From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 00:25:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 00:25:49 -0500
Subject: [LLVMbugs] [Bug 4283] Clang's preprocessor doesn' t support
 multibyte character constants
In-Reply-To: <bug-4283-206@http.llvm.org/bugs/>
Message-ID: <200906010525.n515Pngd019766@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4283


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-01 00:25:48 ---
Fixed in r72686.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 03:29:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 03:29:22 -0500
Subject: [LLVMbugs] [Bug 4289] New: Crash when using K&R-style function
	pointers
Message-ID: <bug-4289-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4289

           Summary: Crash when using K&R-style function pointers
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu





-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 04:16:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 04:16:15 -0500
Subject: [LLVMbugs] [Bug 4286] Assertion failed: (use_empty() && " Uses
 remain when a value is destroyed!"), function ~Value
In-Reply-To: <bug-4286-206@http.llvm.org/bugs/>
Message-ID: <200906010916.n519GFtg005877@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4286


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-01 04:16:00 ---
Fixed in r72688.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 04:28:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 04:28:10 -0500
Subject: [LLVMbugs] [Bug 4287] Varargs + K&R doesn't work nicely
In-Reply-To: <bug-4287-206@http.llvm.org/bugs/>
Message-ID: <200906010928.n519SAnR006445@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4287


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-01 04:27:50 ---
Meh, may as well call it legal. Fixed in r72689.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 05:05:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 05:05:09 -0500
Subject: [LLVMbugs] [Bug 4289] Crash when using function pointers on
	K&R-style functions
In-Reply-To: <bug-4289-206@http.llvm.org/bugs/>
Message-ID: <200906011005.n51A59Va009388@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4289


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-01 05:04:58 ---
Fixed in r72690.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 08:16:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 08:16:15 -0500
Subject: [LLVMbugs] [Bug 4290] New: err_implicit_decl_requires_stdio too
 strict to make autoconf work.
Message-ID: <bug-4290-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4290

           Summary: err_implicit_decl_requires_stdio too strict to make
                    autoconf work.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


The following code doesn't build with Clang:

char vfprintf();

test.c:1:6: error: implicit declaration of 'vfprintf' requires inclusion of the
header <stdio.h>

This means that autoconf is unable to detect it properly and apps miscompile.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 12:39:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 12:39:35 -0500
Subject: [LLVMbugs] [Bug 4291] New: don't emit ".cpu generic" on arm linux
Message-ID: <bug-4291-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4291

           Summary: don't emit ".cpu generic" on arm linux
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


The ".cpu" directive doesn't seem to be documented as being supported in
binutils as, but it works anyways (seemingly supporting the same syntax as the
-mcpu=... flag) which means that ".cpu generic" is illegal:

x.s: Assembler messages:
x.s:2: Error: unknown cpu `generic'

For examples of legal .cpu values, you could see
http://sourceware.org/binutils/docs/as/ARM-Options.html but specifying "arm7"
for example causes this:

x.s: Assembler messages:
x.s:25: Error: selected processor does not support `bx lr'

Just removing the .cpu line produces a binary that actually runs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 13:46:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 13:46:46 -0500
Subject: [LLVMbugs] [Bug 4293] New: PIC is broken on ppc32-linux
Message-ID: <bug-4293-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4293

           Summary: PIC is broken on ppc32-linux
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: PowerPC
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3049)
 --> (http://llvm.org/bugs/attachment.cgi?id=3049)
testcase

This is the source of the one "make check" failure on ppc32-linux. Bugpoint
runs llc -relocation-model=pic and that breaks the reduction of
test/BugPoint/misopt-basictest.ll.

$ Debug/bin/llc -reloaction-model=pic x.bc -f -o x.s
$ gcc x.s -o x.exe
x.s: Assembler messages:
x.s:14: Error: bad expression
x.s:14: Error: syntax error; found `L' but expected `,'
x.s:14: Error: junk at end of line: `L1$pb"@ha'
x.s:14: Error: bad expression
x.s:14: Error: syntax error; found `L' but expected `('
x.s:14: Error: junk at end of line: `L1$pb"@l(3)'

The problem is these:
  lis 3, .LC0-"L1$pb"@ha
  la 3, .LC0-"L1$pb"@l(3)
which probably would work fine if only "L1$pb" were defined anywhere in the .s
file.

By contrast the .s generated by gcc -fPIC looks like:

[...]
  .section ".got2","aw"
.LCTOC1 = .+32768
  .section ".text"
  .section ".got2","aw"
.LC1:
  .long .LC0
[... later on in main]
  addis 30,30,.LCTOC1-.LCF0 at ha
  addi 30,30,.LCTOC1-.LCF0 at l
  lwz 3,.LC1-.LCTOC1(30)
  bl puts+32768 at plt
[... def'n of .LC0 same as llvm's .s]

or with gcc -fpic:

[... no def'n of .LCTOC1 or .LC1]
  addis 30,30,_GLOBAL_OFFSET_TABLE_-.LCF0 at ha
  addi 30,30,_GLOBAL_OFFSET_TABLE_-.LCF0 at l
  lwz 3,.LC0 at got(30)
  bl puts at plt
[... def'n of .LC0 below]


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 14:03:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 14:03:55 -0500
Subject: [LLVMbugs] [Bug 4291] don't emit ".cpu generic" on arm linux
In-Reply-To: <bug-4291-206@http.llvm.org/bugs/>
Message-ID: <200906011903.n51J3tiP029871@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4291


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-06-01 14:03:55 ---
Fixed in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078232.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 20:05:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 20:05:51 -0500
Subject: [LLVMbugs] [Bug 3517] poor handling of extern inline function
In-Reply-To: <bug-3517-206@http.llvm.org/bugs/>
Message-ID: <200906020105.n5215paj009757@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3517


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #10 from Dale Johannesen <dalej at apple.com>  2009-06-01 20:05:51 ---
Fixed in llvm-gcc by fix for 4262, right Duncan?  Extern inline handling looks
OK to me.



-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  1 22:47:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 1 Jun 2009 22:47:04 -0500
Subject: [LLVMbugs] [Bug 137] [llvmg++] Code is generated for empty classes
In-Reply-To: <bug-137-206@http.llvm.org/bugs/>
Message-ID: <200906020347.n523l4xk014659@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=137


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #9 from Chris Lattner <clattner at apple.com>  2009-06-01 22:47:03 ---
We still generate useless code, but I think this is a failing of g++, and we're
not going to fix it.  For example, at -O0 this generates a useless i8 store:

struct x {};
void foo(x *P) {
  *P = x();
}

Closing as fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 02:55:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 02:55:55 -0500
Subject: [LLVMbugs] [Bug 4288] Mixed up # line directives with -E -dD
In-Reply-To: <bug-4288-206@http.llvm.org/bugs/>
Message-ID: <200906020755.n527ttTY003087@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4288


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-02 02:55:55 ---
Fixed in r72724.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 03:16:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 03:16:57 -0500
Subject: [LLVMbugs] [Bug 4296] New: Add parser detection/error recovery for
 nested functions
Message-ID: <bug-4296-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4296

           Summary: Add parser detection/error recovery for nested functions
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Even if clang never ends up supporting nested functions, it would be nice if we
could detect what's going on and give an easily recognizable error message. 
Currently, it gives "expected ';' at end of declaration", which isn't obvious
without going in and looking at the file in question.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 07:03:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 07:03:56 -0500
Subject: [LLVMbugs] [Bug 4297] New: llc: Assertion `castIsValid(getOpcode(),
 S, Ty) && "  Illegal PtrToInt"' failed.
Message-ID: <bug-4297-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4297

           Summary: llc: Assertion `castIsValid(getOpcode(), S, Ty) &&
                    "Illegal PtrToInt"' failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


I'm getting an assertion failure from llc that doesn't seem to be caused by any
instruction in my code (it complains about ptrtoint, which isn't in there).

This is the assertion:

$ llc bugpoint-reduced-simplified.bc 
llc: Instructions.cpp:2463: llvm::PtrToIntInst::PtrToIntInst(llvm::Value*,
const llvm::Type*, const std::string&, llvm::Instruction*): Assertion
`castIsValid(getOpcode(), S, Ty) && "Illegal PtrToInt"' failed.
0   llc             0x0000000000d63cdf
1   llc             0x0000000000d640e9
2   libpthread.so.0 0x00007f5bba7700f0
3   libc.so.6       0x00007f5bb986c015 gsignal + 53
4   libc.so.6       0x00007f5bb986db83 abort + 387
5   libc.so.6       0x00007f5bb9864d89 __assert_fail + 233
6   llc             0x0000000000ce38e8
7   llc             0x0000000000be83ce
8   llc             0x0000000000beb728
9   llc             0x0000000000cfd109
llvm::FPPassManager::runOnFunction(llvm::Function&) + 489
10  llc             0x0000000000cfdb26
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 166
11  llc             0x0000000000cfdd6b
llvm::FunctionPassManager::run(llvm::Function&) + 75
12  llc             0x0000000000516c09 main + 4649
13  libc.so.6       0x00007f5bb9857466 __libc_start_main + 230
14  llc             0x0000000000514df9
Stack dump:
0.      Program arguments: llc bugpoint-reduced-simplified.bc 
1.      Running pass 'Optimize for code generation' on function '@_Dmain'
Aborted


And this is the code:

$ llvm-dis <bugpoint-reduced-simplified.bc 
; ModuleID = '<stdin>'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
target triple = "x86_64-unknown-linux-gnu"
        %"byte[]" = type { i64, i8* }
        %"char[][]" = type { i64, %"byte[]"* }
@.str = external constant [7 x i8]              ; <[7 x i8]*> [#uses=1]

define fastcc i32 @_Dmain(%"char[][]" %unnamed) {
entry:
        %tmp = getelementptr [7 x i8]* @.str, i32 0, i32 0              ; <i8*>
[#uses=1]
        br i1 undef, label %foreachbody, label %foreachend

foreachbody:            ; preds = %entry
        %tmp4 = getelementptr i8* %tmp, i32 undef               ; <i8*>
[#uses=1]
        %tmp5 = load i8* %tmp4          ; <i8> [#uses=0]
        unreachable

foreachend:             ; preds = %entry
        ret i32 0
}


(I'm using r72443, if that matters)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 14:33:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 14:33:39 -0500
Subject: [LLVMbugs] [Bug 4298] New: Strange position-independent code,
 relocatable references problem.
Message-ID: <bug-4298-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4298

           Summary: Strange position-independent code, relocatable
                    references problem.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Hi,

While trying to build the latest (At revision 72733.) svn trunk on AuroraUX (A
OpenSolaris derived distribution; Same result on SXCE snv_110), We hit a
strange position-independent code, relocatable references problem that has been
exceedingly hard to debug so far.

"ld: fatal: relocations remain against allocatable but non-writable sections"

Details of the failed build here:
http://pkg.auroraux.org/llvm_fail.txt

Configure Info:
./configure --prefix=/export/home2/edward/lab/TEST_BUILDS/llvm/test1
--enable-optimized --enable-assertions --enable-expensive-checks

Moving /usr/bin/ld aside 'fixes' the problem. Looking at a diff of the
config.log by moving /usr/bin/ld aside --export-dynamic gets set to 'yes' and
comes into effect, not sure of how much relevance's, although a very
interesting observation.

Tinkering around we figured that something may not have been getting built with
-fPIC somewhere somehow..
So we past --enable-pic=no at the end of the configure line seems to work
around this issue!

Details of the pass build (--enable-pic=no) here:
http://pkg.auroraux.org/llvm_pass_disable-pic.txt

The only other *idea* I could come up with is that there is hand-written
assembler code that has not been coded with the appropriate
position-independent prototypes.

I been trying very hard to try to fix this myself, however, am at a bit of a
loss, any pointers would be greatly appreciated.

System LD Info:
$ /usr/bin/ld --version
ld: Software Generation Utilities - Solaris Link Editors: 5.11-1.1652

$ /usr/sfw/bin/gld --version
GNU ld version 2.15
<snip>

Stock Solaris GCC Info:
$ /usr/sfw/bin/gcc --version
gcc (GCC) 3.4.3 (csl-sol210-3_4-20050802)
<snip>

PATH setup:
$ echo $PATH
/usr/sfw/bin:/usr/gnu/bin:/opt/SUNWspro/bin:/usr/xpg6/bin:/usr/xpg4/bin:/usr/bin:/usr/sbin:/usr/X11/bin:/usr/ccs/bin

Thanks for your time,
Edward O'Callaghan.
-
http://www.auroraux.org/


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 15:05:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 15:05:47 -0500
Subject: [LLVMbugs] [Bug 4299] New: clang doesn' t apply __weak consistently
 to pointer dereferences
Message-ID: <bug-4299-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4299

           Summary: clang doesn't apply __weak consistently to pointer
                    dereferences
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase (compile with "clang-cc -triple i686-apple-darwin -x objective-c
-fobjc-gc -emit-llvm"):
__weak id* g1;
id* __weak g2;
id f1() { return *g1; }
id f2() { return *g2; }

Per the documentation, the compiled versions of both functions should contain
two calls to objc_read_weak, but that isn't true in clang: f2 only contains one
call to objc_read_weak.  (I haven't checked against a Darwin gcc myself, but
Fariborz says that gcc generates the right thing here.)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  2 16:29:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 2 Jun 2009 16:29:46 -0500
Subject: [LLVMbugs] [Bug 4297] llc: Assertion `castIsValid(getOpcode(), S,
 Ty) && "  Illegal PtrToInt"' failed.
In-Reply-To: <bug-4297-206@http.llvm.org/bugs/>
Message-ID: <200906022129.n52LTk67032290@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4297


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dan Gohman <gohman at apple.com>  2009-06-02 16:29:45 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078280.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 04:05:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 04:05:10 -0500
Subject: [LLVMbugs] [Bug 4225] Failed assertion: "Wrong MachineOperand
 accessor" in setIsKill()
In-Reply-To: <bug-4225-206@http.llvm.org/bugs/>
Message-ID: <200906030905.n5395AXT001227@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4225


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Evan Cheng <evan.cheng at apple.com>  2009-06-03 04:04:57 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078295.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 04:09:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 04:09:43 -0500
Subject: [LLVMbugs] [Bug 4303] New: llvm.dbg.declare vs -instcombine
Message-ID: <bug-4303-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4303

           Summary: llvm.dbg.declare vs -instcombine
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


I get a verification failure when I run this test case through -instcombine:

$ cat d.ll
define void @f() {
        %a = alloca {}
        call void @llvm.dbg.declare({}* %a, {}* null)
        ret void
}
declare void @llvm.dbg.declare({}*, {}*)
$ llvm-as -o - d.ll | opt -f -o /dev/null -instcombine
invalid llvm.dbg.declare intrinsic call
        call void @llvm.dbg.declare({ }* null, { }* null)
Broken module found, compilation aborted!
0   opt 0x0844ccc8
Stack dump:
0.      Running pass 'Function Pass Manager' on module '<stdin>'.
1.      Running pass 'Module Verifier' on function '@f'
Aborted

The problem is that -instcombine replaces the alloca with null:

  if (isa<AllocaInst>(AI) && AI.getAllocatedType()->isSized()) {
    // If alloca'ing a zero byte object, replace the alloca with a null
pointer.
    // Note that we only do this for alloca's, because malloc should allocate
    // and return a unique pointer, even for a zero byte allocation.
    if (TD->getTypeAllocSize(AI.getAllocatedType()) == 0)
      return ReplaceInstUsesWith(AI, Constant::getNullValue(AI.getType()));

    // If the alignment is 0 (unspecified), assign it the preferred alignment.
    if (AI.getAlignment() == 0)
      AI.setAlignment(TD->getPrefTypeAlignment(AI.getAllocatedType()));
  }

but -verify asserts that the first arguments to llvm.dbg.declare is non-null:

  case Intrinsic::dbg_declare:  // llvm.dbg.declare
    if (Constant *C = dyn_cast<Constant>(CI.getOperand(1)))
      Assert1(C && !isa<ConstantPointerNull>(C),
              "invalid llvm.dbg.declare intrinsic call", &CI);
    break; 

I don't think -verify should be making ad-hoc checks about debug intrinsics,
because it places too high a burden on optimisation passes to make sure they
don't break these checks. So I would vote for removing that check.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 04:55:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 04:55:36 -0500
Subject: [LLVMbugs] [Bug 4290] err_implicit_decl_requires_stdio too strict
 to make autoconf work.
In-Reply-To: <bug-4290-206@http.llvm.org/bugs/>
Message-ID: <200906030955.n539ta70003361@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4290


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-03 04:55:30 ---
Fixed in r72760.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 05:45:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 05:45:46 -0500
Subject: [LLVMbugs] [Bug 3678] Invalid input/output constraint when
 compiling FreeBSD's libm/ OpenSSL on AMD64
In-Reply-To: <bug-3678-206@http.llvm.org/bugs/>
Message-ID: <200906031045.n53AjkVn005314@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3678


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|REOPENED                    |RESOLVED
         Resolution|                            |INVALID




--- Comment #19 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-03 05:45:35 ---
As far as I can tell, the correct constraint to use here is actually "x".  If
that doesn't work for some reason, please reopen.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 07:38:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 07:38:25 -0500
Subject: [LLVMbugs] [Bug 4304] New: Missing pedantic warning for use of
 designators in c89 mode
Message-ID: <bug-4304-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4304

           Summary: Missing pedantic warning for use of designators in c89
                    mode
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Simple testcase (run with clang -std=c89 -pedantic):
int a[5] = { [2] = 1 };

Should warn in C89 mode because C89 didn't have designators.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 07:41:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 07:41:49 -0500
Subject: [LLVMbugs] [Bug 4305] New: clang missing pedantic warning for use
 of universal character name in c89 mode
Message-ID: <bug-4305-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4305

           Summary: clang missing pedantic warning for use of universal
                    character name in c89 mode
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase (clang -std=c89 -pedantic):
char* x = "\u1234";

Should give a warning because universal character names are not valid in C89.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 07:46:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 07:46:12 -0500
Subject: [LLVMbugs] [Bug 4306] New: clang missing pedantic warning for
 non-constant initializer for aggregate in c89 mode
Message-ID: <bug-4306-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4306

           Summary: clang missing pedantic warning for non-constant
                    initializer for aggregate in c89 mode
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase (clang -c89 -pedantic):
struct x {int x;};
void a(int y) { struct x x = {y}; }

This should warn because it's a constraint violation in C89.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 07:50:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 07:50:04 -0500
Subject: [LLVMbugs] [Bug 4307] New: clang missing pedantic warning for use
 of flexible array member in c89 mode
Message-ID: <bug-4307-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4307

           Summary: clang missing pedantic warning for use of flexible array
                    member in c89 mode
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase (clang -pedantic -std=c89):
struct x { int x,y[]; };

Should warn because this feature does not exist in C89.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 08:26:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 08:26:53 -0500
Subject: [LLVMbugs] [Bug 4308] New: clang: add support for -m32
Message-ID: <bug-4308-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4308

           Summary: clang: add support for -m32
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


This is needed to build the ACPI/realmode code for a 64-bit Linux kernel.

Currently clang does this:
clang: warning: argument unused during compilation: '-m32'
/tmp/cc-j5tdcX.s: Assembler messages:
/tmp/cc-j5tdcX.s:40: Error: bad register name `%r15'
/tmp/cc-j5tdcX.s:41: Error: bad register name `%r14'
/tmp/cc-j5tdcX.s:42: Error: bad register name `%rbx'
/tmp/cc-j5tdcX.s:43: Error: bad register name `%rip)'
/tmp/cc-j5tdcX.s:46: Error: bad register name `%rip)'
/tmp/cc-j5tdcX.s:51: Error: bad register name `%r14d'
/tmp/cc-j5tdcX.s:52: Error: bad register name `%r15d'
/tmp/cc-j5tdcX.s:55: Error: bad register name `%rbx)'
/tmp/cc-j5tdcX.s:56: Error: bad register name `%rbx'
/tmp/cc-j5tdcX.s:57: Error: bad register name `%rbx'
/tmp/cc-j5tdcX.s:65: Error: bad register name `%r15d'
/tmp/cc-j5tdcX.s:87: Error: bad register name `%r14d'
/tmp/cc-j5tdcX.s:109: Error: bad register name `%rip)'
/tmp/cc-j5tdcX.s:120: Error: bad register name `%rip)'
/tmp/cc-j5tdcX.s:126: Error: bad register name `%rip)'
/tmp/cc-j5tdcX.s:131: Error: bad register name `%rbx'
/tmp/cc-j5tdcX.s:132: Error: bad register name `%r14'
/tmp/cc-j5tdcX.s:133: Error: bad register name `%r15'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 10:57:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 10:57:01 -0500
Subject: [LLVMbugs] [Bug 4311] New: Null-pointer dereference in
 Traits::getNext(NodePtr) when buildt optimized.
Message-ID: <bug-4311-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4311

           Summary: Null-pointer dereference in Traits::getNext(NodePtr)
                    when buildt optimized.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: SunOS
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Test Case:
-bash-3.2$ cat > example.c
#include <stdio.h>

int main() {
printf("Hi.\n");
return 0;
}

Trace:
-bash-3.2$ ./test1/bin/clang example.c 
Assertion failed: Traits::getNext(NodePtr) != 0 && "Dereferencing end()!", file
/export/home2/edward/lab/TEST_BUILDS/llvm/llvm/include/llvm/ADT/ilist.h, line
197
0   clang-cc 0x09b37eee PrintStackTrace(void*) + 22
1   clang-cc 0x09e45df0 PrintStackTrace(void*) + 3202840
2   clang-cc 0000000000 PrintStackTrace(void*) + 4132208936
3   clang-cc 0x6834ec83 PrintStackTrace(void*) + 1585540523
Stack dump:
0.      Program arguments:
/export/home2/edward/lab/TEST_BUILDS/llvm/test1/bin/../libexec/clang-cc -triple
i386-pc-solaris2.11 -S -disable-free -main-file-name example.c
--relocation-model static --disable-fp-elim --unwind-tables=0 --mcpu=pentium4
--fmath-errno=1 -fdiagnostics-show-option -o /tmp/cc-Uoaivy.s -x c example.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@main'

Configure Line:
./configure --prefix=/export/home2/edward/lab/TEST_BUILDS/llvm/test1
--enable-optimized --enable-assertions --enable-expensive-checks
--enable-pic=no

Building with, --enable-optimized=no , instead allows one to again build a
Hello world test app.

However tests still fail as can be seen in the following log:
http://pkg.auroraux.org/gmake_check_non-opti.log
Core File:
http://pkg.auroraux.org/core.7z

Anything else be sure to let me know,

Thanks for your time,
Edward O'Callaghan.
-
http://www.auroraux.org/


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 11:27:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 11:27:44 -0500
Subject: [LLVMbugs] [Bug 4311] Null-pointer dereference in
 Traits::getNext(NodePtr) when buildt optimized.
In-Reply-To: <bug-4311-206@http.llvm.org/bugs/>
Message-ID: <200906031627.n53GRip7017221@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4311


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #4 from Duncan Sands <baldrick at free.fr>  2009-06-03 11:27:44 ---
I don't think there is anything to see here.  Various versions of gcc 3.4.x
have proven to miscompile LLVM in the past, see
  http://llvm.org/docs/GettingStarted.html#brokengcc
So this is presumably more of the same, and there's not much we can do about
it.
The testsuite problems should be moved to another bug report.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 11:31:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 11:31:42 -0500
Subject: [LLVMbugs] [Bug 4313] New: tablegen crashes when compiled with LTO
Message-ID: <bug-4313-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4313

           Summary: tablegen crashes when compiled with LTO
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


If I compile LLVM with LTO on Linux, by using instructions from here
http://llvm.org/docs/GoldPlugin.html#lto_autotools, and additionally doing
make ENABLE_OPTIMIZED=1 OPTIMIZE_OPTION=-O4, then tablegen crashes with heap
corruption.

OPTIMIZE_OPTION=-O3 works fine, valgrind is clean; with OPTIMIZE_OPTION=-O4
valgrind reports use-after-free.

I can reproduce this with llvm-ld too, -disable-opt works fine, defaults create
bugged tblgen.

Here's how I've run bugpoint:
bugpoint -run-llc -safe-run-llc -internalize -ipsccp -globalopt -constmerge
-deadargelim -instcombine -basiccg -inline -prune-eh -globalopt -globaldce
-basiccg -argpromotion -instcombine -jump-threading -domtree -domfrontier
-scalarrepl -basiccg -functionattrs -globalsmodref-aa -domtree -loops
-loopsimplify -domfrontier -licm -memdep -gvn -memdep -memcpyopt -dse
-instcombine -jump-threading -domtree -domfrontier -mem2reg -simplifycfg
-globaldce -instcombine -simplifycfg -adce -globaldce -preverify -domtree
-verify /home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/AsmWriterEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/CallingConvEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/ClangDiagnosticsEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/CodeEmitterGen.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/CodeGenDAGPatterns.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/CodeGenInstruction.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/CodeGenTarget.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/DAGISelEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/FastISelEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/InstrEnumEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/InstrInfoEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/IntrinsicEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/LLVMCConfigurationEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/Record.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/RegisterInfoEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/SubtargetEmitter.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TGLexer.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TGParser.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TGSourceMgr.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TGValueTypes.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TableGen.o
/home/edwin/llvm-svn/llvm-lto/utils/TableGen/Release/TableGenBackend.o
-Xlinker=-lLLVMSupport -Xlinker=-lLLVMSystem -Xlinker=/usr/lib/libstdc++.so.6
-Xlinker=-lpthread -Xlinker=-ldl --args -I /home/edwin/llvm-svn/llvm/lib/VMCore
-I /home/edwin/llvm-svn/llvm/include -I /home/edwin/llvm-svn/llvm/include -I
/home/edwin/llvm-svn/llvm/lib/Target
/home/edwin/llvm-svn/llvm/include/llvm/Intrinsics.td -o
/home/edwin/llvm-svn/llvm-lto/lib/VMCore/Release/Intrinsics.gen.tmp
-gen-intrinsic

I haven't got a reduced .bc file yet, will attach when I have one.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 12:22:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 12:22:47 -0500
Subject: [LLVMbugs] [Bug 4314] New: llvm-ld's LTO segfaults in '
 Interprocedural Sparse Conditional Constant Propagation'
Message-ID: <bug-4314-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4314

           Summary: llvm-ld's LTO segfaults in 'Interprocedural Sparse
                    Conditional Constant Propagation'
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: matti.niemenmaa+llvmbugs at iki.fi
                CC: llvmbugs at cs.uiuc.edu


Assembling and linking the following:


%T = type { i32 }

define void @foo() {
        call %T @bar()
        ret void
}

define internal %T @bar() {
        %x = alloca %T
        %y = load %T* %x
        ret %T %y
}


Results in the following:


$ llvm-as arst.ll -f && llvm-ld arst.bc             
0   llvm-ld         0x00000000006e41df
1   llvm-ld         0x00000000006e45cd
2   libpthread.so.0 0x00007f3b9be93850
3   llvm-ld         0x000000000053d3c8
4   llvm-ld         0x00000000005406cb
5   llvm-ld         0x0000000000541b98
6   llvm-ld         0x0000000000542edd
7   llvm-ld         0x000000000067fb10
llvm::MPPassManager::runOnModule(llvm::Module&) + 304
8   llvm-ld         0x000000000068049b
llvm::PassManagerImpl::run(llvm::Module&) + 171
9   llvm-ld         0x000000000048886f llvm::Optimize(llvm::Module*) + 1231
10  llvm-ld         0x000000000048f810 main + 1120
11  libc.so.6       0x00007f3b9af8e9ed __libc_start_main + 253
12  llvm-ld         0x0000000000487ab9
Stack dump:
0.      Program arguments: llvm-ld arst.bc 
1.      Running pass 'Interprocedural Sparse Conditional Constant Propagation'
on module 'a.out'.

With -disable-opt it works fine.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 12:23:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 12:23:15 -0500
Subject: [LLVMbugs] [Bug 4315] New: Assertion failed: " Operands of AddRec
 must be loop-invariant!"
Message-ID: <bug-4315-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4315

           Summary: Assertion failed: "Operands of AddRec must be loop-
                    invariant!"
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=3052)
 --> (http://llvm.org/bugs/attachment.cgi?id=3052)
Reduced testcase

I'm seeing the following crash when building a FreeBSD kernel (with LLVM+Clang
r72770):

cc -O2 -pipe   llvm-crash.c  -o llvm-crash
Assertion failed: (Operands[i]->isLoopInvariant(l) && "Operands of AddRec must
be loop-invariant!"), function SCEVAddRecExpr, file
/home/ed/projects/freebsd-clang/usr.bin/clang/lib/libllvmanalysis/../../../../contrib/llvm/include/llvm/Analysis/ScalarEvolutionExpressions.h,
line 394.
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free -main-file-name llvm-crash.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O2 -fdiagnostics-show-option -o /tmp/cc-ij5lLq.s -x c
llvm-crash.c 
1.      <eof> parser at end of file
2.      Per-module optimization passes
3.      Running pass 'CallGraph Pass Manager' on module 'llvm-crash.c'.
4.      Running pass 'Loop Pass Manager' on function '@crash'
5.      Running pass 'Canonicalize Induction Variables' on basic block
'%for.cond'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 12:37:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 12:37:22 -0500
Subject: [LLVMbugs] [Bug 4316] New: Assertion failed: "Ptr must be a pointer
 to Val type!"
Message-ID: <bug-4316-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4316

           Summary: Assertion failed: "Ptr must be a pointer to Val type!"
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


I'm seeing the following crash with LLVM+Clang r72770:

cc -O2 -pipe   llvm-crash.c  -o llvm-crash
Assertion failed: (getOperand(0)->getType() ==
cast<PointerType>(getOperand(1)->getType())->getElementType() && "Ptr must be a
pointer to Val type!"), function AssertOK, file
.../llvm/lib/VMCore/Instructions.cpp, line 920.
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free -main-file-name llvm-crash.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O2 -fdiagnostics-show-option -o /tmp/cc-YwucJT.s -x c
llvm-crash.c 
1.      <eof> parser at end of file
2.      llvm-crash.c:7:1: LLVM IR generation of declaration 'crash'
3.      llvm-crash.c:8:1: LLVM IR generation of compound statement ('{}')

Code:

union un {
        int i;
        char *c;
};

void
crash(char *buffer)
{

        boom((union un)buffer);
}

I can't seem to generate any LLVM IR.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 14:00:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 14:00:27 -0500
Subject: [LLVMbugs] [Bug 4308] clang: add support for -m32
In-Reply-To: <bug-4308-206@http.llvm.org/bugs/>
Message-ID: <200906031900.n53J0Rhj023598@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4308


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-03 14:00:26 ---
This was working, but I broke it; fixed in r72794.  Sorry for any
inconvenience.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 14:17:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 14:17:30 -0500
Subject: [LLVMbugs] [Bug 4317] New: llc: Assertion `i != PredBlocks.size()
 && " No reachable preds?"' failed.
Message-ID: <bug-4317-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4317

           Summary: llc: Assertion `i != PredBlocks.size() && "No reachable
                    preds?"' failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3054)
 --> (http://llvm.org/bugs/attachment.cgi?id=3054)
bugpoint-reduced-simplified.bc

While reducing a miscompilation with bugpoint, it has found a crash in llc too:
llc: /home/edwin/llvm-svn/llvm/include/llvm/Analysis/Dominators.h:273: void
llvm::DominatorTreeBase<NodeT>::Split(llvm::DominatorTreeBase<typename
GraphT::NodeType>&, typename GraphT::NodeType*) [with N = llvm::BasicBlock*,
GraphT = llvm::GraphTraits<llvm::BasicBlock*>, NodeT = llvm::BasicBlock]:
Assertion `i != PredBlocks.size() && "No reachable preds?"' failed.
0   llc             0x0000000000d6923f
1   llc             0x0000000000d69639
2   libpthread.so.0 0x00007f8ca26a97b0
3   libc.so.6       0x00007f8ca19c9065 gsignal + 53
4   libc.so.6       0x00007f8ca19cc153 abort + 387
5   libc.so.6       0x00007f8ca19c2159 __assert_fail + 233
6   llc             0x0000000000bbe663 void
llvm::DominatorTreeBase<llvm::BasicBlock>::Split<llvm::BasicBlock*,
llvm::GraphTraits<llvm::BasicBlock*>
>(llvm::DominatorTreeBase<llvm::GraphTraits<llvm::BasicBlock*>::NodeType>&,
llvm::GraphTraits<llvm::BasicBlock*>::NodeType*) +3203
7   llc             0x0000000000bbc4ff
8   llc             0x0000000000bbc986
9   llc             0x0000000000d02059
llvm::FPPassManager::runOnFunction(llvm::Function&) + 489
10  llc             0x0000000000d02316
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 166
11  llc             0x0000000000d0255b
llvm::FunctionPassManager::run(llvm::Function&) + 75
12  llc             0x0000000000518189 main + 4633
13  libc.so.6       0x00007f8ca19b55a6 __libc_start_main + 230
14  llc             0x00000000005163b9
Stack dump:
0.      Program arguments: llc bugpoint-reduced-simplified.bc
1.      Running pass 'Exception handling preparation' on function
'@_ZN4llvm16IntrinsicEmitter28EmitIntrinsicToGCCBuiltinMapERKSt6vectorINS_16CodeGenIntrinsicESaIS2_EERSo'
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 15:14:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 15:14:15 -0500
Subject: [LLVMbugs] [Bug 4315] Assertion failed: "Operands of AddRec must be
 loop-invariant!"
In-Reply-To: <bug-4315-206@http.llvm.org/bugs/>
Message-ID: <200906032014.n53KEFnh026586@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4315


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dan Gohman <gohman at apple.com>  2009-06-03 15:14:15 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078324.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 15:45:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 15:45:23 -0500
Subject: [LLVMbugs] [Bug 4316] Assertion failed: "Ptr must be a pointer to
	Val type!"
In-Reply-To: <bug-4316-206@http.llvm.org/bugs/>
Message-ID: <200906032045.n53KjNRs027791@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4316


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-03 15:45:23 ---
Fixed in r72803.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 16:42:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 16:42:38 -0500
Subject: [LLVMbugs] [Bug 4317] llc: Assertion `i != PredBlocks.size() && "No
 reachable preds?"' failed.
In-Reply-To: <bug-4317-206@http.llvm.org/bugs/>
Message-ID: <200906032142.n53Lgcvq030031@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4317


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-03 16:42:37 ---
Fixed in r72810.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 18:48:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 18:48:28 -0500
Subject: [LLVMbugs] [Bug 4320] New: Support building tests in parallel ( at
 least for Nightly Test)
Message-ID: <bug-4320-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4320

           Summary: Support building tests in parallel (at least for Nightly
                    Test)
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Programs Tests
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Most of the time spent doing a nightly test is building the various projects (I
believe). We should extend the infrastructure so there is a way to only build
the targets we don't care about timing. That way we could do the nightlytest in
two stages, the first could build all the input files (in parallel), and then
the second could do the standard -j1 testing of the programs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun  3 20:22:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 3 Jun 2009 20:22:36 -0500
Subject: [LLVMbugs] [Bug 4321] New: linearscan leaves virtual register in
	MBB livein set
Message-ID: <bug-4321-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4321

           Summary: linearscan leaves virtual register in MBB livein set
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


In the attached testcase compiled with

  llc -march=x86 -print-machineinstrs

basic block bb246 has virtual register %reg1025 in its livein
set after register allocation.

When the post-RA scheduler is enabled (with
-disable-post-RA-scheduler=false), this leads to an assertion
failure because it doesn't expect to find virtual registers
in MBB livein sets.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 01:59:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 01:59:13 -0500
Subject: [LLVMbugs] [Bug 4322] New: llc: Assertion
 `FuncInfo->CatchInfoFound.size() == FuncInfo->CatchInfoLost.size() && " Not
 all catch info was assignedto a landing pad!"' failed.
Message-ID: <bug-4322-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4322

           Summary: llc: Assertion `FuncInfo->CatchInfoFound.size() ==
                    FuncInfo->CatchInfoLost.size() && "Not all catch info
                    was assignedto a landing pad!"' failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4313


Created an attachment (id=3056)
 --> (http://llvm.org/bugs/attachment.cgi?id=3056)
bugpoint-reduced-simplified.bc

While reducing a miscompilation with bugpoint it found a LLC crash:
llc bugpoint-reduced-simplified.bc
llc:
/home/edwin/llvm-svn/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:340:
virtual bool llvm::SelectionDAGISel::runOnFunction(llvm::Function&): Assertion
`FuncInfo->CatchInfoFound.size() == FuncInfo->CatchInfoLost.size() && "Not all
catch info was assignedto a landing pad!"' failed.
0   llc             0x0000000000d6923f
1   llc             0x0000000000d69639
2   libpthread.so.0 0x00007f441694a7b0
3   libc.so.6       0x00007f4415c6a065 gsignal + 53
4   libc.so.6       0x00007f4415c6d153 abort + 387
5   libc.so.6       0x00007f4415c63159 __assert_fail + 233
6   llc             0x0000000000a30a07
llvm::SelectionDAGISel::runOnFunction(llvm::Function&) + 3943
7   llc             0x0000000000d02059
llvm::FPPassManager::runOnFunction(llvm::Function&) + 489
8   llc             0x0000000000d02316
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 166
9   llc             0x0000000000d0255b
llvm::FunctionPassManager::run(llvm::Function&) + 75
10  llc             0x0000000000518189 main + 4633
11  libc.so.6       0x00007f4415c565a6 __libc_start_main + 230
12  llc             0x00000000005163b9
Stack dump:
0.      Program arguments: llc bugpoint-reduced-simplified.bc
1.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@_ZN4llvm16IntrinsicEmitter28EmitIntrinsicToGCCBuiltinMapERKSt6vectorINS_16CodeGenIntrinsicESaIS2_EERSo'
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 02:40:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 02:40:25 -0500
Subject: [LLVMbugs] [Bug 4322] llc: Assertion
 `FuncInfo->CatchInfoFound.size() == FuncInfo-> CatchInfoLost.size() && "Not
 all catch info was assignedto a landing pad!"' failed.
In-Reply-To: <bug-4322-206@http.llvm.org/bugs/>
Message-ID: <200906040740.n547ePq7023502@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4322


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2009-06-04 02:40:24 ---
This is a known limitation of llc: eh.selector calls need to be in
landing pads.  Running bugpoint on code with exception handling
often results in eh.selector calls outside of landing pads, causing
this crash.  This is just one in a wide range of tricky problems with
eh.selector that I am working on.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 08:48:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 08:48:26 -0500
Subject: [LLVMbugs] [Bug 4323] New: Tail recursion elimination doesn' t move
 loads above nounwind readonly call
Message-ID: <bug-4323-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4323

           Summary: Tail recursion elimination doesn't move loads above
                    nounwind readonly call
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3061)
 --> (http://llvm.org/bugs/attachment.cgi?id=3061)
Patch to check CI->mayHaveSideEffects() when looking at a load.

opt -tailcallelim refuses to eliminate a tail call if it's followed by load,
even if the call is both nounwind and readonly.

An example where it would be safe to do this:
----
define fastcc i32 @_D4test3sumFPiiiZi(i32* %a_arg, i32 %a_len_arg, i32
%start_arg) nounwind readonly {
entry:
        %tmp2 = icmp slt i32 %start_arg, %a_len_arg             ; <i1>
[#uses=1]
        br i1 %tmp2, label %else, label %if

if:             ; preds = %entry
        ret i32 0

else:           ; preds = %entry
        %tmp4 = sext i32 %start_arg to i64              ; <i64> [#uses=1]
        %tmp6 = getelementptr i32* %a_arg, i64 %tmp4            ; <i32*>
[#uses=1]
        %tmp10 = add i32 %start_arg, 1          ; <i32> [#uses=1]
        %tmp11 = tail call fastcc i32 @_D4test3sumFPiiiZi(i32* %a_arg, i32
%a_len_arg, i32 %tmp10)              ; <i32> [#uses=1]
        %tmp12 = load i32* %tmp6                ; <i32> [#uses=1]
        %tmp13 = add i32 %tmp12, %tmp11         ; <i32> [#uses=1]
        ret i32 %tmp13
}
----

I'm attaching a small patch that remedies the situation.
Note that I used CI->mayHaveSideEffects() instead of CI->onlyReadsMemory() to
correctly handle the case where the call may throw and the load may introduce
undefined behavior[1]. This way, the transformation does not take place in that
case.


[1]: For example, if the pointer being loaded from may be null but the call
will throw in this case it's not safe to move the load before the call since
that would produce undefined behavior instead of an exception.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 13:17:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 13:17:00 -0500
Subject: [LLVMbugs] [Bug 4324] New: GNU89: extern inline may not be followed
 by regular function
Message-ID: <bug-4324-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4324

           Summary: GNU89: extern inline may not be followed by regular
                    function
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


GCC accepts the following code, while Clang does not. Code is taken from ratbox
ircd.

extern void func(void);

extern inline void
func(void)
{
}

void
func(void)
{
}

$ cc llvm-crash.c -o llvm-crash -std=gnu89         
llvm-crash.c:9:1: error: redefinition of 'func'
func(void)
^
llvm-crash.c:4:1: note: previous definition is here
func(void)
^
2 diagnostics generated.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 13:27:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 13:27:51 -0500
Subject: [LLVMbugs] [Bug 4324] GNU89: extern inline may not be followed by
 regular function
In-Reply-To: <bug-4324-206@http.llvm.org/bugs/>
Message-ID: <200906041827.n54IRpJY021951@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4324


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #2 from Douglas Gregor <dgregor at apple.com>  2009-06-04 13:27:50 ---
We're not going to implement this feature.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 13:55:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 13:55:09 -0500
Subject: [LLVMbugs] [Bug 4325] New: clang: Rev 72572 breaks volatile code
Message-ID: <bug-4325-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4325

           Summary: clang: Rev 72572 breaks volatile code
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: miscompilation, regression
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu, mrs at apple.com


after rev 72572, clang does the "while comparison" on the lvalue instead of the
rvalue in the above c code:

-------------Original C Code:
#define DEBUG_BASE 0x2000
void print(const char* text)
{
  unsigned volatile char * const  debug_device=(unsigned char *
const)DEBUG_BASE;
  while(*debug_device=*text++);

}

-------------Clang Rev 72571 llvm generated code (O3):
define void @print(i8* nocapture %text) nounwind noinline {
entry:
        br label %while.cond

while.cond:             ; preds = %while.cond, %entry
        %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %while.cond ]         
; <i32> [#uses=2]
        %text.addr.0 = getelementptr i8* %text, i32 %indvar             ; <i8*>
[#uses=1]
        %tmp1 = load i8* %text.addr.0           ; <i8> [#uses=2]
        volatile store i8 %tmp1, i8* inttoptr (i32 8192 to i8*), align 8192
        %tobool = icmp eq i8 %tmp1, 0           ; <i1> [#uses=1]
        %indvar.next = add i32 %indvar, 1               ; <i32> [#uses=1]
        br i1 %tobool, label %while.end, label %while.cond

while.end:              ; preds = %while.cond
        ret void
}

-------------diff between 72571 and 72572:
@@ -32,9 +32,10 @@
 while.cond:            ; preds = %while.cond, %entry
        %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %while.cond ]         
; <i32> [#uses=2]
        %text.addr.0 = getelementptr i8* %text, i32 %indvar             ; <i8*>
[#uses=1]
-       %tmp1 = load i8* %text.addr.0           ; <i8> [#uses=2]
+       %tmp1 = load i8* %text.addr.0           ; <i8> [#uses=1]
        volatile store i8 %tmp1, i8* inttoptr (i32 8192 to i8*), align 8192
-       %tobool = icmp eq i8 %tmp1, 0           ; <i1> [#uses=1]
+       %tmp3 = volatile load i8* inttoptr (i32 8192 to i8*), align 8192       
        ; <i8> [#uses=1]
+       %tobool = icmp eq i8 %tmp3, 0           ; <i1> [#uses=1]
        %indvar.next = add i32 %indvar, 1               ; <i32> [#uses=1]
        br i1 %tobool, label %while.end, label %while.cond

-------------Clang Rev 72572 llvm generated code (O3):
define void @print(i8* nocapture %text) nounwind noinline {
entry:
        br label %while.cond

while.cond:             ; preds = %while.cond, %entry
        %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %while.cond ]         
; <i32> [#uses=2]
        %text.addr.0 = getelementptr i8* %text, i32 %indvar             ; <i8*>
[#uses=1]
        %tmp1 = load i8* %text.addr.0           ; <i8> [#uses=1]
        volatile store i8 %tmp1, i8* inttoptr (i32 8192 to i8*), align 8192
        %tmp3 = volatile load i8* inttoptr (i32 8192 to i8*), align 8192       
        ; <i8> [#uses=1]
        %tobool = icmp eq i8 %tmp3, 0           ; <i1> [#uses=1]
        %indvar.next = add i32 %indvar, 1               ; <i32> [#uses=1]
        br i1 %tobool, label %while.end, label %while.cond

while.end:              ; preds = %while.cond
        ret void
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 14:24:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 14:24:58 -0500
Subject: [LLVMbugs] [Bug 4326] New: Crash when initializing variable with
 structure offset
Message-ID: <bug-4326-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4326

           Summary: Crash when initializing variable with structure offset
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


This code causes Clang to crash without any message:

struct f {
        int a;
};

int arr = ((void *)&(((struct f *)0)->a)) - (void *)0;

$ make llvm-crash
cc -O2 -pipe -fno-strict-aliasing  llvm-crash.c  -o llvm-crash
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free -main-file-name llvm-crash.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O2 -fdiagnostics-show-option -o /tmp/cc-sNsmLz.s -x c
llvm-crash.c 
1.      llvm-crash.c:5:54: current parser token ';'
*** Error code 248

Stop in /home/ed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 14:44:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 14:44:14 -0500
Subject: [LLVMbugs] [Bug 4325] clang: Rev 72572 breaks volatile code
In-Reply-To: <bug-4325-206@http.llvm.org/bugs/>
Message-ID: <200906041944.n54JiERW024921@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4325


Mike Stump <mrs at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Mike Stump <mrs at apple.com>  2009-06-04 14:44:14 ---
>From the standard:

  An assignment expression has the value of the left operand after the
assignment

If the left operand is volatile, that means that we must read the operand to
know what the value is.  I aware of the controversy and why you might prefer or
think there should be no read.  The C standard should be improved in this area.
 The C++ standard is perfectly clear however.  I'd recommend pursuing this with
the comp.std.c or the C committee.

As it is, very portable code _must_ not use the value of assignment
expressions, if the lhs is volatile and they do not want it read.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 15:04:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 15:04:22 -0500
Subject: [LLVMbugs] [Bug 4326] Crash when initializing variable with
	structure offset
In-Reply-To: <bug-4326-206@http.llvm.org/bugs/>
Message-ID: <200906042004.n54K4Mus025724@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4326


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-04 15:04:21 ---
Fixed in r72886.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 15:54:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 15:54:27 -0500
Subject: [LLVMbugs] [Bug 4321] linearscan leaves virtual register in MBB
	livein set
In-Reply-To: <bug-4321-206@http.llvm.org/bugs/>
Message-ID: <200906042054.n54KsRZL027506@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4321


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |evan.cheng at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-06-04 15:54:27 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078391.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 16:32:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 16:32:14 -0500
Subject: [LLVMbugs] [Bug 4327] New: Missed optimization on
	Benchmarks/Shootout/objinst
Message-ID: <bug-4327-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4327

           Summary: Missed optimization on Benchmarks/Shootout/objinst
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3063)
 --> (http://llvm.org/bugs/attachment.cgi?id=3063)
test script

clang used to not honor -mllvm -disable-llvm-optzns and so would end up running
-O3 level optimizations twice. It turns out this was giving a very dramatic
speedup on Benchmarks/Shootout/objinst:
--
ddunbar at lordcrumb:x$ cat test.sh
#!/bin/bash

for count in 0 1 2 3 4; do
    name="x$count"
    clang -m32 -O0 -c objinst.c -o $name.bc -emit-llvm
    i=0;
    while (($i < $count)); do
        opt -std-compile-opts $name.bc -o $name.bc -f
        i=$((i + 1));
    done
    llvm-ld $name.bc -o $name.linked -lc
    llc -O3 -f $name.linked.bc -o $name.linked.s
    gcc $name.linked.s -o $name -m32 -fomit-frame-pointer -mdynamic-no-pic -lm 

    echo "-- $count iterations (opt -std-compile-opts) --"
    time ./$name > /dev/null
done
ddunbar at lordcrumb:x$ ./test.sh
-- 0 iterations (opt -std-compile-opts) --

real    0m9.705s
user    0m9.699s
sys     0m0.004s
-- 1 iterations (opt -std-compile-opts) --

real    0m9.566s
user    0m9.560s
sys     0m0.005s
-- 2 iterations (opt -std-compile-opts) --

real    0m4.916s
user    0m4.912s
sys     0m0.003s
-- 3 iterations (opt -std-compile-opts) --

real    0m4.658s
user    0m4.652s
sys     0m0.004s
-- 4 iterations (opt -std-compile-opts) --

real    0m0.028s
user    0m0.026s
sys     0m0.002s
--

The full reduction to < .1s is probably not fair, but it might be worth looking
into the difference between running -O3 once and twice.

Drop the attached test in a directory with objinst.c to run.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun  4 16:55:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 4 Jun 2009 16:55:47 -0500
Subject: [LLVMbugs] [Bug 4328] New: Nightly test should have some kind of
 "rotation" mechanism
Message-ID: <bug-4328-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4328

           Summary: Nightly test should have some kind of "rotation"
                    mechanism
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Programs Tests
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


It would be nice if NewNightlyTest.pl had some kind of option for "rotating"
build directories, in conjunction with -noremove. Keeping the previous N
nightly test build directories would make it much easier to quickly analyze
failures.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 00:12:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 00:12:36 -0500
Subject: [LLVMbugs] [Bug 4331] New: Floating point ICE with llvm-gcc 2.5
Message-ID: <bug-4331-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4331

           Summary: Floating point ICE with llvm-gcc 2.5
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jsg at openbsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3065)
 --> (http://llvm.org/bugs/attachment.cgi?id=3065)
preprocessed output

llvm on OpenBSD/i386 -current with llvm-gcc 2.5 fails to compile
s_fmin.c/s_fminf.c/s_fminl.c in libm.

For example

llvm-gcc (GCC) 4.2.1 (Based on Apple Inc. build 5636) (LLVM build)

llvm-gcc -O2 -pipe -g   -I/usr/src/lib/libm/src -I/usr/src/lib/libm/src/ld80 -c
/usr/src/lib/libm/src/s_fmin.c -o s_fmin.o
assertion "Op.isUse() && (Op.isKill() || getFPReg(Op) == FirstFPRegOp ||
MI->killsRegister(Op.getReg())) && "Ret only defs operands, and values aren't
live beyond it"" failed: file
"/data/obj/portbuild/llvm-2.5/llvm-2.5/lib/Target/X86/X86FloatingPoint.cpp",
line 1072, function "handleSpecialFP"
/usr/src/lib/libm/src/s_fmin.c:49: internal compiler error: Abort trap


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 00:29:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 00:29:00 -0500
Subject: [LLVMbugs] [Bug 4332] New: Alignment related ICE with llvm-gcc 2.5
Message-ID: <bug-4332-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4332

           Summary: Alignment related ICE with llvm-gcc 2.5
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jsg at openbsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3066)
 --> (http://llvm.org/bugs/attachment.cgi?id=3066)
preprocessed output

With llvm-gcc 2.5 on OpenBSD/i386 -current compiling libc the following ICE is
seen:

llvm-gcc (GCC) 4.2.1 (Based on Apple Inc. build 5636) (LLVM build)

llvm-gcc -O2 -pipe -g -DLIBC_SCCS -DSYSLIBC_SCCS -I/usr/src/lib/libc/include
-DAPIWARN -DYP -I/usr/src/lib/libc/yp -D__DBINTERFACE_PRIVATE 
-I/usr/src/lib/libc/gdtoa -I/usr/src/lib/libc/arch/i386/gdtoa -DINFNAN_CHECK
-DMULTIPLE_THREADS -DNO_FENV_H -DUSE_LOCALE -I/usr/src/lib/libc -DRESOLVSORT
-DPOSIX_MISTAKE -DFLOATING_POINT -DNLS   -c /usr/src/lib/libc/stdlib/malloc.c
-o malloc.o
assertion "Alignment && "LValue alignment cannot be zero!"" failed: file
".././../gcc/llvm-internal.h", line 275, function "getAlignment"
/usr/src/lib/libc/stdlib/malloc.c: In function 'omalloc':
/usr/src/lib/libc/stdlib/malloc.c:1119: internal compiler error: Abort trap


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 02:48:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 02:48:17 -0500
Subject: [LLVMbugs] [Bug 3256] valgrind reports invalid reads in
	Sema::ActOnPopScope
In-Reply-To: <bug-3256-206@http.llvm.org/bugs/>
Message-ID: <200906050748.n557mHiE021572@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3256


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-05 02:48:07 ---
This doesn't seem to be an issue anymore.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 03:20:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 03:20:30 -0500
Subject: [LLVMbugs] [Bug 3443] recent regression: undeclared function isn' t
 infered the right type
In-Reply-To: <bug-3443-206@http.llvm.org/bugs/>
Message-ID: <200906050820.n558KUY7028687@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3443


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #15 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-05 03:20:30 ---
stpcpy builtin added in r72936.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 12:12:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 12:12:11 -0500
Subject: [LLVMbugs] [Bug 4333] New: configure should support --disable-ffi
Message-ID: <bug-4333-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4333

           Summary: configure should support --disable-ffi
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: autoconf
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


EOM.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 12:53:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 12:53:21 -0500
Subject: [LLVMbugs] [Bug 4334] New: scan-build gives no reports
Message-ID: <bug-4334-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4334

           Summary: scan-build gives no reports
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lbalbalba at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Im trying to build a cpp program called 'desmume' (www.desmume.org) using
llvm/clang in the following manner :

scan-build ./configure
scan-build make

But scan-build then just gives me the following message :
"scan-build: Removing directory '/tmp/scan-build-2009-06-05-1' because it
contains no reports."


Does anyone know what might be going on here ?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 12:56:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 12:56:57 -0500
Subject: [LLVMbugs] [Bug 4334] scan-build gives no reports
In-Reply-To: <bug-4334-206@http.llvm.org/bugs/>
Message-ID: <200906051756.n55HuvR3017690@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4334


John Smith <lbalbalba at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 20:15:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 20:15:23 -0500
Subject: [LLVMbugs] [Bug 4336] New: Iterating over use-def chains doesn' t
 seem to be deterministic.
Message-ID: <bug-4336-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4336

           Summary: Iterating over use-def chains doesn't seem to be
                    deterministic.
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jlerouge at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3069)
 --> (http://llvm.org/bugs/attachment.cgi?id=3069)
Pass to print use-def chains.

Hello,

It seems like iterating over use-def chains is not deterministic. It seems uses
are not always stored in the same order.

The attached patch adds a pass that iterates over the use-def chains, printing
all the instructions in the file "usedef.txt0".

By modifying opt command line, the order in which values are printed is
sometimes different.

Example:

$ ../Debug/bin/opt --load LLVMHello.dylib --load LLVMUseDef.dylib -checkusedef
./MultiSource/Benchmarks/ASCI_Purple/SMG2000/Output/random.bc >/dev/null
$ md5 usedef.txt0 
MD5 (usedef.txt0) = 39fbfb368fc1857d790072a16839124f

$ ../Debug/bin/opt --load LLVMUseDef.dylib -checkusedef
./MultiSource/Benchmarks/ASCI_Purple/SMG2000/Output/random.bc >/dev/null
$ md5 usedef.txt0 
MD5 (usedef.txt0) = 3f79a6ff1668d640f456484d588dfc2d

So, in this case, loading the Hello dylib seems to be enough to expose the non
determinism, even though the hello pass doesn't even run.

In this case, the diff is in the order in which uses for a global are showed:
 diff -u ref1.txt usedef.txt0
--- ref1.txt    2009-06-05 17:51:48.000000000 -0700
+++ usedef.txt0 2009-06-05 17:51:57.000000000 -0700
@@ -1,14 +1,14 @@
 @Seed = internal global i32 13579              ; <i32*> [#uses=4]

 Use Begin:
-       store i32 %3, i32* @Seed, align 4
-
        %4 = load i32* @Seed, align 4           ; <i32> [#uses=1]

        store i32 %3, i32* @Seed, align 4

        %1 = load i32* @Seed, align 4           ; <i32> [#uses=1]

+       store i32 %3, i32* @Seed, align 4
+
 Use End:


I found that from all the bc files in the llvm-test build,
./MultiSource/Benchmarks/ASCI_Purple/random.bc was the smallest one to
reproduce the problem on my machine (Leopard 10.5.7, LLVM svn r72977).

The problem also seems to happen with LLVM 2.5.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun  5 23:23:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 5 Jun 2009 23:23:26 -0500
Subject: [LLVMbugs] [Bug 2598] fptosint fails when asked to convert v2f32
 (0.0, 0, 0) to v2i32 (0  , 0)
In-Reply-To: <bug-2598-206@http.llvm.org/bugs/>
Message-ID: <200906060423.n564NQAh008703@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2598


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-05 23:23:24 ---
Fixed in r72983.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 01:33:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 01:33:13 -0500
Subject: [LLVMbugs] [Bug 4333] configure should support --disable-ffi
In-Reply-To: <bug-4333-206@http.llvm.org/bugs/>
Message-ID: <200906060633.n566XDFV012986@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4333


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Nick Lewycky <nicholas at mxc.ca>  2009-06-06 01:33:12 ---
The flag is "--disable-libffi".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 04:29:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 04:29:41 -0500
Subject: [LLVMbugs] [Bug 4337] New: fprintf makes Clang die
Message-ID: <bug-4337-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4337

           Summary: fprintf makes Clang die
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


The following code makes LLVM/Clang r72995 die:

void fprintf(void);

Assertion failed: (false && "a decl that inherits DeclContext isn't handled"),
function castToDeclContext, file .../llvm/tools/clang/lib/AST/DeclBase.cpp,
line 341.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 09:02:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 09:02:08 -0500
Subject: [LLVMbugs] [Bug 4338] New: Assertion failed:
 (!N1.getValueType().isInteger() && " Illegal setcc for integer!")
Message-ID: <bug-4338-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4338

           Summary: Assertion failed: (!N1.getValueType().isInteger() &&
                    "Illegal setcc for integer!")
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


I'm seeing the following crash when compiling the following code and enabling
-O1 or higher. This is LLVM+Clang r72995:

static const char p[1] = { 1 };

int
crash(const unsigned char *source)
{

        return (p[*source] == 4 && (*source < 0x80 || *source > 0xBF));
}

; ModuleID = 'llvm-crash.c'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "x86_64-undermydesk-freebsd8.0"
@p = internal constant [1 x i8] c"\01", align 1         ; <[1 x i8]*> [#uses=1]

define i32 @crash(i8* nocapture %source) nounwind readonly {
entry:
        %tmp1 = load i8* %source                ; <i8> [#uses=3]
        %idxprom = zext i8 %tmp1 to i64         ; <i64> [#uses=1]
        %arrayidx = getelementptr [1 x i8]* @p, i64 0, i64 %idxprom            
; <i8*> [#uses=1]
        %tmp2 = load i8* %arrayidx              ; <i8> [#uses=1]
        %cmp = icmp eq i8 %tmp2, 4              ; <i1> [#uses=1]
        br i1 %cmp, label %land.rhs, label %land.end

land.rhs:               ; preds = %entry
        %cmp7 = icmp sgt i8 %tmp1, -1           ; <i1> [#uses=1]
        br i1 %cmp7, label %land.end, label %lor.rhs

lor.rhs:                ; preds = %land.rhs
        %cmp12 = icmp ugt i8 %tmp1, -65         ; <i1> [#uses=1]
        %phitmp = zext i1 %cmp12 to i32         ; <i32> [#uses=1]
        ret i32 %phitmp

land.end:               ; preds = %land.rhs, %entry
        %0 = phi i32 [ 0, %entry ], [ 1, %land.rhs ]            ; <i32>
[#uses=1]
        ret i32 %0
}

Assertion failed: (!N1.getValueType().isInteger() && "Illegal setcc for
integer!"), function FoldSetCC, file
.../llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp, line 1404.
Stack dump:
0.      Program arguments: /usr/bin/../libexec/clang-cc -triple
x86_64-undermydesk-freebsd8.0 -S -disable-free -main-file-name llvm-crash.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O1 -fdiagnostics-show-option -o /tmp/cc-5rm5vI.s -x c
llvm-crash.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@crash'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 10:40:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 10:40:49 -0500
Subject: [LLVMbugs] [Bug 4339] New: Clang assertion failure with vector
	operation
Message-ID: <bug-4339-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4339

           Summary: Clang assertion failure with vector operation
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: arplynn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3070)
 --> (http://llvm.org/bugs/attachment.cgi?id=3070)
Simple test-case.

Clang dies with an assertion failure during codegen when compiling the attached
testcase.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 11:41:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 11:41:20 -0500
Subject: [LLVMbugs] [Bug 4337] fprintf makes Clang die
In-Reply-To: <bug-4337-206@http.llvm.org/bugs/>
Message-ID: <200906061641.n56GfKqG012829@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4337


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Ed Schouten <ed at 80386.nl>  2009-06-06 11:41:20 ---
It turns out there was this mysterious miscompilation, somewhere in the system.
After reinstalling my Jail with a system not built by Clang itself, I am not
able to reproduce.

I am closing this ticket for now. Sorry for the inconvenience.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 11:41:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 11:41:49 -0500
Subject: [LLVMbugs] [Bug 4338] Assertion failed:
 (!N1.getValueType().isInteger() && " Illegal setcc for integer!")
In-Reply-To: <bug-4338-206@http.llvm.org/bugs/>
Message-ID: <200906061641.n56GfnhN012877@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4338


Ed Schouten <ed at 80386.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Ed Schouten <ed at 80386.nl>  2009-06-06 11:41:49 ---
Same as 4337; probably caused by a miscompilation earlier on.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 14:09:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 14:09:42 -0500
Subject: [LLVMbugs] [Bug 4339] Clang assertion failure with vector operation
In-Reply-To: <bug-4339-206@http.llvm.org/bugs/>
Message-ID: <200906061909.n56J9g9r017837@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4339


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-06 14:09:41 ---
Fixed in r73005.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 14:31:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 14:31:43 -0500
Subject: [LLVMbugs] [Bug 4340] New: Missed optimisations with vectors.
Message-ID: <bug-4340-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4340

           Summary: Missed optimisations with vectors.
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: arplynn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Here is a very short test-case:

define void @vz(<4 x float>* nocapture %val) nounwind {
entry:
        %tmp1 = load <4 x float>* %val          ; <<4 x float>> [#uses=1]
        %vecins = insertelement <4 x float> %tmp1, float 0.000000e+00, i32 0   
        ; <<4 x float>> [#uses=1]
        %vecins4 = insertelement <4 x float> %vecins, float 0.000000e+00, i32 1
        ; <<4 x float>> [#uses=1]
        %vecins7 = insertelement <4 x float> %vecins4, float 0.000000e+00, i32
2                ; <<4 x float>> [#uses=1]
        %vecins10 = insertelement <4 x float> %vecins7, float 0.000000e+00, i32
3               ; <<4 x float>> [#uses=1]
        store <4 x float> %vecins10, <4 x float>* %val
        ret void
}

Two missed things here: the redundant load and the fact that the whole thing
could be reduced to store <4 x float> <float 0.000000e+00, float 0.000000e+00,
float 0.000000e+00, float 0.000000e+00>, <4 x float>* %val


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 15:08:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 15:08:25 -0500
Subject: [LLVMbugs] [Bug 4340] Missed optimisations with vectors.
In-Reply-To: <bug-4340-206@http.llvm.org/bugs/>
Message-ID: <200906062008.n56K8PPe019800@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4340


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-06 15:08:25 ---
Fixed in r73006.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 18:13:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 18:13:50 -0500
Subject: [LLVMbugs] [Bug 4314] llvm-ld's LTO segfaults in ' Interprocedural
 Sparse Conditional Constant Propagation'
In-Reply-To: <bug-4314-206@http.llvm.org/bugs/>
Message-ID: <200906062313.n56NDoMu025602@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4314


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Nick Lewycky <nicholas at mxc.ca>  2009-06-06 18:13:50 ---
Fixed in r73007,
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078460.html
.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 18:54:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 18:54:43 -0500
Subject: [LLVMbugs] [Bug 4342] New: llvm-gcc frontend passes non-absolute
 path to gold plugin
Message-ID: <bug-4342-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4342

           Summary: llvm-gcc frontend passes non-absolute path to gold
                    plugin
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: rafael.espindola at gmail.com, llvmbugs at cs.uiuc.edu


If I run "llvm-gcc -use-gold-plugin -flto -O2 hello.c -o hello" then llvm-gcc
will run collect2 with "-plugin-opt=gcc=llvm-gcc" which doesn't work. It's
expecting a full path (in my case /usr/local/bin/llvm-gcc).

Can we fix this in the gcc frontend (so it always passes the absolute path to
itself)? Or do we need to make libLTO willing to do a PATH lookup?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 19:17:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 19:17:25 -0500
Subject: [LLVMbugs] [Bug 4039] LegalizeDAG does not allow custom lowering of
 EXTRACT_SUBVECTOR
In-Reply-To: <bug-4039-206@http.llvm.org/bugs/>
Message-ID: <200906070017.n570HPZ5027704@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4039


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-06 19:17:25 ---
This should be fixed with the LegalizeDAG refactoring; please reopen if there
are still issues.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 19:36:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 19:36:30 -0500
Subject: [LLVMbugs] [Bug 3627] vector widening crash
In-Reply-To: <bug-3627-206@http.llvm.org/bugs/>
Message-ID: <200906070036.n570aUHp028356@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3627


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-06 19:36:29 ---
This seems to work now; please reopen if it's still an issue.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun  6 20:08:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 6 Jun 2009 20:08:17 -0500
Subject: [LLVMbugs] [Bug 3628] PPC Backend crashes with cannot select v4i32
 SRA instruction
In-Reply-To: <bug-3628-206@http.llvm.org/bugs/>
Message-ID: <200906070108.n5718HqP029238@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3628


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-06 20:08:15 ---
Fixed in r73009.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun  7 03:31:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 7 Jun 2009 03:31:40 -0500
Subject: [LLVMbugs] [Bug 4343] New: documentation missing for how to emit
 bitcode and machine code
Message-ID: <bug-4343-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4343

           Summary: documentation missing for how to emit bitcode and
                    machine code
           Product: Documentation
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: john at johnnowak.com
                CC: llvmbugs at cs.uiuc.edu


[04:21am] johnnowak: where in the documentation does it discuss how to emit
bitcode and machine code?
[04:21am] johnnowak: i can't seem to track it down
[04:22am] ddunbar__: not sure if its in the actual docs
[04:23am] ddunbar__: see include/llvm/Bitcode/ReaderWriter.h though
[04:23am] johnnowak: you'd think discussing how to generate code would be part
of the documentation for a compiler infrastructure
[04:23am] ddunbar__: patches welcome!
[04:23am] ddunbar__: I thought the tutorial covered this stuff though
[04:24am] johnnowak: it doesn't
[04:24am] edwin: opt -load=yourplugin.so input.bc -o output.bc
[04:26am] johnnowak: i'd hope that you wouldn't want people unfamiliar with the
project submitting essential pieces of documentation for basic yet critical
functionality
[04:27am] ddunbar__: johnnowak: file a docs bug if you like


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun  7 12:52:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 7 Jun 2009 12:52:54 -0500
Subject: [LLVMbugs] [Bug 2562] MMX inserts fail
In-Reply-To: <bug-2562-206@http.llvm.org/bugs/>
Message-ID: <200906071752.n57Hqs7s032066@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2562


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #16 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-07 12:52:53 ---
Okay then, resolving.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun  7 13:05:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 7 Jun 2009 13:05:37 -0500
Subject: [LLVMbugs] [Bug 2544] vector uitofp does not select on x86 with SSE2
In-Reply-To: <bug-2544-206@http.llvm.org/bugs/>
Message-ID: <200906071805.n57I5bIq032469@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2544


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-07 13:05:37 ---
I fixed this recently by adding UINT_TO_FP to the list of operations expanded
on x86.  We could probably generate better code by custom-lowering, but the
current result is good enough for now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun  7 13:36:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 7 Jun 2009 13:36:49 -0500
Subject: [LLVMbugs] [Bug 4344] New: llc asserts on frameaddress intrinsic on
	ARM
Message-ID: <bug-4344-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4344

           Summary: llc asserts on frameaddress intrinsic on ARM
           Product: tools
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: larsivar at igesund.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3071)
 --> (http://llvm.org/bugs/attachment.cgi?id=3071)
reduced and simplified bugpoint bc

Reproduced with a snippet using the frameaddress intrinsic when cross compiling
for ARM.

Relevant arguments could be

-march=arm  
-mtriple=asm-linuxeabi-unknown-gnu 
-gcc=~/android/prebuilt/linux-x86/toolchain/arm-eabi-4.3.1/bin/arm-eabi-gcc

bugpoint reduction attached.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  8 00:50:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 8 Jun 2009 00:50:03 -0500
Subject: [LLVMbugs] [Bug 4345] New: __func__ is a predefined identifier,
	not a keyword
Message-ID: <bug-4345-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4345

           Summary: __func__ is a predefined identifier, not a keyword
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


I think that this should be legal:
void foo() { { int __func__ = 4; }}


__func__ is not a keyword according to C99, it is a predefined identifier.  See
also C99 6.4.2.2.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  8 08:31:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 8 Jun 2009 08:31:00 -0500
Subject: [LLVMbugs] [Bug 4345] __func__ is a predefined identifier,
	not a keyword
In-Reply-To: <bug-4345-206@http.llvm.org/bugs/>
Message-ID: <200906081331.n58DV0CQ016481@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4345


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-06-08 08:30:58 ---
Ah, right!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun  8 17:23:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 8 Jun 2009 17:23:28 -0500
Subject: [LLVMbugs] [Bug 4349] New: Internal compiler error: Assertion `Ty
 && " GEP indices invalid!"' failed
Message-ID: <bug-4349-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4349

           Summary: Internal compiler error: Assertion `Ty && "GEP indices
                    invalid!"' failed
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mcuelenaere at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3072)
 --> (http://llvm.org/bugs/attachment.cgi?id=3072)
Preprocessed source

When compiling the Rockbox source code as simulator, I encountered the
following bug:

cc1: /build/buildd/llvm-2.2/lib/VMCore/Constants.cpp:1862: static
llvm::Constant* llvm::ConstantExpr::getGetElementPtr(llvm::Constant*,
llvm::Value* const*, unsigned int): Assertion `Ty && "GEP indices invalid!"'
failed.
/home/mcuelenaere/rockbox/apps/plugins/rockboy/save.c:155: internal compiler
error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


I attached the preprocessed source as requested.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 09:40:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 09:40:45 -0500
Subject: [LLVMbugs] [Bug 4350] New: clang mishandles the format printf
	attribute
Message-ID: <bug-4350-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4350

           Summary: clang mishandles the format printf attribute
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: bagnara at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3074)
 --> (http://llvm.org/bugs/attachment.cgi?id=3074)
Patch fixing the problem

When clang 72105 parses

#include <stdarg.h>
extern int vprintf(const char*, va_list);

it erroneously produces the invalid format printf attribute

extern int vprintf(const char*, va_list)
  __attribute__ ((__format__ (printf, 1, 2)));

instead of

extern int vprintf(const char*, va_list)
  __attribute__ ((__format__ (printf, 1, 0)));

The attached patch fixes the problem.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 13:01:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 13:01:48 -0500
Subject: [LLVMbugs] [Bug 4351] New: Compile-time constants with comparisons
 of function addresses cannot be determined
Message-ID: <bug-4351-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4351

           Summary: Compile-time constants with comparisons of function
                    addresses cannot be determined
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Clang chokes on the following piece of code:

void
foo(void)
{
}

int p = (foo == (void *)0) ? 1 : 2;

test.c:6:9: error: initializer element is not a compile-time constant
int p = (foo == (void *)0) ? 1 : 2;
        ^~~~~~~~~~~~~~~~~~~~~~~~~~
1 diagnostic generated.

I agree, the code is quite silly, but I'm running into pieces of kernel source
that use this.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 16:24:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 16:24:31 -0500
Subject: [LLVMbugs] [Bug 4342] llvm-gcc frontend passes non-absolute path to
	gold plugin
In-Reply-To: <bug-4342-206@http.llvm.org/bugs/>
Message-ID: <200906092124.n59LOVwQ024859@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4342


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-06-09 16:24:30 ---
Fixed in revision 73146. We now pass the assembler to the plugin. It is the
same assembler that gcc itself would use, so this is probably the correct
solution.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 16:35:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 16:35:31 -0500
Subject: [LLVMbugs] [Bug 4353] New: Escape sequence accepted by GCC and
 triggers warning in clang
Message-ID: <bug-4353-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4353

           Summary: Escape sequence accepted by GCC and triggers warning in
                    clang
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: catfish.man at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Adium uses the \E escape sequence for the escape key, which GCC accepts without
complaint. When building with clang we get warnings along the lines of "Unknown
escape sequence '\E'".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 20:32:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 20:32:53 -0500
Subject: [LLVMbugs] [Bug 4353] Escape sequence accepted by GCC and triggers
 warning in clang
In-Reply-To: <bug-4353-206@http.llvm.org/bugs/>
Message-ID: <200906100132.n5A1WrQb031494@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4353


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-09 20:32:53 ---
Fixed in r73153.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun  9 23:02:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 9 Jun 2009 23:02:15 -0500
Subject: [LLVMbugs] [Bug 4350] clang mishandles the format printf attribute
In-Reply-To: <bug-4350-206@http.llvm.org/bugs/>
Message-ID: <200906100402.n5A42Fj7002982@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4350


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-09 23:02:15 ---
Okay, committed in r73157.  You can usually just send this sort of thing to
cfe-commits, though.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 05:29:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 05:29:58 -0500
Subject: [LLVMbugs] [Bug 4355] New: configure does not pass options to
	BuildTools subdir
Message-ID: <bug-4355-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4355

           Summary: configure does not pass options to BuildTools subdir
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: e0325716 at student.tuwien.ac.at
                CC: llvmbugs at cs.uiuc.edu, e0325716 at student.tuwien.ac.at


When cross-compiling LLVM the build tools are created in BuildTools subdir but
the configure there does not use the same options as the regular build.

When I configure the project with

../llvm-c7/configure --prefix=/nfs/a5/astifter/astifter/llvm/llvm-ppc64-install
--enable-optimized --disable-assertions --host=ppc64-linux

then the configure in BuildTools is done with 

../../llvm-svn/configure

This has the effect that the BuildTools/Debug directory contains the
executables whereas the build is searching for them in
BuildTools/Release-Asserts.

A manual workaround for now is to symlink these two directories in BuildTools.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 12:43:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 12:43:59 -0500
Subject: [LLVMbugs] [Bug 4357] New: Crappy code generated for simple loop
Message-ID: <bug-4357-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4357

           Summary: Crappy code generated for simple loop
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan.cheng at apple.com
                CC: llvmbugs at cs.uiuc.edu


int hcf(int a, int b)
{
        while (a != 0)
        {
                if (a < b) b -= a;
                else a -= b;
        }
        return b;
}

David Majnemer has a piece of code that shows less than optimal code generation
for llvm.

This is what icc generates:

8048624 <hcf>:
 8048624:       8b 54 24 04             mov    0x4(%esp),%edx
 8048628:       8b 44 24 08             mov    0x8(%esp),%eax
 804862c:       85 d2                   test   %edx,%edx
 804862e:       74 0c                   je     804863c <hcf+0x18>
 8048630:       3b d0                   cmp    %eax,%edx
 8048632:       7d 04                   jge    8048638 <hcf+0x14>
 8048634:       2b c2                   sub    %edx,%eax
 8048636:       eb f8                   jmp    8048630 <hcf+0xc>
 8048638:       2b d0                   sub    %eax,%edx
 804863a:       75 f4                   jne    8048630 <hcf+0xc>
 804863c:       c3                      ret

This is what llvm-gcc generates with -fomit-frame-pointer:

_hcf:
        movl    8(%esp), %eax
        movl    4(%esp), %ecx
        jmp     LBB1_3
LBB1_1:
        cmpl    %eax, %ecx
        jl      LBB1_5
        subl    %eax, %ecx
LBB1_3:
        testl   %ecx, %ecx
        jne     LBB1_1
        ret
LBB1_5:
        subl    %ecx, %eax
        jmp     LBB1_3


clang does something even more horrible:

_hcf:
        pushl   %esi
        movl    12(%esp), %eax
        movl    8(%esp), %ecx
        jmp     LBB1_5
        .align  4,0x90
LBB1_1:
        xorl    %edx, %edx
        movl    %eax, %esi
        jmp     LBB1_3
        .align  4,0x90
LBB1_2:
        subl    %ecx, %esi
        addl    %ecx, %edx
LBB1_3:
        cmpl    %esi, %ecx
        jl      LBB1_2
        addl    %edx, %ecx
        subl    %eax, %ecx
        movl    %esi, %eax
LBB1_5:
        testl   %ecx, %ecx
        jne     LBB1_1
        popl    %esi
        ret


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 13:12:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 13:12:02 -0500
Subject: [LLVMbugs] [Bug 4360] New: clang crashes on solaris.
Message-ID: <bug-4360-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4360

           Summary: clang crashes on solaris.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Hi,

llvm/clang;
At revision 73170.

I get the follow trace from clang that crashes instantly on both Solaris &
AuroraUX when run.

(gdb) bt
#0  0xfef60fe0 in malloc_unlocked () from /usr/lib/libmalloc.so.1
#1  0xfef60ab1 in malloc () from /usr/lib/libmalloc.so.1
#2  0xfee99f32 in __cxa_demangle () from /usr/sfw/lib/libstdc++.so.6
#3  0x080d9525 in PrintStackTrace () at Signals.inc:210
#4  0x080d926f in SignalHandler (Sig=11) at Signals.inc:136
#5  0xfecbd0af in __sighndlr () from /lib/libc.so.1
#6  0xfecb01af in call_user_handler () from /lib/libc.so.1
#7  <signal handler called>
#8  0xfef60fe0 in malloc_unlocked () from /usr/lib/libmalloc.so.1
#9  0xfef60ab1 in malloc () from /usr/lib/libmalloc.so.1
#10 0xfeeac0a9 in operator new () from /usr/sfw/lib/libstdc++.so.6
#11 0xfee7d946 in std::string::_Rep::_S_create () from
/usr/sfw/lib/libstdc++.so.6
#12 0x080982b2 in std::string::_S_construct<char const*> (__beg=0x810e2c4
"/export/home2/edward/lab/TEST_BUILDS/llvm/test1/bin", __end=0x810e2f7 "",
__a=@0x80475b0)
    at basic_string.tcc:150
#13 0xfee7fcb3 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string () from /usr/sfw/lib/libstdc++.so.6
#14 0x0809ad9d in Driver (this=0x8047684, _Name=0x810c6a4 "clang",
_Dir=0x810e2c4 "/export/home2/edward/lab/TEST_BUILDS/llvm/test1/bin", 
    _DefaultHostTriple=0x810d324 "i386-pc-solaris2.11",
_DefaultImageName=0x80da5d7 "a.out", _Diags=@0x80476f4) at Driver.cpp:52
#15 0x080962c4 in main (argc=1, argv=0x8047ccc) at driver.cpp:178

Regards,
Edward.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 14:39:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 14:39:26 -0500
Subject: [LLVMbugs] [Bug 4361] New: DAGCombine incorrectly generates 64bit
 ext load from 16 bit load on a platform with 32bit aligned addresses
Message-ID: <bug-4361-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4361

           Summary: DAGCombine incorrectly generates 64bit ext load from 16
                    bit load on a platform with 32bit aligned addresses
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3076)
 --> (http://llvm.org/bugs/attachment.cgi?id=3076)
incorrect extension of 16bit into 64bit

The following test case is a conversion routine from a ushort to a ulong:
define void @test_convert_ulong_ushort_kernel(i16 addrspace(1)* %src, i64
addrspace(1)* %dest) nounwind {
entry:
        %call = tail call i32 @get_global_id(i32 0) nounwind            ; <i32>
[#uses=0]
        %tmp2 = load i16 addrspace(1)* %src             ; <i16> [#uses=1]
        %conv.i.i = zext i16 %tmp2 to i64               ; <i64> [#uses=1]
        store i64 %conv.i.i, i64 addrspace(1)* %dest
        ret void
}

declare i32 @get_global_id(i32)

However, in DAGCombiner.cpp:visitSIGN_EXTEND:3054 and
DAGCombiner.cpp:visitZERO_EXTEND:3148, an incorrect conversion occurs as shown
in incorrectsext.dot which will be attached later. The problem is that it is
converting the 16bit load and an extension to 64bit to a 64bit sign extended
load followed by an and of the lower 16 bits. This works fine if the address
that the 16bit load is from is aligned to the 32bit boundary but does not work
if aligned to the 16bit boundary. 

If I comment out the blocks that contain this conversion:
// fold (sext (sextload x)) -> (sext (truncate (sextload x)))
  // fold (sext ( extload x)) -> (sext (truncate (sextload x)))

and
// fold (zext (load x)) -> (zext (truncate (zextload x)))

then the code gets generated as expected, a 16bit load followed by a zext as
specified in correctsext.dot.

I set all sextload/zextload's to expand via 
setLoadExtAction(ISD::SEXTLOAD, VT, Expand);        
setLoadExtAction(ISD::ZEXTLOAD, VT, Expand);

Where VT is all the value types that I support.
I also set
allowUnalignedMemoryAccesses = false;


The above bitcode given linear values 0-> (N-1) as input to short produces 0,
0, 2, 2, 4, 4, etc... as the second 16bit value does not get correctly shifted
into position.

With my backend, all 8/16 bit loads/stores MUST NOT be removed/combined as I
have custom assembly written that correctly handles the unaligned cases.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 18:22:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 18:22:44 -0500
Subject: [LLVMbugs] [Bug 4366] New: instcombine and simplifycfg mishandle
 stores to null in non-zero address-space
Message-ID: <bug-4366-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4366

           Summary: instcombine and simplifycfg mishandle stores to null in
                    non-zero address-space
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
define void @a() {
  store i32 0, i32 addrspace(1)* null
  ret void
}

simplifycfg illegally turns this into "unreachable", and instcombine illegally
turns this into a store of an undef value.  Neither transformation should touch
this store: there should be a check for the address-space, similar to the check
in InstCombiner::visitLoadInst.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 10 20:32:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 10 Jun 2009 20:32:47 -0500
Subject: [LLVMbugs] [Bug 4361] DAGCombine incorrectly generates 64bit ext
 load from 16 bit load on a platform with 32bit aligned addresses
In-Reply-To: <bug-4361-206@http.llvm.org/bugs/>
Message-ID: <200906110132.n5B1Wlg6022631@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4361


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #10 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-10 20:32:47 ---
Marking this invalid; the report is based on a misunderstanding of the
definition of an extending load.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 05:27:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 05:27:33 -0500
Subject: [LLVMbugs] [Bug 4368] New: Suggest lowering the severity of bad
	quoting
Message-ID: <bug-4368-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4368

           Summary: Suggest lowering the severity of bad quoting
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


I just saw a port (libproxy) that calls the C compiler as follows:

cc -o conftest -DFOO=\\\"bar\\\" conftest.c

GCC:

<command-line>: warning: missing terminating " character

Clang:

In file included from <built-in>:105:
<command line>:1:14: error: missing terminating '"' character
#define FOO \"bar\"
             ^
1 diagnostic generated.

This causes random configure tests to return the wrong result.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 12:44:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 12:44:17 -0500
Subject: [LLVMbugs] [Bug 4370] New: Debug/Function.o is 700k
Message-ID: <bug-4370-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4370

           Summary: Debug/Function.o is 700k
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


This is ridiculous, no?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 12:55:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 12:55:33 -0500
Subject: [LLVMbugs] [Bug 4366] instcombine and simplifycfg mishandle stores
 to null in non-zero address-space
In-Reply-To: <bug-4366-206@http.llvm.org/bugs/>
Message-ID: <200906111755.n5BHtXS1031094@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4366


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-06-11 12:55:32 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090608/078565.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 13:15:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 13:15:23 -0500
Subject: [LLVMbugs] [Bug 4371] New: much bigger resulting code than with gcc
Message-ID: <bug-4371-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4371

           Summary: much bigger resulting code than with gcc
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


lev /tmp$ gcc -c -O2 ansi.c && nm -S ansi.o | grep WriteString
0000000000004970 000000000000190a T WriteString


lev /tmp$ clang -c -O2 ansi.c && nm -S ansi.o | grep WriteString
ansi.c:2891:13: warning: format string is not a string literal (potentially
insecure) [-Wformat-security]
  LMsg(err, str);
            ^~~
1 diagnostic generated.
00000000000003d0 000000000000328e T WriteString

note the size of the WriteString function --- 190a vs 328e, thats quite a lot!
attached test case

this is most probably something in llvm codegen or llvm itself. I dont know


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 13:54:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 13:54:24 -0500
Subject: [LLVMbugs] [Bug 4372] New: Calls to defined functions without a
 prototype aren' t transformed to direct calls
Message-ID: <bug-4372-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4372

           Summary: Calls to defined functions without a prototype aren't
                    transformed to direct calls
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
__attribute((always_inline)) void foo() {}
int main(void) { foo(); return 0; }

Output of clang-cc -emit-llvm -O0:
define void @foo() nounwind alwaysinline {
entry:
        ret void
}

define i32 @main() nounwind {
entry:
        %retval = alloca i32            ; <i32*> [#uses=2]
        call void (...)* bitcast (void ()* @foo to void (...)*)()
        store i32 0, i32* %retval
        %0 = load i32* %retval          ; <i32> [#uses=1]
        ret i32 %0
}

EmitCall needs to be doing something similar to
ReplaceUsesOfNonProtoTypeWithRealFunction.

I think I introduced this issue recently with the fix to bug 4289, which
introduces the bitcast in question.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 14:16:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 14:16:50 -0500
Subject: [LLVMbugs] [Bug 4366] instcombine and simplifycfg mishandle stores
 to null in non-zero address-space
In-Reply-To: <bug-4366-206@http.llvm.org/bugs/>
Message-ID: <200906111916.n5BJGoNC001239@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4366


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-11 14:16:50 ---
The simplifycfg part hasn't been fixed yet.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 15:51:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 15:51:43 -0500
Subject: [LLVMbugs] [Bug 4349] Internal compiler error: Assertion `Ty &&
 "GEP indices invalid!" ' failed
In-Reply-To: <bug-4349-206@http.llvm.org/bugs/>
Message-ID: <200906112051.n5BKphjn004457@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4349


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Dale Johannesen <dalej at apple.com>  2009-06-11 15:51:43 ---
(In reply to comment #2)
> Created an attachment (id=3073)
 --> (http://llvm.org/bugs/attachment.cgi?id=3073) [details]
> reduced .c testcase 
> 
> With current LLVM, this causes a very different error:
> 
> $ llvm-gcc x.c
> cc1: ../../src/gcc/llvm-convert.cpp:7318: static llvm::Constant*
> TreeConstantToLLVM::EmitLV(tree_node*): Assertion `((((enum tree_code)
> (((exp)->common.type))->common.code) == VOID_TYPE) || LV->getType() ==
> ConvertType(((exp)->common.type))->getPointerTo()) && "LValue of constant has
> wrong type!"' failed.

This is fixed:
http://llvm.org/viewvc/llvm-project?view=rev&revision=73205
http://llvm.org/viewvc/llvm-project?view=rev&revision=73206

I realize Rockbox doesn't compile yet, but I'd like to close this one and track
the other problem under its existing report, PR 3518.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 16:03:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 16:03:22 -0500
Subject: [LLVMbugs] [Bug 4371] much bigger resulting code than with gcc
In-Reply-To: <bug-4371-206@http.llvm.org/bugs/>
Message-ID: <200906112103.n5BL3MMS004794@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4371


Roman Divacky <rdivacky at freebsd.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Roman Divacky <rdivacky at freebsd.org>  2009-06-11 16:03:21 ---
yes this fixes my test case completely.... feel free to reopen if it does not
fix this issue for you


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 18:51:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 18:51:06 -0500
Subject: [LLVMbugs] [Bug 4373] New: Aggregate store tickles assert in
 BitcodeReaderValueList:: getValueFwdRef
Message-ID: <bug-4373-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4373

           Summary: Aggregate store tickles assert in
                    BitcodeReaderValueList::getValueFwdRef
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Bitcode Reader
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: markleone at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following code tickles an assert failure in the bitcode reader.
To replicate:  llvm-as -o - test.ll | llvm-dis

@foo = weak global { i32 } zeroinitializer              
@bar = weak global i32 0                

define void @test() {
entry:
        store { i32 } zeroinitializer, { i32 }* @foo
        store i32 1, i32* @bar
        ret void
}

Assertion failed: ((Ty == 0 || Ty == V->getType()) && "Type mismatch in value
table!"), function getValueFwdRef, file BitcodeReader.cpp, line 215.

in llvm::BitcodeReaderValueList::getValueFwdRef at BitcodeReader.cpp:215
in llvm::BitcodeReader::getFnValueByID at BitcodeReader.h:160
in llvm::BitcodeReader::getValue at BitcodeReader.h:196
in llvm::BitcodeReader::ParseFunctionBody at BitcodeReader.cpp:1869


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 11 20:48:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 11 Jun 2009 20:48:55 -0500
Subject: [LLVMbugs] [Bug 1285] Fix consistency of FP operators: add -> fadd,
	etc
In-Reply-To: <bug-1285-206@http.llvm.org/bugs/>
Message-ID: <200906120148.n5C1mtYY013647@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1285


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-11 20:48:54 ---
I think this is fixed, now that we have fadd, fsub, and fmul.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 00:21:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 00:21:38 -0500
Subject: [LLVMbugs] [Bug 4373] Aggregate store tickles assert in
 BitcodeReaderValueList:: getValueFwdRef
In-Reply-To: <bug-4373-206@http.llvm.org/bugs/>
Message-ID: <200906120521.n5C5LcPw019664@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4373


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Nick Lewycky <nicholas at mxc.ca>  2009-06-12 00:21:38 ---
Fixed in r73220:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090608/078599.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 01:11:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 01:11:06 -0500
Subject: [LLVMbugs] [Bug 4374] New: incorrect instcombine optimization
Message-ID: <bug-4374-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4374

           Summary: incorrect instcombine optimization
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: llvm at laurentm.net
                CC: llvmbugs at cs.uiuc.edu


The following transfomation is incorrect: fsub(a,fsub(b,c)) =>
fadd(a,fsub(c,b))

For example:

float
func(float a, float b) {
    return -(a - b);
}

llvm-gcc generates:
define float @func(float %a, float %b) nounwind {
entry:
        %tmp3 = sub float %a, %b                ; <float> [#uses=1]
        %tmp4 = sub float -0.000000e+00, %tmp3          ; <float> [#uses=1]
        ret float %tmp4
}

opt -instcombine generates:
define float @func(float %a, float %b) nounwind {
entry:
        %tmp3 = sub float %b, %a                ; <float> [#uses=1]
        ret float %tmp3
}

The result of func(x,x) should be -0.0f not +0.0f.

InstructionCombining.cpp at 2621
    ...
    if (Op1I->hasOneUse()) {
      // Replace (x - (y - z)) with (x + (z - y)) if the (y - z) subexpression
      // is not used by anyone else...
      //
      if (Op1I->getOpcode() == Instruction::FSub) {
        // Swap the two operands of the subexpr...
        Value *IIOp0 = Op1I->getOperand(0), *IIOp1 = Op1I->getOperand(1);
        Op1I->setOperand(0, IIOp1);
        Op1I->setOperand(1, IIOp0);

        // Create the new top level fadd instruction...
        return BinaryOperator::CreateFAdd(Op0, Op1);
      }
    }
    ...


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 05:44:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 05:44:16 -0500
Subject: [LLVMbugs] [Bug 4375] New: llvm-g++ trips over __extern_inline with
	-O2
Message-ID: <bug-4375-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4375

           Summary: llvm-g++ trips over __extern_inline with -O2
           Product: tools
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: paul at floorball-flamingos.nl
                CC: llvmbugs at cs.uiuc.edu


This is with the RHEL4 binary distribution of llvm-gcc 4.2 (for llvm 2.5), on
Gentoo Linux (w/ 32-bit kernel and libs). GCC 4.3.2 has no problems with same
code.

12:39|melis at juggle2:~/projects/bih> cat t.cxx
#include <Python.h>
12:39|melis at juggle2:~/projects/bih> g++ -c -I /usr/include/python2.5 t.cxx
12:39|melis at juggle2:~/projects/bih> g++ -c -I /usr/include/python2.5 -O2 t.cxx

12:40|melis at juggle2:~/projects/bih> llvm-g++ -c -I /usr/include/python2.5 t.cxx
12:40|melis at juggle2:~/projects/bih> llvm-g++ -c -I /usr/include/python2.5 -O2
t.cxx
In file included from /usr/include/python2.5/Python.h:46,
                 from t.cxx:1:
/usr/include/stdlib.h:278: error: expected constructor, destructor, or type
conversion before 'double'
/usr/include/stdlib.h:283: error: expected constructor, destructor, or type
conversion before 'int'
/usr/include/stdlib.h:288: error: expected constructor, destructor, or type
conversion before 'long'
/usr/include/stdlib.h:297: error: expected constructor, destructor, or type
conversion before 'long'
In file included from /usr/include/python2.5/unicodeobject.h:118,
                 from /usr/include/python2.5/Python.h:88,
                 from t.cxx:1:
/usr/include/wchar.h:334: error: '__extern_inline' does not name a type
/usr/include/wchar.h:340: error: expected constructor, destructor, or type
conversion before 'int'
/usr/include/wchar.h:345: error: '__extern_inline' does not name a type
In file included from /usr/include/signal.h:33,
                 from /usr/include/python2.5/pyfpe.h:129,
                 from /usr/include/python2.5/Python.h:156,
                 from t.cxx:1:
/usr/include/bits/sigset.h:118: error: expected constructor, destructor, or
type conversion before 'int'
/usr/include/bits/sigset.h:119: error: expected constructor, destructor, or
type conversion before 'int'
/usr/include/bits/sigset.h:120: error: expected constructor, destructor, or
type conversion before 'int'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 07:20:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 07:20:45 -0500
Subject: [LLVMbugs] [Bug 4376] New: lli -fast crashes with thread-local
	variables
Message-ID: <bug-4376-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4376

           Summary: lli -fast crashes with thread-local variables
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: marks at dcs.gla.ac.uk
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3085)
 --> (http://llvm.org/bugs/attachment.cgi?id=3085)
Thread local test program

lli -fast seg-faults on the attached input.
lli (no flags) prints OK as it should:

$ lli  thread.bc
OK

$ lli -fast thread.bc
0   lli 0x0874f79e
Segmentation fault

If the thread-local is changed to an ordinary global, then lli -fast runs OK.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 11:31:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 11:31:47 -0500
Subject: [LLVMbugs] [Bug 4278] X86-64 -tailcallopt calling convention
 specification seems out-of-sync with regular cc
In-Reply-To: <bug-4278-206@http.llvm.org/bugs/>
Message-ID: <200906121631.n5CGVlWI021689@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4278


Arnold Schwaighofer <arnold.schwaighofer at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Arnold Schwaighofer <arnold.schwaighofer at gmail.com>  2009-06-12 11:31:43 ---
Fixed see: http://llvm.org/viewvc/llvm-project?view=rev&revision=73233.

Thank you Frits for bringing this to my attention!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 12:02:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 12:02:15 -0500
Subject: [LLVMbugs] [Bug 4377] New: Emitting llvm bitcode and then
 assembling throws multiple definition error
Message-ID: <bug-4377-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4377

           Summary: Emitting llvm bitcode and then assembling throws
                    multiple definition error
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: chster at cs.berkeley.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3087)
 --> (http://llvm.org/bugs/attachment.cgi?id=3087)
Contains three files, rng.cpp, rng.h, annealer_thread.cpp

The source files come from a PARSEC benchmark.

When I emit llvm bitcode and then assemble native code I get the following
error:
+ llvm-g++ -emit-llvm annealer_thread.cpp -c -o annealer_thread.o
+ llvm-g++ -emit-llvm rng.cpp -c -o rng.o
+ llc annealer_thread.o -f -o annealer_thread.s
+ llvm-g++ -c annealer_thread.s -o annealer_thread.o
+ llc rng.o -f -o rng.s
+ llvm-g++ -c rng.s -o rng.o
+ llvm-g++ annealer_thread.o rng.o
rng.o: In function `MTRand::reload()':
rng.s:(.text+0x50): multiple definition of `MTRand::reload()'
annealer_thread.o:annealer_thread.s:(.text+0xa0): first defined here
collect2: ld returned 1 exit status

When I don't everything works:
+ llvm-g++ annealer_thread.cpp -c -o annealer_thread.o
+ llvm-g++ rng.cpp -c -o rng.o
+ llvm-g++ rng.o annealer_thread.o


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 13:43:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 13:43:39 -0500
Subject: [LLVMbugs] [Bug 3518] undefined reference to __compound_literal.*
In-Reply-To: <bug-3518-206@http.llvm.org/bugs/>
Message-ID: <200906121843.n5CIhdET025889@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3518


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Dale Johannesen <dalej at apple.com>  2009-06-12 13:43:39 ---
Fixed here.  I've confirmed that all the tests in the gcc testsuite that were
failing this way now pass.
http://llvm.org/viewvc/llvm-project?view=rev&revision=73238
http://llvm.org/viewvc/llvm-project?view=rev&revision=73239


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 14:08:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 14:08:15 -0500
Subject: [LLVMbugs] [Bug 1285] Fix consistency of FP operators: add -> fadd,
	etc
In-Reply-To: <bug-1285-206@http.llvm.org/bugs/>
Message-ID: <200906121908.n5CJ8Fj0026862@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1285


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-06-12 14:08:15 ---
LLVM still lacks a real fneg.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 14:16:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 14:16:33 -0500
Subject: [LLVMbugs] [Bug 4377] Emitting llvm bitcode and then assembling
 throws multiple definition error
In-Reply-To: <bug-4377-206@http.llvm.org/bugs/>
Message-ID: <200906121916.n5CJGXeV027223@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4377


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-06-12 14:16:33 ---
Works for me on x86-32 linux and x86-64 linux with svn head.  Maybe
Christos is using 2.5 and it was fixed since?  Christos, feel free
to reopen if this is still a problem with latest llvm/llvm-gcc from
the subversion repository.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 14:24:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 14:24:10 -0500
Subject: [LLVMbugs] [Bug 4374] incorrect instcombine optimization
In-Reply-To: <bug-4374-206@http.llvm.org/bugs/>
Message-ID: <200906121924.n5CJOAwB027550@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4374


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dan Gohman <gohman at apple.com>  2009-06-12 14:24:10 ---
No, it predates the add -> fadd change.

Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090608/078631.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 14:38:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 14:38:40 -0500
Subject: [LLVMbugs] [Bug 4378] New: LLVM IR has no floating-point negate
	operator
Message-ID: <bug-4378-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4378

           Summary: LLVM IR has no floating-point negate operator
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


Sign-bit manipulation operations copysign and abs can be done with
libm calls, but there is currently no simple way to get a proper
floating-point negate operation in LLVM IR.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 14:39:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 14:39:08 -0500
Subject: [LLVMbugs] [Bug 1285] Fix consistency of FP operators: add -> fadd,
	etc
In-Reply-To: <bug-1285-206@http.llvm.org/bugs/>
Message-ID: <200906121939.n5CJd8O8028001@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1285


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Dan Gohman <gohman at apple.com>  2009-06-12 14:39:08 ---
It's also not equivalent when x is NaN.

I opened PR4378 to track the lack of fneg.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 16:01:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 16:01:31 -0500
Subject: [LLVMbugs] [Bug 4366] instcombine and simplifycfg mishandle stores
 to null in non-zero address-space
In-Reply-To: <bug-4366-206@http.llvm.org/bugs/>
Message-ID: <200906122101.n5CL1VtM031014@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4366


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-06-12 16:01:30 ---
aha, nice catch, fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090608/078641.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 16:31:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 16:31:58 -0500
Subject: [LLVMbugs] [Bug 2243] PHI coalescing / tail merging missing
	opportunity
In-Reply-To: <bug-2243-206@http.llvm.org/bugs/>
Message-ID: <200906122131.n5CLVwb2031997@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2243


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Dale Johannesen <dalej at apple.com>  2009-06-12 16:31:58 ---
This was fixed, probably a long time ago.  Might have been 52985.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 16:53:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 16:53:19 -0500
Subject: [LLVMbugs] [Bug 4380] New: __builtin_object_size
Message-ID: <bug-4380-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4380

           Summary: __builtin_object_size
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: devang.patel at gmail.com
                CC: llvmbugs at cs.uiuc.edu


In this example, builtin_object_size is not lowered by llvm-gcc.
---
typedef struct _MyFoo {
       int x, y;
} MyFoo;
void foobar() {
int y;
#pragma omp parallel for schedule(dynamic,4)
  for (y=0; y <  64; y++)
    {
      MyFoo *p, *q;
      __builtin___memcpy_chk (q,
                              p,
                              sizeof(p),
                              __builtin_object_size (q, 0));
    }
}
---


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 12 19:27:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 12 Jun 2009 19:27:06 -0500
Subject: [LLVMbugs] [Bug 4372] Calls to defined functions without a
 prototype aren' t transformed to direct calls
In-Reply-To: <bug-4372-206@http.llvm.org/bugs/>
Message-ID: <200906130027.n5D0R6hL005714@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4372


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-06-12 19:27:05 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090608/018117.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 13 18:54:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 13 Jun 2009 18:54:40 -0500
Subject: [LLVMbugs] [Bug 4368] Suggest lowering the severity of bad quoting
In-Reply-To: <bug-4368-206@http.llvm.org/bugs/>
Message-ID: <200906132354.n5DNseAO032486@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4368


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-06-13 18:54:39 ---
I don't see how this can possibly make sense.  I tried:

#define FOO \"bar\"
const char *P = 
FOO;


and both GCC 4.0 and 4.2 reject it with an error.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 13 21:21:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 13 Jun 2009 21:21:14 -0500
Subject: [LLVMbugs] [Bug 4386] New: Constant evaluation doesn' t handle
 converting the address of a weak global to a boolean correctly
Message-ID: <bug-4386-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4386

           Summary: Constant evaluation doesn't handle converting the
                    address of a weak global to a boolean correctly
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
__attribute((weak)) int foo();
int a = !foo;

clang currently evaluates "!foo" to zero; it should give an error.

I'm not sure how this should end up working with the following case:
int s(); int a = s == 0; int s() __attribute((weak));


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 13 21:22:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 13 Jun 2009 21:22:03 -0500
Subject: [LLVMbugs] [Bug 4351] Compile-time constants with comparisons of
 function addresses cannot be determined
In-Reply-To: <bug-4351-206@http.llvm.org/bugs/>
Message-ID: <200906140222.n5E2M3YG004258@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4351


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-13 21:22:02 ---
Fixed in r73321; I filed bug 4386 on the weak global issue.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 03:52:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 03:52:09 -0500
Subject: [LLVMbugs] [Bug 4388] New: clang: preprocessor output missing
	newline
Message-ID: <bug-4388-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4388

           Summary: clang: preprocessor output missing newline
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


clang -E -MD -MF arch/x86/kernel/acpi/realmode/.wakeup.lds.d  -nostdinc
-isystem /usr/lib/gcc/x86_64-linux-gnu/4.3.3/include -Iinclude 
-I/home/edwin/builds/linux-2.6/arch/x86/include -include
include/linux/autoconf.h -D__KERNEL__   -P -C -D__ASSEMBLY__ -o
arch/x86/kernel/acpi/realmode/wakeup.lds
arch/x86/kernel/acpi/realmode/wakeup.lds.S


Creates:
/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.30
 * Sun Jun 14 11:37:11 2009
 */# 2 "<built-in>" 2 3

gcc creates:


/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.30
 * Sun Jun 14 11:37:11 2009
 */


Notice the missing newline before #2.
gcc doesn't output #2 at all.
However even if I manually add a newline, ld still rejects it.
I think no #2 should be output, to be compatible with gcc.

ld:arch/x86/kernel/acpi/realmode/wakeup.lds:6: ignoring invalid character `#'
in expression


Input file:
/*
 * wakeup.ld
 *
 * Linker script for the real-mode wakeup code
 */
#undef i386
#include "wakeup.h"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 04:37:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 04:37:01 -0500
Subject: [LLVMbugs] [Bug 4389] New: clang: wrong scoping: refuses to compile
 global variable name same as local variable name
Message-ID: <bug-4389-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4389

           Summary: clang: wrong scoping: refuses to compile global variable
                    name same as local variable name
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


$ clang testcase-min.i
testcase-min.i:4:12: error: redefinition of 'index' as different kind of symbol
static int index[8];
           ^
testcase-min.i:2:9: note: previous definition is here
    int index = bar();
        ^
2 diagnostics generated.


testcase-min.i:
static void foo () {
    int index = bar();
}
static int index[8];


gcc compiles this code fine, and if I move the 'static int index[8]' before
foo(), then clang compiles it as well.

Why does clang still see the definition of the local index variable at the time
the global index variable is defined?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 05:57:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 05:57:11 -0500
Subject: [LLVMbugs] [Bug 4390] New: clang: wrong struct initializer
Message-ID: <bug-4390-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4390

           Summary: clang: wrong struct initializer
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


clang creates this initializer for a struct:
@sysfs_root = global %0 <{ %struct.anon <{ i32 1 }>, %struct.anon
zeroinitializer, %struct.sysfs_dirent* null, %struct.sysfs_dirent* null, i8*
getelementptr ([1 x i8]* @.str, i32 0, i32 0), %union.anon zeroinitializer,
%union.anon zeroinitializer, i8 0, i8 0, i8 0, i8 0, i32 1, i64 1, i8 0, i8 0,
i8 0, i8 0, i8 0, i8 0, i16 16877 }>, align 8         ; <%0*> [#uses=1]

While llvm-gcc creates this initializer for same struct:
@sysfs_root = global %0 { %struct.atomic_t { i32 1 }, [20 x i8]
zeroinitializer, i8* getelementptr ([1 x i8]* @.str, i32 0, i32 0), [16 x i8]
zeroinitializer, i32 1, i64 1, i16 16877, [14 x i8] zeroinitializer }, align 32
          ; <%0*> [#uses=1]

This causes a panic when trying to boot the linux kernel compiled with clang.

Testcase below, when run with clang it aborts, when run
with llvm-gcc or gcc it doesn't:

struct hlist_node {
 struct hlist_node *next, **pprev;
};                                
struct hlist_head {               
 struct hlist_node *first;        
};                                
typedef unsigned short umode_t;   
typedef unsigned long __kernel_ino_t;
typedef __kernel_ino_t ino_t;        
struct sysfs_elem_dir {              
 struct kobject *kobj;               

 struct sysfs_dirent *children;
};                             
struct sysfs_elem_bin_attr {   
 struct bin_attribute *bin_attr;
 struct hlist_head buffers;     
};                              
typedef struct {                
 volatile int counter;          
} atomic_t;                     
struct sysfs_elem_attr {        
 struct attribute *attr;        
 struct sysfs_open_dirent *open;
};                              
struct sysfs_elem_symlink {     
 struct sysfs_dirent *target_sd;
};                              
struct sysfs_dirent {           
 atomic_t s_count;              
 atomic_t s_active;             
 struct sysfs_dirent *s_parent;
 struct sysfs_dirent *s_sibling;
 const char *s_name;

 union {
  struct sysfs_elem_dir s_dir;
  struct sysfs_elem_symlink s_symlink;
  struct sysfs_elem_attr s_attr;
  struct sysfs_elem_bin_attr s_bin_attr;
 };

 unsigned int s_flags;
 ino_t s_ino;
 umode_t s_mode;
 struct iattr *s_iattr;
};
struct sysfs_dirent sysfs_root = {
 .s_name = "",
 .s_count = { (1) },
 .s_flags = 0x0001,
 .s_mode = 0040000 | 00700 | (00400|00040|00004) | (00100|00010|00001),
 .s_ino = 1,
};
struct iattr *foo()
{
    return sysfs_root.s_iattr;
}
umode_t bar()
{
    return sysfs_root.s_mode;
}
int main()
{
    if (foo() || bar() != 16877)
        abort();
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 09:27:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 09:27:10 -0500
Subject: [LLVMbugs] [Bug 4391] New: Pretty-printing nested unary plus/minus
 prints preincrement /predecrement
Message-ID: <bug-4391-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4391

           Summary: Pretty-printing nested unary plus/minus prints
                    preincrement/predecrement
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: bagnara at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu


$ cat bug.c
int g(void) {
  int a = + +3;
  int b = - -3;
}
$ clang-cc -ast-print bug.c
typedef __int128_t __int128_t;
typedef __uint128_t __uint128_t;
struct __va_list_tag {
    unsigned int gp_offset;
    unsigned int fp_offset;
    void *overflow_arg_area;
    void *reg_save_area;
};
int g() {
    int a = ++3;
    int b = --3;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 09:49:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 09:49:04 -0500
Subject: [LLVMbugs] [Bug 4392] New: llvm-gcc-4.2 configure script doesn't
 recognize makeinfo 4. 11
Message-ID: <bug-4392-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4392

           Summary: llvm-gcc-4.2 configure script doesn't recognize makeinfo
                    4.11
           Product: Build scripts
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: autoconf
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc-4.2/gcc/configure.ac thinks that makeinfo vesion 4.11 is
older than version 4.4, making it too old to use. This is due to
the following code:

# See if makeinfo has been installed and is modern enough
# that we can use it.
gcc_AC_CHECK_PROG_VER(MAKEINFO, makeinfo, --version,
  [GNU texinfo.* \([0-9][0-9.]*\)],
  [4.[4-9]*])

The 4.[4-9]* doesn't match 4.11. As a result, llvm-gcc-4.2 doesn't
build the documentation.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 12:14:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 12:14:35 -0500
Subject: [LLVMbugs] [Bug 4393] New: Default AA used instead of supplied AA
Message-ID: <bug-4393-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4393

           Summary: Default AA used instead of supplied AA
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: dpatel at apple.com, llvmbugs at cs.uiuc.edu


If a pass doesn't preserve AA (such as -simplify-cfg), then the passmanager
will remove the custom AA supplied on the commandline from AvailableAnalysis,
and thus the next time another pass needs AA (for example -gvn) it will
find the immutable basic AA as implementation.
The workaround is to manually rerun the custom AA after a pass that doesn't
preserve AA. This shouldn't be necessarry, the passmanager should remember
what the preferred implementation for an interface/analysisgroup is,
and rerun it if it was invalidated.

If I try to implement a custom AA as an immutablepass, then it is impossible to
use it, even if there is no other pass between -custom-aa and -gvn.
The reason is that the custom AA will get basicAA added as dependency (by the
the parent AliasAnalysis::getAnalysisUsage), and thus
ImmutablePasses will contain BasicAA, then my CustomAA
findAnalysis will thus find BasicAA as implementing AliasAnalysis, before my
pass, thus my pass is never chosen as implementing the interface.

The reason is that dependent analysis are added to various lists
(immutablepasses), before the analysis itself.

I think the passmanager shouldn't rely on the order analysis are added to
immutablepasses (since there is no way to add your own AA as first),
but instead use another Map that specifies the preferred implementation.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 15:58:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 15:58:18 -0500
Subject: [LLVMbugs] [Bug 4388] clang: preprocessor output missing newline
In-Reply-To: <bug-4388-206@http.llvm.org/bugs/>
Message-ID: <200906142058.n5EKwItK011156@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4388


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-14 15:58:17 ---


*** This bug has been marked as a duplicate of bug 2741 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 16:42:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 16:42:57 -0500
Subject: [LLVMbugs] [Bug 4390] clang: wrong struct initializer
In-Reply-To: <bug-4390-206@http.llvm.org/bugs/>
Message-ID: <200906142142.n5ELgv1A012800@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4390


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-14 16:42:56 ---
Thanks for the nice testcase!  Fixed in r73349.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 17:27:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 17:27:12 -0500
Subject: [LLVMbugs] [Bug 4332] Alignment related ICE with llvm-gcc 2.5
In-Reply-To: <bug-4332-206@http.llvm.org/bugs/>
Message-ID: <200906142227.n5EMRC2Z014407@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4332


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Duncan Sands <baldrick at free.fr>  2009-06-14 17:27:11 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090608/078682.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 17:42:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 17:42:25 -0500
Subject: [LLVMbugs] [Bug 4391] Pretty-printing nested unary plus/minus
 prints preincrement/ predecrement
In-Reply-To: <bug-4391-206@http.llvm.org/bugs/>
Message-ID: <200906142242.n5EMgPAJ014949@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4391


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-14 17:42:24 ---
Fixed in r73356, but -ast-print has a lot of broken edge cases at the moment if
you expect the output to be valid.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 19:52:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 19:52:53 -0500
Subject: [LLVMbugs] [Bug 4393] Default AA used instead of supplied AA
In-Reply-To: <bug-4393-206@http.llvm.org/bugs/>
Message-ID: <200906150052.n5F0qrjf019328@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4393


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-06-14 19:52:53 ---
This is correct behavior.  Not doing this would require rerunning the analysis
group implementation, which would be bad.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 14 20:25:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 14 Jun 2009 20:25:44 -0500
Subject: [LLVMbugs] [Bug 2741] PP output loses line count with -C
In-Reply-To: <bug-2741-206@http.llvm.org/bugs/>
Message-ID: <200906150125.n5F1Pii6020339@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2741


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-06-14 20:25:43 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090608/018166.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 01:57:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 01:57:06 -0500
Subject: [LLVMbugs] [Bug 4388] clang: preprocessor output missing newline
In-Reply-To: <bug-4388-206@http.llvm.org/bugs/>
Message-ID: <200906150657.n5F6v6qQ031161@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4388


T??r??k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|DUPLICATE                   |




--- Comment #2 from T??r??k Edwin <edwintorok at gmail.com>  2009-06-15 01:57:05 ---
This was not fixed when PR2741 was fixed:

$ clang -E -nostdinc -include x.h  -P -C foo.S
/* foo */# 2 "<built-in>" 2 3
/*
 * wakeup.ld
 *
 * Linker script for the real-mode wakeup code
 */

$ cat x.h
/* foo */
$ cat foo.S
/*
 * wakeup.ld
 *
 * Linker script for the real-mode wakeup code
 */


The #2 is still there in clang's output.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 04:58:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 04:58:35 -0500
Subject: [LLVMbugs] [Bug 4388] clang: preprocessor output missing newline
In-Reply-To: <bug-4388-206@http.llvm.org/bugs/>
Message-ID: <200906150958.n5F9wZcp016426@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4388


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-15 04:58:32 ---
Sorry about that; should actually be fixed in r73382.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 05:30:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 05:30:08 -0500
Subject: [LLVMbugs] [Bug 4395] New: clang: extra spaces in preprocessor
	output
Message-ID: <bug-4395-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4395

           Summary: clang: extra spaces in preprocessor output
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


When compiling the linux kernel, vmlinux.lds gets extra spaces in the output,
causing the linker to reject it.

Testcase:
#define NOTES .notes : AT(ADDR(.notes) - LOAD_OFFSET) {
VMLINUX_SYMBOL(__start_notes) = .; *(.note.*) VMLINUX_SYMBOL(__stop_notes) = .;
}

$ clang -E -P -C foo.h
.notes : AT(ADDR(.notes) - LOAD_OFFSET) { VMLINUX_SYMBOL(__start_notes) = .;
*(.note. *) VMLINUX_SYMBOL(__stop_notes) = .; }
$ gcc -E -P -C foo.h



.notes : AT(ADDR(.notes) - LOAD_OFFSET) { VMLINUX_SYMBOL(__start_notes) = .;
*(.note.*) VMLINUX_SYMBOL(__stop_notes) = .; }

Notice the difference: (.note.*) vs (.note. *)

If I manually remove those spaces (they occur in 3 places), then the linker
accepts vmlinux.lds.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 08:25:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 08:25:35 -0500
Subject: [LLVMbugs] [Bug 4396] New: Assertion `(PhysReg == 0 || !isChain) &&
 " Chain dependence via physreg data?"' failed.
Message-ID: <bug-4396-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4396

           Summary: Assertion `(PhysReg == 0 || !isChain) && "Chain
                    dependence via physreg data?"' failed.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3094)
 --> (http://llvm.org/bugs/attachment.cgi?id=3094)
The bitcode in question.

The attached (reduced) bitcode file produces the following assert in llc
-tailcallopt:
llc: ScheduleDAGSDNodes.cpp:201: void
llvm::ScheduleDAGSDNodes::AddSchedEdges(): Assertion `(PhysReg == 0 ||
!isChain) && "Chain dependence via physreg data?"' failed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 09:32:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 09:32:45 -0500
Subject: [LLVMbugs] [Bug 4360] clang crashes on solaris.
In-Reply-To: <bug-4360-206@http.llvm.org/bugs/>
Message-ID: <200906151432.n5FEWjIC025910@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4360


Edward O'Callaghan <eocallaghan at auroraux.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #3 from Edward O'Callaghan <eocallaghan at auroraux.org>  2009-06-15 09:32:45 ---
It looks like malloc_unlocked is a leaf routine which is attempting to fetch
some data from the parent's stack slot and that slot contains 'zero'.. which
malloc_unlocked isn't expecting.

After *much* work, this looks to be a mis-compile by Sun's old GCC.

Unfortunately there is not much I can do about that, Sun should update there
GCC to 4.x.x already ;(

> This page should include:

$ /usr/sfw/bin/gcc -v
Reading specs from /usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/specs
Configured with: /builds2/sfwnv-gate/usr/src/cmd/gcc/gcc-3.4.3/configure
--prefix=/usr/sfw --with-as=/usr/sfw/bin/gas --with-gnu-as
--with-ld=/usr/ccs/bin/ld --without-gnu-ld --enable-languages=c,c++,f77,objc
--enable-shared
Thread model: posix
gcc version 3.4.3 (csl-sol210-3_4-20050802)


< http://llvm.org/docs/GettingStarted.html#brokengcc

Closing as "WONTFIX", Unless someone else has a say, then please reopen.

Cheers,
Edward.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 09:46:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 09:46:29 -0500
Subject: [LLVMbugs] [Bug 4396] Assertion `(PhysReg == 0 || !isChain) && "
 Chain dependence via physreg data?"' failed.
In-Reply-To: <bug-4396-206@http.llvm.org/bugs/>
Message-ID: <200906151446.n5FEkT5H026437@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4396


Arnold Schwaighofer <arnold.schwaighofer at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Arnold Schwaighofer <arnold.schwaighofer at gmail.com>  2009-06-15 09:46:29 ---
CheckTailCallReturnConstraints is missing a check on the
incomming chain of the RETURN node. The incomming chain must
be the outgoing chain of the CALL node. This causes the
backend to identify tail calls that are not tail calls.

Fixed in <http://llvm.org/viewvc/llvm-project?view=rev&revision=73387>.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 10:47:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 10:47:46 -0500
Subject: [LLVMbugs] [Bug 4397] New: 2008-06-03-DomFrontier.ll test fails on
	Solaris.
Message-ID: <bug-4397-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4397

           Summary: 2008-06-03-DomFrontier.ll test fails on Solaris.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Good day,

FAIL:
/export/home/edward/lab/llvm/llvm/test/Transforms/LoopIndexSplit/2008-06-03-DomFrontier.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/export/home/edward/lab/llvm/llvm/test/Transforms/LoopIndexSplit/2008-06-03-DomFrontier.ll
|  opt -loop-rotate -loop-unswitch -loop-index-split -instcombine
-disable-output
Dominator Information for WirelessCreatePSK
Pass 'Rotate Loops'
----- Valid -----
  DomFrontier for BB%entry is:  
  DomFrontier for BB%bb52 is:   %bb52
  DomFrontier for BB%bb63 is:   %bb63%bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb68 is:   %bb63
  DomFrontier for BB%bb.nph is: %bb142.loopexit
  DomFrontier for BB%bb131 is:  %bb63
  DomFrontier for BB%bb142 is:  %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb142.loopexit is: %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb52.loopexit is:  %bb52.backedge
  DomFrontier for BB%bb63.preheader is: %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb142.preheader is:        %bb52.backedge
  DomFrontier for BB%bb52.backedge is:  %bb52
  DomFrontier for BB%bb131.clone is:    %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb131.bb63.clone_crit_edge is:    
%bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb89.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb63.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb.nph7 is:        %bb52.loopexit
  DomFrontier for BB%bb63.clone.bb142.loopexit_crit_edge is:    %bb142.loopexit
  DomFrontier for BB%bb142.bb52.loopexit_crit_edge is:  %bb52.loopexit
----- Invalid -----
  DomFrontier for BB%entry is:  
  DomFrontier for BB%bb52 is:   %bb52
  DomFrontier for BB%bb63 is:   %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb68 is:   %bb63
  DomFrontier for BB%bb.nph is: %bb142.loopexit
  DomFrontier for BB%bb131 is:  %bb63
  DomFrontier for BB%bb142 is:  %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb142.loopexit is: %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb52.loopexit is:  %bb52.backedge
  DomFrontier for BB%bb63.preheader is: %bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb142.preheader is:        %bb52.backedge
  DomFrontier for BB%bb52.backedge is:  %bb52
  DomFrontier for BB%bb131.clone is:    %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb131.bb63.clone_crit_edge is:    
%bb52.loopexit%bb63.preheader
  DomFrontier for BB%bb89.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb63.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb.nph7 is:        %bb52.loopexit
  DomFrontier for BB%bb63.clone.bb142.loopexit_crit_edge is:    %bb142.loopexit
  DomFrontier for BB%bb142.bb52.loopexit_crit_edge is:  %bb52.loopexit
Assertion failed: 0 && "Invalid dominator info", file PassManager.cpp, line 727
0   opt       0x089d8812 char const* std::find<char const*, char>(char const*,
char const*, char const&) + 376
1   opt       0x089d8fd0 llvm::sys::RemoveFileOnSignal(llvm::sys::Path const&,
std::string*) + 596
2   libc.so.1 0xfeb5d1df __sighndlr + 15
3   libc.so.1 0xfeb5025f call_user_handler + 687
4   libc.so.1 0xfeb623b5 _lwp_kill + 21
5   libc.so.1 0xfeb0ac2a raise + 34
6   libc.so.1 0xfeae203c abort + 116
7   libc.so.1 0xfeae230a __assert + 130
8   opt       0x08921d61 llvm::PMDataManager::verifyDomInfo(llvm::Pass&,
llvm::Function&) + 855
9   opt       0x087c71e7 llvm::LPPassManager::runOnFunction(llvm::Function&) +
1309
10  opt       0x08921eaf llvm::FPPassManager::runOnFunction(llvm::Function&) +
289
11  opt       0x0892205c llvm::FPPassManager::runOnModule(llvm::Module&) + 110
12  opt       0x08920cc8 llvm::MPPassManager::runOnModule(llvm::Module&) + 236
13  opt       0x0892230c llvm::PassManagerImpl::run(llvm::Module&) + 124
14  opt       0x08922371 llvm::PassManager::run(llvm::Module&) + 39
15  opt       0x084c060d main + 4771
16  opt       0x084997a0 _start + 128
Stack dump:
0.      Running pass 'Function Pass Manager' on module '<stdin>'.
1.      Running pass 'Loop Pass Manager' on function '@WirelessCreatePSK'

(gdb) bt
#0  0xfeb623b5 in _lwp_kill () from /lib/libc.so.1
#1  0xfeb5aadc in thr_kill () from /lib/libc.so.1
#2  0xfeb0ac2a in raise () from /lib/libc.so.1
#3  0xfeae20ba in abort () from /lib/libc.so.1
#4  0xfeae230a in _assert () from /lib/libc.so.1
#5  0x08932dbd in llvm::PMDataManager::verifyDomInfo (this=0x8ad9d98,
P=@0x8acf938, F=@0x8ad5180) at PassManager.cpp:727
#6  0x087d3683 in llvm::LPPassManager::runOnFunction (this=0x8ad9d88,
F=@0x8ad5180) at LoopPass.cpp:245
#7  0x08932f0b in llvm::FPPassManager::runOnFunction (this=0x8ad8268,
F=@0x8ad5180) at PassManager.cpp:1343
#8  0x089330b8 in llvm::FPPassManager::runOnModule (this=0x8ad8268,
M=@0x8accc40) at PassManager.cpp:1366
#9  0x08931d24 in llvm::MPPassManager::runOnModule (this=0x8ad6e78,
M=@0x8accc40) at PassManager.cpp:1411
#10 0x08933368 in llvm::PassManagerImpl::run (this=0x8ad4118, M=@0x8accc40) at
PassManager.cpp:1480
#11 0x089333cd in llvm::PassManager::run (this=0x8046eb0, M=@0x8accc40) at
PassManager.cpp:1509
#12 0x084c873d in main (argc=6, argv=0x8047088) at opt.cpp:483

bash-3.2$ uname -sv
SunOS snv_115
bash-3.2$ echo $PATH
/opt/gcc4/bin:/usr/xpg6/bin:/usr/xpg4/bin:/usr/bin:/usr/sbin:/usr/X11/bin:/usr/ccs/bin:/usr/sfw/bin:/usr/gnu/bin:/opt/csw/bin:/opt/clang/bin
bash-3.2$ which gcc
/opt/gcc4/bin/gcc
bash-3.2$ which llvm-as
/opt/clang/bin/llvm-as
bash-3.2$ llvm-as --version
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.6svn
  DEBUG build with assertions.
  Built Jun 14 2009(19:20:10).
bash-3.2$ gcc --version
gcc (GCC) 4.2.4

Regards,
Edward.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 14:26:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 14:26:52 -0500
Subject: [LLVMbugs] [Bug 4397] 2008-06-03-DomFrontier.ll test fails on
	Solaris.
In-Reply-To: <bug-4397-206@http.llvm.org/bugs/>
Message-ID: <200906151926.n5FJQqbU003983@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4397


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-06-15 14:26:52 ---
This is due to turning on expensive checks - it is not solaris
specific.  This is the remaining unsolved problem in PR4238.

*** This bug has been marked as a duplicate of bug 4238 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 14:48:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 14:48:10 -0500
Subject: [LLVMbugs] [Bug 4375] llvm-g++ trips over __extern_inline starting
	from -O1
In-Reply-To: <bug-4375-206@http.llvm.org/bugs/>
Message-ID: <200906151948.n5FJmAJx004942@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4375


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #7 from Duncan Sands <baldrick at free.fr>  2009-06-15 14:48:07 ---
I'm closing this because it's a gcc-4.2 problem that I doubt we want
to do anything about (especially as there's a workaround!).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 15 14:49:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 15 Jun 2009 14:49:06 -0500
Subject: [LLVMbugs] [Bug 4395] clang: extra spaces in preprocessor output
In-Reply-To: <bug-4395-206@http.llvm.org/bugs/>
Message-ID: <200906151949.n5FJn6gX005007@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4395


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-15 14:49:05 ---
Fixed in r73408.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 00:16:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 00:16:51 -0500
Subject: [LLVMbugs] [Bug 4336] Iterating over use-def chains doesn't seem to
 be deterministic.
In-Reply-To: <bug-4336-206@http.llvm.org/bugs/>
Message-ID: <200906160516.n5G5GpkA024728@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4336


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-06-16 00:16:50 ---
Fixed here, many thanks Julien!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090615/078780.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 02:00:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 02:00:31 -0500
Subject: [LLVMbugs] [Bug 4201] llvm.smul.with.overflow.i32 not implemented
	on PPC
In-Reply-To: <bug-4201-206@http.llvm.org/bugs/>
Message-ID: <200906160700.n5G70V4L028054@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4201


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-16 02:00:30 ---
Fixed in r73477.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 02:18:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 02:18:16 -0500
Subject: [LLVMbugs] [Bug 4398] New: passManager should always call
 releaseMemory to avoid use-after-free bugs
Message-ID: <bug-4398-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4398

           Summary: passManager should always call releaseMemory to avoid
                    use-after-free bugs
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: gohman at apple.com, llvmbugs at cs.uiuc.edu


Currently some passes call releaseMemory (or their equivalent) in their run
function, some don't (for example ScalarEvolution).

According to the documentation "This method is called after the run* method for
the class, before the next call of run* in your pass." 

However the PassManager doesn't do this for on-the-fly passes, which can cause
use-after-free bugs if you are using ScalarEvolution from a ModulePass 
(LoopInfo is rerun, but ScalarEvolution still has stale data from the previous
LoopInfo run).

As discusses on llvm-commits the following should happen:
 - PassManager should call releaseMemory for on-the-fly passes too
 - PassManager should assert if it forgot to call releaseMemory
 - passes that have a releaseMemory equivalent should have the method renamed
to releaseMemory
 - passes should no longer call releaseMemory in their run function


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 08:24:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 08:24:31 -0500
Subject: [LLVMbugs] [Bug 4400] New: Structure Alignment in MSVC
Message-ID: <bug-4400-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4400

           Summary: Structure Alignment in MSVC
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Keywords: portability
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: manoel at fonetica.com.br
                CC: llvmbugs at cs.uiuc.edu


In Windows plataform with  MSVC:

target datalayout =
"p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:1-f80:32:32"
        %FIRSTSTRU = type { i32,  double, i32, i8* }

If I define a strucutre like %FIRSTSTRU and a function like :

define i32 @STRUFUN(%FIRSTSTRU* %OUTRASTRU) {
entry:



   ret i32 1
}

 in the C side I do

struct  xfirstStru 
{
        int     var1;
        double  var2;
        int     var3;
        char    *var4;
};
typedef  struct xfirstStru      osX ;
typedef  osX* PosX ;

PosX    xf      = (PosX)        malloc( sizeof(osX) );
char *paux      = (char*)       malloc( 256 );
strcpy(paux,"teste C - LLVM");
memset(xf,0,sizeof(osX));
and we call the LLVM function from the C function,
fp_StruFun(xf);

the fields , from the double one, are not at the expected position.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 08:34:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 08:34:12 -0500
Subject: [LLVMbugs] [Bug 4400] Structure Alignment in MSVC
In-Reply-To: <bug-4400-206@http.llvm.org/bugs/>
Message-ID: <200906161334.n5GDYCJ0023173@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4400


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-06-16 08:33:54 ---
You have to insert necessary padding for the fields. For example, llvm-gcc for
mingw32 produces the following code:

        %struct.osX = type { i32, [4 x i8], double, i32, i8* }


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 12:40:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 12:40:53 -0500
Subject: [LLVMbugs] [Bug 4401] New: Assertion failure in ScheduleDAGRList:
 instruction released too many times
Message-ID: <bug-4401-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4401

           Summary: Assertion failure in ScheduleDAGRList: instruction
                    released too many times
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: reid.kleckner at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I'm working on unladen-swallow and I was running our regression suite with
python configured to always JIT compile everything, and I ran into this
assertion error:

*** Scheduling failed! ***
SU(1): 0x7f89479a2548: i64,ch = CopyFromReg 0x1238918, 0x7f8955387510
 has been released too many times!
python: ScheduleDAGRRList.cpp:205:
void<unnamed>::ScheduleDAGRRList::ReleasePred(llvm::SUnit*, const llvm::SDep*):
Assertion `0' failed.

My platform triple is: x86_64-unknown-linux-gnu

The suite takes hours to run because it tests a lot of Python edge cases by
generating absurdly large fragments of code and having Python execute them and
LLVM doesn't like absurdly large code fragments, so I ran it under GDB in case
it crashed.  I currently have it stopped at the assertion failure, and I'm
going to try to dig out the LLVM IR from the program so I can provide a stand
alone test case.

Here's the backtrace:

#0  0x00007f897a04f095 in raise () from /lib/libc.so.6
#1  0x00007f897a050af0 in abort () from /lib/libc.so.6
#2  0x00007f897a0482df in __assert_fail () from /lib/libc.so.6
#3  0x0000000000939418 in (anonymous namespace)::ScheduleDAGRRList::ReleasePred
    ()
#4  0x000000000093945a in (anonymous
namespace)::ScheduleDAGRRList::ReleasePredecessors ()
#5  0x000000000093a4bf in (anonymous
namespace)::ScheduleDAGRRList::ScheduleNodeBottomUp ()
#6  0x000000000093eb9c in (anonymous
namespace)::ScheduleDAGRRList::ListScheduleBottomUp ()
#7  0x000000000093fb60 in (anonymous namespace)::ScheduleDAGRRList::Schedule ()
#8  0x0000000000a1b241 in llvm::ScheduleDAG::Run ()
#9  0x00000000008c5fb4 in llvm::SelectionDAGISel::CodeGenAndEmitDAG ()
#10 0x00000000008c7339 in llvm::SelectionDAGISel::SelectBasicBlock ()
#11 0x00000000008c9615 in llvm::SelectionDAGISel::SelectAllBasicBlocks ()
#12 0x00000000008caa21 in llvm::SelectionDAGISel::runOnFunction ()
#13 0x0000000000ca4f55 in llvm::FPPassManager::runOnFunction ()
#14 0x0000000000ca538d in llvm::FunctionPassManagerImpl::run ()
#15 0x0000000000ca55b6 in llvm::FunctionPassManager::run ()
#16 0x00000000009a7ab9 in llvm::JIT::runJITOnFunctionUnlocked ()
#17 0x00000000009a82d1 in llvm::JIT::getPointerToFunction ()
#18 0x000000000057bcc1 in PyEval_EvalCodeEx (co=0x7f897b315160, 
    globals=0x7f88d965cda0, locals=0x7f88d965cda0, args=0x0, argcount=0, 
    kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0)
    at Python/eval.cc:3763
#19 0x000000000057cca2 in PyEval_EvalCode (co=0xd22, globals=0xd22, locals=0x6)
    at Python/eval.cc:541
#20 0x0000000000591bd0 in PyImport_ExecCodeModuleEx (
    name=0x7fff83354c05 "longlist", co=0x7f897b315160, 
    pathname=0x7fff83353b40 "longlist.pyo") at Python/import.c:687
#21 0x0000000000592065 in load_compiled_module (
    name=0x7fff83354c05 "longlist", cpathname=0x7fff83353b40 "longlist.pyo", 
    fp=<value optimized out>) at Python/import.c:815
#22 0x0000000000592c79 in import_submodule (mod=0x11cbc00, 
    subname=0x7fff83354c05 "longlist", fullname=0x7fff83354c05 "longlist")
    at Python/import.c:2552
#23 0x00000000005931cd in load_next (mod=0x527e050, altmod=0x11cbc00, 
    p_name=<value optimized out>, buf=0x7fff83354c00 "test.longlist", 
    p_buflen=0x7fff83354bf0) at Python/import.c:2376
#24 0x0000000000593539 in PyImport_ImportModuleLevel (
    name=<value optimized out>, globals=0x7f88da336d80, 
    locals=<value optimized out>, fromlist=0x11cbc00, level=-1)
    at Python/import.c:2094
#25 0x0000000000573549 in builtin___import__ (self=<value optimized out>, 
    args=<value optimized out>, kwds=<value optimized out>)
    at Python/bltinmodule.c:49
#26 0x00000000004fd81d in PyObject_Call (func=0x7f897b30a518, 
    arg=0x7f88d50e5f18, kw=0x0) at Objects/abstract.c:2487
#27 0x0000000000573d66 in PyEval_CallObjectWithKeywords (func=0x7f897b30a518, 
    arg=0x7f88d50e5f18, kw=0x0) at Python/eval.cc:42
#28 0x000000000056e980 in builtin_import_name (self=<value optimized out>, 
    args=<value optimized out>) at Python/bltinmodule.c:1170
#29 0x000000000057c8a2 in _PyEval_CallFunction (pp_stack=0x7fff83355de8, 
    oparg=<value optimized out>) at Python/eval.cc:3843
#30 0x00007f8908651cad in ?? ()
#31 0x00007f88da336d80 in ?? ()
#32 0x00007f8922969559 in ?? () from /usr/lib/libGL.so.1
#33 0x00007f897b32ca48 in ?? ()
#34 0x00007f891a72e3b0 in ?? ()
#35 0x0000000000000000 in ?? ()

As you can see it trails off into nonsense, because GDB can't make backtraces
for JITed code yet.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 13:02:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 13:02:47 -0500
Subject: [LLVMbugs] [Bug 4228] Duplicate debug symbol in recursively
	included file
In-Reply-To: <bug-4228-206@http.llvm.org/bugs/>
Message-ID: <200906161802.n5GI2lnT000695@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4228


devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from devang.patel <devang.patel at gmail.com>  2009-06-16 13:02:46 ---
Fixed.
revision 73520.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 13:12:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 13:12:27 -0500
Subject: [LLVMbugs] [Bug 4240] clang: broken debug info for restrict
In-Reply-To: <bug-4240-206@http.llvm.org/bugs/>
Message-ID: <200906161812.n5GICRMb001134@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4240


devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |devang.patel at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #3 from devang.patel <devang.patel at gmail.com>  2009-06-16 13:12:26 ---
This is a gdb bug.
Darwin gdb was recently fixed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 19:44:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 19:44:20 -0500
Subject: [LLVMbugs] [Bug 4402] New: Add pedantic checking for constant
 expressions to clang
Message-ID: <bug-4402-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4402

           Summary: Add pedantic checking for constant expressions to clang
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


>From Sema::CheckForConstantInitializer:
  // FIXME: Need strict checking.  In C89, we need to check for
  // any assignment, increment, decrement, function-calls, or
  // commas outside of a sizeof.  In C99, it's the same list,
  // except that the aforementioned are allowed in unevaluated
  // expressions.  Everything else falls under the
  // "may accept other forms of constant expressions" exception.
  // (We never end up here for C++, so the constant expression
  // rules there don't matter.)

Filing here for completeness.  This is relatively low-priority because it's a
non-trivial amount of work, and it's a warning nobody would care about outside
of -pedantic mode.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 20:01:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 20:01:14 -0500
Subject: [LLVMbugs] [Bug 4403] New: clang x86-64 ABI doesn' t set alignment
 correctly for byval struct containing SSE vector
Message-ID: <bug-4403-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4403

           Summary: clang x86-64 ABI doesn't set alignment correctly for
                    byval struct containing SSE vector
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
#include <emmintrin.h>
struct x {__m128 x,y,z;};
__m128 a(int a, int b, int c, int d, int e, int f, int g, struct x x) { return
x.x; }

The -emit-llvm output doesn't set the alignment for argument x, and as a result
the -S result contains an misaligned load.

r73306 is the parallel fix for x86-32; I think that the code for x86-64 should
be doing something similar.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 16 21:58:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 16 Jun 2009 21:58:12 -0500
Subject: [LLVMbugs] [Bug 3439] math wrong-code bug
In-Reply-To: <bug-3439-206@http.llvm.org/bugs/>
Message-ID: <200906170258.n5H2wC7i020504@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3439


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-16 21:58:08 ---
Fixed in r73598.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 00:39:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 00:39:32 -0500
Subject: [LLVMbugs] [Bug 4404] New: Incorrect X86 code for trivial LL
	assembly
Message-ID: <bug-4404-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4404

           Summary: Incorrect X86 code for trivial LL assembly
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: major
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan at fallingsnow.net
                CC: llvmbugs at cs.uiuc.edu


Given this C:

int load() {
  int* ptr = (int*)0xabcfed;
  return *ptr;
}


Generates LL (using llvm-gcc -emit-llvm -O2 -S -o scratch/load.ll
scratch/load.c):

; ModuleID = 'scratch/load.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i386-apple-darwin9"

define i32 @load() nounwind  {
entry:
        load i32* inttoptr (i64 11259885 to i32*), align 4              ;
<i32>:0 [#uses=1]
        ret i32 %0
}

Generates this x86 (using llvm-as < scratch/load.ll | llc -march=x86):

        .text
        .align  4,0x90
        .globl  _load
_load:
LBB1_0: ## entry
        movl    11259885, %eax
        ret

        .subsections_via_symbols


This seems quite wrong. The x86 code contains no load from the address, just
the address itself left in %eax.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 00:41:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 00:41:38 -0500
Subject: [LLVMbugs] [Bug 3455] another spurious divide by zero
In-Reply-To: <bug-3455-206@http.llvm.org/bugs/>
Message-ID: <200906170541.n5H5fcJw025780@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3455


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-17 00:41:37 ---
This seems to be working now.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 00:42:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 00:42:07 -0500
Subject: [LLVMbugs] [Bug 4404] Incorrect X86 code for trivial LL assembly
In-Reply-To: <bug-4404-206@http.llvm.org/bugs/>
Message-ID: <200906170542.n5H5g7Ii025818@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4404


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-06-17 00:42:07 ---
That is a load.  $ indicates a constant.  With -print-machineinstrs you can see
this is a load:

        %EAX<def> = MOV32rm %reg0, 1, %reg0, 11259885, %reg0, Mem:LD(4,4)
[0x190200c + 0]

Alternatively, with intel syntax:

        mov     EAX, DWORD PTR [11259885]


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 00:53:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 00:53:24 -0500
Subject: [LLVMbugs] [Bug 790] Portability issues for Win32
In-Reply-To: <bug-790-206@http.llvm.org/bugs/>
Message-ID: <200906170553.n5H5rORV026136@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=790


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #13 from Chris Lattner <clattner at apple.com>  2009-06-17 00:53:23 ---
Yes, please file new bugzillas about remaining individual issues.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 08:35:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 08:35:28 -0500
Subject: [LLVMbugs] [Bug 4405] New: Miniperl dies with segmentation fault
 when optimizer is turned on
Message-ID: <bug-4405-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4405

           Summary: Miniperl dies with segmentation fault when optimizer is
                    turned on
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Someone (me?) still needs to track down the exact issue, but right now I'm
seeing segmentation faults of miniperl when building Perl 5.8.9 on
FreeBSD/amd64 when using LLVM and Clang r73340.

The issue can easily be reproduced by building Perl with CFLAGS=-O2.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 11:58:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 11:58:34 -0500
Subject: [LLVMbugs] [Bug 4406] New: BrainF example fails with assertion
	failure
Message-ID: <bug-4406-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4406

           Summary: BrainF example fails with assertion failure
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Running hello.bf with -jit works (it prints the hello world message), but there
is an assertion failure when it quits:

$ Release/examples/BrainF  hello.bf -jit
------- Running JIT -------
Hello World!
While deleting: i32 (i32)* %putchar
An asserting value handle still pointed to this value!
Aborted

Running it without the JIT doesn't print anything, and there is no assertion
failure either:
$ Release/examples/BrainF  hello.bf

hello.bf contains this (as a single line), which is the example from Wikipedia:
++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>.

I don't know when this bug was introduced, I just noticed it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 13:06:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 13:06:23 -0500
Subject: [LLVMbugs] [Bug 4408] New: Incorrect inlining on function with
 complex parameters
Message-ID: <bug-4408-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4408

           Summary: Incorrect inlining on function with complex parameters
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


The inliner seems to generate incorrect code, the initial icmp on length in
memcpy_custom is removed after inlining.

--orignal c source code ------------
extern char __DATA_START,__TEXT_END;
extern char __BSS_START,__BSS_LENGTH;
extern int  __DATA_LENGTH;

void* memcpy_custom(void* dst,const void* src, unsigned int length);
void a() 
{
  memcpy_custom(&__DATA_START,&__TEXT_END,(unsigned int)&__DATA_LENGTH);
}

void* memcpy_custom(void* dst,const void* src, unsigned int length)
{
  char* dst_char=dst;
  const char* src_char=src;
  while(length--)
  {
    *dst_char++=*src_char++;
  }
  return 0;
}

----------llvm code after clang-cc --emit-llvm test.c -O3

; ModuleID = 'test.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
target triple = "i386-pc-linux-gnu"
@__DATA_START = external global i8              ; <i8*> [#uses=1]
@__TEXT_END = external global i8                ; <i8*> [#uses=1]
@__DATA_LENGTH = external global i32            ; <i32*> [#uses=1]

define void @a() nounwind {
entry:
        br label %while.body.i

while.body.i:           ; preds = %while.body.i, %entry
        %indvar.i = phi i32 [ 0, %entry ], [ %indvar.next.i, %while.body.i ]   
        ; <i32> [#uses=3]
        %dst_char.09.i = getelementptr i8* @__DATA_START, i32 %indvar.i        
; <i8*> [#uses=1]
        %src_char.010.i = getelementptr i8* @__TEXT_END, i32 %indvar.i         
; <i8*> [#uses=1]
        %tmp5.i = load i8* %src_char.010.i              ; <i8> [#uses=1]
        store i8 %tmp5.i, i8* %dst_char.09.i
        %indvar.next.i = add i32 %indvar.i, 1           ; <i32> [#uses=2]
        %exitcond = icmp eq i32 %indvar.next.i, ptrtoint (i32* @__DATA_LENGTH
to i32)           ; <i1> [#uses=1]
        br i1 %exitcond, label %memcpy_custom.exit, label %while.body.i

memcpy_custom.exit:             ; preds = %while.body.i
        ret void
}

define noalias i8* @memcpy_custom(i8* nocapture %dst, i8* nocapture %src, i32
%length) nounwind {
entry:
        %tobool13 = icmp eq i32 %length, 0              ; <i1> [#uses=1]
        br i1 %tobool13, label %while.end, label %while.body

while.body:             ; preds = %while.body, %entry
        %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %while.body ]         
; <i32> [#uses=3]
        %src_char.010 = getelementptr i8* %src, i32 %indvar             ; <i8*>
[#uses=1]
        %dst_char.09 = getelementptr i8* %dst, i32 %indvar              ; <i8*>
[#uses=1]
        %tmp5 = load i8* %src_char.010          ; <i8> [#uses=1]
        store i8 %tmp5, i8* %dst_char.09
        %indvar.next = add i32 %indvar, 1               ; <i32> [#uses=2]
        %exitcond = icmp eq i32 %indvar.next, %length           ; <i1>
[#uses=1]
        br i1 %exitcond, label %while.end, label %while.body

while.end:              ; preds = %while.body, %entry
        ret i8* null
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 16:08:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 16:08:08 -0500
Subject: [LLVMbugs] [Bug 4408] Incorrect inlining on function with complex
	parameters
In-Reply-To: <bug-4408-206@http.llvm.org/bugs/>
Message-ID: <200906172108.n5HL88XB005240@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4408


Sylv??re Teissier <quickslyver at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 16:22:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 16:22:38 -0500
Subject: [LLVMbugs] [Bug 4409] New: LLVM needs a tool for pattern match LLVM
	IR
Message-ID: <bug-4409-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4409

           Summary: LLVM needs a tool for pattern match LLVM IR
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DejaGNU
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Trying to match patterns in LLVM IR using grep is painful and error prone,
especially when trying to look for a pattern that spans multiple instructions.

We should implement some kind of basic tool for match parts of the LLVM IR
against a pattern for use in writing simple test cases.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 19:59:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 19:59:32 -0500
Subject: [LLVMbugs] [Bug 4298] Strange position-independent code,
 relocatable references problem.
In-Reply-To: <bug-4298-206@http.llvm.org/bugs/>
Message-ID: <200906180059.n5I0xW1L013485@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4298


Edward O'Callaghan <eocallaghan at auroraux.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #4 from Edward O'Callaghan <eocallaghan at auroraux.org>  2009-06-17 19:59:30 ---
Looks like this is a broken GCC thus a side effect of yet another GCC bug :p

Unless anyone else objects, closing as WONTFIX.

Sorry to wast time,
Thanks,
Edward.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 17 20:55:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 17 Jun 2009 20:55:20 -0500
Subject: [LLVMbugs] [Bug 4410] New: incorrect expansion for uint_to_fp
Message-ID: <bug-4410-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4410

           Summary: incorrect expansion for uint_to_fp
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: benedict.gaster at amd.com
                CC: llvmbugs at cs.uiuc.edu


Consider the following C code:

float convert(uint64_t u)
{
    return (float)u;
}

which as LLVM IR is:

LL:
define float @convert(i64 %a) nounwind {
entry:
        %tmp12 = uitofp i64 %a to float         ; <float> [#uses=1]
        ret float %tmp12
}

Compiling this with llc -march=x86-64 gives:

ASM (amd64):
.LCPI1_0:                                       # i64
        .quad   6881500230622117888
convert:
        cvtsi2ssq       %rdi, %xmm0
        testq   %rdi, %rdi
        sets    %al
        movzbl  %al, %eax
        addss   .LCPI1_0(,%rax,4), %xmm0
        ret

The ???C??? equivalent code is:

float convert(uint64_t u)
{
    static float fudge[] = { 0.0f, 0x1.0p64f };
    float f = (float)(int64_t)u;
    f += fudge[msb(u)];
    return f;
}

The adjustment by 2^64 to counteract the effect of converting treating ???u???
as a signed value does not address rounding properly.
For example, in rounding to the nearest even mode, converting
0xffffff7fffffffffull with this code returns 0x1p+64 instead of the expected
0x1.fffffep+63.

To produce the correct values in all rounding modes, the following expansion
should be used instead:

float convert(uint64_t u)
{
    if (msb(u)) {
        return 2.0f * (float)(int64_t)((u >> 1) | (u & 1));
    }
    else {
        return (float)(int64_t)f;
    }
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 18 07:17:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 18 Jun 2009 07:17:25 -0500
Subject: [LLVMbugs] [Bug 4411] New: Durring compilation from GCC front end
 to IR signed/ unsigned type information gets lost
Message-ID: <bug-4411-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4411

           Summary: Durring compilation from GCC front end to IR
                    signed/unsigned type information gets lost
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: hisu00 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3101)
 --> (http://llvm.org/bugs/attachment.cgi?id=3101)
Example

Take a look at the attached C source code (typetest.c).

The function open takes a typed function pointer as parameter (nfy parameter
with type nfy_fct). nfy_fct function type takes an unsigned int and a short
parameter.

link_nfy_fct function (which is from type nfy_fct) is compiled to:
define internal void @link_nfy_fct(i32 %event, i16 signext %cond) nounwind

The header of open function is transformed to:
define i32 @open(i32 %linktype, i8* %linkname, i8* %linkoptions, i32 %timeout,
void (i32, i16)* %nfy) nounwind

where the type for nfy should be void (i32, i16 signext)* instead of void (i32,
i16)*. With this the information gets lost in C to IR phase and the C Backend
produces code which can not be compiled with some other strict compilers.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 18 08:30:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 18 Jun 2009 08:30:45 -0500
Subject: [LLVMbugs] [Bug 4412] New: Cairo library is broken -- library
 cannot be linked against properly
Message-ID: <bug-4412-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4412

           Summary: Cairo library is broken -- library cannot be linked
                    against properly
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


I still have to figure this out in more detail. nm(1) shows no difference, but
it's still fishy.

Anyway, here it goes: some Pango libs/apps won't link against cairo properly if
cairo is compiled with Clang. Cairo uses the following code to export its
functions. It looks quite complicated if you ask me.

void 
cairo_rel_line_to(void)
{

        puts("Hi");
}
extern __typeof (cairo_rel_line_to) cairo_rel_line_to __asm__ ("_"
"INT_cairo_rel_line_to") __attribute__((__visibility__("hidden")));
extern __typeof (cairo_rel_line_to) EXT_cairo_rel_line_to __asm__("_"
"cairo_rel_line_to") __attribute__((__alias__("_" "INT_cairo_rel_line_to")));


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 18 23:23:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 18 Jun 2009 23:23:34 -0500
Subject: [LLVMbugs] [Bug 4323] Tail recursion elimination doesn' t move
 loads above nounwind readonly call
In-Reply-To: <bug-4323-206@http.llvm.org/bugs/>
Message-ID: <200906190423.n5J4NYXc007086@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4323


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #13 from Chris Lattner <clattner at apple.com>  2009-06-18 23:23:34 ---
Patch applied here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090615/079051.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 01:20:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 01:20:13 -0500
Subject: [LLVMbugs] [Bug 4413] New: Missed optimization in select on
 correlated phi nodes
Message-ID: <bug-4413-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4413

           Summary: Missed optimization in select on correlated phi nodes
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3102)
 --> (http://llvm.org/bugs/attachment.cgi?id=3102)
Assembly with missed optimization.

`opt -O3` at head fails to optimize the following function:

declare i32 @ext()

define i32 @foo(i1 %a) {
entry:
        br i1 %a, label %bb1, label %bb2

bb1:            ; preds = %entry
        %0 = tail call i32 @ext()               ; <i32> [#uses=1]
        br label %bb2

bb2:            ; preds = %bb1, %entry
        %cond = phi i1 [ true, %bb1 ], [ false, %entry ]                ; <i1>
[#uses=1]
        %val = phi i32 [ %0, %bb1 ], [ 0, %entry ]              ; <i32>
[#uses=1]
        %res = select i1 %cond, i32 %val, i32 0         ; <i32> [#uses=1]
        ret i32 %res
}


Despite the fact that %res is equal to %val for every in-edge to %bb2, so the
select could be removed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 02:01:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 02:01:24 -0500
Subject: [LLVMbugs] [Bug 2484] <4 x float> SSE shuffle breaks without SSE2
In-Reply-To: <bug-2484-206@http.llvm.org/bugs/>
Message-ID: <200906190701.n5J71OLP013094@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2484


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2009-06-19 02:01:19 ---
Fixed in r73760.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 08:25:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 08:25:35 -0500
Subject: [LLVMbugs] [Bug 4414] New: SelectionDAG:: getMemset generate
 incorrect libcall for platforms with int!=32b
Message-ID: <bug-4414-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4414

           Summary: SelectionDAG::getMemset generate incorrect libcall for
                    platforms with int!=32b
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


function SelectionDAG::getMemset (SelectionDAG.cpp)
generate a lib call to 
void *memset(void *s, int c, size_t n);
assuming that 'int c' is an int32, that's not true for all target !


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 10:02:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 10:02:40 -0500
Subject: [LLVMbugs] [Bug 3976] Assert in TargetLowering::SimplifySetCC when
 using intergers >= 256 bits
In-Reply-To: <bug-3976-206@http.llvm.org/bugs/>
Message-ID: <200906191502.n5JF2eXL010482@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3976


Damien <ddneff at hotmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Damien <ddneff at hotmail.com>  2009-06-19 10:02:31 ---
Since llvm 2.5 the offending code has been changed so that it checks to ensure
the value can fit in 64 bits before proceeding.  This fixes the assertion so it
will no longer fire, so I am resolving the bug, but I don't like the fix.  It
completely disables the optimization for types over 64 bits.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 12:27:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 12:27:12 -0500
Subject: [LLVMbugs] [Bug 4415] New: crash with -indvars
Message-ID: <bug-4415-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4415

           Summary: crash with -indvars
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: arplynn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3106)
 --> (http://llvm.org/bugs/attachment.cgi?id=3106)
Test case.

The following testcase (run with opt -indvars bugpoint-reduced-simplified.bc)
causes the following death in opt:

0   opt               0x004927f8 std::_Rb_tree<llvm::sys::Path,
llvm::sys::Path, std::_Identity<llvm::sys::Path>, std::less<llvm::sys::Path>,
std::allocator<llvm::sys::Path> >::insert_unique(llvm::sys::Path const&) + 8504
1   opt               0x00492d7a std::_Rb_tree<llvm::sys::Path,
llvm::sys::Path, std::_Identity<llvm::sys::Path>, std::less<llvm::sys::Path>,
std::allocator<llvm::sys::Path> >::insert_unique(llvm::sys::Path const&) + 9914
2   libSystem.B.dylib 0x970f62bb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1760599407
4   opt               0x002c8f06 llvm::Pass* llvm::callDefaultCtor<(anonymous
namespace)::LoaderPass>() + 56950
5   opt               0x002c9489 llvm::Pass* llvm::callDefaultCtor<(anonymous
namespace)::LoaderPass>() + 58361
6   opt               0x002c971c llvm::Pass* llvm::callDefaultCtor<(anonymous
namespace)::LoaderPass>() + 59020
7   opt               0x000afc12 llvm::DenseMap<llvm::BasicBlock*,
llvm::Value*, llvm::DenseMapInfo<llvm::BasicBlock*>,
llvm::DenseMapInfo<llvm::Value*> >::find(llvm::BasicBlock* const&) + 22114
8   opt               0x002a0c16 llvm::LoopInfo::print(std::ostream&,
llvm::Module const*) const + 4662
9   opt               0x00404139 llvm::FunctionPass::~FunctionPass() + 38825
10  opt               0x0040464d llvm::FunctionPass::~FunctionPass() + 40125
11  opt               0x00404d4e llvm::FunctionPass::~FunctionPass() + 41918
12  opt               0x00405224 llvm::FunctionPass::~FunctionPass() + 43156
13  opt               0x0040530b llvm::FunctionPass::~FunctionPass() + 43387
14  opt               0x00009d33 llvm::scc_iterator<llvm::CallGraphNode*,
llvm::GraphTraits<llvm::CallGraphNode*> >
llvm::scc_end<llvm::CallGraphNode*>(llvm::CallGraphNode*) + 7251
15  opt               0x00002166 _mh_execute_header + 4454
Stack dump:
0.      Running pass 'Function Pass Manager' on module
'bugpoint-reduced-simplified.bc'.
1.      Running pass 'Loop Pass Manager' on function '@foo'
2.      Running pass 'Canonicalize Induction Variables' on basic block
'%while.cond'
Bus error


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 14:28:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 14:28:12 -0500
Subject: [LLVMbugs] [Bug 4416] New: Assertion while compiling llvm-gcc
Message-ID: <bug-4416-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4416

           Summary: Assertion while compiling llvm-gcc
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bharadwajy at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3107)
 --> (http://llvm.org/bugs/attachment.cgi?id=3107)
mf-heuristics.bc

I encounter the following error during cross compilation of compilation of
llvm-gcc (version 73777). I am using llvm (version 73777).

cc1:
<some-dir>llvm-root/src/llvm/include/llvm/CodeGen/LiveIntervalAnalysis.h:134:
llvm::LiveInterval& llvm::LiveIntervals::getInterval(unsigned int): Assertion
`I != r2iMap_.end() && "Interval does not exist for register"' failed.
../../../../../src/llvm-gcc-4.2/libmudflap/mf-heuristics.c:175: internal
compiler error: Aborted

I am attaching mf-heuristics.bc obtainied by compiling it with -emit-llvm.
Unfortunately running bugpoint on mf-heuristics.bc yielded in simplified-*.bc
files which exhibited a different failure.

Thanks for your help.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 17:11:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 17:11:04 -0500
Subject: [LLVMbugs] [Bug 4418] New: JIT broken on MingW for
 available_externally functions.
Message-ID: <bug-4418-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4418

           Summary: JIT broken on MingW for available_externally functions.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jlerouge at apple.com
                CC: llvmbugs at cs.uiuc.edu


The following C code does not work with the JIT anymore on MingW, since rev
72619.

$ cat t.c                                                                       
#include <stdio.h>                                                              

int main() {                                                                    
    return getchar();                                                           
}               

$ llvm-gcc -c -emit-llvm -o t.bc t.c  
$ lli t.bc                                                                      
Assertion failed: Addr && "Code generation didn't add function to GlobalAddress
table!", file                                                                   
c:/cygwin/home/jlerouge/buildbot/llvm-test-fixed-src/lib/ExecutionEngine/JIT/JIT.cpp,
line 603        

An explanation / discussion was started on the LLVM commit:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090601/078411.html

Thanks,
Julien


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 17:52:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 17:52:05 -0500
Subject: [LLVMbugs] [Bug 4419] New: Register Scavenger assertion
Message-ID: <bug-4419-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4419

           Summary: Register Scavenger assertion
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Register Allocator
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3108)
 --> (http://llvm.org/bugs/attachment.cgi?id=3108)
Reduced testcases

Consider the bytecode attached.

./llc bugpoint-reduced-simplified.bc -mattr=+vfp2 -mtriple=arm-linux-eabi -f
-float-abi=hard

yields:

llc: /home/asl/proj/llvm/src_arm/lib/CodeGen/RegisterScavenging.cpp:279: void
llvm::RegScavenger::forward(): Assertion `(isReserved(Reg) || isUnused(Reg) ||
IsImpDef || isImplicitlyDefined(Reg) || isLiveInButUnusedBefore(Reg, MI, MBB,
TRI, MRI)) && "Re-defining a live register!"' failed.

That's due to the following code:

        BL <ga:__ieee754_sqrtf>, %S0<kill>, %R0<imp-def,dead>,
%R1<imp-def,dead>, %R2<imp-def,dead>, %R3<imp-def,dead>, %R12<imp-def,dead>,
%LR<imp-def,dead>, %D0<imp-def>, %D1<imp-def,dead>, %D2<imp-def,dead>,
%D3<imp-def,dead>, %D4<imp-def,dead>, %D5<imp-def,dead>, %D6<imp-def,dead>,
%D7<imp-def,dead>, %CPSR<imp-def,dead>
        ADJCALLSTACKUP 0, 0, 14, %reg0, %SP<imp-def>, %SP<imp-use>
        %S1<def> = FLDS <cp#0>, 0, 14, %reg0, Mem:LD(4,4) [ConstantPool + 0]

S1 is imp-def on call due to imp-def on D0. It's redef'ed later on FLDS. This
sounds like livevars bug which triggers regscavenger assertion later.

I'm attaching both orioginal and reduced testcases.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 20:08:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 20:08:07 -0500
Subject: [LLVMbugs] [Bug 4420] New: Missed const propagation from condition
 of branch to phi node at landing
Message-ID: <bug-4420-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4420

           Summary: Missed const propagation from condition of branch to phi
                    node at landing
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Created an attachment (id=3110)
 --> (http://llvm.org/bugs/attachment.cgi?id=3110)
Assembly with missed optimization.

In the following assembly (which doesn't change under head's `opt -O3`:

declare i1 @ext()
define i1 @foo() {
entry:
        %cond = tail call i1 @ext()             ; <i1> [#uses=2]
        br i1 %cond, label %bb1, label %bb2

bb1:            ; preds = %entry
        %cond2 = tail call i1 @ext()            ; <i1> [#uses=1]
        br i1 %cond2, label %bb3, label %bb2

bb2:            ; preds = %bb1, %entry
        %cond_merge = phi i1 [ %cond, %entry ], [ false, %bb1 ]         ; <i1>
[#uses=1]
        ret i1 %cond_merge

bb3:            ; preds = %bb1
        %res = tail call i1 @ext()              ; <i1> [#uses=1]
        ret i1 %res
}



%entry has "br i1 %cond, label %bb1, label %bb2". If this jumps to %bb2, we
know %cond was false. But %bb2 has "%cond_merge = phi i1 [ %cond, %entry ]...",
in which it could have inlined %cond as false.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 19 23:35:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 19 Jun 2009 23:35:19 -0500
Subject: [LLVMbugs] [Bug 4419] Register Scavenger assertion
In-Reply-To: <bug-4419-206@http.llvm.org/bugs/>
Message-ID: <200906200435.n5K4ZJN9004890@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4419


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-06-19 23:35:16 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090615/079121.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 20 02:10:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 20 Jun 2009 02:10:51 -0500
Subject: [LLVMbugs] [Bug 4421] New: LICM doesn't preserve AA
Message-ID: <bug-4421-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4421

           Summary: LICM doesn't preserve AA
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


LICM declares itself as requiring AliasAnalysis but not preserving it. That's
odd, it should probably preserve AA.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 20 10:42:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 20 Jun 2009 10:42:01 -0500
Subject: [LLVMbugs] [Bug 4422] New: Incorrect alignment from alloca
Message-ID: <bug-4422-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4422

           Summary: Incorrect alignment from alloca
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lennart at augustsson.net
                CC: llvmbugs at cs.uiuc.edu


When allocating memory with alloca for vectors on i386 the returned memory is
not always on 16 byte boundary as it has to be.
E.g.
   a = alloca <2 x i64>
sometimes returns memory on 8 byte boundary, which means that using MMS
instructions to load and store will cause a fault.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 20 20:03:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 20 Jun 2009 20:03:25 -0500
Subject: [LLVMbugs] [Bug 4424] New: instcombine creates constexpr sdiv(0,
	x) instead of
Message-ID: <bug-4424-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4424

           Summary: instcombine creates constexpr sdiv(0, x) instead of 0
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: nicholas at mxc.ca
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu


$ cat ~/tmp/constfold_missed.ll 
; ModuleID = '<stdin>'

declare void @ext()

define i32 @foo(i32 %ptr) {
entry:
        %zero = sub i32 %ptr, %ptr              ; <i32> [#uses=1]
        %div_zero = sdiv i32 %zero, ptrtoint (i32* getelementptr (i32* null,
i32 1) to i32)             ; <i32> [#uses=1]
        ret i32 %div_zero
}

$ llvm-as < ~/tmp/constfold_missed.ll |opt -instcombine|llvm-dis
; ModuleID = '<stdin>'

declare void @ext()

define i32 @foo(i32 %ptr) {
entry:
        ret i32 sdiv (i32 0, i32 ptrtoint (i32* getelementptr (i32* null, i32
1) to i32))
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 20 20:16:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 20 Jun 2009 20:16:40 -0500
Subject: [LLVMbugs] [Bug 4424] instcombine creates constexpr sdiv(0,
	x) instead of 0
In-Reply-To: <bug-4424-206@http.llvm.org/bugs/>
Message-ID: <200906210116.n5L1Geov021640@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4424


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-06-20 20:16:39 ---
Implemented here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090615/079135.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 21 07:05:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 21 Jun 2009 07:05:04 -0500
Subject: [LLVMbugs] [Bug 4185] x87 inline asm assertion
In-Reply-To: <bug-4185-206@http.llvm.org/bugs/>
Message-ID: <200906211205.n5LC54BM012010@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4185


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-06-21 07:04:58 ---
Fixed in revision 73850.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 03:53:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 03:53:13 -0500
Subject: [LLVMbugs] [Bug 4427] New: LLVM_NATIVE_ARCH in config.h conflicts
 with compiler/ buildsystem platform arch defines
Message-ID: <bug-4427-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4427

           Summary: LLVM_NATIVE_ARCH in config.h conflicts with
                    compiler/buildsystem platform arch defines
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3111)
 --> (http://llvm.org/bugs/attachment.cgi?id=3111)
config.h.cmake.patch

When compiling icedtea6 that have beenm modified to include TargetSelect.h and
use InitializeNativeTarget() my build fail with this error when compiling on
ia32:

In file included from
/home/xerxes/icedtea6/openjdk/hotspot/src/share/vm/shark/llvmHeaders.hpp:39,
                 from ../generated/incls/_compileBroker.cpp.incl:9,
                 from
/home/xerxes/icedtea6/openjdk/hotspot/src/share/vm/compiler/compileBroker.cpp:26:
/usr/local/include/llvm/Target/TargetSelect.h: In function 'bool
llvm::InitializeNativeTarget()':
/usr/local/include/llvm/Target/TargetSelect.h:55: error: 'InitializeTarget' is
not a member of 'llvm'

The cause are that the icedtea6 buildsystem have defined X86 and this conflicts
with the LLVM_NATIVE_ARCH define in config.h.
The conflict makes LLVM_NATIVE_ARCH contan an empty string instead of the X86
string and this in turn makes the macro in config.h create InitializeTarget
instead of InitializeX86Target


The attached patch includes a workaround for this issue when building llvm
using cmake.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 05:35:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 05:35:29 -0500
Subject: [LLVMbugs] [Bug 4428] New: Assertion failed: (BitWidth ==
 RHS.BitWidth && " Bit widths must be same for comparison"), function ult,
 file llvm/lib/  Support/APInt.cpp, line 492.
Message-ID: <bug-4428-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4428

           Summary: Assertion failed: (BitWidth == RHS.BitWidth && "Bit
                    widths must be same for comparison"), function ult, file
                    llvm/lib/Support/APInt.cpp, line 492.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


The following code causes Clang to crash when pulled through the loop
optimizer:

void
boom(void)
{
        char buf[64];
        char l;
        int zext;
        char *p, *pe;

        p = 1;
        pe = buf;
        zext = p - pe;
        l += zext;

        while (zext > 0 && p > buf) {
                p--;
                zext--;
        }

        while (l > 0 && p > buf);
}

; ModuleID = 'llvm-crash.c'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "x86_64-undermydesk-freebsd8.0"
@p = internal constant [1 x i8] c"\01", align 1         ; <[1 x i8]*> [#uses=1]

define i32 @crash(i8* nocapture %source) nounwind readonly {
entry:  
        %tmp1 = load i8* %source                ; <i8> [#uses=3]
        %idxprom = zext i8 %tmp1 to i64         ; <i64> [#uses=1]
        %arrayidx = getelementptr [1 x i8]* @p, i64 0, i64 %idxprom            
; <i8*> [#uses=1]
        %tmp2 = load i8* %arrayidx              ; <i8> [#uses=1]
        %cmp = icmp eq i8 %tmp2, 4              ; <i1> [#uses=1]
        br i1 %cmp, label %land.rhs, label %land.end

land.rhs:               ; preds = %entry
        %cmp7 = icmp sgt i8 %tmp1, -1           ; <i1> [#uses=1]
        br i1 %cmp7, label %land.end, label %lor.rhs

lor.rhs:                ; preds = %land.rhs
        %cmp12 = icmp ugt i8 %tmp1, -65         ; <i1> [#uses=1]
        %phitmp = zext i1 %cmp12 to i32         ; <i32> [#uses=1]
        ret i32 %phitmp

land.end:               ; preds = %land.rhs, %entry
        %0 = phi i32 [ 0, %entry ], [ 1, %land.rhs ]            ; <i32>
[#uses=1]
        ret i32 %0
}

cc: warning: argument unused during compilation: '-c'
llvm-crash.c:9:4: warning: incompatible integer to pointer conversion assigning
'int', expected 'char *'
        p = 1;
          ^ ~
Assertion failed: (BitWidth == RHS.BitWidth && "Bit widths must be same for
comparison"), function ult, file llvm/lib/Support/APInt.cpp, line 492.
Stack dump:
0.      Program arguments: clang-cc -triple x86_64-undermydesk-freebsd8.0 -S
-disable-free -main-file-name llvm-crash.c --relocation-model static
--disable-fp-elim --unwind-tables=1 --mcpu=x86-64 --fmath-errno=1 -O2
-std=gnu89 -fdiagnostics-show-option -o llvm-crash -x c llvm-crash.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Loop Pass Manager' on function '@boom'
4.      Running pass 'Loop Strength Reduction' on basic block '%while.cond'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 06:00:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 06:00:17 -0500
Subject: [LLVMbugs] [Bug 4429] New: ExecutionEngine JIT and Interpreter
 generated as . o when using cmake and makes icedtea6 linking fail.
Message-ID: <bug-4429-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4429

           Summary: ExecutionEngine JIT and Interpreter generated as .o when
                    using cmake and makes icedtea6 linking fail.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3112)
 --> (http://llvm.org/bugs/attachment.cgi?id=3112)
xecutionEngine_JIT_Interpreter_o_to_a.patch cmake fix

When building icedtea6 (OpenJDK) using shark i have started to get a linking
error after llvm r73709 

icedtea6 fist creates a libjvm.so library (Linking vm...) containing the llvm
shark JIT, this links ok.

Then icedtea6 links a launcher binary against this .so library and here the
linking fail, the launcher binary dont contain any llvm calls. 

The attached patch fixes this issue by making llvm generate .a files for
LLVMExecutionEngine JIT and Interpreter.


{                                                    \
            echo Linking vm...;                                         \
                                                 \
            gcc -m32 -Xlinker -O1 -shared                                 \
                       -Xlinker --version-script=mapfile_reorder -Xlinker
-soname=libjvm.so -L/usr/local/lib  -lpthread -lffi -ldl -lm  -o libjvm.so
abstractCompiler.o accessFlags.o adaptiveSizePolicy.o adjoiningGenerations.o
adjoiningVirtualSpaces.o ageTable.o allocation.o allocationStats.o aprofiler.o
arguments.o array.o arrayKlass.o arrayKlassKlass.o arrayOop.o asPSOldGen.o
asPSYoungGen.o asParNewGeneration.o assembler.o assembler_linux_zero.o
assembler_zero.o atomic.o attachListener.o attachListener_linux.o barrierSet.o
bcEscapeAnalyzer.o biasedLocking.o binaryTreeDictionary.o bitMap.o
blockOffsetTable.o bytecode.o bytecodeHistogram.o bytecodeInterpreter.o
bytecodeInterpreterWithChecks.o bytecodeInterpreter_zero.o bytecodeStream.o
bytecodeTracer.o bytecodes.o bytecodes_zero.o cSpaceCounters.o
cardTableExtension.o cardTableModRefBS.o cardTableRS.o ciArray.o ciArrayKlass.o
ciConstant.o ciConstantPoolCache.o ciEnv.o ciExceptionHandler.o ciField.o
ciFlags.o ciInstance.o ciInstanceKlass.o ciInstanceKlassKlass.o ciKlass.o
ciKlassKlass.o ciMethod.o ciMethodBlocks.o ciMethodData.o ciMethodKlass.o
ciNullObject.o ciObjArray.o ciObjArrayKlass.o ciObjArrayKlassKlass.o ciObject.o
ciObjectFactory.o ciSignature.o ciStreams.o ciSymbol.o ciSymbolKlass.o ciType.o
ciTypeArray.o ciTypeArrayKlass.o ciTypeArrayKlassKlass.o ciTypeFlow.o
ciUtilities.o classFileError.o classFileParser.o classFileStream.o
classLoader.o classLoadingService.o classify.o cmsAdaptiveSizePolicy.o
cmsCollectorPolicy.o cmsGCAdaptivePolicyCounters.o cmsLockVerifier.o
cmsPermGen.o coTracker.o codeBlob.o codeBuffer.o codeCache.o collectedHeap.o
collectionSetChooser.o collectorCounters.o collectorPolicy.o
compactibleFreeListSpace.o compactingPermGenGen.o compilationPolicy.o
compileBroker.o compileLog.o compiledIC.o compiledICHolderKlass.o
compiledICHolderOop.o compilerOracle.o compressedStream.o concurrentG1Refine.o
concurrentG1RefineThread.o concurrentGCThread.o concurrentMark.o
concurrentMarkSweepGeneration.o concurrentMarkSweepThread.o
concurrentMarkThread.o concurrentZFThread.o constMethodKlass.o constMethodOop.o
constantPoolKlass.o constantPoolOop.o constantTag.o copy.o cpCacheKlass.o
cpCacheOop.o cppInterpreter.o cppInterpreter_zero.o debug.o debugInfo.o
debugInfoRec.o debug_zero.o defNewGeneration.o deoptimization.o
depChecker_zero.o dependencies.o dictionary.o dirtyCardQueue.o disassembler.o
disassembler_zero.o dtraceAttacher.o dtraceJSDT.o dtraceJSDT_linux.o dump.o
dump_zero.o events.o evmCompat.o exceptionHandlerTable.o exceptions.o
fieldDescriptor.o fieldType.o filemap.o forte.o fprofiler.o frame.o
frame_zero.o freeBlockDictionary.o freeChunk.o freeList.o g1BlockOffsetTable.o
g1CollectedHeap.o g1CollectorPolicy.o g1MMUTracker.o g1MarkSweep.o g1RemSet.o
g1SATBCardTableModRefBS.o g1_globals.o gSpaceCounters.o
gcAdaptivePolicyCounters.o gcCause.o gcLocker.o gcOverheadReporter.o
gcPolicyCounters.o gcStats.o gcTaskManager.o gcTaskThread.o gcUtil.o
genCollectedHeap.o genMarkSweep.o genRemSet.o generateOopMap.o generation.o
generationCounters.o generationSpec.o globalDefinitions.o globals.o
growableArray.o handles.o hashtable.o heap.o heapDumper.o heapInspection.o
heapRegion.o heapRegionRemSet.o heapRegionSeq.o histogram.o hpi.o hpi_linux.o
icBuffer.o icBuffer_zero.o icache.o icache_zero.o immutableSpace.o init.o
instanceKlass.o instanceKlassKlass.o instanceOop.o instanceRefKlass.o
intHisto.o interfaceSupport.o interp_masm_zero.o interpreter.o
interpreterRT_zero.o interpreterRuntime.o interpreter_zero.o
invocationCounter.o iterator.o java.o javaAssertions.o javaCalls.o
javaClasses.o jni.o jniCheck.o jniFastGetField.o jniFastGetField_zero.o
jniHandles.o jniPeriodicChecker.o jvm.o jvm_linux.o
jvmtiClassFileReconstituter.o jvmtiCodeBlobEvents.o jvmtiEnter.o
jvmtiEnterTrace.o jvmtiEnv.o jvmtiEnvBase.o jvmtiEnvThreadState.o
jvmtiEventController.o jvmtiExport.o jvmtiExtensions.o jvmtiGetLoadedClasses.o
jvmtiImpl.o jvmtiManageCapabilities.o jvmtiRedefineClasses.o jvmtiTagMap.o
jvmtiThreadState.o jvmtiTrace.o jvmtiUtil.o klass.o klassKlass.o klassOop.o
klassVtable.o linkResolver.o loaderConstraints.o location.o lowMemoryDetector.o
management.o markOop.o markSweep.o memRegion.o memoryManager.o memoryPool.o
memoryService.o memprofiler.o methodComparator.o methodDataKlass.o
methodDataOop.o methodKlass.o methodLiveness.o methodOop.o monitorChunk.o
mutableNUMASpace.o mutableSpace.o mutex.o mutexLocker.o mutex_linux.o
nativeInst_zero.o nativeLookup.o nmethod.o numberSeq.o objArrayKlass.o
objArrayKlassKlass.o objArrayOop.o objectMonitor_linux.o objectStartArray.o
oop.o oopFactory.o oopMap.o oopMapCache.o oopRecorder.o oopsHierarchy.o
orderAccess.o os.o osThread.o osThread_linux.o os_linux.o os_linux_zero.o
ostream.o parCardTableModRefBS.o parGCAllocBuffer.o parMarkBitMap.o
parNewGeneration.o parallelScavengeHeap.o pcDesc.o pcTasks.o perf.o perfData.o
perfMemory.o perfMemory_linux.o permGen.o placeholders.o preserveException.o
privilegedStack.o psAdaptiveSizePolicy.o psCompactionManager.o
psGCAdaptivePolicyCounters.o psGenerationCounters.o psMarkSweep.o
psMarkSweepDecorator.o psMemoryPool.o psOldGen.o psParallelCompact.o
psPermGen.o psPromotionLAB.o psPromotionManager.o psScavenge.o psTasks.o
psVirtualspace.o psYoungGen.o ptrQueue.o referencePolicy.o referenceProcessor.o
reflection.o reflectionUtils.o register.o register_definitions_zero.o
register_zero.o relocInfo.o relocInfo_zero.o relocator.o resolutionErrors.o
resourceArea.o restore.o rewriter.o rframe.o runtimeService.o safepoint.o
satbQueue.o scopeDesc.o serialize.o sharedHeap.o sharedRuntime.o
sharedRuntimeTrans.o sharedRuntimeTrig.o sharedRuntime_zero.o sharkBlock.o
sharkBuilder.o sharkCacheDecache.o sharkCompiler.o sharkConstant.o sharkEntry.o
sharkFunction.o sharkInliner.o sharkIntrinsics.o sharkMemoryManager.o
sharkRuntime.o sharkState.o sharkStateScanner.o sharkTopLevelBlock.o
sharkType.o shark_globals.o signature.o sizes.o space.o spaceCounters.o
spaceDecorator.o sparsePRT.o specialized_oop_closures.o stackMapFrame.o
stackMapTable.o stackValue.o stackValueCollection.o statSampler.o
stubCodeGenerator.o stubGenerator_zero.o stubRoutines.o stubRoutines_linux.o
stubRoutines_zero.o stubs.o survRateGroup.o sweeper.o symbolKlass.o symbolOop.o
symbolTable.o synchronizer.o systemDictionary.o task.o taskqueue.o
templateInterpreter.o templateInterpreter_zero.o templateTable.o
templateTable_zero.o tenuredGeneration.o thread.o threadCritical_linux.o
threadLS_linux_zero.o threadLocalAllocBuffer.o threadLocalStorage.o
threadService.o thread_linux_zero.o timer.o typeArrayKlass.o
typeArrayKlassKlass.o typeArrayOop.o unhandledOops.o universe.o unsafe.o utf8.o
verificationType.o verifier.o vframe.o vframeArray.o vframe_hp.o virtualspace.o
vmCMSOperations.o vmError.o vmError_linux.o vmGCOperations.o vmPSOperations.o
vmStructs.o vmSymbols.o vmThread.o vm_operations.o vm_operations_g1.o
vm_version.o vm_version_linux_zero.o vm_version_zero.o vmreg.o vmreg_zero.o
vtableStubs.o vtableStubs_zero.o vtune_linux.o workgroup.o xmlstream.o
yieldingWorkgroup.o sharkValue.o  -lstdc++ -lm -ldl -lpthread -lffi
-lLLVMX86AsmPrinter -lLLVMX86CodeGen -lLLVMSelectionDAG -lLLVMAsmPrinter
-lLLVMJIT -lLLVMExecutionEngine -lLLVMCodeGen -lLLVMScalarOpts
-lLLVMTransformUtils -lLLVMipa -lLLVMAnalysis -lLLVMTarget -lLLVMCore
-lLLVMSupport -lLLVMSystem;       \
                                                \
            rm -f libjvm.so.1; ln -s libjvm.so libjvm.so.1;                    
             \
            if [ -x /usr/sbin/selinuxenabled ] ; then                   \
              /usr/sbin/selinuxenabled;                                 \
              if [ $? = 0 ] ; then                                      \
                /usr/bin/chcon -t textrel_shlib_t libjvm.so;                  
\
                if [ $? != 0 ]; then                                   \
                  echo "ERROR: Cannot chcon libjvm.so";                 \
                fi                                                      \
              fi                                                        \
            fi                                                          \
        }
Linking vm...
{ \
            echo Linking launcher...; \
             \
            gcc -m32 -Xlinker -O1 -m32 -export-dynamic  -L `pwd` -o gamma
launcher.o -ljvm -lm -ldl -lpthread; \
             \
        }
Linking launcher...
/home/xerxes/icedtea6/openjdk-ecj/build/linux-i586/hotspot/outputdir/linux_zero_shark/product/libjvm.so:
undefined reference to `llvm::sys::Mutex::Mutex(bool)'
/home/xerxes/icedtea6/openjdk-ecj/build/linux-i586/hotspot/outputdir/linux_zero_shark/product/libjvm.so:
undefined reference to `llvm::sys::Mutex::~Mutex()'
/home/xerxes/icedtea6/openjdk-ecj/build/linux-i586/hotspot/outputdir/linux_zero_shark/product/libjvm.so:
undefined reference to `llvm::sys::Mutex::release()'
/home/xerxes/icedtea6/openjdk-ecj/build/linux-i586/hotspot/outputdir/linux_zero_shark/product/libjvm.so:
undefined reference to `llvm::sys::Mutex::acquire()'
collect2: ld returned 1 exit status
make[7]: *** [gamma] Error 1


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 11:10:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 11:10:36 -0500
Subject: [LLVMbugs] [Bug 4430] New: Too weak memory operation alignment from
 llvm-gcc with -O0
Message-ID: <bug-4430-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4430

           Summary: Too weak memory operation alignment from llvm-gcc with -
                    O0
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: hkultala at cs.tut.fi
                CC: llvmbugs at cs.uiuc.edu


C-code

__attribute__((noinline))
void resInput(int *restrict input_buf, int *restrict output_buf) {
    for (int i = 0; i < 4; ++i) {
        output_buf[i] = input_buf[i];
    }
}

gets compiled into


define void @resInput(i32* noalias %input_buf, i32* noalias %output_buf)
nounwind noinline {
entry:  
        %input_buf_addr = alloca i32*           ; <i32**> [#uses=2]
        %output_buf_addr = alloca i32*          ; <i32**> [#uses=2]
        %i = alloca i32         ; <i32*> [#uses=6]
        %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
        store i32* %input_buf, i32** %input_buf_addr
        store i32* %output_buf, i32** %output_buf_addr
        store i32 0, i32* %i, align 4
        br label %bb1

bb:             ; preds = %bb1
        %0 = load i32** %input_buf_addr, align 4                ; <i32*>
[#uses=1]
        %1 = load i32* %i, align 4              ; <i32> [#uses=1]
        %2 = getelementptr i32* %0, i32 %1              ; <i32*> [#uses=1]
        %3 = load i32* %2, align 1              ; <i32> [#uses=1]
        %4 = load i32** %output_buf_addr, align 4               ; <i32*>
[#uses=1]
        %5 = load i32* %i, align 4              ; <i32> [#uses=1]
        %6 = getelementptr i32* %4, i32 %5              ; <i32*> [#uses=1]
        store i32 %3, i32* %6, align 1
        %7 = load i32* %i, align 4              ; <i32> [#uses=1]
        %8 = add i32 %7, 1              ; <i32> [#uses=1]
        store i32 %8, i32* %i, align 4
        br label %bb1

bb1:            ; preds = %bb, %entry
        %9 = load i32* %i, align 4              ; <i32> [#uses=1]
        %10 = icmp sle i32 %9, 3                ; <i1> [#uses=1]
        br i1 %10, label %bb, label %bb2

bb2:            ; preds = %bb1
        br label %return

return:         ; preds = %bb2
        ret void
}

when compiled with -O0. Some of those load i32 and store i32 operations have
align 1 instead of align 4.


With -O1 alignment is correct, the code becomes:

define void @resInput(i32* noalias %input_buf, i32* noalias %output_buf)
nounwind noinline {
bb1.thread:
        br label %bb1

bb1:            ; preds = %bb1, %bb1.thread
        %i.0.reg2mem.0 = phi i32 [ 0, %bb1.thread ], [ %indvar.next, %bb1 ]   
; <i32> [#uses=3]
        %0 = getelementptr i32* %input_buf, i32 %i.0.reg2mem.0          ;
<i32*> [#uses=1]
        %1 = load i32* %0, align 4              ; <i32> [#uses=1]
        %2 = getelementptr i32* %output_buf, i32 %i.0.reg2mem.0         ;
<i32*> [#uses=1]
        store i32 %1, i32* %2, align 4
        %indvar.next = add i32 %i.0.reg2mem.0, 1                ; <i32>
[#uses=2]
        %exitcond = icmp eq i32 %indvar.next, 4         ; <i1> [#uses=1]
        br i1 %exitcond, label %return, label %bb1

return:         ; preds = %bb1
        ret void
}

Here the load and store have align 4 as it should have.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 13:22:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 13:22:03 -0500
Subject: [LLVMbugs] [Bug 4416] Assertion while compiling llvm-gcc
In-Reply-To: <bug-4416-206@http.llvm.org/bugs/>
Message-ID: <200906221822.n5MIM34H027283@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4416


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #4 from Evan Cheng <evan.cheng at apple.com>  2009-06-22 13:22:01 ---
Dup of PR4344.

*** This bug has been marked as a duplicate of bug 4344 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 13:39:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 13:39:16 -0500
Subject: [LLVMbugs] [Bug 4344] llc asserts on frameaddress intrinsic on ARM
In-Reply-To: <bug-4344-206@http.llvm.org/bugs/>
Message-ID: <200906221839.n5MIdGNj027938@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4344


Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Evan Cheng <evan.cheng at apple.com>  2009-06-22 13:39:14 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079184.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 15:24:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 15:24:04 -0500
Subject: [LLVMbugs] [Bug 4331] Floating point ICE with llvm-gcc 2.5
In-Reply-To: <bug-4331-206@http.llvm.org/bugs/>
Message-ID: <200906222024.n5MKO4Gn031653@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4331


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Dale Johannesen <dalej at apple.com>  2009-06-22 15:24:03 ---
It seems like we can close this as something that's already been fixed then.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 20:31:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 20:31:59 -0500
Subject: [LLVMbugs] [Bug 4431] New: Can't declare a property as an alias of
	another.
Message-ID: <bug-4431-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4431

           Summary: Can't declare a property as an alias of another.
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: stevencanfield at gmail.com
                CC: llvmbugs at cs.uiuc.edu


If I declare two properties and synthesize both to the same instance variable,
clang will report an error. The error is of the form 'synthesized properties %0
and %1 both claim ivar %2' (and is called error_duplicate_ivar_use internally).
I am using this heavily in my current project to interface with legacy code.
For example,

@interface Person : NSObject {
        NSString* first_name;
}
@property (nonatomic,retain) NSString* first_name;
@property (nonatomic,retain) NSString* firstName; // New Style alias
@end

@implementation Person
@synthesize firstName=first_name, first_name;
@end

The goal here is that older code in the (large) codebase uses the first_name
accessor or in some places is dependent on database column names. At some point
these will be cleaned out but for the foreseeable future I would like to keep
first_name available. However, it is not the proper style so for new code I'd
like to be able to reference firstName. This has been allowed in GCC through
4.2 and is only an error in Clang.

I'm happy to provide a patch to change this (it's trivial) but I wanted to see
people's opinions first. Is this a dangerous 'feature' ? I couldn't find
anything in the (admittedly limited) Objective C 2.0 reference to decide either
way.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 22 21:34:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 22 Jun 2009 21:34:42 -0500
Subject: [LLVMbugs] [Bug 4431] Can't declare a property as an alias of
	another.
In-Reply-To: <bug-4431-206@http.llvm.org/bugs/>
Message-ID: <200906230234.n5N2Ygac012564@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4431


Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Fariborz Jahanian <fjahanian at apple.com>  2009-06-22 21:34:42 ---
This is part of the language spec. and cannot be changed. Two properties cannot
use the same 'ivar'
for their setter/getter synthesis. This is deemed necessary because it will be
extremely error prone to
allow the ivar changed via unrelated property operations.




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 08:48:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 08:48:36 -0500
Subject: [LLVMbugs] [Bug 4432] New: OCaml binding use-after-free
Message-ID: <bug-4432-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4432

           Summary: OCaml binding use-after-free
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Running the test/Bindings/Ocaml/vmcore.ml test under valgrind results in

==8604== Invalid read of size 8
==8604==    at 0x577EDF0: std::_Rb_tree_increment(std::_Rb_tree_node_base*) (in
/usr/lib/libstdc++.so.6.0.12)
==8604==    by 0x45438A: std::_Rb_tree_iterator<std::pair<std::string const,
llvm::Type const*> >::operator++() (stl_tree.h:184)
==8604==    by 0x44AA30: LLVMDeleteTypeName (Core.cpp:78)
==8604==    by 0x5503A0: llvm_delete_type_name (llvm_ocaml.c:138)
==8604==    by 0x5639CE: caml_interprete (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)
==8604==    by 0x56527A: caml_main (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)
==8604==    by 0x562607: main (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)
==8604==  Address 0x61114f8 is 24 bytes inside a block of size 48 free'd
==8604==    at 0x4C25A5D: operator delete(void*) (vg_replace_malloc.c:344)
==8604==    by 0x4CCF2D:
__gnu_cxx::new_allocator<std::_Rb_tree_node<std::pair<std::string const,
llvm::Type const*> > >::deallocate(std::_Rb_tree_node<std::pair<std::string
const, llvm::Type const*> >*, unsigned long) (new_allocator.h:95)
==8604==    by 0x4CCE27: std::_Rb_tree<std::string const, std::pair<std::string
const, llvm::Type const*>, std::_Select1st<std::pair<std::string const,
llvm::Type const*> >, std::less<std::string const>,
std::allocator<std::pair<std::string const, llvm::Type const*> >
>::_M_put_node(std::_Rb_tree_node<std::pair<std::string const, llvm::Type
const*> >*) (stl_tree.h:363)
==8604==    by 0x4CCACB: std::_Rb_tree<std::string const, std::pair<std::string
const, llvm::Type const*>, std::_Select1st<std::pair<std::string const,
llvm::Type const*> >, std::less<std::string const>,
std::allocator<std::pair<std::string const, llvm::Type const*> >
>::_M_destroy_node(std::_Rb_tree_node<std::pair<std::string const, llvm::Type
const*> >*) (stl_tree.h:384)
==8604==    by 0x4CC78B: std::_Rb_tree<std::string const, std::pair<std::string
const, llvm::Type const*>, std::_Select1st<std::pair<std::string const,
llvm::Type const*> >, std::less<std::string const>,
std::allocator<std::pair<std::string const, llvm::Type const*> >
>::erase(std::_Rb_tree_iterator<std::pair<std::string const, llvm::Type const*>
>) (stl_tree.h:1348)
==8604==    by 0x4CC486: std::map<std::string const, llvm::Type const*,
std::less<std::string const>, std::allocator<std::pair<std::string const,
llvm::Type const*> > >::erase(std::_Rb_tree_iterator<std::pair<std::string
const, llvm::Type const*> >) (stl_map.h:567)
==8604==    by 0x4CBDCD:
llvm::TypeSymbolTable::remove(std::_Rb_tree_iterator<std::pair<std::string
const, llvm::Type const*> >) (TypeSymbolTable.cpp:73)
==8604==    by 0x44AA24: LLVMDeleteTypeName (Core.cpp:80)
==8604==    by 0x5503A0: llvm_delete_type_name (llvm_ocaml.c:138)
==8604==    by 0x5639CE: caml_interprete (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)
==8604==    by 0x56527A: caml_main (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)
==8604==    by 0x562607: main (in
llvm-objects/test/Bindings/Ocaml/Output/vmcore.ml.tmp)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 11:06:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 11:06:53 -0500
Subject: [LLVMbugs] [Bug 4433] New: Clang always dies,
	even when building empty C files
Message-ID: <bug-4433-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4433

           Summary: Clang always dies, even when building empty C files
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Support Libraries
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


On FreeBSD, Clang dies with the following error, even when building empty C
files:

Assertion failed: (NumTimers == 0 && "TimerGroup destroyed before all contained
timers!"), function ~TimerGroup, file llvm/include/llvm/Support/Timer.h, line
159.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 11:36:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 11:36:48 -0500
Subject: [LLVMbugs] [Bug 4433] Clang always dies,
	even when building empty C files
In-Reply-To: <bug-4433-206@http.llvm.org/bugs/>
Message-ID: <200906231636.n5NGamqm024862@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4433


Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Owen Anderson <resistor at mac.com>  2009-06-23 11:36:46 ---
Reverted in r73957.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 12:22:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 12:22:51 -0500
Subject: [LLVMbugs] [Bug 4427] LLVM_NATIVE_ARCH in config.h conflicts with
 compiler/ buildsystem platform arch defines
In-Reply-To: <bug-4427-206@http.llvm.org/bugs/>
Message-ID: <200906231722.n5NHMpjY026525@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4427


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Douglas Gregor <dgregor at apple.com>  2009-06-23 12:22:39 ---
Thanks for the fix! Committed here:
 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079247.html
 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079248.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 13:05:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 13:05:10 -0500
Subject: [LLVMbugs] [Bug 4429] ExecutionEngine JIT and Interpreter generated
 as . o when using cmake and makes icedtea6 linking fail.
In-Reply-To: <bug-4429-206@http.llvm.org/bugs/>
Message-ID: <200906231805.n5NI5AQb027954@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4429


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Douglas Gregor <dgregor at apple.com>  2009-06-23 13:05:04 ---
This patch was a good start. Thanks! I've completely eliminated the relinking
logic from the CMake-based build system:

 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079257.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 14:41:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 14:41:49 -0500
Subject: [LLVMbugs] [Bug 4434] New: ICE with inline FP asm with
	llvm-gcc/llvm 2.6svn
Message-ID: <bug-4434-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4434

           Summary: ICE with inline FP asm with llvm-gcc/llvm 2.6svn
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jsg at openbsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3116)
 --> (http://llvm.org/bugs/attachment.cgi?id=3116)
source file

llvm-gcc -O2 -pipe -g -DLIBC_SCCS -DSYSLIBC_SCCS -I/usr/src/lib/libc/include
-DAPIWARN
 -DYP -I/usr/src/lib/libc/yp -D__DBINTERFACE_PRIVATE -I/usr/src/lib/libc/gdtoa
-I/usr/src/lib/libc/arch/i386/gdtoa -DINFNAN_CHECK -DMULTIPLE_THREADS
-DNO_FENV_H -DUSE_LOCALE -I/usr/src/lib/libc -DRESOLVSORT -DPOSIX_MISTAKE
-DFLOATING_POINT -DNLS   -c /usr/src/lib/libc/arch/i386/gen/ldexp.c -o ldexp.o
assertion "Reg >= X86::FP0 && Reg <= X86::FP6 && "Expected FP register!""
failed: file

"/usr/obj/ports/llvm-2.6pre20090621p0/llvm-2.6pre20090621/lib/Target/X86/X86FloatingPoint.cpp",
line 176, function "getFPReg"
/usr/src/lib/libc/arch/i386/gen/ldexp.c:53: internal compiler error: Abort trap

gcc version 4.2.1 (Based on Apple Inc. build 5646) (LLVM build)
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.6svn
  Optimized build with assertions.
  Built Jun 23 2009(10:50:32).

#define __GNUC__ 4
is automatically defined by llvm-gcc in this case, so it takes the first
codepath.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 15:21:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 15:21:20 -0500
Subject: [LLVMbugs] [Bug 4435] New: missed optimization,
 not folding away two equivalent BBs
Message-ID: <bug-4435-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4435

           Summary: missed optimization, not folding away two equivalent BBs
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu, jyasskin at google.com


The attached hand-written .ll file contains four BBs in an if-statement
diamond. The %cond_true and %cond_false blocks are both very similar; all
instructions are the same except for the parameter of the GEP which is the same
type in either case. The optimization I'd like to see is that one change being
folded away into a phi node and then the blocks getting merged into the %done
block.

This is very similar to an optimization simplify-cfg does, and is also an
example of partial redundancy elimination and instruction sinking. Where should
this optimization belong?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 16:37:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 16:37:19 -0500
Subject: [LLVMbugs] [Bug 4380] __builtin_object_size
In-Reply-To: <bug-4380-206@http.llvm.org/bugs/>
Message-ID: <200906232137.n5NLbJ52003559@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4380


Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dalej at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Dale Johannesen <dalej at apple.com>  2009-06-23 16:37:19 ---
Actually, it is lowered to -1, and then the memcpy_chk expansion decides it is
OK to lower that to a plain memcpy since nothing about the object size is
known, like this:
        call void @llvm.memcpy.i32(i8* %2, i8* %3, i32 4, i32 1)
The final 1 is alignment,  not anything to do with object-size.  I think this
is OK.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 17:20:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 17:20:54 -0500
Subject: [LLVMbugs] [Bug 4380] __builtin_object_size
In-Reply-To: <bug-4380-206@http.llvm.org/bugs/>
Message-ID: <200906232220.n5NMKsq5005118@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4380


devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #3 from devang.patel <devang.patel at gmail.com>  2009-06-23 17:20:53 ---
Are you sure it is lowered when -fopnemp is used ?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 23 23:57:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 23 Jun 2009 23:57:56 -0500
Subject: [LLVMbugs] [Bug 4436] New: scev expander produces invalid code
Message-ID: <bug-4436-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4436

           Summary: scev expander produces invalid code
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: gohman at apple.com, llvmbugs at cs.uiuc.edu


Created an attachment (id=3118)
 --> (http://llvm.org/bugs/attachment.cgi?id=3118)
testcase

The attached bugpoint-reduced testcase demonstrates some breakage:

$ opt -indvars b.bc
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

Instruction does not dominate all uses!
        %0 = load i32* undef, align 4           ; <i32> [#uses=3]
        %tmp = sub i32 0, %0            ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp = sub i32 0, %0            ; <i32> [#uses=1]
        %tmp2 = add i32 %tmp, -1                ; <i32> [#uses=2]
Instruction does not dominate all uses!
        %tmp2 = add i32 %tmp, -1                ; <i32> [#uses=2]
        %tmp3 = icmp ugt i32 -1, %tmp2          ; <i1> [#uses=1]
Instruction does not dominate all uses!
        %tmp3 = icmp ugt i32 -1, %tmp2          ; <i1> [#uses=1]
        %umax = select i1 %tmp3, i32 -1, i32 %tmp2              ; <i32>
[#uses=1]
Instruction does not dominate all uses!
        %umax = select i1 %tmp3, i32 -1, i32 %tmp2              ; <i32>
[#uses=1]
        %tmp4 = sub i32 0, %umax                ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp4 = sub i32 0, %umax                ; <i32> [#uses=1]
        %tmp5 = add i32 %tmp4, -1               ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp5 = add i32 %tmp4, -1               ; <i32> [#uses=1]
        %tmp56 = inttoptr i32 %tmp5 to i8*              ; <i8*> [#uses=1]
Broken module found, compilation aborted!
0   opt 0x08427892
1   opt 0x08427dc5
2       0xffffe400 __kernel_sigreturn + 0
Stack dump:
0.      Running pass 'Function Pass Manager' on module 'b.bc'.
1.      Running pass 'Module Verifier' on function '@string_expandtabs'
Aborted


You can see what it's producing with "opt -indvars b.bc --disable-verify |
llvm-dis".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 02:29:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 02:29:20 -0500
Subject: [LLVMbugs] [Bug 4437] New: llvm optimization pass misoptimizes
	plumhall
Message-ID: <bug-4437-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4437

           Summary: llvm optimization pass misoptimizes plumhall
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: critical
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dodohack at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3119)
 --> (http://llvm.org/bugs/attachment.cgi?id=3119)
2 unoptimized bitcode

I try to optimize the timeseq.bc in attached file with opt (svn version 74068)

The following is my trace for this bug:

[hwalin at earth plumhall]$ ll
?????? 108K
-rw-r--r-- 1 hwalin jxhk-rd 55K 06-24 15:20 timeseq.bc
-rw-r--r-- 1 hwalin jxhk-rd 39K 06-24 15:20 util.bc
[hwalin at earth plumhall]$ /temp/llvm-svn/Debug/bin/opt -O3 timeseq.bc -o
timeseq.o3.bc
[hwalin at earth plumhall]$ /temp/llvm-svn/Debug/bin/llvm-ld -disable-opt
timeseq.bc util.bc -o x
[hwalin at earth plumhall]$ /temp/llvm-svn/Debug/bin/llvm-ld -disable-opt
timeseq.o3.bc util.bc -o y
[hwalin at earth plumhall]$ /temp/llvm-svn/Debug/bin/lli x.bc
***** Reached first test *****
***** 576 successful test cases in  timeseq.c ( auto pscalar2 auto array1
timeseq )
 *****
***** 0 errors detected in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 *****
***** 0 skipped sections in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 *****
[hwalin at earth plumhall]$ /temp/llvm-svn/Debug/bin/lli y.bc
***** Reached first test *****
ERROR in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 at line 386: (127) != (-120)
ERROR in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 at line 388: (127) != (-120)
......
<lots of error message>
.....
***** 554 successful test cases in  timeseq.c ( auto pscalar2 auto array1
timeseq )
 *****
***** 22 errors detected in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 *****
***** 0 skipped sections in  timeseq.c ( auto pscalar2 auto array1 timeseq )
 *****
[hwalin at earth plumhall]$


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 03:13:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 03:13:03 -0500
Subject: [LLVMbugs] [Bug 4438] New: Using the preprocessor from within
	macros fails
Message-ID: <bug-4438-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4438

           Summary: Using the preprocessor from within macros fails
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


The following code doesn't seem to build with Clang, but it does work with GCC:

#include <stdio.h>

#define DOIT(x, y, z)   puts(x y z)
#define YES 1

int
main(int argc, char *argv[])
{
        DOIT("Hello ",
#if YES >= 1
             "World",
#endif
#if YES > 1
             "?",
#endif
             "!");

        return (0);
}

test.c:16:10: error: too few arguments provided to function-like macro
invocation
             "!");
                ^
test.c:9:2: error: use of undeclared identifier 'DOIT'
        DOIT("Hello ",
        ^
2 diagnostics generated.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 07:36:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 07:36:06 -0500
Subject: [LLVMbugs] [Bug 4439] New: Cannot yet select error (llvm from
	subversion)
Message-ID: <bug-4439-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4439

           Summary: Cannot yet select error (llvm from subversion)
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: veldema at cs.fau.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3120)
 --> (http://llvm.org/bugs/attachment.cgi?id=3120)
bitcode file

I have now idea what insn it can't 'match' here.
What I did (effectively) was:
      MemoryBuffer *Buffer = MemoryBuffer::getFile(out, &errMsg);
      Module *module = ParseBitcodeFile(Buffer, &errMsg);
      TheExecutionEngine = ExecutionEngine::createJIT(OurModuleProvider,
&errMsg);
      Function *LF = module->getFunction(name);
      void *FPtr = TheExecutionEngine->getPointerToFunction(LF);

using the attached file. 
Running llc over it works however. 
The JIT gives me:

Cannot yet select: 0x7fffd8439fc0: i64 = X86ISD::Wrapper 0x7fffd84336f0

Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7fffeab1d910 (LWP 9023)]
0x00000039a56332f5 in *__GI_raise (sig=<value optimized out>) at
../nptl/sysdeps/unix/sysv/linux/raise.c:64
64        return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
Current language:  auto; currently minimal
(gdb) where
#0  0x00000039a56332f5 in *__GI_raise (sig=<value optimized out>) at
../nptl/sysdeps/unix/sysv/linux/raise.c:64
#1  0x00000039a5634b20 in *__GI_abort () at abort.c:88
#2  0x0000000000f0f119 in (anonymous
namespace)::X86DAGToDAGISel::CannotYetSelect (this=0x7fffd814ee20, N=
      {Node = 0x7fffd8439fc0, ResNo = 0}) at X86GenDAGISel.inc:55853
#3  0x0000000000f0b707 in (anonymous
namespace)::X86DAGToDAGISel::Select_X86ISD_Wrapper_i64 (
    this=0x7fffd814ee20, N=@0x7fffeab137f8) at X86GenDAGISel.inc:54197
#4  0x0000000000f0eed5 in (anonymous namespace)::X86DAGToDAGISel::SelectCode
(this=0x7fffd814ee20, N=
      {Node = 0x7fffd8439fc0, ResNo = 0}) at X86GenDAGISel.inc:55831
#5  0x0000000000f159c9 in (anonymous namespace)::X86DAGToDAGISel::Select
(this=0x7fffd814ee20, N=
      {Node = 0x7fffd8439fc0, ResNo = 0}) at X86ISelDAGToDAG.cpp:1711
#6  0x0000000000daa548 in (anonymous
namespace)::X86DAGToDAGISel::SelectRoot(struct llvm::SelectionDAG &) (
    this=0x7fffd814ee20, DAG=@0x7fffd814fc50)
    at /home/veldema/Projects/llvm/include/llvm/CodeGen/DAGISelHeader.h:112
#7  0x0000000000f11010 in (anonymous
namespace)::X86DAGToDAGISel::InstructionSelect (this=0x7fffd814ee20)
    at X86ISelDAGToDAG.cpp:634
#8  0x00000000011ba3b6 in llvm::SelectionDAGISel::CodeGenAndEmitDAG
(this=0x7fffd814ee20)
    at SelectionDAGISel.cpp:682
#9  0x00000000011b908f in llvm::SelectionDAGISel::SelectBasicBlock
(this=0x7fffd814ee20, LLVMBB=0x7fffd87fb670, 
    Begin=
    {<bidirectional_iterator<llvm::Instruction, long>> =
{<std::iterator<std::bidirectional_iterator_tag, llvm::Instruction, long,
llvm::Instruction*, llvm::Instruction&>> = {<No data fields>}, <No data
fields>}, NodePtr = 0x7fffd8808af8}, End=
      {<bidirectional_iterator<llvm::Instruction, long>> =
{<std::iterator<std::bidirectional_iterator_tag, llvm::Instruction, long,
llvm::Instruction*, llvm::Instruction&>> = {<No data fields>}, <No data
fields>}, NodePtr = 0x7fffd87fb670}) at SelectionDAGISel.cpp:496
#10 0x00000000011bb1aa in llvm::SelectionDAGISel::SelectAllBasicBlocks
(this=0x7fffd814ee20, Fn=
    @0x7fffd87127a0, MF=@0x7fffd15b8080, MMI=0x7fffd8381850, DW=0x7fffd8154400,
TII=@0x7fffd80fa2b8)
    at SelectionDAGISel.cpp:886
#11 0x00000000011b82bf in llvm::SelectionDAGISel::runOnFunction
(this=0x7fffd814ee20, Fn=@0x7fffd87127a0)
    at SelectionDAGISel.cpp:326
#12 0x0000000001307c08 in llvm::FPPassManager::runOnFunction
(this=0x7fffd80feef0, F=@0x7fffd87127a0)
    at PassManager.cpp:1351
#13 0x0000000001307939 in llvm::FunctionPassManagerImpl::run
(this=0x7fffd8004180, F=@0x7fffd87127a0)
    at PassManager.cpp:1304
#14 0x0000000001307695 in llvm::FunctionPassManager::run (this=0x7fffd8004140,
F=@0x7fffd87127a0)
    at PassManager.cpp:1246
#15 0x0000000000cddcae in llvm::JIT::runJITOnFunctionUnlocked
(this=0x7fffd8004040, F=0x7fffd87127a0, 
    locked=@0x7fffeab14980) at JIT.cpp:532
#16 0x0000000000cde113 in llvm::JIT::getPointerToFunction (this=0x7fffd8004040,
F=0x7fffd87127a0) at JIT.cpp:605
#17 0x0000000000ce2ee2 in (anonymous namespace)::JITResolver::JITCompilerFn
(Stub=0x7fffe919ce5c)
    at JITEmitter.cpp:391
#18 0x0000000000f7ce7a in X86CompilationCallback2 (StackPtr=0x7fffeab14bf8,
RetAddr=140737104170588)
    at X86JITInfo.cpp:365
#19 0x0000000000f7cd7a in X86CompilationCallback () at X86JITInfo.cpp:39
#20 0x00007fffe919ce5d in ?? ()
#21 0x00007fffe919ffad in ?? ()
---Type <return> to continue, or q <return> to quit---
#22 0x0000000002443420 in ?? ()
#23 0x00007fffe919ff10 in ?? ()
#24 0x0000000002443420 in ?? ()
#25 0x000000000484a4a0 in ?? ()
#26 0x00000000048c9948 in ?? ()
#27 0x00007fffe4356730 in ?? ()
#28 0x00007fffeab14c90 in ?? ()
#29 0x0000000000936143 in migrate_to_jit(struct LVM_LabelInstruction *,
LVM_State *, LVM_Thread *) (label=
    0x7fffe427e2d0, state=0x2443420, thread=0x484a4a0)
    at /home/veldema/Projects/jackal-main/Manta/manta/LVM/interp.cc:461
#30 0x00000000009363f9 in LVM_Thread::interp_loop (this=0x484a4a0,
migrated=true, from_jit=true)
    at
/home/veldema/vm_build/jackal-x86_64-1.0/compiler/lasm/LVM/vm-interp.cc:3
#31 0x00007fffe91a289e in ?? ()
#32 0x00007fffe91a27b0 in ?? ()
#33 0x00007fffeb51dde0 in ?? ()
#34 0x00007fffeab1d910 in ?? ()
#35 0x0000000000000000 in ?? ()


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 10:04:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 10:04:31 -0500
Subject: [LLVMbugs] [Bug 4436] scev expander produces invalid code
In-Reply-To: <bug-4436-206@http.llvm.org/bugs/>
Message-ID: <200906241504.n5OF4V5H019854@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4436


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-06-24 10:04:28 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079348.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 14:56:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 14:56:06 -0500
Subject: [LLVMbugs] [Bug 4440] New: llvm-config --libs should not include
	libLLVMgold.so
Message-ID: <bug-4440-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4440

           Summary: llvm-config --libs should not include libLLVMgold.so
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Currently llvm-config --libs prints out all .a libraries and also
libLLVMgold.so.  The problem is that libLLVMgold.so links with
libLTO which itself links with vast numbers of LLVM libraries.
The result is that you can get symbols defined twice.  I noticed
this because the domtree constructor was registering domtree
twice with the pass manager (causing an assertion failure).
In fact there were two domtree constructors, one from the .a and
one from libLTO.

Presumably llvm-config should get a special option for printing
out libLLVMgold.so.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 15:24:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 15:24:11 -0500
Subject: [LLVMbugs] [Bug 4442] New: Format string modifier functions with
 format_arg attribute
Message-ID: <bug-4442-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4442

           Summary: Format string modifier functions with format_arg
                    attribute
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: tss at iki.fi
                CC: llvmbugs at cs.uiuc.edu


gcc has this nice feature where it (apparently) assumes that when a function is
of type

const char *function(const char *) __attribute__((format_arg(1)))

it assumes that the return value is also a format string. So for example gcc
gives a proper warning about the following code:

const char * __attribute__((format_arg(1)))
format(const char *format) { return format; }

int main(int argc, char **argv)
{
        printf(format("%s"), 1234); // warning
        printf(format("%s"), "str"); // no warning even with
-Wformat-nonliteral
        return 0;
}

Would be nice if clang did the same thing, or implemented some other way to
mark the return value as a format string.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 20:06:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 20:06:43 -0500
Subject: [LLVMbugs] [Bug 4437] llvm optimization pass misoptimizes plumhall
In-Reply-To: <bug-4437-206@http.llvm.org/bugs/>
Message-ID: <200906250106.n5P16hUR010503@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4437


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-06-24 20:06:43 ---
(In reply to comment #2)
>  Id[0] = 17;                                                
>  **ppCsc = 8;                                               
> 
>  **ppCsc *= Id[0];                                          

This code invokes undefined behavior, as it is attempting to convert
the double value 136.0 to signed char, in which type the value cannot
be represented.

(In reply to comment #3)

> BTW, it is duo to the optimization pass where may misconvert the value, because
> without any optimization, lli gave the correct answer as I mentioned in #1.

Undefined behavior is permitted to vary between optimization levels,
and it's even permitted to do precisely what is expected.

>From a quality-of-implementation perspective one may have a
preference for fptosi's undefined behavior working in a particular way, but
a conformance test such as Plum Hall shouldn't depend on undefined
behavior.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Wed Jun 24 22:00:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 24 Jun 2009 22:00:46 -0500
Subject: [LLVMbugs] [Bug 4443] New: macro expansion Expr has wrong source
	range
Message-ID: <bug-4443-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4443

           Summary: macro expansion Expr has wrong source range
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Consider the following code:

#define FOO(x) (x)

int main(void) {
  int y = FOO(10);
  return y;
}

If you look at the Expr corresponding to FOO(10) that initializes y, a
dumpAll() will look like:

(ParenExpr 0x1a15580 'int'
  (IntegerLiteral 0x1a15560 'int' 10))

However, if you look at the source range between getLocStart() and getLocEnd(),
it will correspond only to "FOO" in the initializer, not to "FOO(10)".
Considering that Expr encompasses the ParenExpr and IntegerLiteral, the range
should include the complete "FOO(10)" expression.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 25 11:08:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 25 Jun 2009 11:08:04 -0500
Subject: [LLVMbugs] [Bug 4446] New: Constructing recursive type { \1* }
 fails: unnecessarily resolves to abstract type
Message-ID: <bug-4446-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4446

           Summary: Constructing recursive type { \1* } fails: unnecessarily
                    resolves to abstract type
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tinmanmoves at gmail.com
                CC: llvmbugs at cs.uiuc.edu


In pseudo-code:

%p = type opaque*
%q = type %q
%r = type %q*
resolve(%q, %r)
... leads to type %t = type \1*

Using the same method for { \1* } :

%p = type opaque*
%q = type { %p }
%r = type { %p* }
resolve(%q, %r)
... leads to type { opaque** }, when I expect something equivalent to { \1* }

Below is the C code to reproduce this problem:

#include <llvm-c/Core.h>

int main (int argc, const char * argv[]) {
  LLVMTypeHandleRef handle;
  LLVMTypeRef concrete;

  // type \1*
  LLVMTypeRef back1star = LLVMPointerType(LLVMOpaqueType(), 0);
  concrete = LLVMPointerType(back1star, 0);
  handle = LLVMCreateTypeHandle(back1star);
  LLVMRefineType(back1star, concrete);
  back1star = LLVMResolveTypeHandle(handle);
  LLVMDisposeTypeHandle(handle);

  // type { \1* }
  LLVMTypeRef inner_back1star = LLVMPointerType(LLVMOpaqueType(), 0);
  LLVMTypeRef struct_back1star = LLVMStructType(&inner_back1star, 1, 0);
  LLVMTypeRef inner_concrete = LLVMPointerType(inner_back1star, 0);
  concrete = LLVMStructType(&inner_concrete, 1, 0);
  handle = LLVMCreateTypeHandle(struct_back1star);
  LLVMRefineType(struct_back1star, concrete);
  struct_back1star = LLVMResolveTypeHandle(handle);
  LLVMDisposeTypeHandle(handle);

  /* output:

   ; ModuleID = 'type-info'
     type opaque                ; type %0
     %back1star = type %back1star*
     %struct_back1star = type { %0** }

   expected:

   ; ModuleID = 'type-info'
     %back1star = type %back1star*
     %struct_back1star = type { \1* }

   or:

   ; ModuleID = 'type-info'
     type \1*           ; type %0
     %back1star = type %back1star*
     %struct_back1star = type { %0 }

   */
  LLVMModuleRef m = LLVMModuleCreateWithName("type-info");
  LLVMAddTypeName(m, "back1star", back1star);
  LLVMAddTypeName(m, "struct_back1star", struct_back1star);
  LLVMDumpModule(m);
  LLVMDisposeModule(m);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 25 11:25:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 25 Jun 2009 11:25:31 -0500
Subject: [LLVMbugs] [Bug 4443] macro expansion Expr has wrong source range
In-Reply-To: <bug-4443-206@http.llvm.org/bugs/>
Message-ID: <200906251625.n5PGPVfe019813@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4443


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-06-25 11:25:26 ---
Clang tracks this information, but doesn't print it in the dump output.  The )
token that gets expanded through the macro will have a sourcelocation that is a
macroloc.  On that location you can us SM.getInstantiationRange(Loc)  which
returns a pair.  The first element of the pair should be the FOO token, the
second element of the range should be the ) of the foo instantiation.  You can
see that the diagnostic printer uses this in cases like this:

extern struct x y;
#define FOO(X) X
void bar() {
  1 + FOO(y);
}

t.c:8:5: error: invalid operands to binary expression ('int' and 'struct x')
  1 + FOO(y);
  ~ ^ ~~~~~~

In the ast we just have the 1, +, and y tokens.  The y token has an
instantiation range that covers the whole foo invocation (4 tokens).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 25 12:10:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 25 Jun 2009 12:10:33 -0500
Subject: [LLVMbugs] [Bug 4447] New: llvm-gcc asserts on gcc extension
Message-ID: <bug-4447-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4447

           Summary: llvm-gcc asserts on gcc extension
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


extern inline int foo(int x) { return x&1; }
int bar (int x, int y) { return foo(x+y); }
int foo (int x) { return x&1; }

llvm-gcc -O0 z.c
Assertion failed: (FnEntry->isDeclaration() && "Multiple fns with same name and
neither are external!"), function StartFunctionBody

gcc "extern inline" == C99 "inline", of course.  My reading of C99 is that we
don't have to accept this, but it is not explicitly stated, and I seem to
recall a flame war on the gcc lists when the matter arose; it is possible the
gcc maintainers came to a different conclusion.   In any event, gcc does accept
this, and llvm-gcc accepts it at any opt level above -O0; clang rejects it. 
Regardless of legality, asserting is bad.
(This showed up in 4 test failures in the gcc testsuite, not in real code.)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 25 15:07:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 25 Jun 2009 15:07:38 -0500
Subject: [LLVMbugs] [Bug 4449] New: CppBackend writes incorrect C strings if
 input file name contains a backslash
Message-ID: <bug-4449-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4449

           Summary: CppBackend writes incorrect C strings if input file name
                    contains a backslash
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: abbeyj at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3121)
 --> (http://llvm.org/bugs/attachment.cgi?id=3121)
Escape mName when printing

The CppBackend will produce incorrect C strings for the module identifier if
the input file name contains a backslash.  This is more easily shown on Windows
where backslash is the path separator.  Patch attached.


Steps to reproduce:

C:\Temp>cat temp.ll
@x = common global i32 0

C:\Temp>llvm-as temp.ll

C:\Temp>llc -march=cpp .\temp.bc -o temp.cpp

C:\Temp>llc -march=cpp -cppgen=contents .\temp.bc -o temp-contents.cpp


Actual Results:
Note that the generated strings contain a tab character (\t).

C:\Temp>grep "temp.bc" temp.cpp
  Module* mod = new Module(".\temp.bc");

C:\Temp>grep "temp.bc" temp-contents.cpp
mod->setModuleIdentifier(".\temp.bc");


Expected Results:
The backslash should be escaped.

C:\Temp>grep "temp.bc" temp.cpp
  Module* mod = new Module(".\x5Ctemp.bc");

C:\Temp>grep "temp.bc" temp-contents.cpp
mod->setModuleIdentifier(".\x5Ctemp.bc");


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Thu Jun 25 23:33:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 25 Jun 2009 23:33:59 -0500
Subject: [LLVMbugs] [Bug 4449] CppBackend writes incorrect C strings if
 input file name contains a backslash
In-Reply-To: <bug-4449-206@http.llvm.org/bugs/>
Message-ID: <200906260433.n5Q4XxfQ012530@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4449


Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Nick Lewycky <nicholas at mxc.ca>  2009-06-25 23:33:59 ---
Patch applied in r74265! Thanks!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 02:21:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 02:21:15 -0500
Subject: [LLVMbugs] [Bug 4453] New: TableGen asserting on Win32
Message-ID: <bug-4453-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4453

           Summary: TableGen asserting on Win32
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: TableGen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: public at alisdairm.net
                CC: llvmbugs at cs.uiuc.edu


Trying to build Clang with Visual C++ 2008 Express edition, running on Windows
Vista.

TableGen currently asserts, making all progress impossible.

I am triggering the assert in line 46 of ...\lib\system\win32\threadlocal.inc. 
This appears to be due to a change within the preceding 24-48 hours of this bug
report.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 04:07:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 04:07:29 -0500
Subject: [LLVMbugs] [Bug 4454] New: clang reject empty character constants
 in the preprocessor
Message-ID: <bug-4454-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4454

           Summary: clang reject empty character constants in the
                    preprocessor
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


lev /tmp$ cat clang-bug1.c
#define _TAG MKTAG('', '')

lev /tmp$ clang clang-bug1.c
clang-bug1.c:1:20: error: empty character constant
#define _TAG MKTAG('', '')
                   ^
clang-bug1.c:1:24: error: empty character constant
#define _TAG MKTAG('', '')
                       ^
2 diagnostics generated.

lev /tmp$ gcc clang-bug1.c
/usr/lib/crt1.o(.text+0x8a): In function `_start':
: undefined reference to `main'

as you can see clang does not compile this trivial thing but gcc does. this
prevents mplayer from compiling


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 05:21:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 05:21:20 -0500
Subject: [LLVMbugs] [Bug 4456] New: static float __attribute__((aligned(16)))
Message-ID: <bug-4456-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4456

           Summary: static float __attribute__((aligned(16)))
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


lev /tmp$ cat clang-bug2.c                                                     
                                   static float __attribute__((aligned(16)))
sseSinCos1c[256];

lev /tmp$ clang -c clang-bug2.c && nm clang-bug2.o

lev /tmp$ gcc -c clang-bug2.c && nm clang-bug2.o
0000000000000000 b sseSinCos1c

lev /tmp$ 

as you can see the static float __attribute((aligned(16))) variable is compiled
out of the object file. gcc puts it in the object file just fine.

this is mplayer thing


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 06:51:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 06:51:56 -0500
Subject: [LLVMbugs] [Bug 4456] static float __attribute__((aligned(16)))
In-Reply-To: <bug-4456-206@http.llvm.org/bugs/>
Message-ID: <200906261151.n5QBpusQ005554@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4456


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #3 from Anton Korobeynikov <asl at math.spbu.ru>  2009-06-26 06:51:51 ---
(In reply to comment #2)
> static float __attribute__((aligned(16))) sseSinCos1c[256];
That's bug in the source code. The array should be marked as "used".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 10:32:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 10:32:56 -0500
Subject: [LLVMbugs] [Bug 4457] New: Array size error when building with LLVM
Message-ID: <bug-4457-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4457

           Summary: Array size error when building with LLVM
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: goatboy160 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


I receive the following error when building my code under LLVM.

 error: size of array 'class_id_list' is not an integral constant-expression

Here is the line of code it is pointing at

u_int8_t class_id_list[CMSG_SPACE(1024 * sizeof(u_int32_t))];

This was found in version 3.1.2 of XCode and version 3.2 shipping with the
Developer Preview of Snow Leopard.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 11:21:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 11:21:09 -0500
Subject: [LLVMbugs] [Bug 4459] New: Stackifier asserts
Message-ID: <bug-4459-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4459

           Summary: Stackifier asserts
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Running llc -march=x86 on 

---------------------------
declare double @ceil(double)

declare void @test(i32)

define void @test2() {
entry:
        %0 = call double @ceil(double undef)
        %1 = fptosi double %0 to i32
        call void asm sideeffect "fistpl $0",
"=*m,{st},~{dirflag},~{fpsr},~{flags},~{st}"(i32* undef, double undef)
        call void @test(i32 %1 )
        unreachable
}
----------------------------

causes

/usr/local/espindola/llvm/llvm/lib/Target/X86/X86FloatingPoint.cpp:129:
void<unnamed>::FPS::moveToTop(unsigned int,
llvm::ilist_iterator<llvm::MachineInstr>): Assertion `RegMap[RegOnTop] <
StackTop' failed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 12:51:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 12:51:02 -0500
Subject: [LLVMbugs] [Bug 4461] New: clang: duplicate asm symbol error when
	using -O1 -g
Message-ID: <bug-4461-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4461

           Summary: clang: duplicate asm symbol error when using -O1 -g
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3123)
 --> (http://llvm.org/bugs/attachment.cgi?id=3123)
reduced testcase

Compiling the attached preprocessed file fails:
$ clang -g testcase-min.i -c -O1
/tmp/cc-i6uVJr.s: Assembler messages:
/tmp/cc-i6uVJr.s:223: Error: symbol `.Linfo_begin1' is already defined
/tmp/cc-i6uVJr.s:253: Error: symbol `.Linfo_end1' is already defined
/tmp/cc-i6uVJr.s:364: Error: symbol `.Lpubnames_begin1' is already defined
/tmp/cc-i6uVJr.s:371: Error: symbol `.Lpubnames_end1' is already defined

It compiles fine at -O0 -g, and also at -O1 and -O2 without -g.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 13:34:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 13:34:33 -0500
Subject: [LLVMbugs] [Bug 4228] Duplicate debug symbol in recursively
	included file
In-Reply-To: <bug-4228-206@http.llvm.org/bugs/>
Message-ID: <200906261834.n5QIYXSP019976@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4228


devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




--- Comment #8 from devang.patel <devang.patel at gmail.com>  2009-06-26 13:34:32 ---
I reverted the patch, it was causing other failures. 74304.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 15:14:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 15:14:19 -0500
Subject: [LLVMbugs] [Bug 4461] clang: duplicate asm symbol error when using
	-O1 -g
In-Reply-To: <bug-4461-206@http.llvm.org/bugs/>
Message-ID: <200906262014.n5QKEJV8023181@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4461


devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|dpatel at apple.com            |devang.patel at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from devang.patel <devang.patel at gmail.com>  2009-06-26 15:14:18 ---
Reverting PR 4228 patch helps. Pl. try agin.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 15:29:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 15:29:26 -0500
Subject: [LLVMbugs] [Bug 4392] llvm-gcc-4.2 configure script doesn't
 recognize makeinfo 4.11
In-Reply-To: <bug-4392-206@http.llvm.org/bugs/>
Message-ID: <200906262029.n5QKTQn3023635@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4392


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bob.wilson at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-06-26 15:29:25 ---
Fixed in r74318.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 15:37:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 15:37:35 -0500
Subject: [LLVMbugs] [Bug 4463] New: ARM AAPCS-VFP ABI does not support v2f64
	arguments
Message-ID: <bug-4463-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4463

           Summary: ARM AAPCS-VFP ABI does not support v2f64 arguments
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


Passing v2f64 values (and other vector types of the same size) as arguments and
return values should work for the APCS and base AAPCS ABIs.  Handling those
types in the AAPCS-VFP ABI is not yet implemented.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 15:44:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 15:44:21 -0500
Subject: [LLVMbugs] [Bug 4464] New: PRAGMA_STRUCT_ALIGN defined but #pragma
 options align doesn 't work
Message-ID: <bug-4464-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4464

           Summary: PRAGMA_STRUCT_ALIGN defined but #pragma options align
                    doesn't work
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following code produces warnings:

#if PRAGMA_STRUCT_ALIGN
#pragma options align=mac68k
#endif

typedef struct {
        short vRefNum;
        long parID;
        unsigned char name[64];
} FileDesc; /* FileDesc==FileDescriptor */

#if PRAGMA_STRUCT_ALIGN
#pragma options align=reset
#endif

The warning is:

warning: unknown pragma ignored [-Wunknown-pragmas]
#pragma options align=mac68k

The same code does not produce any warnings with Apple's GCC 4.0 used by Xcode
by default.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 17:52:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 17:52:28 -0500
Subject: [LLVMbugs] [Bug 4465] New: offsetof() inconsistent with gcc on
 ParamBlockRec from Carbon headers
Message-ID: <bug-4465-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4465

           Summary: offsetof() inconsistent with gcc on ParamBlockRec from
                    Carbon headers
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


If you include Carbon headers, and try the following with GCC and clang:

 offsetof(struct IOParam, ioVRefNum)

The result with GCC is 22, while the result with clang is 24.

This bug makes it impossible to call, for example PBReadSync(), which takes a
ParamBlockRec parameter, because the user application fills in that structure
at different offsets (clang) than the library code (gcc).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 19:22:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 19:22:26 -0500
Subject: [LLVMbugs] [Bug 4466] New: assertion failure in X86 AsmPrinter for
	fast isel
Message-ID: <bug-4466-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4466

           Summary: assertion failure in X86 AsmPrinter for fast isel
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


This assertion failure is breaking the llvm-gcc self-hosted buildbot.  I've got
2 separate bugpoint-reduced testcases (one from my build and one from Bill). 
Compiling them with:

llc -disable-fp-elim -relocation-model=pic -O0 -regalloc=local crash1.bc

causes this assertion:

Assertion failed: (!Subtarget->isPICStyleStub() && !Subtarget->isPICStyleGOT()
&& "Should have operand flag!"), function printOperand, file
/Volumes/LocalHD/bwilson/build/llvmCore.roots/llvmCore~obj/src/lib/Target/X86/AsmPrinter/X86ATTAsmPrinter.cpp,
line 499.
0   llc               0x007fa0e5 llvm::sys::SetInterruptFunction(void (*)()) +
81
1   llc               0x007fa65b llvm::sys::RemoveFileOnSignal(llvm::sys::Path
const&, std::string*) + 595
2   libSystem.B.dylib 0x901da2bb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1877106031
4   libSystem.B.dylib 0x9024e23a raise + 26
5   libSystem.B.dylib 0x9025a679 abort + 73
6   libSystem.B.dylib 0x9024f3db __assert_rtn + 101
7   llc               0x0013507c llvm::ThumbInstrInfo::~ThumbInstrInfo() + 6894
8   llc               0x00135cf6 llvm::ThumbInstrInfo::~ThumbInstrInfo() +
10088
9   llc               0x0013604e llvm::ThumbInstrInfo::~ThumbInstrInfo() +
10944
10  llc               0x0013c649 std::map<llvm::Function const*,
llvm::X86MachineFunctionInfo, std::less<llvm::Function const*>,
std::allocator<std::pair<llvm::Function const* const,
llvm::X86MachineFunctionInfo> > >::operator[](llvm::Function const* const&) +
703
11  llc               0x00137476 llvm::ThumbInstrInfo::~ThumbInstrInfo() +
16104
12  llc               0x00138c84 llvm::ThumbInstrInfo::~ThumbInstrInfo() +
22262
13  llc               0x00139447 llvm::ThumbInstrInfo::~ThumbInstrInfo() +
24249
14  llc               0x00562e24
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 64
15  llc               0x0077512e
llvm::FPPassManager::runOnFunction(llvm::Function&) + 288
16  llc               0x00775650
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 112
17  llc               0x00775820
llvm::FunctionPassManager::run(llvm::Function&) + 156
18  llc               0x000031e1 main + 2981
19  llc               0x00001936 start + 54
Stack dump:
0.      Program arguments:
/Volumes/LocalHD/bwilson/puzzlebox/build/llvmCore.roots/llvmCore~dst/Developer/usr/local/bin/llc
-disable-fp-elim -relocation-model=pic -O0 -regalloc=local crash2.bc 
1.      Running pass 'X86 AT&T-Style Assembly Printer' on function '@main'
Abort trap

The Thumb stuff in the stack trace is very weird.  I would suspect this might
be related to some of the recent Thumb changes, except that when Bill runs it,
he gets SPARC instruction destructors in his stack trace.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 20:33:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 20:33:48 -0500
Subject: [LLVMbugs] [Bug 4466] assertion failure in X86 AsmPrinter for fast
	isel
In-Reply-To: <bug-4466-206@http.llvm.org/bugs/>
Message-ID: <200906270133.n5R1XmPa001826@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4466


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Chris Lattner <clattner at apple.com>  2009-06-26 20:33:48 ---
Fixed, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079601.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 23:06:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 23:06:12 -0500
Subject: [LLVMbugs] [Bug 4442] Format string modifier functions with
 format_arg attribute
In-Reply-To: <bug-4442-206@http.llvm.org/bugs/>
Message-ID: <200906270406.n5R46C0T006309@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4442


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anders Carlsson <andersca at mac.com>  2009-06-26 23:06:10 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090622/018569.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Fri Jun 26 23:18:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 26 Jun 2009 23:18:58 -0500
Subject: [LLVMbugs] [Bug 4446] Constructing recursive type { \1* } fails:
 unnecessarily resolves to abstract type
In-Reply-To: <bug-4446-206@http.llvm.org/bugs/>
Message-ID: <200906270418.n5R4Iw2S006673@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4446


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #5 from Chris Lattner <clattner at apple.com>  2009-06-26 23:18:57 ---
Right, you can only refine opaque, not random abstract types.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 00:14:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 00:14:48 -0500
Subject: [LLVMbugs] [Bug 4468] New: printf format attribute is not
	implemented
Message-ID: <bug-4468-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4468

           Summary: printf format attribute is not implemented
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: tss at iki.fi
                CC: llvmbugs at cs.uiuc.edu


This code should give a warning:

void foo(const char *format, ...) __attribute__((format(printf, 1, 2)));

void test(void)
{
        foo("%s", 1234);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 13:02:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 13:02:36 -0500
Subject: [LLVMbugs] [Bug 4470] New: format printf and format_arg attributes
 don't work together
Message-ID: <bug-4470-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4470

           Summary: format printf and format_arg attributes don't work
                    together
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: tss at iki.fi
                CC: llvmbugs at cs.uiuc.edu


This code shouldn't give a warning (bug 4442 fixed this partially):

const char *foo(const char *format) __attribute__((format_arg(1)));

void __attribute__((format(printf, 1, 0)))
foo2(const char *fmt, va_list va)
{
        vprintf(foo(fmt), va);
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 17:10:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 17:10:24 -0500
Subject: [LLVMbugs] [Bug 4471] New: Assertion failed: (0 && "Unexpected exit
 condition!"), function splitLoop, file LoopIndexSplit.cpp, line 980.
Message-ID: <bug-4471-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4471

           Summary: Assertion failed: (0 && "Unexpected exit condition!"),
                    function splitLoop, file LoopIndexSplit.cpp, line 980.
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Created an attachment (id=3128)
 --> (http://llvm.org/bugs/attachment.cgi?id=3128)
Pre-processed source

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 74396M)

Crashes at -O2 when building stmt.c from gcc 4.2.1

Assertion failed: (0 && "Unexpected exit condition!"), function splitLoop, file
LoopIndexSplit.cpp, line 980.
Stack dump:
0.  Program arguments: /usr/opt/llvm/bin/../libexec/clang-cc -triple
x86_64-unknown-freebsd8.0 -S -disable-free -main-file-name stmt.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -DIN_GCC -DHAVE_CONFIG_H -DPREFIX="/usr/obj/usr/src/tmp/usr"
-I/usr/obj/usr/src/tmp/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools
-I/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools
-I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc
-I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/config
-I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/include
-I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/libcpp/include
-I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/libdecnumber
-I/usr/obj/usr/src/tmp/legacy/usr/include -O2 -fdiagnostics-show-option -o
/tmp/cc-P6NGzj.s -x c
/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/stmt.c 
1.  <eof> parser at end of file
2.  Per-module optimization passes
3.  Running pass 'CallGraph Pass Manager' on module
'/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/stmt.c'.
4.  Running pass 'Loop Pass Manager' on function '@expand_case'
5.  Running pass 'Index Split Loops' on basic block '%for.body73.i'
*** Error code 250


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 18:00:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 18:00:29 -0500
Subject: [LLVMbugs] [Bug 4471] Assertion failed: (0 && "Unexpected exit
 condition!"), function splitLoop, file LoopIndexSplit.cpp, line 980.
In-Reply-To: <bug-4471-206@http.llvm.org/bugs/>
Message-ID: <200906272300.n5RN0Toe021426@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4471


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-06-27 18:00:29 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090622/079661.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 18:09:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 18:09:15 -0500
Subject: [LLVMbugs] [Bug 4472] New: GVN pass really slow on this file
Message-ID: <bug-4472-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4472

           Summary: GVN pass really slow on this file
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu


When building emacs with clang the build of src/dbusbind.c seems to be taking
forever, this was tracked down to the GVN pass on the xd_append_arg function,
further bugpoint reduction causes strange crashers in opt.

0>pwo at one ~/bugpoint-emacs> time opt -gvn bugpoint-reduced-function.bc
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

364.180u 15.899s 6:21.51 99.6%  6730+3145k 0+0io 0pf+0w

Is this expected to be taking 6 minutes ?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 18:13:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 18:13:50 -0500
Subject: [LLVMbugs] [Bug 4473] New: Assertion failed: (T && "getTerminator
 returned null!"), function SuccIterator,
 file /usr/src-local/llvm/include/llvm/Support/CFG.h  , line 98.
Message-ID: <bug-4473-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4473

           Summary: Assertion failed: (T && "getTerminator returned null!"),
                    function SuccIterator, file /usr/src-
                    local/llvm/include/llvm/Support/CFG.h, line 98.
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3130)
 --> (http://llvm.org/bugs/attachment.cgi?id=3130)
bugpoint-reduced-simplified.bc

Building emacs results in this assertion when compiling src/fringe.c

0>pwo at one ~/bugpoint-emacs> opt bugpoint-reduced-simplified.bc -loopsimplify
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

Assertion failed: (T && "getTerminator returned null!"), function SuccIterator,
file /usr/src-local/llvm/include/llvm/Support/CFG.h, line 98.
Stack dump:
0.      Running pass 'Function Pass Manager' on module
'bugpoint-reduced-simplified.bc'.
1.      Running pass 'Canonicalize natural loops' on function
'@Fdefine_fringe_bitmap'
Abort (core dumped)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sat Jun 27 18:16:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 27 Jun 2009 18:16:07 -0500
Subject: [LLVMbugs] [Bug 4473] Assertion failed: (T && "getTerminator
 returned null!"), function SuccIterator,
 file /usr/src-local/llvm/include/llvm/Support/CFG.h  , line 98.
In-Reply-To: <bug-4473-206@http.llvm.org/bugs/>
Message-ID: <200906272316.n5RNG7AQ021946@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4473


Pawel Worach <pawel.worach at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Pawel Worach <pawel.worach at gmail.com>  2009-06-27 18:16:07 ---
So this was fixed in the next svn revision (74399).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Sun Jun 28 14:56:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 28 Jun 2009 14:56:48 -0500
Subject: [LLVMbugs] [Bug 4470] format printf and format_arg attributes don't
 work together
In-Reply-To: <bug-4470-206@http.llvm.org/bugs/>
Message-ID: <200906281956.n5SJum3W026419@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4470


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Anders Carlsson <andersca at mac.com>  2009-06-28 14:56:43 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090622/018606.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 05:43:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 05:43:54 -0500
Subject: [LLVMbugs] [Bug 4474] New: PATCH: add OpenBSD to triple
Message-ID: <bug-4474-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4474

           Summary: PATCH: add OpenBSD to triple
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jsg at openbsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3133)
 --> (http://llvm.org/bugs/attachment.cgi?id=3133)
OpenBSD triple addition diff

attached diff adds OpenBSD to triple, which is required for at least clang
when adding OpenBSD specific linker etc configuration.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 06:02:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 06:02:10 -0500
Subject: [LLVMbugs] [Bug 4475] New: clang: gdb warning for -g -O0: warning:
 Unmapped DWARF Register #61 encountered.
Message-ID: <bug-4475-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4475

           Summary: clang: gdb warning for -g -O0: warning: Unmapped DWARF
                    Register #61 encountered.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


When compiling code with -g -O0, and breakpointing in gdb I get warnings
every time the breakpoint is hit / every time I step in the code.
I can still print variables though, so this may be a benign warning for the -O0
case.

However if I compile with -O1 -g, I cannot print variables (see below), so this
may be a problem for the -O1 case (I assume the variable is assigned to a
register, but the DWARF information is missing this info?)

Here is what happens with -O0:
$ clang -g -O0 x.c -o a.out
$ gdb /a.out
GNU gdb 6.8-debian
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...

(gdb) b readString
Breakpoint 1 at 0x400582: file x.c, line 9.
(gdb) r
Starting program: /home/edwin/clam/git/builds/clang/a.out
Breakpoint 1, readString (p=warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.                            
warning: Unmapped DWARF Register #50 encountered.                            
warning: Unmapped DWARF Register #51 encountered.                            
warning: Unmapped DWARF Register #52 encountered.                            
warning: Unmapped DWARF Register #53 encountered.                            
warning: Unmapped DWARF Register #54 encountered.                            
warning: Unmapped DWARF Register #55 encountered.                            
0x7ffff16b9512 "k??????\177", off=0x7ffff16b950c, len=10, ok=0x7ffff16b950b
"\001") at x.c:9
9           unsigned stringlen;
Current language:  auto; currently minimal
(gdb) n
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
10          char *str = (char*)readData(p, off, len, ok, &stringlen);
(gdb) n
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
11          if (*ok && stringlen && str[stringlen-1] != '\0') {
(gdb) p str
$1 = 0x2095010 ""


With -O1 -g:

$ gdb ./a.out
GNU gdb 6.8-debian
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...
(gdb) b readString
Breakpoint 1 at 0x400543: file x.c, line 10.
(gdb) r
Starting program: /home/edwin/clam/git/builds/clang/a.out

Breakpoint 1, readString () at x.c:10
10          char *str = (char*)readData(p, off, len, ok, &stringlen);
Current language:  auto; currently minimal
(gdb) n

warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
warning: Unmapped DWARF Register #48 encountered.
warning: Unmapped DWARF Register #49 encountered.
warning: Unmapped DWARF Register #50 encountered.
warning: Unmapped DWARF Register #51 encountered.
warning: Unmapped DWARF Register #52 encountered.
warning: Unmapped DWARF Register #53 encountered.
warning: Unmapped DWARF Register #54 encountered.
warning: Unmapped DWARF Register #55 encountered.
11          if (*ok && stringlen && str[stringlen-1] != '\0') {
(gdb) p str
No symbol "str" in current context.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 08:27:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 08:27:08 -0500
Subject: [LLVMbugs] [Bug 4476] New: cmake fails to find pthread_rwlock_init
 and pthread_getspecific using check_symbol_exists when they exists
Message-ID: <bug-4476-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4476

           Summary: cmake fails to find pthread_rwlock_init and
                    pthread_getspecific using check_symbol_exists when they
                    exists
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3134)
 --> (http://llvm.org/bugs/attachment.cgi?id=3134)
workaround fix: change implementation from check_symbol_exists to
check_library_exists .patch

cd llvm
svn up
cd ..
mkdir llvm-build
cd llvm-build
cmake ../llvm
...
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
...
-- Looking for pthread_mutex_lock
-- Looking for pthread_mutex_lock - found
-- Looking for pthread_rwlock_init
-- Looking for pthread_rwlock_init - not found.
-- Looking for pthread_getspecific
-- Looking for pthread_getspecific - not found. 

The problem are that pthread_rwlock_init and pthread_getspecific are not found
on a system that should have them.

The attached patch includes a workaround that changes the implementation of
these tests to use check_library_exists instead of check_symbol_exists.
with this change the output from cmake are:

-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Looking for pthread_getspecific in pthread
-- Looking for pthread_getspecific in pthread - found
-- Looking for pthread_rwlock_init in pthread
-- Looking for pthread_rwlock_init in pthread - found
...
-- Looking for pthread_mutex_lock
-- Looking for pthread_mutex_lock - found 

There are one open cmake check_symbol_exists bugs on the cmake mantis
bugtracker this might be related:
http://www.cmake.org/Bug/view.php?id=8758 - CheckSymbolExists does not find a
symbol although it exists


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 08:46:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 08:46:50 -0500
Subject: [LLVMbugs] [Bug 4474] PATCH: add OpenBSD to triple
In-Reply-To: <bug-4474-206@http.llvm.org/bugs/>
Message-ID: <200906291346.n5TDkokt017714@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4474


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2009-06-29 08:46:42 ---
Patch applied.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 09:14:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 09:14:24 -0500
Subject: [LLVMbugs] [Bug 4477] New: indvars wrong loop trip count
Message-ID: <bug-4477-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4477

           Summary: indvars wrong loop trip count
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3135)
 --> (http://llvm.org/bugs/attachment.cgi?id=3135)
testcase .ll

On the attached testcase indvars seems to think the trip count is more
than it is.  This can be seen by looking at the loop exit values it
calculates:

cc70a02__complex_multiplication.170.exit:
        %.lcssa1 = phi i8 [ 108, %cc70a02__complex_integers__Oadd.153.exit.i ] 
        ; <i8> [#uses=1]
        %.lcssa = phi i8 [ -48, %cc70a02__complex_integers__Oadd.153.exit.i ]  
        ; <i8> [#uses=1]

The correct values are 63 and -28 respectively (these are 9*7 and
-4*7, while indvars has calculated 9*12 and -4*12).  Reduced from Ada
ACATS test cc70a02.

Reproduce using: opt -indvars


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 11:29:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 11:29:28 -0500
Subject: [LLVMbugs] [Bug 4465] offsetof() inconsistent with gcc on
 ParamBlockRec from Carbon headers
In-Reply-To: <bug-4465-206@http.llvm.org/bugs/>
Message-ID: <200906291629.n5TGTSkx022947@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4465


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Daniel Dunbar <daniel at zuster.org>  2009-06-29 11:29:25 ---


*** This bug has been marked as a duplicate of bug 4464 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 11:39:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 11:39:12 -0500
Subject: [LLVMbugs] [Bug 4478] New: Assertion failed: (i == ti ||
 !mi->getOperand(i).isReg() || mi->getOperand(i).getReg() != regA),
 function runOnMachineFunction, file TwoAddressInstructionPass.cpp, line 812.
Message-ID: <bug-4478-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4478

           Summary: Assertion failed: (i == ti || !mi->getOperand(i).isReg()
                    || mi->getOperand(i).getReg() != regA), function
                    runOnMachineFunction, file
                    TwoAddressInstructionPass.cpp, line 812.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


when compiling the attached program I get this:

lev /tmp$ clang -O2 tftp-proxy.c                                               
                                                                       
Assertion failed: (i == ti || !mi->getOperand(i).isReg() ||
mi->getOperand(i).getReg() != regA), function runOnMachineFunction, file
TwoAddressInstructionPass.cpp, line 812.
Stack dump:
0.      Program arguments: /usr/local/bin/../libexec/clang-cc -triple
x86_64-unknown-freebsd7.2 -S -disable-free -main-file-name tftp-proxy.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O2 -fdiagnostics-show-option -o /tmp/cc-z1bJ7B.s -x c
tftp-proxy.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Two-Address instruction pass' on function '@main'

without the -O2 it works.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 12:22:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 12:22:19 -0500
Subject: [LLVMbugs] [Bug 4454] clang reject empty character constants in the
 preprocessor
In-Reply-To: <bug-4454-206@http.llvm.org/bugs/>
Message-ID: <200906291722.n5THMJvD024628@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4454


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #2 from Chris Lattner <clattner at apple.com>  2009-06-29 12:22:18 ---
GCC rejects this with an error when it's not in a #define.  It is apparently
deferring lexical checks to later than clang is.  Unless this is affecting a
large number of applications, I don't think we should worry about it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 12:26:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 12:26:47 -0500
Subject: [LLVMbugs] [Bug 4476] cmake fails to find pthread_rwlock_init and
 pthread_getspecific using check_symbol_exists when they exists
In-Reply-To: <bug-4476-206@http.llvm.org/bugs/>
Message-ID: <200906291726.n5THQlEC024787@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4476


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Douglas Gregor <dgregor at apple.com>  2009-06-29 12:26:46 ---
This was fixed in r74426 (with the submitted patch):

 
http://llvm.org/viewvc/llvm-project/llvm/trunk/cmake/config-ix.cmake?r1=74284&r2=74426&diff_format=h


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 12:31:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 12:31:50 -0500
Subject: [LLVMbugs] [Bug 4428] Assertion failed: (BitWidth == RHS.BitWidth
 && " Bit widths must be same for comparison"), function ult,
 file llvm/lib/  Support/APInt.cpp, line 492.
In-Reply-To: <bug-4428-206@http.llvm.org/bugs/>
Message-ID: <200906291731.n5THVoOn024973@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4428


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-06-29 12:31:50 ---
Yes, thanks for confirming it.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 12:33:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 12:33:38 -0500
Subject: [LLVMbugs] [Bug 2913] MultiSource/Benchmarks/MallocBench/gs/ gs
 segfaults with native compiler
In-Reply-To: <bug-2913-206@http.llvm.org/bugs/>
Message-ID: <200906291733.n5THXcVm025068@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2913


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #8 from Dan Gohman <gohman at apple.com>  2009-06-29 12:33:35 ---
With r74425, both the "native" and llvm-compiled executables execute
without segfaulting on x86-64.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 14:02:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 14:02:28 -0500
Subject: [LLVMbugs] [Bug 4479] New: register scavenger assertion failure
Message-ID: <bug-4479-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4479

           Summary: register scavenger assertion failure
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3138)
 --> (http://llvm.org/bugs/attachment.cgi?id=3138)
testcase

This attached bugpoint test causes an assertion failure in the register
scavenger when compiling with "llc -O3 -relocation-model=pic
-mcpu=arm1136jf-s":

Assertion failed: (isUsed(Reg) && "Using an undefined register!"), function
forward, file
/Users/bwilson/local/llvm/llvm/lib/CodeGen/RegisterScavenging.cpp, line 266.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 14:02:28 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 14:02:28 -0500
Subject: [LLVMbugs] [Bug 4480] New: register scavenger assertion failure
Message-ID: <bug-4480-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4480

           Summary: register scavenger assertion failure
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3139)
 --> (http://llvm.org/bugs/attachment.cgi?id=3139)
testcase

This attached bugpoint test causes an assertion failure in the register
scavenger when compiling with "llc -O3 -relocation-model=pic
-mcpu=arm1136jf-s":

Assertion failed: (isUsed(Reg) && "Using an undefined register!"), function
forward, file
/Users/bwilson/local/llvm/llvm/lib/CodeGen/RegisterScavenging.cpp, line 266.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 15:32:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 15:32:44 -0500
Subject: [LLVMbugs] [Bug 4459] Stackifier asserts
In-Reply-To: <bug-4459-206@http.llvm.org/bugs/>
Message-ID: <200906292032.n5TKWigJ030576@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4459


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 15:35:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 15:35:13 -0500
Subject: [LLVMbugs] [Bug 4477] indvars wrong loop trip count
In-Reply-To: <bug-4477-206@http.llvm.org/bugs/>
Message-ID: <200906292035.n5TKZDrC030682@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4477


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-06-29 15:35:11 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090629/079729.html


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 15:46:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 15:46:11 -0500
Subject: [LLVMbugs] [Bug 4480] register scavenger assertion failure
In-Reply-To: <bug-4480-206@http.llvm.org/bugs/>
Message-ID: <200906292046.n5TKkBG7031055@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4480


Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-06-29 15:46:09 ---
It looks like I must have hit submit twice.  This is the same as 4479.

*** This bug has been marked as a duplicate of bug 4479 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 16:37:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 16:37:26 -0500
Subject: [LLVMbugs] [Bug 4481] New: llvm-config with src!=obj omits
	-I<obj_dir>/include
Message-ID: <bug-4481-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4481

           Summary: llvm-config with src!=obj omits -I<obj_dir>/include
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-config
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu


jyasskin at enki:~/opensource/llvm/oprof/dbg$ cat test_config.cc 
#include "llvm/Module.h"

int main() {
}

jyasskin at enki:~/opensource/llvm/oprof/dbg$ ./Debug/bin/llvm-config --cxxflags
-I/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include  -D_DEBUG 
-D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -g -fPIC 
-Woverloaded-virtual

jyasskin at enki:~/opensource/llvm/oprof/dbg$ g++ -c test_config.cc
`./Debug/bin/llvm-config --cxxflags` 2>&1|head -20
In file included from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/Value.h:18,
                 from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/User.h:22,
                 from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/Constant.h:17,
                 from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/GlobalValue.h:21,
                 from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/Function.h:21,
                 from
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/Module.h:18,
                 from test_config.cc:1:
/usr/local/google/jyasskin/llvm/oprof/dbg/../src/include/llvm/Use.h:29:31:
error: llvm/ADT/iterator.h: No such file or directory
...
jyasskin at enki:~/opensource/llvm/oprof/dbg$ ls include/llvm/ADT/iterator.h 
include/llvm/ADT/iterator.h



If llvm-config --cxxflags included
"-I/usr/local/google/jyasskin/llvm/oprof/dbg/include", this would compile
successfully.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 16:47:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 16:47:20 -0500
Subject: [LLVMbugs] [Bug 4482] New: Wrong codegen for calls on darwin ppc32
Message-ID: <bug-4482-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4482

           Summary: Wrong codegen for calls on darwin ppc32
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dalej at apple.com
                CC: llvmbugs at cs.uiuc.edu


Try the following in llvm-gcc on darwin-ppc32 with -O0 -S:

extern int exact_log2 (unsigned long long);
extern __inline__ int
exact_log2 (unsigned long long x)
{
  return x == (x & -x) && x ? (int) __builtin_ctzll (x) : -1;
}
int foo(unsigned long long x) {
  return exact_log2(x);
}

The call looks like:

bl exact_log2$non_lazy_ptr

which is wrong.  If you want to generate a $stub, you can, but there is no
reason to do either on Leopard (this should be controlled by
-mmacosx-version-min).

This causes ppc32 bootstrap to fail, and is a recent regression.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Mon Jun 29 18:22:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 29 Jun 2009 18:22:29 -0500
Subject: [LLVMbugs] [Bug 4483] New: The JIT allocates global data inside of
 function bodies, which can be freed
Message-ID: <bug-4483-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4483

           Summary: The JIT allocates global data inside of function bodies,
                    which can be freed
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: reid.kleckner at gmail.com
        ReportedBy: reid.kleckner at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Consider the program:

int counter = 0;
void foo() { return ++counter; }
void bar() { return ++counter; }

If we use the JIT, and make a call to foo, free the code for foo, and then call
bar, we'll access freed memory.  Of course, you won't notice the problem until
something else is allocated in the freed region, but if you poison freed memory
it's easier to detect.

This also interferes with reattempting to JIT machine code, which I was working
on when I found this bug.  I ran into it because I free the machine code before
retrying, and if the function allocated any globals before reattempting, the
old addresses are saved in the GlobalValue to address map.

I've attached a failing test case, and I'm working on a fix.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 30 06:31:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 30 Jun 2009 06:31:52 -0500
Subject: [LLVMbugs] [Bug 4484] New: yet another stackifier bug with inline
	asm
Message-ID: <bug-4484-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4484

           Summary: yet another stackifier bug with inline asm
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: rafael.espindola at gmail.com
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


the test

---------------------------------------------
declare x86_fp80 @ceil()

declare void @test(x86_fp80)

define void @test2(x86_fp80 %a) {
entry:
        %0 = call x86_fp80 @ceil()
        call void asm sideeffect "fistpl $0", "{st},~{st}"(x86_fp80 %a)
        call void @test(x86_fp80 %0)
        ret void
}
---------------------------------------------

Fails with

llc: /usr/local/espindola/llvm/llvm/lib/Target/X86/X86FloatingPoint.cpp:1149:
void<unnamed>::FPS::handleSpecialFP(llvm::ilist_iterator<llvm::MachineInstr>&):
Assertion `StackTop == 1 && FirstFPRegOp == getStackEntry(0) && "Top of stack
not the right register for RET!"' failed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 30 07:20:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 30 Jun 2009 07:20:06 -0500
Subject: [LLVMbugs] [Bug 4484] yet another stackifier bug with inline asm
In-Reply-To: <bug-4484-206@http.llvm.org/bugs/>
Message-ID: <200906301220.n5UCK6VQ009541@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4484


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 30 10:09:34 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 30 Jun 2009 10:09:34 -0500
Subject: [LLVMbugs] [Bug 4485] New: stackifier bug with inline asm
Message-ID: <bug-4485-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4485

           Summary: stackifier bug with inline asm
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: rafael.espindola at gmail.com
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


the test

-------------------------------------------------------
define void @test(x86_fp80* %a) {
entry:
        %0 = load x86_fp80* %a, align 16
        %1 = fmul x86_fp80 %0, 0xK4006B400000000000000
        %2 = fmul x86_fp80 %1, 0xK4012F424000000000000
        tail call void asm sideeffect "fistpl $0", "{st},~{st}"(x86_fp80 %2)
        %3 = load x86_fp80* %a, align 16
        %4 = fmul x86_fp80 %3, 0xK4006B400000000000000
        %5 = fmul x86_fp80 %4, 0xK4012F424000000000000
        tail call void asm sideeffect "fistpl $0", "{st},~{st}"(x86_fp80 %5)
        ret void
}
----------------------------------------------------

Fails with

Assertion `StackTop == 1 && FirstFPRegOp == getStackEntry(0) && "Top of stack
not the right register for RET!"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 30 11:42:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 30 Jun 2009 11:42:41 -0500
Subject: [LLVMbugs] [Bug 4485] stackifier bug with inline asm
In-Reply-To: <bug-4485-206@http.llvm.org/bugs/>
Message-ID: <200906301642.n5UGgfRD018435@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4485


Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at cs.uiuc.edu  Tue Jun 30 23:14:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 30 Jun 2009 23:14:51 -0500
Subject: [LLVMbugs] [Bug 4486] New: llvm-ld: Assertion `writers == 0 && "
 Writer lock already acquired!"'
Message-ID: <bug-4486-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4486

           Summary: llvm-ld: Assertion `writers == 0 && "Writer lock already
                    acquired!"'
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: haohui.mai at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3142)
 --> (http://llvm.org/bugs/attachment.cgi?id=3142)
test case

It seems that llvm-ld has acquire a lock twice.

llvm-ld mm/page_alloc.bc mm/page-writeback.bc -o a
llvm-ld: /home/mai4/work/src/llvm/include/llvm/System/RWMutex.h:119: bool
llvm::sys::SmartRWMutex<mt_only>::writer_acquire() [with bool mt_only = true]:
Assertion `writers == 0 && "Writer lock already acquired!"' failed.
0   llvm-ld 0x083694c8
Stack dump:
0.      Program arguments: llvm-ld mm/page_alloc.bc mm/page-writeback.bc -o a 
Aborted


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.