From bugzilla-daemon at cs.uiuc.edu  Sat Mar  1 00:05:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 00:05:24 -0600
Subject: [LLVMbugs] [Bug 2114] New: LLVM returns vector in mm0 register on
 X86, llvm-gcc uses ax:dx
Message-ID: <bug-2114-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2114

           Summary: LLVM returns vector in mm0 register on X86, llvm-gcc
                    uses ax:dx
           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: andersca at mac.com
                CC: llvmbugs at cs.uiuc.edu


The following code

__m64 f()
{
    return (__m64){0L};
}

is compiled by llvm-gcc to

        pxor    %mm0, %mm0
        ret

whereas gcc compiles it to

        pushl   %ebp
        xorl    %eax, %eax
        movl    %esp, %ebp
        subl    $8, %esp
        leave
        xorl    %edx, %edx
        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  Sat Mar  1 02:18:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 02:18:12 -0600
Subject: [LLVMbugs] [Bug 2115] New: please document llvm::Use
Message-ID: <bug-2115-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2115

           Summary: please document llvm::Use
           Product: Documentation
           Version: trunk
          Platform: All
               URL: http://llvm.org/doxygen/classllvm_1_1Use.html#a8
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Doxygen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


The methods on Use are all undocumented, except for the default constructor
which states that the only thing you can do with a default-constructed Use is
call init() on 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  Sat Mar  1 02:33:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 02:33:01 -0600
Subject: [LLVMbugs] [Bug 2089] sizeof('a') in C++ is == 1 and in C == 4
In-Reply-To: <bug-2089-206@http.llvm.org/bugs/>
Message-ID: <200803010833.m218X1jW018309@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2089


Chris Lattner <sabre at nondot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
          Component|LLVM Codegen                |Semantic Analyzer
         Resolution|                            |FIXED




--- Comment #2 from Chris Lattner <sabre at nondot.org>  2008-03-01 02:33:01 ---
Looks great!  Applied:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080225/004500.html

Thanks Nuno,

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 03:02:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 03:02:26 -0600
Subject: [LLVMbugs] [Bug 2113] verifyModule accepts "alloca void ()"
In-Reply-To: <bug-2113-206@http.llvm.org/bugs/>
Message-ID: <200803010902.m2192Q91025988@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2113


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-01 03:02:25 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059026.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 Mar  1 03:23:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 03:23:59 -0600
Subject: [LLVMbugs] [Bug 2116] New: Add support for 'transactions' in the
	pass manager
Message-ID: <bug-2116-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2116

           Summary: Add support for 'transactions' in the pass manager
           Product: libraries
           Version: 2.2
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: new-feature
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sabre at nondot.org
                CC: llvmbugs at cs.uiuc.edu


Lots of passes operate by making incremental changes to a function: instcombine
iterates through a worklist, mem2reg promotes allocas one at a time, and the
inliner inlines a function one at a time.  Before and after each
transformation, the IR is valid, and could be snapshotted.  It would be really
awesome to expose this behavior through to the passmanager to allow the
verifier to be run after each transaction and for bugpoint to be able to reduce
bugs down to the level of a single transaction.

I haven't thought fully about how this would work, but it seems that a pass
should be able to opt-in to the transaction system.  Each time it does
something, it should bump the per-pass counter (which is owned/maintained by
the passmgr) and before it does a transaction it checks some counter or
something.  If the counter is reaches zero, then it stops making
transformations.

This is a fairly large and open ended project, but would be incredibly valuable
for tracking down nasty bugs, particularly miscompilations.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 09:22:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 09:22:51 -0600
Subject: [LLVMbugs] [Bug 2117] New: Codegen constraint improvements
Message-ID: <bug-2117-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2117

           Summary: Codegen constraint improvements
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Since a number of operations produce results in certain registers or take input
in certain registers, it would be very nice to be able to say:

let Constraints = "$dst = %eax, $val = %ecx" in
let Constraints = "$dst = %eax:%edx" in

etc

and have the selection dag insert the to and from reg nodes automatically. 
This seems like it would cut down on a lot of custom lowering (dag op lowered
to copyto, op, copyfrom).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 20:48:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 20:48:38 -0600
Subject: [LLVMbugs] [Bug 318] [testsuite] Should diff against SPEC reference
 outputs, not results from native run
In-Reply-To: <bug-318-206@http.llvm.org/bugs/>
Message-ID: <200803020248.m222mcYl025375@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=318


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #3 from Chris Lattner <sabre at nondot.org>  2008-03-01 20:48:38 ---
I don't think this is desirable.  This breaks a lot of the functionality in the
testsuite.  I think we should use the official spec harness for official runs. 
llvm-test is about getting an idea of performance and correctness.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 20:52:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 20:52:09 -0600
Subject: [LLVMbugs] [Bug 717] undef through phi doesn't collapse to undef
	when it ought
In-Reply-To: <bug-717-206@http.llvm.org/bugs/>
Message-ID: <200803020252.m222q9O7025556@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=717


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #4 from Chris Lattner <sabre at nondot.org>  2008-03-01 20:52:08 ---
I moved this minor optzn into the README.txt file:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059042.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 Mar  1 20:57:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 20:57:49 -0600
Subject: [LLVMbugs] [Bug 744] Incorrect fallthrough when no cpu help is
	available
In-Reply-To: <bug-744-206@http.llvm.org/bugs/>
Message-ID: <200803020257.m222vnfP025714@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=744


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #4 from Chris Lattner <sabre at nondot.org>  2008-03-01 20:57:48 ---
the mcpu and mattr facilities are a feature of the native llvm code generators
(libcodegen) and aren't used by other targets like the CBE and MSIL backends. 
This is like setting unswitch-threshold in opt when not running loop unswitch.

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 21:03:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 21:03:01 -0600
Subject: [LLVMbugs] [Bug 1172] Vector -> Integer and int->vec Bitcast Not
	Supported
In-Reply-To: <bug-1172-206@http.llvm.org/bugs/>
Message-ID: <200803020303.m22331cr025909@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1172


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #7 from Chris Lattner <sabre at nondot.org>  2008-03-01 21:03:01 ---
These work.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 23:41:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 23:41:22 -0600
Subject: [LLVMbugs] [Bug 1930] cbe doesn't handle insertelement,
 extractelement and shufflevector
In-Reply-To: <bug-1930-206@http.llvm.org/bugs/>
Message-ID: <200803020541.m225fM4w030408@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1930


Chris Lattner <sabre at nondot.org> changed:

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




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  1 23:42:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Mar 2008 23:42:05 -0600
Subject: [LLVMbugs] [Bug 1930] cbe doesn't handle insertelement,
 extractelement and shufflevector
In-Reply-To: <bug-1930-206@http.llvm.org/bugs/>
Message-ID: <200803020542.m225g5j0030445@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1930


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #2 from Chris Lattner <sabre at nondot.org>  2008-03-01 23:42:04 ---
These have been implemented, patches here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059050.html
...
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059054.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 Mar  2 02:07:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 02:07:47 -0600
Subject: [LLVMbugs] [Bug 2065] cbe missing a address of operation
In-Reply-To: <bug-2065-206@http.llvm.org/bugs/>
Message-ID: <200803020807.m2287lsB002308@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2065


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #8 from Chris Lattner <sabre at nondot.org>  2008-03-02 02:07:46 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059058.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 Mar  2 02:21:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 02:21:40 -0600
Subject: [LLVMbugs] [Bug 1126] CBE doesn't support vectors yet
In-Reply-To: <bug-1126-206@http.llvm.org/bugs/>
Message-ID: <200803020821.m228LeHr002751@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1126


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #6 from Chris Lattner <sabre at nondot.org>  2008-03-02 02:21:39 ---
This is now done, the CBE generally supports vectors.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  2 02:54:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 02:54:53 -0600
Subject: [LLVMbugs] [Bug 1481] CBE doesn't support some target intrinsics
In-Reply-To: <bug-1481-206@http.llvm.org/bugs/>
Message-ID: <200803020854.m228sr24011587@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1481


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #3 from Chris Lattner <sabre at nondot.org>  2008-03-02 02:54:52 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080225/059065.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 Mar  2 13:47:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 13:47:44 -0600
Subject: [LLVMbugs] [Bug 2118] New: Loop Optimizer Problems
Message-ID: <bug-2118-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2118

           Summary: Loop Optimizer Problems
           Product: libraries
           Version: 2.2
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-quality, slow-compile
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sabre at nondot.org
                CC: llvmbugs at cs.uiuc.edu


Consider this code:
void test(double vertices[][3], double KeKe[][12], double Me[], double
jacobian[3][3]) {
  double ds[3][4];
  double sum[3];
  double det;
  double volume;
  double E, nu;
  double c1, c2, c3;
  double tt, ts;
  int i, j, m, n, row, column;
  shape_ders(ds);
   i = 0;
    for (j = 0; j < 3; j++) {
      double sum = 0.0;
      for (m = 0; m < 4; m++)
        sum = sum + ds[i][m] * vertices[m][j];
      jacobian[j][i] = sum;         /* compute Jacobian */
    }
}

When compiled with llvm-gcc, there are loads of ds that are left in the loop. 
Peeling the inner loop means that the loads are all from LIV indices and could
be hoisted.  This is a problem, however, this points out other problems.  I've
attached a bc file from right before the loop optimizer.  The loop optimizer
consists of:

$opt loop_test.bc -loopsimplify -loop-rotate -licm -loop-unswitch
-loop-index-split -instcombine -indvars -loop-unroll 

When you run this with -debug-pass=Executions, it becomes clear that LCSSA is
being rerun before each pass, this is very bad.  Do the loop passes need to
explicitly state that they preserve LCSSA?

0xc031a8     Executing Pass 'Loop Pass Manager' on Function
'element_matrices'...
0xc04028       Executing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028       Executing Pass 'Index Split Loops' on Loop '...
0xc04028        Freeing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028        Freeing Pass 'Index Split Loops' on Loop '...
0xc04028       Executing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028       Executing Pass 'Index Split Loops' on Loop '...
0xc04028        Freeing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028        Freeing Pass 'Index Split Loops' on Loop '...
0xc04028       Executing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028       Executing Pass 'Index Split Loops' on Loop '...
0xc04028        Freeing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028        Freeing Pass 'Index Split Loops' on Loop '...
0xc04028       Executing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028       Executing Pass 'Index Split Loops' on Loop '...
0xc04028        Freeing Pass 'Loop-Closed SSA Form Pass' on Loop '...
0xc04028        Freeing Pass 'Index Split Loops' on Loop '...

Also, the 'not hoisting after unrolling' is an apparent failure of the new
LoopPassManager or a problem with the pass ordering in .  All these loop
optimizations should be run on the inner loop first (which unrolls it) then on
the outer loop (which can hoist the loads).

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  2 21:29:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 21:29:42 -0600
Subject: [LLVMbugs] [Bug 642] [PowerPC] Miscompilation of some ordered
	comparisons
In-Reply-To: <bug-642-206@http.llvm.org/bugs/>
Message-ID: <200803030329.m233Tg05002451@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=642


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #10 from Nick Lewycky <nicholas at mxc.ca>  2008-03-02 21:29:41 ---
There's still tons of fixme's in getPredicateForSetCC referencing this bug.

For example:

  case ISD::SETOEQ:    // FIXME: This is incorrect see PR642.
  case ISD::SETUEQ:
  case ISD::SETEQ:  return PPC::PRED_EQ;
  case ISD::SETONE:    // FIXME: This is incorrect see PR642.
  case ISD::SETUNE:

...


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  2 23:59:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Mar 2008 23:59:48 -0600
Subject: [LLVMbugs] [Bug 1535] SVN Conversion Related Bugs
In-Reply-To: <bug-1535-206@http.llvm.org/bugs/>
Message-ID: <200803030559.m235xmVL006732@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1535


Tanya Lattner <tonic at nondot.org> changed:

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




--- Comment #31 from Tanya Lattner <tonic at nondot.org>  2008-03-02 23:59:48 ---
This is fixed with the exception of llvm-test not working if its named
something else. If we care, a separate PR should be filed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  3 14:48:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Mar 2008 14:48:01 -0600
Subject: [LLVMbugs] [Bug 2119] New: compiling llvm-gcc4.2 fails:
 GOMP_critical_name_start: Assertion `(i >= FTy->getNumParams() ||
 FTy->getParamType(i) == Params[i]- >getType()) && "Calling a function with
 a bad signature!"' failed.
Message-ID: <bug-2119-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2119

           Summary: compiling llvm-gcc4.2 fails: GOMP_critical_name_start:
                    Assertion `(i >= FTy->getNumParams() || FTy-
                    >getParamType(i) == Params[i]->getType()) && "Calling a
                    function with a bad signature!"' 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=1487)
 --> (http://llvm.org/bugs/attachment.cgi?id=1487)
reduced testcase

Using llvm-gcc4.2 SVN r47853, llvm SVN r47853 I get an internal error.
I could only get a preprocessed .i file, even -O0 gives the error.
I reduced critical.i to testcase-min.i, looks like __sync_bool_compare_and_swap
is the cause of this.

make[4]: Entering directory
`/home/edwin/llvm-svn/obj42/x86_64-unknown-linux-gnu/libgomp'
if /bin/sh ./libtool --mode=compile /home/edwin/llvm-svn/obj42/./gcc/xgcc
-B/home/edwin/llvm-svn/obj42/./gcc/
-B/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/bin/
-B/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/lib/ -isystem
/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/include -isystem
/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/sys-include
-DHAVE_CONFIG_H -I. -I../../../llvm-gcc4.2/libgomp -I. 
-I../../../llvm-gcc4.2/libgomp/config/linux/x86
-I../../../llvm-gcc4.2/libgomp/config/linux
-I../../../llvm-gcc4.2/libgomp/config/posix -I../../../llvm-gcc4.2/libgomp 
-Wall -Werror -Wc,-pthread -O2 -g -O2  -MT critical.lo -MD -MP -MF
".deps/critical.Tpo" -c -o critical.lo ../../../llvm-gcc4.2/libgomp/critical.c;
\
        then mv -f ".deps/critical.Tpo" ".deps/critical.Plo"; else rm -f
".deps/critical.Tpo"; exit 1; fi
/home/edwin/llvm-svn/obj42/./gcc/xgcc -B/home/edwin/llvm-svn/obj42/./gcc/
-B/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/bin/
-B/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/lib/ -isystem
/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/include -isystem
/home/edwin/llvm-svn/obj42/../install/x86_64-unknown-linux-gnu/sys-include
-DHAVE_CONFIG_H -I. -I../../../llvm-gcc4.2/libgomp -I.
-I../../../llvm-gcc4.2/libgomp/config/linux/x86
-I../../../llvm-gcc4.2/libgomp/config/linux
-I../../../llvm-gcc4.2/libgomp/config/posix -I../../../llvm-gcc4.2/libgomp
-Wall -pthread -Werror -O2 -g -O2 -MT critical.lo -MD -MP -MF
.deps/critical.Tpo -c ../../../llvm-gcc4.2/libgomp/critical.c -o critical.o
cc1: Instructions.cpp:266: void llvm::CallInst::init(llvm::Value*, llvm::Value*
const*, unsigned int): Assertion `(i >= FTy->getNumParams() ||
FTy->getParamType(i) == Params[i]->getType()) && "Calling a function with a bad
signature!"' failed.
../../../llvm-gcc4.2/libgomp/critical.c: In function
'GOMP_critical_name_start':
../../../llvm-gcc4.2/libgomp/critical.c:54: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[4]: *** [critical.lo] Error 1
make[4]: Leaving directory
`/home/edwin/llvm-svn/obj42/x86_64-unknown-linux-gnu/libgomp'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory
`/home/edwin/llvm-svn/obj42/x86_64-unknown-linux-gnu/libgomp'
make[2]: *** [all] Error 2
make[2]: Leaving directory
`/home/edwin/llvm-svn/obj42/x86_64-unknown-linux-gnu/libgomp'
make[1]: *** [all-target-libgomp] Error 2
make[1]: Leaving directory `/home/edwin/llvm-svn/obj42'
make: *** [all] Error 2

This is on an x86-64 platform, so I compile without bootstrapping (it was
building successfully a while ago, can't remember which SVN version):

$ ../llvm-gcc4.2/configure   --prefix=/home/edwin/llvm-svn/obj42/../install
--program-prefix=llvm- --enable-llvm=/home/edwin/llvm-svn/llvm/
--disable-shared --disable-multilib --disable-bootstrap
--enable-languages=c,c++

$ uname -a
Linux lightspeed2 2.6.24-1-amd64 #1 SMP Mon Feb 11 13:47:43 UTC 2008 x86_64
GNU/Linux

$ gcc -v
Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v
--enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr
--enable-shared --with-system-zlib --libexecdir=/usr/lib
--without-included-gettext --enable-threads=posix --enable-nls
--with-gxx-include-dir=/usr/include/c++/4.2 --program-suffix=-4.2
--enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-mpfr
--disable-libmudflap --enable-checking=release --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.2.3 (Debian 4.2.3-1)


-- 
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 Mar  3 18:49:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Mar 2008 18:49:12 -0600
Subject: [LLVMbugs] [Bug 1501] Critical loop backedges are split and copies
 placed on the edge
In-Reply-To: <bug-1501-206@http.llvm.org/bugs/>
Message-ID: <200803040049.m240nCdr020227@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1501


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

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




--- Comment #4 from Evan Cheng <evan.cheng at apple.com>  2008-03-03 18:49:12 ---
This is done. Coalescer can commute addss to eliminate the copy.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  4 09:24:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Mar 2008 09:24:14 -0600
Subject: [LLVMbugs] [Bug 2119] compiling llvm-gcc4.2 fails:
 GOMP_critical_name_start: Assertion `(i >= FTy->getNumParams() ||
 FTy->getParamType(i) == Params[i]- >getType()) && "Calling a function with
 a bad signature!"' failed.
In-Reply-To: <bug-2119-206@http.llvm.org/bugs/>
Message-ID: <200803041524.m24FOEuJ026315@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2119


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

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




--- Comment #1 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-03-04 09:24:13 ---
fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059141.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 Mar  4 16:02:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Mar 2008 16:02:39 -0600
Subject: [LLVMbugs] [Bug 2120] New: test/lib/llvm.exp uses wrong
	substitution order
Message-ID: <bug-2120-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2120

           Summary: test/lib/llvm.exp uses wrong substitution order
           Product: Test Suite
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: major
          Priority: P2
         Component: DejaGNU
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jo at durchholz.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1491)
 --> (http://llvm.org/bugs/attachment.cgi?id=1491)
Replacement order reversal

When replacing command names from site.exp, the code first replaces %compile_c,
then %compile_cxx, but it will never find %compile_cxx since its prefix
%compile_c was already replaced.

Fix: revert the replacement order.

(Note on severity: inconsequential right now since the %compile_* macros aren't
used in test/, but the bug fix I'm about to report next will depend on 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 Mar  4 16:07:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Mar 2008 16:07:05 -0600
Subject: [LLVMbugs] [Bug 2121] New: Test suite uses wrong compiler
Message-ID: <bug-2121-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2121

           Summary: Test suite uses wrong compiler
           Product: Test Suite
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: major
          Priority: P2
         Component: DejaGNU
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jo at durchholz.org
                CC: llvmbugs at cs.uiuc.edu
        Depends on: 2120


Created an attachment (id=1492)
 --> (http://llvm.org/bugs/attachment.cgi?id=1492)
Replace calls to gcc and as with %compile_c, calls to g++ with %compile_cxx.

The test suite has hardcoded calls to gcc, g++ and as.

This means the (potentially unsuitable) default compiler will be used, instead
of the compiler that was selected for ./configure.

Fix: See patch.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  4 18:14:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Mar 2008 18:14:45 -0600
Subject: [LLVMbugs] [Bug 2014] triggered assertion in getAlignment()
In-Reply-To: <bug-2014-206@http.llvm.org/bugs/>
Message-ID: <200803050014.m250Ejml011218@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2014


Lauro Venancio <lauro.venancio at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lauro.venancio at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Lauro Venancio <lauro.venancio at gmail.com>  2008-03-04 18:14:45 ---


*** This bug has been marked as a duplicate of bug 2081 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  5 09:41:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Mar 2008 09:41:08 -0600
Subject: [LLVMbugs] [Bug 2122] New: frem does not work with floats on x86
	without SSE
Message-ID: <bug-2122-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2122

           Summary: frem does not work with floats on x86 without SSE
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas.geoffray at lip6.fr
                CC: llvmbugs at cs.uiuc.edu


The frem instruction does not work when it operates on float on X86 without
SSE.

Here's the testcase:

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


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  5 09:47:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Mar 2008 09:47:50 -0600
Subject: [LLVMbugs] [Bug 2054] llvm-ld crashes with Assertion `0 &&
 "Unmapped global?"' failed. when linking code with weakrefs on linux amd64
In-Reply-To: <bug-2054-206@http.llvm.org/bugs/>
Message-ID: <200803051547.m25FloVo015542@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2054


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

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




--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru>  2008-03-05 09:47:49 ---
Trivially fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059212.html

Testcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059213.html

Currently alias linking is heavily broken. Some amount of work will be needed
:)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  5 11:50:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Mar 2008 11:50:50 -0600
Subject: [LLVMbugs] [Bug 1993] Variants are inappropriately pruned
In-Reply-To: <bug-1993-206@http.llvm.org/bugs/>
Message-ID: <200803051750.m25HooZi019118@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1993


Scott Michel <scottm at aero.org> 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  Wed Mar  5 19:10:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Mar 2008 19:10:42 -0600
Subject: [LLVMbugs] [Bug 2114] llvm-gcc returns vector in mm0 register on
 X86, gcc uses eax:edx
In-Reply-To: <bug-2114-206@http.llvm.org/bugs/>
Message-ID: <200803060110.m261Agqh032361@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2114


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #10 from Dale Johannesen <dalej at apple.com>  2008-03-05 19:10:41 ---
OK, thanks for looking.  I'll close this 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  Thu Mar  6 00:49:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Mar 2008 00:49:18 -0600
Subject: [LLVMbugs] [Bug 1933] Invalid remainder instcombine
In-Reply-To: <bug-1933-206@http.llvm.org/bugs/>
Message-ID: <200803060649.m266nIse009137@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1933


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #21 from Nick Lewycky <nicholas at mxc.ca>  2008-03-06 00:49:17 ---
Fixed. Patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059253.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 Mar  6 15:04:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Mar 2008 15:04:02 -0600
Subject: [LLVMbugs] [Bug 2123] New: recent linker changes can drop section
	on globals
Message-ID: <bug-2123-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2123

           Summary: recent linker changes can drop section on globals
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Linker
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1498)
 --> (http://llvm.org/bugs/attachment.cgi?id=1498)
a.bc

llvm-ld -link-as-library a.bc b.bc -o c.bc

drops the section ".data.init_task" from the global init_task_union in a.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 Mar  6 19:41:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Mar 2008 19:41:18 -0600
Subject: [LLVMbugs] [Bug 2124] New: llvm-dis exhausts all memory
Message-ID: <bug-2124-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2124

           Summary: llvm-dis exhausts all memory
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-dis
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1500)
 --> (http://llvm.org/bugs/attachment.cgi?id=1500)
memory hog

llvm-dis (on both amd64 and x86) exhausted all available memory (>2 GB min)
trying to disassemble the attached file.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  6 22:08:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Mar 2008 22:08:54 -0600
Subject: [LLVMbugs] [Bug 2125] New: C backend mishandles C++ struct return
Message-ID: <bug-2125-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2125

           Summary: C backend mishandles C++ struct return
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: C
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


Compiling the following C++ testcase:

#include <string>
#include <sstream>
using namespace std;

int main()
{
  ostringstream  oss(string(1162*9, 'x'));
  oss.str();
  return 0;
}

with llvm-g++ -emit-llvm and using llc -march=c to produce a C
source file, compiling that and executing it, gets a Segmentation
Fault, at least on x86 Darwin. The problem is the call to str(), which returns
a std::string by value. The generated C source has str() returning a
small struct by value, however the library is expecting it to be returned
by hidden-argument because std::string has a copy constructor.

This is causing bigfib and possibly other llvm-test tests to fail.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  7 00:14:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Mar 2008 00:14:54 -0600
Subject: [LLVMbugs] [Bug 2125] C backend mishandles C++ struct return
In-Reply-To: <bug-2125-206@http.llvm.org/bugs/>
Message-ID: <200803070614.m276EsaW024706@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2125


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-07 00:14:54 ---
yep :(

*** This bug has been marked as a duplicate of bug 1658 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  7 00:37:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Mar 2008 00:37:01 -0600
Subject: [LLVMbugs] [Bug 2122] frem does not work with floats on x86 without
	SSE
In-Reply-To: <bug-2122-206@http.llvm.org/bugs/>
Message-ID: <200803070637.m276b1iw025548@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2122


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #2 from Chris Lattner <sabre at nondot.org>  2008-03-07 00:37:01 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059281.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 Mar  7 03:29:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Mar 2008 03:29:18 -0600
Subject: [LLVMbugs] [Bug 2126] New: something funny with the symbol table
	lookup
Message-ID: <bug-2126-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2126

           Summary: something funny with the symbol table lookup
           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: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1501)
 --> (http://llvm.org/bugs/attachment.cgi?id=1501)
optimized linux simple_throw bytecode

For some reason, on Linux the JIT can resolve a symbol while the interpreter
can't. This is SingleSource/Regression/C++/EH/simple_throw:

$ lli --force-interpreter=true x.bc
Could not resolve external global address: _ZTIi
llvm-commit/Debug/bin/lli[0x87e0322]
llvm-commit/Debug/bin/lli[0x87e0454]
[0xffffe500]
Aborted
$ lli x.bc
terminate called after throwing an instance of 'int'
llvm-commit/Debug/bin/lli[0x87e0322]
llvm-commit/Debug/bin/lli[0x87e0454]
[0xffffe500]
Aborted

(Neither behaviour is entirely correct, but the JIT is closer. The program does
throw 5, but it ought to catch it too.)

I added debugging to the JIT and it is resolving _ZTIi.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  7 13:40:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Mar 2008 13:40:51 -0600
Subject: [LLVMbugs] [Bug 2127] New: Build Problem on Open Suse x86_64
Message-ID: <bug-2127-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2127

           Summary: Build Problem on Open Suse x86_64
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: santosh.nagarakatte at gmail.com
                CC: llvmbugs at cs.uiuc.edu


LLVM-gcc won't build on this platform. I am using the latest svn check out of
llvm-gcc.

I am using Open Suse with gcc version 4.1.2 20061115 (prerelease) (SUSE Linux).

I configure it using the following 

../llvm-gcc/configure --prefix=`pwd`/../install
--enable-llvm=/mnt/eclipse/acg/users/santoshn/LLVM/latest/llvm
--enable-languages=c,c++ --disable-bootstrap --disable-shared
--disable-multilib


The error is as follows.

cc1:
/mnt/eclipse/acg/users/santoshn/LLVM/latest/llvm/include/llvm/ADT/APInt.h:894:
uint64_t llvm::APInt::getZExtValue() const: Assertion `getActiveBits() <= 64 &&
"Too many bits for uint64_t"' failed.
../../llvm-gcc/gcc/libgcc2.c:126: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[3]: *** [libgcc/./_addvdi3.o] Error 1
make[3]: Leaving directory `/mnt/eclipse/acg/users/santoshn/LLVM/obj/gcc'
make[2]: *** [stmp-multilib] Error 2
make[2]: Leaving directory `/mnt/eclipse/acg/users/santoshn/LLVM/obj/gcc'
make[1]: *** [all-gcc] Error 2
make[1]: Leaving directory `/mnt/eclipse/acg/users/santoshn/LLVM/obj'
make: *** [all] Error 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  Fri Mar  7 20:50:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Mar 2008 20:50:26 -0600
Subject: [LLVMbugs] [Bug 2126] symbol table lookup in interpreter constructor
In-Reply-To: <bug-2126-206@http.llvm.org/bugs/>
Message-ID: <200803080250.m282oQV1003834@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2126


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #4 from Nick Lewycky <nicholas at mxc.ca>  2008-03-07 20:50:25 ---
Fixed. Patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080303/059340.html

I don't think a portable testcase can be written, so no 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 Mar  8 01:06:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Mar 2008 01:06:44 -0600
Subject: [LLVMbugs] [Bug 2128] New: interpreter should bind functions by type
Message-ID: <bug-2128-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2128

           Summary: interpreter should bind functions by type
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interpreter
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Currently the Interpreter has a short list of functions it can bind to.

Instead, it should take the FunctionType, collapse it down to the ABI-relevant
bits (getTypeID already does this for each param) then cast the funcptr and
call through it.

This would mean having large switching statements (or template abuse?) to
create the casts for each possible combination of types. Or we could use
libffi, http://sources.redhat.com/libffi/ , which is very liberally licensed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  8 03:06:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Mar 2008 03:06:11 -0600
Subject: [LLVMbugs] [Bug 2130] New: Out of memory while compiling
 dsputil_mmx.c from ffmpeg
Message-ID: <bug-2130-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2130

           Summary: Out of memory while compiling dsputil_mmx.c from ffmpeg
           Product: tools
           Version: 2.2
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ismail at namtrac.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1508)
 --> (http://llvm.org/bugs/attachment.cgi?id=1508)
Preprocessed source code

Using llvm-gcc 4.2 from iPhone SDK, I get out of memory while compiling
dsputil_mmx.c, according to top cc1 allocated up to 2.5GB VSIZE and died
afterwards with out of memory 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 Mar  8 22:56:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Mar 2008 22:56:30 -0600
Subject: [LLVMbugs] [Bug 2131] New: clang -fsyntax only fires assertion on
	inline asm
Message-ID: <bug-2131-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2131

           Summary: clang -fsyntax only fires assertion on inline asm
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kremenek at cs.stanford.edu
                CC: llvmbugs at cs.uiuc.edu


LLVM version: 48064
clang version:48082

For the following input, an assertion fires using "clang -fsyntax only":

  int main(  int argc,  char **argv) {
    __asm("#DEFINITION##define " "FOO" "\t%0" : : "n" ((unsigned int)(0x04)));
  }

$ clang -fsyntax-only t.c

Assertion failed: (0 && "Unknown input constraint type!"), function
validateInputConstraint, file TargetInfo.cpp, line 186.
...

This input compiles fine with 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  Sun Mar  9 00:03:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 00:03:17 -0600
Subject: [LLVMbugs] [Bug 2131] clang -fsyntax only fires assertion on inline
	asm
In-Reply-To: <bug-2131-206@http.llvm.org/bugs/>
Message-ID: <200803090603.m2963HNC026037@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2131


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #1 from Anders Carlsson <andersca at mac.com>  2008-03-09 00:03:17 ---
Fixed in
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080303/004633.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 Mar  9 10:37:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 10:37:05 -0500
Subject: [LLVMbugs] [Bug 2132] New: x86_64: Compiling
 gcc.c-torture/compile/20050622-1. c exhausts physical memory
Message-ID: <bug-2132-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2132

           Summary: x86_64: Compiling gcc.c-torture/compile/20050622-1.c
                    exhausts physical memory
           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


I tried to run the gcc testsuite on llvm-gcc (SVN r48109, both llvm-gcc and
llvm).

When it tried to compile gcc.c-torture/compile/20050622-1.c it began using huge
amounts of memory (1.7g+), that causes excessive swapout, and I had to kill the
test. 
FWIW clang can compile it.

This only occurs on x86-64, compiling the same on x86 works fine.
$ uname -a
Linux lightspeed2 2.6.25-rc3-00347-gcad226b #2 Mon Mar 3 23:19:49 EET 2008
x86_64 GNU/Linux

To reproduce:
$ llvm-gcc gcc/testsuite/gcc.c-torture/compile/20050622-1.c -S -O0

Or even:
$ llvm-gcc gcc.c-torture/compile/20050622-1.c -fsyntax-only

Memory usage starts to climb fast, and compilation doesn't finish even after
minutes of running (on a system with 2 Gb physical RAM).

My system's gcc can compile these fine:

$ gcc -v
Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v
--enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr
--enable-shared --with-system-zlib --libexecdir=/usr/lib
--without-included-gettext --enable-threads=posix --enable-nls
--with-gxx-include-dir=/usr/include/c++/4.2 --program-suffix=-4.2
--enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-mpfr
--disable-libmudflap --enable-checking=release --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.2.3 (Debian 4.2.3-2)

$ time gcc gcc.c-torture/compile/20050622-1.c -S

real    0m0.414s
user    0m0.317s
sys     0m0.080s

$ time gcc gcc.c-torture/compile/20050622-1.c -S

real    0m0.414s
user    0m0.317s
sys     0m0.080s

I can't get bitcode, so here is the preprocessed source:
llvm-gcc -E gcc.c-torture/compile/20050622-1.c
# 1 "gcc.c-torture/compile/20050622-1.c"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "gcc.c-torture/compile/20050622-1.c"

struct S { char buf[72*1024*1024]; };




extern void bar (struct S);

struct S s;

int
foo (void)
{
  bar (s);
  return 0;
}


There are lots of test suite failures on "test for warning, test for excess
errors". Do you want a bugreport opened for that?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 10:45:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 10:45:41 -0500
Subject: [LLVMbugs] [Bug 2133] New: llvm: make check fails on SVN r48109
Message-ID: <bug-2133-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2133

           Summary: llvm: make check fails on SVN r48109
           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
        Depends on: 2123


Make check fails on x86.

===  Summary ===

# of expected passes            2289
# of unexpected failures        7
# of expected failures          5


FAIL: /home/edwin/llvm/llvm-svn/trunk/test/CodeGen/CellSPU/immed64.ll
FAIL:
/home/edwin/llvm/llvm-svn/trunk/test/CodeGen/IA64/2005-08-22-LegalizerCrash.ll
FAIL: /home/edwin/llvm/llvm-svn/trunk/test/CodeGen/IA64/2005-10-29-shladd.ll
FAIL: /home/edwin/llvm/llvm-svn/trunk/test/CodeGen/IA64/ret-0.ll
FAIL:
/home/edwin/llvm/llvm-svn/trunk/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
FAIL:
/home/edwin/llvm/llvm-svn/trunk/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll
FAIL:
/home/edwin/llvm/llvm-svn/trunk/test/Linker/2008-03-07-DroppedSection_a.ll

Full make check output 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  Sun Mar  9 11:59:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 11:59:53 -0500
Subject: [LLVMbugs] [Bug 2134] New: llc -relocation-model=pic: hangs while
 generating code for glibc-2.7/posix/wordexp.c
Message-ID: <bug-2134-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2134

           Summary: llc -relocation-model=pic: hangs while generating code
                    for glibc-2.7/posix/wordexp.c
           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=1513)
 --> (http://llvm.org/bugs/attachment.cgi?id=1513)
bugpoint reduced testcase

Using SVN r48109 llvm, and llvm-gcc4.2 on x86 compilation hangs while building
glibc-2.7/posix/wordexp.c. 
I narrowed this down to be a problem in llc.

It was tricky to get bugpoint to reduce the testcase, since I was dealing with
a hang and not a crash, so I put an alarm(1) in llc.cpp, and I got bugpoint
reducing to a small testcase.

To reproduce:
$ llc bugpoint-reduced-simplified.bc -f -relocation-model=pic

gdb shows this (sorry this is a release build, without symbols):

(gdb) r bugpoint-reduced-simplified.bc -f -relocation-model=pic
Starting program: /home/edwin/llvm/llvm-svn/trunk/Release/bin/llc
bugpoint-reduced-simplified.bc -f -relocation-model=pic
[Thread debugging using libthread_db enabled]
^C[New Thread 0xb7d586c0 (LWP 11110)]

Program received signal SIGINT, Interrupt.
[Switching to Thread 0xb7d586c0 (LWP 11110)]
0x0864fe1f in (anonymous namespace)::RALinScan::assignRegOrStackSlotAtInterval
()
(gdb) bt
#0  0x0864fe1f in (anonymous
namespace)::RALinScan::assignRegOrStackSlotAtInterval ()
#1  0x08f54ba0 in ?? ()
#2  0x00000001 in ?? ()
#3  0xb801eff4 in ?? () from /lib/ld-linux.so.2
#4  0x7c9c61e1 in ?? ()
#5  0xb801f820 in ?? ()
#6  0xbfc1ecd0 in ?? ()
#7  0xb800d5bf in ?? () from /lib/ld-linux.so.2
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
(gdb) c

I will follow-up when the debug build is complete with a better stacktrace.

I am attaching the bugpoint reduced testcase, and the preprocessed C source
too.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 12:43:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 12:43:13 -0500
Subject: [LLVMbugs] [Bug 2135] New: Wrong x86-32 lowering of select + 64 bit
	multiply
Message-ID: <bug-2135-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2135

           Summary: Wrong x86-32 lowering of select + 64 bit multiply
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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=1515)
 --> (http://llvm.org/bugs/attachment.cgi?id=1515)
testcase .ll

Running llc x.1.bc results in the assembler:
        movl    $3, %ecx
        xorl    %eax, %eax
        mull    %ecx
What happened to the select?  The problem can
also be seen with lli:

$ lli x.1.bc
... backtrace ...
Aborted (core dumped)
$ lli -force-interpreter x.1.bc
$

The wrong code seems to be generated during
legalization.  This is the cause of three recent
failures in the Ada ACATS testsuite.


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

From akyrtzi at gmail.com  Sun Mar  9 13:21:51 2008
From: akyrtzi at gmail.com (Argiris Kirtzidis)
Date: Sun, 09 Mar 2008 11:21:51 -0700
Subject: [LLVMbugs] Compiler errors on msvc
Message-ID: <47D42ABF.5040101@gmail.com>

Hi,

The attached patch fixes a couple of compiler errors on msvc:

-    std::map<const Value*,Value*>::const_iterator I = 
ValueMap.find(SAliasee);
+    std::map<const Value*,Value*>::const_iterator VMI = 
ValueMap.find(SAliasee);

error: redefinition of 'I'

+#include <algorithm>
so that std::find is usable.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: msvc-fix.zip
Type: application/octet-stream
Size: 623 bytes
Desc: not available
Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080309/7d191aa9/attachment.obj 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: msvc-fix.patch
Url: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080309/7d191aa9/attachment.pl 

From kremenek at apple.com  Sun Mar  9 13:34:26 2008
From: kremenek at apple.com (Ted Kremenek)
Date: Sun, 9 Mar 2008 11:34:26 -0700
Subject: [LLVMbugs] Compiler errors on msvc
In-Reply-To: <47D42ABF.5040101@gmail.com>
References: <47D42ABF.5040101@gmail.com>
Message-ID: <2DD33276-8417-4FDC-853F-902B8EF2BC5E@apple.com>

Applied!

http://llvm.org/viewvc/llvm-project?rev=48115&view=rev

Thanks!

On Mar 9, 2008, at 11:21 AM, Argiris Kirtzidis wrote:

> Hi,
>
> The attached patch fixes a couple of compiler errors on msvc:
>
> -    std::map<const Value*,Value*>::const_iterator I =  
> ValueMap.find(SAliasee);
> +    std::map<const Value*,Value*>::const_iterator VMI =  
> ValueMap.find(SAliasee);
>
> error: redefinition of 'I'
>
> +#include <algorithm>
> so that std::find is usable.
>
> <msvc-fix.zip>Index: lib/Linker/LinkModules.cpp
> ===================================================================
> --- lib/Linker/LinkModules.cpp	(revision 48113)
> +++ lib/Linker/LinkModules.cpp	(working copy)
> @@ -597,9 +597,9 @@
>
>     // Globals were already linked, thus we can just query ValueMap  
> for variant
>     // of SAliasee in Dest
> -    std::map<const Value*,Value*>::const_iterator I =  
> ValueMap.find(SAliasee);
> -    assert(I != ValueMap.end() && "Aliasee not linked");
> -    GlobalValue* DAliasee = cast<GlobalValue>(I->second);
> +    std::map<const Value*,Value*>::const_iterator VMI =  
> ValueMap.find(SAliasee);
> +    assert(VMI != ValueMap.end() && "Aliasee not linked");
> +    GlobalValue* DAliasee = cast<GlobalValue>(VMI->second);
>
>     // Try to find something 'similar' to SGA in destination module.
>     if (GlobalAlias *DGA = Dest->getNamedAlias(SGA->getName())) {
> Index: lib/Transforms/IPO/ExtractGV.cpp
> ===================================================================
> --- lib/Transforms/IPO/ExtractGV.cpp	(revision 48113)
> +++ lib/Transforms/IPO/ExtractGV.cpp	(working copy)
> @@ -17,6 +17,7 @@
> #include "llvm/Constants.h"
> #include "llvm/Transforms/IPO.h"
> #include "llvm/Support/Compiler.h"
> +#include <algorithm>
> using namespace llvm;
>
> namespace {
> _______________________________________________
> LLVMbugs mailing list
> LLVMbugs at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080309/730cea04/attachment.html 

From bugzilla-daemon at cs.uiuc.edu  Sun Mar  9 15:51:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 15:51:23 -0500
Subject: [LLVMbugs] [Bug 2133] llvm: make check fails on SVN r48109
In-Reply-To: <bug-2133-206@http.llvm.org/bugs/>
Message-ID: <200803092051.m29KpN4B018389@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2133


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #6 from Chris Lattner <sabre at nondot.org>  2008-03-09 15:51:23 ---
fixed, sorry for the breakage!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 16:43:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 16:43:19 -0500
Subject: [LLVMbugs] [Bug 2129] New: Out of memory while compiling
 dsputil_mmx.c from ffmpeg
Message-ID: <bug-2129-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2129

           Summary: Out of memory while compiling dsputil_mmx.c from ffmpeg
           Product: tools
           Version: 2.2
          Platform: Macintosh
        OS/Version: MacOS X
            Status: RESOLVED
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ismail at namtrac.org
                CC: llvmbugs at cs.uiuc.edu



??smail "cartman" D??nmez <ismail at namtrac.org> changed:

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


Using llvm-gcc 4.2 from iPhone SDK, I get out of memory while compiling
dsputil_mmx.c, according to top cc1 allocated up to 2.5GB VSIZE and died
afterwards with out of memory error.


--- Comment #1 from ??smail "cartman" D??nmez <ismail at namtrac.org>  2008-03-09 16:43:18 ---


*** This bug has been marked as a duplicate of bug 2130 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 16:44:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 16:44:32 -0500
Subject: [LLVMbugs] [Bug 2136] New: x86_64 make check: 2008-02-13-sret.cpp
	fails
Message-ID: <bug-2136-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2136

           Summary: x86_64 make check: 2008-02-13-sret.cpp fails
           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=1516)
 --> (http://llvm.org/bugs/attachment.cgi?id=1516)
/usr/local/bin/llvm-gcc -emit-llvm -S -O0 -emit-llvm
/home/edwin/llvm-svn/llvm/test/C++Frontend/2008-02-13-sret.cpp -o
2008-02-13-sret.ll

[Maybe this is known already, but I couldn't find any open bugs using
bugzilla's search.]

I get this with r48120 on x86-64/Linux (everything OK on x86 Linux):

XPASS: /home/edwin/llvm-svn/llvm/test/C++Frontend/2006-11-06-StackTrace.cpp
FAIL: /home/edwin/llvm-svn/llvm/test/C++Frontend/2008-02-13-sret.cpp
Failed with exit(1) at line 1
while running: /usr/local/bin/llvm-gcc -emit-llvm -S -O0 -emit-llvm
/home/edwin/llvm-svn/llvm/test/C++Frontend/2008-02-13-sret.cpp -o - | grep
{retval\|memtmp} | grep S242 |  grep {i32 1} | count 1
count: expected 1 lines and got 0.
child process exited abnormally

And this fails but that is already known:
FAIL: /home/edwin/llvm-svn/llvm/test/CFrontend/2008-01-25-ByValReadNone.c

# of expected passes            2435
# of unexpected failures        2
# of unexpected successes       1
# of expected failures          5

$ usr/local/bin/llvm-gcc -emit-llvm -S -O0 -emit-llvm
/home/edwin/llvm-svn/llvm/test/C++Frontend/2008-02-13-sret.cpp -o -|grep -E
"retval|memtmp"
        %retval = alloca %struct.S92, align 4           ; <%struct.S92*>
[#uses=2]
        %tmp14 = getelementptr %struct.S92* %retval, i32 0, i32 0              
; <i16*> [#uses=1]
        %retval17 = bitcast %struct.S92* %retval to i32*                ;
<i32*> [#uses=1]
        %retval18 = load i32* %retval17         ; <i32> [#uses=1]
        ret i32 %retval18
        %retval = alloca %struct.S93            ; <%struct.S93*> [#uses=2]
        %tmp8 = getelementptr %struct.S93* %retval, i32 0, i32 0               
; <i8**> [#uses=1]
        %retval11 = bitcast %struct.S93* %retval to i8**                ;
<i8**> [#uses=1]
        %retval12 = load i8** %retval11         ; <i8*> [#uses=1]
        ret i8* %retval12


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 18:43:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 18:43:26 -0500
Subject: [LLVMbugs] [Bug 2130] Out of memory while compiling dsputil_mmx.c
	from ffmpeg
In-Reply-To: <bug-2130-206@http.llvm.org/bugs/>
Message-ID: <200803092343.m29NhQhN023139@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2130


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #10 from Chris Lattner <sabre at nondot.org>  2008-03-09 18:43:26 ---
Both the reduced and original testcase builds file on darwin/x86 with TOT.  I
think this was a bug in the regalloc that was fixed a couple weeks ago.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar  9 21:50:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Mar 2008 21:50:30 -0500
Subject: [LLVMbugs] [Bug 2022] 2.2 does not build cleanly with GCC 4.3
In-Reply-To: <bug-2022-206@http.llvm.org/bugs/>
Message-ID: <200803100250.m2A2oULP028020@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2022


Bryan O'Sullivan <bos at serpentine.com> changed:

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




--- Comment #9 from Bryan O'Sullivan <bos at serpentine.com>  2008-03-09 21:50:30 ---
Yes, it's working now 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  Mon Mar 10 01:46:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 01:46:10 -0500
Subject: [LLVMbugs] [Bug 2120] test/lib/llvm.exp uses wrong substitution
	order
In-Reply-To: <bug-2120-206@http.llvm.org/bugs/>
Message-ID: <200803100646.m2A6kAlS009122@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2120


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-10 01:46:09 ---
Makes sense, patch looks great, applied:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059425.html

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  Mon Mar 10 01:54:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 01:54:14 -0500
Subject: [LLVMbugs] [Bug 2121] Test suite uses wrong compiler
In-Reply-To: <bug-2121-206@http.llvm.org/bugs/>
Message-ID: <200803100654.m2A6sEsU009416@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2121


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-10 01:54:14 ---
Looks great, thanks for the patch (and sorry for the delay)!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059427.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 Mar 10 12:57:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 12:57:36 -0500
Subject: [LLVMbugs] [Bug 2136] x86_64 make check: 2008-02-13-sret.cpp fails
In-Reply-To: <bug-2136-206@http.llvm.org/bugs/>
Message-ID: <200803101757.m2AHvabD005691@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2136


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #6 from Dale Johannesen <dalej at apple.com>  2008-03-10 12:57:36 ---
OK, I've checked that in 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 Mar 10 15:47:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 15:47:08 -0500
Subject: [LLVMbugs] [Bug 2135] Wrong x86-32 lowering of select + 64 bit
	multiply
In-Reply-To: <bug-2135-206@http.llvm.org/bugs/>
Message-ID: <200803102047.m2AKl8x9011558@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2135


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-03-10 15:47:06 ---

(In reply to comment #2)
> This bug was introduced by dan's apintification.  This logic:

Indeed. Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059490.html

I haven't run the ACATS tests; if they are still failing, please reopen this PR
or file a new one. 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  Mon Mar 10 17:39:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 17:39:47 -0500
Subject: [LLVMbugs] [Bug 2123] recent linker changes can drop section on
	globals
In-Reply-To: <bug-2123-206@http.llvm.org/bugs/>
Message-ID: <200803102239.m2AMdlQ1015114@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2123


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

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




--- Comment #14 from Anton Korobeynikov <asl at math.spbu.ru>  2008-03-10 17:39:46 ---
Really 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 Mar 10 18:08:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Mar 2008 18:08:00 -0500
Subject: [LLVMbugs] [Bug 2068] PPC should use byval for large byval arguments
In-Reply-To: <bug-2068-206@http.llvm.org/bugs/>
Message-ID: <200803102308.m2AN80M8016052@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2068


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #10 from Dale Johannesen <dalej at apple.com>  2008-03-10 18:08:00 ---
Yep.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 11 07:32:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 07:32:44 -0500
Subject: [LLVMbugs] [Bug 2134] llc -relocation-model=pic: hangs while
 generating code for glibc-2.7/posix/wordexp.c
In-Reply-To: <bug-2134-206@http.llvm.org/bugs/>
Message-ID: <200803111232.m2BCWiAQ014966@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2134


T??r??k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |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 Mar 11 08:07:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 08:07:34 -0500
Subject: [LLVMbugs] [Bug 2137] New: link failed glibc-2.7: libc_pic.os:
 __libc_resp accessed both as normal and thread local symbol
Message-ID: <bug-2137-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2137

           Summary: link failed glibc-2.7: libc_pic.os: __libc_resp accessed
                    both as normal and thread local symbol
           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


llvm-gcc4.2 is now able to _compile_ all of glibc (as I stated in a previous
bugreport), however I get a link failure:

make[2]: Leaving directory `/home/edwin/glibc-2.7/build-tree/glibc-2.7/elf'
/home/edwin/llvm/install/bin/llvm-gcc   -shared -static-libgcc -Wl,-O1 
-Wl,-z,defs -Wl,-dynamic-linker=/lib/ld-linux.so.2 
-B/home/edwin/glibc-2.7/build-tree/i386-libc/csu/ 
-Wl,--version-script=/home/edwin/glibc-2.7/build-tree/i386-libc/libc.map
-Wl,-soname=libc.so.6 -Wl,-z,combreloc -Wl,-z,relro -Wl,--hash-style=both
-nostdlib -nostartfiles -e __libc_main
-L/home/edwin/glibc-2.7/build-tree/i386-libc
-L/home/edwin/glibc-2.7/build-tree/i386-libc/math
-L/home/edwin/glibc-2.7/build-tree/i386-libc/elf
-L/home/edwin/glibc-2.7/build-tree/i386-libc/dlfcn
-L/home/edwin/glibc-2.7/build-tree/i386-libc/nss
-L/home/edwin/glibc-2.7/build-tree/i386-libc/nis
-L/home/edwin/glibc-2.7/build-tree/i386-libc/rt
-L/home/edwin/glibc-2.7/build-tree/i386-libc/resolv
-L/home/edwin/glibc-2.7/build-tree/i386-libc/crypt
-L/home/edwin/glibc-2.7/build-tree/i386-libc/nptl
-Wl,-rpath-link=/home/edwin/glibc-2.7/build-tree/i386-libc:/home/edwin/glibc-2.7/build-tree/i386-libc/math:/home/edwin/glibc-2.7/build-tree/i386-libc/elf:/home/edwin/glibc-2.7/build-tree/i386-libc/dlfcn:/home/edwin/glibc-2.7/build-tree/i386-libc/nss:/home/edwin/glibc-2.7/build-tree/i386-libc/nis:/home/edwin/glibc-2.7/build-tree/i386-libc/rt:/home/edwin/glibc-2.7/build-tree/i386-libc/resolv:/home/edwin/glibc-2.7/build-tree/i386-libc/crypt:/home/edwin/glibc-2.7/build-tree/i386-libc/nptl
-o /home/edwin/glibc-2.7/build-tree/i386-libc/libc.so -T
/home/edwin/glibc-2.7/build-tree/i386-libc/shlib.lds
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/abi-note.o
/home/edwin/glibc-2.7/build-tree/i386-libc/elf/soinit.os
/home/edwin/glibc-2.7/build-tree/i386-libc/libc_pic.os
/home/edwin/glibc-2.7/build-tree/i386-libc/elf/sofini.os
/home/edwin/glibc-2.7/build-tree/i386-libc/elf/interp.os
/home/edwin/glibc-2.7/build-tree/i386-libc/elf/ld.so -lgcc
/usr/bin/ld: /home/edwin/glibc-2.7/build-tree/i386-libc/libc_pic.os:
`__libc_resp' accessed both as normal and thread local symbol
/home/edwin/glibc-2.7/build-tree/i386-libc/libc_pic.os: could not read symbols:
File format not recognized
collect2: ld returned 1 exit status
make[1]: *** [/home/edwin/glibc-2.7/build-tree/i386-libc/libc.so] Error 1
make[1]: Leaving directory `/home/edwin/glibc-2.7/build-tree/glibc-2.7'
make: *** [all] Error 2


__libc_resp related things I found in the sources:
#define __resp __libc_resp

#if USE__THREAD
#undef __Resp
__thread struct __res_state *__resp = &_res;
extern __thread struct __res_state *__libc_resp
 __attribute__((alias ("__resp"))) attribute_hidden;
#endif

$ readelf -a libc_pic.os|grep resp:
000af232  00189a12 R_386_TLS_GD      00000004   __libc_resp
000afb20  00189a12 R_386_TLS_GD      00000004   __libc_resp
000d8b47  00189a03 R_386_GOT32       00000004   __libc_resp
000d8b5a  00189a03 R_386_GOT32       00000004   __libc_resp
000d8b6f  00189a03 R_386_GOT32       00000004   __libc_resp
000d8b9c  00189a03 R_386_GOT32       00000004   __libc_resp
000d8c03  00189a03 R_386_GOT32       00000004   __libc_resp
000d8c2c  00189a03 R_386_GOT32       00000004   __libc_resp
000d8d73  00189a12 R_386_TLS_GD      00000004   __libc_resp
000da2b5  00189a12 R_386_TLS_GD      00000004   __libc_resp
000da4c5  00189a12 R_386_TLS_GD      00000004   __libc_resp
000da923  00189a12 R_386_TLS_GD      00000004   __libc_resp
000daa1d  00189a12 R_386_TLS_GD      00000004   __libc_resp
000daaf9  00189a12 R_386_TLS_GD      00000004   __libc_resp
000dacb6  00189a12 R_386_TLS_GD      00000004   __libc_resp
000dae5d  00189a12 R_386_TLS_GD      00000004   __libc_resp
000df8fd  00189a12 R_386_TLS_GD      00000004   __libc_resp
000e0b7d  00189a12 R_386_TLS_GD      00000004   __libc_resp
000e134d  00189a12 R_386_TLS_GD      00000004   __libc_resp
000fc38a  00189a12 R_386_TLS_GD      00000004   __libc_resp
000000aa  00189a12 R_386_TLS_GD      00000004   __libc_resp
000000c0  00189a12 R_386_TLS_GD      00000004   __libc_resp
  6298: 00000004     4 TLS     GLOBAL DEFAULT   70 __libc_resp
  7696: 00000004     4 TLS     GLOBAL DEFAULT   70 __resp


I guess that R_386_GOT32 and R_386_TLS_GD conflict, how can I find out _why_
some symbols are R_386_GOT32, and others are R_386_TLS_GD?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 11 14:46:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 14:46:50 -0500
Subject: [LLVMbugs] [Bug 2138] New: llc: assertion fail while compiling
	wcstoll_l.c
Message-ID: <bug-2138-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2138

           Summary: llc: assertion fail while compiling wcstoll_l.c
           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=1521)
 --> (http://llvm.org/bugs/attachment.cgi?id=1521)
bugpoint reduced testcase

Sometimes when compiling wcstoll_l.c I get an internal compiler error, I
tracked it down to llc.
The error occurs approximately 1 out of 10 times when compiling the same file.
I run valgrind on it, but showed no warnings.

The llvm-gcc ICE:
$ /home/edwin/llvm/install/bin/llvm-gcc wcstoll_l.c -c -std=gnu99
-fgnu89-inline -O2 -Wall -Winline -Wwrite-strings -fmerge-all-constants
-fnested-functions -fstrict-aliasing -g -pipe -Wstrict-prototypes
-mpreferred-stack-boundary=2 -pg -I../include   -I../include
-I/home/edwin/glibc-2.7/build-tree/i386-libc/wcsmbs
-I/home/edwin/glibc-2.7/build-tree/i386-libc -I../sysdeps/i386/elf
-I../nptl/sysdeps/unix/sysv/linux/i386/i486
-I../nptl/sysdeps/unix/sysv/linux/i386 -I../sysdeps/unix/sysv/linux/i386
-I../nptl/sysdeps/unix/sysv/linux -I../nptl/sysdeps/pthread
-I../sysdeps/pthread -I../sysdeps/unix/sysv/linux -I../sysdeps/gnu
-I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sysdeps/unix/inet
-I../sysdeps/unix/sysv/i386 -I../nptl/sysdeps/unix/sysv -I../sysdeps/unix/sysv
-I../sysdeps/unix/i386 -I../nptl/sysdeps/unix -I../sysdeps/unix
-I../sysdeps/posix -I../nptl/sysdeps/i386/i486 -I../sysdeps/i386/i486
-I../sysdeps/i386/fpu -I../nptl/sysdeps/i386 -I../sysdeps/i386
-I../sysdeps/wordsize-32 -I../sysdeps/ieee754/ldbl-96
-I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754
-I../sysdeps/generic/elf -I../sysdeps/generic -I../nptl  -I.. -I../libio -I.
-nostdinc -isystem
/home/edwin/llvm/install/bin/../lib/gcc/i686-pc-linux-gnu/4.2.1/include
-isystem
/home/edwin/llvm/install/bin/../lib/gcc/i686-pc-linux-gnu/4.2.1/include-fixed
-isystem /home/edwin/glibc-2.7/debian/include -D_LIBC_REENTRANT -include
../include/libc-symbols.h  -DPROF      -D_IO_MTSAFE_IO 
cc1: VirtRegMap.h:190: int llvm::VirtRegMap::getStackSlot(unsigned int) const:
Assertion `TargetRegisterInfo::isVirtualRegister(virtReg)' failed.
wcstoll_l.c:30: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

LLC assertion failure:
$ llc bugpoint-reduced-simplified.bc -f
llc: /home/edwin/llvm/llvm-svn/trunk/include/llvm/ADT/IndexedMap.h:54: typename
std::vector<T, std::allocator<_CharT> >::const_reference llvm::IndexedMap<T,
ToIndexT>::operator[](typename ToIndexT::argument_type) const [with T = int,
ToIndexT = llvm::VirtReg2IndexFunctor]: Assertion `toIndex_(n) <
storage_.size() && "index out of bounds!"' failed.
llc[0x87eda4d]
/lib/i686/cmov/libc.so.6(abort+0x188)[0x4df0dee8]
/lib/i686/cmov/libc.so.6(__assert_fail+0xee)[0x4df0557e]
llc[0x8665b44]
[0x93b64d0]
Aborted

System is x86-32 Linux, running llvm SVN r48224.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 11 15:01:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 15:01:28 -0500
Subject: [LLVMbugs] [Bug 2139] New: sext i16 to i32 being turned into movzwl
Message-ID: <bug-2139-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2139

           Summary: sext i16 to i32 being turned into movzwl
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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=1522)
 --> (http://llvm.org/bugs/attachment.cgi?id=1522)
testcase .ll

$ llc testcase.bc
$ gcc -o testcase testcase.s
$ ./testcase
Aborted (core dumped)

Inspection of the .ll shows that it should not abort.
This seems to be because an sext is being emitted as a zext.
Reduced from a recent Ada testsuite failure.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 11 16:30:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 16:30:59 -0500
Subject: [LLVMbugs] [Bug 2139] sext i16 to i32 being turned into movzwl
In-Reply-To: <bug-2139-206@http.llvm.org/bugs/>
Message-ID: <200803112130.m2BLUxWP032044@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2139


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #2 from Dan Gohman <gohman at apple.com>  2008-03-11 16:30:59 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059578.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 Mar 11 17:45:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Mar 2008 17:45:41 -0500
Subject: [LLVMbugs] [Bug 2137] link failed glibc-2.7: libc_pic.os:
 __libc_resp accessed both as normal and thread local symbol
In-Reply-To: <bug-2137-206@http.llvm.org/bugs/>
Message-ID: <200803112245.m2BMjfBQ001876@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2137


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

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




--- Comment #8 from Anton Korobeynikov <asl at math.spbu.ru>  2008-03-11 17:45:41 ---
Fixed in:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059590.html

Testcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059591.html

Current TLS implementation for PIC is far from ideal, but hope the stuff will
work. 

PS: This is really nasty alias usage :)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 02:52:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 02:52:52 -0500
Subject: [LLVMbugs] [Bug 2140] New: missing memcpy simplification:
 forwarding into a call slot
Message-ID: <bug-2140-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2140

           Summary: missing memcpy simplification: forwarding into a call
                    slot
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


The GVN testcase 2008-02-24-MultipleUseofSRet.ll could be further
optimized (this testcase was reduced from some complex arithmetic):
currently it is optimized to:

define fastcc void @badly_optimized() nounwind  {
entry:
        %z = alloca { x86_fp80, x86_fp80 }
        %tmp = alloca { x86_fp80, x86_fp80 }
        call fastcc void @initialize( { x86_fp80, x86_fp80 }* noalias sret 
%tmp )
        %tmp1 = bitcast { x86_fp80, x86_fp80 }* %tmp to i8*
        %z3 = bitcast { x86_fp80, x86_fp80 }* %z to i8*
        call void @llvm.memcpy.i32( i8* %z3, i8* %tmp1, i32 24, i32 8 )
        %tmp5 = call fastcc x86_fp80 @passed_uninitialized( { x86_fp80,
x86_fp80 }* %z )
        ret void
}

The memcpy could be avoided by using %tmp in the call to @passed_uninitialized.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 06:19:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 06:19:04 -0500
Subject: [LLVMbugs] [Bug 2141] New: Could raise libcalls to intrinsics using
 parameter attributes
Message-ID: <bug-2141-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2141

           Summary: Could raise libcalls to intrinsics using parameter
                    attributes
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


The simplify libcalls pass could automatically turn a
call to @sin into a call to llvm.sin by checking the
profile and (this is the important bit which was lacking
before) that the call has the "readnone" attribute set.
The attribute needs to be checked because llvm.sin has
this attribute, and the @sin call might set errno or
use the rounding mode (in which case it will usually
be marked readonly).  Similarly for a bunch of other
functions which have LLVM intrinsic equivalents.  In
general I suppose the transform could be done if the
libcall has at least all the same attributes as the
intrinsic.

That said, it's not completely clear that it makes
sense to do this - what does using the intrinsic buy
us?  I guess it could simplify some code.  Also, I
notice that codegen turns any function called @sin
into a SIN node (and presumably into a processor sin
instruction if it has one).  This seems bogus since
it ignores all these issues about errno and rounding
modes, and whether @sin should be identified with the
sin function at all, etc.  However if the proposed
transform was implemented then codegen could just
convert llvm.sin, and know that any @sin function
will already have been turned into llvm.sin if that
is 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  Wed Mar 12 08:40:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 08:40:43 -0500
Subject: [LLVMbugs] [Bug 2142] New: A compiler error happened.
Message-ID: <bug-2142-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2142

           Summary: A compiler error happened.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yonggangluo at hotmail.com
                CC: llvmbugs at cs.uiuc.edu, yonggangluo at hotmail.com


Created an attachment (id=1528)
 --> (http://llvm.org/bugs/attachment.cgi?id=1528)
This is the source code file that contains the corrent source code line. I
corrent it.

In the file llvm\lib\System\Win32\Path.inc
on the line 48
the current code is 
 extern const char sys::PathSeparator = ';';


but it in 

namespace llvm {
namespace sys {
 extern const char sys::PathSeparator = ';';

****
}
}

so i think it's necessay to move 
 extern const char sys::PathSeparator = ';';
outside namespace sys.
or just remove sys:: from 
 extern const char sys::PathSeparator = ';';
Do you think so?

Then I think the right source code must be
 extern const char PathSeparator = ';';


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 10:51:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 10:51:18 -0500
Subject: [LLVMbugs] [Bug 2143] New: glibc-2.7/csu/initfini. s generated by
 llvm-gcc not understood by glibc build system
Message-ID: <bug-2143-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2143

           Summary: glibc-2.7/csu/initfini.s generated by llvm-gcc not
                    understood by glibc build system
           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=1529)
 --> (http://llvm.org/bugs/attachment.cgi?id=1529)
initfini.c

Generating crti.S, crtn.S, initfini.s is done via a call to compiler -S, and
postprocessed using sed.
Unfortunately llvm's output is not what glibc's build system expects, and the
result is wrong crti.S, crtn.S with a linker error about duplicate symbols.

As a comment in csu/Makefile says:
"In ELF we also install crti.o and crtn.o, special "initializer" and
"finalizer" files used in the link to make the .init and .fini sections work
right; both these files are built (in an arcane manner) from initfini.c"

This is the linker error:
/home/edwin/llvm/install/bin/llvm-gcc -nostdlib -nostartfiles -o
/home/edwin/glibc-2.7/build-tree/i386-libc/iconv/iconvconfig 
-Wl,-dynamic-linker=/lib/ld-linux.so.2   -Wl,-z,combreloc -Wl,-z,relro
-Wl,--hash-style=both /home/edwin/glibc-2.7/build-tree/i386-libc/csu/crt1.o
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.o
`/home/edwin/llvm/install/bin/llvm-gcc --print-file-name=crtbegin.o`
/home/edwin/glibc-2.7/build-tree/i386-libc/iconv/iconvconfig.o
/home/edwin/glibc-2.7/build-tree/i386-libc/iconv/strtab.o
/home/edwin/glibc-2.7/build-tree/i386-libc/iconv/xmalloc.o
/home/edwin/glibc-2.7/build-tree/i386-libc/iconv/hash-string.o 
-Wl,-rpath-link=/home/edwin/glibc-2.7/build-tree/i386-libc:/home/edwin/glibc-2.7/build-tree/i386-libc/math:/home/edwin/glibc-2.7/build-tree/i386-libc/elf:/home/edwin/glibc-2.7/build-tree/i386-libc/dlfcn:/home/edwin/glibc-2.7/build-tree/i386-libc/nss:/home/edwin/glibc-2.7/build-tree/i386-libc/nis:/home/edwin/glibc-2.7/build-tree/i386-libc/rt:/home/edwin/glibc-2.7/build-tree/i386-libc/resolv:/home/edwin/glibc-2.7/build-tree/i386-libc/crypt:/home/edwin/glibc-2.7/build-tree/i386-libc/nptl
/home/edwin/glibc-2.7/build-tree/i386-libc/libc.so.6
/home/edwin/glibc-2.7/build-tree/i386-libc/libc_nonshared.a -lgcc
-Wl,--as-needed -lgcc_s  -Wl,--no-as-needed
`/home/edwin/llvm/install/bin/llvm-gcc --print-file-name=crtend.o`
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o: In function `dummy':
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.S:15: multiple definition
of `dummy'
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.o:/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.S:36:
first defined here
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o: In function `_init':
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.S:41: multiple definition
of `_init'
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.o:/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.S:88:
first defined here
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o: In function `_fini':
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.S:100: multiple definition
of `_fini'
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.o:/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.S:152:
first defined here
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crti.o: In function `_fini':
(.fini+0x14): undefined reference to `i_am_not_a_leaf'
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o: In function `_fini':
(.fini+0x27): undefined reference to `i_am_not_a_leaf'
/home/edwin/glibc-2.7/build-tree/i386-libc/csu/crtn.o: In function `_fini':
(.fini+0x2c): undefined reference to `i_am_not_a_leaf'
collect2: ld returned 1 exit status

I will attach initfini.c, and the initfini.s, crti.S, crtn.S generated by gcc
and llvm.

At a first glance it looks like inline assembly comments are emitted at the
wrong place (too early):

csu/initfini.s is the gcc one, ../i386-libc/csu/initfini.s is llvm's.

--- csu/initfini.s      2008-03-12 17:36:26.837217313 +0200
+++ ../i386-libc/csu/initfini.s 2008-03-12 17:12:39.480693562 +0200
@@ -1,36 +1,57 @@
-       .file   "initfini.c"
-#APP
+# Start of file scope inline assembly

 #include "defs.h"

 /*@HEADER_ENDS*/

 /*@TESTS_BEGIN*/
-#NO_APP
+
+/*@TESTS_END*/
+
+/*@_init_PROLOG_BEGINS*/
+.section .init
+END_INIT
+
+/*@_init_EPILOG_ENDS*/
+
+/*@_fini_PROLOG_BEGINS*/
+.section .fini
+END_FINI
+
+/*@_fini_EPILOG_ENDS*/
+
+/*@TRAILER_BEGINS*/
+# End of file scope inline assembly
+
+
        .text
-       .p2align 2,,3
+       .align  16
 .globl dummy
        .type   dummy, @function
 dummy:
        pushl   %ebp
        movl    %esp, %ebp
-       subl    $8, %esp
+       pushl   %ebx
+       subl    $4, %esp
+       call    .Lllvm$1.$piclabel
+.Lllvm$1.$piclabel:
+       popl    %ebx
+       addl    $_GLOBAL_OFFSET_TABLE_ + [.-.Lllvm$1.$piclabel], %ebx
        movl    8(%ebp), %eax
        testl   %eax, %eax
-       je      .L4
+       je      .LBB1_2 # UnifiedReturnBlock
+.LBB1_1:       # bb
        call    *%eax
-.L4:
-       leave
+.LBB1_2:       # UnifiedReturnBlock
+       addl    $4, %esp
+       popl    %ebx
+       popl    %ebp
        ret
-#APP
+       .size   dummy, .-dummy

-/*@TESTS_END*/

-/*@_init_PROLOG_BEGINS*/
        .section .init
-#NO_APP
-       .section        .init,"ax", at progbits
-       .p2align 2,,3
+       .align  16
 .globl _init
        .type   _init, @function
 _init:
@@ -38,38 +59,41 @@
        movl    %esp, %ebp
        pushl   %ebx
        subl    $4, %esp
-       call    .L10
-.L10:
+       call    .Lllvm$2.$piclabel
+.Lllvm$2.$piclabel:
        popl    %ebx
-       addl    $_GLOBAL_OFFSET_TABLE_+[.-.L10], %ebx
-       movl    __gmon_start__ at GOT(%ebx), %edx
-       testl   %edx, %edx
-       je      .L7
+       addl    $_GLOBAL_OFFSET_TABLE_ + [.-.Lllvm$2.$piclabel], %ebx
+       cmpl    $0, __gmon_start__ at GOT(%ebx)
+       je      .LBB2_2 # bb4
+.LBB2_1:       # bb
        call    __gmon_start__ at PLT
-.L7:
+.LBB2_2:       # bb4
 #APP
        ALIGN
+       #NO_APP
+       #APP
        END_INIT
+       #NO_APP
+       #APP

 /*@_init_PROLOG_ENDS*/
+       #NO_APP
+       #APP

 /*@_init_EPILOG_BEGINS*/
+       #NO_APP
+       #APP
        .section .init
 #NO_APP
-       popl    %eax
+       addl    $4, %esp
        popl    %ebx
-       leave
+       popl    %ebp
        ret
-#APP
-       END_INIT
+       .size   _init, .-_init

-/*@_init_EPILOG_ENDS*/

-/*@_fini_PROLOG_BEGINS*/
        .section .fini
-#NO_APP
-       .section        .fini,"ax", at progbits
-       .p2align 2,,3
+       .align  16
 .globl _fini
        .type   _fini, @function
 _fini:
@@ -77,13 +101,17 @@
        movl    %esp, %ebp
        pushl   %ebx
        subl    $4, %esp
-       call    .L13
-.L13:
+       call    .Lllvm$3.$piclabel
+.Lllvm$3.$piclabel:
        popl    %ebx
-       addl    $_GLOBAL_OFFSET_TABLE_+[.-.L13], %ebx
+       addl    $_GLOBAL_OFFSET_TABLE_ + [.-.Lllvm$3.$piclabel], %ebx
 #APP
        ALIGN
+       #NO_APP
+       #APP
        END_FINI
+       #NO_APP
+       #APP

 /*@_fini_PROLOG_ENDS*/
 #NO_APP
@@ -91,18 +119,15 @@
 #APP

 /*@_fini_EPILOG_BEGINS*/
+       #NO_APP
+       #APP
        .section .fini
 #NO_APP
-       popl    %ecx
+       addl    $4, %esp
        popl    %ebx
-       leave
+       popl    %ebp
        ret
-#APP
-       END_FINI
-       
-/*@_fini_EPILOG_ENDS*/
-       
-/*@TRAILER_BEGINS*/
+       .size   _fini, .-_fini
        .weak   __gmon_start__
-       .ident  "GCC: (GNU) 4.2.3 (Debian 4.2.3-2)"
-       .section        .note.GNU-stack,"", at progbits
+
+       .ident  "GCC: (GNU) 4.2.1 (Based on Apple Inc. build 5546) (LLVM
build)"


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

From daniel at zuster.org  Wed Mar 12 13:51:07 2008
From: daniel at zuster.org (Daniel Dunbar)
Date: Wed, 12 Mar 2008 11:51:07 -0700 (PDT)
Subject: [LLVMbugs] PATCH: Unbreak DynamicLibrary::GetAddressOfSymbol from
	ltdl removal
Message-ID: <443980.66815.qm@web54602.mail.re2.yahoo.com>

var YAHOO = {'Shortcuts' : {}};
YAHOO.Shortcuts.hasSensitiveText = false;
YAHOO.Shortcuts.sensitivityType = [];
YAHOO.Shortcuts.doUlt = false;
YAHOO.Shortcuts.location = "us";
YAHOO.Shortcuts.document_id = 0;
YAHOO.Shortcuts.document_type = "";
YAHOO.Shortcuts.document_title = "PATCH: Unbreak DynamicLibrary::GetAddressOfSymbol for ltdl removal";
YAHOO.Shortcuts.document_publish_date = "";
YAHOO.Shortcuts.document_author = "daniel at zuster.org";
YAHOO.Shortcuts.document_url = "";
YAHOO.Shortcuts.document_tags = "";
YAHOO.Shortcuts.annotationSet = {

};
<!-- DIV {margin:0px;}-->Unbreak DynamicLibrary::GetAddressOfSymbol from ltdl removal
 - Seems to be broken by rev47065
 - Removed (unused, always zero) handle member variable and always
    failing assert.

This function doesn't seem to be used by LLVM, perhaps it should be 
removed or indicate in the documentation that SearchForAddressOfSymbol
is preferred.

 - Daniel








-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080312/3da9916a/attachment.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DynamicLibrary-GetAddressOfSymbol-Fix.patch
Type: application/octet-stream
Size: 1267 bytes
Desc: not available
Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080312/3da9916a/attachment.obj 

From bugzilla-daemon at cs.uiuc.edu  Wed Mar 12 17:35:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 17:35:56 -0500
Subject: [LLVMbugs] [Bug 2144] New: Verifier allows invalid aliases
Message-ID: <bug-2144-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2144

           Summary: Verifier allows invalid aliases
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Keywords: accepts-invalid
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


The verifier check for aliases is not strong enough, in particular it does not
check that the aliasee is a function or variable. Patch and test case attached.

I'm assuming that the (assembly language) documentation is correct and that
aliases to aliases are illegal. However, there is an issue in that the
documentation says that bitcasts of global *values* are legal. I think should
be bitcasts of global variables or functions, but this is not reflected in my
patch.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 17:51:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 17:51:46 -0500
Subject: [LLVMbugs] [Bug 2145] New: llvm-gcc emits invalid aliases
Message-ID: <bug-2145-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2145

           Summary: llvm-gcc emits invalid aliases
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1538)
 --> (http://llvm.org/bugs/attachment.cgi?id=1538)
problem input

The attached code emits an alias that references another alias, which generates
invalid bitcode (according to the assembly documentation). It isn't clear to me
whether or not this is intended to work. The x86 backend compiles it fine but
AsmWriter fails on it, and linking it as bytecode does not seem to work (more
testcases available on request).

--
ddunbar at ozzy:test$ llvm-gcc -c test.c && nm test.o
00000000 T a
00000000 T b
00000000 T c
ddunbar at ozzy:test$ llvm-gcc -O4 -c test.c && llvm-nm test.o
         T a
         T b
         ? c
ddunbar at ozzy:test$ llvm-gcc -O4 -c test.c && llc test.o
ddunbar at ozzy:test$ llvm-gcc -O4 -c test.c && llvm-dis test.o
llvm-dis: AsmWriter.cpp:1012: void llvm::AssemblyWriter::printAlias(const
llvm::GlobalAlias*): Assertion `0 && "Unsupported aliasee"' failed.
llvm-dis(_ZN40_GLOBAL__N_Signals.cpp_00000000_644EB5E213SignalHandlerEi+0x20f)[0x814f5ff]
/lib/tls/i686/cmov/libc.so.6(abort+0x109)[0xb7cbdfb9]
/lib/tls/i686/cmov/libc.so.6(__assert_fail+0x10f)[0xb7cb5fbf]
llvm-dis(_ZN4llvm14AssemblyWriter10printAliasEPKNS_11GlobalAliasE+0x14d)[0x8099b1d]
[0xbfeb6934]
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 Mar 12 19:27:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 19:27:32 -0500
Subject: [LLVMbugs] [Bug 2146] New: Linker does not merge aliases and
 prototypes correctly
Message-ID: <bug-2146-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2146

           Summary: Linker does not merge aliases and prototypes correctly
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Linker
        AssignedTo: asl at math.spbu.ru
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


If a symbol is declared as an alias in one linkage unit and external in another
then the linking fails (because the linking first includes the proto, then
tries to link the alias and use force renaming, which complains that the
conflicting symbol is external). I suspect this bug exists for global variables
as well?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 20:32:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 20:32:18 -0500
Subject: [LLVMbugs] [Bug 2147] New: Comment is incorrect
Message-ID: <bug-2147-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2147

           Summary: Comment is incorrect
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: trivial
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: overminddl1 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The comment in the fibonacci.cpp file is incorrect and needs to be changed from
// Create the "if (arg < 2) goto exitbb"
to
// Create the "if (arg <= 2) goto exitbb"

Got me confused on an instruction until I actually looked at it, simple fix,
line 61 as of the date of this bug submit in SVN.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 21:10:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 21:10:45 -0500
Subject: [LLVMbugs] [Bug 2148] New: 3 VC8 Build Scripts (Projects) are out
 of date, patches included
Message-ID: <bug-2148-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2148

           Summary: 3 VC8 Build Scripts (Projects) are out of date, patches
                    included
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: critical
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: overminddl1 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1540)
 --> (http://llvm.org/bugs/attachment.cgi?id=1540)
Patches to fix the VC8 projects.

Three VC8 projects are out of date with the SVN head as of the date of this bug
addition.

lib\Transforms\Utils\LowerSelect.cpp lib\Transforms\Scalar\LowerPacked.cpp
lib\Transforms\Scalar\CorrelatedExprs.cpp need to be *removed* from the VC8
project 'Transforms'
lib\CodeGen\PseudoSourceValue.cpp and lib\CodeGen\LoopAligner.cpp need to be
*added* to the VC8 project 'CodeGen'
lib\Transforms\IPO\StructRetPromotion.cpp needs to be *added* to the VC8
project 'Transforms'
lib\ExecutionEngine\JIT\JITDwarfEmitter.h and
lib\ExecutionEngine\JIT\JITDwarfEmitter.cpp need to be *added* to the VC8
project 'ExecutionEngine'

And the dobison.cmd file needs some work to always work when it detects that
bison exists (as it did with me), but it is non-vital and I need to look at
that one some more (I barely noticed the issue, just copied things manually
when I noticed that it did not work), thus it is not blocking, but if anyone
wants to look at it, feel free.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 12 22:30:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Mar 2008 22:30:12 -0500
Subject: [LLVMbugs] [Bug 2147] Comment is incorrect
In-Reply-To: <bug-2147-206@http.llvm.org/bugs/>
Message-ID: <200803130330.m2D3UCPD028258@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2147


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-12 22:30:12 ---
Thanks, fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059663.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 Mar 13 00:18:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 00:18:24 -0500
Subject: [LLVMbugs] [Bug 2142] A compiler error happened.
In-Reply-To: <bug-2142-206@http.llvm.org/bugs/>
Message-ID: <200803130518.m2D5IOUC031320@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2142


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-13 00:18:24 ---
Fixed, thanks:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059670.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 Mar 13 05:30:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 05:30:12 -0500
Subject: [LLVMbugs] [Bug 2149] New: inlining unwind breaks eh intrinsics
Message-ID: <bug-2149-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2149

           Summary: inlining unwind breaks eh intrinsics
           Product: libraries
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


The comment at the top of HandleInlinedInvoke() in
lib/Transforms/Utils/InlineFunction.cpp says:

/// HandleInlinedInvoke - If we inlined an invoke site, we need to convert
calls
/// in the body of the inlined function into invokes and turn unwind
/// instructions into branches to the invoke unwind dest.

I don't think it's safe to inline an unwind instruction and turn it into a
branch to the invoke unwind destination, because any calls to llvm.eh.exception
and llvm.eh.selector at the destination won't work.

(I'm assuming that the code generator implements these eh intrinsics by getting
the exception pointer and exception selector from well-known native registers
which are set up as part of the stack unwinding process. If you branch directly
to the invoke unwind destination then these registers won't be set up.)


My suggested solution is not to inline any function containing an unwind
instruction.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 13 05:50:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 05:50:02 -0500
Subject: [LLVMbugs] [Bug 2149] inlining unwind breaks eh intrinsics
In-Reply-To: <bug-2149-206@http.llvm.org/bugs/>
Message-ID: <200803131050.m2DAo2RN016264@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2149


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2008-03-13 05:49:59 ---
Currently the unwind instruction is not handled at all by codegen,
so there can't be a problem due to this transform!  Unwinding is
actually handled by calls to runtime functions (these calls are
generated by llvm-gccc), for which everything works correctly.
I actually have a plan to make it possible to use the unwind
instruction, which does indeed require copying the contents of
the eh registers into a variable at the start of each landing
pad, and having the intrinsics grab the value of the variable
not of the register directly.  I'm marking this report invalid,
since it is.  Please feel free to re-open if you have a 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  Thu Mar 13 09:13:54 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 09:13:54 -0500
Subject: [LLVMbugs] [Bug 2150] New: No warning when reaching end of a
	non-void function.
Message-ID: <bug-2150-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2150

           Summary: No warning when reaching end of a non-void function.
           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: richard.smith at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


$ cat foo.c
int foo() {}
$ gcc foo.c -c -Wall
foo.c: In function ???foo???:
foo.c:1: warning: control reaches end of non-void function
$ llvm-gcc foo.c -c -Wall
$

In gcc that warning comes out of the tree optimisation passes, which llvm does
not use. It would be nice if it did an equivalent check, 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  Thu Mar 13 10:54:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 10:54:18 -0500
Subject: [LLVMbugs] [Bug 2149] inlining unwind breaks eh intrinsics
In-Reply-To: <bug-2149-206@http.llvm.org/bugs/>
Message-ID: <200803131554.m2DFsIEP025244@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2149


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |alenhar2 at uiuc.edu
         Resolution|INVALID                     |WONTFIX




--- Comment #2 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-03-13 10:54:17 ---
Runtime calls are how throwing is currently implemented, but unwind is in some
sense the official llvm way to throw an exception (as specified by the
LangRef).  That it has been completely disabled seems to indicate a serious
bug.  Either unwind should work (in some situations), or it should be removed
from the instruction set.  If the LangRef had a large warning that you have to
run the LowerInvoke pass for unwind to behave as specified, that might also be
an acceptable solution.

So it sounds much more like won't fix than 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  Thu Mar 13 11:08:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 11:08:10 -0500
Subject: [LLVMbugs] [Bug 2149] inlining unwind breaks eh intrinsics
In-Reply-To: <bug-2149-206@http.llvm.org/bugs/>
Message-ID: <200803131608.m2DG8AUn025710@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2149


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|WONTFIX                     |INVALID




--- Comment #4 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-03-13 11:08:10 ---
fair enough.


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

From dpatel at apple.com  Thu Mar 13 11:56:15 2008
From: dpatel at apple.com (Devang Patel)
Date: Thu, 13 Mar 2008 09:56:15 -0700
Subject: [LLVMbugs] PATCH: Unbreak DynamicLibrary::GetAddressOfSymbol
	from	ltdl removal
In-Reply-To: <443980.66815.qm@web54602.mail.re2.yahoo.com>
References: <443980.66815.qm@web54602.mail.re2.yahoo.com>
Message-ID: <3103309E-4DC5-4C4B-98C9-3BC1EFAAAB37@apple.com>


On Mar 12, 2008, at 11:51 AM, Daniel Dunbar wrote:

> Unbreak DynamicLibrary::GetAddressOfSymbol from ltdl removal
>  - Seems to be broken by rev47065
>  - Removed (unused, always zero) handle member variable and always
>     failing assert.
>
> This function doesn't seem to be used by LLVM, perhaps it should be
> removed or indicate in the documentation that SearchForAddressOfSymbol
> is preferred.
>
>  - Daniel

Done.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059691.html

-
Devang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20080313/5e9d5e6b/attachment.html 

From bugzilla-daemon at cs.uiuc.edu  Thu Mar 13 16:28:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 16:28:59 -0500
Subject: [LLVMbugs] [Bug 2151] New: clang -fsyntax-only runs forever on
	provided input
Message-ID: <bug-2151-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2151

           Summary: clang -fsyntax-only runs forever on provided input
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kremenek at cs.stanford.edu
                CC: llvmbugs at cs.uiuc.edu, snaroff at apple.com


I've attached an input file that causes "clang -fsyntax-only" to not terminate.
 The file was the intermediate byproduct of running Delta Debugging to generate
a reduced test case for a bug I was encountering in the static analyzer.  I
cannot isolate that bug, however, because the semantic analyzer never finishes
processing the file (which I doubt is valid source code).  I don't know if I
can make the input file more reduced (or legible); it's a reduced code snippet
from a truly hairy function from PCRE.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 13 23:37:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Mar 2008 23:37:48 -0500
Subject: [LLVMbugs] [Bug 2152] New: g++ 4.3.0 report a compiling error
Message-ID: <bug-2152-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2152

           Summary: g++ 4.3.0 report a compiling error
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yonggangluo at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


The environment that i used is msys+mingw.
gcc compiler version is 4.3.0
binutils-2.18.50-20080109.tar.gz
mingw32-make-3.81-2.tar.gz
mingw-runtime-3.14.tar.gz
w32api-3.11.tar.gz
coreutils-5.97.tar.gz

$ make
make[1]: Entering directory `/d/llvm/lib/System'
llvm[1]: Compiling Alarm.cpp for Debug build 
llvm[1]: Compiling Disassembler.cpp for Debug build 
llvm[1]: Compiling DynamicLibrary.cpp for Debug build 
In file included from DynamicLibrary.cpp:35:
Win32/DynamicLibrary.inc: In constructor
'llvm::sys::DynamicLibrary::DynamicLibr                                        
                                    ary()':
Win32/DynamicLibrary.inc:80: error: class 'llvm::sys::DynamicLibrary' does not
h                                                                            
ave any field named 'handle'
Win32/DynamicLibrary.inc:81: error: 'handle' was not declared in this scope
Win32/DynamicLibrary.inc: In destructor
'llvm::sys::DynamicLibrary::~DynamicLibr                                       
                                     ary()':
Win32/DynamicLibrary.inc:86: error: 'handle' was not declared in this scope
Win32/DynamicLibrary.inc:91: error: 'handle' was not declared in this scope
Win32/DynamicLibrary.inc:93: error: 'handle' was not declared in this scope
Win32/DynamicLibrary.inc: At global scope:
Win32/DynamicLibrary.inc:179: error: no 'void*
llvm::sys::DynamicLibrary::GetAdd                                              
                              ressOfSymbol(const char*)' member function
declared in class 'llvm::sys::Dynamic                                          
                                  Library'
make[1]: *** [/d/llvm/lib/System/Debug/DynamicLibrary.o] Error 1
make[1]: Leaving directory `/d/llvm/lib/System'
make: *** [all] 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  Fri Mar 14 12:56:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 12:56:07 -0500
Subject: [LLVMbugs] [Bug 2153] New: simd-4.C fails wrongly at -O2
Message-ID: <bug-2153-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2153

           Summary: simd-4.C fails wrongly at -O2
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


simd-4.C from the gcc testsuite (g++.dg/eh) fails differently at
-O0 and at -O2 while with mainline you always get the llvm-gcc -O0
behaviour:

$ llvm-g++ -o simd-4 -O0 simd-4.C
$ ./simd-4
terminate called after throwing an instance of 'int'
Aborted (core dumped)

$ llvm-g++ -o simd-4 -O2 simd-4.C
$ ./simd-4
Aborted (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  Fri Mar 14 13:00:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 13:00:02 -0500
Subject: [LLVMbugs] [Bug 2154] New: invoke+terminate could be simplified at
	codegen time
Message-ID: <bug-2154-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2154

           Summary: invoke+terminate could be simplified at codegen time
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


A call in a gcc must-not-throw region results in an invoke
that calls terminate in the unwind basic block.  This could
be turned into a call at codegen time as long as a "nothrow"
entry is entered into the dwarf eh table, resulting in a
reduction in code size.  Since this is C++ specific, the
transform should only be done when the llvm.selector call
uses the C++ personality function.  We used not to need this
transform because we enforced must-not-throw region by adding
funky meaning to the 'nounwind' attribute, but we don't do
that 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 Mar 14 15:47:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 15:47:07 -0500
Subject: [LLVMbugs] [Bug 2138] llc: assertion fail while compiling
	wcstoll_l.c
In-Reply-To: <bug-2138-206@http.llvm.org/bugs/>
Message-ID: <200803142047.m2EKl7ro000721@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2138


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

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




--- Comment #15 from Evan Cheng <evan.cheng at apple.com>  2008-03-14 15:47:06 ---
Ok, std::multimap iterator can be invalidated with changes to elements of a
different key. Increments the iterator first fixed the crash. Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080310/059733.html

Thanks for the test case!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 14 16:45:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 16:45:19 -0500
Subject: [LLVMbugs] [Bug 2059] Missed sret memcpy optimization
In-Reply-To: <bug-2059-206@http.llvm.org/bugs/>
Message-ID: <200803142145.m2ELjJxf002715@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2059


Owen Anderson <resistor at mac.com> changed:

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




--- Comment #21 from Owen Anderson <resistor at mac.com>  2008-03-14 16:45:18 ---
GVN doesn't attempt to do DCE.  There are other passes that do that. ;-)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 14 17:04:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 17:04:41 -0500
Subject: [LLVMbugs] [Bug 2155] New: Debug+Checks make check fails on x86
Message-ID: <bug-2155-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2155

           Summary: Debug+Checks make check fails on x86
           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=1545)
 --> (http://llvm.org/bugs/attachment.cgi?id=1545)
make check output from SVN HEAD

llvm SVN r48371, on x86 Linux running Debian sid, gcc/g++ 4.2.3, libstdc++6-4.2

I run:
$ make check ENABLE_OPTIMIZED=0 ENABLE_EXPENSIVE_CHECKS=1
...
                ===  Summary ===

# of expected passes            2279
# of unexpected failures        3
# of expected failures          4

I apply David Greene's ScheduleDAG patch from the mailing list:

$ make check ENABLE_OPTIMIZED=0 ENABLE_EXPENSIVE_CHECKS=1
...
                ===  Summary ===

# of expected passes            2279
# of unexpected failures        3
# of expected failures          4

This is one additional failure, however I am not sure if it exposes a
previously latent bug, or it is a new bug.

I am attaching make check outputs. 
Please let me know if you want me to run bugpoint on any of the failures (and
which build, w/ or w/o ScheduleDAG patch).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 14 17:13:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Mar 2008 17:13:09 -0500
Subject: [LLVMbugs] [Bug 2156] New: x86-64
 test/CodeGen/X86/2006-05-02-InstrSched2.ll fails
Message-ID: <bug-2156-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2156

           Summary: x86-64  test/CodeGen/X86/2006-05-02-InstrSched2.ll fails
           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=1547)
 --> (http://llvm.org/bugs/attachment.cgi?id=1547)
fail.bc to reproduce

When running make check on r48371, Linux x86-64:

FAIL: /home/edwin/llvm-svn/llvm/test/CodeGen/X86/2006-05-02-InstrSched2.ll
Failed with exit(1) at line 1
while running: llvm-upgrade <
/home/edwin/llvm-svn/llvm/test/CodeGen/X86/2006-05-02-InstrSched2.ll | llvm-as
| llc -march=x86 -stats |&  grep asm-printer | grep 13
child killed: segmentation violation
child process exited abnormally

If I run it manually I get:

$ llvm-upgrade <
/home/edwin/llvm-svn/llvm/test/CodeGen/X86/2006-05-02-InstrSched2.ll | llvm-as
| llc -march=x86 -stats |  grep asm-printer | grep 13
llc[0x9aed66]
/lib/libc.so.6[0x3932831f70]
llc[0x82db2f]
llc(_ZN4llvm19MachineFunctionPass13runOnFunctionERNS_8FunctionE+0x29)[0x5f00e9]
llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x231)[0x94ca11]
llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x1c9)[0x94d639]
llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0x31)[0x94d791]
llc(main+0x993)[0x4b42d3]
/lib/libc.so.6(__libc_start_main+0xfa)[0x393281e1ea]
llc(cos+0xf1)[0x4b1cd9]

If I remove either of '-stats' or 'o -' option from llc it doesn't crash
anymore.

I tried to get a reduced testcase with bugpoint, but running llc on
bugpoint-reduced-simplified.bc didn't reproduce the crash (although it was
crashing when called from bugpoint).

Here's a backtrace from a debug build (on x86-64, Linux):
(gdb) r -march=x86 fail.bc -f -o - -stats
Starting program: /home/edwin/llvm-svn/llvm/Debug/bin/llc -march=x86 fail.bc -f
-o - -stats
[Thread debugging using libthread_db enabled]
[New Thread 0x7f294d0c76f0 (LWP 14326)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f294d0c76f0 (LWP 14326)]
0x000000000089e16a in llvm::ilist_traits<llvm::MachineInstr>::getNext (N=0x91)
    at /home/edwin/llvm-svn/llvm/include/llvm/CodeGen/MachineBasicBlock.h:37
37        static MachineInstr* getNext(MachineInstr* N) { return N->Next; }
(gdb) bt
#0  0x000000000089e16a in llvm::ilist_traits<llvm::MachineInstr>::getNext
(N=0x91)
    at /home/edwin/llvm-svn/llvm/include/llvm/CodeGen/MachineBasicBlock.h:37
#1  0x000000000089f98c in llvm::ilist_iterator<llvm::MachineInstr>::operator*
(this=0x7fff550ed080)
    at /home/edwin/llvm-svn/llvm/include/llvm/ADT/ilist:121
#2  0x000000000089f9d3 in llvm::ilist_iterator<llvm::MachineInstr>::operator->
(this=0x7fff550ed080)
    at /home/edwin/llvm-svn/llvm/include/llvm/ADT/ilist:124
#3  0x0000000000ae0837 in runOnMachineFunction (this=0x1157a40, MF=@0x115f7a0)
at TwoAddressInstructionPass.cpp:204
#4  0x0000000000898ae9 in llvm::MachineFunctionPass::runOnFunction
(this=0x1157a40, F=@0x11454c0)
    at /home/edwin/llvm-svn/llvm/include/llvm/CodeGen/MachineFunctionPass.h:41
#5  0x0000000000c17e76 in llvm::FPPassManager::runOnFunction (this=0x1149760,
F=@0x11454c0) at PassManager.cpp:1167
#6  0x0000000000c1809c in llvm::FunctionPassManagerImpl::run (this=0x1144830,
F=@0x11454c0) at PassManager.cpp:1125
#7  0x0000000000c181fa in llvm::FunctionPassManager::run (this=0x7fff550ed320,
F=@0x11454c0) at PassManager.cpp:1070
#8  0x000000000077d1dc in main (argc=7, argv=0x7fff550ed548) at llc.cpp:296
(gdb)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 15 12:41:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Mar 2008 12:41:58 -0500
Subject: [LLVMbugs] [Bug 2157] New: Filter expressions could (should?) be
 removed from LLVM exception handling
Message-ID: <bug-2157-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2157

           Summary: Filter expressions could (should?) be removed from LLVM
                    exception handling
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


Since we are currently eliminating magic semantics from LLVM's
exception handling, how about eliminating filters?  The unwinder
allows for two ways of comparing an exception with a list of
typeinfos (typeinfos correspond to classes in C++): (a) does the
exception match a typeinfo in the list, or (the opposite) (b)
does the exception fail to match any typeinfo in the list?  Case
(b) is the filter case, case (a) is the selector case.

Note how case (b) can be implemented using case (a): just
invert the result!  And in fact when I started working on
exception handling, one of the first things I did was to
eliminate filters and only make use of selectors.  However
it didn't work because in C++ filters have a magic property
(reminiscent of must-not-throw regions, which can be thought
of as a filter with an empty list of typeinfos): if the handler
for a filter throws an exception, the runtime checks whether
the new exception fails to match the filter, and if so calls
terminate (this should sound familiar to anyone who followed
the recent 'nounwind' saga).

So I'm proposing eliminating filters, and instead implementing
them using selectors and explicit extra invoke+terminate code,
eliminating this magic property.  This would be a conceptual
simplification of LLVM exception handling.

There are some disadvantages: due to the way dwarf eh actually
works, it is more efficient to use a filter than to use a
selector and inverting the result, because "inverting the
result" requires performing comparing N comparisons where N
is the number of typeinfos, while using a filter only needs
one.  Also, this will increase code size.  Note that all of
this only matters for C++ which I think is the only user of
filters right now.

On the other hand, we could try to mitigate these problems
at codegen time by identifying code sequences which could
be realized by a filter entry in the dwarf tables, i.e. by
auto-generating filters from selectors.  I'm afraid that this
seems much harder than PR2154, probably requiring heroic
optimization.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 16 15:48:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Mar 2008 15:48:53 -0500
Subject: [LLVMbugs] [Bug 2158] New: clang has no doxygen
Message-ID: <bug-2158-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2158

           Summary: clang has no doxygen
           Product: Documentation
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Doxygen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Please provide doxygen auto-documentation for clang on the website.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 16 15:52:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Mar 2008 15:52:39 -0500
Subject: [LLVMbugs] [Bug 2159] New: don't break doxygen every night
Message-ID: <bug-2159-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2159

           Summary: don't break doxygen every night
           Product: Documentation
           Version: trunk
          Platform: PC
               URL: http://llvm.org/doxygen
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Doxygen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Doxygen runs nightly (midnight in PST, if I recall). What happens is that the
directory that's live serving from the website gets wiped out and doxygen
starts to refill it.

This causes us to serve 404s to our users who are looking for the
documentation. That in turn leads to things like this:
http://pablotron.org/?cid=1539 . This will only become a larger problem as we
have non-US users where the midnight in Pacific time is the middle of the day
for them.

Please make doxygen fill a new directory and then pivot that into place,
minimizing the daily downtime.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 16 21:40:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Mar 2008 21:40:32 -0500
Subject: [LLVMbugs] [Bug 816] [ppc] _Complex double ABI mismatch
In-Reply-To: <bug-816-206@http.llvm.org/bugs/>
Message-ID: <200803170240.m2H2eW3D002255@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=816


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #1 from Dale Johannesen <dalej at apple.com>  2008-03-16 21:40:31 ---
Fixed here.  (Also _Complex long long and _Complex long double which had
similar problems.)
http://llvm.org/viewvc/llvm-project?view=rev&revision=48439
http://llvm.org/viewvc/llvm-project?view=rev&revision=48440


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 17 05:08:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Mar 2008 05:08:12 -0500
Subject: [LLVMbugs] [Bug 2160] New: anders-aa wrong result on external
	function call
Message-ID: <bug-2160-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2160

           Summary: anders-aa wrong result on external function call
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


On this testcase "opt -anders-aa -gvn" results in the function
returning undef, because Andersen's says that the call to @f
cannot modify the memory pointed to by %tmp.

declare void @f(i32*)

define i32 @g() {
entry:
        %tmp = alloca i32               ; <i32*> [#uses=2]
        call void @f( i32* %tmp )
        %tmp2 = load i32* %tmp          ; <i32> [#uses=1]
        ret i32 %tmp2
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 17 11:51:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Mar 2008 11:51:06 -0500
Subject: [LLVMbugs] [Bug 2161] New: code fragment from gcc 4. 2 sources
 crashes clang -fsyntax-only
Message-ID: <bug-2161-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2161

           Summary: code fragment from gcc 4.2 sources crashes clang -
                    fsyntax-only
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kremenek at cs.stanford.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1548)
 --> (http://llvm.org/bugs/attachment.cgi?id=1548)
Test cases that crashes clang -fsyntax-only

The following code fragment crashes clang -fsyntax-only (also attached):

     void decimal_real_maxval (struct real_value *r, int sign, enum
machine_mode mode) {
               switch (mode)     { 

I see the following in gdb:

(gdb) r
Starting program: /Volumes/MyBookScratch/Debug_clang_2/clang -fsyntax-only
t.c.ok
Reading symbols for shared libraries +++. done
t.c.ok:3:2: error: expected '}'

 ^

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_PROTECTION_FAILURE at address: 0x00000004
0x00118801 in clang::Type::getCanonicalTypeInternal (this=0x0) at Type.h:360
360       QualType getCanonicalTypeInternal() const { return CanonicalType; }
(gdb) bt
#0  0x00118801 in clang::Type::getCanonicalTypeInternal (this=0x0) at
Type.h:360
#1  0x0011883e in clang::QualType::getCanonicalType (this=0xbfffe108) at
Type.h:1102
#2  0x000e7a01 in clang::ASTContext::getTypeInfo (this=0xbfffee30, T={ThePtr =
0}) at ASTContext.cpp:184
#3  0x000e81f0 in clang::ASTContext::getTypeInfo (this=0xbfffee30, T={ThePtr =
9462992}) at ASTContext.cpp:302
#4  0x00057ad3 in clang::ASTContext::getTypeSize (this=0xbfffee30, T={ThePtr =
9462992}) at ASTContext.h:272
#5  0x000d9066 in clang::Sema::ActOnFinishSwitchStmt (this=0x902620,
SwitchLoc={ID = 131182}, Switch=0x9066d0, Body=0x901410) at SemaStmt.cpp:310
#6  0x00141ba6 in clang::Parser::ParseSwitchStatement (this=0xbfffefe0) at
ParseStmt.cpp:622
#7  0x001407f1 in clang::Parser::ParseStatementOrDeclaration (this=0xbfffefe0,
OnlyStatement=false) at ParseStmt.cpp:131
#8  0x0013fdb4 in clang::Parser::ParseCompoundStatementBody (this=0xbfffefe0,
isStmtExpr=false) at ParseStmt.cpp:425
#9  0x0014019e in clang::Parser::ParseFunctionStatementBody (this=0xbfffefe0,
Decl=0x906510, L={ID = 131159}, R={ID = 131159}) at ParseStmt.cpp:1149
#10 0x00144166 in clang::Parser::ParseFunctionDefinition (this=0xbfffefe0,
D=@0xbfffec20) at Parser.cpp:490
#11 0x001446a9 in clang::Parser::ParseDeclarationOrFunctionDefinition
(this=0xbfffefe0) at Parser.cpp:436
#12 0x00144803 in clang::Parser::ParseExternalDeclaration (this=0xbfffefe0) at
Parser.cpp:350
#13 0x00144a21 in clang::Parser::ParseTopLevelDecl (this=0xbfffefe0,
Result=@0xbffff08c) at Parser.cpp:267
#14 0x000b39ab in clang::ParseAST (PP=@0xbffff184, Consumer=0x901220,
PrintStats=false) at ParseAST.cpp:48
#15 0x00024148 in ProcessInputFile (PP=@0xbffff184, InFile=@0x901450,
OurDiagnosticClient=@0x901470) at clang.cpp:1111
#16 0x00026858 in main (argc=3, argv=0xbffff488) at clang.cpp:1305
(gdb)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 17 18:42:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Mar 2008 18:42:39 -0500
Subject: [LLVMbugs] [Bug 2067] Loop unroller heuristics are over-aggressive
	with calls
In-Reply-To: <bug-2067-206@http.llvm.org/bugs/>
Message-ID: <200803172342.m2HNgde8015025@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2067


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #3 from Devang Patel <dpatel at apple.com>  2008-03-17 18:42:39 ---
Done.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/059811.html

I ran MultiSource and SingleSource tests before and after this change. Cursory
glance did not reveal any significant performance drop. We'll see if nightly
tester reports anything tomorrow.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 17 19:40:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Mar 2008 19:40:13 -0500
Subject: [LLVMbugs] [Bug 1441] Pass Manager reruns analysis pass if
	requested explicitly
In-Reply-To: <bug-1441-206@http.llvm.org/bugs/>
Message-ID: <200803180040.m2I0eDE3016507@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1441


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #3 from Devang Patel <dpatel at apple.com>  2008-03-17 19:40:12 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/059813.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 Mar 18 00:49:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Mar 2008 00:49:20 -0500
Subject: [LLVMbugs] [Bug 2162] New: Migrate away from Value::getName()
Message-ID: <bug-2162-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2162

           Summary: Migrate away from Value::getName()
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: slow-compile
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sabre at nondot.org
                CC: llvmbugs at cs.uiuc.edu


Internally value names are stored in a StringMap, which is an efficient data
structure for finding and uniquing and autorenaming names that are in the
symbol table.  This is goodness and it's FAR more efficient than something like
std::map<std::string, Value*> which is what we used to use.

The disadvantage of this is that the map doesn't hold an std::string, which
means that (in particular) Value::getName() is extremely expensive.  It is
expensive because it *copies* the name of the value into an std::string to
return it.  This means that all clients of it that matter should move off of it
to reduce our compile times.  This will significantly reduce heap thrashing.

There are many ways to improve the situation.  For example:

1) This like that asmwriter that need to print out the string can move to using
Value::getNameStart()+Value::getNameLen(), which are both very efficient.

2) The nasty std:string Tmp = V1->getName(); V1->setName(""); V2->setName(Tmp)
idiom can be replaced with the *far* more efficient and nicer V2->takeName(V1)
call.

3) I common pattern is copy the name of some value with a suffix.  We should
add a helper to value for this, something like V1->copySuffixedName(V2, "i") or
something, which adds ".i".

4) It is also very common to use a numeric suffix, with this pattern:
V2->setName(V1->getName()+"."+utostr(i))   This is expensive for many reasons:
utostr allocates a string, the concatenations do as well, etc.  This is
generally horrible for memory use and performance.  It would be better to add a
V1->copySuffixedName(V2, i) to handle this, and implement it efficiently.

5) setName(std::string) is also fairly expensive if the input is not a string. 
Clients should start using the setName version that takes a const char* +
length for full generality, or just an const char* if they don't care about
nul's in the string.

6) Any uses of getName() that occur in debug code or in other places that we
don't care about should migrate over to using Value::getNameStr().  This will
allow us to identify places that we haven't look at yet.

When the full audit + migration is done, we can remove Value::getName()
entirely.

-Chris


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 18 17:03:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Mar 2008 17:03:39 -0500
Subject: [LLVMbugs] [Bug 2163] New: Codegen illegally reorders division by
 zero before function call
Message-ID: <bug-2163-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2163

           Summary: Codegen illegally reorders division by zero before
                    function call
           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:

#include <stdlib.h>
int x() {exit(0);}
int main(void) {x(); return 1/0;}

Running this through llvm-gcc -O0, generates a program that crashes.  This
program should not crash.  (The semantics of this program are well-defined
because the division by zero should not end up being executed.)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 18 17:23:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Mar 2008 17:23:17 -0500
Subject: [LLVMbugs] [Bug 2160] anders-aa wrong result on external function
	call
In-Reply-To: <bug-2160-206@http.llvm.org/bugs/>
Message-ID: <200803182223.m2IMNHiU031741@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2160


Daniel Berlin <dberlin at dberlin.org> changed:

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




--- Comment #1 from Daniel Berlin <dberlin at dberlin.org>  2008-03-18 17:23:17 ---
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  Wed Mar 19 16:19:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 16:19:15 -0500
Subject: [LLVMbugs] [Bug 2164] New: Codegen doesn't optimize 16-bit divide
	into multiply
Message-ID: <bug-2164-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2164

           Summary: Codegen doesn't optimize 16-bit divide into multiply
           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


Created an attachment (id=1549)
 --> (http://llvm.org/bugs/attachment.cgi?id=1549)
Patch to apintify magicu

Simple testcase:
unsigned t(unsigned short x) {return x/255;}

llvm-gcc currently produces:
t:
        movw    $255, %cx
        xorw    %dx, %dx
        movw    4(%esp), %ax
        divw    %cx
        movzwl  %ax, %eax
        ret

vs. gcc:

t:
        movzwl  4(%esp), %eax
        imull   $32897, %eax, %eax
        shrl    $23, %eax
        ret

Attached is a patch that improves this case to the following:
t:
        movw    $32897, %ax
        mulw    4(%esp)
        shrw    $7, %dx
        movzwl  %dx, %eax
        ret

This isn't perfect, but the rest probably belongs in lib/Target/README.txt.
(I'm not really how to deal with this.)

I haven't converted the signed division case 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  Wed Mar 19 18:11:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 18:11:47 -0500
Subject: [LLVMbugs] [Bug 2161] code fragment from gcc 4.2 sources crashes
 clang -fsyntax-only
In-Reply-To: <bug-2161-206@http.llvm.org/bugs/>
Message-ID: <200803192311.m2JNBlUK022465@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2161


snaroff at apple.com <snaroff at apple.com> changed:

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




--- Comment #1 from snaroff at apple.com  2008-03-19 18:11:46 ---
Sema::ActOnStartOfFunctionDef() needed to diagnose parmdecls with incomplete
types (C99 6.7.5.3p4).

Fixed by http://llvm.org/viewvc/llvm-project?view=rev&revision=48568


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 19 18:23:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 18:23:21 -0500
Subject: [LLVMbugs] [Bug 2091] referenced structure should? be implicitly
	defined
In-Reply-To: <bug-2091-206@http.llvm.org/bugs/>
Message-ID: <200803192323.m2JNNLjH022785@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2091


snaroff at apple.com <snaroff at apple.com> changed:

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




--- Comment #3 from snaroff at apple.com  2008-03-19 18:23:21 ---
I don't understand. Since clang doesn't support C++ yet, why is this a bug?

For example, the preprocessed file you forwarded has declarations with "throw"
specifiers...

extern int _IO_feof (_IO_FILE *__fp) throw ();

Confused.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 19 18:48:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 18:48:34 -0500
Subject: [LLVMbugs] [Bug 2165] New: Constant folding should fold "<1 x i64>
 bitcast (i64 63 to <1 x i64>)"
Message-ID: <bug-2165-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2165

           Summary: Constant folding should fold "<1 x i64> bitcast (i64 63
                    to <1 x i64>)"
           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


Created an attachment (id=1550)
 --> (http://llvm.org/bugs/attachment.cgi?id=1550)
Patch

Issue per summary: we should fold "<1 x i64> bitcast (i64 63 to <1 x i64>)" to
"<1 x i64> <i64 63>".  This is more consistent, and helps codegen generate more
efficient code on x86.

Patch attached; I'm not sure if this is the right approach (actually, I'm
pretty sure it's wrong in the sense that it doesn't catch all the cases it
should... I'll have to look at it a bit more).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 19 20:10:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 20:10:21 -0500
Subject: [LLVMbugs] [Bug 2034] module-level analysis invalidated but not
	re-run
In-Reply-To: <bug-2034-206@http.llvm.org/bugs/>
Message-ID: <200803200110.m2K1ALgB026672@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2034


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #2 from Devang Patel <dpatel at apple.com>  2008-03-19 20:10:20 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/059944.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 Mar 19 23:21:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 23:21:26 -0500
Subject: [LLVMbugs] [Bug 2166] New: Compiling unoptimised 445.gobmk (from
 spec2006) crashes LLC
Message-ID: <bug-2166-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2166

           Summary: Compiling unoptimised 445.gobmk (from spec2006) crashes
                    LLC
           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: lhames at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1552)
 --> (http://llvm.org/bugs/attachment.cgi?id=1552)
bugpoint simplified bc file which produces crash

Using LLVM 2.2 (debug build) on an x86 machine running Debian etch, and using
llvm-gcc 4.0:

Compiled all 445.gobmk sources with

llvm-gcc -O0 -c -emit-llvm -DSPEC_CPU -DSPEC_CPU2006 -DHAVE_CONFIG_H
-D_GNU_SOURCE -D__STDC_LIMIT_MACROS -DNDEBUG -D__NO_MATH_INLINES <foo>.c -o
<foo>.bc

The resulting bytecodes were linked "llvm-ld -disable-opt -link-as-library -o
445.gobmk.bc *.bc"

Running LLC on the resulting bytecode file causes an assertion failure:

llc: SelectionDAG.cpp:476: void llvm::SelectionDAG::DeleteNode(llvm::SDNode*):
Assertion `N->use_empty() && "Cannot delete a node that is not dead!"' failed.
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc((anonymous
namespace)::PrintStackTrace()+0x22)[0x8a2d4a4]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc((anonymous
namespace)::SignalHandler(int)+0x110)[0x8a2d768]
[0x4001d420]
/lib/libc.so.6(abort+0x101)[0x401995b1]
/lib/libc.so.6(__assert_fail+0xfd)[0x4019146d]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::SelectionDAG::DeleteNode(llvm::SDNode*)+0x61)[0x8746b57]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x870b552]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x871ba6a]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x871bbf2]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x871cfa7]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::SelectionDAG::Combine(bool,
llvm::AliasAnalysis&)+0xad)[0x871d3af]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::SelectionDAGISel::CodeGenAndEmitDAG(llvm::SelectionDAG&)+0x7b)[0x8752ddb]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::SelectionDAGISel::SelectBasicBlock(llvm::BasicBlock*,
llvm::MachineFunction&, llvm::FunctionLoweringInfo&)+0xbe)[0x8768408]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::SelectionDAGISel::runOnFunction(llvm::Function&)+0x263)[0x876a47b]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x8630d1b]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::FPPassManager::runOnFunction(llvm::Function&)+0x122)[0x89bd4ca]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::FunctionPassManagerImpl::run(llvm::Function&)+0x6e)[0x89bd732]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(llvm::FunctionPassManager::run(llvm::Function&)+0x88)[0x89bd880]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc(main+0x998)[0x8411da4]
/lib/libc.so.6(__libc_start_main+0xdc)[0x40183ebc]
/home/lhames/devel/llvm/plimits/trunk/install/llvm/bin/llc[0x8410591]


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 19 23:44:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Mar 2008 23:44:17 -0500
Subject: [LLVMbugs] [Bug 2166] Compiling unoptimised 445.gobmk (from
 spec2006) crashes LLC
In-Reply-To: <bug-2166-206@http.llvm.org/bugs/>
Message-ID: <200803200444.m2K4iH6M032661@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2166


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #2 from Chris Lattner <sabre at nondot.org>  2008-03-19 23:44:16 ---
Ok, this has probably been fixed since llvm 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  Thu Mar 20 04:47:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Mar 2008 04:47:12 -0500
Subject: [LLVMbugs] [Bug 2167] New: llvm-gcc generates debug info for
 compiler-generated temporaries
Message-ID: <bug-2167-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2167

           Summary: llvm-gcc generates debug info for compiler-generated
                    temporaries
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1553)
 --> (http://llvm.org/bugs/attachment.cgi?id=1553)
source file

With the attached source:

$ ~/llvm-gcc4.2-2.2-x86-linux-RHEL4/bin/llvm-gcc -g -o main main.c
$ gdb main GNU gdb 6.4.90-debian
[...]
(gdb) b main
Breakpoint 1 at 0x80483b5: file /home/foad//main.c, line 12.
(gdb) run
[...]
Breakpoint 1, main (argc=1, argv=0xbfb86c74) at /home/foad//main.c:12
12        for (i = 0; i < 256 ; i++)
(gdb) info locals
k = 0
j = -1209057292
i = -1210246004
val = "..."
buf = "..."
iftmp.5 = 0
(gdb) 

It seems wrong that "iftmp.5" is listed as a local variable here, because there
is no such variable in my program. Presumably it's some temporary value
generated by the compiler.


My suggested fix is not to emit any debug info for "artificial" variables:

Index: llvm-debug.cpp
===================================================================
--- llvm-debug.cpp      (revision 48589)
+++ llvm-debug.cpp      (working copy)
@@ -315,6 +315,10 @@
 /// region - "llvm.dbg.declare."
 void DebugInfo::EmitDeclare(tree decl, unsigned Tag, const char *Name,
                             tree type, Value *AI, BasicBlock *CurBB) {
+  // Ignore compiler temporaries.
+  if (DECL_ARTIFICIAL(decl))
+    return;
+
   // Lazily construct llvm.dbg.declare function.
   const PointerType *EmpPtr = SR.getEmptyStructPtrType();
   if (!DeclareFn)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 20 09:24:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Mar 2008 09:24:11 -0500
Subject: [LLVMbugs] [Bug 2168] New: spurious '!' in gcc/Makefile.in
Message-ID: <bug-2168-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2168

           Summary: spurious '!' in gcc/Makefile.in
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


In the GCC 4.2 gcc/Makefile.in, there's a spurious '!' character at the start
of this line:

stage1_build:
        if test "${SET_DYLIB_PATH}" = "1"; then \
          DYLD_LIBRARY_PATH="`${PWD_COMMAND}`"; \
        fi; \
        $(MAKE) CC="$(CC)" libdir=$(libdir) LANGUAGES="$(BOOT_LANGUAGES)" \
                CFLAGS="$(STAGE1_CFLAGS) $(STAGE1_CHECKING_CFLAGS)" \
                MAKEINFO="$(MAKEINFO)" MAKEINFOFLAGS="$(MAKEINFOFLAGS)" \
                COVERAGE_FLAGS= OBJS-onestep="$(OBJS)" \
!               LLVM_STAGE="-stage1"
        $(STAMP) stage1_build
        echo stage1_build > stage_last

This thwarted my attempt to do a "make bootstrap" from the LLVM 2.2 GCC 4.2
source release.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 20 16:52:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Mar 2008 16:52:30 -0500
Subject: [LLVMbugs] [Bug 2168] spurious '!' in gcc/Makefile.in
In-Reply-To: <bug-2168-206@http.llvm.org/bugs/>
Message-ID: <200803202152.m2KLqUxU008638@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2168


Bill Wendling <isanbard at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |isanbard at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bill Wendling <isanbard at gmail.com>  2008-03-20 16:52:30 ---
Fixed:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/059995.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 Mar 21 04:36:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Mar 2008 04:36:01 -0500
Subject: [LLVMbugs] [Bug 2167] llvm-gcc generates debug info for
 compiler-generated temporaries
In-Reply-To: <bug-2167-206@http.llvm.org/bugs/>
Message-ID: <200803210936.m2L9a1cO004956@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2167


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #10 from Duncan Sands <baldrick at free.fr>  2008-03-21 04:36:01 ---
Fixed, patch here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/060015.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 Mar 21 15:43:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Mar 2008 15:43:32 -0500
Subject: [LLVMbugs] [Bug 2169] New: anders-aa wrong result with memcpy
Message-ID: <bug-2169-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2169

           Summary: anders-aa wrong result with memcpy
           Product: new-bugs
           Version: unspecified
          Platform: Other
        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


For the following testcase,
  opt -anders-aa -gvn
results in %tmp being replaced with undef:

declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind
declare void @use(i8)

define void @f(i8* %x) {
entry:
        %copy = alloca i8               ; <i8*> [#uses=6]
        call void @llvm.memcpy.i32( i8* %copy, i8* %x, i32 1, i32 4 )
        %tmp = load i8* %copy           ; <i8> [#uses=1]
        call void @use(i8 %tmp)
        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  Sat Mar 22 03:19:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Mar 2008 03:19:01 -0500
Subject: [LLVMbugs] [Bug 2145] llvm-gcc emits invalid aliases
In-Reply-To: <bug-2145-206@http.llvm.org/bugs/>
Message-ID: <200803220819.m2M8J1i6020278@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2145


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

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




--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru>  2008-03-22 03:19:01 ---
Fixed in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/060064.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 Mar 22 03:38:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Mar 2008 03:38:44 -0500
Subject: [LLVMbugs] [Bug 2144] Verifier allows invalid aliases
In-Reply-To: <bug-2144-206@http.llvm.org/bugs/>
Message-ID: <200803220838.m2M8ci5W020840@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2144


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

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




--- Comment #5 from Anton Korobeynikov <asl at math.spbu.ru>  2008-03-22 03:38:44 ---
Documentation and Verifier was updated in:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/060065.html

Testcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080317/060066.html

Please reopen, if something 'bad' with chained aliases will be found.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 23 20:08:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Mar 2008 20:08:11 -0500
Subject: [LLVMbugs] [Bug 2170] New: Please add a link for Passes. html to
 the main documentation index
Message-ID: <bug-2170-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2170

           Summary: Please add a link for Passes.html to the main
                    documentation index
           Product: Website
           Version: unspecified
          Platform: All
               URL: http://llvm.cs.uiuc.edu/docs/
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: Documentation
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: viridia at gmail.com
                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 Mar 24 12:40:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Mar 2008 12:40:46 -0500
Subject: [LLVMbugs] [Bug 2171] New: using iplist<T>::iterator operator[]
 should cause a compile error
Message-ID: <bug-2171-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2171

           Summary: using iplist<T>::iterator operator[] should cause a
                    compile error
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


On IRC we saw this problem:

  Function::arg_iterator AI = Method->arg_begin();
  AI[0].setName("foo");
  AI[1].setName("foo");

which causes a crash at run-time because AI[1] is invalid. There's no need to
support this kind of iteration, as the STL list::iterator doesn't either. We
should make sure that this is a compile error. Perhaps by defining a private
operator[]?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 24 13:31:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Mar 2008 13:31:28 -0500
Subject: [LLVMbugs] [Bug 2170] Please add a link for Passes. html to the
 main documentation index
In-Reply-To: <bug-2170-206@http.llvm.org/bugs/>
Message-ID: <200803241831.m2OIVS1Q005482@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2170


Tanya Lattner <tonic at nondot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tonic at nondot.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Tanya Lattner <tonic at nondot.org>  2008-03-24 13:31:27 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060133.html

It won't update on llvm.org til tonight.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 24 13:40:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Mar 2008 13:40:43 -0500
Subject: [LLVMbugs] [Bug 1462] CodeGen doesn't fully support i128
In-Reply-To: <bug-1462-206@http.llvm.org/bugs/>
Message-ID: <200803241840.m2OIehpr006245@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1462


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #10 from Dan Gohman <gohman at apple.com>  2008-03-24 13:40:43 ---
This is now implemented, for 64-bit targets, here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060117.html

The workaround from #6 is reverted, and all the testcases in
this PR are now handled without 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  Mon Mar 24 17:14:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Mar 2008 17:14:20 -0500
Subject: [LLVMbugs] [Bug 2172] New: Dwarf emitter assumes all functions have
	debug info
Message-ID: <bug-2172-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2172

           Summary: Dwarf emitter assumes all functions have debug info
           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: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1554)
 --> (http://llvm.org/bugs/attachment.cgi?id=1554)
file with some debug info

compiling this with -g:
typedef int (*ft)(int);

extern int bar(int);

ft foo(int x) {
 return bar;
}


and this without
int bar(int x ) {
  return x;
}


Then linking (with -disable opt) yields the attached bytecode file, which
crashes llc.  The crash happens when llc tries to emit dwarf info for bar,
which doesn't exist.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 25 04:13:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 04:13:45 -0500
Subject: [LLVMbugs] [Bug 2062] build: cannot build llvm from source on
 read-only file system
In-Reply-To: <bug-2062-206@http.llvm.org/bugs/>
Message-ID: <200803250913.m2P9DjmV008406@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2062


Bill Wendling <isanbard at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |isanbard at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Bill Wendling <isanbard at gmail.com>  2008-03-25 04:13:44 ---
Fixed with: Committed revision 48766.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 25 09:31:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 09:31:56 -0500
Subject: [LLVMbugs] [Bug 2173] New: llvm::cin operator >> broken
Message-ID: <bug-2173-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2173

           Summary: llvm::cin operator >> broken
           Product: libraries
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Support Libraries
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dekruijf at cs.wisc.edu
                CC: llvmbugs at cs.uiuc.edu


The >> operator for the class BaseStream in include/llvm/Support/Streams.h does
not work because the argument type is declared as 'const'.  It must *not* be
'const', because the >> operator modifies its argument.  With the patch below,
it now works.

Index: include/llvm/Support/Streams.h
===================================================================
--- include/llvm/Support/Streams.h      (revision 1)
+++ include/llvm/Support/Streams.h      (working copy)
@@ -46,7 +46,7 @@
     }

     template <typename Ty>
-    BaseStream &operator >> (const Ty &Thing) {
+    BaseStream &operator >> (Ty &Thing) {
       if (Stream) *Stream >> Thing;
       return *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 Mar 25 13:17:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 13:17:42 -0500
Subject: [LLVMbugs] [Bug 2173] llvm::cin operator >> broken
In-Reply-To: <bug-2173-206@http.llvm.org/bugs/>
Message-ID: <200803251817.m2PIHgUF026064@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2173


Bill Wendling <isanbard at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |isanbard at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bill Wendling <isanbard at gmail.com>  2008-03-25 13:17:42 ---
Applied:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060196.html

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  Tue Mar 25 14:53:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 14:53:15 -0500
Subject: [LLVMbugs] [Bug 2174] New: llvm generated code ~20% slower than gcc
	for sqlite3
Message-ID: <bug-2174-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2174

           Summary: llvm generated code ~20% slower than gcc for sqlite3
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          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=1555)
 --> (http://llvm.org/bugs/attachment.cgi?id=1555)
bzip2 packed sqlite.c

I adapted sqlite3's build system to llvm-test.

When running the test I get results like this:
Program | GCCAS  Bytecode LLC compile LLC-BETA compile JIT codegen | GCC    
CBE     LLC     LLC-BETA JIT     | GCC/CBE GCC/LLC GCC/LLC-BETA LLC/LLC-BETA
sqlite3 | 4.8396 1744588  5.3496      *                3.6930      |    7.08   
8.29    8.68 *          13.63 | 0.85    0.82    n/a          n/a

To reproduce you can either take the package from here and run 'make
TEST=nightly report':
http://edwintorok.googlepages.com/sqlite_lemon_llvmtest.tar.gz

Or follow these steps (I am attaching all needed files):
$ llvm-gcc -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1
-DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
-DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=0
-DHAVE_FDATASYNC=0 -DHAVE_USLEEP=0 -DHAVE_LOCALTIME_R=0 -DHAVE_GMTIME_R=0
-DHAVE_READLINE=0 -I. -DSQLITE_THREADSAFE=0 -DSQLITE_OMIT_LOAD_EXTENSION=1
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/MultiSource/Applications/sqlite3
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/MultiSource/Applications/sqlite3
-I/home/edwin/llvm-svn/llvm/include
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/include -I../../..//include
-I/home/edwin/llvm-svn/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS
-DNDEBUG  -O3 sqlite.c  -emit-llvm -c -o sqlite3.bc && opt -std-compile-opts -f
sqlite3.bc -o sqlite3_opt.bc && llvm-ld -O2 -native sqlite3_opt.bc -o
sqlite3_llvm
$ gcc-4.3 -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1
-DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
-DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=0
-DHAVE_FDATASYNC=0 -DHAVE_USLEEP=0 -DHAVE_LOCALTIME_R=0 -DHAVE_GMTIME_R=0
-DHAVE_READLINE=0 -I. -DSQLITE_THREADSAFE=0 -DSQLITE_OMIT_LOAD_EXTENSION=1
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/MultiSource/Applications/sqlite3
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/MultiSource/Applications/sqlite3
-I/home/edwin/llvm-svn/llvm/include
-I/home/edwin/llvm-svn/llvm/projects/llvm-test/include -I../../..//include
-I/home/edwin/llvm-svn/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS
-DNDEBUG  -O3 sqlite.c -o sqlite_native

$ time ./sqlite_native :memory: <tst.sql >/dev/null
real    0m1.589s
user    0m1.493s
sys     0m0.010s

$ time ./sqlite_llvm :memory: <tst.sql >/dev/null
real    0m1.970s
user    0m1.890s
sys     0m0.013s


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 25 16:27:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 16:27:49 -0500
Subject: [LLVMbugs] [Bug 1576] Need a mechanism to tell codegen to optimize
	for size
In-Reply-To: <bug-1576-206@http.llvm.org/bugs/>
Message-ID: <200803252127.m2PLRnEX032228@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1576


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #3 from Devang Patel <dpatel at apple.com>  2008-03-25 16:27:49 ---
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060209.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 Mar 25 18:36:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 18:36:46 -0500
Subject: [LLVMbugs] [Bug 2175] New: llvm-gcc abort on tricky VLA cast
Message-ID: <bug-2175-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2175

           Summary: llvm-gcc abort on tricky VLA cast
           Product: tools
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc aborts on this:

void *p(void);
long d(void);

float foo(long i0, long i1, long i2, long i3, long i4, long i5) {
  return ((float (*(*(*)[d()])[])[])p())[i0][i1][i2][i3][i4][i5];
}

The abort looks like this:

Unhandled expression!
TREE_CODE: 117
 <save_expr 0x41d14990
    type <integer_type 0x41d0c3f0 long int public SI
        size <integer_cst 0x41d046c0 constant invariant 32>
        unit size <integer_cst 0x41d041e0 constant invariant 4>
        align 32 symtab 2 alias set -1 precision 32 min <integer_cst 0x41d047e0
-2147483648> max <integer_cst 0x41d04810 2147483647>
        LLVM:  i32
        pointer_to_this <pointer_type 0x41d16b60>>
    side-effects invariant
    arg 0 <call_expr 0x41d14900 type <integer_type 0x41d0c3f0 long int>
        side-effects
        arg 0 <addr_expr 0x41d148d0 type <pointer_type 0x41dbacb0>
            constant invariant arg 0 <function_decl 0x41dbb300 d>>>>
t.c: In function ???foo???:
t.c:4: internal compiler error: in Emit, at llvm-convert.cpp:851


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 25 23:17:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Mar 2008 23:17:44 -0500
Subject: [LLVMbugs] [Bug 2176] New: llvmc2 fails to build (source code error)
Message-ID: <bug-2176-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2176

           Summary: llvmc2 fails to build (source code error)
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: llvmc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: aatyrael at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Something is propably wrong in the autogenerated code that is included in
Tools.cpp

Mingw 4.3/MSYS enviroment
WinXP
SVN trunk head Revision of 26/3/2008
Error Message:

make[2]: Entering directory `/c/llvm/tools/llvmc2'
llvm[2]: Building LLVMCC configuration library with tblgen
llvm[2]: Compiling Core.cpp for Release build 
llvm[2]: Compiling Tools.cpp for Release build 
In file included from Tools.cpp:28:
Tools.inc: In member function 'virtual bool llvm_gcc_assembler::IsLast()
const':
Tools.inc:167: error: could not convert 'AutoGeneratedPrefixListWa' to 'bool'
make[2]: *** [/c/llvm/tools/llvmc2/Release/Tools.o] Error 1
make[2]: Leaving directory `/c/llvm/tools/llvmc2'
make[1]: *** [llvmc2/.makeall] Error 2
make[1]: Leaving directory `/c/llvm/tools'
make: *** [all] 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  Fri Mar 28 18:21:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Mar 2008 18:21:45 -0500
Subject: [LLVMbugs] [Bug 2177] New: "Invalid token factor!" SelectionDAG.cpp,
	line 1953
Message-ID: <bug-2177-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2177

           Summary: "Invalid token factor!" SelectionDAG.cpp, line 1953
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dpatel at apple.com
                CC: llvmbugs at cs.uiuc.edu


Attached byte code causes llc assert.

ssertion failed: (VT == MVT::Other && N1.getValueType() == MVT::Other &&
N2.getValueType() == MVT::Other && "Invalid token factor!"), function getNode,
file
/Volumes/Sandbox/llvm/llvmCore-2022.roots/llvmCore-2022~obj/src/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp,
line 1953.
0   llc                                 0x0058012e
_ZNSt8_Rb_treeIN4llvm3sys4PathES2_St9_IdentityIS2_ESt4lessIS2_ESaIS2_EE8_M_eraseEPSt13_Rb_tree_nodeIS2_E
+ 9038
1   libSystem.B.dylib                   0x930bd97b _sigtramp + 43
2   ???                                 0xffffffff 0x0 + 4294967295
3   libSystem.B.dylib                   0x93136782 raise + 26
4   libSystem.B.dylib                   0x93145d3f abort + 73
5   libSystem.B.dylib                   0x93137923 __assert_rtn + 101
6   llc                                 0x002c6f25
_ZN4llvm8DenseMapIPNS_6SDNodeESt6vectorIPNS_5SUnitESaIS5_EENS_12DenseMapInfoIS2_EENS8_IS7_EEE16FindAndConstructERKS2_
+ 126485
7   llc                                 0x002c33bd
_ZN4llvm8DenseMapIPNS_6SDNodeESt6vectorIPNS_5SUnitESaIS5_EENS_12DenseMapInfoIS2_EENS8_IS7_EEE16FindAndConstructERKS2_
+ 111277
8   llc                                 0x0012a01e
_ZN4llvm2cl4listISsbNS0_6parserISsEEED0Ev + 1198286
9   llc                                 0x00143bf6
_ZN4llvm2cl4listISsbNS0_6parserISsEEED0Ev + 1303718
10  llc                                 0x002412a1
_ZN4llvm15BitstreamReader18ReadBlockInfoBlockEv + 160833
11  llc                                 0x00242805
_ZN4llvm15BitstreamReader18ReadBlockInfoBlockEv + 166309
12  llc                                 0x00242f9b
_ZN4llvm15BitstreamReader18ReadBlockInfoBlockEv + 168251
13  llc                                 0x002d3df7 _ZN4llvm11StoreSDNodeD1Ev +
1495
14  llc                                 0x003083b7 _ZN4llvm11StoreSDNodeD1Ev +
215959
15  llc                                 0x0030ad6e _ZN4llvm11StoreSDNodeD1Ev +
226638
16  llc                                 0x000153c2
_ZN4llvm2cl4listISsbNS0_6parserISsEEED0Ev + 64626
17  llc                                 0x005145d5
_ZNSt8_Rb_treeIlSt4pairIKlPN4llvm8PassInfoEESt10_Select1stIS5_ESt4lessIlESaIS5_EE13insert_uniqueERKS5_
+ 27365
18  llc                                 0x00514b07
_ZNSt8_Rb_treeIlSt4pairIKlPN4llvm8PassInfoEESt10_Select1stIS5_ESt4lessIlESaIS5_EE13insert_uniqueERKS5_
+ 28695
19  llc                                 0x00514c88
_ZNSt8_Rb_treeIlSt4pairIKlPN4llvm8PassInfoEESt10_Select1stIS5_ESt4lessIlESaIS5_EE13insert_uniqueERKS5_
+ 29080
20  llc                                 0x00003046 _mh_execute_header + 8262
21  llc                                 0x00001836 _mh_execute_header + 2102
22  ???                                 0x00000002 0x0 + 2
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 Mar 28 18:46:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Mar 2008 18:46:13 -0500
Subject: [LLVMbugs] [Bug 2177] "Invalid token factor!" SelectionDAG.cpp,
	line 1953
In-Reply-To: <bug-2177-206@http.llvm.org/bugs/>
Message-ID: <200803282346.m2SNkDkQ026703@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2177


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-03-28 18:46:13 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060382.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 Mar 29 00:37:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Mar 2008 00:37:57 -0500
Subject: [LLVMbugs] [Bug 2178] New: byval ignored by llvm2cpp.
Message-ID: <bug-2178-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2178

           Summary: byval ignored by llvm2cpp.
           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: ofv at wanadoo.es
                CC: llvmbugs at cs.uiuc.edu


On the demo cgi:

http://www.llvm.org/demo/index.cgi

enter this code:

struct S {
double d;
};

S foo(S);

double bar() {
  S s;
  s = foo(s);
  return s.d;
}

Check C++ as source language and check "Show LLVM C++ API code".

The C++ API code shown does not show any use of ParamAttr::ByVal, despite `foo'
being declared as

declare void @_Z3foo1S(%struct.S* sret , %struct.S* byval)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 29 06:26:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Mar 2008 06:26:42 -0500
Subject: [LLVMbugs] [Bug 2178] byval ignored by llvm2cpp.
In-Reply-To: <bug-2178-206@http.llvm.org/bugs/>
Message-ID: <200803291126.m2TBQgrc022815@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2178


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>  2008-03-29 06:26:41 ---
This was fixed in TOT:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060392.html

updating llvm-gcc / llvm2cpp on website is another 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  Sat Mar 29 11:16:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Mar 2008 11:16:22 -0500
Subject: [LLVMbugs] [Bug 2176] llvmc2 fails to build (source code error)
In-Reply-To: <bug-2176-206@http.llvm.org/bugs/>
Message-ID: <200803291616.m2TGGMFT030775@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2176


Mikhail Glushenkov <foldr at codedgers.com> changed:

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




--- Comment #3 from Mikhail Glushenkov <foldr at codedgers.com>  2008-03-29 11:16:20 ---
> Fixed in SVN(48874).

That is,
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080324/060308.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 Mar 29 15:42:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Mar 2008 15:42:47 -0500
Subject: [LLVMbugs] [Bug 2179] New: llvm-gcc checkingenabled_flag is busted.
Message-ID: <bug-2179-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2179

           Summary: llvm-gcc checkingenabled_flag is busted.
           Product: tools
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc doesn't always find the built LLVM tree. Steps to reproduce:

1. Configure a LLVM with "configure --enable-optimized=1". Build it with
"make". Ensure you don't have a Debug/ directory in the tree when you're done.

2. Configure llvm-gcc 4.2. The configure will fail to find the llvm directory
even though --enable-llvm is set correctly.

The problem is caused by this snippet:

  if test -x "$LLVMBASEPATH/Release/bin/llc$EXEEXT"; then
    if test x$checkingenabled_flag  = x ; then        # always false
      echo Found Release LLVM Tree in $LLVMBASEPATH
      LLVMBUILDMODE="Release"
   fi

$checkingenabled_flag will be blank. Even though there's code to set it, I
can't see that it will ever be set.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 29 21:13:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Mar 2008 21:13:53 -0500
Subject: [LLVMbugs] [Bug 2180] New: llvm-gcc-4.2 cannot compile firefox 3 on
 Mac OS X 10.5 x86
Message-ID: <bug-2180-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2180

           Summary: llvm-gcc-4.2 cannot compile firefox 3 on Mac OS X 10.5
                    x86
           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: devchan1 at bellsouth.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1559)
 --> (http://llvm.org/bugs/attachment.cgi?id=1559)
bugpoint-reduced-simplified.bc

llvm version is from svn head
llvm-gcc-4.2 is from svn head
firefox version is from mozilla cvs head

I have attached the 'bugpoint-reduced-simplified.bc' file.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 30 02:41:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 02:41:57 -0500
Subject: [LLVMbugs] [Bug 2181] New: Error in Command Line Library page
Message-ID: <bug-2181-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2181

           Summary: Error in Command Line Library page
           Product: Website
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: trivial
          Priority: P2
         Component: Documentation
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: grrussel at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


http://llvm.org/docs/CommandLine.html

has the following snippet of code

  ofstream Output(OutputFilename.c_str());
  if (Out.good()) .


Out should probably be 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  Sun Mar 30 12:00:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 12:00:13 -0500
Subject: [LLVMbugs] [Bug 2181] Error in Command Line Library page
In-Reply-To: <bug-2181-206@http.llvm.org/bugs/>
Message-ID: <200803301700.m2UH0DTd007476@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2181


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #1 from Chris Lattner <sabre at nondot.org>  2008-03-30 12:00:12 ---
Fixed, 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  Sun Mar 30 12:45:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 12:45:31 -0500
Subject: [LLVMbugs] [Bug 2148] 3 VC8 Build Scripts (Projects) are out of
 date, patches included
In-Reply-To: <bug-2148-206@http.llvm.org/bugs/>
Message-ID: <200803301745.m2UHjVcj008837@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2148


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #6 from Chris Lattner <sabre at nondot.org>  2008-03-30 12:45:31 ---
I think that this has been fixed, please reopen if not.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 30 12:55:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 12:55:43 -0500
Subject: [LLVMbugs] [Bug 780] tools only pull in some symbols from library
	archive
In-Reply-To: <bug-780-206@http.llvm.org/bugs/>
Message-ID: <200803301755.m2UHthTP009258@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=780


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #23 from Chris Lattner <sabre at nondot.org>  2008-03-30 12:55:43 ---
This has been fixed by Devang's change of the passmgr from using RTTI to using
an explicit ID symbol for each pass.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 30 13:04:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 13:04:45 -0500
Subject: [LLVMbugs] [Bug 1808] CBE generates code that will SIGILL: "ote: if
 this code is reached, the program will abort"
In-Reply-To: <bug-1808-206@http.llvm.org/bugs/>
Message-ID: <200803301804.m2UI4jf8009627@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1808


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #7 from Chris Lattner <sabre at nondot.org>  2008-03-30 13:04:44 ---
Unfortunately, there isn't really a good solution for this.  I'd suggest asking
the GCC developers how to handle 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  Sun Mar 30 13:07:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 13:07:24 -0500
Subject: [LLVMbugs] [Bug 1680] x86 inline asm generates wrong code
In-Reply-To: <bug-1680-206@http.llvm.org/bugs/>
Message-ID: <200803301807.m2UI7OB7009761@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1680


Chris Lattner <sabre at nondot.org> changed:

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




--- Comment #6 from Chris Lattner <sabre at nondot.org>  2008-03-30 13:07:24 ---
I fixed this quite awhile back.  llvm-gcc now produces:

        # InlineAsm Start
        %eax %edi %ebp
        # InlineAsm End


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 30 14:07:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Mar 2008 14:07:12 -0500
Subject: [LLVMbugs] [Bug 1570] instcombine doesn't collapse zext/xor/zext
In-Reply-To: <bug-1570-206@http.llvm.org/bugs/>
Message-ID: <200803301907.m2UJ7Cid011692@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1570


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #7 from Nick Lewycky <nicholas at mxc.ca>  2008-03-30 14:07:11 ---
Moved to lib/Target/README.txt.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 31 00:09:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 00:09:00 -0500
Subject: [LLVMbugs] [Bug 2182] New: wrong code emitted for volatile accesses
Message-ID: <bug-2182-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2182

           Summary: wrong code emitted for volatile accesses
           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: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


This is for:

regehr at home-linux:~/volatile/tmp0$ llvm-gcc -v
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/llvm-gcc4.2-2.2.source/obj/../install
--enable-llvm=/home/regehr/llvm-2.2 --enable-languages=c,c++
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5546) (LLVM build)


This code:

volatile int x;
void loop_2 (void) {
  int i;
  for (i=0; i<4; i++) {
    x = x+3;
  }
}

Compiled like this:

llvm-gcc -O2 -S foo.c

Turns into:

loop_2:
        pushl   %ebp
        movl    %esp, %ebp
        movl    x, %eax
        leal    3(%eax), %ecx
        movl    %ecx, x
        leal    6(%eax), %ecx
        movl    %ecx, x
        leal    9(%eax), %ecx
        movl    %ecx, x
        addl    $12, %eax
        movl    %eax, x
        popl    %ebp
        ret

x must be loaded and stored at each loop iteration.  The loads are incorrectly
being optimized away.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 31 10:54:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 10:54:34 -0500
Subject: [LLVMbugs] [Bug 2179] llvm-gcc checkingenabled_flag is busted.
In-Reply-To: <bug-2179-206@http.llvm.org/bugs/>
Message-ID: <200803311554.m2VFsYps022873@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2179


Tanya Lattner <tonic at nondot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tonic at nondot.org
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Tanya Lattner <tonic at nondot.org>  2008-03-31 10:54:32 ---


*** This bug has been marked as a duplicate of bug 2025 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 31 11:03:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 11:03:14 -0500
Subject: [LLVMbugs] [Bug 2183] New: Cannot run make test: ERROR: tcl error
 sourcing /home/edwin /llvm-svn/llvm/test/site.exp
Message-ID: <bug-2183-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2183

           Summary: Cannot run make test: ERROR: tcl error sourcing
                    /home/edwin/llvm-svn/llvm/test/site.exp
           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


On Linux x86-64 I cannot run make check, same on 32-bit.
On x86-64 I have ocamlc installed, on 32-bit I don't. Yet it fails on both.

tclsh is 8.5, but I also have 8.4 installed.

edwin at lightspeed2:~/llvm-svn/llvm$ make check
llvm[0]: Running test suite
make[1]: Entering directory `/home/edwin/llvm-svn/llvm/test'
Making a new site.exp file...
( ulimit -t 600 ; ulimit -d 512000 ; \
         
PATH="/home/edwin/llvm-svn/llvm/Release/bin:/home/edwin/llvm-svn/llvm/test/Scripts:/usr/lib/ccache:/home/edwin/bin:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games"
\
          /usr/bin/runtest --ignore "llvm2cpp.exp" )
ERROR: tcl error sourcing /home/edwin/llvm-svn/llvm/test/site.exp.
extra characters after close-quote
    while executing
"set ocamlc ""/usr/bin/ocamlc.opt" -cc g++ -I
/home/edwin/llvm-svn/llvm/Release/lib/ocaml"
## All variables above are generated by configure. Do Not Ed..."
    (file "/home/edwin/llvm-svn/llvm/test/site.exp" line 24)
    invoked from within
"source /home/edwin/llvm-svn/llvm/test/site.exp"
    ("uplevel" body line 1)
    invoked from within
"uplevel #0 source /home/edwin/llvm-svn/llvm/test/site.exp"
    invoked from within
"catch "uplevel #0 source $file""
make[1]: *** [check-local] Error 1
make[1]: Leaving directory `/home/edwin/llvm-svn/llvm/test'
make: *** [check] Error 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 Mar 31 11:51:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 11:51:07 -0500
Subject: [LLVMbugs] [Bug 2183] Cannot run make test: ERROR: tcl error
 sourcing /home/edwin/ llvm-svn/llvm/test/site.exp
In-Reply-To: <bug-2183-206@http.llvm.org/bugs/>
Message-ID: <200803311651.m2VGp7vu024679@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2183


Gordon Henriksen <gordonhenriksen at mac.com> changed:

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




--- Comment #3 from Gordon Henriksen <gordonhenriksen at mac.com>  2008-03-31 11:51:07 ---
I've reverted Erick's patch until a more thoroughly tested version is
available. Do not install ocaml on a path with spaces.

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080331/060443.html

LLVM's makefiles (as with most makefiles) do not support spaces in pathnames.
This is unfortunate, but I don't think that fixing it is realistic due to
'make's limitations.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 31 11:49:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 11:49:46 -0500
Subject: [LLVMbugs] [Bug 2184] New: internal compiler error in
 lhd_set_decl_assembler_name
Message-ID: <bug-2184-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=2184

           Summary: internal compiler error in lhd_set_decl_assembler_name
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=1562)
 --> (http://llvm.org/bugs/attachment.cgi?id=1562)
test case

With the attached source (from the GCC 4.2.1 testsuite) I get:

$ ~/llvm/llvm-gcc4.2-2.2-x86-linux-RHEL4/bin/llvm-gcc -w -O3 pr22061-1.c
pr22061-1.c: In function ???main???:
pr22061-1.c:6: internal compiler error: in lhd_set_decl_assembler_name, at
langhooks.c:165
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Mar 31 15:55:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 15:55:57 -0500
Subject: [LLVMbugs] [Bug 2182] wrong code emitted for volatile accesses
In-Reply-To: <bug-2182-206@http.llvm.org/bugs/>
Message-ID: <200803312055.m2VKtvHf000420@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2182


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #3 from Dan Gohman <gohman at apple.com>  2008-03-31 15:55:57 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080331/060458.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 Mar 31 19:59:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Mar 2008 19:59:45 -0500
Subject: [LLVMbugs] [Bug 2180] llvm-gcc-4.2 cannot compile firefox 3 on Mac
 OS X 10.5 x86
In-Reply-To: <bug-2180-206@http.llvm.org/bugs/>
Message-ID: <200804010059.m310xjr0008684@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2180


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #5 from Dale Johannesen <dalej at apple.com>  2008-03-31 19:59:45 ---
Crash is fixed here.
http://llvm.org/viewvc/llvm-project?view=rev&revision=49011
That may not be enough to get the code to work, though.  In the shrunk example
the movq's are deleted, which is legitimate, but I advise checking the output
of the original program.


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