse = 1;




From lattner at cs.uiuc.edu  Thu Dec 18 23:15:04 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 23:15:04 2003
Subject: [llvm-commits] CVS: llvm-www/Publications.html
Message-ID: <200312190514.XAA27886@zion.cs.uiuc.edu>


Changes in directory llvm-www:

Publications.html updated: 1.1 -> 1.2

---
Log message:

Note that this paper will be publicshed in CGO


---
Diffs of the changes:  (+1 -1)

Index: llvm-www/Publications.html
diff -u llvm-www/Publications.html:1.1 llvm-www/Publications.html:1.2
--- llvm-www/Publications.html:1.1	Tue Nov 18 17:04:53 2003
+++ llvm-www/Publications.html	Thu Dec 18 23:13:58 2003
@@ -16,7 +16,7 @@
                         Analysis & Transformation",
                         Chris Lattner & Vikram Adve.

                         Technical Report #UIUCDCS-R-2003-2380, Computer Science
-                        Dept., Univ. of Illinois, Sep. 2003.
+                        Dept., Univ. of Illinois, Sep. 2003.  (to appear in CGO'04)
                     
"
                         Memory Safety Without Runtime Checks or Garbage
                         Collection"





From lattner at cs.uiuc.edu  Thu Dec 18 23:17:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 23:17:01 2003
Subject: [llvm-commits] CVS: llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html
Message-ID: <200312190516.XAA28141@zion.cs.uiuc.edu>


Changes in directory llvm-www/pubs:

2003-09-30-LifelongOptimizationTR.html updated: 1.3 -> 1.4

---
Log message:

add note


---
Diffs of the changes:  (+2 -0)

Index: llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html
diff -u llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html:1.3 llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html:1.4
--- llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html:1.3	Tue Oct 21 19:26:00 2003
+++ llvm-www/pubs/2003-09-30-LifelongOptimizationTR.html	Thu Dec 18 23:15:51 2003
@@ -49,6 +49,8 @@
   Illinois, Sep. 2003.
 
 
+This paper is the submitted version of a paper to appear in CGO'04.
+
 Download:
 
   LLVM: A Compilation




From lattner at cs.uiuc.edu  Thu Dec 18 23:58:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 23:58:02 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Utils/Local.cpp
Message-ID: <200312190556.XAA29362@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Utils:

Local.cpp updated: 1.12 -> 1.13

---
Log message:

Add new function


---
Diffs of the changes:  (+41 -10)

Index: llvm/lib/Transforms/Utils/Local.cpp
diff -u llvm/lib/Transforms/Utils/Local.cpp:1.12 llvm/lib/Transforms/Utils/Local.cpp:1.13
--- llvm/lib/Transforms/Utils/Local.cpp:1.12	Tue Nov 11 16:41:34 2003
+++ llvm/lib/Transforms/Utils/Local.cpp	Thu Dec 18 23:56:28 2003
@@ -15,18 +15,18 @@
 #include "llvm/Transforms/Utils/Local.h"
 #include "llvm/iTerminators.h"
 #include "llvm/iOperators.h"
+#include "llvm/iPHINode.h"
 #include "llvm/ConstantHandling.h"
-
-namespace llvm {
+using namespace llvm;
 
 //===----------------------------------------------------------------------===//
 //  Local constant propagation...
 //
 
-// ConstantFoldInstruction - If an instruction references constants, try to fold
-// them together...
-//
-bool doConstantPropagation(BasicBlock::iterator &II) {
+/// doConstantPropagation - If an instruction references constants, try to fold
+/// them together...
+///
+bool llvm::doConstantPropagation(BasicBlock::iterator &II) {
   if (Constant *C = ConstantFoldInstruction(II)) {
     // Replaces all of the uses of a variable with uses of the constant.
     II->replaceAllUsesWith(C);
@@ -43,7 +43,7 @@
 // constant value, convert it into an unconditional branch to the constant
 // destination.
 //
-bool ConstantFoldTerminator(BasicBlock *BB) {
+bool llvm::ConstantFoldTerminator(BasicBlock *BB) {
   TerminatorInst *T = BB->getTerminator();
       
   // Branch - See if we are conditional jumping on constant
@@ -165,7 +165,7 @@
 //  Local dead code elimination...
 //
 
-bool isInstructionTriviallyDead(Instruction *I) {
+bool llvm::isInstructionTriviallyDead(Instruction *I) {
   return I->use_empty() && !I->mayWriteToMemory() && !isa(I);
 }
 
@@ -174,7 +174,7 @@
 // to point to the instruction that immediately succeeded the original
 // instruction.
 //
-bool dceInstruction(BasicBlock::iterator &BBI) {
+bool llvm::dceInstruction(BasicBlock::iterator &BBI) {
   // Look for un"used" definitions...
   if (isInstructionTriviallyDead(BBI)) {
     BBI = BBI->getParent()->getInstList().erase(BBI);   // Bye bye
@@ -183,4 +183,35 @@
   return false;
 }
 
-} // End llvm namespace
+//===----------------------------------------------------------------------===//
+//  PHI Instruction Simplification
+//
+
+/// hasConstantValue - If the specified PHI node always merges together the same
+/// value, return the value, otherwise return null.
+///
+Value *llvm::hasConstantValue(PHINode *PN) {
+  // If the PHI node only has one incoming value, eliminate the PHI node...
+  if (PN->getNumIncomingValues() == 1)
+    return PN->getIncomingValue(0);
+
+  // Otherwise if all of the incoming values are the same for the PHI, replace
+  // the PHI node with the incoming value.
+  //
+  Value *InVal = 0;
+  for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
+    if (PN->getIncomingValue(i) != PN)  // Not the PHI node itself...
+      if (InVal && PN->getIncomingValue(i) != InVal)
+        return 0;  // Not the same, bail out.
+      else
+        InVal = PN->getIncomingValue(i);
+
+  // The only case that could cause InVal to be null is if we have a PHI node
+  // that only has entries for itself.  In this case, there is no entry into the
+  // loop, so kill the PHI.
+  //
+  if (InVal == 0) InVal = Constant::getNullValue(PN->getType());
+
+  // All of the incoming values are the same, return the value now.
+  return InVal;
+}




From lattner at cs.uiuc.edu  Thu Dec 18 23:58:11 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 23:58:11 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/Transforms/Utils/Local.h
Message-ID: <200312190556.XAA29040@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/Transforms/Utils:

Local.h updated: 1.12 -> 1.13

---
Log message:

Add new function


---
Diffs of the changes:  (+10 -0)

Index: llvm/include/llvm/Transforms/Utils/Local.h
diff -u llvm/include/llvm/Transforms/Utils/Local.h:1.12 llvm/include/llvm/Transforms/Utils/Local.h:1.13
--- llvm/include/llvm/Transforms/Utils/Local.h:1.12	Tue Nov 11 16:41:31 2003
+++ llvm/include/llvm/Transforms/Utils/Local.h	Thu Dec 18 23:56:20 2003
@@ -20,6 +20,7 @@
 namespace llvm {
 
 class Pass;
+class PHINode;
 
 //===----------------------------------------------------------------------===//
 //  Local constant propagation...
@@ -54,6 +55,15 @@
 /// instruction.
 ///
 bool dceInstruction(BasicBlock::iterator &BBI);
+
+//===----------------------------------------------------------------------===//
+//  PHI Instruction Simplification
+//
+
+/// hasConstantValue - If the specified PHI node always merges together the same
+/// value, return the value, otherwise return null.
+///
+Value *hasConstantValue(PHINode *PN);
 
 
 //===----------------------------------------------------------------------===//




From lattner at cs.uiuc.edu  Thu Dec 18 23:59:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Thu Dec 18 23:59:02 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/InstructionCombining.cpp
Message-ID: <200312190558.XAA30260@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

InstructionCombining.cpp updated: 1.144 -> 1.145

---
Log message:

Factor code out into the Utils library


---
Diffs of the changes:  (+3 -23)

Index: llvm/lib/Transforms/Scalar/InstructionCombining.cpp
diff -u llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.144 llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.145
--- llvm/lib/Transforms/Scalar/InstructionCombining.cpp:1.144	Sat Dec  6 19:24:23 2003
+++ llvm/lib/Transforms/Scalar/InstructionCombining.cpp	Thu Dec 18 23:58:40 2003
@@ -1906,29 +1906,9 @@
 // PHINode simplification
 //
 Instruction *InstCombiner::visitPHINode(PHINode &PN) {
-  // If the PHI node only has one incoming value, eliminate the PHI node...
-  if (PN.getNumIncomingValues() == 1)
-    return ReplaceInstUsesWith(PN, PN.getIncomingValue(0));
-  
-  // Otherwise if all of the incoming values are the same for the PHI, replace
-  // the PHI node with the incoming value.
-  //
-  Value *InVal = 0;
-  for (unsigned i = 0, e = PN.getNumIncomingValues(); i != e; ++i)
-    if (PN.getIncomingValue(i) != &PN)  // Not the PHI node itself...
-      if (InVal && PN.getIncomingValue(i) != InVal)
-        return 0;  // Not the same, bail out.
-      else
-        InVal = PN.getIncomingValue(i);
-
-  // The only case that could cause InVal to be null is if we have a PHI node
-  // that only has entries for itself.  In this case, there is no entry into the
-  // loop, so kill the PHI.
-  //
-  if (InVal == 0) InVal = Constant::getNullValue(PN.getType());
-
-  // All of the incoming values are the same, replace the PHI node now.
-  return ReplaceInstUsesWith(PN, InVal);
+  if (Value *V = hasConstantValue(&PN))
+    return ReplaceInstUsesWith(PN, V);
+  return 0;
 }
 
 




From lattner at cs.uiuc.edu  Fri Dec 19 00:28:03 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 00:28:03 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/LoopSimplify.cpp
Message-ID: <200312190627.AAA03541@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

LoopSimplify.cpp updated: 1.28 -> 1.29

---
Log message:

Generalize a special case to fix PR187


---
Diffs of the changes:  (+28 -18)

Index: llvm/lib/Transforms/Scalar/LoopSimplify.cpp
diff -u llvm/lib/Transforms/Scalar/LoopSimplify.cpp:1.28 llvm/lib/Transforms/Scalar/LoopSimplify.cpp:1.29
--- llvm/lib/Transforms/Scalar/LoopSimplify.cpp:1.28	Wed Dec 10 11:20:35 2003
+++ llvm/lib/Transforms/Scalar/LoopSimplify.cpp	Fri Dec 19 00:27:08 2003
@@ -33,12 +33,12 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Transforms/Scalar.h"
-#include "llvm/Analysis/Dominators.h"
-#include "llvm/Analysis/LoopInfo.h"
 #include "llvm/Function.h"
 #include "llvm/iTerminators.h"
 #include "llvm/iPHINode.h"
 #include "llvm/Constant.h"
+#include "llvm/Analysis/Dominators.h"
+#include "llvm/Analysis/LoopInfo.h"
 #include "llvm/Support/CFG.h"
 #include "Support/SetOperations.h"
 #include "Support/Statistic.h"
@@ -164,18 +164,22 @@
   // incoming edges in BB into new PHI nodes in NewBB.
   //
   if (!Preds.empty()) {  // Is the loop not obviously dead?
-    if (Preds.size() == 1) {
-      // No need to insert one operand PHI nodes!  Instead, just update the
-      // incoming block ID's.
-      for (BasicBlock::iterator I = BB->begin();
-           PHINode *PN = dyn_cast(I); ++I) {
-        unsigned i = PN->getBasicBlockIndex(Preds[0]);
-        PN->setIncomingBlock(i, NewBB);
-      }
-    } else {
-      for (BasicBlock::iterator I = BB->begin();
-           PHINode *PN = dyn_cast(I); ++I) {
-        
+    // Check to see if the values being merged into the new block need PHI
+    // nodes.  If so, insert them.
+    for (BasicBlock::iterator I = BB->begin();
+         PHINode *PN = dyn_cast(I); ++I) {
+      
+      // Check to see if all of the values coming in are the same.  If so, we
+      // don't need to create a new PHI node.
+      Value *InVal = PN->getIncomingValueForBlock(Preds[0]);
+      for (unsigned i = 1, e = Preds.size(); i != e; ++i)
+        if (InVal != PN->getIncomingValueForBlock(Preds[i])) {
+          InVal = 0;
+          break;
+        }
+      
+      // If the values coming into the block are not the same, we need a PHI.
+      if (InVal == 0) {
         // Create the new PHI node, insert it into NewBB at the end of the block
         PHINode *NewPHI = new PHINode(PN->getType(), PN->getName()+".ph", BI);
         
@@ -184,11 +188,17 @@
           Value *V = PN->removeIncomingValue(Preds[i]);
           NewPHI->addIncoming(V, Preds[i]);
         }
-        
-        // Add an incoming value to the PHI node in the loop for the preheader
-        // edge.
-        PN->addIncoming(NewPHI, NewBB);
+        InVal = NewPHI;
+      } else {
+        // Remove all of the edges coming into the PHI nodes from outside of the
+        // block.
+        for (unsigned i = 0, e = Preds.size(); i != e; ++i)
+          PN->removeIncomingValue(Preds[i], false);
       }
+
+      // Add an incoming value to the PHI node in the loop for the preheader
+      // edge.
+      PN->addIncoming(InVal, NewBB);
     }
     
     // Now that the PHI nodes are updated, actually move the edges from




From lattner at cs.uiuc.edu  Fri Dec 19 00:33:05 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 00:33:05 2003
Subject: [llvm-commits] CVS: llvm/docs/ReleaseNotes.html
Message-ID: <200312190632.AAA04240@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

ReleaseNotes.html updated: 1.88 -> 1.89

---
Log message:

Turn QOI section into CQ section (QUI can be added later if needed)
Add PR187


---
Diffs of the changes:  (+3 -4)

Index: llvm/docs/ReleaseNotes.html
diff -u llvm/docs/ReleaseNotes.html:1.88 llvm/docs/ReleaseNotes.html:1.89
--- llvm/docs/ReleaseNotes.html:1.88	Thu Dec 18 10:43:14 2003
+++ llvm/docs/ReleaseNotes.html	Fri Dec 19 00:31:56 2003
@@ -113,13 +113,12 @@
 
 
 
-In this release, the following Quality of Implementation issues were
-fixed:
+In this release, the following Code Quality  issues were fixed:
 
 
 
 
-
+[loopsimplify] Many pointless phi nodes are created
 
 
 
@@ -582,7 +581,7 @@
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" />
 
   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/18 16:43:14 $
+  Last modified: $Date: 2003/12/19 06:31:56 $
 
 
 




From brukman at cs.uiuc.edu  Fri Dec 19 00:34:01 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Fri Dec 19 00:34:01 2003
Subject: [llvm-commits] CVS: llvm-www/header.incl
Message-ID: <200312190633.AAA09160@zion.cs.uiuc.edu>


Changes in directory llvm-www:

header.incl updated: 1.9 -> 1.10

---
Log message:

Small optimization: since "/doxygen" becomes a redirect to "/doxygen/", there
are 2 hits to the webserver for every access (so says w3c link checker). So,
let's just send the browser there directly.


---
Diffs of the changes:  (+1 -1)

Index: llvm-www/header.incl
diff -u llvm-www/header.incl:1.9 llvm-www/header.incl:1.10
--- llvm-www/header.incl:1.9	Wed Dec 17 21:10:11 2003
+++ llvm-www/header.incl	Fri Dec 19 00:33:01 2003
@@ -72,7 +72,7 @@
   


 
   Dev. Resources:

-  doxygen
+  doxygen
   cvsweb

   Nightly Tester
 




From brukman at cs.uiuc.edu  Fri Dec 19 00:43:03 2003
From: brukman at cs.uiuc.edu (Misha Brukman)
Date: Fri Dec 19 00:43:03 2003
Subject: [llvm-commits] CVS: llvm-www/releases/index.html
Message-ID: <200312190642.AAA14179@zion.cs.uiuc.edu>


Changes in directory llvm-www/releases:

index.html updated: 1.11 -> 1.12

---
Log message:

The license link should go to release 1.1, not 1.0 by now.


---
Diffs of the changes:  (+1 -1)

Index: llvm-www/releases/index.html
diff -u llvm-www/releases/index.html:1.11 llvm-www/releases/index.html:1.12
--- llvm-www/releases/index.html:1.11	Wed Dec 17 16:48:50 2003
+++ llvm-www/releases/index.html	Fri Dec 19 00:42:07 2003
@@ -28,7 +28,7 @@
 
 
 
-LLVM is distributed under the University of Illinois
+
LLVM is distributed under the University of Illinois
 Open Source License, an OSI-approved
 license.




From lattner at cs.uiuc.edu  Fri Dec 19 00:55:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 00:55:01 2003
Subject: [llvm-commits] CVS: llvm/test/Regression/Transforms/LICM/sink_multiple.ll
Message-ID: <200312190654.AAA14374@zion.cs.uiuc.edu>


Changes in directory llvm/test/Regression/Transforms/LICM:

sink_multiple.ll added (r1.1)

---
Log message:

New testcase


---
Diffs of the changes:  (+19 -0)

Index: llvm/test/Regression/Transforms/LICM/sink_multiple.ll
diff -c /dev/null llvm/test/Regression/Transforms/LICM/sink_multiple.ll:1.1
*** /dev/null	Fri Dec 19 00:54:47 2003
--- llvm/test/Regression/Transforms/LICM/sink_multiple.ll	Fri Dec 19 00:54:37 2003
***************
*** 0 ****
--- 1,19 ----
+ ; The loop sinker was running from the bottom of the loop to the top, causing
+ ; it to miss opportunities to sink instructions that depended on sinking other
+ ; instructions from the loop.  Instead they got hoisted, which is better than
+ ; leaving them in the loop, but increases register pressure pointlessly.
+ 
+ ; RUN: llvm-as < %s | opt -licm | llvm-dis | grep -C1 getelementptr | grep Out:
+ 
+ %Ty = type { int, int }
+ %X = external global %Ty
+ 
+ int %test() {
+         br label %Loop
+ Loop:
+         %dead = getelementptr %Ty* %X, long 0, ubyte 0
+         %sunk2 = load int* %dead
+         br bool false, label %Loop, label %Out
+ Out:
+         ret int %sunk2
+ }




From lattner at cs.uiuc.edu  Fri Dec 19 01:23:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 01:23:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/LICM.cpp
Message-ID: <200312190722.BAA19206@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

LICM.cpp updated: 1.52 -> 1.53

---
Log message:

Implement LICM/sink_multiple.ll, by sinking all possible instructions in the 
loop before hoisting any.


---
Diffs of the changes:  (+55 -20)

Index: llvm/lib/Transforms/Scalar/LICM.cpp
diff -u llvm/lib/Transforms/Scalar/LICM.cpp:1.52 llvm/lib/Transforms/Scalar/LICM.cpp:1.53
--- llvm/lib/Transforms/Scalar/LICM.cpp:1.52	Thu Dec 18 02:12:32 2003
+++ llvm/lib/Transforms/Scalar/LICM.cpp	Fri Dec 19 01:22:45 2003
@@ -11,7 +11,7 @@
 // code from the body of a loop as possible.  It does this by either hoisting
 // code into the preheader block, or by sinking code to the exit blocks if it is
 // safe.  This pass also promotes must-aliased memory locations in the loop to
-// live in registers.
+// live in registers, thus hoisting and sinking "invariant" loads and stores.
 //
 // This pass uses alias analysis for two purposes:
 //
@@ -92,6 +92,14 @@
     ///
     void visitLoop(Loop *L, AliasSetTracker &AST);
 
+    /// SinkRegion - Walk the specified region of the CFG (defined by all blocks
+    /// dominated by the specified block, and that are in the current loop) in
+    /// reverse depth first order w.r.t the DominatorTree.  This allows us to
+    /// visit uses before definitions, allowing us to sink a loop body in one
+    /// pass without iteration.
+    ///
+    void SinkRegion(DominatorTree::Node *N);
+
     /// HoistRegion - Walk the specified region of the CFG (defined by all
     /// blocks dominated by the specified block, and that are in the current
     /// loop) in depth first order w.r.t the DominatorTree.  This allows us to
@@ -258,8 +266,10 @@
   //
   // Traverse the body of the loop in depth first order on the dominator tree so
   // that we are guaranteed to see definitions before we see uses.  This allows
-  // us to perform the LICM transformation in one pass, without iteration.
+  // us to sink instructions in one pass, without iteration.  AFter sinking
+  // instructions, we perform another pass to hoist them out of the loop.
   //
+  SinkRegion(DT->getNode(L->getHeader()));
   HoistRegion(DT->getNode(L->getHeader()));
 
   // Now that all loop invariants have been removed from the loop, promote any
@@ -272,6 +282,42 @@
   Preheader = 0;
 }
 
+/// SinkRegion - Walk the specified region of the CFG (defined by all blocks
+/// dominated by the specified block, and that are in the current loop) in
+/// reverse depth first order w.r.t the DominatorTree.  This allows us to visit
+/// uses before definitions, allowing us to sink a loop body in one pass without
+/// iteration.
+///
+void LICM::SinkRegion(DominatorTree::Node *N) {
+  assert(N != 0 && "Null dominator tree node?");
+  BasicBlock *BB = N->getBlock();
+
+  // If this subregion is not in the top level loop at all, exit.
+  if (!CurLoop->contains(BB)) return;
+
+  // We are processing blocks in reverse dfo, so process children first...
+  const std::vector &Children = N->getChildren();
+  for (unsigned i = 0, e = Children.size(); i != e; ++i)
+    SinkRegion(Children[i]);
+
+  // Only need to process the contents of this block if it is not part of a
+  // subloop (which would already have been processed).
+  if (inSubLoop(BB)) return;
+
+  for (BasicBlock::iterator II = BB->begin(), E = BB->end(); II != E; ) {
+    Instruction &I = *II++;
+    
+    // Check to see if we can sink this instruction to the exit blocks
+    // of the loop.  We can do this if the all users of the instruction are
+    // outside of the loop.  In this case, it doesn't even matter if the
+    // operands of the instruction are loop invariant.
+    //
+    if (canSinkOrHoistInst(I) && isNotUsedInLoop(I))
+      sink(I);
+  }
+}
+
+
 /// HoistRegion - Walk the specified region of the CFG (defined by all blocks
 /// dominated by the specified block, and that are in the current loop) in depth
 /// first order w.r.t the DominatorTree.  This allows us to visit definitions
@@ -289,26 +335,15 @@
   if (!inSubLoop(BB))
     for (BasicBlock::iterator II = BB->begin(), E = BB->end(); II != E; ) {
       Instruction &I = *II++;
-
-      // We can only handle simple expressions and loads with this code.
-      if (canSinkOrHoistInst(I)) {
-        // First check to see if we can sink this instruction to the exit blocks
-        // of the loop.  We can do this if the only users of the instruction are
-        // outside of the loop.  In this case, it doesn't even matter if the
-        // operands of the instruction are loop invariant.
-        //
-        if (isNotUsedInLoop(I))
-          sink(I);
-        
-        // If we can't sink the instruction, try hoisting it out to the
-        // preheader.  We can only do this if all of the operands of the
-        // instruction are loop invariant and if it is safe to hoist the
-        // instruction.
-        //
-        else if (isLoopInvariantInst(I) && isSafeToExecuteUnconditionally(I))
+      
+      // Try hoisting the instruction out to the preheader.  We can only do this
+      // if all of the operands of the instruction are loop invariant and if it
+      // is safe to hoist the instruction.
+      //
+      if (isLoopInvariantInst(I) && canSinkOrHoistInst(I) && 
+          isSafeToExecuteUnconditionally(I))
           hoist(I);
       }
-    }
 
   const std::vector &Children = N->getChildren();
   for (unsigned i = 0, e = Children.size(); i != e; ++i)




From lattner at cs.uiuc.edu  Fri Dec 19 01:52:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 01:52:01 2003
Subject: [llvm-commits] CVS: llvm/runtime/GCCLibraries/crtend/crtend.c
Message-ID: <200312190751.BAA21747@zion.cs.uiuc.edu>


Changes in directory llvm/runtime/GCCLibraries/crtend:

crtend.c updated: 1.2 -> 1.3

---
Log message:

Add support for people calling main recursively


---
Diffs of the changes:  (+7 -0)

Index: llvm/runtime/GCCLibraries/crtend/crtend.c
diff -u llvm/runtime/GCCLibraries/crtend/crtend.c:1.2 llvm/runtime/GCCLibraries/crtend/crtend.c:1.3
--- llvm/runtime/GCCLibraries/crtend/crtend.c:1.2	Wed Jun 25 23:20:38 2003
+++ llvm/runtime/GCCLibraries/crtend/crtend.c	Fri Dec 19 01:51:46 2003
@@ -34,6 +34,13 @@
   /* Loop over all of the constructor records, calling each function pointer. */
   TorRec *R = __llvm_getGlobalCtors();
 
+  /* Recursively calling main is not legal C, but lots of people do it for
+   * testing stuff.  We might as well work for them.
+   */
+  static _Bool Initialized = 0;
+  if (Initialized) return;
+  Initialized = 1;
+
   /* Only register the global dtor handler if there is at least one global
    * dtor!
    */




From lattner at cs.uiuc.edu  Fri Dec 19 02:19:00 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 02:19:00 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/LICM.cpp
Message-ID: <200312190818.CAA00882@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

LICM.cpp updated: 1.53 -> 1.54

---
Log message:

Remove the wierd "Operands" loop, by traversing basicblocks in reverse order


---
Diffs of the changes:  (+14 -14)

Index: llvm/lib/Transforms/Scalar/LICM.cpp
diff -u llvm/lib/Transforms/Scalar/LICM.cpp:1.53 llvm/lib/Transforms/Scalar/LICM.cpp:1.54
--- llvm/lib/Transforms/Scalar/LICM.cpp:1.53	Fri Dec 19 01:22:45 2003
+++ llvm/lib/Transforms/Scalar/LICM.cpp	Fri Dec 19 02:18:16 2003
@@ -304,16 +304,18 @@
   // subloop (which would already have been processed).
   if (inSubLoop(BB)) return;
 
-  for (BasicBlock::iterator II = BB->begin(), E = BB->end(); II != E; ) {
-    Instruction &I = *II++;
+  for (BasicBlock::iterator II = BB->end(); II != BB->begin(); ) {
+    Instruction &I = *--II;
     
     // Check to see if we can sink this instruction to the exit blocks
     // of the loop.  We can do this if the all users of the instruction are
     // outside of the loop.  In this case, it doesn't even matter if the
     // operands of the instruction are loop invariant.
     //
-    if (canSinkOrHoistInst(I) && isNotUsedInLoop(I))
+    if (canSinkOrHoistInst(I) && isNotUsedInLoop(I)) {
+      ++II;
       sink(I);
+    }
   }
 }
 
@@ -403,14 +405,14 @@
 }
 
 /// sink - When an instruction is found to only be used outside of the loop,
-/// this function moves it to the exit blocks and patches up SSA form as
-/// needed.
+/// this function moves it to the exit blocks and patches up SSA form as needed.
+/// This method is guaranteed to remove the original instruction from its
+/// position, and may either delete it or move it to outside of the loop.
 ///
 void LICM::sink(Instruction &I) {
   DEBUG(std::cerr << "LICM sinking instruction: " << I);
 
   const std::vector &ExitBlocks = CurLoop->getExitBlocks();
-  std::vector Operands(I.op_begin(), I.op_end());
 
   if (isa(I)) ++NumMovedLoads;
   ++NumSunk;
@@ -512,20 +514,18 @@
         }
       }
     }
+
+    // If the instruction doesn't dominate any exit blocks, it must be dead.
+    if (InsertedBlocks.empty()) {
+      CurAST->remove(&I);
+      I.getParent()->getInstList().erase(&I);
+    }
       
     // Finally, promote the fine value to SSA form.
     std::vector Allocas;
     Allocas.push_back(AI);
     PromoteMemToReg(Allocas, *DT, *DF, AA->getTargetData());
   }
-  
-  // Since we just sunk an instruction, check to see if any other instructions
-  // used by this instruction are now sinkable.  If so, sink them too.
-  for (unsigned i = 0, e = Operands.size(); i != e; ++i)
-    if (Instruction *OpI = dyn_cast(Operands[i]))
-      if (CurLoop->contains(OpI->getParent()) && canSinkOrHoistInst(*OpI) &&
-          isNotUsedInLoop(*OpI) && isSafeToExecuteUnconditionally(*OpI))
-        sink(*OpI);
 }
 
 /// hoist - When an instruction is found to only use loop invariant operands




From lattner at cs.uiuc.edu  Fri Dec 19 02:44:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 02:44:01 2003
Subject: [llvm-commits] CVS: llvm/docs/AliasAnalysis.html
Message-ID: <200312190843.CAA01831@zion.cs.uiuc.edu>


Changes in directory llvm/docs:

AliasAnalysis.html updated: 1.14 -> 1.15

---
Log message:

Expand on the AliasSetTracker a bit, so I can remember this next time I come back to it.  :)


---
Diffs of the changes:  (+33 -6)

Index: llvm/docs/AliasAnalysis.html
diff -u llvm/docs/AliasAnalysis.html:1.14 llvm/docs/AliasAnalysis.html:1.15
--- llvm/docs/AliasAnalysis.html:1.14	Wed Dec 17 17:10:49 2003
+++ llvm/docs/AliasAnalysis.html	Fri Dec 19 02:43:07 2003
@@ -366,9 +366,9 @@
 begin()/end() methods.
 
 The AliasSets formed by the AliasSetTracker are guaranteed
-to be disjoint, calculate mod/ref information for the set, and keep track of
-whether or not all of the pointers in the set are Must aliases.  The
-AliasSetTracker also makes sure that sets are properly folded due to call
+to be disjoint, calculate mod/ref information and volatility for the set, and
+keep track of whether or not all of the pointers in the set are Must aliases.
+The AliasSetTracker also makes sure that sets are properly folded due to call
 instructions, and can provide a list of pointers in each set.
 
 As an example user of this, the Loop
@@ -376,11 +376,38 @@
 about each loop nest.  If an AliasSet in a loop is not modified, then all load
 instructions from that set may be hoisted out of the loop.  If any alias sets
 are stored and are must alias sets, then the stores may be sunk to
-outside of the loop.  Both of these transformations obviously only apply if the
-pointer argument is loop-invariant.

+outside of the loop, promoting the memory location to a register for the
+duration of the loop nest.  Both of these transformations obviously only apply
+if the pointer argument is loop-invariant.
 
 

 
+
+  The AliasSetTracker implementation
+
+
+
+
+The AliasSetTracker class is implemented to be as efficient as possible.  It
+uses the union-find algorithm to efficiently merge AliasSets when a pointer is
+inserted into the AliasSetTracker that aliases multiple sets.  The primary data
+structure is a hash table mapping pointers to the AliasSet they are in.
+
+The AliasSetTracker class must maintain a list of all of the LLVM Value*'s
+that are in each AliasSet.  Since the hash table already has entries for each
+LLVM Value* of interest, the AliasesSets thread the linked list through these
+hash-table nodes to avoid having to allocate memory unnecessarily, and to make
+merging alias sets extremely efficient (the linked list merge is constant time).
+
+
+You shouldn't need to understand these details if you are just a client of
+the AliasSetTracker, but if you look at the code, hopefully this brief
+description will help make sense of why things are designed the way they
+are.
+
+
+
+
 
 
   Using the AliasAnalysis interface directly
@@ -486,7 +513,7 @@
 
   Chris Lattner

   The LLVM Compiler Infrastructure

-  Last modified: $Date: 2003/12/17 23:10:49 $
+  Last modified: $Date: 2003/12/19 08:43:07 $
 
 
 




From lattner at cs.uiuc.edu  Fri Dec 19 03:09:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 03:09:02 2003
Subject: [llvm-commits] CVS: llvm/lib/Transforms/Scalar/ADCE.cpp
Message-ID: <200312190908.DAA19502@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Transforms/Scalar:

ADCE.cpp updated: 1.69 -> 1.70

---
Log message:

Fix ADCE/2003-12-19-MergeReturn.llx


---
Diffs of the changes:  (+8 -7)

Index: llvm/lib/Transforms/Scalar/ADCE.cpp
diff -u llvm/lib/Transforms/Scalar/ADCE.cpp:1.69 llvm/lib/Transforms/Scalar/ADCE.cpp:1.70
--- llvm/lib/Transforms/Scalar/ADCE.cpp:1.69	Fri Nov 21 20:13:08 2003
+++ llvm/lib/Transforms/Scalar/ADCE.cpp	Fri Dec 19 03:08:34 2003
@@ -14,21 +14,21 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Transforms/Scalar.h"
-#include "llvm/Transforms/Utils/Local.h"
-#include "llvm/Transforms/Utils/BasicBlockUtils.h"
 #include "llvm/Type.h"
 #include "llvm/Analysis/PostDominators.h"
 #include "llvm/iTerminators.h"
 #include "llvm/iPHINode.h"
 #include "llvm/Constant.h"
 #include "llvm/Support/CFG.h"
+#include "llvm/Transforms/Utils/BasicBlockUtils.h"
+#include "llvm/Transforms/Utils/Local.h"
+#include "llvm/Transforms/Utils/UnifyFunctionExitNodes.h"
 #include "Support/Debug.h"
 #include "Support/DepthFirstIterator.h"
 #include "Support/Statistic.h"
 #include "Support/STLExtras.h"
 #include 
-
-namespace llvm {
+using namespace llvm;
 
 namespace {
   Statistic<> NumBlockRemoved("adce", "Number of basic blocks removed");
@@ -61,6 +61,9 @@
   // getAnalysisUsage - We require post dominance frontiers (aka Control
   // Dependence Graph)
   virtual void getAnalysisUsage(AnalysisUsage &AU) const {
+    // We require that all function nodes are unified, because otherwise code
+    // can be marked live that wouldn't necessarily be otherwise.
+    AU.addRequired();
     AU.addRequired();
     AU.addRequired();
   }
@@ -101,7 +104,7 @@
   RegisterOpt X("adce", "Aggressive Dead Code Elimination");
 } // End of anonymous namespace
 
-Pass *createAggressiveDCEPass() { return new ADCE(); }
+Pass *llvm::createAggressiveDCEPass() { return new ADCE(); }
 
 void ADCE::markBlockAlive(BasicBlock *BB) {
   // Mark the basic block as being newly ALIVE... and mark all branches that
@@ -474,5 +477,3 @@
 
   return MadeChanges;
 }
-
-} // End llvm namespace




From lattner at cs.uiuc.edu  Fri Dec 19 03:09:11 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 03:09:11 2003
Subject: [llvm-commits] CVS: llvm/test/Regression/Transforms/ADCE/2003-12-19-MergeReturn.llx
Message-ID: <200312190908.DAA18924@zion.cs.uiuc.edu>


Changes in directory llvm/test/Regression/Transforms/ADCE:

2003-12-19-MergeReturn.llx added (r1.1)

---
Log message:

New testcase for missed optimization


---
Diffs of the changes:  (+28 -0)

Index: llvm/test/Regression/Transforms/ADCE/2003-12-19-MergeReturn.llx
diff -c /dev/null llvm/test/Regression/Transforms/ADCE/2003-12-19-MergeReturn.llx:1.1
*** /dev/null	Fri Dec 19 03:08:25 2003
--- llvm/test/Regression/Transforms/ADCE/2003-12-19-MergeReturn.llx	Fri Dec 19 03:08:15 2003
***************
*** 0 ****
--- 1,28 ----
+ ; This testcase was failing because without merging the return blocks, ADCE
+ ; didn't know that it could get rid of the then.0 block.
+ 
+ ; RUN: llvm-as < %s | opt -adce | llvm-dis | not grep load
+ 
+ 
+ implementation   ; Functions:
+ 
+ void %main(uint %argc, sbyte** %argv) {
+ entry:
+ 	call void %__main( )
+ 	%tmp.1 = setle uint %argc, 5		;  [#uses=1]
+ 	br bool %tmp.1, label %then.0, label %return
+ 
+ then.0:		; preds = %entry
+ 	%tmp.8 = load sbyte** %argv		;  [#uses=1]
+ 	%tmp.10 = load sbyte* %tmp.8		;  [#uses=1]
+ 	%tmp.11 = seteq sbyte %tmp.10, 98		;  [#uses=1]
+ 	br bool %tmp.11, label %then.1, label %return
+ 
+ then.1:		; preds = %then.0
+ 	ret void
+ 
+ return:		; preds = %entry, %then.0
+ 	ret void
+ }
+ 
+ declare void %__main()




From criswell at cs.uiuc.edu  Fri Dec 19 11:23:02 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Fri Dec 19 11:23:02 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.html
Message-ID: <200312191722.LAA03246@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.html updated: 1.4 -> 1.5

---
Log message:

All of the previous downloaders (except me!) are in the database now.



---
Diffs of the changes:  (+1 -2)

Index: llvm-www/releases/register.html
diff -u llvm-www/releases/register.html:1.4 llvm-www/releases/register.html:1.5
--- llvm-www/releases/register.html:1.4	Thu Dec 18 15:52:44 2003
+++ llvm-www/releases/register.html	Fri Dec 19 11:22:25 2003
@@ -16,8 +16,7 @@
 


 We'd like to collect some information about people and organizations who are
 interested in LLVM.  Please complete the following form below to download a copy
-of the LLVM software, or use the Quick Download if you have registered since
-December 19, 2003.
+of the LLVM software, or use the Quick Download if you have already registered.
 
 
 
 




From criswell at cs.uiuc.edu  Fri Dec 19 11:24:01 2003
From: criswell at cs.uiuc.edu (John Criswell)
Date: Fri Dec 19 11:24:01 2003
Subject: [llvm-commits] CVS: llvm-www/releases/register.html
Message-ID: <200312191723.LAA03354@choi.cs.uiuc.edu>


Changes in directory llvm-www/releases:

register.html updated: 1.5 -> 1.6

---
Log message:

Added clarification.



---
Diffs of the changes:  (+2 -1)

Index: llvm-www/releases/register.html
diff -u llvm-www/releases/register.html:1.5 llvm-www/releases/register.html:1.6
--- llvm-www/releases/register.html:1.5	Fri Dec 19 11:22:25 2003
+++ llvm-www/releases/register.html	Fri Dec 19 11:23:13 2003
@@ -16,7 +16,8 @@
 


 We'd like to collect some information about people and organizations who are
 interested in LLVM.  Please complete the following form below to download a copy
-of the LLVM software, or use the Quick Download if you have already registered.
+of the LLVM software, or use the Quick Download if you have already registered
+previously.
 
 
 




From lattner at cs.uiuc.edu  Fri Dec 19 13:49:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 13:49:02 2003
Subject: [llvm-commits] CVS: llvm/utils/NightlyTest.pl
Message-ID: <200312191948.NAA18843@zion.cs.uiuc.edu>


Changes in directory llvm/utils:

NightlyTest.pl updated: 1.41 -> 1.42

---
Log message:

Eliminate a spurious warning


---
Diffs of the changes:  (+1 -1)

Index: llvm/utils/NightlyTest.pl
diff -u llvm/utils/NightlyTest.pl:1.41 llvm/utils/NightlyTest.pl:1.42
--- llvm/utils/NightlyTest.pl:1.41	Thu Dec 18 21:47:31 2003
+++ llvm/utils/NightlyTest.pl	Fri Dec 19 13:48:43 2003
@@ -132,7 +132,7 @@
   if (/^-notest$/)         { $NOTEST     = 1; $NORUNNINGTESTS = 1; next; }
   if (/^-norunningtests$/) { $NORUNNINGTESTS = 1; next; }
   if (/^-parallel$/)       { $MAKEOPTS   = "-j2 -l3.0"; next; }
-  if (/^-enable-linscan$/) { $ENABLELINEARSCAN = "ENABLE_LINEARSCAN=1"; }
+  if (/^-enable-linscan$/) { $ENABLELINEARSCAN = "ENABLE_LINEARSCAN=1"; next; }
 
   print "Unknown option: $_ : ignoring!\n";
 }




From lattner at cs.uiuc.edu  Fri Dec 19 17:26:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 17:26:01 2003
Subject: [llvm-commits] CVS: llvm/test/Programs/SingleSource/CustomChecked/Makefile
Message-ID: <200312192325.RAA32562@zion.cs.uiuc.edu>


Changes in directory llvm/test/Programs/SingleSource/CustomChecked:

Makefile updated: 1.9 -> 1.10

---
Log message:

Add linear scan targets


---
Diffs of the changes:  (+10 -1)

Index: llvm/test/Programs/SingleSource/CustomChecked/Makefile
diff -u llvm/test/Programs/SingleSource/CustomChecked/Makefile:1.9 llvm/test/Programs/SingleSource/CustomChecked/Makefile:1.10
--- llvm/test/Programs/SingleSource/CustomChecked/Makefile:1.9	Thu Oct 23 10:12:18 2003
+++ llvm/test/Programs/SingleSource/CustomChecked/Makefile	Fri Dec 19 17:25:15 2003
@@ -12,8 +12,9 @@
 LLI_RUN  := $(addsuffix .run-lli, $(PREFIXED_PROGRAMS_TO_TEST))
 JIT_RUN  := $(addsuffix .run-jit, $(PREFIXED_PROGRAMS_TO_TEST))
 LLC_RUN  := $(addsuffix .run-llc, $(PREFIXED_PROGRAMS_TO_TEST))
+LLCLS_RUN  := $(addsuffix .run-llc-ls, $(PREFIXED_PROGRAMS_TO_TEST))
 CBE_RUN  := $(addsuffix .run-cbe, $(PREFIXED_PROGRAMS_TO_TEST))
-.PRECIOUS: $(LLI_RUN) $(JIT_RUN) $(LLC_RUN) $(CBE_RUN)
+.PRECIOUS: $(LLI_RUN) $(JIT_RUN) $(LLC_RUN) $(LLCLS_RUN) $(CBE_RUN)
 
 ifndef DISABLE_LLC_DIFFS
 all:: $(LLC_RUN)
@@ -45,6 +46,10 @@
 Output/%.run-llc: Output/%.llc
 	-$(TESTRUNR) $(SourceDir)/$(filter $*.%, $(Source)) "$< $(RUN_OPTIONS)" $@
 
+$(PROGRAMS_TO_TEST:%=Output/%.run-llc-ls): \
+Output/%.run-llc-ls: Output/%.llc-ls
+	-$(TESTRUNR) $(SourceDir)/$(filter $*.%, $(Source)) "$< $(RUN_OPTIONS)" $@
+
 $(PROGRAMS_TO_TEST:%=Output/%.run-cbe): \
 Output/%.run-cbe: Output/%.cbe
 	-$(TESTRUNR) $(SourceDir)/$(filter $*.%, $(Source)) "$< $(RUN_OPTIONS)" $@
@@ -60,6 +65,10 @@
 
 $(PROGRAMS_TO_TEST:%=Output/%.exe-llc): \
 Output/%.exe-llc: Output/%.run-llc
+	-cp $< $@
+
+$(PROGRAMS_TO_TEST:%=Output/%.exe-llc-ls): \
+Output/%.exe-llc-ls: Output/%.run-llc-ls
 	-cp $< $@
 
 $(PROGRAMS_TO_TEST:%=Output/%.exe-cbe): \




From vnc61ut at bigfoot.com  Fri Dec 19 17:54:02 2003
From: vnc61ut at bigfoot.com (Sheri Peoples)
Date: Fri Dec 19 17:54:02 2003
Subject: [llvm-commits] speed up low carb weight loss 9b  kzjmvlsyla g
Message-ID: <0a4-0hh-71g-sr@on843r>

An HTML attachment was scrubbed...
URL: http://lists.cs.uiuc.edu/pipermail/llvm-commits/attachments/20031219/8058f635/attachment.html 

From b366rtc at lycos.de  Fri Dec 19 18:13:02 2003
From: b366rtc at lycos.de (Viola Self)
Date: Fri Dec 19 18:13:02 2003
Subject: [llvm-commits] guaranteed Weight loss program xjjuwmnp ept
Message-ID: 

An HTML attachment was scrubbed...
URL: http://lists.cs.uiuc.edu/pipermail/llvm-commits/attachments/20031219/1354c676/attachment.html 

From lattner at cs.uiuc.edu  Fri Dec 19 19:22:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:22:01 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/Target/TargetJITInfo.h TargetMachine.h
Message-ID: <200312200121.TAA24509@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/Target:

TargetJITInfo.h added (r1.1)
TargetMachine.h updated: 1.40 -> 1.41

---
Log message:

Rip JIT specific stuff out of TargetMachine, as per PR176


---
Diffs of the changes:  (+62 -25)

Index: llvm/include/llvm/Target/TargetJITInfo.h
diff -c /dev/null llvm/include/llvm/Target/TargetJITInfo.h:1.1
*** /dev/null	Fri Dec 19 19:21:48 2003
--- llvm/include/llvm/Target/TargetJITInfo.h	Fri Dec 19 19:21:37 2003
***************
*** 0 ****
--- 1,54 ----
+ //===- Target/TargetJITInfo.h - Target Information for JIT ------*- C++ -*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // This file exposes an abstract interface used by the Just-In-Time code
+ // generator to perform target-specific activities, such as emitting stubs.  If
+ // a TargetMachine supports JIT code generation, it should provide one of these
+ // objects through the getJITInfo() method.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef LLVM_TARGET_TARGETJITINFO_H
+ #define LLVM_TARGET_TARGETJITINFO_H
+ 
+ namespace llvm {
+   class Function;
+   class FunctionPassManager;
+   class MachineCodeEmitter;
+ 
+   /// TargetJITInfo - Target specific information required by the Just-In-Time
+   /// code generator.
+   struct TargetJITInfo {
+     virtual ~TargetJITInfo() {}
+     
+     /// addPassesToJITCompile - Add passes to the specified pass manager to
+     /// implement a fast code generator for this target.
+     ///
+     virtual void addPassesToJITCompile(FunctionPassManager &PM) = 0;
+     
+     /// replaceMachineCodeForFunction - Make it so that calling the function
+     /// whose machine code is at OLD turns into a call to NEW, perhaps by
+     /// overwriting OLD with a branch to NEW.  This is used for self-modifying
+     /// code.
+     ///
+     virtual void replaceMachineCodeForFunction (void *Old, void *New) = 0;
+     
+     /// getJITStubForFunction - Create or return a stub for the specified
+     /// function.  This stub acts just like the specified function, except that
+     /// it allows the "address" of the function to be taken without having to
+     /// generate code for it.  Targets do not need to implement this method, but
+     /// doing so will allow for faster startup of the JIT.
+     ///
+     virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE) {
+       return 0;
+     }
+   };
+ } // End llvm namespace
+ 
+ #endif


Index: llvm/include/llvm/Target/TargetMachine.h
diff -u llvm/include/llvm/Target/TargetMachine.h:1.40 llvm/include/llvm/Target/TargetMachine.h:1.41
--- llvm/include/llvm/Target/TargetMachine.h:1.40	Fri Dec 12 01:10:32 2003
+++ llvm/include/llvm/Target/TargetMachine.h	Fri Dec 19 19:21:37 2003
@@ -21,6 +21,7 @@
 
 class TargetInstrInfo;
 class TargetInstrDescriptor;
+class TargetJITInfo;
 class TargetSchedInfo;
 class TargetRegInfo;
 class TargetFrameInfo;
@@ -79,16 +80,16 @@
   ///
   virtual const MRegisterInfo*          getRegisterInfo() const { return 0; }
 
-  // Data storage information
+  /// getJITInfo - If this target supports a JIT, return information for it,
+  /// otherwise return null.
+  ///
+  virtual TargetJITInfo *getJITInfo() { return 0; }
+
+  // Data storage information.  FIXME, this should be moved out to sparc
+  // specific code.
   // 
   virtual unsigned findOptimalStorageSize(const Type* ty) const;
   
-  /// addPassesToJITCompile - Add passes to the specified pass manager to
-  /// implement a fast dynamic compiler for this target.  Return true if this is
-  /// not supported for this target.
-  ///
-  virtual bool addPassesToJITCompile(FunctionPassManager &PM) { return true; }
-
   /// addPassesToEmitAssembly - Add passes to the specified pass manager to get
   /// assembly langage code emitted.  Typically this will involve several steps
   /// of code generation.  This method should return true if assembly emission
@@ -107,24 +108,6 @@
   virtual bool addPassesToEmitMachineCode(FunctionPassManager &PM,
                                           MachineCodeEmitter &MCE) {
     return true;
-  }
-
-  /// replaceMachineCodeForFunction - Make it so that calling the
-  /// function whose machine code is at OLD turns into a call to NEW,
-  /// perhaps by overwriting OLD with a branch to NEW.
-  ///
-  /// FIXME: this is JIT-specific.
-  ///
-  virtual void replaceMachineCodeForFunction (void *Old, void *New) {
-    assert (0 && "Current target cannot replace machine code for functions");
-  }
-
-  /// getJITStubForFunction - Create or return a stub for the specified
-  /// function.  This stub acts just like the specified function, except that it
-  /// allows the "address" of the function to be taken without having to
-  /// generate code for it.
-  virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE) {
-    return 0;
   }
 };
 




From lattner at cs.uiuc.edu  Fri Dec 19 19:23:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:23:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/X86/X86JITInfo.h X86CodeEmitter.cpp X86InstrInfo.h X86TargetMachine.cpp X86TargetMachine.h
Message-ID: <200312200122.TAA24548@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/X86:

X86JITInfo.h added (r1.1)
X86CodeEmitter.cpp updated: 1.42 -> 1.43
X86InstrInfo.h updated: 1.28 -> 1.29
X86TargetMachine.cpp updated: 1.39 -> 1.40
X86TargetMachine.h updated: 1.18 -> 1.19

---
Log message:

Rip JIT specific stuff out of TargetMachine, as per PR176



---
Diffs of the changes:  (+74 -38)

Index: llvm/lib/Target/X86/X86JITInfo.h
diff -c /dev/null llvm/lib/Target/X86/X86JITInfo.h:1.1
*** /dev/null	Fri Dec 19 19:22:15 2003
--- llvm/lib/Target/X86/X86JITInfo.h	Fri Dec 19 19:22:04 2003
***************
*** 0 ****
--- 1,47 ----
+ //===- X86JITInfo.h - X86 implementation of the JIT interface  --*- C++ -*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // This file contains the X86 implementation of the TargetJITInfo class.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef X86JITINFO_H
+ #define X86JITINFO_H
+ 
+ #include "llvm/Target/TargetJITInfo.h"
+ 
+ namespace llvm {
+   class TargetMachine;
+   class X86JITInfo : public TargetJITInfo {
+     TargetMachine &TM;
+   public:
+     X86JITInfo(TargetMachine &tm) : TM(tm) {}
+ 
+     /// addPassesToJITCompile - Add passes to the specified pass manager to
+     /// implement a fast dynamic compiler for this target.  Return true if this
+     /// is not supported for this target.
+     ///
+     virtual void addPassesToJITCompile(FunctionPassManager &PM);
+     
+     /// replaceMachineCodeForFunction - Make it so that calling the function
+     /// whose machine code is at OLD turns into a call to NEW, perhaps by
+     /// overwriting OLD with a branch to NEW.  This is used for self-modifying
+     /// code.
+     ///
+     virtual void replaceMachineCodeForFunction (void *Old, void *New);
+     
+     /// getJITStubForFunction - Create or return a stub for the specified
+     /// function.  This stub acts just like the specified function, except that
+     /// it allows the "address" of the function to be taken without having to
+     /// generate code for it.
+     virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE);
+   };
+ }
+ 
+ #endif


Index: llvm/lib/Target/X86/X86CodeEmitter.cpp
diff -u llvm/lib/Target/X86/X86CodeEmitter.cpp:1.42 llvm/lib/Target/X86/X86CodeEmitter.cpp:1.43
--- llvm/lib/Target/X86/X86CodeEmitter.cpp:1.42	Fri Dec 12 13:57:48 2003
+++ llvm/lib/Target/X86/X86CodeEmitter.cpp	Fri Dec 19 19:22:04 2003
@@ -53,11 +53,19 @@
   JITResolver *TheJITResolver;
 }
 
-void *X86TargetMachine::getJITStubForFunction(Function *F,
-                                              MachineCodeEmitter &MCE) {
+void *X86JITInfo::getJITStubForFunction(Function *F, MachineCodeEmitter &MCE) {
   if (TheJITResolver == 0)
     TheJITResolver = new JITResolver(MCE);
   return (void*)((unsigned long)TheJITResolver->getLazyResolver(F));
+}
+
+void X86JITInfo::replaceMachineCodeForFunction (void *Old, void *New) {
+  char *OldByte = (char *) Old;
+  *OldByte++ = 0xE9;                // Emit JMP opcode.
+  int32_t *OldWord = (int32_t *) OldByte;
+  int32_t NewAddr = (intptr_t) New;
+  int32_t OldAddr = (intptr_t) OldWord;
+  *OldWord = NewAddr - OldAddr - 4; // Emit PC-relative addr of New code.
 }
 
 /// addFunctionReference - This method is called when we need to emit the


Index: llvm/lib/Target/X86/X86InstrInfo.h
diff -u llvm/lib/Target/X86/X86InstrInfo.h:1.28 llvm/lib/Target/X86/X86InstrInfo.h:1.29
--- llvm/lib/Target/X86/X86InstrInfo.h:1.28	Tue Nov 11 16:41:33 2003
+++ llvm/lib/Target/X86/X86InstrInfo.h	Fri Dec 19 19:22:04 2003
@@ -1,4 +1,4 @@
-//===- X86InstructionInfo.h - X86 Instruction Information ---------*-C++-*-===//
+//===- X86InstrInfo.h - X86 Instruction Information ------------*- C++ -*- ===//
 // 
 //                     The LLVM Compiler Infrastructure
 //


Index: llvm/lib/Target/X86/X86TargetMachine.cpp
diff -u llvm/lib/Target/X86/X86TargetMachine.cpp:1.39 llvm/lib/Target/X86/X86TargetMachine.cpp:1.40
--- llvm/lib/Target/X86/X86TargetMachine.cpp:1.39	Fri Dec 12 23:36:19 2003
+++ llvm/lib/Target/X86/X86TargetMachine.cpp	Fri Dec 19 19:22:04 2003
@@ -21,8 +21,7 @@
 #include "llvm/Transforms/Scalar.h"
 #include "Support/CommandLine.h"
 #include "Support/Statistic.h"
-
-namespace llvm {
+using namespace llvm;
 
 namespace {
   cl::opt PrintCode("print-machineinstrs",
@@ -36,7 +35,7 @@
 // allocateX86TargetMachine - Allocate and return a subclass of TargetMachine
 // that implements the X86 backend.
 //
-TargetMachine *allocateX86TargetMachine(const Module &M) {
+TargetMachine *llvm::allocateX86TargetMachine(const Module &M) {
   return new X86TargetMachine(M);
 }
 
@@ -45,7 +44,8 @@
 ///
 X86TargetMachine::X86TargetMachine(const Module &M)
   : TargetMachine("X86", true, 4, 4, 4, 4, 4),
-    FrameInfo(TargetFrameInfo::StackGrowsDown, 8/*16 for SSE*/, 4) {
+    FrameInfo(TargetFrameInfo::StackGrowsDown, 8/*16 for SSE*/, 4),
+    JITInfo(*this) {
 }
 
 
@@ -108,7 +108,7 @@
 /// implement a fast dynamic compiler for this target.  Return true if this is
 /// not supported for this target.
 ///
-bool X86TargetMachine::addPassesToJITCompile(FunctionPassManager &PM) {
+void X86JITInfo::addPassesToJITCompile(FunctionPassManager &PM) {
   // FIXME: Implement the switch instruction in the instruction selector!
   PM.add(createLowerSwitchPass());
 
@@ -120,9 +120,9 @@
   PM.add(createCFGSimplificationPass());
 
   if (NoPatternISel)
-    PM.add(createX86SimpleInstructionSelector(*this));
+    PM.add(createX86SimpleInstructionSelector(TM));
   else
-    PM.add(createX86PatternInstructionSelector(*this));
+    PM.add(createX86PatternInstructionSelector(TM));
 
   // Run optional SSA-based machine code optimizations next...
   if (!NoSSAPeephole)
@@ -156,18 +156,6 @@
   PM.add(createX86PeepholeOptimizerPass());
 
   if (PrintCode)  // Print the register-allocated code
-    PM.add(createX86CodePrinterPass(std::cerr, *this));
-  return false; // success!
-}
-
-void X86TargetMachine::replaceMachineCodeForFunction (void *Old, void *New) {
-  // FIXME: This code could perhaps live in a more appropriate place.
-  char *OldByte = (char *) Old;
-  *OldByte++ = 0xE9;                // Emit JMP opcode.
-  int32_t *OldWord = (int32_t *) OldByte;
-  int32_t NewAddr = (intptr_t) New;
-  int32_t OldAddr = (intptr_t) OldWord;
-  *OldWord = NewAddr - OldAddr - 4; // Emit PC-relative addr of New code.
+    PM.add(createX86CodePrinterPass(std::cerr, TM));
 }
 
-} // End llvm namespace


Index: llvm/lib/Target/X86/X86TargetMachine.h
diff -u llvm/lib/Target/X86/X86TargetMachine.h:1.18 llvm/lib/Target/X86/X86TargetMachine.h:1.19
--- llvm/lib/Target/X86/X86TargetMachine.h:1.18	Fri Dec 12 01:11:18 2003
+++ llvm/lib/Target/X86/X86TargetMachine.h	Fri Dec 19 19:22:04 2003
@@ -18,12 +18,14 @@
 #include "llvm/Target/TargetFrameInfo.h"
 #include "llvm/PassManager.h"
 #include "X86InstrInfo.h"
+#include "X86JITInfo.h"
 
 namespace llvm {
 
 class X86TargetMachine : public TargetMachine {
-  X86InstrInfo InstrInfo;
+  X86InstrInfo    InstrInfo;
   TargetFrameInfo FrameInfo;
+  X86JITInfo      JITInfo;
 public:
   X86TargetMachine(const Module &M);
 
@@ -33,16 +35,15 @@
     return &InstrInfo.getRegisterInfo();
   }
 
+  virtual TargetJITInfo *getJITInfo() {
+    return &JITInfo;
+  }
+
+
   virtual const TargetSchedInfo &getSchedInfo()  const { abort(); }
   virtual const TargetRegInfo   &getRegInfo()    const { abort(); }
   virtual const TargetCacheInfo  &getCacheInfo() const { abort(); }
 
-  /// addPassesToJITCompile - Add passes to the specified pass manager to
-  /// implement a fast dynamic compiler for this target.  Return true if this is
-  /// not supported for this target.
-  ///
-  virtual bool addPassesToJITCompile(FunctionPassManager &PM);
-
   /// addPassesToEmitMachineCode - Add passes to the specified pass manager to
   /// get machine code emitted.  This uses a MachineCodeEmitter object to handle
   /// actually outputting the machine code and resolving things like the address
@@ -53,14 +54,6 @@
                                           MachineCodeEmitter &MCE);
   
   virtual bool addPassesToEmitAssembly(PassManager &PM, std::ostream &Out);
-
-  virtual void replaceMachineCodeForFunction (void *Old, void *New);
-
-  /// getJITStubForFunction - Create or return a stub for the specified
-  /// function.  This stub acts just like the specified function, except that it
-  /// allows the "address" of the function to be taken without having to
-  /// generate code for it.
-  virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE);
 };
 
 } // End llvm namespace




From lattner at cs.uiuc.edu  Fri Dec 19 19:23:10 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:23:10 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/Sparc/SparcJITInfo.h SparcTargetMachine.cpp SparcTargetMachine.h SparcV9CodeEmitter.cpp
Message-ID: <200312200122.TAA24530@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/Sparc:

SparcJITInfo.h added (r1.1)
SparcTargetMachine.cpp updated: 1.92 -> 1.93
SparcTargetMachine.h updated: 1.1 -> 1.2
SparcV9CodeEmitter.cpp updated: 1.47 -> 1.48

---
Log message:

Rip JIT specific stuff out of TargetMachine, as per PR176


---
Diffs of the changes:  (+61 -25)

Index: llvm/lib/Target/Sparc/SparcJITInfo.h
diff -c /dev/null llvm/lib/Target/Sparc/SparcJITInfo.h:1.1
*** /dev/null	Fri Dec 19 19:22:09 2003
--- llvm/lib/Target/Sparc/SparcJITInfo.h	Fri Dec 19 19:21:59 2003
***************
*** 0 ****
--- 1,47 ----
+ //===- SparcJITInfo.h - Sparc implementation of the JIT interface -*-C++-*-===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // This file contains the Sparc implementation of the TargetJITInfo class.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #ifndef SPARCJITINFO_H
+ #define SPARCJITINFO_H
+ 
+ #include "llvm/Target/TargetJITInfo.h"
+ 
+ namespace llvm {
+   class TargetMachine;
+   class SparcJITInfo : public TargetJITInfo {
+     TargetMachine &TM;
+   public:
+     SparcJITInfo(TargetMachine &tm) : TM(tm) {}
+ 
+     /// addPassesToJITCompile - Add passes to the specified pass manager to
+     /// implement a fast dynamic compiler for this target.  Return true if this
+     /// is not supported for this target.
+     ///
+     virtual void addPassesToJITCompile(FunctionPassManager &PM);
+     
+     /// replaceMachineCodeForFunction - Make it so that calling the function
+     /// whose machine code is at OLD turns into a call to NEW, perhaps by
+     /// overwriting OLD with a branch to NEW.  This is used for self-modifying
+     /// code.
+     ///
+     virtual void replaceMachineCodeForFunction (void *Old, void *New);
+     
+     /// getJITStubForFunction - Create or return a stub for the specified
+     /// function.  This stub acts just like the specified function, except that
+     /// it allows the "address" of the function to be taken without having to
+     /// generate code for it.
+     //virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE);
+   };
+ }
+ 
+ #endif


Index: llvm/lib/Target/Sparc/SparcTargetMachine.cpp
diff -u llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.92 llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.93
--- llvm/lib/Target/Sparc/SparcTargetMachine.cpp:1.92	Wed Dec 17 16:04:00 2003
+++ llvm/lib/Target/Sparc/SparcTargetMachine.cpp	Fri Dec 19 19:21:59 2003
@@ -73,7 +73,8 @@
     schedInfo(*this),
     regInfo(*this),
     frameInfo(*this),
-    cacheInfo(*this) {
+    cacheInfo(*this),
+    jitInfo(*this) {
 }
 
 // addPassesToEmitAssembly - This method controls the entire code generation
@@ -152,8 +153,8 @@
 // addPassesToJITCompile - This method controls the JIT method of code
 // generation for the UltraSparc.
 //
-bool SparcTargetMachine::addPassesToJITCompile(FunctionPassManager &PM) {
-  const TargetData &TD = getTargetData();
+void SparcJITInfo::addPassesToJITCompile(FunctionPassManager &PM) {
+  const TargetData &TD = TM.getTargetData();
 
   PM.add(new TargetData("lli", TD.isLittleEndian(), TD.getPointerSize(),
                         TD.getPointerAlignment(), TD.getDoubleAlignment()));
@@ -173,11 +174,11 @@
   PM.add(createDecomposeMultiDimRefsPass());
   
   // Construct and initialize the MachineFunction object for this fn.
-  PM.add(createMachineCodeConstructionPass(*this));
+  PM.add(createMachineCodeConstructionPass(TM));
 
   // Specialize LLVM code for this target machine and then
   // run basic dataflow optimizations on LLVM code.
-  PM.add(createPreSelectionPass(*this));
+  PM.add(createPreSelectionPass(TM));
   // Run basic dataflow optimizations on LLVM code
   PM.add(createReassociatePass());
 
@@ -185,15 +186,13 @@
   //PM.add(createLICMPass());
   //PM.add(createGCSEPass());
 
-  PM.add(createInstructionSelectionPass(*this));
+  PM.add(createInstructionSelectionPass(TM));
 
-  PM.add(getRegisterAllocator(*this));
+  PM.add(getRegisterAllocator(TM));
   PM.add(createPrologEpilogInsertionPass());
 
   if (!DisablePeephole)
-    PM.add(createPeepholeOptsPass(*this));
-
-  return false; // success!
+    PM.add(createPeepholeOptsPass(TM));
 }
 
 //----------------------------------------------------------------------------
@@ -201,10 +200,6 @@
 // that implements the Sparc backend. (the llvm/CodeGen/Sparc.h interface)
 //----------------------------------------------------------------------------
 
-namespace llvm {
-
-TargetMachine *allocateSparcTargetMachine(const Module &M) {
+TargetMachine *llvm::allocateSparcTargetMachine(const Module &M) {
   return new SparcTargetMachine();
-}
-
 }


Index: llvm/lib/Target/Sparc/SparcTargetMachine.h
diff -u llvm/lib/Target/Sparc/SparcTargetMachine.h:1.1 llvm/lib/Target/Sparc/SparcTargetMachine.h:1.2
--- llvm/lib/Target/Sparc/SparcTargetMachine.h:1.1	Wed Dec 17 16:04:00 2003
+++ llvm/lib/Target/Sparc/SparcTargetMachine.h	Fri Dec 19 19:21:59 2003
@@ -22,6 +22,7 @@
 #include "SparcInternals.h"
 #include "SparcRegInfo.h"
 #include "SparcFrameInfo.h"
+#include "SparcJITInfo.h"
 
 namespace llvm {
 
@@ -31,6 +32,7 @@
   SparcRegInfo   regInfo;
   SparcFrameInfo frameInfo;
   SparcCacheInfo cacheInfo;
+  SparcJITInfo   jitInfo;
 public:
   SparcTargetMachine();
 
@@ -39,19 +41,11 @@
   virtual const TargetRegInfo    &getRegInfo()   const { return regInfo; }
   virtual const TargetFrameInfo  &getFrameInfo() const { return frameInfo; }
   virtual const TargetCacheInfo  &getCacheInfo() const { return cacheInfo; }
+  virtual       TargetJITInfo    *getJITInfo()         { return &jitInfo; }
 
   virtual bool addPassesToEmitAssembly(PassManager &PM, std::ostream &Out);
-  virtual bool addPassesToJITCompile(FunctionPassManager &PM);
   virtual bool addPassesToEmitMachineCode(FunctionPassManager &PM,
                                           MachineCodeEmitter &MCE);
-  virtual void replaceMachineCodeForFunction(void *Old, void *New);
-
-  /// getJITStubForFunction - Create or return a stub for the specified
-  /// function.  This stub acts just like the specified function, except that it
-  /// allows the "address" of the function to be taken without having to
-  /// generate code for it.
-  ///
-  ///virtual void *getJITStubForFunction(Function *F, MachineCodeEmitter &MCE);
 };
 
 } // End llvm namespace


Index: llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp
diff -u llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.47 llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.48
--- llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp:1.47	Wed Dec 17 16:04:00 2003
+++ llvm/lib/Target/Sparc/SparcV9CodeEmitter.cpp	Fri Dec 19 19:21:59 2003
@@ -582,7 +582,7 @@
   }
 }
 
-void SparcTargetMachine::replaceMachineCodeForFunction (void *Old, void *New) {
+void SparcJITInfo::replaceMachineCodeForFunction (void *Old, void *New) {
   assert (TheJITResolver &&
 	"Can only call replaceMachineCodeForFunction from within JIT");
   uint64_t Target = (uint64_t)(intptr_t)New;




From lattner at cs.uiuc.edu  Fri Dec 19 19:23:19 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:23:19 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/JIT/JIT.cpp VM.cpp VM.h
Message-ID: <200312200122.TAA24562@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine/JIT:

JIT.cpp updated: 1.26 -> 1.27
VM.cpp updated: 1.16 -> 1.17
VM.h updated: 1.18 -> 1.19

---
Log message:

Rip JIT specific stuff out of TargetMachine, as per PR176


---
Diffs of the changes:  (+16 -15)

Index: llvm/lib/ExecutionEngine/JIT/JIT.cpp
diff -u llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.26 llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.27
--- llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.26	Mon Dec  8 02:06:28 2003
+++ llvm/lib/ExecutionEngine/JIT/JIT.cpp	Fri Dec 19 19:22:19 2003
@@ -80,19 +80,20 @@
   // Allocate a target...
   TargetMachine *Target = TargetMachineAllocator(*MP->getModule());
   assert(Target && "Could not allocate target machine!");
-  
-  // Create the virtual machine object...
-  return new VM(MP, Target);
+
+  // If the target supports JIT code generation, return a new JIT now.
+  if (TargetJITInfo *TJ = Target->getJITInfo())
+    return new VM(MP, *Target, *TJ);
+  return 0;
 }
 
-VM::VM(ModuleProvider *MP, TargetMachine *tm) : ExecutionEngine(MP), TM(*tm),
-  PM(MP)
-{
+VM::VM(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji)
+  : ExecutionEngine(MP), TM(tm), TJI(tji), PM(MP) {
   setTargetData(TM.getTargetData());
 
   // Initialize MCE
   MCE = createEmitter(*this);
-
+  
   setupPassManager();
 
   emitGlobals();


Index: llvm/lib/ExecutionEngine/JIT/VM.cpp
diff -u llvm/lib/ExecutionEngine/JIT/VM.cpp:1.16 llvm/lib/ExecutionEngine/JIT/VM.cpp:1.17
--- llvm/lib/ExecutionEngine/JIT/VM.cpp:1.16	Fri Dec 12 01:12:02 2003
+++ llvm/lib/ExecutionEngine/JIT/VM.cpp	Fri Dec 19 19:22:19 2003
@@ -18,6 +18,7 @@
 #include "llvm/CodeGen/MachineCodeEmitter.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/Target/TargetMachine.h"
+#include "llvm/Target/TargetJITInfo.h"
 using namespace llvm;
 
 VM::~VM() {
@@ -30,11 +31,7 @@
 ///
 void VM::setupPassManager() {
   // Compile LLVM Code down to machine code in the intermediate representation
-  if (TM.addPassesToJITCompile(PM)) {
-    std::cerr << "lli: target '" << TM.getName()
-              << "' doesn't support JIT compilation!\n";
-    abort();
-  }
+  TJI.addPassesToJITCompile(PM);
 
   // Turn the machine code intermediate representation into bytes in memory that
   // may be executed.
@@ -87,7 +84,7 @@
   if (I != GlobalAddress.end()) return I->second;
 
   // If the target supports "stubs" for functions, get a stub now.
-  if (void *Ptr = TM.getJITStubForFunction(F, *MCE))
+  if (void *Ptr = TJI.getJITStubForFunction(F, *MCE))
     return Ptr;
 
   // Otherwise, if the target doesn't support it, just codegen the function.
@@ -112,6 +109,6 @@
   MachineFunction::destruct(F);
   runJITOnFunction(F);
   assert(Addr && "Code generation didn't add function to GlobalAddress table!");
-  TM.replaceMachineCodeForFunction(OldAddr, Addr);
+  TJI.replaceMachineCodeForFunction(OldAddr, Addr);
   return Addr;
 }


Index: llvm/lib/ExecutionEngine/JIT/VM.h
diff -u llvm/lib/ExecutionEngine/JIT/VM.h:1.18 llvm/lib/ExecutionEngine/JIT/VM.h:1.19
--- llvm/lib/ExecutionEngine/JIT/VM.h:1.18	Fri Dec 12 01:12:02 2003
+++ llvm/lib/ExecutionEngine/JIT/VM.h	Fri Dec 19 19:22:19 2003
@@ -24,15 +24,18 @@
 class GlobalValue;
 class Constant;
 class TargetMachine;
+class TargetJITInfo;
 class MachineCodeEmitter;
 
 class VM : public ExecutionEngine {
   TargetMachine &TM;       // The current target we are compiling to
+  TargetJITInfo &TJI;      // The JITInfo for the target we are compiling to
+  
   FunctionPassManager PM;  // Passes to compile a function
   MachineCodeEmitter *MCE; // MCE object
 
+  VM(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji);
 public:
-  VM(ModuleProvider *MP, TargetMachine *tm);
   ~VM();
 
   /// create - Create an return a new JIT compiler if there is one available




From lattner at cs.uiuc.edu  Fri Dec 19 19:46:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:46:01 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
Message-ID: <200312200145.TAA27517@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine:

ExecutionEngine.cpp updated: 1.39 -> 1.40

---
Log message:

Update for changes in the JIT


---
Diffs of the changes:  (+2 -2)

Index: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
diff -u llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.39 llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.40
--- llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.39	Mon Dec  8 02:22:48 2003
+++ llvm/lib/ExecutionEngine/ExecutionEngine.cpp	Fri Dec 19 19:45:17 2003
@@ -14,7 +14,7 @@
 
 #define DEBUG_TYPE "jit"
 #include "Interpreter/Interpreter.h"
-#include "JIT/VM.h"
+#include "JIT/JIT.h"
 #include "llvm/Constants.h"
 #include "llvm/DerivedTypes.h"
 #include "llvm/Module.h"
@@ -55,7 +55,7 @@
 
   // Unless the interpreter was explicitly selected, make a JIT.
   if (!ForceInterpreter)
-    EE = VM::create(MP);
+    EE = JIT::create(MP);
 
   // If we can't make a JIT, make an interpreter instead.
   try {




From lattner at cs.uiuc.edu  Fri Dec 19 19:47:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 19:47:01 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/JIT/TargetSelect.cpp Emitter.cpp Intercept.cpp JIT.cpp JIT.h VM.cpp VM.h
Message-ID: <200312200146.TAA27541@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine/JIT:

TargetSelect.cpp added (r1.1)
Emitter.cpp updated: 1.36 -> 1.37
Intercept.cpp updated: 1.11 -> 1.12
JIT.cpp updated: 1.27 -> 1.28
JIT.h updated: 1.19 -> 1.20
VM.cpp (r1.17) removed
VM.h (r1.19) removed

---
Log message:

Cleanup the JIT as per PR176.  This renames the VM class to JIT, and merges the
VM.cpp and JIT.cpp files into JIT.cpp.  This also splits some nasty code out
into TargetSelect.cpp so that people hopefully won't notice it.  :)


---
Diffs of the changes:  (+217 -107)

Index: llvm/lib/ExecutionEngine/JIT/TargetSelect.cpp
diff -c /dev/null llvm/lib/ExecutionEngine/JIT/TargetSelect.cpp:1.1
*** /dev/null	Fri Dec 19 19:46:37 2003
--- llvm/lib/ExecutionEngine/JIT/TargetSelect.cpp	Fri Dec 19 19:46:27 2003
***************
*** 0 ****
--- 1,92 ----
+ //===-- TargetSelect.cpp - Target Chooser Code ----------------------------===//
+ // 
+ //                     The LLVM Compiler Infrastructure
+ //
+ // This file was developed by the LLVM research group and is distributed under
+ // the University of Illinois Open Source License. See LICENSE.TXT for details.
+ // 
+ //===----------------------------------------------------------------------===//
+ //
+ // This file contains the hideously gross code that is currently used to select
+ // a particular TargetMachine for the JIT to use.  This should obviously be
+ // improved in the future, probably by having the TargetMachines register
+ // themselves with the runtime, and then have them choose themselves if they
+ // match the current machine.
+ //
+ //===----------------------------------------------------------------------===//
+ 
+ #include "JIT.h"
+ #include "llvm/Module.h"
+ #include "llvm/ModuleProvider.h"
+ #include "llvm/Target/TargetMachine.h"
+ #include "llvm/Target/TargetMachineImpls.h"
+ #include "Support/CommandLine.h"
+ using namespace llvm;
+ 
+ #if !defined(ENABLE_X86_JIT) && !defined(ENABLE_SPARC_JIT)
+ #define NO_JITS_ENABLED
+ #endif
+ 
+ namespace {
+   enum ArchName { x86, Sparc };
+ 
+ #ifndef NO_JITS_ENABLED
+   cl::opt
+   Arch("march", cl::desc("Architecture to JIT to:"), cl::Prefix,
+        cl::values(
+ #ifdef ENABLE_X86_JIT
+                   clEnumVal(x86, "  IA-32 (Pentium and above)"),
+ #endif
+ #ifdef ENABLE_SPARC_JIT
+                   clEnumValN(Sparc, "sparc", "  Sparc-V9"),
+ #endif
+                   0),
+ #if defined(ENABLE_X86_JIT)
+   cl::init(x86)
+ #elif defined(ENABLE_SPARC_JIT)
+   cl::init(Sparc)
+ #endif
+        );
+ #endif /* NO_JITS_ENABLED */
+ }
+ 
+ /// create - Create an return a new JIT compiler if there is one available
+ /// for the current target.  Otherwise, return null.
+ ///
+ ExecutionEngine *JIT::create(ModuleProvider *MP) {
+   TargetMachine* (*TargetMachineAllocator)(const Module &) = 0;
+ 
+   // Allow a command-line switch to override what *should* be the default target
+   // machine for this platform. This allows for debugging a Sparc JIT on X86 --
+   // our X86 machines are much faster at recompiling LLVM and linking LLI.
+ #ifndef NO_JITS_ENABLED
+ 
+   switch (Arch) {
+ #ifdef ENABLE_X86_JIT
+   case x86:
+     TargetMachineAllocator = allocateX86TargetMachine;
+     break;
+ #endif
+ #ifdef ENABLE_SPARC_JIT
+   case Sparc:
+     TargetMachineAllocator = allocateSparcTargetMachine;
+     break;
+ #endif
+   default:
+     assert(0 && "-march flag not supported on this host!");
+   }
+ #else
+   return 0;
+ #endif
+ 
+   // Allocate a target...
+   TargetMachine *Target = TargetMachineAllocator(*MP->getModule());
+   assert(Target && "Could not allocate target machine!");
+ 
+   // If the target supports JIT code generation, return a new JIT now.
+   if (TargetJITInfo *TJ = Target->getJITInfo())
+     return new JIT(MP, *Target, *TJ);
+   return 0;
+ }
+ 
+ 


Index: llvm/lib/ExecutionEngine/JIT/Emitter.cpp
diff -u llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.36 llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.37
--- llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.36	Mon Dec  8 02:06:28 2003
+++ llvm/lib/ExecutionEngine/JIT/Emitter.cpp	Fri Dec 19 19:46:27 2003
@@ -16,7 +16,7 @@
 #ifndef _POSIX_MAPPED_FILES
 #define _POSIX_MAPPED_FILES
 #endif
-#include "VM.h"
+#include "JIT.h"
 #include "llvm/Constant.h"
 #include "llvm/Module.h"
 #include "llvm/CodeGen/MachineCodeEmitter.h"
@@ -31,7 +31,7 @@
 
 namespace {
   Statistic<> NumBytes("jit", "Number of bytes of machine code compiled");
-  VM *TheVM = 0;
+  JIT *TheJIT = 0;
 
   /// JITMemoryManager - Manage memory for the JIT code generation in a logical,
   /// sane way.  This splits a large block of MAP_NORESERVE'd memory into two
@@ -142,7 +142,7 @@
     // constant pool.
     std::vector ConstantPoolAddresses;
   public:
-    Emitter(VM &vm) { TheVM = &vm; }
+    Emitter(JIT &jit) { TheJIT = &jit; }
 
     virtual void startFunction(MachineFunction &F);
     virtual void finishFunction(MachineFunction &F);
@@ -166,13 +166,13 @@
   };
 }
 
-MachineCodeEmitter *VM::createEmitter(VM &V) {
-  return new Emitter(V);
+MachineCodeEmitter *JIT::createEmitter(JIT &jit) {
+  return new Emitter(jit);
 }
 
 void Emitter::startFunction(MachineFunction &F) {
   CurByte = CurBlock = MemMgr.startFunctionBody();
-  TheVM->addGlobalMapping(F.getFunction(), CurBlock);
+  TheJIT->addGlobalMapping(F.getFunction(), CurBlock);
 }
 
 void Emitter::finishFunction(MachineFunction &F) {
@@ -197,8 +197,8 @@
   unsigned TotalSize = 0;
   for (unsigned i = 0, e = Constants.size(); i != e; ++i) {
     const Type *Ty = Constants[i]->getType();
-    unsigned Size      = TheVM->getTargetData().getTypeSize(Ty);
-    unsigned Alignment = TheVM->getTargetData().getTypeAlignment(Ty);
+    unsigned Size      = TheJIT->getTargetData().getTypeSize(Ty);
+    unsigned Alignment = TheJIT->getTargetData().getTypeAlignment(Ty);
     // Make sure to take into account the alignment requirements of the type.
     TotalSize = (TotalSize + Alignment-1) & ~(Alignment-1);
 
@@ -213,7 +213,7 @@
   // Actually output all of the constants, and remember their addresses.
   for (unsigned i = 0, e = Constants.size(); i != e; ++i) {
     void *Addr = Pool + ConstantOffset[i];
-    TheVM->InitializeMemory(Constants[i], Addr);
+    TheJIT->InitializeMemory(Constants[i], Addr);
     ConstantPoolAddresses.push_back(Addr);
   }
 }
@@ -248,10 +248,10 @@
 uint64_t Emitter::getGlobalValueAddress(GlobalValue *V) {
   // Try looking up the function to see if it is already compiled, if not return
   // 0.
-  return (intptr_t)TheVM->getPointerToGlobalIfAvailable(V);
+  return (intptr_t)TheJIT->getPointerToGlobalIfAvailable(V);
 }
 uint64_t Emitter::getGlobalValueAddress(const std::string &Name) {
-  return (intptr_t)TheVM->getPointerToNamedFunction(Name);
+  return (intptr_t)TheJIT->getPointerToNamedFunction(Name);
 }
 
 // getConstantPoolEntryAddress - Return the address of the 'ConstantNum' entry
@@ -272,19 +272,19 @@
 }
 
 uint64_t Emitter::forceCompilationOf(Function *F) {
-  return (intptr_t)TheVM->getPointerToFunction(F);
+  return (intptr_t)TheJIT->getPointerToFunction(F);
 }
 
 // getPointerToNamedFunction - This function is used as a global wrapper to
-// VM::getPointerToNamedFunction for the purpose of resolving symbols when
+// JIT::getPointerToNamedFunction for the purpose of resolving symbols when
 // bugpoint is debugging the JIT. In that scenario, we are loading an .so and
 // need to resolve function(s) that are being mis-codegenerated, so we need to
 // resolve their addresses at runtime, and this is the way to do it.
 extern "C" {
   void *getPointerToNamedFunction(const char *Name) {
-    Module &M = TheVM->getModule();
+    Module &M = TheJIT->getModule();
     if (Function *F = M.getNamedFunction(Name))
-      return TheVM->getPointerToFunction(F);
-    return TheVM->getPointerToNamedFunction(Name);
+      return TheJIT->getPointerToFunction(F);
+    return TheJIT->getPointerToNamedFunction(Name);
   }
 }


Index: llvm/lib/ExecutionEngine/JIT/Intercept.cpp
diff -u llvm/lib/ExecutionEngine/JIT/Intercept.cpp:1.11 llvm/lib/ExecutionEngine/JIT/Intercept.cpp:1.12
--- llvm/lib/ExecutionEngine/JIT/Intercept.cpp:1.11	Mon Dec  8 02:06:28 2003
+++ llvm/lib/ExecutionEngine/JIT/Intercept.cpp	Fri Dec 19 19:46:27 2003
@@ -15,7 +15,7 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "VM.h"
+#include "JIT.h"
 #include "Support/DynamicLinker.h"
 #include 
 using namespace llvm;
@@ -28,7 +28,7 @@
 /// calls to atexit(3), which we intercept and store in
 /// AtExitHandlers.
 ///
-void VM::runAtExitHandlers() {
+void JIT::runAtExitHandlers() {
   while (!AtExitHandlers.empty()) {
     void (*Fn)() = AtExitHandlers.back();
     AtExitHandlers.pop_back();
@@ -45,7 +45,7 @@
 
 // jit_exit - Used to intercept the "exit" library call.
 static void jit_exit(int Status) {
-  VM::runAtExitHandlers();   // Run atexit handlers...
+  JIT::runAtExitHandlers();   // Run atexit handlers...
   exit(Status);
 }
 
@@ -61,7 +61,7 @@
 /// function by using the dynamic loader interface.  As such it is only useful 
 /// for resolving library symbols, not code generated symbols.
 ///
-void *VM::getPointerToNamedFunction(const std::string &Name) {
+void *JIT::getPointerToNamedFunction(const std::string &Name) {
   // Check to see if this is one of the functions we want to intercept...
   if (Name == "exit") return (void*)&jit_exit;
   if (Name == "atexit") return (void*)&jit_atexit;


Index: llvm/lib/ExecutionEngine/JIT/JIT.cpp
diff -u llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.27 llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.28
--- llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.27	Fri Dec 19 19:22:19 2003
+++ llvm/lib/ExecutionEngine/JIT/JIT.cpp	Fri Dec 19 19:46:27 2003
@@ -1,4 +1,4 @@
-//===-- JIT.cpp - LLVM Just-In-Time Compiler ------------------------------===//
+//===-- JIT.cpp - LLVM Just in Time Compiler ------------------------------===//
 // 
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,102 +7,50 @@
 // 
 //===----------------------------------------------------------------------===//
 //
-// This file implements the top-level support for creating a Just-In-Time
-// compiler for the current architecture.
+// This tool implements a just-in-time compiler for LLVM, allowing direct
+// execution of LLVM bytecode in an efficient manner.
 //
 //===----------------------------------------------------------------------===//
 
-#include "VM.h"
-#include "llvm/Module.h"
+#include "JIT.h"
+#include "llvm/Function.h"
 #include "llvm/ModuleProvider.h"
+#include "llvm/CodeGen/MachineCodeEmitter.h"
+#include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/ExecutionEngine/GenericValue.h"
 #include "llvm/Target/TargetMachine.h"
-#include "llvm/Target/TargetMachineImpls.h"
-#include "Support/CommandLine.h"
+#include "llvm/Target/TargetJITInfo.h"
 using namespace llvm;
 
-#if !defined(ENABLE_X86_JIT) && !defined(ENABLE_SPARC_JIT)
-#define NO_JITS_ENABLED
-#endif
-
-namespace {
-  enum ArchName { x86, Sparc };
-
-#ifndef NO_JITS_ENABLED
-  cl::opt
-  Arch("march", cl::desc("Architecture to JIT to:"), cl::Prefix,
-       cl::values(
-#ifdef ENABLE_X86_JIT
-                  clEnumVal(x86, "  IA-32 (Pentium and above)"),
-#endif
-#ifdef ENABLE_SPARC_JIT
-                  clEnumValN(Sparc, "sparc", "  Sparc-V9"),
-#endif
-                  0),
-#if defined(ENABLE_X86_JIT)
-  cl::init(x86)
-#elif defined(ENABLE_SPARC_JIT)
-  cl::init(Sparc)
-#endif
-       );
-#endif /* NO_JITS_ENABLED */
-}
-
-/// create - Create an return a new JIT compiler if there is one available
-/// for the current target.  Otherwise, return null.
-///
-ExecutionEngine *VM::create(ModuleProvider *MP) {
-  TargetMachine* (*TargetMachineAllocator)(const Module &) = 0;
-
-  // Allow a command-line switch to override what *should* be the default target
-  // machine for this platform. This allows for debugging a Sparc JIT on X86 --
-  // our X86 machines are much faster at recompiling LLVM and linking LLI.
-#ifndef NO_JITS_ENABLED
-
-  switch (Arch) {
-#ifdef ENABLE_X86_JIT
-  case x86:
-    TargetMachineAllocator = allocateX86TargetMachine;
-    break;
-#endif
-#ifdef ENABLE_SPARC_JIT
-  case Sparc:
-    TargetMachineAllocator = allocateSparcTargetMachine;
-    break;
-#endif
-  default:
-    assert(0 && "-march flag not supported on this host!");
-  }
-#else
-  return 0;
-#endif
-
-  // Allocate a target...
-  TargetMachine *Target = TargetMachineAllocator(*MP->getModule());
-  assert(Target && "Could not allocate target machine!");
-
-  // If the target supports JIT code generation, return a new JIT now.
-  if (TargetJITInfo *TJ = Target->getJITInfo())
-    return new VM(MP, *Target, *TJ);
-  return 0;
-}
-
-VM::VM(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji)
+JIT::JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji)
   : ExecutionEngine(MP), TM(tm), TJI(tji), PM(MP) {
   setTargetData(TM.getTargetData());
 
   // Initialize MCE
   MCE = createEmitter(*this);
   
-  setupPassManager();
+  // Compile LLVM Code down to machine code in the intermediate representation
+  TJI.addPassesToJITCompile(PM);
+
+  // Turn the machine code intermediate representation into bytes in memory that
+  // may be executed.
+  if (TM.addPassesToEmitMachineCode(PM, *MCE)) {
+    std::cerr << "lli: target '" << TM.getName()
+              << "' doesn't support machine code emission!\n";
+    abort();
+  }
 
   emitGlobals();
 }
 
+JIT::~JIT() {
+  delete MCE;
+  delete &TM;
+}
+
 /// run - Start execution with the specified function and arguments.
 ///
-GenericValue VM::run(Function *F, const std::vector &ArgValues)
-{
+GenericValue JIT::run(Function *F, const std::vector &ArgValues) {
   assert (F && "Function *F was null at entry to run()");
 
   int (*PF)(int, char **, const char **) =
@@ -119,4 +67,75 @@
   GenericValue rv;
   rv.IntVal = ExitCode;
   return rv;
+}
+
+/// runJITOnFunction - Run the FunctionPassManager full of
+/// just-in-time compilation passes on F, hopefully filling in
+/// GlobalAddress[F] with the address of F's machine code.
+///
+void JIT::runJITOnFunction(Function *F) {
+  static bool isAlreadyCodeGenerating = false;
+  assert(!isAlreadyCodeGenerating && "Error: Recursive compilation detected!");
+
+  // JIT the function
+  isAlreadyCodeGenerating = true;
+  PM.run(*F);
+  isAlreadyCodeGenerating = false;
+}
+
+/// getPointerToFunction - This method is used to get the address of the
+/// specified function, compiling it if neccesary.
+///
+void *JIT::getPointerToFunction(Function *F) {
+  void *&Addr = GlobalAddress[F];   // Check if function already code gen'd
+  if (Addr) return Addr;
+
+  // Make sure we read in the function if it exists in this Module
+  MP->materializeFunction(F);
+
+  if (F->isExternal())
+    return Addr = getPointerToNamedFunction(F->getName());
+
+  runJITOnFunction(F);
+  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
+  return Addr;
+}
+
+// getPointerToFunctionOrStub - If the specified function has been
+// code-gen'd, return a pointer to the function.  If not, compile it, or use
+// a stub to implement lazy compilation if available.
+//
+void *JIT::getPointerToFunctionOrStub(Function *F) {
+  // If we have already code generated the function, just return the address.
+  std::map::iterator I = GlobalAddress.find(F);
+  if (I != GlobalAddress.end()) return I->second;
+
+  // If the target supports "stubs" for functions, get a stub now.
+  if (void *Ptr = TJI.getJITStubForFunction(F, *MCE))
+    return Ptr;
+
+  // Otherwise, if the target doesn't support it, just codegen the function.
+  return getPointerToFunction(F);
+}
+
+/// recompileAndRelinkFunction - This method is used to force a function
+/// which has already been compiled, to be compiled again, possibly
+/// after it has been modified. Then the entry to the old copy is overwritten
+/// with a branch to the new copy. If there was no old copy, this acts
+/// just like JIT::getPointerToFunction().
+///
+void *JIT::recompileAndRelinkFunction(Function *F) {
+  void *&Addr = GlobalAddress[F];   // Check if function already code gen'd
+
+  // If it's not already compiled (this is kind of weird) there is no
+  // reason to patch it up.
+  if (!Addr) { return getPointerToFunction (F); }
+
+  void *OldAddr = Addr;
+  Addr = 0;
+  MachineFunction::destruct(F);
+  runJITOnFunction(F);
+  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
+  TJI.replaceMachineCodeForFunction(OldAddr, Addr);
+  return Addr;
 }


Index: llvm/lib/ExecutionEngine/JIT/JIT.h
diff -u llvm/lib/ExecutionEngine/JIT/JIT.h:1.19 llvm/lib/ExecutionEngine/JIT/JIT.h:1.20
--- llvm/lib/ExecutionEngine/JIT/JIT.h:1.19	Fri Dec 19 19:22:19 2003
+++ llvm/lib/ExecutionEngine/JIT/JIT.h	Fri Dec 19 19:46:27 2003
@@ -1,4 +1,4 @@
-//===-- VM.h - Definitions for Virtual Machine ------------------*- C++ -*-===//
+//===-- JIT.h - Class definition for the JIT --------------------*- C++ -*-===//
 // 
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,12 +7,12 @@
 // 
 //===----------------------------------------------------------------------===//
 //
-// This file defines the top-level Virtual Machine data structure.
+// This file defines the top-level JIT data structure.
 //
 //===----------------------------------------------------------------------===//
 
-#ifndef VM_H
-#define VM_H
+#ifndef JIT_H
+#define JIT_H
 
 #include "llvm/ExecutionEngine/ExecutionEngine.h"
 #include "llvm/PassManager.h"
@@ -27,16 +27,16 @@
 class TargetJITInfo;
 class MachineCodeEmitter;
 
-class VM : public ExecutionEngine {
+class JIT : public ExecutionEngine {
   TargetMachine &TM;       // The current target we are compiling to
   TargetJITInfo &TJI;      // The JITInfo for the target we are compiling to
   
   FunctionPassManager PM;  // Passes to compile a function
   MachineCodeEmitter *MCE; // MCE object
 
-  VM(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji);
+  JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji);
 public:
-  ~VM();
+  ~JIT();
 
   /// create - Create an return a new JIT compiler if there is one available
   /// for the current target.  Otherwise, return null.
@@ -79,13 +79,12 @@
   /// which has already been compiled, to be compiled again, possibly
   /// after it has been modified. Then the entry to the old copy is overwritten
   /// with a branch to the new copy. If there was no old copy, this acts
-  /// just like VM::getPointerToFunction().
+  /// just like JIT::getPointerToFunction().
   ///
   void *recompileAndRelinkFunction(Function *F);
 
 private:
-  static MachineCodeEmitter *createEmitter(VM &V);
-  void setupPassManager();
+  static MachineCodeEmitter *createEmitter(JIT &J);
   void runJITOnFunction (Function *F);
 };
 




From lattner at cs.uiuc.edu  Fri Dec 19 20:04:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 20:04:01 2003
Subject: [llvm-commits] CVS: llvm/lib/Target/X86/X86CodeEmitter.cpp
Message-ID: <200312200203.UAA28188@zion.cs.uiuc.edu>


Changes in directory llvm/lib/Target/X86:

X86CodeEmitter.cpp updated: 1.43 -> 1.44

---
Log message:

A minor cleanup for better encapsulation


---
Diffs of the changes:  (+13 -12)

Index: llvm/lib/Target/X86/X86CodeEmitter.cpp
diff -u llvm/lib/Target/X86/X86CodeEmitter.cpp:1.43 llvm/lib/Target/X86/X86CodeEmitter.cpp:1.44
--- llvm/lib/Target/X86/X86CodeEmitter.cpp:1.43	Fri Dec 19 19:22:04 2003
+++ llvm/lib/Target/X86/X86CodeEmitter.cpp	Fri Dec 19 20:03:14 2003
@@ -50,13 +50,17 @@
     unsigned resolveFunctionReference(unsigned RetAddr);
   };
 
-  JITResolver *TheJITResolver;
+  static JITResolver &getResolver(MachineCodeEmitter &MCE) {
+    static JITResolver *TheJITResolver = 0;
+    if (TheJITResolver == 0)
+      TheJITResolver = new JITResolver(MCE);
+    return *TheJITResolver;
+  }
 }
 
+
 void *X86JITInfo::getJITStubForFunction(Function *F, MachineCodeEmitter &MCE) {
-  if (TheJITResolver == 0)
-    TheJITResolver = new JITResolver(MCE);
-  return (void*)((unsigned long)TheJITResolver->getLazyResolver(F));
+  return (void*)((unsigned long)getResolver(MCE).getLazyResolver(F));
 }
 
 void X86JITInfo::replaceMachineCodeForFunction (void *Old, void *New) {
@@ -126,7 +130,8 @@
   // Sanity check to make sure this really is a call instruction...
   assert(((unsigned char*)(intptr_t)RetAddr)[-1] == 0xE8 &&"Not a call instr!");
   
-  unsigned NewVal = TheJITResolver->resolveFunctionReference(RetAddr);
+  JITResolver &JR = getResolver(*(MachineCodeEmitter*)0);
+  unsigned NewVal = JR.resolveFunctionReference(RetAddr);
 
   // Rewrite the call target... so that we don't fault every time we execute
   // the call.
@@ -267,10 +272,8 @@
   
   if (Address == 0) {
     // FIXME: this is JIT specific!
-    if (TheJITResolver == 0)
-      TheJITResolver = new JITResolver(MCE);
-    Address = TheJITResolver->addFunctionReference(MCE.getCurrentPCValue(),
-                                                   cast(GV));
+    Address = getResolver(MCE).addFunctionReference(MCE.getCurrentPCValue(),
+                                                    cast(GV));
   }
   emitMaybePCRelativeValue(Address, true);
 }
@@ -288,9 +291,7 @@
   //
   if (Address == 0) {
     // FIXME: this is JIT specific!
-    if (TheJITResolver == 0)
-      TheJITResolver = new JITResolver(MCE);
-    Address = TheJITResolver->getLazyResolver((Function*)GV);
+    Address = getResolver(MCE).getLazyResolver((Function*)GV);
   }
 
   emitMaybePCRelativeValue(Address, false);




From lattner at cs.uiuc.edu  Fri Dec 19 20:46:02 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 20:46:02 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/ExecutionEngine/ExecutionEngine.h
Message-ID: <200312200245.UAA32317@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/ExecutionEngine:

ExecutionEngine.h updated: 1.20 -> 1.21

---
Log message:

Simple refactorings to prepare for lazy global emission


---
Diffs of the changes:  (+10 -3)

Index: llvm/include/llvm/ExecutionEngine/ExecutionEngine.h
diff -u llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.20 llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.21
--- llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.20	Fri Dec 12 00:31:42 2003
+++ llvm/include/llvm/ExecutionEngine/ExecutionEngine.h	Fri Dec 19 20:45:16 2003
@@ -21,9 +21,10 @@
 
 namespace llvm {
 
+union GenericValue;
 class Constant;
 class Function;
-union GenericValue;
+class GlobalVariable;
 class GlobalValue;
 class Module;
 class ModuleProvider;
@@ -60,8 +61,8 @@
 
   static ExecutionEngine *create(ModuleProvider *MP, bool ForceInterpreter);
 
-  void addGlobalMapping(const Function *F, void *Addr) {
-    void *&CurVal = GlobalAddress[(const GlobalValue*)F];
+  void addGlobalMapping(const GlobalValue *GV, void *Addr) {
+    void *&CurVal = GlobalAddress[GV];
     assert(CurVal == 0 && "GlobalMapping already established!");
     CurVal = Addr;
   }
@@ -107,6 +108,12 @@
 
 protected:
   void emitGlobals();
+
+  // EmitGlobalVariable - This method emits the specified global variable to the
+  // address specified in GlobalAddresses, or allocates new memory if it's not
+  // already in the map.
+  void EmitGlobalVariable(GlobalVariable *GV);
+
   GenericValue getConstantValue(const Constant *C);
   GenericValue LoadValueFromMemory(GenericValue *Ptr, const Type *Ty);
 };




From lattner at cs.uiuc.edu  Fri Dec 19 20:46:11 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 20:46:11 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
Message-ID: <200312200245.UAA32327@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine:

ExecutionEngine.cpp updated: 1.40 -> 1.41

---
Log message:

Simple refactorings to prepare for lazy global emission
Also, add a stat for the number of globals emitted


---
Diffs of the changes:  (+16 -2)

Index: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
diff -u llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.40 llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.41
--- llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.40	Fri Dec 19 19:45:17 2003
+++ llvm/lib/ExecutionEngine/ExecutionEngine.cpp	Fri Dec 19 20:45:37 2003
@@ -30,6 +30,7 @@
 
 namespace {
   Statistic<> NumInitBytes("lli", "Number of bytes of global vars initialized");
+  Statistic<> NumGlobals  ("lli", "Number of global vars initialized");
 }
 
 ExecutionEngine::ExecutionEngine(ModuleProvider *P) : 
@@ -372,7 +373,7 @@
       
       // Allocate some memory for it!
       unsigned Size = TD.getTypeSize(Ty);
-      GlobalAddress[I] = new char[Size];
+      addGlobalMapping(I, new char[Size]);
       NumInitBytes += Size;
 
       DEBUG(std::cerr << "Global '" << I->getName() << "' -> "
@@ -394,5 +395,18 @@
   for (Module::giterator I = getModule().gbegin(), E = getModule().gend();
        I != E; ++I)
     if (!I->isExternal())
-      InitializeMemory(I->getInitializer(), GlobalAddress[I]);
+      EmitGlobalVariable(I);
+}
+
+// EmitGlobalVariable - This method emits the specified global variable to the
+// address specified in GlobalAddresses, or allocates new memory if it's not
+// already in the map.
+void ExecutionEngine::EmitGlobalVariable(GlobalVariable *GV) {
+  void *&GA = GlobalAddress[GV];
+  if (GA == 0) {
+    // If it's not already specified, allocate memory for the global.
+    GA = new char[getTargetData().getTypeSize(GV->getType()->getElementType())];
+  }
+  InitializeMemory(GV->getInitializer(), GA);
+  ++NumGlobals;
 }




From lattner at cs.uiuc.edu  Fri Dec 19 21:37:01 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 21:37:01 2003
Subject: [llvm-commits] CVS: llvm/include/llvm/ExecutionEngine/ExecutionEngine.h
Message-ID: <200312200336.VAA10317@zion.cs.uiuc.edu>


Changes in directory llvm/include/llvm/ExecutionEngine:

ExecutionEngine.h updated: 1.21 -> 1.22

---
Log message:

Cleanups to implement PR135


---
Diffs of the changes:  (+11 -4)

Index: llvm/include/llvm/ExecutionEngine/ExecutionEngine.h
diff -u llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.21 llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.22
--- llvm/include/llvm/ExecutionEngine/ExecutionEngine.h:1.21	Fri Dec 19 20:45:16 2003
+++ llvm/include/llvm/ExecutionEngine/ExecutionEngine.h	Fri Dec 19 21:35:50 2003
@@ -35,11 +35,11 @@
   Module &CurMod;
   const TargetData *TD;
 
-protected:
-  ModuleProvider *MP;
   // GlobalAddress - A mapping between LLVM global values and their actualized
   // version...
   std::map GlobalAddress;
+protected:
+  ModuleProvider *MP;
 
   void setTargetData(const TargetData &td) {
     TD = &td;
@@ -63,7 +63,7 @@
 
   void addGlobalMapping(const GlobalValue *GV, void *Addr) {
     void *&CurVal = GlobalAddress[GV];
-    assert(CurVal == 0 && "GlobalMapping already established!");
+    assert((CurVal == 0 || Addr == 0) && "GlobalMapping already established!");
     CurVal = Addr;
   }
 
@@ -106,13 +106,20 @@
   ///
   virtual void *recompileAndRelinkFunction(Function *F) = 0;
 
+  /// getOrEmitGlobalVariable - Return the address of the specified global
+  /// variable, possibly emitting it to memory if needed.  This is used by the
+  /// Emitter.
+  virtual void *getOrEmitGlobalVariable(const GlobalVariable *GV) {
+    return getPointerToGlobal((GlobalValue*)GV);
+  }
+
 protected:
   void emitGlobals();
 
   // EmitGlobalVariable - This method emits the specified global variable to the
   // address specified in GlobalAddresses, or allocates new memory if it's not
   // already in the map.
-  void EmitGlobalVariable(GlobalVariable *GV);
+  void EmitGlobalVariable(const GlobalVariable *GV);
 
   GenericValue getConstantValue(const Constant *C);
   GenericValue LoadValueFromMemory(GenericValue *Ptr, const Type *Ty);




From lattner at cs.uiuc.edu  Fri Dec 19 21:37:11 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 21:37:11 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/JIT/Emitter.cpp JIT.cpp JIT.h
Message-ID: <200312200336.VAA10444@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine/JIT:

Emitter.cpp updated: 1.37 -> 1.38
JIT.cpp updated: 1.28 -> 1.29
JIT.h updated: 1.20 -> 1.21

---
Log message:

Implement PR135, lazy emission of global variables



---
Diffs of the changes:  (+78 -15)

Index: llvm/lib/ExecutionEngine/JIT/Emitter.cpp
diff -u llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.37 llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.38
--- llvm/lib/ExecutionEngine/JIT/Emitter.cpp:1.37	Fri Dec 19 19:46:27 2003
+++ llvm/lib/ExecutionEngine/JIT/Emitter.cpp	Fri Dec 19 21:36:47 2003
@@ -248,7 +248,11 @@
 uint64_t Emitter::getGlobalValueAddress(GlobalValue *V) {
   // Try looking up the function to see if it is already compiled, if not return
   // 0.
-  return (intptr_t)TheJIT->getPointerToGlobalIfAvailable(V);
+  if (isa(V))
+    return (intptr_t)TheJIT->getPointerToGlobalIfAvailable(V);
+  else {
+    return (intptr_t)TheJIT->getOrEmitGlobalVariable(cast(V));
+  }
 }
 uint64_t Emitter::getGlobalValueAddress(const std::string &Name) {
   return (intptr_t)TheJIT->getPointerToNamedFunction(Name);


Index: llvm/lib/ExecutionEngine/JIT/JIT.cpp
diff -u llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.28 llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.29
--- llvm/lib/ExecutionEngine/JIT/JIT.cpp:1.28	Fri Dec 19 19:46:27 2003
+++ llvm/lib/ExecutionEngine/JIT/JIT.cpp	Fri Dec 19 21:36:47 2003
@@ -13,13 +13,16 @@
 //===----------------------------------------------------------------------===//
 
 #include "JIT.h"
+#include "llvm/DerivedTypes.h"
 #include "llvm/Function.h"
+#include "llvm/GlobalVariable.h"
 #include "llvm/ModuleProvider.h"
 #include "llvm/CodeGen/MachineCodeEmitter.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/ExecutionEngine/GenericValue.h"
 #include "llvm/Target/TargetMachine.h"
 #include "llvm/Target/TargetJITInfo.h"
+#include "Support/DynamicLinker.h"
 using namespace llvm;
 
 JIT::JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji)
@@ -39,8 +42,6 @@
               << "' doesn't support machine code emission!\n";
     abort();
   }
-
-  emitGlobals();
 }
 
 JIT::~JIT() {
@@ -81,22 +82,36 @@
   isAlreadyCodeGenerating = true;
   PM.run(*F);
   isAlreadyCodeGenerating = false;
+
+  // If the function referred to a global variable that had not yet been
+  // emitted, it allocates memory for the global, but doesn't emit it yet.  Emit
+  // all of these globals now.
+  while (!PendingGlobals.empty()) {
+    const GlobalVariable *GV = PendingGlobals.back();
+    PendingGlobals.pop_back();
+    EmitGlobalVariable(GV);
+  }
 }
 
 /// getPointerToFunction - This method is used to get the address of the
 /// specified function, compiling it if neccesary.
 ///
 void *JIT::getPointerToFunction(Function *F) {
-  void *&Addr = GlobalAddress[F];   // Check if function already code gen'd
-  if (Addr) return Addr;
+  if (void *Addr = getPointerToGlobalIfAvailable(F))
+    return Addr;   // Check if function already code gen'd
 
   // Make sure we read in the function if it exists in this Module
   MP->materializeFunction(F);
 
-  if (F->isExternal())
-    return Addr = getPointerToNamedFunction(F->getName());
+  if (F->isExternal()) {
+    void *Addr = getPointerToNamedFunction(F->getName());
+    addGlobalMapping(F, Addr);
+    return Addr;
+  }
 
   runJITOnFunction(F);
+
+  void *Addr = getPointerToGlobalIfAvailable(F);
   assert(Addr && "Code generation didn't add function to GlobalAddress table!");
   return Addr;
 }
@@ -107,8 +122,8 @@
 //
 void *JIT::getPointerToFunctionOrStub(Function *F) {
   // If we have already code generated the function, just return the address.
-  std::map::iterator I = GlobalAddress.find(F);
-  if (I != GlobalAddress.end()) return I->second;
+  if (void *Addr = getPointerToGlobalIfAvailable(F))
+    return Addr;
 
   // If the target supports "stubs" for functions, get a stub now.
   if (void *Ptr = TJI.getJITStubForFunction(F, *MCE))
@@ -118,6 +133,33 @@
   return getPointerToFunction(F);
 }
 
+/// getOrEmitGlobalVariable - Return the address of the specified global
+/// variable, possibly emitting it to memory if needed.  This is used by the
+/// Emitter.
+void *JIT::getOrEmitGlobalVariable(const GlobalVariable *GV) {
+  void *Ptr = getPointerToGlobalIfAvailable(GV);
+  if (Ptr) return Ptr;
+
+  // If the global is external, just remember the address.
+  if (GV->isExternal()) {
+    Ptr = GetAddressOfSymbol(GV->getName().c_str());
+    if (Ptr == 0) {
+      std::cerr << "Could not resolve external global address: "
+                << GV->getName() << "\n";
+      abort();
+    }
+  } else {
+    // If the global hasn't been emitted to memory yet, allocate space.  We will
+    // actually initialize the global after current function has finished
+    // compilation.
+    Ptr =new char[getTargetData().getTypeSize(GV->getType()->getElementType())];
+    PendingGlobals.push_back(GV);
+  }
+  addGlobalMapping(GV, Ptr);
+  return Ptr;
+}
+
+
 /// recompileAndRelinkFunction - This method is used to force a function
 /// which has already been compiled, to be compiled again, possibly
 /// after it has been modified. Then the entry to the old copy is overwritten
@@ -125,16 +167,23 @@
 /// just like JIT::getPointerToFunction().
 ///
 void *JIT::recompileAndRelinkFunction(Function *F) {
-  void *&Addr = GlobalAddress[F];   // Check if function already code gen'd
+  void *OldAddr = getPointerToGlobalIfAvailable(F);
 
-  // If it's not already compiled (this is kind of weird) there is no
-  // reason to patch it up.
-  if (!Addr) { return getPointerToFunction (F); }
+  // If it's not already compiled there is no reason to patch it up.
+  if (OldAddr == 0) { return getPointerToFunction(F); }
 
-  void *OldAddr = Addr;
-  Addr = 0;
+  // Delete the old function mapping.
+  addGlobalMapping(F, 0);
+
+  // Destroy the machine code for this function.  FIXME: this should be
+  // incorporated into the code generator!
   MachineFunction::destruct(F);
+
+  // Recodegen the function
   runJITOnFunction(F);
+
+  // Update state, forward the old function to the new function.
+  void *Addr = getPointerToGlobalIfAvailable(F);
   assert(Addr && "Code generation didn't add function to GlobalAddress table!");
   TJI.replaceMachineCodeForFunction(OldAddr, Addr);
   return Addr;


Index: llvm/lib/ExecutionEngine/JIT/JIT.h
diff -u llvm/lib/ExecutionEngine/JIT/JIT.h:1.20 llvm/lib/ExecutionEngine/JIT/JIT.h:1.21
--- llvm/lib/ExecutionEngine/JIT/JIT.h:1.20	Fri Dec 19 19:46:27 2003
+++ llvm/lib/ExecutionEngine/JIT/JIT.h	Fri Dec 19 21:36:47 2003
@@ -34,6 +34,11 @@
   FunctionPassManager PM;  // Passes to compile a function
   MachineCodeEmitter *MCE; // MCE object
 
+  /// PendingGlobals - Global variables which have had memory allocated for them
+  /// while a function was code generated, but which have not been initialized
+  /// yet.
+  std::vector PendingGlobals;
+
   JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji);
 public:
   ~JIT();
@@ -68,6 +73,11 @@
   /// compiling it if necessary.
   ///
   void *getPointerToFunction(Function *F);
+
+  /// getOrEmitGlobalVariable - Return the address of the specified global
+  /// variable, possibly emitting it to memory if needed.  This is used by the
+  /// Emitter.
+  void *getOrEmitGlobalVariable(const GlobalVariable *GV);
 
   /// getPointerToFunctionOrStub - If the specified function has been
   /// code-gen'd, return a pointer to the function.  If not, compile it, or use




From lattner at cs.uiuc.edu  Fri Dec 19 21:38:00 2003
From: lattner at cs.uiuc.edu (Chris Lattner)
Date: Fri Dec 19 21:38:00 2003
Subject: [llvm-commits] CVS: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
Message-ID: <200312200337.VAA10450@zion.cs.uiuc.edu>


Changes in directory llvm/lib/ExecutionEngine:

ExecutionEngine.cpp updated: 1.41 -> 1.42

---
Log message:

Implement PR135, lazy emission of global variables


---
Diffs of the changes:  (+7 -4)

Index: llvm/lib/ExecutionEngine/ExecutionEngine.cpp
diff -u llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.41 llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.42
--- llvm/lib/ExecutionEngine/ExecutionEngine.cpp:1.41	Fri Dec 19 20:45:37 2003
+++ llvm/lib/ExecutionEngine/ExecutionEngine.cpp	Fri Dec 19 21:36:25 2003
@@ -155,7 +155,8 @@
           const_cast(dyn_cast(CPR->getValue())))
         Result = PTOGV(getPointerToFunctionOrStub(F));
       else 
-        Result = PTOGV(getPointerToGlobal(CPR->getValue()));
+        Result = PTOGV(getOrEmitGlobalVariable(
+                           cast(CPR->getValue())));
 
     } else {
       assert(0 && "Unknown constant pointer type!");
@@ -374,7 +375,6 @@
       // Allocate some memory for it!
       unsigned

First Name (required):
Last Name (required):
Email Address (required):
Organizational Affiliation:
Position/Title:

Program	Status	Notes
llvm	1.1	15 Dec 2003	WORKS	Just did some simple touch testing after 'make tools'. Things seem to work, but I didn't try running the testsuite or anything. +
hbd	0.2.3	26 Nov 2003	26 Nov 2003	WORKS	Did not try the install target.
FALSE	FALSE	-- b	The boolean value FALSE (0) is pushed onto the stack.	The boolean value FALSE (0) is pushed on to the stack.
TRUE	TRUE	-- b	The boolean value TRUE (-1) is pushed onto the stack.	The boolean value TRUE (-1) is pushed on to the stack.
BITWISE OPERATIONS
Word	Name	Operation	Description
ABS	w -- \|w\|	One value s popped off the stack; its absolute value is computed - and then pushed onto the stack. If w1 is -1 then w2 is 1. If w1 is + and then pushed on to the stack. If w1 is -1 then w2 is 1. If w1 is 1 then w2 is also 1.
NEG	NEG	w -- -w	One value is popped off the stack which is negated and then - pushed back onto the stack. If w1 is -1 then w2 is 1. If w1 is + pushed back on to the stack. If w1 is -1 then w2 is 1. If w1 is 1 then w2 is -1.
+	ADD	w1 w2 -- w2+w1	Two values are popped off the stack. Their sum is pushed back - onto the stack
-	SUB	w1 w2 -- w2-w1	Two values are popped off the stack. Their difference is pushed back - onto the stack
*	MUL	w1 w2 -- w2*w1	Two values are popped off the stack. Their product is pushed back - onto the stack
/	DIV	w1 w2 -- w2/w1	Two values are popped off the stack. Their quotient is pushed back - onto the stack
MOD	MOD	w1 w2 -- w2%w1	Two values are popped off the stack. Their remainder after division - of w1 by w2 is pushed back onto the stack
*/	STAR_SLAH	w1 w2 w3 -- (w3*w2)/w1	Three values are popped off the stack. The product of w1 and w2 is - divided by w3. The result is pushed back onto the stack.
++	INCR	w -- w+1	One value is popped off the stack. It is incremented by one and then - pushed back onto the stack.
--	DECR	w -- w-1	One value is popped off the stack. It is decremented by one and then - pushed back onto the stack.
MIN	MIN	w1 w2 -- (w2<w1?w2:w1)	Two values are popped off the stack. The larger one is pushed back - onto the stack.
MAX	MAX	w1 w2 -- (w2>w1?w2:w1)	Two values are popped off the stack. The larger value is pushed back - onto the stack.
STACK MANIPULATION OPERATIONS
Word	Name	Operation	Description
DUP	DUP	w1 -- w1 w1	One value is popped off the stack. That value is then pushed onto - the stack twice to duplicate the top stack vaue.	One value is popped off the stack. That value is then pushed on to + the stack twice to duplicate the top stack value.
DUP2	DUP2	w1 w2 -- w1 w2 w1 w2	The top two values on the stack are duplicated. That is, two vaues +	The top two values on the stack are duplicated. That is, two values are popped off the stack. They are alternately pushed back on the stack twice each.
SWAP	w1 w2 -- w2 w1	The top two stack items are reversed in their order. That is, two - values are popped off the stack and pushed back onto the stack in + values are popped off the stack and pushed back on to the stack in the opposite order they were popped.
SWAP2	w1 w2 w3 w4 -- w3 w4 w2 w1	The top four stack items are swapped in pairs. That is, two values are popped and retained. Then, two more values are popped and retained. - The values are pushed back onto the stack in the reverse order but + The values are pushed back on to the stack in the reverse order but in pairs.
OVER	OVER	w1 w2-- w1 w2 w1	Two values are popped from the stack. They are pushed back - onto the stack in the order w1 w2 w1. This seems to cause the + on to the stack in the order w1 w2 w1. This seems to cause the top stack element to be duplicated "over" the next value.
OVER2	OVER2	w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2	The third and fourth values on the stack are replicated onto the +	The third and fourth values on the stack are replicated on to the top of the stack
ROT	ROT	w1 w2 w3 -- w2 w3 w1	The top three values are rotated. That is, three value are popped - off the stack. They are pushed back onto the stack in the order + off the stack. They are pushed back on to the stack in the order w1 w3 w2.
ROT2	One value is popped off the stack. The value is used as the size of a memory block to allocate. The size is in bytes, not words. The memory allocation is completed and the address of the memory - block is pushed onto the stack.
FREE	FREE	The boolean value on the top of the stack is examined. If it is non-zero then the "words..." between WHILE and END are executed. Execution then begins again at the WHILE where another boolean is popped off the stack. To prevent this operation from eating up the entire - stack, you should push onto the stack (just before the END) a boolean value that indicates + stack, you should push on to the stack (just before the END) a boolean value that indicates whether to terminate. Note that since booleans and integers can be coerced you can use the following "for loop" idiom: `(push count) WHILE (words...) -- END` @@ -960,19 +960,19 @@	IN_STR	-- s	A string is read from the input via the scanf(3) format string " %as". The - resulting string is pushed onto the stack.
<d	IN_STR	-- w	An integer is read from the input via the scanf(3) format string " %d". The - resulting value is pushed onto the stack
<c	IN_CHR	-- w	A single character is read from the input via the scanf(3) format string - " %c". The value is converted to an integer and pushed onto the stack.
DUMP	DUMP

') - print ('Error in form:' + error) + print ('Error in form: ' + error) print ('

') + print ('Error in form: ' + error) + print ('

Sorry; not registered

Download: