From bugzilla-daemon at cs.uiuc.edu  Sat Aug  1 14:05:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 14:05:37 -0500
Subject: [LLVMbugs] [Bug 3742] Missing warning: "noreturn function does
	return"
In-Reply-To: <bug-3742-206@http.llvm.org/bugs/>
Message-ID: <200908011905.n71J5bkt008572@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3742


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-01 14:05:36 ---
Mike did 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  Sat Aug  1 14:36:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 14:36:10 -0500
Subject: [LLVMbugs] [Bug 4405] Miniperl dies with segmentation fault when
 optimizer is turned on
In-Reply-To: <bug-4405-206@http.llvm.org/bugs/>
Message-ID: <200908011936.n71JaAT8009896@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4405


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #7 from Chris Lattner <clattner at apple.com>  2009-08-01 14:36:09 ---
Yes, lets assume so.  Please reopen or file a new bug 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  Sat Aug  1 16:34:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 16:34:00 -0500
Subject: [LLVMbugs] [Bug 4665] New: no warning when assigning signed value
 to an unsigned variable
Message-ID: <bug-4665-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4665

           Summary: no warning when assigning signed value to an unsigned
                    variable
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


witten /tmp# cat unsigned.c
void foo(void) {
   unsigned int i = -10;
}
witten /tmp# clang -Wall -c unsigned.c

witten /tmp# 

as you can see there's no warning when I try to assign -10 to an unsigned
variable


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  1 17:42:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 17:42:11 -0500
Subject: [LLVMbugs] [Bug 4661] Assertion `(Ty == 0 || Ty == V->getType()) &&
 " Type mismatch in value table!"' failed. On file with named metadata
In-Reply-To: <bug-4661-206@http.llvm.org/bugs/>
Message-ID: <200908012242.n71MgBjK016063@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4661


Tomas Lindquist Olsen <tomas.l.olsen at gmail.com> changed:

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




--- Comment #2 from Tomas Lindquist Olsen <tomas.l.olsen at gmail.com>  2009-08-01 17:42:11 ---
I think you may be right, as I'm not able to reproduce this right now. Sorry
about that, I'll reopen if I hit this again.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  1 17:50:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 17:50:32 -0500
Subject: [LLVMbugs] [Bug 4666] New: Assertion `ResolveConstants.empty() && "
 Constants not resolved?"' failed. on bitcode with metadata
Message-ID: <bug-4666-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4666

           Summary: Assertion `ResolveConstants.empty() && "Constants not
                    resolved?"' failed. on bitcode with metadata
           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: tomas.l.olsen at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3271)
 --> (http://llvm.org/bugs/attachment.cgi?id=3271)
.ll file that doesn't work as .bc

The attached .ll file can be assembled to .bc , but the resulting .bc is not
usable:

$ llvm-as BitManip.ll
$ llc BitManip.bc
llc: /home/tomas/devel/llvm/lib/Bitcode/Reader/BitcodeReader.h:62: void
llvm::BitcodeReaderValueList::clear(): Assertion `ResolveConstants.empty() &&
"Constants not resolved?"' failed.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug  1 20:00:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 20:00:27 -0500
Subject: [LLVMbugs] [Bug 4083] Using Clang on iPhone
In-Reply-To: <bug-4083-206@http.llvm.org/bugs/>
Message-ID: <200908020100.n7210RRe020746@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4083


Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|llvmbugs at cs.uiuc.edu        |
             Status|NEW                         |ASSIGNED




--- Comment #6 from Ted Kremenek <kremenek at apple.com>  2009-08-01 20:00:26 ---
It's about me updating the clang-analyzer website.  I still plan on doing 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  Sat Aug  1 20:01:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 20:01:56 -0500
Subject: [LLVMbugs] [Bug 4667] New: LoopIndexSplit problem with
	ENABLE_EXPENSIVE_CHECKS
Message-ID: <bug-4667-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4667

           Summary: LoopIndexSplit problem with ENABLE_EXPENSIVE_CHECKS
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: dpatel at apple.com, llvmbugs at cs.uiuc.edu


This test fails:
  llvm/test/Transforms/LoopIndexSplit/2008-06-03-DomFrontier.ll
when I run with ENABLE_EXPENSIVE_CHECKS=1 (in a Debug build)

--
FAIL:
/Volumes/Data/Users/ddunbar/llvm/test/Transforms/LoopIndexSplit/2008-06-03-DomFrontier.ll
Failed with signal(SIGABRT) at line 1
while running:  
Dominator Information for WirelessCreatePSK
Pass 'Rotate Loops'
----- Valid -----
  DomFrontier for BB%entry is:  
  DomFrontier for BB%bb52 is:   %bb52
  DomFrontier for BB%bb63 is:   %bb63%bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb68 is:   %bb63
  DomFrontier for BB%bb131 is:  %bb63
  DomFrontier for BB%bb142 is:  %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb63.clone.bb142.loopexit_crit_edge is:    %bb142.loopexit
  DomFrontier for BB%bb63.preheader is: %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb142.loopexit is: %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb142.preheader is:        %bb52.backedge
  DomFrontier for BB%bb52.loopexit is:  %bb52.backedge
  DomFrontier for BB%bb52.backedge is:  %bb52
  DomFrontier for BB%bb63.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb.nph is: %bb142.loopexit
  DomFrontier for BB%bb131.clone is:    %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb89.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb131.bb63.clone_crit_edge is:    
%bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb.nph7 is:        %bb52.loopexit
  DomFrontier for BB%bb142.bb52.loopexit_crit_edge is:  %bb52.loopexit
----- Invalid -----
  DomFrontier for BB%entry is:  
  DomFrontier for BB%bb52 is:   %bb52
  DomFrontier for BB%bb63 is:   %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb68 is:   %bb63
  DomFrontier for BB%bb131 is:  %bb63
  DomFrontier for BB%bb142 is:  %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb63.clone.bb142.loopexit_crit_edge is:    %bb142.loopexit
  DomFrontier for BB%bb63.preheader is: %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb142.loopexit is: %bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb142.preheader is:        %bb52.backedge
  DomFrontier for BB%bb52.loopexit is:  %bb52.backedge
  DomFrontier for BB%bb52.backedge is:  %bb52
  DomFrontier for BB%bb63.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb.nph is: %bb142.loopexit
  DomFrontier for BB%bb131.clone is:    %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb89.clone is:     %bb142.loopexit%bb89.clone
  DomFrontier for BB%bb131.bb63.clone_crit_edge is:    
%bb63.preheader%bb52.loopexit
  DomFrontier for BB%bb.nph7 is:        %bb52.loopexit
  DomFrontier for BB%bb142.bb52.loopexit_crit_edge is:  %bb52.loopexit
Invalid dominator info
UNREACHABLE executed at PassManager.cpp:744!
0   opt               0x000000010050f0dd PrintStackTrace(void*) + 38
1   opt               0x000000010050f769 SignalHandler(int) + 294
2   libSystem.B.dylib 0x00007fff84b3c2fa _sigtramp + 26
3   libstdc++.6.dylib 0x00007fff824eb472
__gnu_debug::_Safe_iterator_base::_M_detach() + 134
4   opt               0x0000000100060387 llvm::llvm_report_error(llvm::Twine
const&) + 0
5   opt               0x000000010046993c
llvm::PMDataManager::verifyDomInfo(llvm::Pass&, llvm::Function&) + 830
6   opt               0x000000010032c4d7
llvm::LPPassManager::runOnFunction(llvm::Function&) + 1365
7   opt               0x0000000100469abd
llvm::FPPassManager::runOnFunction(llvm::Function&) + 341
8   opt               0x0000000100469c9e
llvm::FPPassManager::runOnModule(llvm::Module&) + 82
9   opt               0x00000001004686a0
llvm::MPPassManager::runOnModule(llvm::Module&) + 418
10  opt               0x0000000100469f99
llvm::PassManagerImpl::run(llvm::Module&) + 111
11  opt               0x0000000100469ffb llvm::PassManager::run(llvm::Module&)
+ 33
12  opt               0x000000010007e616 main + 4423
13  opt               0x0000000100063d7c start + 52
14  opt               0x0000000000000006 start + 4294558398
Stack dump:
0.      Running pass 'Function Pass Manager' on module '<stdin>'.
1.      Running pass 'Loop Pass Manager' on function '@WirelessCreatePSK'

--


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  1 20:24:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 20:24:59 -0500
Subject: [LLVMbugs] [Bug 4668] New: "Ran out of registers during register
 allocation" on x86-32 with asm using only 6 registers
Message-ID: <bug-4668-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4668

           Summary: "Ran out of registers during register allocation" on
                    x86-32 with asm using only 6 registers
           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=3272)
 --> (http://llvm.org/bugs/attachment.cgi?id=3272)
Reduced testcase

Per summary, incorrect error; to reproduce, run llc (with no special arguments)
over attached.  Seems to work with "-regalloc=local".

Testcase reduced from ffmpeg (the source of all fun inline asm issues).


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug  1 22:42:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 22:42:13 -0500
Subject: [LLVMbugs] [Bug 3800] asm inline : Ran out of registers at -O0
In-Reply-To: <bug-3800-206@http.llvm.org/bugs/>
Message-ID: <200908020342.n723gDcS025817@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3800


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #4 from Anders Carlsson <andersca at mac.com>  2009-08-01 22:42:10 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090727/019917.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 Aug  1 22:49:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 22:49:38 -0500
Subject: [LLVMbugs] [Bug 4669] New: Assertion failure
Message-ID: <bug-4669-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4669

           Summary: Assertion failure
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The code

--------
declare <1 x i64> @llvm.x86.mmx.pslli.q(<1 x i64>, i32) nounwind readnone

define <1 x i64> @test(i64 %t) nounwind {
entry:
        %t1 = insertelement <1 x i64> undef, i64 %t, i32 0
        %t2 = tail call <1 x i64> @llvm.x86.mmx.pslli.q(<1 x i64> %t1, i32 48)
nounwind readnone
        ret <1 x i64> %t2
}
----------

Couses llc -march=x86-64 to fail with

SelectionDAG.cpp:2407: llvm::SDValue llvm::SelectionDAG::getNode(unsigned int,
llvm::DebugLoc, llvm::MVT, llvm::SDValue): Assertion
`Operand.getValueType().bitsLT(VT) && "Invalid anyext node, dst < src!"'
failed.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug  1 23:18:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Aug 2009 23:18:19 -0500
Subject: [LLVMbugs] [Bug 4670] New: printf-style parameter types not checked
Message-ID: <bug-4670-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4670

           Summary: printf-style parameter types not checked
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: ASSIGNED
          Severity: minor
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: parseerror at gmail.com
        ReportedBy: parseerror at gmail.com
                CC: llvmbugs at cs.uiuc.edu


clang does not warn about a mismatch between format string and parameter types,
allowing things like, i.e.

printf("%s", 1);

i will work on getting some basic checking running, with warnings based on
GCC's behavior.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug  2 08:50:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 08:50:01 -0500
Subject: [LLVMbugs] [Bug 4671] New: `.gnu.linkonce.t.XXX' referenced in
 section `.rodata' of XXX.o: defined in discarded section
 `.gnu.linkonce.t.X' of XXX. o regression with weak_odr function using
 switch table
Message-ID: <bug-4671-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4671

           Summary: `.gnu.linkonce.t.XXX' referenced in section `.rodata' of
                    XXX.o: defined in discarded section `.gnu.linkonce.t.X'
                    of XXX.o regression with weak_odr function using switch
                    table
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: regression
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tomas.l.olsen at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3273)
 --> (http://llvm.org/bugs/attachment.cgi?id=3273)
bc 1

This was a little hard to reproduce, but seems to be caused by:

* a weak_odr uses a switch, and the switch is big enough to result in a
switch-table during codegen.

* the weak function itself goes in a .linkonce section
* the switch table in .rodata

if two modules contain the same function (and its switch table), once linking
(natively), one of the linkonce sections is removed, but the rodata is not.

Reproduce like this:
$ llc foo1.bc
$ llc foo2.bc
$ llc foo3.bc
$ gcc foo1.s foo2.s foo3.s -o foo

This produces some linker errors, most are due to these being D module and gcc
isn't linking the D runtime library, but the bug is shown as well:

`.gnu.linkonce.t._D3foo9__T2fnTiZ2fnFiZi' referenced in section `.rodata' of
/tmp/ccA6toMe.o: defined in discarded section
`.gnu.linkonce.t._D3foo9__T2fnTiZ2fnFiZi' of /tmp/ccA6toMe.o

I hope this is not too hard to fix, and took a little while to track down a
reasonable(?) 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  Sun Aug  2 14:16:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 14:16:15 -0500
Subject: [LLVMbugs] [Bug 4672] New: CodeGen/SystemZ ,
 CodeGen/Thumb & CodeGen/  Thumb2 fail test on Solaris.
Message-ID: <bug-4672-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4672

           Summary: CodeGen/SystemZ , CodeGen/Thumb & CodeGen/Thumb2 fail
                    test on Solaris.
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DejaGNU
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Good day,

The follow target tests now fail on Solaris/AuroraUX hosts.

FAIL:
/export/home2/edward/lab/llvm/llvm/test/CodeGen/SystemZ/05-MemImmStores.ll
Failed with exit(1) at line 5
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/SystemZ/05-MemImmStores.ll | 
llc | /usr/bin/ggrep mviy  | count 1
count: expected 1 lines and got        2.
child process exited abnormally
FAIL:
/export/home2/edward/lab/llvm/llvm/test/CodeGen/SystemZ/05-MemLoadsStores.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/SystemZ/05-MemLoadsStores.ll | 
llc | /usr/bin/ggrep ly     | count 2
count: expected 2 lines and got        4.
child process exited abnormally
Running /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb/dg.exp ...
Running /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/dg.exp ...
FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll
Failed with exit(1) at line 3
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll |  llc
-march=thumb -mattr=+thumb2 | /usr/bin/ggrep mov\.w | /usr/bin/ggrep 1
child process exited abnormally
FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll
Failed with exit(1) at line 3
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll |  llc
-march=thumb -mattr=+thumb2 | /usr/bin/ggrep mov\.w | /usr/bin/ggrep 1
child process exited abnormally
FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll
Failed with exit(1) at line 3
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll |  llc
-march=thumb -mattr=+thumb2 | /usr/bin/ggrep mov\.w | /usr/bin/ggrep 1
child process exited abnormally
FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll |  llc
-march=thumb -mattr=+thumb2 | /usr/bin/ggrep {mov\.w\W*r[0-9],\W*#[0-9]*} |
/usr/bin/ggrep {#171\|#1179666\|#872428544\|#1448498774\|#66846720} | count 5
count: expected 5 lines and got        4.
child process exited abnormally

-bash-3.2$ /opt/clang/bin/llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/SystemZ/05-MemImmStores.ll | 
/opt/clang/bin/llc


        .text
        .align  2
        .globl  foo1
        .type   foo1, at function
foo1:
.LBB1_0:        # entry
        lghi    %r1, 1
        stg     %r1, 8(%r2)
        br      %r14
        .size   foo1, .-foo1


        .align  2
        .globl  foo2
        .type   foo2, at function
foo2:
.LBB2_0:        # entry
        lhi     %r1, 2
        sty     %r1, 4(%r2)
        br      %r14
        .size   foo2, .-foo2


        .align  2
        .globl  foo3
        .type   foo3, at function
foo3:
.LBB3_0:        # entry
        lhi     %r1, 3
        sth     %r1, 2(%r2)
        br      %r14
        .size   foo3, .-foo3


        .align  2
        .globl  foo4
        .type   foo4, at function
foo4:
.LBB4_0:        # entry
        mviy    1(%r2), 4
        br      %r14
        .size   foo4, .-foo4


        .align  2
        .globl  foo5
        .type   foo5, at function
foo5:
.LBB5_0:        # entry
        mviy    -1(%r2), 4
        br      %r14
        .size   foo5, .-foo5


        .align  2
        .globl  foo6
        .type   foo6, at function
foo6:
.LBB6_0:        # entry
        lhi     %r1, 3
        sthy    %r1, -2(%r2)
        br      %r14
        .size   foo6, .-foo6

        .section        .note.GNU-stack,"", at progbits


Let me know what other information I can provide.

Cheers,
Edward O'Callaghan.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  2 15:38:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 15:38:40 -0500
Subject: [LLVMbugs] [Bug 4673] New: Empty classes have positive DataSize
Message-ID: <bug-4673-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4673

           Summary: Empty classes have positive DataSize
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rideau3 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


While implementing the __is_empty type trait, I chose to use the DataSize
member of ASTRecordLayout. As it turns out, a completely empty class ("struct
foo {};") has a DataSize equal to its Size. If the class contains a
constructor, destructor, or copy assignment operator, DataSize is 0 as
expected. Other member functions do not seem to affect this. Those three
functions are the definition of a trivial class - is that the issue?


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug  2 16:10:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 16:10:24 -0500
Subject: [LLVMbugs] [Bug 4673] Empty classes have positive DataSize
In-Reply-To: <bug-4673-206@http.llvm.org/bugs/>
Message-ID: <200908022110.n72LAOxK028849@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4673


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-02 16:10:22 ---
This is working as intended; please use some other mechanism to determine
__is_empty.  See http://www.codesourcery.com/public/cxx-abi/abi.html#POD .


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  2 16:38:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 16:38:13 -0500
Subject: [LLVMbugs] [Bug 4675] New: lli: symbol lookup error:
 /usr/lib/oprofile/libopagent.so.1 : undefined symbol: bfd_init
Message-ID: <bug-4675-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4675

           Summary: lli: symbol lookup error:
                    /usr/lib/oprofile/libopagent.so.1: undefined symbol:
                    bfd_init
           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, jyasskin at google.com


The JIT's not working. Here's a sample of trying to run it on the 'hello world'
program:

$ llvm-commit/Debug/bin/lli hello.bc
llvm-commit/Debug/bin/lli: symbol lookup error:
/usr/lib/oprofile/libopagent.so.1: undefined symbol: bfd_init

I don't see how libopagent.so.1 requires "bfd_init".

$ ldd /usr/lib/oprofile/libopagent.so.1
        linux-gate.so.1 =>  (0xffffe000)
        libdl.so.2 => /lib/i686/cmov/libdl.so.2 (0xf7f44000)
        libc.so.6 => /lib/i686/cmov/libc.so.6 (0xf7de6000)
        /lib/ld-linux.so.2 (0xf7f86000)
$ nm /usr/lib/oprofile/libopagent.so.1 | grep bfd
nm: /usr/lib/oprofile/libopagent.so.1: no symbols

Here's how lli was built:

make[2]: Entering directory `/home/nicholas/llvm-commit/tools/lli'
llvm[2]: Linking Debug executable lli
g++-4.4 -I/home/nicholas/llvm-commit/include
-I/home/nicholas/llvm-commit/tools/lli  -D_DEBUG  -D_GNU_SOURCE
-D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -g -fno-exceptions -fPIC 
-Woverloaded-virtual   -pedantic -Wno-long-long -Wall -W -Wno-unused-parameter
-Wwrite-strings  -g  -Wl,-R -Wl,/home/nicholas/llvm-commit/Debug/bin
-Wl,-export-dynamic -L/home/nicholas/llvm-commit/Debug/lib
-L/home/nicholas/llvm-commit/Debug/lib  -o
/home/nicholas/llvm-commit/Debug/bin/lli 
/home/nicholas/llvm-commit/tools/lli/Debug/lli.o  \
        -lLLVMBitReader -lLLVMX86CodeGen -lLLVMX86Info -lLLVMSelectionDAG
-lLLVMAsmPrinter -lLLVMInterpreter -lLLVMJIT -lLLVMExecutionEngine
-lLLVMCodeGen -lLLVMScalarOpts -lLLVMTransformUtils -lLLVMipa -lLLVMAnalysis
-lLLVMTarget -lLLVMMC -lLLVMCore -lLLVMSupport -lLLVMSystem   -lopagent
-lpthread -lffi -ldl -lm  -L/usr/lib/oprofile -Wl,-rpath,/usr/lib/oprofile
llvm[2]: ======= Finished Linking Debug Executable lli
make[2]: Leaving directory `/home/nicholas/llvm-commit/tools/lli'

Here's what the loader thinks about how lli was linked:

$ ldd Debug/bin/lli
        linux-gate.so.1 =>  (0xffffe000)
        libopagent.so.1 => /usr/lib/oprofile/libopagent.so.1 (0xf7f37000)
        libpthread.so.0 => /lib/i686/cmov/libpthread.so.0 (0xf7ee5000)
        libffi.so.5 => /usr/lib/libffi.so.5 (0xf7ede000)
        libdl.so.2 => /lib/i686/cmov/libdl.so.2 (0xf7eda000)
        libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xf7dea000)
        libm.so.6 => /lib/i686/cmov/libm.so.6 (0xf7dc4000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xf7d99000)
        libc.so.6 => /lib/i686/cmov/libc.so.6 (0xf7c3b000)
        /lib/ld-linux.so.2 (0xf7f3c000)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  2 22:01:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Aug 2009 22:01:41 -0500
Subject: [LLVMbugs] [Bug 4669] Assertion failure
In-Reply-To: <bug-4669-206@http.llvm.org/bugs/>
Message-ID: <200908030301.n7331fXt008231@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4669


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

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




--- Comment #2 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-08-02 22:01:41 ---
Fixed in 77940.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  3 05:34:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 05:34:47 -0500
Subject: [LLVMbugs] [Bug 4677] New: Unsupported asm: input constraint with a
 matching output constraint of incompatible type! in x86_64 mode
Message-ID: <bug-4677-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4677

           Summary: Unsupported asm: input constraint with a matching output
                    constraint of incompatible type! in x86_64 mode
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: regression
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: benny.kra at gmail.com
                CC: llvmbugs at cs.uiuc.edu


$ cat t.c
int main() {
  unsigned _data = 42;
  __asm__("bswap   %0":"+r"(_data));
  /* asm from /usr/include/libkern/i386/_OSByteOrder.h */
}

$ clang -m64 t.c
LLVM ERROR: llvm: error: Unsupported asm: input constraint with a matching
output constraint of incompatible type!

I'm using LLVM r77969 on OS X 10.5.


--- Comment #1 from Benjamin Kramer <benny.kra at gmail.com>  2009-08-03 05:34:32 ---
Created an attachment (id=3279)
 --> (http://llvm.org/bugs/attachment.cgi?id=3279)
bitcode


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  3 18:30:05 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 18:30:05 -0500
Subject: [LLVMbugs] [Bug 4678] New: "arguments were dropped" waring
Message-ID: <bug-4678-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4678

           Summary: "arguments were dropped" waring
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Compiling the c++ program

void a(...) {
}
void b() {
  a(0);
}

produces the warning
WARNING: While resolving call to function '_Z1az' arguments were dropped!

At the llvm level this looks like


define void @_Z1az() {
entry:
        ret void
}
define void @_Z1bv() nounwind {
entry:
        call void (...)* bitcast (void ()* @_Z1az to void (...)*)(i32 0)
nounwind
        ret void
}

and the warning can be produced with

opt -std-compile-opts test.bc -disable-output


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug  3 19:06:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 19:06:14 -0500
Subject: [LLVMbugs] [Bug 4679] New: Expect to see machine code sinking,
	not seeing it.
Message-ID: <bug-4679-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4679

           Summary: Expect to see machine code sinking, not seeing it.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan at fallingsnow.net
                CC: llvmbugs at cs.uiuc.edu


Give the following IR and x86:

[[[ LLVM Optimized IR: bleh ]]]

define %"struct.rubinius::Object"* @0(%"struct.rubinius::VM"* %state,
%"struct.rubinius::CallFrame"* %previous, %"struct.rubinius::Dispatch"* %msg,
%"struct.rubinius::Arguments"* %args) {
entry:
        %cfstk94 = alloca [9 x %"struct.rubinius::Object"*], align 4           
; <[9 x %"struct.rubinius::Object"*]*> [#uses=8]
        %var_mem93 = alloca [5 x %"struct.rubinius::Object"*], align 4         
; <[5 x %"struct.rubinius::Object"*]*> [#uses=6]
        %cfstk94.sub = getelementptr [9 x %"struct.rubinius::Object"*]*
%cfstk94, i32 0, i32 0          ; <%"struct.rubinius::Object"**> [#uses=1]
        %var_mem93.sub = getelementptr [5 x %"struct.rubinius::Object"*]*
%var_mem93, i32 0, i32 0              ; <%"struct.rubinius::Object"**>
[#uses=1]
        %stack = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32
0, i32 8                ; <%"struct.rubinius::Object"**> [#uses=1]
        %0 = getelementptr %"struct.rubinius::Dispatch"* %msg, i32 0, i32 2    
        ; <%"struct.rubinius::Executable"**> [#uses=1]
        %msg.exec = load %"struct.rubinius::Executable"** %0            ;
<%"struct.rubinius::Executable"*> [#uses=1]
        %1 = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32 0,
i32 3            ; <%"struct.rubinius::Object"**> [#uses=1]
        %previous.c = bitcast %"struct.rubinius::CallFrame"* %previous to
%"struct.rubinius::Object"*           ; <%"struct.rubinius::Object"*> [#uses=1]
        store %"struct.rubinius::Object"* %previous.c,
%"struct.rubinius::Object"** %cfstk94.sub
        %2 = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32 0,
i32 2            ; <%"struct.rubinius::Object"**> [#uses=1]
        %msg.c = bitcast %"struct.rubinius::Dispatch"* %msg to
%"struct.rubinius::Object"*              ; <%"struct.rubinius::Object"*>
[#uses=1]
        store %"struct.rubinius::Object"* %msg.c, %"struct.rubinius::Object"**
%2
        %cm.c = getelementptr %"struct.rubinius::Executable"* %msg.exec, i32 0,
i32 0           ; <%"struct.rubinius::Object"*> [#uses=1]
        store %"struct.rubinius::Object"* %cm.c, %"struct.rubinius::Object"**
%1
        %3 = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32 0,
i32 4            ; <%"struct.rubinius::Object"**> [#uses=1]
        store %"struct.rubinius::Object"* inttoptr (i32 16 to
%"struct.rubinius::Object"*), %"struct.rubinius::Object"** %3
        %4 = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32 0,
i32 5            ; <%"struct.rubinius::Object"**> [#uses=1]
        %5 = getelementptr [9 x %"struct.rubinius::Object"*]* %cfstk94, i32 0,
i32 7            ; <%"struct.rubinius::Object"**> [#uses=1]
        %vars.c = bitcast [5 x %"struct.rubinius::Object"*]* %var_mem93 to
%"struct.rubinius::Object"*          ; <%"struct.rubinius::Object"*> [#uses=1]
        store %"struct.rubinius::Object"* %vars.c, %"struct.rubinius::Object"**
%5
        store %"struct.rubinius::Object"* null, %"struct.rubinius::Object"**
%var_mem93.sub
        %6 = getelementptr %"struct.rubinius::Arguments"* %args, i32 0, i32 0  
        ; <%"struct.rubinius::Object"**> [#uses=1]
        %args.recv = load %"struct.rubinius::Object"** %6               ;
<%"struct.rubinius::Object"*> [#uses=3]
        %7 = getelementptr [5 x %"struct.rubinius::Object"*]* %var_mem93, i32
0, i32 2          ; <%"struct.rubinius::Object"**> [#uses=1]
        store %"struct.rubinius::Object"* %args.recv,
%"struct.rubinius::Object"** %7
        %8 = getelementptr %"struct.rubinius::Dispatch"* %msg, i32 0, i32 1    
        ; <%"struct.rubinius::Module"**> [#uses=1]
        %msg.module = load %"struct.rubinius::Module"** %8              ;
<%"struct.rubinius::Module"*> [#uses=1]
        %9 = getelementptr [5 x %"struct.rubinius::Object"*]* %var_mem93, i32
0, i32 4          ; <%"struct.rubinius::Object"**> [#uses=1]
        %msg.module.c = getelementptr %"struct.rubinius::Module"* %msg.module,
i32 0, i32 0             ; <%"struct.rubinius::Object"*> [#uses=1]
        store %"struct.rubinius::Object"* %msg.module.c,
%"struct.rubinius::Object"** %9
        %10 = getelementptr %"struct.rubinius::Arguments"* %args, i32 0, i32 1 
        ; <%"struct.rubinius::Object"**> [#uses=1]
        %args.block = load %"struct.rubinius::Object"** %10             ;
<%"struct.rubinius::Object"*> [#uses=1]
        %11 = getelementptr [5 x %"struct.rubinius::Object"*]* %var_mem93, i32
0, i32 3         ; <%"struct.rubinius::Object"**> [#uses=1]
        store %"struct.rubinius::Object"* %args.block,
%"struct.rubinius::Object"** %11
        %12 = getelementptr [5 x %"struct.rubinius::Object"*]* %var_mem93, i32
0, i32 1         ; <%"struct.rubinius::Object"**> [#uses=1]
        store %"struct.rubinius::Object"* null, %"struct.rubinius::Object"**
%12
        store %"struct.rubinius::Object"* %args.recv,
%"struct.rubinius::Object"** %stack
        store %"struct.rubinius::Object"* inttoptr (i32 2 to
%"struct.rubinius::Object"*), %"struct.rubinius::Object"** %4
        %class_pos = getelementptr %"struct.rubinius::Object"* %args.recv, i32
0, i32 0, i32 1          ; <%"struct.rubinius::Class"**> [#uses=1]
        %ref_class = load %"struct.rubinius::Class"** %class_pos               
; <%"struct.rubinius::Class"*> [#uses=1]
        %class_id_pos = getelementptr %"struct.rubinius::Class"* %ref_class,
i32 0, i32 3               ; <i32*> [#uses=1]
        %class_id = load i32* %class_id_pos             ; <i32> [#uses=1]
        %check_class_id = icmp eq i32 %class_id, 1076           ; <i1>
[#uses=1]
        br i1 %check_class_id, label %correct_class, label %use_send

correct_class:          ; preds = %entry
        ret %"struct.rubinius::Object"* inttoptr (i32 23 to
%"struct.rubinius::Object"*)

use_send:               ; preds = %entry
        %call_frame = bitcast [9 x %"struct.rubinius::Object"*]* %cfstk94 to
%"struct.rubinius::CallFrame"*             ; <%"struct.rubinius::CallFrame"*>
[#uses=1]
        %13 = call %"struct.rubinius::Object"*
@rbx_continue_uncommon(%"struct.rubinius::VM"* %state,
%"struct.rubinius::CallFrame"* %call_frame, %"struct.rubinius::Arguments"*
%args, i32 0)          ; <%"struct.rubinius::Object"*> [#uses=1]
        ret %"struct.rubinius::Object"* %13
}

[[[ JIT Machine Code ]]]
 0x26b1010  sub $0x4c, %esp         
 0x26b1013  mov 0x58(%esp), %eax    
 0x26b1017  mov 0x8(%eax), %ecx     
 0x26b101a  mov 0x54(%esp), %edx    
 0x26b101e  mov %edx, 0x28(%esp)    
 0x26b1022  mov %eax, 0x30(%esp)    
 0x26b1026  mov %ecx, 0x34(%esp)    
 0x26b102a  mov $0x10, 0x38(%esp)   
 0x26b1032  lea 0x14(%esp), %ecx    
 0x26b1036  mov %ecx, 0x44(%esp)    
 0x26b103a  mov $0x0, 0x14(%esp)    
 0x26b1042  mov 0x5c(%esp), %ecx    
 0x26b1046  mov (%ecx), %edx        
 0x26b1048  mov %edx, 0x1c(%esp)    
 0x26b104c  mov 0x4(%eax), %eax     
 0x26b104f  mov %eax, 0x24(%esp)    
 0x26b1053  mov 0x4(%ecx), %eax     
 0x26b1056  mov %eax, 0x20(%esp)    
 0x26b105a  mov $0x0, 0x18(%esp)    
 0x26b1062  mov %edx, 0x48(%esp)    
 0x26b1066  mov $0x2, 0x3c(%esp)    
 0x26b106e  mov 0x4(%edx), %eax     
 0x26b1071  cmp $0x434, 0x24(%eax)  
 0x26b1078  jnz 0x77                 ; 0x26b1087
 0x26b107e  mov $0x17, %eax         
 0x26b1083  add $0x4c, %esp         
 0x26b1086  ret                     
 0x26b1087  mov 0x50(%esp), %eax    
 0x26b108b  mov %ecx, 0x8(%esp)     
 0x26b108f  lea 0x28(%esp), %ecx    
 0x26b1093  mov %ecx, 0x4(%esp)     
 0x26b1097  mov %eax, (%esp)        
 0x26b109a  mov $0x0, 0xc(%esp)     
 0x26b10a2  call 0xfffffffffda6c5c0  ; 0x11d5d0 rbx_continue_uncommon
 0x26b10a7  add $0x4c, %esp         
 0x26b10aa  ret                     


The vast majority of those mov instructions before 0x26b106e could be sunk to
around 0x26b1087, as their values are unused in the terminating branch starting
on 0x26b107e.

Is there a construct in here that is preventing sinking? Is the sinker not
working correctly? Please advise.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  3 19:27:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 19:27:40 -0500
Subject: [LLVMbugs] [Bug 4643] pathological inlining/unrolling causes
 SelectionDAG:: ComputeMaskedBits to eat time
In-Reply-To: <bug-4643-206@http.llvm.org/bugs/>
Message-ID: <200908040027.n740RekN000352@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4643


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #5 from Dan Gohman <gohman at apple.com>  2009-08-03 19:27:29 ---
With opt -O3 and then llc on the result,  the problem becomes very noticeable.
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083460.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 Aug  3 22:37:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 22:37:55 -0500
Subject: [LLVMbugs] [Bug 4171] darwin-string-literals test failure on ppc32
 for utf-16 string
In-Reply-To: <bug-4171-206@http.llvm.org/bugs/>
Message-ID: <200908040337.n743btXX006289@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4171


Mark Cianciosa <cianci66 at mac.com> changed:

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




--- Comment #5 from Mark Cianciosa <cianci66 at mac.com>  2009-08-03 22:37:53 ---
I can confirm this test now passes on powerpc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  3 22:45:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Aug 2009 22:45:27 -0500
Subject: [LLVMbugs] [Bug 4678] "arguments were dropped" warning
In-Reply-To: <bug-4678-206@http.llvm.org/bugs/>
Message-ID: <200908040345.n743jRbi006545@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4678


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

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




--- Comment #5 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-08-03 22:45:18 ---
Fixed in revision 78039.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 01:00:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 01:00:50 -0500
Subject: [LLVMbugs] [Bug 4666] Assertion `ResolveConstants.empty() &&
 "Constants not resolved?" ' failed. on bitcode with metadata
In-Reply-To: <bug-4666-206@http.llvm.org/bugs/>
Message-ID: <200908040600.n7460ohE011252@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4666


devang.patel <devang.patel at gmail.com> changed:

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




--- Comment #1 from devang.patel <devang.patel at gmail.com>  2009-08-04 01:00:50 ---
Fixed. Rev - 78056.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 04:23:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 04:23:01 -0500
Subject: [LLVMbugs] [Bug 4680] New: Assembler error due to invalid 'orn'
	instruction
Message-ID: <bug-4680-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4680

           Summary: Assembler error due to invalid 'orn' instruction
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3281)
 --> (http://llvm.org/bugs/attachment.cgi?id=3281)
Bytecode

Consider the attached bitcode (md5 from libiberty).

Assembler rejects the generated assembly with the following messages:

md5.s: Assembler messages:
md5.s:563: Error: bad instruction `orn r7,r5,r3'
md5.s:572: Error: bad instruction `orn r7,lr,r12'
md5.s:580: Error: bad instruction `orn r6,r3,r5'
md5.s:588: Error: bad instruction `orn r6,r12,lr'
md5.s:597: Error: bad instruction `orn r6,r5,r3'
md5.s:605: Error: bad instruction `orn lr,r2,r12'
md5.s:614: Error: bad instruction `orn lr,r3,r5'
md5.s:622: Error: bad instruction `orn lr,r12,r2'
md5.s:631: Error: bad instruction `orn r5,lr,r3'
md5.s:640: Error: bad instruction `orn r5,r2,r12'
md5.s:648: Error: bad instruction `orn r5,r3,lr'
md5.s:657: Error: bad instruction `orn r5,r12,r2'
md5.s:665: Error: bad instruction `orn r4,lr,r3'
md5.s:674: Error: bad instruction `orn r4,r2,r12'
md5.s:683: Error: bad instruction `orn r4,r3,lr'
md5.s:693: Error: bad instruction `orn r4,r12,r2'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 06:25:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 06:25:36 -0500
Subject: [LLVMbugs] [Bug 4681] New: Live intervals assertions durting
 compilation of libmudflap
Message-ID: <bug-4681-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4681

           Summary: Live intervals assertions durting compilation of
                    libmudflap
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3282)
 --> (http://llvm.org/bugs/attachment.cgi?id=3282)
Bugpoint reduced testcase

Consider the attached bytecode. llc currently asserts with:

./llc bugpoint-reduced-simplified.bc 
llc:
/home/asl/proj/llvm/src_arm/include/llvm/CodeGen/LiveIntervalAnalysis.h:136:
llvm::LiveInterval& llvm::LiveIntervals::getInterval(unsigned int): Assertion
`I != r2iMap_.end() && "Interval does not exist for register"' failed.
0   llc             0x0000000000fce3e4
1   llc             0x0000000000fce902
2   libpthread.so.0 0x00007f8a403d36f0
3   libc.so.6       0x00007f8a3f70f9c5 gsignal + 53
4   libc.so.6       0x00007f8a3f710e70 abort + 272
5   libc.so.6       0x00007f8a3f708edf __assert_fail + 239
6   llc             0x0000000000d4b4ff
llvm::LiveIntervals::getInterval(unsigned int) + 169
7   llc             0x0000000000d87eb7
llvm::SimpleRegisterCoalescing::JoinCopy(llvm::CopyRec&, bool&) + 5293
8   llc             0x0000000000d8941c
llvm::SimpleRegisterCoalescing::CopyCoalesceInMBB(llvm::MachineBasicBlock*,
std::vector<llvm::CopyRec, std::allocator<llvm::CopyRec> >&) + 1318
9   llc             0x0000000000d8960a
llvm::SimpleRegisterCoalescing::joinIntervals() + 234
10  llc             0x0000000000d89cce
llvm::SimpleRegisterCoalescing::runOnMachineFunction(llvm::MachineFunction&) +
746
11  llc             0x0000000000d13877
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 83
12  llc             0x0000000000f4fa6d
llvm::FPPassManager::runOnFunction(llvm::Function&) + 341
13  llc             0x0000000000f5005d
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 111
14  llc             0x0000000000f50204
llvm::FunctionPassManager::run(llvm::Function&) + 112
15  llc             0x000000000080b94e main + 3164
16  libc.so.6       0x00007f8a3f6fca44 __libc_start_main + 244
17  llc             0x0000000000809d49
Stack dump:
0.      Program arguments: ./llc bugpoint-reduced-simplified.bc 
1.      Running pass 'Simple Register Coalescing' on function
'@__mf_heuristic_check'

This breaks llvm-gcc build on ARM


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 11:18:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 11:18:59 -0500
Subject: [LLVMbugs] [Bug 4680] Assembler error due to invalid 'orn'
	instruction
In-Reply-To: <bug-4680-206@http.llvm.org/bugs/>
Message-ID: <200908041618.n74GIxae010606@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4680


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

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




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2009-08-04 11:18:58 ---
That's an assembler bug. Can you get a newer version of gas or you can disable
it in the td 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  Tue Aug  4 11:55:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 11:55:26 -0500
Subject: [LLVMbugs] [Bug 4528] Internal compiler error in
	RegisterScavenging.cpp
In-Reply-To: <bug-4528-206@http.llvm.org/bugs/>
Message-ID: <200908041655.n74GtQxb012003@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4528


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

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




--- Comment #7 from Evan Cheng <evan.cheng at apple.com>  2009-08-04 11:54:16 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083523.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 Aug  4 12:41:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 12:41:57 -0500
Subject: [LLVMbugs] [Bug 4671] `.gnu.linkonce.t.XXX' referenced in section
 `.rodata' of XXX.o: defined in discarded section `.gnu.linkonce.t.X' of
 XXX. o regression with weak_odr function using switch table
In-Reply-To: <bug-4671-206@http.llvm.org/bugs/>
Message-ID: <200908041741.n74Hfvvr013678@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4671


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #9 from Chris Lattner <clattner at apple.com>  2009-08-04 12:40:17 ---
Yes, I think this was a binutils bug, the generated code looks ok to me. 
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 Aug  4 13:19:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 13:19:43 -0500
Subject: [LLVMbugs] [Bug 3800] asm inline : Ran out of registers at -O0
In-Reply-To: <bug-3800-206@http.llvm.org/bugs/>
Message-ID: <200908041819.n74IJhEL014916@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3800


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #5 from Anders Carlsson <andersca at mac.com>  2009-08-04 13:19:14 ---
The fix for this lead to PR4677 so I backed it out.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 13:19:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 13:19:48 -0500
Subject: [LLVMbugs] [Bug 4677] Unsupported asm: input constraint with a
 matching output constraint of incompatible type! in x86_64 mode
In-Reply-To: <bug-4677-206@http.llvm.org/bugs/>
Message-ID: <200908041819.n74IJmiU014929@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4677


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #3 from Anders Carlsson <andersca at mac.com>  2009-08-04 13:19:37 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090803/019990.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 Aug  4 13:47:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 13:47:59 -0500
Subject: [LLVMbugs] [Bug 4681] Live intervals assertions durting compilation
 of libmudflap
In-Reply-To: <bug-4681-206@http.llvm.org/bugs/>
Message-ID: <200908041847.n74IlxlR015789@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4681


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

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




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2009-08-04 13:47:24 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083540.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 Aug  4 16:13:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 16:13:19 -0500
Subject: [LLVMbugs] [Bug 4682] New: Invalid assembler emitted for some thumb
 instructions
Message-ID: <bug-4682-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4682

           Summary: Invalid assembler emitted for some thumb instructions
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3284)
 --> (http://llvm.org/bugs/attachment.cgi?id=3284)
picadd.bc

Consider the attached 2 testcases

1. Reproduce with llc picadd.bc -relocation-model=pic

Invalid "add.w r0, r0, pc" instruction is emitted. It has 3 register operands
and 'w' suffix thus it's treated as T3 instruction by assembler (no other
variant is possible due to 'w' suffix - assembler *must* use wide variant).
However It's not possible to use sp and pc registers as dest and rhs registers
for this encoding. 

It's really should be T1 or T2 instruction (and thus - narrow) with dst = lhs.

2. Reproduce with llc subsp.bc

Invalid "sub.w sp, r7, #8" instruction is emitted.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  4 22:25:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Aug 2009 22:25:54 -0500
Subject: [LLVMbugs] [Bug 4684] New: Unknown loc info!
Message-ID: <bug-4684-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4684

           Summary: Unknown loc info!
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: wendling at apple.com
                CC: llvmbugs at cs.uiuc.edu


$cat foo.ll
; ModuleID = 'bugpoint-reduced-simplified.bc'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
target triple = "x86_64-apple-darwin9.8"

declare i64 @func2(<1 x i64>)

define void @func1(i32* %dest, i32 %length, i32 %src, i32 %const_alpha) {
entry:
        br i1 undef, label %bb, label %bb6

bb:             ; preds = %entry
        br label %bb4

bb3:            ; preds = %bb4
        %0 = call i64 @func2(<1 x i64> undef)
        br label %bb4

bb4:            ; preds = %bb3, %bb
        br i1 undef, label %bb3, label %bb9

bb6:            ; preds = %entry
        unreachable

bb9:            ; preds = %bb4
        ret void
}

$ llvm-as < foo.ll | llc -O0 -mattr=+mmx
Unknown loc info!
UNREACHABLE executed at
/Users/void/llvm/llvm.src/lib/Target/X86/X86FastISel.cpp:1322!
0   llc               0x00a70a50 llvm::sys::SetInterruptFunction(void (*)()) +
82
1   llc               0x00a70e3f llvm::sys::SetInterruptFunction(void (*)()) +
1089
2   libSystem.B.dylib 0x91bf62bb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1849728367
4   libSystem.B.dylib 0x91c6a23a raise + 26
5   libSystem.B.dylib 0x91c76679 abort + 73
6   llc               0x00a522f7 llvm::llvm_unreachable_internal(char const*,
char const*, unsigned int) + 201
7   llc               0x004442c8 (anonymous
namespace)::X86FastISel::X86SelectCall(llvm::Instruction*) + 2274
8   llc               0x004453ff (anonymous
namespace)::X86FastISel::TargetSelectInstruction(llvm::Instruction*) + 439
9   llc               0x00709b67
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function&,
llvm::MachineFunction&, llvm::MachineModuleInfo*, llvm::DwarfWriter*,
llvm::TargetInstrInfo const&) + 1495
10  llc               0x0070a22d
llvm::SelectionDAGISel::runOnMachineFunction(llvm::MachineFunction&) + 903
11  llc               0x007cc2b1
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 77
12  llc               0x009e90b8
llvm::FPPassManager::runOnFunction(llvm::Function&) + 304
13  llc               0x009e9328
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 112
14  llc               0x009e943c
llvm::FunctionPassManager::run(llvm::Function&) + 130
15  llc               0x000033ba main + 3664
16  llc               0x00001f36 start + 54
Stack dump:
0.      Program arguments: llc -O0 -mattr=+mmx 
1.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@func1'
cat foo.ll
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  Wed Aug  5 00:53:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 00:53:30 -0500
Subject: [LLVMbugs] [Bug 4684] Unknown loc info!
In-Reply-To: <bug-4684-206@http.llvm.org/bugs/>
Message-ID: <200908050553.n755rUIi006327@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4684


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-05 00:53:29 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083621.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 Aug  5 01:38:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 01:38:31 -0500
Subject: [LLVMbugs] [Bug 4657] register scavenger asserts with subreg
	lowering
In-Reply-To: <bug-4657-206@http.llvm.org/bugs/>
Message-ID: <200908050638.n756cUUb007752@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4657


Jakob Stoklund Olesen <stoklund at 2pi.dk> changed:

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




--- Comment #4 from Jakob Stoklund Olesen <stoklund at 2pi.dk>  2009-08-05 01:38:25 ---
Fixed: http://llvm.org/viewvc/llvm-project?rev=77989&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  5 07:05:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 07:05:51 -0500
Subject: [LLVMbugs] [Bug 4685] New: Assertion failed: ((Other.isNull() ||
 Other->isCanonical()) && "Type is not canonical!")
Message-ID: <bug-4685-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4685

           Summary: Assertion failed: ((Other.isNull() || Other-
                    >isCanonical()) && "Type is not canonical!")
           Product: clang
           Version: unspecified
          Platform: Other
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Code from gcc/libiberty/xmalloc.c.

# clang -c t.c
Assertion failed: ((Other.isNull() || Other->isCanonical()) && "Type is not
canonical!"), function CreateUnsafe, file
/usr/src-local/llvm/tools/clang/lib/AST/../../include/clang/AST/CanonicalType.h,
line 708.
Stack dump:
0.  Program arguments: /usr/opt/llvm/bin/../libexec/clang-cc -triple
x86_64-unknown-freebsd8.0 -S -disable-free -main-file-name t.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -fdiagnostics-show-option -o /tmp/cc-MBwYsq.s -x c t.c 
1.  t.c:7:1: current parser token '{'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)
# cat t.c
typedef unsigned long size_t;

extern void xmalloc_failed (size_t) __attribute__ ((__noreturn__));

void
xmalloc_failed (size_t size)
{

}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  5 12:44:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 12:44:06 -0500
Subject: [LLVMbugs] [Bug 4572] X86: invalid scale by ESP formed
In-Reply-To: <bug-4572-206@http.llvm.org/bugs/>
Message-ID: <200908051744.n75Hi6Tr009807@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4572


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #16 from Dan Gohman <gohman at apple.com>  2009-08-05 12:43:53 ---
The fix is now enabled with this commit:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083684.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 Aug  5 14:01:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 14:01:04 -0500
Subject: [LLVMbugs] [Bug 4686] New: Another 2 regscavenger assertions
Message-ID: <bug-4686-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4686

           Summary: Another 2 regscavenger assertions
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3287)
 --> (http://llvm.org/bugs/attachment.cgi?id=3287)
regscav2.ll

Consider two attached .ll files. Both causes different assertions in
regscavenger.

Reproduce with llc -mattr=+vfp2


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  5 14:05:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 14:05:55 -0500
Subject: [LLVMbugs] [Bug 4685] Assertion failed: ((Other.isNull() ||
 Other->isCanonical()) && " Type is not canonical!")
In-Reply-To: <bug-4685-206@http.llvm.org/bugs/>
Message-ID: <200908051905.n75J5tGK012437@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4685


Douglas Gregor <dgregor at apple.com> changed:

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




--- Comment #2 from Douglas Gregor <dgregor at apple.com>  2009-08-05 14:05:34 ---
Fixed here:

 
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090803/020018.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 Aug  5 22:08:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Aug 2009 22:08:26 -0500
Subject: [LLVMbugs] [Bug 4687] New: arm10tdmi (armv5) does not support strd
Message-ID: <bug-4687-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4687

           Summary: arm10tdmi (armv5) does not support strd
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3289)
 --> (http://llvm.org/bugs/attachment.cgi?id=3289)
testcase

LLVM is emitting bad assembly for the attached testcase:

$ llc -O2 -mcpu=arm10tdmi x.bc -o - | grep cpu
        .cpu arm10tdmi
$ llc -O2 -mcpu=arm10tdmi x.bc -o - | grep strd
        strd r0, [r6]
        strd r0, [r6, #+8]
$ llc -O2 -mcpu=arm10tdmi x.bc -o - | as
{standard input}: Assembler messages:
{standard input}:545: Error: selected processor does not support `strd r0,[r6]'
{standard input}:551: Error: selected processor does not support `strd
r0,[r6,#+8]'

Apparently ARMv5 is too old for the 'strd' instruction. (Or perhaps my binutils
just thinks it is.)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 01:24:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 01:24:33 -0500
Subject: [LLVMbugs] [Bug 4670] printf-style parameter types not checked
In-Reply-To: <bug-4670-206@http.llvm.org/bugs/>
Message-ID: <200908060624.n766OXOJ003819@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4670


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-06 01:24:32 ---


*** This bug has been marked as a duplicate of bug 4468 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 04:15:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 04:15:39 -0500
Subject: [LLVMbugs] [Bug 4688] New: Fallout from recent changes in call
	lowering
Message-ID: <bug-4688-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4688

           Summary: Fallout from recent changes in call lowering
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: gohman at apple.com, llvmbugs at cs.uiuc.edu


Created an attachment (id=3290)
 --> (http://llvm.org/bugs/attachment.cgi?id=3290)
testcase

Consider the attached testcase (reduction from Qt). Running llc yields:

$ ./llc bugpoint-reduced-simplified.bc 
llc:
/home/asl/proj/llvm/src/lib/CodeGen/SelectionDAG/SelectionDAGBuild.cpp:4473:
void llvm::SelectionDAGLowering::LowerCallTo(llvm::CallSite, llvm::SDValue,
bool, llvm::MachineBasicBlock*): Assertion `(isTailCall || CS.getType() ==
Type::VoidTy || Result.first.getNode()) && "Non-null value expected with
non-void non-tail call!"' failed.
0   llc             0x0000000000e77880
1   llc             0x0000000000e77d9e
2   libpthread.so.0 0x00007fe2649986f0
3   libc.so.6       0x00007fe263cd49c5 gsignal + 53
4   libc.so.6       0x00007fe263cd5e70 abort + 272
5   libc.so.6       0x00007fe263ccdedf __assert_fail + 239
6   llc             0x0000000000aa71c5
llvm::SelectionDAGLowering::LowerCallTo(llvm::CallSite, llvm::SDValue, bool,
llvm::MachineBasicBlock*) + 1763
7   llc             0x0000000000ab6ad2
llvm::SelectionDAGLowering::visitCall(llvm::CallInst&) + 2950
8   llc             0x0000000000ac12a6
llvm::SelectionDAGLowering::visit(unsigned int, llvm::User&) + 1008
9   llc             0x0000000000ac136a
llvm::SelectionDAGLowering::visit(llvm::Instruction&) + 44
10  llc             0x0000000000ad77c8
llvm::SelectionDAGISel::SelectBasicBlock(llvm::BasicBlock*,
llvm::ilist_iterator<llvm::Instruction>,
llvm::ilist_iterator<llvm::Instruction>) + 96
11  llc             0x0000000000ad8363
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function&,
llvm::MachineFunction&, llvm::MachineModuleInfo*, llvm::DwarfWriter*,
llvm::TargetInstrInfo const&) + 2383
12  llc             0x0000000000ad925f
llvm::SelectionDAGISel::runOnMachineFunction(llvm::MachineFunction&) + 947
13  llc             0x0000000000bc1d77
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 83
14  llc             0x0000000000dfaeb5
llvm::FPPassManager::runOnFunction(llvm::Function&) + 341
15  llc             0x0000000000dfb4a5
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 111
16  llc             0x0000000000dfb64c
llvm::FunctionPassManager::run(llvm::Function&) + 112
17  llc             0x0000000000401e4e main + 3164
18  libc.so.6       0x00007fe263cc1a44 __libc_start_main + 244
19  llc             0x0000000000400249
Stack dump:
0.      Program arguments: ./llc bugpoint-reduced-simplified.bc 
1.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@_ZN15QtSharedPointer22internalSafetyCheckAddEPVKv'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 06:38:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 06:38:06 -0500
Subject: [LLVMbugs] [Bug 4689] New: Regalloc problems due to subregs?
Message-ID: <bug-4689-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4689

           Summary: Regalloc problems due to subregs?
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: evan.cheng at apple.com, gohman at apple.com,
                    llvmbugs at cs.uiuc.edu


Created an attachment (id=3291)
 --> (http://llvm.org/bugs/attachment.cgi?id=3291)
Testcase

Recent changes to support kernel code model uncovered interesting bug (however,
it's not connected with it at all, since medium code model is also broken!)

Consider the testcase attached. The bb in question is for.body170

In medium code model we have in that file:

        movabsq $Xrsvd, %r10d

in kernel -
        movq $Xrsvd, %r9d

which is definitely wrong.

I verified - all stuff are of i64 types during isel and instructions selected
in questions are MOV64ri and MOV64ri32 correspondingly having 64 bit output
reg. I haven't looked more, maybe remat introduces invalid reg or something
like this.

The problem seems to have been unnoticed because only small code model is
widely use.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 10:38:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 10:38:37 -0500
Subject: [LLVMbugs] [Bug 4688] Fallout from recent changes in call lowering
In-Reply-To: <bug-4688-206@http.llvm.org/bugs/>
Message-ID: <200908061538.n76FcbdP004432@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4688


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-06 10:38:36 ---
This is fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083751.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 Aug  6 11:49:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 11:49:35 -0500
Subject: [LLVMbugs] [Bug 4638] Missing simplification of calls
In-Reply-To: <bug-4638-206@http.llvm.org/bugs/>
Message-ID: <200908061649.n76GnZrc007472@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4638


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #7 from Chris Lattner <clattner at apple.com>  2009-08-06 11:49:34 ---
The front-ends shouldn't be in the business of deciding whether they can cheat
on calling conventions.  We expect the optimizers to clean things up after
them.  It looks like it is happening in this case, so closing as 'works
correctly'.  Feel free to file optimizer enhancement requests if other cases
are not handled so 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  Thu Aug  6 14:54:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 14:54:37 -0500
Subject: [LLVMbugs] [Bug 4693] New: arm llvm-gcc build missing symbol
 ___aeabi_cdcmple but __aeabi_cdcmple is present
Message-ID: <bug-4693-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4693

           Summary: arm llvm-gcc build missing symbol ___aeabi_cdcmple but
                    __aeabi_cdcmple is present
           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: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Building llvm-gcc on armv7 linux natively makes it to stage two after
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083768.html
is applied. Once there, we get this problem in configure:

configure:2118: checking for C compiler default output file name
configure:2121:  /home/nlewycky/llvm-gcc/build/./prev-gcc/xgcc
-B/home/nlewycky/llvm-gcc/build/./prev-gcc/
-B/home/nlewycky/llvm-gcc/install/armv7l-unknown-linux-gnueabi/bin/ -g -O2  
conftest.c  >&5
/home/nlewycky/llvm-gcc/build/./prev-gcc/cc1: symbol lookup error:
/home/nlewycky/llvm-gcc/build/./prev-gcc/libgcc_s.so.1: undefined symbol:
___aeabi_cdcmple
configure:2124: $? = 1
[snip trivial program]

Checking libgcc_s.so.1 with nm, we see:
         U ___aeabi_cdcmple
         U ___aeabi_cdrcmple
         U ___aeabi_cfcmple
         U ___aeabi_cfrcmple
         U ___cmpdf2
         U ___cmpsf2
[...]
0000316c T __aeabi_cdcmpeq
0000316c T __aeabi_cdcmple
0000315c T __aeabi_cdrcmple
00003878 T __aeabi_cfcmpeq
00003878 T __aeabi_cfcmple
00003870 T __aeabi_cfrcmple
0000329c T __aeabi_d2f
0000320c T __aeabi_d2iz
00005fa8 T __aeabi_d2lz
[many many more defined]

The difference is that the undefined symbols have three underscores while the
defined ones have two. How'd that happen?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 18:01:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 18:01:49 -0500
Subject: [LLVMbugs] [Bug 4626] UNREACHABLE executed at BranchFolding.cpp:465!
In-Reply-To: <bug-4626-206@http.llvm.org/bugs/>
Message-ID: <200908062301.n76N1nq3023635@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4626


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #5 from Dale Johannesen <dalej at apple.com>  2009-08-06 18:01:48 ---
Fixed here.
http://llvm.org/viewvc/llvm-project?rev=78351&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 19:04:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 19:04:52 -0500
Subject: [LLVMbugs] [Bug 4668] "Ran out of registers during register
 allocation" on x86-32 with asm using only 6 registers
In-Reply-To: <bug-4668-206@http.llvm.org/bugs/>
Message-ID: <200908070004.n7704qSP031455@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4668


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>  2009-08-06 19:04:52 ---
This works for me (not my doing).  I've put the example in the testsuite to
keep it from regressing; this is a fragile area.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 19:35:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 19:35:14 -0500
Subject: [LLVMbugs] [Bug 4659] thumb-2 emits invalid instruction sub.w sp, rX,
	#xx
In-Reply-To: <bug-4659-206@http.llvm.org/bugs/>
Message-ID: <200908070035.n770ZE35002893@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4659


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

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




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-08-06 19:35:14 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083797.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 Aug  6 19:35:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 19:35:25 -0500
Subject: [LLVMbugs] [Bug 4682] Invalid assembler emitted for some thumb
	instructions
In-Reply-To: <bug-4682-206@http.llvm.org/bugs/>
Message-ID: <200908070035.n770ZPWm002938@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4682


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

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




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-08-06 19:35:24 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083797.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 Aug  6 20:33:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 20:33:10 -0500
Subject: [LLVMbugs] [Bug 4696] New: [meta] The AuroraUX source tree does not
 fully compile with Clang
Message-ID: <bug-4696-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4696

           Summary: [meta] The AuroraUX source tree does not fully compile
                    with Clang
           Product: clang
           Version: unspecified
          Platform: PC
               URL: http://auroraux.org
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu
        Depends on: 3679,4672,4690,4695


I'm creating this bug report, so we can keep track of all the reports that need
to be resolved before we can fully compile AuroraUX with Clang.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  6 23:17:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Aug 2009 23:17:25 -0500
Subject: [LLVMbugs] [Bug 4668] "Ran out of registers during register
 allocation" on x86-32 with asm using only 6 registers
In-Reply-To: <bug-4668-206@http.llvm.org/bugs/>
Message-ID: <200908070417.n774HPBG031486@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4668


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-06 23:17:24 ---
The original is still breaking; I'll attach it in a moment.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  7 02:58:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Aug 2009 02:58:43 -0500
Subject: [LLVMbugs] [Bug 4686] Another 2 regscavenger assertions
In-Reply-To: <bug-4686-206@http.llvm.org/bugs/>
Message-ID: <200908070758.n777whrB005304@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4686


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

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




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-08-07 02:58:41 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090803/083812.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 Aug  7 11:21:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Aug 2009 11:21:00 -0500
Subject: [LLVMbugs] [Bug 3333] No warning when shifting invalid amount
In-Reply-To: <bug-3333-206@http.llvm.org/bugs/>
Message-ID: <200908071621.n77GL0C4007638@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3333


Ryan Flynn <parseerror at gmail.com> changed:

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




--- Comment #1 from Ryan Flynn <parseerror at gmail.com>  2009-08-07 11:20:59 ---
r78385


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  7 13:57:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Aug 2009 13:57:41 -0500
Subject: [LLVMbugs] [Bug 4667] LoopIndexSplit problem with
	ENABLE_EXPENSIVE_CHECKS
In-Reply-To: <bug-4667-206@http.llvm.org/bugs/>
Message-ID: <200908071857.n77IvfUC027986@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4667


devang.patel <devang.patel at gmail.com> changed:

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




--- Comment #2 from devang.patel <devang.patel at gmail.com>  2009-08-07 13:57:41 ---
Jakob fixed 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  Fri Aug  7 20:42:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Aug 2009 20:42:07 -0500
Subject: [LLVMbugs] [Bug 3444] unterminated switch statement crashes clang
In-Reply-To: <bug-3444-206@http.llvm.org/bugs/>
Message-ID: <200908080142.n781g7Bm013920@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3444


Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dgregor at apple.com
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #14 from Douglas Gregor <dgregor at apple.com>  2009-08-07 20:42:06 ---
Fixed in r78450.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  7 23:25:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Aug 2009 23:25:42 -0500
Subject: [LLVMbugs] [Bug 3156] Couldn't allocate output reg for constraint
 'r' (x86_32/OSX)
In-Reply-To: <bug-3156-206@http.llvm.org/bugs/>
Message-ID: <200908080425.n784PgR7002118@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3156


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

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




--- Comment #6 from Evan Cheng <evan.cheng at apple.com>  2009-08-07 23:25:41 ---
Must have been fixed a while 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  Sat Aug  8 11:35:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Aug 2009 11:35:35 -0500
Subject: [LLVMbugs] [Bug 4699] New: Assertion failed: (N->use_empty() && "
 Cannot delete a node that is not dead!"), function DeleteNodeNotInCSEMaps,
 file SelectionDAG.cpp, line 594.
Message-ID: <bug-4699-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4699

           Summary: Assertion failed: (N->use_empty() && "Cannot delete a
                    node that is not dead!"), function
                    DeleteNodeNotInCSEMaps, file SelectionDAG.cpp, line 594.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: benny.kra at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3296)
 --> (http://llvm.org/bugs/attachment.cgi?id=3296)
reduced testcase

Assertion failed: (N->use_empty() && "Cannot delete a node that is not dead!"),
function DeleteNodeNotInCSEMaps, file SelectionDAG.cpp, line 594.

Program received signal SIGABRT, Aborted.
0x96ba2e42 in __kill ()
(gdb) bt
#0  0x96ba2e42 in __kill ()
#1  0x96ba2e34 in kill$UNIX2003 ()
#2  0x96c1523a in raise ()
#3  0x96c21679 in abort ()
#4  0x96c163db in __assert_rtn ()
#5  0x0041a2d3 in llvm::SelectionDAG::DeleteNodeNotInCSEMaps (this=0x15059c0,
N=0x1815228) at SelectionDAG.cpp:594
#6  0x0042a554 in llvm::SelectionDAG::DeleteNode (this=0x15059c0, N=0x1815228)
at SelectionDAG.cpp:589
#7  0x003915e4 in (anonymous namespace)::DAGCombiner::visitMERGE_VALUES
(this=0xbfffedac, N=0x1815228) at DAGCombiner.cpp:931
#8  0x003b42f7 in (anonymous namespace)::DAGCombiner::visit (this=0xbfffedac,
N=0x1815228) at DAGCombiner.cpp:729
#9  0x003b49c9 in (anonymous namespace)::DAGCombiner::combine (this=0xbfffedac,
N=0x1815228) at DAGCombiner.cpp:793
#10 0x003b5f64 in (anonymous namespace)::DAGCombiner::Run (this=0xbfffedac,
AtLevel=llvm::Unrestricted) at DAGCombiner.cpp:667
#11 0x003b63ac in llvm::SelectionDAG::Combine (this=0x15059c0,
Level=llvm::Unrestricted, AA=@0x1505f5c, OptLevel=llvm::CodeGenOpt::Default) at
DAGCombiner.cpp:6239
#12 0x0047fa57 in llvm::SelectionDAGISel::CodeGenAndEmitDAG (this=0x15058b0) at
SelectionDAGISel.cpp:480
#13 0x004821ec in llvm::SelectionDAGISel::SelectBasicBlock (this=0x15058b0,
LLVMBB=0x1502640, Begin={<bidirectional_iterator<llvm::Instruction,ptrdiff_t>>
= {<> = {<No data fields>}, <No data fields>}, NodePtr = 0x150b07c},
End={<bidirectional_iterator<llvm::Instruction,ptrdiff_t>> = {<> = {<No data
fields>}, <No data fields>}, NodePtr = 0x1502640}) at SelectionDAGISel.cpp:400
#14 0x00482abf in llvm::SelectionDAGISel::SelectAllBasicBlocks (this=0x15058b0,
Fn=@0x1501d70, MF=@0x150b0e0, MMI=0x1509200, DW=0x1506180, TII=@0x18086f0) at
SelectionDAGISel.cpp:790
#15 0x00482e99 in llvm::SelectionDAGISel::runOnMachineFunction (this=0x15058b0,
mf=@0x150b0e0) at SelectionDAGISel.cpp:334
#16 0x00546835 in llvm::MachineFunctionPass::runOnFunction (this=0x15058b0,
F=@0x1501d70) at MachineFunctionPass.cpp:31
#17 0x007707a0 in llvm::FPPassManager::runOnFunction (this=0x1501b40,
F=@0x1501d70) at PassManager.cpp:1372
#18 0x00770a10 in llvm::FunctionPassManagerImpl::run (this=0x1501960,
F=@0x1501d70) at PassManager.cpp:1324
#19 0x00770b24 in llvm::FunctionPassManager::run (this=0xbffff1dc,
F=@0x1501d70) at PassManager.cpp:1254
#20 0x000033ba in main (argc=2, argv=0xbffff2dc) at llc.cpp:394


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  8 13:24:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Aug 2009 13:24:59 -0500
Subject: [LLVMbugs] [Bug 4700] New: "operation has no effect" for macros
 that expand to 0 << 1
Message-ID: <bug-4700-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4700

           Summary: "operation has no effect" for macros that expand to 0 <<
                    1
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ssen at apple.com
                CC: llvmbugs at cs.uiuc.edu


llvm/clang r78480

The following is typical for code that generates enums or error codes where the
top bits are a subsystem ID and the low bits are an individual error code:

[ssen at virgon]$ cat foo.c
#define errcode(x, y) (x << 16 | y)
int a = errcode(0, 0);
int b = errcode(1, 0);
[ssen at virgon]$ clang -c foo.c -Wall
foo.c:2:9: warning: operation has no effect [-Wall]
int a = errcode(0, 0);
        ^~~~~~~~~~~~~
foo.c:1:26: note: instantiated from:
#define errcode(x, y) (x << 16 | y)
                         ^  ~~
1 diagnostic generated.
[ssen at virgon]$ 

This code should not warn. It's expected behavior that sometimes there will be
0 in the top bits.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug  8 14:19:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Aug 2009 14:19:22 -0500
Subject: [LLVMbugs] [Bug 4700] "operation has no effect" for macros that
 expand to 0 << 1
In-Reply-To: <bug-4700-206@http.llvm.org/bugs/>
Message-ID: <200908081919.n78JJMdn031002@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4700


Ryan Flynn <parseerror at gmail.com> changed:

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




--- Comment #6 from Ryan Flynn <parseerror at gmail.com>  2009-08-08 14:19:22 ---
r78488


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  8 14:56:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Aug 2009 14:56:21 -0500
Subject: [LLVMbugs] [Bug 4701] New: Definitions of id and Class ignored
Message-ID: <bug-4701-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4701

           Summary: Definitions of id and Class ignored
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: csdavec at swan.ac.uk
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3297)
 --> (http://llvm.org/bugs/attachment.cgi?id=3297)
Test case.

Accessing fields in id and Class no longer works.  Any definition of these
types in code is now completely ignored.  This breaks any code on the GNU
runtime that includes objc/objc-api.h.  

The attached file is the result of preprocessing this trivial program on a
platform with the GNU platform:

#import <objc/objc.h>

const void *testSel(SEL a)
{
    return a->sel_types;
}
const void *testClass(Class a)
{
    return a->name;
}
const void *testId(id a)
{
    return a->class_pointer;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  8 22:37:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Aug 2009 22:37:30 -0500
Subject: [LLVMbugs] [Bug 4702] New: "make check" fails,
	claiming lack of dejagnu
Message-ID: <bug-4702-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4702

           Summary: "make check" fails, claiming lack of dejagnu
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jgarzik at pobox.com
                CC: llvmbugs at cs.uiuc.edu


In current SVN (r78520), attempting to run "make check" on Fedora 10/x86-64
fails, even though dejagnu is installed:


[jgarzik at bd llvm]$ make check
llvm[0]: Running test suite
make[1]: Entering directory `/garz/repo/llvm/test'
Making a new site.exp file...
*** dejagnu not found.  Make sure 'runtest' is in your PATH, then reconfigure
LLVM.
make[1]: Leaving directory `/garz/repo/llvm/test'


[jgarzik at bd llvm]$ rpm -q dejagnu
dejagnu-1.4.4-13.fc10.noarch


[jgarzik at bd llvm]$ which runtest
/usr/bin/runtest


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug  9 02:02:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Aug 2009 02:02:39 -0500
Subject: [LLVMbugs] [Bug 4702] "make check" fails, claiming lack of dejagnu
In-Reply-To: <bug-4702-206@http.llvm.org/bugs/>
Message-ID: <200908090702.n7972ddN022227@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4702


Bill Wendling <wendling at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |wendling at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Bill Wendling <wendling at apple.com>  2009-08-09 02:02:39 ---
Check the config.log to determine why it didn't find the "runtest" during
configuration. If it wasn't found, re-configure with your PATH set to where
runtest is.

We really can't help with 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 Aug  9 18:40:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Aug 2009 18:40:43 -0500
Subject: [LLVMbugs] [Bug 4704] New: cast from ptr to scalar crashes clang
Message-ID: <bug-4704-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4704

           Summary: cast from ptr to scalar crashes clang
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


recent regression when generating assembly for the following code:

// clang-cc -S -O2
void f() {
        int a = (int)("") - 0;
        int b = (int)(""); // possibly different bug
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 10 11:54:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 11:54:23 -0500
Subject: [LLVMbugs] [Bug 4618] AliasSetTracker needs to use CallbackVH/
 WeakVH or be updated more correctly by loop passes
In-Reply-To: <bug-4618-206@http.llvm.org/bugs/>
Message-ID: <200908101654.n7AGsN2f002472@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4618


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #3 from Dan Gohman <gohman at apple.com>  2009-08-10 11:54:21 ---
This is fixed in r77623-r77696.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 10 14:11:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 14:11:31 -0500
Subject: [LLVMbugs] [Bug 4705] New: msp430: llc : Assertion failed: (Reg &&
 " this is not a register!")
Message-ID: <bug-4705-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4705

           Summary: msp430: llc : Assertion failed: (Reg && "this is not a
                    register!")
           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: karthiksun at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3300)
 --> (http://llvm.org/bugs/attachment.cgi?id=3300)
bugpoint -run-llc foo.bc --tool-args -march=msp430

Building a simple.c file with llvm r78578 to generate msp430 assembly file.

The command used is 
   llc -f -march=msp430 -o foo.s foo.bc

Crashes llc with 
   Assertion failed: (Reg && "this is not a register!").

This is probably related to bug #4587


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 10 16:15:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 16:15:13 -0500
Subject: [LLVMbugs] [Bug 4705] msp430: llc : Assertion failed: (Reg && "this
 is not a register! ")
In-Reply-To: <bug-4705-206@http.llvm.org/bugs/>
Message-ID: <200908102115.n7ALFDG8007211@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4705


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-10 16:15:12 ---
This file is not suitable for msp430 at all: it contains targetdata and all
stuff from 32-bit CPU


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 10 16:41:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 16:41:24 -0500
Subject: [LLVMbugs] [Bug 4705] msp430: llc : Assertion failed: (Reg && "this
 is not a register! ")
In-Reply-To: <bug-4705-206@http.llvm.org/bugs/>
Message-ID: <200908102141.n7ALfO1D010869@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4705


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-08-10 16:41:23 ---
LLVM IR emitted by a C front-end is target-specific, in general.

However, I don't see anything in the given testcase that should preclude it
from at least being codegen'd for msp430 or any other target, so this does seem
to be a CodeGen bug.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 10 19:30:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 19:30:23 -0500
Subject: [LLVMbugs] [Bug 4706] New: Bitcode reader fails after rev# 78258
Message-ID: <bug-4706-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4706

           Summary: Bitcode reader fails after rev# 78258
           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: gareth_morgan77 at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3301)
 --> (http://llvm.org/bugs/attachment.cgi?id=3301)
IR file that will crash bitcode reader when compiled and read

The attached IR, when compiled to bitcode using LLVM-AS, will crash the
BitcodeReader when the BC is read.  If the source tree is rolled back to 78257
will work fine, but at 78258 or later it will cause an assert here:
       
LLVMTest.exe!llvm::TypeMap<llvm::StructValType,llvm::StructType>::RefineAbstractType(llvm::StructType
* Ty=0x01aed0d0, const llvm::DerivedType * OldType=0x01aeb328, const llvm::Type
* NewType=0x01aeb290)  Line 323 + 0x26 bytes      C++
        LLVMTest.exe!llvm::StructType::refineAbstractType(const
llvm::DerivedType * OldType=0x01aeb328, const llvm::Type * NewType=0x01aeb290) 
Line 1152       C++
        LLVMTest.exe!llvm::DerivedType::unlockedRefineAbstractTypeTo(const
llvm::Type * NewType=0x01aeb290)  Line 1055 + 0x1c bytes     C++
        LLVMTest.exe!llvm::DerivedType::refineAbstractTypeTo(const llvm::Type *
NewType=0x01aeb290)  Line 1076  C++
        LLVMTest.exe!llvm::BitcodeReader::ParseTypeTable()  Line 624    C++
        LLVMTest.exe!llvm::BitcodeReader::ParseModule(const
std::basic_string<char,std::char_traits<char>,std::allocator<char> > &
ModuleID="C:\svn_trunk\Code\Tools\llvm-gcc-helpers\TS.bc")  Line 1270 + 0x8
bytes    C++
        LLVMTest.exe!llvm::BitcodeReader::ParseBitcode()  Line 1521 + 0x42
bytes        C++
        LLVMTest.exe!llvm::getBitcodeModuleProvider(llvm::MemoryBuffer *
Buffer=0x01ae7928, llvm::LLVMContext & Context={...},
std::basic_string<char,std::char_traits<char>,std::allocator<char> > *
ErrMsg="")  Line 2222 + 0x8 bytes C++
        LLVMTest.exe!llvm::ParseBitcodeFile(llvm::MemoryBuffer *
Buffer=0x01ae7928, llvm::LLVMContext & Context={...},
std::basic_string<char,std::char_traits<char>,std::allocator<char> > *
ErrMsg="")  Line 2240 + 0x11 bytes        C++


The code for reading it looks like this:
        memBuf = llvm::MemoryBuffer::getFile("TS.bc");

        std::string errStr;
        mainModule = llvm::ParseBitcodeFile(memBuf, LLVMGetContext(), &errStr);
        if(NULL == mainModule)
        {
                printf("Module errStr: %s\n", errStr.c_str());
                exit(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 Aug 10 19:33:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Aug 2009 19:33:26 -0500
Subject: [LLVMbugs] [Bug 4675] lli: symbol lookup error:
 /usr/lib/oprofile/libopagent.so.1: undefined symbol: bfd_init
In-Reply-To: <bug-4675-206@http.llvm.org/bugs/>
Message-ID: <200908110033.n7B0XQ6I002600@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4675


Jeffrey Yasskin <jyasskin at google.com> changed:

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




--- Comment #3 from Jeffrey Yasskin <jyasskin at google.com>  2009-08-10 19:33:25 ---
Nick and I think this is a problem with the Debian oprofile package he
installed, and not with the way libopagent is linked in LLVM.  He found
http://bugs.debian.org/537744 explaining the Debian problem.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 11 14:36:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 14:36:25 -0500
Subject: [LLVMbugs] [Bug 4707] New: Use after free in thumb2 tests
Message-ID: <bug-4707-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4707

           Summary: Use after free in thumb2 tests
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: benny.kra at gmail.com
                CC: llvmbugs at cs.uiuc.edu


the following tests show errors in valgrind:
test/CodeGen/Thumb2/thumb2-ldr.ll
test/CodeGen/Thumb2/thumb2-ldrb.ll
test/CodeGen/Thumb2/thumb2-ldrh.ll
test/CodeGen/Thumb2/thumb2-mov3.ll
test/CodeGen/Thumb2/thumb2-str.ll
test/CodeGen/Thumb2/thumb2-strb.ll
test/CodeGen/Thumb2/thumb2-strh.ll

example valgrind log (thumb2-str.ll) on osx 10.5.8:
 Invalid read of size 1
    at 0x3B183D: llvm::MachineOperand::isReg() const (MachineOperand.h:144)
    by 0x388A64: UpdateCPSRLiveness(llvm::MachineInstr&, bool)
(Thumb2SizeReduction.cpp:446)
    by 0x38A087: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:513)
    by 0x38A17E: (anonymous
namespace)::Thumb2SizeReduce::runOnMachineFunction(llvm::MachineFunction&)
(Thumb2SizeReduction.cpp:528)
    by 0x579D74: llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(MachineFunctionPass.cpp:31)
    by 0x7AA35B: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1372)
    by 0x7AA5CB: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1324)
    by 0x7AA6DF: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1254)
    by 0x33B9: main (llc.cpp:394)
  Address 0x1966dd0 is 0 bytes inside a block of size 120 free'd
    at 0x1598B1A: operator delete(void*) (vg_replace_malloc.c:346)
    by 0x52A292:
__gnu_cxx::new_allocator<llvm::MachineOperand>::deallocate(llvm::MachineOperand*,
unsigned long) (ext/new_allocator.h:94)
    by 0x52A2BC: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::_M_deallocate(llvm::MachineOperand*,
unsigned long) (stl_vector.h:123)
    by 0x52A2FB: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~_Vector_base() (stl_vector.h:109)
    by 0x52A372: std::vector<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~vector() (stl_vector.h:273)
    by 0x57CD76: llvm::MachineInstr::~MachineInstr() (MachineInstr.cpp:438)
    by 0x5727A4: llvm::MachineFunction::DeleteMachineInstr(llvm::MachineInstr*)
(MachineFunction.cpp:198)
    by 0x568092:
llvm::ilist_traits<llvm::MachineInstr>::deleteNode(llvm::MachineInstr*)
(MachineBasicBlock.cpp:120)
    by 0x4B502E: llvm::iplist<llvm::MachineInstr,
llvm::ilist_traits<llvm::MachineInstr>
>::erase(llvm::ilist_iterator<llvm::MachineInstr>) (ilist.h:463)
    by 0x4B504E:
llvm::MachineBasicBlock::erase(llvm::ilist_iterator<llvm::MachineInstr>)
(MachineBasicBlock.h:272)
    by 0x3892D4: (anonymous
namespace)::Thumb2SizeReduce::ReduceToNarrow(llvm::MachineBasicBlock&,
llvm::MachineInstr*, (anonymous namespace)::ReduceEntry const&, bool)
(Thumb2SizeReduction.cpp:437)
    by 0x38A057: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:508)

 Invalid read of size 1
    at 0x3B183D: llvm::MachineOperand::isReg() const (MachineOperand.h:144)
    by 0x5427E9: llvm::MachineOperand::isUndef() const (MachineOperand.h:205)
    by 0x388A76: UpdateCPSRLiveness(llvm::MachineInstr&, bool)
(Thumb2SizeReduction.cpp:446)
    by 0x38A087: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:513)
    by 0x38A17E: (anonymous
namespace)::Thumb2SizeReduce::runOnMachineFunction(llvm::MachineFunction&)
(Thumb2SizeReduction.cpp:528)
    by 0x579D74: llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(MachineFunctionPass.cpp:31)
    by 0x7AA35B: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1372)
    by 0x7AA5CB: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1324)
    by 0x7AA6DF: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1254)
    by 0x33B9: main (llc.cpp:394)
  Address 0x1966dd0 is 0 bytes inside a block of size 120 free'd
    at 0x1598B1A: operator delete(void*) (vg_replace_malloc.c:346)
    by 0x52A292:
__gnu_cxx::new_allocator<llvm::MachineOperand>::deallocate(llvm::MachineOperand*,
unsigned long) (ext/new_allocator.h:94)
    by 0x52A2BC: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::_M_deallocate(llvm::MachineOperand*,
unsigned long) (stl_vector.h:123)
    by 0x52A2FB: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~_Vector_base() (stl_vector.h:109)
    by 0x52A372: std::vector<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~vector() (stl_vector.h:273)
    by 0x57CD76: llvm::MachineInstr::~MachineInstr() (MachineInstr.cpp:438)
    by 0x5727A4: llvm::MachineFunction::DeleteMachineInstr(llvm::MachineInstr*)
(MachineFunction.cpp:198)
    by 0x568092:
llvm::ilist_traits<llvm::MachineInstr>::deleteNode(llvm::MachineInstr*)
(MachineBasicBlock.cpp:120)
    by 0x4B502E: llvm::iplist<llvm::MachineInstr,
llvm::ilist_traits<llvm::MachineInstr>
>::erase(llvm::ilist_iterator<llvm::MachineInstr>) (ilist.h:463)
    by 0x4B504E:
llvm::MachineBasicBlock::erase(llvm::ilist_iterator<llvm::MachineInstr>)
(MachineBasicBlock.h:272)
    by 0x3892D4: (anonymous
namespace)::Thumb2SizeReduce::ReduceToNarrow(llvm::MachineBasicBlock&,
llvm::MachineInstr*, (anonymous namespace)::ReduceEntry const&, bool)
(Thumb2SizeReduction.cpp:437)
    by 0x38A057: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:508)

 Invalid read of size 4
    at 0x542830: llvm::MachineOperand::isUndef() const (MachineOperand.h:206)
    by 0x388A76: UpdateCPSRLiveness(llvm::MachineInstr&, bool)
(Thumb2SizeReduction.cpp:446)
    by 0x38A087: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:513)
    by 0x38A17E: (anonymous
namespace)::Thumb2SizeReduce::runOnMachineFunction(llvm::MachineFunction&)
(Thumb2SizeReduction.cpp:528)
    by 0x579D74: llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(MachineFunctionPass.cpp:31)
    by 0x7AA35B: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1372)
    by 0x7AA5CB: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1324)
    by 0x7AA6DF: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1254)
    by 0x33B9: main (llc.cpp:394)
  Address 0x1966dd0 is 0 bytes inside a block of size 120 free'd
    at 0x1598B1A: operator delete(void*) (vg_replace_malloc.c:346)
    by 0x52A292:
__gnu_cxx::new_allocator<llvm::MachineOperand>::deallocate(llvm::MachineOperand*,
unsigned long) (ext/new_allocator.h:94)
    by 0x52A2BC: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::_M_deallocate(llvm::MachineOperand*,
unsigned long) (stl_vector.h:123)
    by 0x52A2FB: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~_Vector_base() (stl_vector.h:109)
    by 0x52A372: std::vector<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~vector() (stl_vector.h:273)
    by 0x57CD76: llvm::MachineInstr::~MachineInstr() (MachineInstr.cpp:438)
    by 0x5727A4: llvm::MachineFunction::DeleteMachineInstr(llvm::MachineInstr*)
(MachineFunction.cpp:198)
    by 0x568092:
llvm::ilist_traits<llvm::MachineInstr>::deleteNode(llvm::MachineInstr*)
(MachineBasicBlock.cpp:120)
    by 0x4B502E: llvm::iplist<llvm::MachineInstr,
llvm::ilist_traits<llvm::MachineInstr>
>::erase(llvm::ilist_iterator<llvm::MachineInstr>) (ilist.h:463)
    by 0x4B504E:
llvm::MachineBasicBlock::erase(llvm::ilist_iterator<llvm::MachineInstr>)
(MachineBasicBlock.h:272)
    by 0x3892D4: (anonymous
namespace)::Thumb2SizeReduce::ReduceToNarrow(llvm::MachineBasicBlock&,
llvm::MachineInstr*, (anonymous namespace)::ReduceEntry const&, bool)
(Thumb2SizeReduction.cpp:437)
    by 0x38A057: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:508)

 Invalid read of size 1
    at 0x3B183D: llvm::MachineOperand::isReg() const (MachineOperand.h:144)
    by 0x37026B: llvm::MachineOperand::getReg() const (MachineOperand.h:170)
    by 0x388A97: UpdateCPSRLiveness(llvm::MachineInstr&, bool)
(Thumb2SizeReduction.cpp:448)
    by 0x38A087: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:513)
    by 0x38A17E: (anonymous
namespace)::Thumb2SizeReduce::runOnMachineFunction(llvm::MachineFunction&)
(Thumb2SizeReduction.cpp:528)
    by 0x579D74: llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(MachineFunctionPass.cpp:31)
    by 0x7AA35B: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1372)
    by 0x7AA5CB: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1324)
    by 0x7AA6DF: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1254)
    by 0x33B9: main (llc.cpp:394)
  Address 0x1966dd0 is 0 bytes inside a block of size 120 free'd
    at 0x1598B1A: operator delete(void*) (vg_replace_malloc.c:346)
    by 0x52A292:
__gnu_cxx::new_allocator<llvm::MachineOperand>::deallocate(llvm::MachineOperand*,
unsigned long) (ext/new_allocator.h:94)
    by 0x52A2BC: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::_M_deallocate(llvm::MachineOperand*,
unsigned long) (stl_vector.h:123)
    by 0x52A2FB: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~_Vector_base() (stl_vector.h:109)
    by 0x52A372: std::vector<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~vector() (stl_vector.h:273)
    by 0x57CD76: llvm::MachineInstr::~MachineInstr() (MachineInstr.cpp:438)
    by 0x5727A4: llvm::MachineFunction::DeleteMachineInstr(llvm::MachineInstr*)
(MachineFunction.cpp:198)
    by 0x568092:
llvm::ilist_traits<llvm::MachineInstr>::deleteNode(llvm::MachineInstr*)
(MachineBasicBlock.cpp:120)
    by 0x4B502E: llvm::iplist<llvm::MachineInstr,
llvm::ilist_traits<llvm::MachineInstr>
>::erase(llvm::ilist_iterator<llvm::MachineInstr>) (ilist.h:463)
    by 0x4B504E:
llvm::MachineBasicBlock::erase(llvm::ilist_iterator<llvm::MachineInstr>)
(MachineBasicBlock.h:272)
    by 0x3892D4: (anonymous
namespace)::Thumb2SizeReduce::ReduceToNarrow(llvm::MachineBasicBlock&,
llvm::MachineInstr*, (anonymous namespace)::ReduceEntry const&, bool)
(Thumb2SizeReduction.cpp:437)
    by 0x38A057: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:508)

 Invalid read of size 4
    at 0x3702B2: llvm::MachineOperand::getReg() const (MachineOperand.h:171)
    by 0x388A97: UpdateCPSRLiveness(llvm::MachineInstr&, bool)
(Thumb2SizeReduction.cpp:448)
    by 0x38A087: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:513)
    by 0x38A17E: (anonymous
namespace)::Thumb2SizeReduce::runOnMachineFunction(llvm::MachineFunction&)
(Thumb2SizeReduction.cpp:528)
    by 0x579D74: llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
(MachineFunctionPass.cpp:31)
    by 0x7AA35B: llvm::FPPassManager::runOnFunction(llvm::Function&)
(PassManager.cpp:1372)
    by 0x7AA5CB: llvm::FunctionPassManagerImpl::run(llvm::Function&)
(PassManager.cpp:1324)
    by 0x7AA6DF: llvm::FunctionPassManager::run(llvm::Function&)
(PassManager.cpp:1254)
    by 0x33B9: main (llc.cpp:394)
  Address 0x1966dd8 is 8 bytes inside a block of size 120 free'd
    at 0x1598B1A: operator delete(void*) (vg_replace_malloc.c:346)
    by 0x52A292:
__gnu_cxx::new_allocator<llvm::MachineOperand>::deallocate(llvm::MachineOperand*,
unsigned long) (ext/new_allocator.h:94)
    by 0x52A2BC: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::_M_deallocate(llvm::MachineOperand*,
unsigned long) (stl_vector.h:123)
    by 0x52A2FB: std::_Vector_base<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~_Vector_base() (stl_vector.h:109)
    by 0x52A372: std::vector<llvm::MachineOperand,
std::allocator<llvm::MachineOperand> >::~vector() (stl_vector.h:273)
    by 0x57CD76: llvm::MachineInstr::~MachineInstr() (MachineInstr.cpp:438)
    by 0x5727A4: llvm::MachineFunction::DeleteMachineInstr(llvm::MachineInstr*)
(MachineFunction.cpp:198)
    by 0x568092:
llvm::ilist_traits<llvm::MachineInstr>::deleteNode(llvm::MachineInstr*)
(MachineBasicBlock.cpp:120)
    by 0x4B502E: llvm::iplist<llvm::MachineInstr,
llvm::ilist_traits<llvm::MachineInstr>
>::erase(llvm::ilist_iterator<llvm::MachineInstr>) (ilist.h:463)
    by 0x4B504E:
llvm::MachineBasicBlock::erase(llvm::ilist_iterator<llvm::MachineInstr>)
(MachineBasicBlock.h:272)
    by 0x3892D4: (anonymous
namespace)::Thumb2SizeReduce::ReduceToNarrow(llvm::MachineBasicBlock&,
llvm::MachineInstr*, (anonymous namespace)::ReduceEntry const&, bool)
(Thumb2SizeReduction.cpp:437)
    by 0x38A057: (anonymous
namespace)::Thumb2SizeReduce::ReduceMBB(llvm::MachineBasicBlock&)
(Thumb2SizeReduction.cpp:508)

This might be the cause for test failures on solaris.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 11 14:41:12 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 14:41:12 -0500
Subject: [LLVMbugs] [Bug 4708] New: llvm-prof has cast warnings after r78477
Message-ID: <bug-4708-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4708

           Summary: llvm-prof has cast warnings after r78477
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Global Analyses
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: idadesub at users.sourceforge.net
                CC: llvmbugs at cs.uiuc.edu, daniel at zuster.org


Created an attachment (id=3303)
 --> (http://llvm.org/bugs/attachment.cgi?id=3303)
patch to fix warnings in llvm-prof.cpp

It looks like r78477 changed a couple of the return types for ProfileInfo to
return doubles instead of unsigned ints, and llvm-prof.cpp wasn't updated to
reflect this. I'm now getting a couple cast warnings from this (I know
gcc-4.1.1 is bad, but the warnings appear to be valid):

/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_pair.h:
In constructor ???std::pair<_T1, _T2>::pair(const std::pair<_U1, _U2>&) [with
_U1 = llvm::Function*, _U2 = double, _T1 = llvm::Function*, _T2 = unsigned
int]???:
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2477:
  instantiated from ???void std::partial_sort(_RandomAccessIterator,
_RandomAccessIterator, _RandomAccessIterator, _Compare) [with
_RandomAccessIterator = __gnu_cxx::__normal_iterator<std::pair<llvm::Function*,
double>*, std::vector<std::pair<llvm::Function*, double>,
std::allocator<std::pair<llvm::Function*, double> > > >, _Compare =
PairSecondSortReverse<llvm::Function*>]???
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2666:
  instantiated from ???void std::__introsort_loop(_RandomAccessIterator,
_RandomAccessIterator, _Size, _Compare) [with _RandomAccessIterator =
__gnu_cxx::__normal_iterator<std::pair<llvm::Function*, double>*,
std::vector<std::pair<llvm::Function*, double>,
std::allocator<std::pair<llvm::Function*, double> > > >, _Size = long int,
_Compare = PairSecondSortReverse<llvm::Function*>]???
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2749:
  instantiated from ???void std::sort(_RandomAccessIterator,
_RandomAccessIterator, _Compare) [with _RandomAccessIterator =
__gnu_cxx::__normal_iterator<std::pair<llvm::Function*, double>*,
std::vector<std::pair<llvm::Function*, double>,
std::allocator<std::pair<llvm::Function*, double> > > >, _Compare =
PairSecondSortReverse<llvm::Function*>]???
/scratch/llvm/llvm/tools/llvm-prof/llvm-prof.cpp:169:   instantiated from here
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_pair.h:90:
warning: converting to ???unsigned int??? from ???const double???
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_pair.h:
In constructor ???std::pair<_T1, _T2>::pair(const std::pair<_U1, _U2>&) [with
_U1 = llvm::BasicBlock*, _U2 = double, _T1 = llvm::BasicBlock*, _T2 = unsigned
int]???:
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2477:
  instantiated from ???void std::partial_sort(_RandomAccessIterator,
_RandomAccessIterator, _RandomAccessIterator, _Compare) [with
_RandomAccessIterator =
__gnu_cxx::__normal_iterator<std::pair<llvm::BasicBlock*, double>*,
std::vector<std::pair<llvm::BasicBlock*, double>,
std::allocator<std::pair<llvm::BasicBlock*, double> > > >, _Compare =
PairSecondSortReverse<llvm::BasicBlock*>]???
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2666:
  instantiated from ???void std::__introsort_loop(_RandomAccessIterator,
_RandomAccessIterator, _Size, _Compare) [with _RandomAccessIterator =
__gnu_cxx::__normal_iterator<std::pair<llvm::BasicBlock*, double>*,
std::vector<std::pair<llvm::BasicBlock*, double>,
std::allocator<std::pair<llvm::BasicBlock*, double> > > >, _Size = long int,
_Compare = PairSecondSortReverse<llvm::BasicBlock*>]???
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_algo.h:2749:
  instantiated from ???void std::sort(_RandomAccessIterator,
_RandomAccessIterator, _Compare) [with _RandomAccessIterator =
__gnu_cxx::__normal_iterator<std::pair<llvm::BasicBlock*, double>*,
std::vector<std::pair<llvm::BasicBlock*, double>,
std::allocator<std::pair<llvm::BasicBlock*, double> > > >, _Compare =
PairSecondSortReverse<llvm::BasicBlock*>]???
/scratch/llvm/llvm/tools/llvm-prof/llvm-prof.cpp:212:   instantiated from here
/usr/lib/gcc/x86_64-redhat-linux/4.1.1/../../../../include/c++/4.1.1/bits/stl_pair.h:90:
warning: converting to ???unsigned int??? from ???const double


I've attached a patch that fixes this, but I'm not sure what the consequences
of this patch will do since I've got a bad gcc installed.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 11 14:53:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 14:53:01 -0500
Subject: [LLVMbugs] [Bug 4699] Assertion failed: (N->use_empty() && " Cannot
 delete a node that is not dead!"), function DeleteNodeNotInCSEMaps,
 file SelectionDAG.cpp, line 594.
In-Reply-To: <bug-4699-206@http.llvm.org/bugs/>
Message-ID: <200908111953.n7BJr1kh009477@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4699


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-11 14:53:01 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090810/084091.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 Aug 11 17:42:20 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 17:42:20 -0500
Subject: [LLVMbugs] [Bug 2247] llvm needs an 'exact divide' flag/instruction
In-Reply-To: <bug-2247-206@http.llvm.org/bugs/>
Message-ID: <200908112242.n7BMgKmV000351@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2247


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #9 from Dan Gohman <gohman at apple.com>  2009-08-11 17:42:19 ---
This is now fixed. LLVM's sdiv operator now has an "exact" flag, and
clang and llvm-gcc are now both using it for pointer difference
expressions.  On the original testcase, test and test2 are now
compiled to the same code.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 11 19:34:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 19:34:48 -0500
Subject: [LLVMbugs] [Bug 4710] New: Assert fires in the Verifier when called
 from the PassManager in 'opt'
Message-ID: <bug-4710-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4710

           Summary: Assert fires in the Verifier when called from the
                    PassManager in 'opt'
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: major
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Good day,

/opt/clang/bin/opt asserts and aborts with the following traces below;

Suspected files in question:
 lib/VMCore/Verifier.cpp
 lib/VMCore/PassManager.cpp

== In gdb on Linux ==
Instruction does not dominate all uses!
        %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [ %tmp12, %BB ]       
        ; <i32> [#uses=1]    %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [
%tmp12, %BB ]         ; <i32> [#uses=1]Broken module found, compilation
aborted!
0   opt               0x003c3356 llvm::sys::SetInterruptFunction(void (*)()) +
82
1   opt               0x003c3745 llvm::sys::SetInterruptFunction(void (*)()) +
1089
2   libSystem.B.dylib 0x96ba12bb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1766190447
4   libSystem.B.dylib 0x96c1523a raise + 26
5   libSystem.B.dylib 0x96c21679 abort + 73
6   opt               0x0037a687 llvm::BaseStream<std::ostream>&
llvm::BaseStream<std::ostream>::operator<< <char [2]>(char const (&) [2]) + 351
7   opt               0x0037e09b llvm::InstVisitor<(anonymous
namespace)::Verifier, void>::visit(llvm::Function&) + 185
8   opt               0x003424a0
llvm::FPPassManager::runOnFunction(llvm::Function&) + 304
9   opt               0x00342652
llvm::FPPassManager::runOnModule(llvm::Module&) + 84
10  opt               0x00342089
llvm::MPPassManager::runOnModule(llvm::Module&) + 385
11  opt               0x003422de llvm::PassManagerImpl::run(llvm::Module&) +
112
12  opt               0x00342337 llvm::PassManager::run(llvm::Module&) + 27
13  opt               0x0000d340 main + 4634
14  opt               0x00001eb6 start + 54
Stack dump:
0.      Running pass 'Function Pass Manager' on module '<stdin>'.
1.      Running pass 'Module Verifier' on function '@func_105'

======8<=========

Instruction does not dominate all uses!
        %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [ %tmp12, %BB ]       
        ; <i32> [#uses=1]    %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [
%tmp12, %BB ]         ; <i32> [#uses=1]Broken module found, compilation
aborted!

Program received signal SIGABRT, Aborted.
0x96ba2e42 in __kill ()
(gdb) bt
#0  0x96ba2e42 in __kill ()
#1  0x96ba2e34 in kill$UNIX2003 ()
#2  0x96c1523a in raise ()
#3  0x96c21679 in abort ()
#4  0x0037a687 in (anonymous namespace)::Verifier::abortIfBroken
(this=0xe02480) at Verifier.cpp:216
#5  0x0037e09b in (anonymous namespace)::Verifier::runOnFunction
(this=0xe02480, F=@0xe02790) at Verifier.cpp:171
#6  0x003424a0 in llvm::FPPassManager::runOnFunction (this=0xe021e0,
F=@0xe02790) at PassManager.cpp:1372
#7  0x00342652 in llvm::FPPassManager::runOnModule (this=0xe021e0, M=@0xe02b00)
at PassManager.cpp:1395
#8  0x00342089 in llvm::MPPassManager::runOnModule (this=0xe01bd0, M=@0xe02b00)
at PassManager.cpp:1448
#9  0x003422de in llvm::PassManagerImpl::run (this=0xe01a00, M=@0xe02b00) at
PassManager.cpp:1529
#10 0x00342337 in llvm::PassManager::run (this=0xbffff128, M=@0xe02b00) at
PassManager.cpp:1558
#11 0x0000d340 in main (argc=4, argv=0xbffff1ac) at opt.cpp:522

== On solaris from llvm/test/Transforms/SimplifyCFG/2009-01-18-PHIPropCrash.ll
test data. ==

-bash-3.2$ /opt/clang/bin/llvm-as <
/export/home2/edward/lab/llvm/llvm/test/Transforms/SimplifyCFG/2009-01-18-PHIPropCrash.ll
| /opt/clang/bin/opt -simplifycfg
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

Instruction does not dominate all uses!
        %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [ %tmp12, %BB ]       
        ; <i32> [#uses=1]       %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9
], [ %tmp12, %BB ]                ; <i32> [#uses=1]Broken module found,
compilation aborted!
0   opt       0x086f8778 llvm::sys::SetInterruptFunction(void (*)()) + 88
1   opt       0x086f8d82 llvm::sys::RemoveFileOnSignal(llvm::sys::Path const&,
std::string*) + 596
2   libc.so.1 0xfeb6d0af __sighndlr + 15
3   libc.so.1 0xfeb601af call_user_handler + 687
4   libc.so.1 0xfeb71997 _lwp_kill + 7
5   libc.so.1 0xfeb1ab7a raise + 34
6   libc.so.1 0xfeaf1f8c abort + 116
7   opt       0x0869a587 llvm::EVT::EVT() + 1445
8   opt       0x086a7e4d llvm::verifyFunction(llvm::Function const&,
llvm::VerifierFailureAction) + 53485
9   opt       0x08672ec2 llvm::FPPassManager::runOnFunction(llvm::Function&) +
302
10  opt       0x0867308a llvm::FPPassManager::runOnModule(llvm::Module&) + 110
11  opt       0x086711b1 llvm::MPPassManager::runOnModule(llvm::Module&) + 393
12  opt       0x0867333a llvm::PassManagerImpl::run(llvm::Module&) + 124
13  opt       0x0867339f llvm::PassManager::run(llvm::Module&) + 39
14  opt       0x0830f982 main + 4606
15  opt       0x082fe070 _start + 128
Stack dump:
0.      Running pass 'Function Pass Manager' on module '<stdin>'.
1.      Running pass 'Module Verifier' on function '@func_105'
Abort (core dumped)

== Using gdb on Solaris ==

-bash-3.2$ /opt/clang/bin/llvm-as <
/export/home2/edward/lab/llvm/llvm/test/Transforms/SimplifyCFG/2009-01-18-PHIPropCrash.ll
> output-from-llvm-as
-bash-3.2$ more output-from-llvm-as 
-bash-3.2$ gdb /opt/clang/bin/opt 
GNU gdb 6.3.50_2004-11-23-cvs
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "i386-pc-solaris2.11"...
(gdb) run -simplifycfg <output-from-llvm-as
Starting program: /opt/clang/bin/opt -simplifycfg <output-from-llvm-as
warning: Unable to find dynamic linker breakpoint function.
GDB will be unable to debug shared library initializers
and track explicitly loaded dynamic code.
warning: Lowest section in /lib/libpthread.so.1 is .dynamic at 00000074
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

Instruction does not dominate all uses!
        %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9 ], [ %tmp12, %BB ]       
        ; <i32> [#uses=1]       %tmp12 = phi i32 [ 0, %BB13 ], [ %tmp10, %BB9
], [ %tmp12, %BB ]                ; <i32> [#uses=1]Broken module found,
compilation aborted!

Program received signal SIGABRT, Aborted.
0xfeb71997 in _lwp_kill () from /lib/libc.so.1
(gdb) bt
#0  0xfeb71997 in _lwp_kill () from /lib/libc.so.1
#1  0xfeb6aa0c in thr_kill () from /lib/libc.so.1
#2  0xfeb1ab7a in raise () from /lib/libc.so.1
#3  0xfeaf1f8c in abort () from /lib/libc.so.1
#4  0x0869a587 in abortIfBroken (this=0x8796c70) at Verifier.cpp:216
#5  0x086a7e4d in runOnFunction (this=0x8796c70, F=@0x87970d0) at
Verifier.cpp:171
#6  0x08672ec2 in llvm::FPPassManager::runOnFunction (this=0x8793360,
F=@0x87970d0) at PassManager.cpp:1372
#7  0x0867308a in llvm::FPPassManager::runOnModule (this=0x8793360,
M=@0x8796a58) at PassManager.cpp:1395
#8  0x086711b1 in llvm::MPPassManager::runOnModule (this=0x8798a10,
M=@0x8796a58) at PassManager.cpp:1448
#9  0x0867333a in llvm::PassManagerImpl::run (this=0x87964e8, M=@0x8796a58) at
PassManager.cpp:1529
#10 0x0867339f in llvm::PassManager::run (this=0x8047cd0, M=@0x8796a58) at
PassManager.cpp:1558
#11 0x0830f982 in main (argc=2, argv=0x8047d50) at opt.cpp:522

I tried my best to collect all the data I know of here. I don't really
understand the problem..

$ /opt/clang/bin/opt --version
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.6svn
  DEBUG build with assertions.
  Built Aug 11 2009(21:04:11).

  Registered Targets:
    (none)
$ /opt/clang/bin/clang --version
clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 78700M)
Target: i386-pc-solaris2.11
Thread model: posix


Best Regards,
Edward O'Callaghan.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 11 20:05:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Aug 2009 20:05:24 -0500
Subject: [LLVMbugs] [Bug 4711] New: Thumb2 tests fail horrifyingly on
	Solaris.
Message-ID: <bug-4711-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4711

           Summary: Thumb2 tests fail horrifyingly on Solaris.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Good day,

No idea where to start debugging this:

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldr.ll:23:10:
error: expected string not found in input
; CHECK: mov.w r1, #4096
         ^
<stdin>:36:1: note: scanning from here
 movw r1, #4096
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrb.ll:23:10:
error: expected string not found in input
; CHECK: mov.w r1, #4096
         ^
<stdin>:36:1: note: scanning from here
 movw r1, #4096
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-ldrh.ll:23:10:
error: expected string not found in input
; CHECK: mov.w r1, #4096
         ^
<stdin>:36:1: note: scanning from here
 movw r1, #4096
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-mov3.ll:6:10:
error: expected string not found in input
; CHECK: movs r0, #171
         ^
<stdin>:16:1: note: scanning from here
 movw r0, #171
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-str.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-str.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-str.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-str.ll:28:10:
error: expected string not found in input
; CHECK: mov.w r2, #4096
         ^
<stdin>:46:1: note: scanning from here
 movw r2, #4096
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strb.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strb.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strb.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strb.ll:28:10:
error: expected string not found in input
; CHECK: mov.w r2, #4096
         ^
<stdin>:46:1: note: scanning from here
 movw r2, #4096
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strh.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strh.ll |  llc
-march=thumb -mattr=+thumb2 | FileCheck
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strh.ll
/export/home2/edward/lab/llvm/llvm/test/CodeGen/Thumb2/thumb2-strh.ll:28:10:
error: expected string not found in input
; CHECK: mov.w r2, #4096
         ^
<stdin>:46:1: note: scanning from here
 movw r2, #4096
^


Cheers,
Edward.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Aug 12 12:13:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Aug 2009 12:13:53 -0500
Subject: [LLVMbugs] [Bug 4704] fallout from AltiVec patch
In-Reply-To: <bug-4704-206@http.llvm.org/bugs/>
Message-ID: <200908121713.n7CHDrhv030384@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4704


Nate Begeman <natebegeman at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |natebegeman at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #7 from Nate Begeman <natebegeman at mac.com>  2009-08-12 12:13:52 ---
Attempt 2 of this work fixed these issues.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Aug 12 20:55:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Aug 2009 20:55:54 -0500
Subject: [LLVMbugs] [Bug 4708] llvm-prof has cast warnings after r78477
In-Reply-To: <bug-4708-206@http.llvm.org/bugs/>
Message-ID: <200908130155.n7D1tsC6002472@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4708


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-12 20:55:53 ---
Thanks, applied as r78880


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 12 21:21:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Aug 2009 21:21:58 -0500
Subject: [LLVMbugs] [Bug 4712] New: internal compiler error when compiling
 ObjC 2. 0 code with fast enumeration
Message-ID: <bug-4712-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4712

           Summary: internal compiler error when compiling ObjC 2.0 code
                    with fast enumeration
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: aguynamedryan+llvm at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3304)
 --> (http://llvm.org/bugs/attachment.cgi?id=3304)
Testcase code that causes internal compiler error, reduced as per:
http://llvm.org/docs/HowToSubmitABug.html

While using llvm-gcc to compile some code that uses Objective-C 2.0's fast
enumeration, I encountered an internal compiler error.

I am running a version of llvm-gcc that I built from the svn repo, revision
78845.  I fully recompiled both llvm and llvm-gcc using this revision number.

The command I am using, under Ubuntu 9.04, to compile the testcase-min.mi file
that I've attached is:

/usr/local/bin/gcc -I/usr/include/GNUstep testcase-min.mi

and the error is:

internal compiler error: in objc_build_foreach_components, at
objc/objc-act.c:19778


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 13 00:34:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 00:34:17 -0500
Subject: [LLVMbugs] [Bug 4706] Bitcode reader fails after rev# 78258
In-Reply-To: <bug-4706-206@http.llvm.org/bugs/>
Message-ID: <200908130534.n7D5YHJa030643@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4706


Owen Anderson <resistor at mac.com> changed:

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




--- Comment #4 from Owen Anderson <resistor at mac.com>  2009-08-13 00:34:16 ---
I can't reproduce this issue on TOT with either the .ll or the .bc input.

If this is still occurring for you, please reopen and attach a full stack trace
of the crash.


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

From bugzilla-daemon at cs.uiuc.edu  Thu Aug 13 08:18:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 08:18:38 -0500
Subject: [LLVMbugs] [Bug 4713] New: xgcc fails silently,
	cannot build crtbegin.o
Message-ID: <bug-4713-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4713

           Summary: xgcc fails silently, cannot build crtbegin.o
           Product: Build scripts
           Version: 2.5
          Platform: PC
        OS/Version: other
            Status: NEW
          Keywords: build-problem
          Severity: blocker
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dave.hein at aspentech.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3306)
 --> (http://llvm.org/bugs/attachment.cgi?id=3306)
Log of make demonstrating the failure.

I'm running on Windows 7 RC (64-bit) using Cygwin. I'm trying to build the gcc
front-end and get a failure building crtbegin.o -- but there is no error
message.

I've attached a log of the stdout/stderr from the make.

So there are two bugs:

1. xgxx fails silently
2. xgxx fails to build crtbegin.o.

Note: Source trees for llvm and llvm-gcc were pulled via SVN from the
RELEASE_25 tag. I built llvm just fine (using a build target of tools-only) and
installed it. The config and make of llvm-gcc worked without any problem up to
this failure.

Note2: This version info for xgcc and gcc might be helpful:

$ gcc/xgcc -v
Using built-in specs.
Target: i686-pc-cygwin
Configured with: /cygdrive/c/archives/llvm/2.5/src/llvm-gcc-4.2/configure
--prog
ram-prefix=llvm-
--enable-llvm=/cygdrive/c/Users/heind/cygwin/trash/llvm-objects
 --enable-languages=c,c++ --prefix=/cygdrive/c/Users/heind/cygwin/local
Thread model: single
gcc version 4.2.1 (Based on Apple Inc. build 5636) (LLVM build 2.5)

$ gcc -v
Using built-in specs.
Target: i686-pc-cygwin
Configured with: /gnu/gcc/package/gcc4-4.3.2-2/src/gcc-4.3.2/configure
--srcdir=
/gnu/gcc/package/gcc4-4.3.2-2/src/gcc-4.3.2 --prefix=/usr --exec-prefix=/usr
--b
indir=/usr/bin --sbindir=/usr/sbin --libexecdir=/usr/sbin --datadir=/usr/share
-
-localstatedir=/var --sysconfdir=/etc --infodir=/usr/share/info
--mandir=/usr/sh
are/man --datadir=/usr/share --infodir=/usr/share/info --mandir=/usr/share/man
-
v --with-gmp=/usr --with-mpfr=/usr --enable-bootstrap
--enable-version-specific-
runtime-libs --with-slibdir=/usr/bin --libexecdir=/usr/lib --enable-static
--ena
ble-shared --enable-shared-libgcc --enable-__cxa_atexit --with-gnu-ld
--with-gnu
-as --with-dwarf2 --disable-sjlj-exceptions
--enable-languages=ada,c,c++,fortran
,java,objc,obj-c++ --disable-symvers --enable-libjava --program-suffix=-4
--enab
le-libgomp --enable-libssp --enable-libada --enable-threads=posix
AS=/opt/gcc-to
ols/bin/as.exe AS_FOR_TARGET=/opt/gcc-tools/bin/as.exe
LD=/opt/gcc-tools/bin/ld.
exe LD_FOR_TARGET=/opt/gcc-tools/bin/ld.exe
Thread model: posix
gcc version 4.3.2 20080827 (beta) 2 (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  Thu Aug 13 12:29:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 12:29:38 -0500
Subject: [LLVMbugs] [Bug 4714] New: compiler warnings on windows
Message-ID: <bug-4714-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4714

           Summary: compiler warnings on windows
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Generic Execution Engine Support
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: micah.villmow at amd.com
                CC: llvmbugs at cs.uiuc.edu


llvm\include\llvm/Operator.h(104) : warning C4624:
'llvm::OverflowingBinaryOperator' : destructor could not be generated because a
base class destructor is inaccessible
llvm\include\llvm/Operator.h(121) : warning C4624: 'llvm::AddOperator' :
destructor could not be generated because a base class destructor is
inaccessible
llvm\include\llvm/Operator.h(138) : warning C4624: 'llvm::SubOperator' :
destructor could not be generated because a base class destructor is
inaccessible
llvm\include\llvm/Operator.h(155) : warning C4624: 'llvm::MulOperator' :
destructor could not be generated because a base class destructor is
inaccessible
llvm\include\llvm/Operator.h(182) : warning C4624: 'llvm::SDivOperator' :
destructor could not be generated because a base class destructor is
inaccessible
llvm\include\llvm/Operator.h(250) : warning C4624: 'llvm::GEPOperator' :
destructor could not be generated because a base class destructor is
inaccessible


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 13 13:53:51 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 13:53:51 -0500
Subject: [LLVMbugs] [Bug 4710] simplifycfg horks phi node
In-Reply-To: <bug-4710-206@http.llvm.org/bugs/>
Message-ID: <200908131853.n7DIrp9F005929@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4710


Nick Lewycky <nlewycky at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE
            Summary|Assert fires in the Verifier|simplifycfg horks phi node
                   |when called from the        |
                   |PassManager in 'opt'        |




--- Comment #1 from Nick Lewycky <nlewycky at google.com>  2009-08-13 13:53:51 ---
Asserting is exactly what the verifier is designed to do: check whether the IR
is valid, and if it's not, assert. What you're seeing here is the -simplifycfg
pass producing illegal IR.

This is exactly what test/Transforms/SimplifyCFG/2009-01-18-PHIPropCrash.ll is
testing for. This test is known failing, which is why it's XFAILed (see the
"XFAIL: *" at the top) and contains a reference to the real problem, bug 3016.


*** This bug has been marked as a duplicate of bug 3016 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 13 17:29:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 17:29:22 -0500
Subject: [LLVMbugs] [Bug 4706] Bitcode reader fails after rev# 78258
In-Reply-To: <bug-4706-206@http.llvm.org/bugs/>
Message-ID: <200908132229.n7DMTMhj028255@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4706


Gareth <gareth_morgan77 at hotmail.com> changed:

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




--- Comment #5 from Gareth <gareth_morgan77 at hotmail.com>  2009-08-13 17:29:21 ---
I have synced my tree to rev 78927 and call still reproduce this bug. 
Callstack is identical to one shown above.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 13 19:02:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Aug 2009 19:02:55 -0500
Subject: [LLVMbugs] [Bug 4714] compiler warnings on windows
In-Reply-To: <bug-4714-206@http.llvm.org/bugs/>
Message-ID: <200908140002.n7E02tDe007599@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4714


Micah Villmow <micah.villmow at amd.com> changed:

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




--- Comment #2 from Micah Villmow <micah.villmow at amd.com>  2009-08-13 19:02:55 ---
This fixed the warnings.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 02:18:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 02:18:26 -0500
Subject: [LLVMbugs] [Bug 4716] New: Another regscavenger assertion
Message-ID: <bug-4716-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4716

           Summary: Another regscavenger assertion
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: evan.cheng at apple.com, llvmbugs at cs.uiuc.edu,
                    ojomojo at gmail.com, stoklund at 2pi.dk


Created an attachment (id=3310)
 --> (http://llvm.org/bugs/attachment.cgi?id=3310)
regscav.ll

Consider the attached testcase. It currently causes the following assertion:

Assertion failed: ((KillRegs.test(Reg) || isUnused(Reg) ||
isLiveInButUnusedBefore(Reg, MI, MBB, TRI, MRI)) && "Re-defining a live
register!"), function forward, file
/Users/asl/Projects/llvm/src/lib/CodeGen/RegisterScavenging.cpp, line 245.
0   llc               0x00adf040 llvm::sys::SetInterruptFunction(void (*)()) +
82
1   llc               0x00adf42f llvm::sys::SetInterruptFunction(void (*)()) +
1089
2   libSystem.B.dylib 0x9329c2bb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1825979759
4   libSystem.B.dylib 0x9331023a raise + 26
5   libSystem.B.dylib 0x9331c679 abort + 73
6   libSystem.B.dylib 0x933113db __assert_rtn + 101
7   llc               0x0088f03c llvm::RegScavenger::forward() + 1418
8   llc               0x0037f94d
llvm::RegScavenger::forward(llvm::ilist_iterator<llvm::MachineInstr>) + 107
9   llc               0x008558dd
llvm::PEI::replaceFrameIndices(llvm::MachineFunction&) + 1221
10  llc               0x00857881
llvm::PEI::runOnMachineFunction(llvm::MachineFunction&) + 407
11  llc               0x00823f8f
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 77
12  llc               0x00a56db8
llvm::FPPassManager::runOnFunction(llvm::Function&) + 304
13  llc               0x00a57028
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 112
14  llc               0x00a5713c
llvm::FunctionPassManager::run(llvm::Function&) + 130
15  llc               0x000038ba main + 3664
16  llc               0x00002436 start + 54
17  llc               0x00000003 start + 18446744073709542403
Stack dump:
0.      Program arguments: ./llc regscav4.bc -debug 
1.      Running pass 'Prolog/Epilog Insertion & Frame Finalization' on function
'@_start'


I believe that this testcase worked fine ~month ago, so this should be some
fallout from recent changes


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 05:25:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 05:25:22 -0500
Subject: [LLVMbugs] [Bug 4717] New: Rejects valid c++
Message-ID: <bug-4717-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4717

           Summary: Rejects valid c++
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc rejects

       class C {
         public:
         enum { a = 3};
       };
       template<class C2>
       class C3 : public C2 {
         public:
         C2::a;
         class C4;
       };
       template<class C2>
       class C3<C2>::C4 {
         int v[a];
         public:
         C4(int a2);
       };
       void f(int a3) {
         C3<C>::C4 a4(a3);
       }

This was fixed in http://gcc.gnu.org/ml/gcc-cvs/2007-09/msg00616.html, but
unfortunately that patch is GPL3.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 05:29:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 05:29:33 -0500
Subject: [LLVMbugs] [Bug 4718] New: Rejects valid c++
Message-ID: <bug-4718-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4718

           Summary: Rejects valid c++
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


llvm-gcc rejects

        class A {
        public:
          A();
          private:
         A(const A&);
        };
        void B(const A& arg);
        void C() {
          B(A());
        }

This was fixed in gcc in http://gcc.gnu.org/ml/gcc-cvs/2007-10/msg00701.html,
but
unfortunately that patch is GPL3.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 17:07:14 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 17:07:14 -0500
Subject: [LLVMbugs] [Bug 3959] [Mips] Return of a double constant causes an
	assert.
In-Reply-To: <bug-3959-206@http.llvm.org/bugs/>
Message-ID: <200908142207.n7EM7Ekw017172@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3959


Bruno Cardoso Lopes <bruno.cardoso at gmail.com> changed:

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




--- Comment #3 from Bruno Cardoso Lopes <bruno.cardoso at gmail.com>  2009-08-14 17:07:14 ---
This works perfectly for me on TOT.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 17:27:07 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 17:27:07 -0500
Subject: [LLVMbugs] [Bug 4238] loop rotate do not preserve domtree and/ or
 domfrontier info properly
In-Reply-To: <bug-4238-206@http.llvm.org/bugs/>
Message-ID: <200908142227.n7EMR74d019857@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4238


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #14 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-14 17:27:07 ---
No, that was it... so I suppose this is 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  Fri Aug 14 18:30:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 18:30:59 -0500
Subject: [LLVMbugs] [Bug 4720] New: clang should not error on different main
Message-ID: <bug-4720-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4720

           Summary: clang should not error on different main
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: OpenBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jsg at openbsd.org
                CC: llvmbugs at cs.uiuc.edu


Recent versions of clang now error on main defined differently,
this is not in keeping with the behaviour of gcc.

clang:
../../../../kern/init_main.c:137:5: error: first argument of 'main' should be
of type 'int'
int     main(void *);
        ^
../../../../kern/init_main.c:184:1: error: first argument of 'main' should be
of type 'int'
main(void *framep)
^

gcc:
../../../../kern/init_main.c:185: warning: first argument of `main' should be
`int'
../../../../kern/init_main.c:185: warning: `main' takes only zero or two
arguments

The gcc warning can be disabled with -Wno-main, clang does not recognise this
option.  With this recent change clang can no longer compile the OpenBSD kernel
and likely other code besides.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 19:03:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 19:03:33 -0500
Subject: [LLVMbugs] [Bug 4721] New: memory bugs and codegen crash with vlas
Message-ID: <bug-4721-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4721

           Summary: memory bugs and codegen crash with vlas
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


This testcase:

void sum1(int rb) {
  typedef unsigned char imgrow[rb];
  typedef imgrow img[rb];

  const img *br;
  int y;

   (*br)[y];
}

Crashes codegen with an assert, and crashes -ast-print/-ast-dump with valgrind
errors about invalid reads.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 19:47:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 19:47:50 -0500
Subject: [LLVMbugs] [Bug 4722] New: GVN doesn't look through select for
	alloca
Message-ID: <bug-4722-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4722

           Summary: GVN doesn't look through select for alloca
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


This probably isn't GVN's fault directly. Somebody isn't peering through select
when checking whether a given value really is a select or not. This means we
miss this optimization:

%struct.X = type { i32 }

define void @_Z1fb(i1 %b) nounwind {
entry:
  %x1 = alloca %struct.X, align 8
  %x2 = alloca %struct.X, align 8
  %iftmp.0.0 = select i1 %b, %struct.X* %x2, %struct.X* %x1
  %tmp1 = getelementptr %struct.X* %iftmp.0.0, i64 0, i32 0
  %tmp2 = load i32* %tmp1, align 8
  %tmp3 = add i32 %tmp2, 1
  store i32 %tmp3, i32* %tmp1, align 8
  ret void
}

We need to make sure we limit the depth of any such search though, because it
can grow N^2 easily.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 21:16:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 21:16:53 -0500
Subject: [LLVMbugs] [Bug 4720] clang should not error on different main
In-Reply-To: <bug-4720-206@http.llvm.org/bugs/>
Message-ID: <200908150216.n7F2GrbH017364@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4720


John McCall <rjmccall at apple.com> changed:

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




--- Comment #6 from John McCall <rjmccall at apple.com>  2009-08-14 21:16:53 ---
I've changed -ffreestanding to disable all special handling of main() in
r79070.  Please feel free to reopen this bug if you feel that switching to
-ffreestanding is unreasonable.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 21:19:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 21:19:22 -0500
Subject: [LLVMbugs] [Bug 4723] New: invalid relocation for PIC emitted in
 libstdc++ build
Message-ID: <bug-4723-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4723

           Summary: invalid relocation for PIC emitted in libstdc++ build
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3313)
 --> (http://llvm.org/bugs/attachment.cgi?id=3313)
preprocessed C++ source

The llvm-gcc build fails on x86_64-unknown-linux-gnu with this message:

/usr/bin/ld: .libs/bitmap_allocator.o: relocation R_X86_64_PC32 against symbol
`__gxx_personality_v0' can not be used when making a shared object; recompile
with -fPIC

It appears that llvm-g++ is emitting a non-PIC relocation in PIC mode,
in the .eh_frame section.  Attached is the preprocessed source for 
bitmap_allocator.o. Compiling it with both g++-4.2 and llvm-g++ shows
the difference:

g++-4.2 -fpic -c bitmap_allocator.i && readelf -r bitmap_allocator.o |grep
personality
Relocation section '.rela.data.DW.ref.__gxx_personality_v0' at offset 0x1b998
contains 1 entries:
000000000000  018500000001 R_X86_64_64       0000000000000000
__gxx_personality_v0 + 0

llvm-g++ -fpic -c bitmap_allocator.i && readelf -r bitmap_allocator.o |grep
personality
00000000002b  01b800000002 R_X86_64_PC32     0000000000000000
__gxx_personality_v0 + 0


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

From bugzilla-daemon at cs.uiuc.edu  Fri Aug 14 21:56:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 21:56:54 -0500
Subject: [LLVMbugs] [Bug 4721] memory bugs and codegen crash with vlas
In-Reply-To: <bug-4721-206@http.llvm.org/bugs/>
Message-ID: <200908150256.n7F2usB6022339@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4721


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-14 21:56:54 ---
Fixed in r79071.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 14 23:21:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Aug 2009 23:21:38 -0500
Subject: [LLVMbugs] [Bug 4724] New: "relocation R_X86_64_PC32 against
 undefined symbol" building C++ shared libraries
Message-ID: <bug-4724-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4724

           Summary: "relocation R_X86_64_PC32 against undefined symbol"
                    building C++ shared libraries
           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: ssen at apple.com
                CC: llvmbugs at cs.uiuc.edu


I'm trying to build llvm/llvm-gcc-4.2 trunk (r79073) on Ubuntu 9.0.4 (x86_64)

libstdc++-v3 fails to link with:
/usr/bin/ld: .libs/bitmap_allocator.o: relocation R_X86_64_PC32 against symbol
`__gxx_personality_v0' can not be used when making a shared object; recompile
with -fPIC
/usr/bin/ld: final link failed: Bad value
collect2: ld returned 1 exit status


I managed to reduce this to something fairly simple (i think the difference in
mesage is because libstdc++.so actually provides the named symbol):

$ cat foo.cc
extern void foo(void);

void bar(void) throw() {
  foo();
}
$ ./gcc/xgcc -Bgcc -c foo.cc -fPIC
$ /usr/bin/g++ -shared -o foo.so foo.o
/usr/bin/ld: foo.o: relocation R_X86_64_PC32 against undefined symbol
`__gxx_personality_v0@@CXXABI_1.3' can not be used when making a shared object;
recompile with -fPIC
/usr/bin/ld: final link failed: Bad value
collect2: ld returned 1 exit status
$ g++ -c foo.cc -fPIC
$ /usr/bin/g++ -shared -o foo.so foo.o
$ 

Based on looking at the assembly, it looks like bad EH frame data, where
llvm-g++ is creating:
.LEH_frame1:
.Lsection_eh_frame1:
.Leh_frame_common1:
        .long   .Leh_frame_common_end1-.Leh_frame_common_begin1
.Leh_frame_common_begin1:
        .long   0x0
        .byte   0x1
        .asciz  "zPLR"
        .uleb128        1
        .sleb128        -8
        .byte   0x10
        .uleb128        7
        .byte   0x1B
        .long   __gxx_personality_v0-.

The ".long" creates a 32-bit relocation in a 64-bit ELF object, which can't be
linked into a shared library because the difference might be greater than
32-bits.

g++-4.3.3 creates:
        .uleb128 0x7
        .byte   0x9b
        .long   DW.ref.__gxx_personality_v0-.
        .byte   0x1b
...
.LEFDE1:
        .hidden DW.ref.__gxx_personality_v0
        .weak   DW.ref.__gxx_personality_v0
        .section       
.data.DW.ref.__gxx_personality_v0,"awG", at progbits,DW.ref
.__gxx_personality_v0,comdat
        .align 8
        .type   DW.ref.__gxx_personality_v0, @object
        .size   DW.ref.__gxx_personality_v0, 8
DW.ref.__gxx_personality_v0:
        .quad   __gxx_personality_v0


Using -fno-exceptions does eliminate the issue, which reinforces bad EH data


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 04:06:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 04:06:39 -0500
Subject: [LLVMbugs] [Bug 4723] invalid relocation for PIC emitted in
	libstdc++ build
In-Reply-To: <bug-4723-206@http.llvm.org/bugs/>
Message-ID: <200908150906.n7F96dCw015891@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4723


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-15 04:06:39 ---


*** This bug has been marked as a duplicate of bug 4724 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 08:21:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 08:21:52 -0500
Subject: [LLVMbugs] [Bug 4716] Another regscavenger assertion
In-Reply-To: <bug-4716-206@http.llvm.org/bugs/>
Message-ID: <200908151321.n7FDLqiD018689@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4716


Jakob Stoklund Olesen <stoklund at 2pi.dk> changed:

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




--- Comment #2 from Jakob Stoklund Olesen <stoklund at 2pi.dk>  2009-08-15 08:21:51 ---
Fixed here: http://llvm.org/viewvc/llvm-project?rev=79096&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 13:34:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 13:34:36 -0500
Subject: [LLVMbugs] [Bug 4724] "relocation R_X86_64_PC32 against undefined
 symbol" building C++ shared libraries
In-Reply-To: <bug-4724-206@http.llvm.org/bugs/>
Message-ID: <200908151834.n7FIYaje025658@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4724


Shantonu Sen <ssen at apple.com> changed:

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




--- Comment #7 from Shantonu Sen <ssen at apple.com>  2009-08-15 13:34:36 ---
bootstrap works! (including libstdc++-v3)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 13:42:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 13:42:22 -0500
Subject: [LLVMbugs] [Bug 4725] New: SoftenFloatOp_FP_TO_SINT assertion
Message-ID: <bug-4725-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4725

           Summary: SoftenFloatOp_FP_TO_SINT assertion
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: stoklund at 2pi.dk
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3315)
 --> (http://llvm.org/bugs/attachment.cgi?id=3315)
Test case for Blackfin back-end

The attached test case fails with

Assertion failed: (LC != RTLIB::UNKNOWN_LIBCALL && "Unsupported FP_TO_SINT!"),
function SoftenFloatOp_FP_TO_SINT, file
/Users/stoklund/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp, line 693.

I believe this happens when using soft-float on any target with legal i8 or i16
types:

llc -march=x86 -soft-float
llc -march=x86-64 -soft-float
llc -march=pic16
llc -march=bfin
llc -march=msp430


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 15:29:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 15:29:36 -0500
Subject: [LLVMbugs] [Bug 4726] New: Patch series to get APInt and APFloat to
 take StringRefs
Message-ID: <bug-4726-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4726

           Summary: Patch series to get APInt and APFloat to take StringRefs
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: idadesub at users.sourceforge.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3316)
 --> (http://llvm.org/bugs/attachment.cgi?id=3316)
unit tests for APInt

Hi all,

I sent these patches to llvm-commits, but I didn't want to get them lost in the
shuffle so I'm also putting them here. This series does a couple things. First,
it converts APFloat to take a StringRef instead of a raw c string. The patch to
change APInt was already committed. Second, it adds a bunch of unit tests to
APInt and APFloat to validate string parsing. Third, it exposes constructing
ConstantInts and ConstantFPs from llvm-c through strings. Fourth, it then
exposes this to ocaml.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 16:34:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 16:34:42 -0500
Subject: [LLVMbugs] [Bug 4727] New: Expose most of IRBuilder and Constant*
 to llvm-c and ocaml
Message-ID: <bug-4727-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4727

           Summary: Expose most of IRBuilder and Constant* to llvm-c and
                    ocaml
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: idadesub at users.sourceforge.net
                CC: llvmbugs at cs.uiuc.edu


This patch series exposes most of the functionality in IRBuilder and the
Constant functions to llvm-c and ocaml. Specifically, it exposes:

IRBuilder::CreateAggregateRet
IRBuilder::CreateNSWAdd
IRBuilder::CreateFAdd
IRBuilder::CreateFSub
IRBuilder::CreateFMul
IRBuilder::CreateExactSDiv
IRBuilder::CreateInBoundsGEP
IRBuilder::CreateStructGEP
IRBuilder::CreateGlobalString
IRBuilder::CreateGlobalStringPtr
IRBuilder::CreateZExtOrBitCast
IRBuilder::CreateSExtOrBitCast
IRBuilder::CreateTruncOrBitCast
IRBuilder::CreatePointerCast
IRBuilder::CreateIntCast
IRBuilder::CreateFPCast
IRBuilder::CreateIsNull
IRBuilder::CreateIsNotNull
IRBuilder::CreatePtrDiff
ConstantExpr::getFNeg
ConstantExpr::getNSWAdd
ConstantExpr::getFAdd
ConstantExpr::getFSub
ConstantExpr::getFMul
ConstantExpr::getExactSDiv
ConstantExpr::getInBoundsGEP
ConstantExpr::getZExtOrBitCast
ConstantExpr::getSExtOrBitCast
ConstantExpr::getTruncOrBitCast
ConstantExpr::getPointerCast
ConstantExpr::getIntegerCast
ConstantExpr::getFPCast


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 15 21:18:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 21:18:18 -0500
Subject: [LLVMbugs] [Bug 4727] Expose most of IRBuilder and Constant* to
 llvm-c and ocaml
In-Reply-To: <bug-4727-206@http.llvm.org/bugs/>
Message-ID: <200908160218.n7G2IIcF018428@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4727


Erick Tryzelaar <idadesub at users.sourceforge.net> changed:

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




--- Comment #8 from Erick Tryzelaar <idadesub at users.sourceforge.net>  2009-08-15 21:18:17 ---
Great, I'll commit it 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  Sat Aug 15 21:26:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Aug 2009 21:26:26 -0500
Subject: [LLVMbugs] [Bug 4722] GVN doesn't look through select for alloca
In-Reply-To: <bug-4722-206@http.llvm.org/bugs/>
Message-ID: <200908160226.n7G2QQ7P019555@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4722


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-08-15 21:26:25 ---
I don't think this is really worthwhile.  Loads from undef values are
incredibly rare, usually the result of jump threading.  This case is not worth
making memdep more complex for imo.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 01:36:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 01:36:06 -0500
Subject: [LLVMbugs] [Bug 4728] New: msvc buildbot broken due to r79177
Message-ID: <bug-4728-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4728

           Summary: msvc buildbot broken due to r79177
           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, ofv at wanadoo.es


See http://google1.osuosl.org:8011/builders/clang-i686-xp-msvc9


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 01:38:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 01:38:24 -0500
Subject: [LLVMbugs] [Bug 3016] simplifycfg creates dead code and then
 creates invalid code because of it
In-Reply-To: <bug-3016-206@http.llvm.org/bugs/>
Message-ID: <200908160638.n7G6cOea018108@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3016


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #9 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-16 01:38:24 ---
Fixed in r79174.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 01:50:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 01:50:09 -0500
Subject: [LLVMbugs] [Bug 4031] no LLVM intrinsic for
 __builtin_ia32_ptestz128 SSE4. 1 instruction
In-Reply-To: <bug-4031-206@http.llvm.org/bugs/>
Message-ID: <200908160650.n7G6o9Ok019653@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4031


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-16 01:50:09 ---
This got fixed recently.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 02:45:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 02:45:42 -0500
Subject: [LLVMbugs] [Bug 4728] msvc buildbot broken due to r79177
In-Reply-To: <bug-4728-206@http.llvm.org/bugs/>
Message-ID: <200908160745.n7G7jgeY032273@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4728


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #1 from ??scar Fuentes <ofv at wanadoo.es>  2009-08-16 02:45:41 ---
r79180 should fix 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  Sun Aug 16 04:17:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 04:17:22 -0500
Subject: [LLVMbugs] [Bug 4728] msvc buildbot broken due to r79177
In-Reply-To: <bug-4728-206@http.llvm.org/bugs/>
Message-ID: <200908160917.n7G9HMSp018158@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4728


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-16 04:17:21 ---
It's still broken...


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 04:53:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 04:53:48 -0500
Subject: [LLVMbugs] [Bug 4728] msvc buildbot broken due to r79177
In-Reply-To: <bug-4728-206@http.llvm.org/bugs/>
Message-ID: <200908160953.n7G9rml2011805@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4728


Benjamin Kramer <benny.kra at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |benny.kra at gmail.com
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Benjamin Kramer <benny.kra at gmail.com>  2009-08-16 04:53:48 ---
fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090810/084696.html

Oscar, please make sure I didn't break your parallel builds.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 16 06:58:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 06:58:32 -0500
Subject: [LLVMbugs] [Bug 4707] Use after free in Thumb2SizeReduction
In-Reply-To: <bug-4707-206@http.llvm.org/bugs/>
Message-ID: <200908161158.n7GBwWdN027198@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4707


Benjamin Kramer <benny.kra at gmail.com> changed:

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




--- Comment #2 from Benjamin Kramer <benny.kra at gmail.com>  2009-08-16 06:58:32 ---
Fixed
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090810/084700.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 Aug 16 22:23:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Aug 2009 22:23:50 -0500
Subject: [LLVMbugs] [Bug 4731] New: NEON vbslq_f32() intrinsic triggers an
	assert
Message-ID: <bug-4731-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4731

           Summary: NEON vbslq_f32() intrinsic triggers an assert
           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: deeppatel1987 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3329)
 --> (http://llvm.org/bugs/attachment.cgi?id=3329)
test case

The attached code triggers an assert because the NEON_builtin_vbsl handling
(llvm_arm.cpp:1869) is trying to create an and <4 x float>, <4 x i32>.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 17 00:06:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 00:06:53 -0500
Subject: [LLVMbugs] [Bug 4731] NEON vbslq_f32() intrinsic triggers an assert
In-Reply-To: <bug-4731-206@http.llvm.org/bugs/>
Message-ID: <200908170506.n7H56rJX024621@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4731


Bob Wilson <bob.wilson at apple.com> changed:

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




--- Comment #2 from Bob Wilson <bob.wilson at apple.com>  2009-08-17 00:06:53 ---
I took the fix one step further and canonicalized all the types to use i32
elements.  We need to do that for all the bitwise NEON operations to make sure
things get CSEd, so we might as well start with this one.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 17 03:55:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 03:55:38 -0500
Subject: [LLVMbugs] [Bug 4726] Patch series to get APInt and APFloat to take
	StringRefs
In-Reply-To: <bug-4726-206@http.llvm.org/bugs/>
Message-ID: <200908170855.n7H8tcfW000827@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4726


Erick Tryzelaar <idadesub at users.sourceforge.net> changed:

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




--- Comment #7 from Erick Tryzelaar <idadesub at users.sourceforge.net>  2009-08-17 03:55:38 ---
Yep, I committed it earlier today. Closing!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 17 10:36:31 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 10:36:31 -0500
Subject: [LLVMbugs] [Bug 4678] "arguments were dropped" warning
In-Reply-To: <bug-4678-206@http.llvm.org/bugs/>
Message-ID: <200908171536.n7HFaVN2018277@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4678


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-08-17 10:36:30 ---
Duncan reverted the 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  Mon Aug 17 11:32:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 11:32:53 -0500
Subject: [LLVMbugs] [Bug 4732] New: spurious undefined reference
Message-ID: <bug-4732-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4732

           Summary: spurious undefined reference
           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


Seen on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp181$ clang -O3 small.c -o foo
/tmp/cc-LeSGPe.o: In function `int321':
small.c:(.text+0x8a): undefined reference to `.LBB5_1'
small.c:(.text+0x90): undefined reference to `.LBB5_1'
collect2: ld returned 1 exit status
clang: error: linker command failed with exit code 1 (use -v to see invocation)

regehr at john-home:~/volatile/tmp181$ clang -v

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 79189)
Target: i386-pc-linux-gnu
Thread model: posix

regehr at john-home:~/volatile/tmp181$ cat small.c

char foo1 (int _si1, char _si2)
{
  return 0 > 1 || _si1 <= 0 && _si1 < _si2 || 0 ? : 1;
}

char foo2 (int _si1, char _si2)
{
  return 0;
}

char foo3 (int _si1, char _si2)
{
  return 0 && 1 || _si2 > 0 && (-2147483647 - 1 / _si2) ? : _si2;
}

char func_16 (int z, int a)
{
  return 0;
}

void int321 (char p_103, int uint8p_104)
{
  char l_106;
  func_16 (foo1
           (p_103 || foo3 (l_106, p_103), 1), 1);
  for (; p_103 <= 1; foo2(0,0))
    {
    }
}

int main (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  Mon Aug 17 11:46:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 11:46:36 -0500
Subject: [LLVMbugs] [Bug 4733] New: Ran out of registers during register
	allocation!
Message-ID: <bug-4733-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4733

           Summary: Ran out of registers during register allocation!
           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


regehr at john-home:~/volatile/tmp182$ llvm-gcc -c -O small.c

LLVM ERROR: Ran out of registers during register allocation!

regehr at john-home:~/volatile/tmp182$ llvm-gcc -v

Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/z/tmp/llvm-gcc-r79189-install
--program-prefix=llvm-r79189- --enable-languages=c,c++
--enable-llvm=/home/regehr/z/tmp/llvm-r79189 --enable-checking=release
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)

regehr at john-home:~/volatile/tmp182$ cat small.c

static char foo1 (int _si1, char _si2)
{
  return _si1 > 32767 - _si2 < -_si2 ? : 1;
}

char foo2 (char _ui1, char _ui2)
{
  return _ui2;
}

static char foo3 (char _left, int _right)
{
  return _right >= 1 * 8 ? : _left;
}

static char foo4 (long long int _left, int _right)
{
  return _right >= 1 * 8 ? : _left;
}

struct S0
{
  char f0;
  char f1;
  char f3;
};

struct S1
{
  char f1;
  char f2;
};

struct S0 g_65 = {
  -1, 1
};

long long int g_102;

int safe (int);
int func_13 (int);

void uint80 (int p_11)
{
  struct S0 l_17 = {
    1, 0x3
  };
  struct S1 l_14 = {
    3, 0x41
  };
  func_13 (foo2 (l_17.f0, p_11));
  if (func_53
      (foo1
       (foo4 (g_102, p_11),
        foo3 (l_14.f2, g_65.f3)), safe (g_65.f1),
       safe (l_14.f1), 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 Aug 17 11:47:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 11:47:06 -0500
Subject: [LLVMbugs] [Bug 4734] New: -loop-rotate produces invalid code
Message-ID: <bug-4734-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4734

           Summary: -loop-rotate produces invalid code
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3331)
 --> (http://llvm.org/bugs/attachment.cgi?id=3331)
testcase

Running opt -loop-rotate transforms a program that runs correctly into one that
produces a segmentation fault.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 17 18:19:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Aug 2009 18:19:11 -0500
Subject: [LLVMbugs] [Bug 4735] New: LLVM-GCC miscompiles libstdc++
Message-ID: <bug-4735-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4735

           Summary: LLVM-GCC miscompiles libstdc++
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3333)
 --> (http://llvm.org/bugs/attachment.cgi?id=3333)
Possible solution

This is what it generates when compiling string-inst.cc:

        .section        .gnu.linkonce.t._ZNKSs7_M_dataEv,"a", at progbits


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 18 09:06:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 09:06:47 -0500
Subject: [LLVMbugs] [Bug 4735] LLVM-GCC miscompiles libstdc++
In-Reply-To: <bug-4735-206@http.llvm.org/bugs/>
Message-ID: <200908181406.n7IE6lpm005503@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4735


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

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




--- Comment #4 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-18 09:06:47 ---
Applied, 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 Aug 18 10:08:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 10:08:58 -0500
Subject: [LLVMbugs] [Bug 4736] New: ice: Binary operator types must match!
Message-ID: <bug-4736-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4736

           Summary: ice: Binary operator types must match!
           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


Seen on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp184$ clang -O2 small.c

clang-cc: SelectionDAG.cpp:2538: llvm::SDValue
llvm::SelectionDAG::getNode(unsigned int, llvm::DebugLoc, llvm::EVT,
llvm::SDValue, llvm::SDValue): Assertion `VT.isInteger() && N1.getValueType()
== N2.getValueType() && N1.getValueType() == VT && "Binary operator types must
match!"' failed.
0   clang-cc 0x09320ef8
Stack dump:
0.      Program arguments:
/home/regehr/z/tmp/llvm-gcc-r79306-install/bin/../libexec/clang-cc -triple
i386-pc-linux-gnu -S -disable-free -main-file-name small.c --relocation-model
static --disable-fp-elim --unwind-tables=0 --mcpu=pentium4 --fmath-errno=1 -O2
-fdiagnostics-show-option -o /tmp/cc-DaVIq9.s -x c small.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@int87'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/volatile/tmp184$ clang -v

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 79306)
Target: i386-pc-linux-gnu
Thread model: posix

regehr at john-home:~/volatile/tmp184$ cat small.c

struct S0
{
  int f0;
  long long int f1;
  long long int f2;
  long long int f3;
};

struct S1
{
  int f0;
  char f1;
  struct S0 f2;
};

volatile int g_4;

struct S1 g_144 = {
  0L, 1L
};

int func_72 (char p_74)
{
  return g_4;
}

int func_70 (struct S1 p_71)
{
  int l_77;
  if (func_72 (p_71.f2.f3))
    return p_71.f2.f1;
  for (l_77 = 0; l_77; )
    {
      return 1;
    }
  return p_71.f2.f2;
}

int int87 (uint64p_8)
{
  for (; ; )
    if (func_70 (g_144))
      {
        return 1;
      }
  return 0;
}


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 18 10:38:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 10:38:22 -0500
Subject: [LLVMbugs] [Bug 4737] New: possible clang wrong code bug
Message-ID: <bug-4737-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4737

           Summary: possible clang wrong code bug
           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


Sorry for the size of the test input, I wasn't able to reduce it any more.  I
think the test input is well-formed and that clang is wrong at -O2.

Seen on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp183$ clang -O small.c -o small
regehr at john-home:~/volatile/tmp183$ ./small
checksum = 1
regehr at john-home:~/volatile/tmp183$ clang -O2 small.c -o small
regehr at john-home:~/volatile/tmp183$ ./small
checksum = 4294967297
regehr at john-home:~/volatile/tmp183$ clang -v
clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 79306)
Target: i386-pc-linux-gnu
Thread model: posix
regehr at john-home:~/volatile/tmp183$ cat small.c
#include <stdint.h>
#include <stdio.h>

uint8_t crc32_tab[256];
static uint8_t crc32_context = 1;

static void 
crc32_gentab (void)
{
  uint32_t crc;
  const uint32_t poly = 0xEDB88320UL;
  int i, j;

  for (i = 0; i < 256; i++) {
    crc = i;
    for (j = 8; j > 0; j--) {
      if (crc & 1) {
        crc = (crc >> 1) ^ poly;
      } else {
        crc >>= 1;
      }
    }
    crc32_tab[i] = crc;
  }
}

static void 
crc32_byte (uint8_t b) {
  crc32_context = 
    (((int)crc32_context >> 8) & 0x00FFFFFF) ^ 
    crc32_tab[(crc32_context ^ b) & 0xFF];
}

struct S0
{
  uint8_t f1;
  uint32_t f2;
  uint32_t f3;
};

volatile uint8_t g_3 = 1;
uint8_t g_14;
uint64_t g_17;

struct S0 g_174 = {
  -4L, 0xE212550BL, 1
};

int32_t func_12 (void);
int32_t func_12 (void)
{
  for (g_14 = 0; 0; )
    {
    }
  return 1;
}

uint64_t func_72 (struct S0 p_82);
uint64_t func_72 (struct S0 p_82)
{
  if (func_12 ())
    {
    }
  return p_82.f3;
}

void func_1 (void);
void func_1 (void)
{
  if (g_3)
    {
    }
  else
    {
      return;
    }
  g_17 = func_72 (g_174);
}

int main (void)
{
  crc32_gentab ();
  func_1 ();
  crc32_byte (g_17);
  printf ("checksum = %lld\n", g_17);
  return 0;
}


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 18 11:04:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 11:04:06 -0500
Subject: [LLVMbugs] [Bug 4734] -loop-rotate produces invalid code
In-Reply-To: <bug-4734-206@http.llvm.org/bugs/>
Message-ID: <200908181604.n7IG46dc020506@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4734


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

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




--- Comment #3 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2009-08-18 11:04:06 ---
In the end it looks like this is a duplicate of 4735. It is fixed now.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 18 12:22:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 12:22:19 -0500
Subject: [LLVMbugs] [Bug 4732] codegen pass breaks machine cfg
In-Reply-To: <bug-4732-206@http.llvm.org/bugs/>
Message-ID: <200908181722.n7IHMJWe030753@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4732


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-08-18 12:22:18 ---
This is fixed with r79338.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 18 12:23:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 12:23:19 -0500
Subject: [LLVMbugs] [Bug 2998] llvm-g++ EH code doesn't work with GCC 4.3
	libgcc_s.so.1
In-Reply-To: <bug-2998-206@http.llvm.org/bugs/>
Message-ID: <200908181723.n7IHNJfU030892@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2998


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #6 from Duncan Sands <baldrick at free.fr>  2009-08-18 12:23:19 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084856.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 Aug 18 13:01:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 13:01:55 -0500
Subject: [LLVMbugs] [Bug 4732] codegen pass breaks machine cfg
In-Reply-To: <bug-4732-206@http.llvm.org/bugs/>
Message-ID: <200908181801.n7II1t93003350@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4732


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #5 from Dale Johannesen <dalej at apple.com>  2009-08-18 13:01:54 ---
"invalid cfg"?  You didn't think so when reviewing 78351, which introduced the
simpler CFG, so I'm skeptical this is well defined.  There are lots of other
places in that area that remove duplicate edges, and IMO that is the way it
should be throughout, although right now only BranchFolding does that.  It is
more efficient, possibly noticeably so in the case of large switches.

The real problem here is that the redundant conditional branches were not
removed when the edges were fixed up.  That's because AnalyzeBranch is not
handling JNE followed by JLE, and there are probably other bugs (missing
optimizations at least) lurking that are associated with that.  Its comments
say:

    // If [the branch conditions] differ, see if they fit one of the known
patterns.
    // Theoretically we could handle more patterns here, but
    // we shouldn't expect to see them if instruction selection
    // has done a reasonable job.

so where did JNE followed by JLE come from?


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 18 14:04:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 14:04:39 -0500
Subject: [LLVMbugs] [Bug 4738] New: Invalid strcmp optimization involving
 weak GV initializers
Message-ID: <bug-4738-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4738

           Summary: Invalid strcmp optimization involving weak GV
                    initializers
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


If a GlobalVariable is weak and has a constant initializer some optimizations
may make optimizations based on faulty assumptions, since the initializer may
be overridden at link time by a non-weak symbol.

One such case is simplifylibcalls, optimizing strcmp, but there may be more.

Here is the testcase for wrong strcmp optimization:
foo.c:
#include <string.h>
#include <stdlib.h>
const char fake_init[] __attribute__((weak)) = "y";

int main()
{
    if (strcmp(fake_init+1, "y") != 0)
        abort();
    return 0;
}

foo2.c:
const char fake_init[] = "yy";

llvm-gcc and clang fails the test:

$ llvm-gcc foo.c foo2.c && ./a.out
$ llvm-gcc foo.c foo2.c -O1 && ./a.out
Aborted
$ clang foo.c foo2.c && ./a.out
$ clang foo.c foo2.c -O1 && ./a.out
Aborted
$ gcc foo.c foo2.c && ./a.out
$ gcc foo.c foo2.c -O1 && ./a.out
$ gcc foo.c foo2.c -O2 && ./a.out

I think the easiest fix would be to make getInitializer() return the
initializer only if the GV cannot be overridden. Code that needs the
initializer for purposes other than optimization (such as backend) could use a
different method that gives you the real initializer.
Or we could check each use of getInitializer() to make sure its guarded by
hasDefinitiveInitializer, or !mayBeOverridden.

Some stats:
$ git grep getInitializer |grep -E Analysis\|Transforms|wc -l
46
$ git grep mayBeOverridden |wc -l
11
git grep hasDefinitiveInitializer |wc -l
7

So it seems there are 28 potential locations where GV->getInitializer is used
for optimization wrongly.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 18 15:09:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 15:09:18 -0500
Subject: [LLVMbugs] [Bug 4739] New: GlobalAlias of ConstantExpr emits
	incorrect code
Message-ID: <bug-4739-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4739

           Summary: GlobalAlias of ConstantExpr emits incorrect code
           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: csdavec at swan.ac.uk
                CC: llvmbugs at cs.uiuc.edu


A global alias may be defined as a GEP pointing to an element in the middle of
a structure, in which case the generated assembly is incorrect.  Consider this
simple example:

%0 = type { i32, i32 }
@structure = internal global %0 { i32 0, i32 1 }
@element1 = alias getelementptr( %0* @structure, i32 0, i32 1)
@.str = private constant [12 x i8] c"Value:  %d\0A\00"

declare i32 @printf(i8*, ...)

define i32 @main() {
entry:
  %tmp1 = load i32* getelementptr( %0* @structure, i32 0, i32 1)
  call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([12 x i8]* @.str,
i32 0, i32 0), i32 %tmp1); <i32> [#uses=0]
  %tmp2 = load i32* @element1
  call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([12 x i8]* @.str,
i32 0, i32 0), i32 %tmp2); <i32> [#uses=0]
  ret i32 0
}

Running this prints:
Value:  1
Value:  0

This is because the @element1 alias is being resolved at to the global
referenced in the GEP, not to the result of the GEP.  Looking at the generated
assembly, this is obvious:

    .globl  element1
    .set     element1, structure

This ought to be:
    .globl  element1
    .set     element1, structure+4

Making this change in the asm gives:
Value:  1
Value:  1

This is the correct output.

If there is some other way of emitting a symbol in the middle of an LLVM
structure so I can work around this bug, please let me know.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 18 19:23:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 19:23:24 -0500
Subject: [LLVMbugs] [Bug 4738] Invalid strcmp optimization involving weak GV
 initializers
In-Reply-To: <bug-4738-206@http.llvm.org/bugs/>
Message-ID: <200908190023.n7J0NOEb018862@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4738


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>  2009-08-18 19:23:24 ---
(In reply to comment #2)
> (In reply to comment #1)
> > The bug may exist, but the testcase is wrong.  Take a look at the IR:
> > 
> > @fake_init = weak_odr constant [2 x i8] c"y\00"   ; <[2 x i8]*> [#uses=2]
> > 
> > The linkage is ODR linkage, meaning that the initializer is reliable.
> > Why is llvm-gcc producing this?  It does it because gcc itself is happy
> > to do constant folding and other optimizations using the initial value of
> > weak constants.  Since constant folding occurs before IR generation, it
> > seemed pointless not to allow LLVM to do further optimizations using the
> > initial value.  There's been some discussion about this kind of transform
> > on the gcc mailing lists, and the conclusion was that it is valid.  LLVM
> > gives you more choice: generate weak_odr linkage if the front-end wants to
> > allow this kind of thing, otherwise generate weak linkage.
> 
> Well if this optimization is valid I'd like to at least see a warning about it,
> at least when using clang -Wall. 

The optimizer isn't currently well-positioned to emit warnings, but I
agree that it'd be nice if there were a way to warn when an invalid
assumption is encountered.

> > In the LLVM context it is valid to use the initializer if the linkage is
> > weak_odr, but not if it is weak.  Does simplifylibcalls still do the
> > transform if the linkage type of @fake_init is weak?
> > 
> 
> It does a different optimization then, which still aborts:
> -  %2 = call i32 @strcmp(i8* getelementptr ([2 x i8]* @fake_init, i32 0, i64
> 1), i8* getelementptr ([2 x i8]* @.str, i64 0, i64 0)) nounwind readonly; <i32>
> [#uses=1]
> +  %strcmpload = load i8* getelementptr ([2 x i8]* @.str, i64 0, i64 0); <i8>
> [#uses=1]
> +  %2 = zext i8 %strcmpload to i32                 ; <i32> [#uses=1]
> 
> It checks only the first byte (because it knowns the string length).

This is fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084902.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 Aug 18 19:27:08 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 19:27:08 -0500
Subject: [LLVMbugs] [Bug 4736] ice: Binary operator types must match!
In-Reply-To: <bug-4736-206@http.llvm.org/bugs/>
Message-ID: <200908190027.n7J0R8Ua019344@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4736


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-18 19:27:07 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084897.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 Aug 18 20:23:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 20:23:56 -0500
Subject: [LLVMbugs] [Bug 4740] New: typo in llvm-gcc/config/arm/lib1funcs.asm
Message-ID: <bug-4740-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4740

           Summary: typo in llvm-gcc/config/arm/lib1funcs.asm
           Product: tools
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: timm at apple.com
                CC: llvmbugs at cs.uiuc.edu


line 187: #if definded (__thumb2__)

discovered while trying to cross compile for arm


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 18 20:39:23 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Aug 2009 20:39:23 -0500
Subject: [LLVMbugs] [Bug 4740] typo in llvm-gcc/config/arm/lib1funcs.asm
In-Reply-To: <bug-4740-206@http.llvm.org/bugs/>
Message-ID: <200908190139.n7J1dNcD028681@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4740


Bill Wendling <wendling at apple.com> changed:

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




--- Comment #1 from Bill Wendling <wendling at apple.com>  2009-08-18 20:39:23 ---
Fixed. Thanks!

Sending        llvm-gcc.src/gcc/config/arm/lib1funcs.asm
Transmitting file data .
Committed revision 79398.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 00:29:02 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 00:29:02 -0500
Subject: [LLVMbugs] [Bug 4741] New: libpcre-7.9-r1 build errors gentoo
Message-ID: <bug-4741-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4741

           Summary: libpcre-7.9-r1 build errors gentoo
           Product: libraries
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fedor.sakharov at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Hello. 

As I have been trying to update world on my gentoo box, I encoutered this build
error, that asked to submit a full bug report here. 

i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I.     -O2 -march=i686 -pipe -MT
pcregrep.o -MD -MP -MF .deps/pcregrep.Tpo -c -o pcregrep.o pcregrep.c           
pcrecpp.cc: In member function 'void pcrecpp::RE::Init(const std::string&,
const pcrecpp::RE_Options*)':                        
pcrecpp.cc:85: internal compiler error: in set_mem_attributes_minus_bitpos, at
emit-rtl.c:1533                                  
Please submit a full bug report,                                                
with preprocessed source if appropriate.                                        
See <URL:http://llvm.org/bugs/> for instructions.                               
make[1]: *** [pcrecpp.lo] Error 1                                               
make[1]: *** Waiting for unfinished jobs....                                    
mv -f .deps/pcregrep.Tpo .deps/pcregrep.Po                                      
mv -f .deps/pcretest.Tpo .deps/pcretest.Po                                      
make[1]: Leaving directory
`/var/tmp/portage/dev-libs/libpcre-7.9-r1/work/pcre-7.9'                        
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  Wed Aug 19 03:54:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 03:54:09 -0500
Subject: [LLVMbugs] [Bug 4737] possible clang wrong code bug
In-Reply-To: <bug-4737-206@http.llvm.org/bugs/>
Message-ID: <200908190854.n7J8s9Ea031093@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4737


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #4 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-19 03:54:09 ---
Fixed in r79415.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 07:33:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 07:33:36 -0500
Subject: [LLVMbugs] [Bug 4742] New: [inline asm] `%a' template does not work
Message-ID: <bug-4742-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4742

           Summary: [inline asm] `%a' template does not work
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3338)
 --> (http://llvm.org/bugs/attachment.cgi?id=3338)
ZIP archive with C and BC files

In LLVM r79290:

LLVM ERROR: Invalid operand found in inline asm: 'movw  %gs:${1:a}, ${0:w}'
INLINEASM <es:movw      %gs:${1:a}, ${0:w}>, 10, %AX<def>, 11, 6, 14,
%EFLAGS<earlyclobber,def,dead>


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 08:44:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 08:44:22 -0500
Subject: [LLVMbugs] [Bug 4743] New: Linker warnings "resolving _Xxx by
 linking to _Xxx@nn" when compiling W32 programs
Message-ID: <bug-4743-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4743

           Summary: Linker warnings "resolving _Xxx by linking to _Xxx at nn"
                    when compiling W32 programs
           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: d.g.gorbachev at gmail.com
                CC: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


Created an attachment (id=3339)
 --> (http://llvm.org/bugs/attachment.cgi?id=3339)
ZIP archive with examples

LLVM-GCC and LLVM r79290.

Warning: resolving _ExitProcess by linking to _ExitProcess at 4
Use --enable-stdcall-fixup to disable these warnings
Use --disable-stdcall-fixup to disable these fixups

Notice ".def _ExitProcess;", while MinGW32-GCC generates ".def _ExitProcess at 4;"
or nothing.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 09:12:13 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 09:12:13 -0500
Subject: [LLVMbugs] [Bug 4743] Linker warnings "resolving _Xxx by linking to
 _Xxx@nn" when compiling W32 programs
In-Reply-To: <bug-4743-206@http.llvm.org/bugs/>
Message-ID: <200908191412.n7JECDYH010117@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4743


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

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




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-19 09:12:13 ---


*** This bug has been marked as a duplicate of bug 2936 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 12:58:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 12:58:09 -0500
Subject: [LLVMbugs] [Bug 4742] [inline asm] `%a' template does not work
In-Reply-To: <bug-4742-206@http.llvm.org/bugs/>
Message-ID: <200908191758.n7JHw9nE010369@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4742


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #1 from Dale Johannesen <dalej at apple.com>  2009-08-19 12:58:08 ---
The .c and .bc files don't match (most obviously, main vs foo) and I can't see
how you could get that asm IR out of that source line.  Translating "r,r" to
"ir" is wrong, and I'm 99% sure llvm-gcc doesn't do that.  Anyway, you should
probably look at your front end.  Darwin produces this, for example:
%asmtmp = call i32 asm sideeffect "mov $1, $0\0A",
"=r,r,~{dirflag},~{fpsr},~{flags}"(i32 %1) nounwind


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 15:11:38 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 15:11:38 -0500
Subject: [LLVMbugs] [Bug 4742] [inline asm] `%a' template does not work
In-Reply-To: <bug-4742-206@http.llvm.org/bugs/>
Message-ID: <200908192011.n7JKBcX7027368@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4742


Dmitry Gorbachev <d.g.gorbachev at gmail.com> changed:

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




--- Comment #2 from Dmitry Gorbachev <d.g.gorbachev at gmail.com>  2009-08-19 15:11:38 ---
Hi,

The .c and .bc files do match, there's no any foo.
Please try to compile it here: http://llvm.org/demo/index.cgi

; ModuleID = '/tmp/webcompile/_23568_0.bc'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
target triple = "i386-pc-linux-gnu"

define i32 @main() nounwind {
entry:
        %asmtmp.i = tail call i16 asm "movw\09%gs:${1:a}, ${0:w}",
"=r,ir,~{dirflag},~{fpsr},~{flags}"(i32 6) nounwind          ; <i16> [#uses=1]
        %0 = zext i16 %asmtmp.i to i32          ; <i32> [#uses=1]
        ret i32 %0
}

Both llvm-gcc and clang give analogous results.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 16:32:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 16:32:18 -0500
Subject: [LLVMbugs] [Bug 4744] New: error compiling multilib for llvm-gcc as
 a cross-compiler for arm
Message-ID: <bug-4744-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4744

           Summary: error compiling multilib for llvm-gcc as a cross-
                    compiler for arm
           Product: tools
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: timm at apple.com
                CC: llvmbugs at cs.uiuc.edu


Using binutils 2.18 and newlib 1.17

Configure args:
timMac:llvm-gcc-build timm$ ../llvm-gcc-4.2.1/configure
--prefix=/usr/local/llvm/armtools --program-prefix=llvm-
--enable-llvm=/Users/timm/armtools-build/llvm-build/ --enable-languages=c++
--host=i686-apple-darwin9 --build=i686-apple-darwin9 --target=arm-elf
--with-gxx-include-dir=/usr/include/c++/4.0.0 --enable-interwork
--enable-multilib --with-newlib
--with-header=../newlib-1.17.0/newlib/libc/include

Results:
timMac:llvm-gcc-build timm$ make all-gcc BUILDOPTIONS=LLVM_VERSION_INFO=2.5
...
/usr/local/llvm/armtools/arm-elf/bin/ar  rc ./libgcov.a libgcc/./_gcov.o
libgcc/./_gcov_merge_add.o libgcc/./_gcov_merge_single.o
libgcc/./_gcov_merge_delta.o libgcc/./_gcov_fork.o libgcc/./_gcov_execl.o
libgcc/./_gcov_execlp.o libgcc/./_gcov_execle.o libgcc/./_gcov_execv.o
libgcc/./_gcov_execvp.o libgcc/./_gcov_execve.o
libgcc/./_gcov_interval_profiler.o libgcc/./_gcov_pow2_profiler.o
libgcc/./_gcov_one_value_profiler.o
/usr/local/llvm/armtools/arm-elf/bin/ranlib ./libgcov.a
make GCC_FOR_TARGET="/Users/timm/armtools-build/llvm-gcc-build/./gcc/xgcc
-B/Users/timm/armtools-build/llvm-gcc-build/./gcc/
-B/usr/local/llvm/armtools/arm-elf/bin/ -B/usr/local/llvm/armtools/arm-elf/lib/
-isystem /usr/local/llvm/armtools/arm-elf/include -isystem
/usr/local/llvm/armtools/arm-elf/sys-include" \
          AR_FOR_TARGET="/usr/local/llvm/armtools/arm-elf/bin/ar" \
          AR_CREATE_FOR_TARGET="/usr/local/llvm/armtools/arm-elf/bin/ar  rc" \
          AR_EXTRACT_FOR_TARGET="/usr/local/llvm/armtools/arm-elf/bin/ar  x" \
          AR_FLAGS_FOR_TARGET="" \
          CC="gcc -m32" CFLAGS="-g -O2 -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition
-Wmissing-format-attribute  " \
          BUILD_PREFIX="" \
          BUILD_PREFIX_1="" \
          LANGUAGES="" \
          LIBGCC2_CFLAGS="-O2  -O2 -g -O2  -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE
  -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fno-inline -g  -DIN_LIBGCC2
-D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -mthumb "  \
          LIBGCC2_STATIC_CFLAGS=""  \
          MULTILIB_CFLAGS=" -mthumb" T=thumb/ thumb/crtbegin.o thumb/crtend.o
thumb/crti.o thumb/crtn.o
make[3]: `thumb/crtbegin.o' is up to date.
make[3]: `thumb/crtend.o' is up to date.
make[3]: `thumb/crti.o' is up to date.
make[3]: `thumb/crtn.o' is up to date.
/Users/timm/armtools-build/llvm-gcc-build/./gcc/xgcc
-B/Users/timm/armtools-build/llvm-gcc-build/./gcc/
-B/usr/local/llvm/armtools/arm-elf/bin/ -B/usr/local/llvm/armtools/arm-elf/lib/
-isystem /usr/local/llvm/armtools/arm-elf/include -isystem
/usr/local/llvm/armtools/arm-elf/sys-include  -O2  -O2 -g -O2  -DIN_GCC
-DCROSS_DIRECTORY_STRUCTURE   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -Wold-style-definition  -isystem ./include  -fno-inline -g
 -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -I. -I.
-I../../llvm-gcc-4.2.1/gcc -I../../llvm-gcc-4.2.1/gcc/.
-I../../llvm-gcc-4.2.1/gcc/../include -I./../intl
-I../../llvm-gcc-4.2.1/gcc/../libcpp/include 
-I../../llvm-gcc-4.2.1/gcc/../libdecnumber -I../libdecnumber
-I/Users/timm/armtools-build/llvm-build//include
-I/Users/timm/armtools-build/llvm/include -mthumb  -fexceptions -c
../../llvm-gcc-4.2.1/gcc/unwind-dw2-fde.c -o libgcc/thumb/unwind-dw2-fde.o
/var/folders/bG/bGtEFVXRFvunPKmCQBG6o++++TI/-Tmp-//cc2YwPhz.s: Assembler
messages:
/var/folders/bG/bGtEFVXRFvunPKmCQBG6o++++TI/-Tmp-//cc2YwPhz.s:96: Error:
selected processor does not support `sxtb r5,r5'
/var/folders/bG/bGtEFVXRFvunPKmCQBG6o++++TI/-Tmp-//cc2YwPhz.s:537: Error:
selected processor does not support `sxtb r6,r6'
/var/folders/bG/bGtEFVXRFvunPKmCQBG6o++++TI/-Tmp-//cc2YwPhz.s:705: Error:
selected processor does not support `sxtb r1,r1'
/var/folders/bG/bGtEFVXRFvunPKmCQBG6o++++TI/-Tmp-//cc2YwPhz.s:711: Error:
selected processor does not support `sxtb r1,r1'
make[2]: *** [libgcc/thumb/unwind-dw2-fde.o] Error 1
make[1]: *** [stmp-multilib] Error 2
make: *** [all-gcc] Error 2

Let me know if you need more info or if I am just configuring something
incorrectly


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 19 17:52:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Aug 2009 17:52:41 -0500
Subject: [LLVMbugs] [Bug 4742] [inline asm] `%a' template does not work
In-Reply-To: <bug-4742-206@http.llvm.org/bugs/>
Message-ID: <200908192252.n7JMqfte015304@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4742


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #4 from Dale Johannesen <dalej at apple.com>  2009-08-19 17:52:41 ---
Having done that, I thought I should try to fix it:)  Patch 79484 gets the
syntax right for Darwin.  If Linux is not quite the same this is where to
change 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  Thu Aug 20 04:08:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Aug 2009 04:08:24 -0500
Subject: [LLVMbugs] [Bug 4745] New: Add recoverable string parsing errors to
 APInt and APFloat
Message-ID: <bug-4745-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4745

           Summary: Add recoverable string parsing errors to APInt and
                    APFloat
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: idadesub at users.sourceforge.net
                CC: llvmbugs at cs.uiuc.edu


This is a bug to track the discussion we had on llvm-dev and irc about adding
recoverable string parsing to APInt and APFloat. I've attached the first
version to implementing this for APFloat. It converts most of the asserts in
the string parsing code to return "opInvalidString" if it finds an invalid
character.

Two open questions. First, in order to be efficient, I created a constructor
that can create an empty APFloat. Then, you can call "convertFromString" to
parse the string. This is a little ugly, however. Does anyone have a better
suggestion?

Second, it only has one error code, opInvalidString. Should we expose more
details, like "does not have exponent", or "too many dots in number", or is
opInvalidString sufficient?

Fortunately, it looks like most of the exposure of this patch is pretty small,
as only ConstantFP::get actually creates a APFloat from a string in the
codebase.

If this looks good, I'll do the same with APInt, which I expect will be a bit
simpler to do.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 20 10:43:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Aug 2009 10:43:26 -0500
Subject: [LLVMbugs] [Bug 4747] New: assertion failed: setOperand() out of
	range!
Message-ID: <bug-4747-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4747

           Summary: assertion failed: setOperand() out of range!
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3342)
 --> (http://llvm.org/bugs/attachment.cgi?id=3342)
testcase (from SingleSource/CoyoteBench/fftbench)

This is a regression from something that changed yesterday.  It hit twice in
the armv6 nightly tests.  The attached testcase is not bugpoint-reduced because
when I reduce it, it hits another problem.  Run this with llc (no options
needed) and you get:

Assertion failed: (i_nocapture < OperandTraits<PHINode>::operands(this) &&
"setOperand() out of range!"), function setOperand, file
/Users/bwilson/local/llvm/llvm/include/llvm/Instructions.h, line 2002.
0   llc               0x00b03284 PrintStackTrace(void*) + 45
1   llc               0x00b0388c SignalHandler(int) + 410
2   libSystem.B.dylib 0x98115bfb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1743692847
4   libSystem.B.dylib 0x981a3ba5 raise + 26
5   libSystem.B.dylib 0x981b9c5c abort + 93
6   libSystem.B.dylib 0x981a6804 __pthread_markcancel + 0
7   llc               0x00a026c1 llvm::PHINode::setOperand(unsigned int,
llvm::Value*) + 89
8   llc               0x00a02714 llvm::PHINode::setIncomingBlock(unsigned int,
llvm::BasicBlock*) + 34
9   llc               0x0093da5d llvm::SplitCriticalEdge(llvm::TerminatorInst*,
unsigned int, llvm::Pass*, bool) + 619
10  llc               0x0091ebcd SplitEdgeNicely(llvm::TerminatorInst*,
unsigned int, llvm::SmallSet<std::pair<llvm::BasicBlock const*,
llvm::BasicBlock const*>, 8u>&, llvm::Pass*) + 1139
11  llc               0x0091eff0 (anonymous
namespace)::CodeGenPrepare::OptimizeBlock(llvm::BasicBlock&) + 228
12  llc               0x0091ff63 (anonymous
namespace)::CodeGenPrepare::runOnFunction(llvm::Function&) + 137
13  llc               0x00a846c3
llvm::FPPassManager::runOnFunction(llvm::Function&) + 297
14  llc               0x00a84bf6
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 112
15  llc               0x00a84dc2
llvm::FunctionPassManager::run(llvm::Function&) + 130
16  llc               0x000301d0 main + 3577
17  llc               0x0002ecc9 start + 53
Stack dump:
0.      Program arguments: /Users/bwilson/local/llvm/install/bin/llc
fftbench.llvm.bc 
1.      Running pass 'Optimize for code generation' on function '@main'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 20 10:45:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Aug 2009 10:45:45 -0500
Subject: [LLVMbugs] [Bug 4739] GlobalAlias of ConstantExpr emits incorrect
	code
In-Reply-To: <bug-4739-206@http.llvm.org/bugs/>
Message-ID: <200908201545.n7KFjjdk026132@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4739


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #19 from Chris Lattner <clattner at apple.com>  2009-08-20 10:45:45 ---
You can't do stuff like this on the mac.  Public symbols are considered to
separate blocks of code and data, and the linker reorders those blocks.  This
is why the mac doesn't support "multiple entry points" in the obvious way etc. 
In any case, I think this bug is invalid, GlobalAlias should only except an
offset of zero from the aliasee.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 20 12:31:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Aug 2009 12:31:17 -0500
Subject: [LLVMbugs] [Bug 4748] New: Instcombine transform confuse basicaa
Message-ID: <bug-4748-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4748

           Summary: Instcombine transform confuse basicaa
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan at fallingsnow.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3343)
 --> (http://llvm.org/bugs/attachment.cgi?id=3343)
File with bad alias info

Attached is an .ll file with 2 functinons which show the problem. @good shows
code as it's commonly emitted by a frontend,  a BC to a GEP. There is an
instcombine transform that converts this to a form like @bad, a GEP to a BC.

The issue is that the GEP to BC form isn't understood by basicaa, so quite a
bit of alias information is lost by this transform. Here is the AA results for
the file:

kendall :: git/rbx ?? llvm-as < bad.ll | opt -aa-eval
-print-all-alias-modref-info -disable-output
Function: bad: 5 pointers, 0 call sites
  MustAlias:    %struct.A* %A, %struct.B* %B
  MayAlias:     %struct.A* %A, %struct.B* %C
  MayAlias:     %struct.B* %B, %struct.B* %C
  MayAlias:     %struct.A* %A, i32* %X
  MayAlias:     %struct.B* %B, i32* %X
  MustAlias:    %struct.B* %C, i32* %X
  MayAlias:     %struct.A* %A, i32* %Y
  NoAlias:      %struct.B* %B, i32* %Y
  MayAlias:     %struct.B* %C, i32* %Y
  MayAlias:     i32* %X, i32* %Y
Function: good: 5 pointers, 0 call sites
  MustAlias:    %struct.A* %A, %struct.B* %B
  MustAlias:    %struct.A* %A, %struct.A* %Q
  MustAlias:    %struct.A* %Q, %struct.B* %B
  MayAlias:     %struct.A* %A, i32* %X
  MayAlias:     %struct.B* %B, i32* %X
  MayAlias:     %struct.A* %Q, i32* %X
  MayAlias:     %struct.A* %A, i32* %Y
  NoAlias:      %struct.B* %B, i32* %Y
  MayAlias:     %struct.A* %Q, i32* %Y
  MustAlias:    i32* %X, i32* %Y
===== Alias Analysis Evaluator Report =====
  20 Total Alias Queries Performed
  2 no alias responses (10.0%)
  12 may alias responses (60.0%)
  6 must alias responses (30.0%)
  Alias Analysis Evaluator Pointer Alias Summary: 10%/60%/30%
  Alias Analysis Mod/Ref Evaluator Summary: no mod/ref!

You can see that critical alias information between %X and %Y is lost in the
transform. I've been advised that the instcombine transform is actually
advantageous in same cases, so I think that basicaa needs to be taught about
this transformed form so it can produce the same aliasing information as the
original form.

In my case, this is an ugly bug, as code that would normally be nicely
optimized is now not, because a number of passes depend on replacing the
redundancy of %X and %Y.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 21 01:37:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 01:37:42 -0500
Subject: [LLVMbugs] [Bug 3039] llvm-gcc42's README.LLVM has the wrong target
 triple for leopard
In-Reply-To: <bug-3039-206@http.llvm.org/bugs/>
Message-ID: <200908210637.n7L6bgLY010020@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3039


Erick Tryzelaar <idadesub at users.sourceforge.net> 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  Fri Aug 21 06:11:16 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 06:11:16 -0500
Subject: [LLVMbugs] [Bug 4530] llvm-ld does not merge debug information
	correctly
In-Reply-To: <bug-4530-206@http.llvm.org/bugs/>
Message-ID: <200908211111.n7LBBGaV024849@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4530


Richard Pennington <rich at pennware.com> changed:

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




--- Comment #6 from Richard Pennington <rich at pennware.com>  2009-08-21 06:11:16 ---
Resolved in r79616


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 21 15:01:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 15:01:41 -0500
Subject: [LLVMbugs] [Bug 4749] New: StrongPhiElimination losts a copy
Message-ID: <bug-4749-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4749

           Summary: StrongPhiElimination losts a copy
           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: yuanlin2004 at gmail.com
                CC: llvmbugs at cs.uiuc.edu, yuanlin2004 at gmail.com


Take the following test case tt.c,

--------------------------------------
void foo( int *pa,  int *pb)
{
    int x;
    int y;

    x = *pa;
    y = *pb;

    while(*pa) {
        y = x;
        x = 0;
        pa ++;
    }

    *pb = y;
}

--------------------------------------

>~/llvm/trunk/llvm/Debug/bin/clang-cc tt.c --emit-llvm-bc -o tt.bc
>~/llvm/trunk/llvm/Debug/bin/opt -O3 tt.bc -o tt.opt.bc
>~/llvm/trunk/llvm/Debug/bin/llc -march=x86 tt.opt.bc -o - -strong-phi-elim
        .file   "tt.opt.bc"

        .text
        .align  16
        .globl  foo
        .type   foo, at function
foo:                                                        # @foo
.LBB1_0:                                                    # %entry
        pushl   %esi
        movl    8(%esp), %eax
        movl    (%eax), %ecx
        testl   %ecx, %ecx
        movl    12(%esp), %edx
        je      .LBB1_3
        .align  16
.LBB1_1:                                                    # %while.body
                                                            # Loop Depth 1
                                                            # Loop Header
                                                            # Inner Loop
        movl    %ecx, %esi
        cmpl    $0, 4(%eax)
        leal    4(%eax), %eax
        movl    $0, %ecx               <--------------- clears ecx
        jne     .LBB1_1
.LBB1_2:                                                    # %while.end
        movl    %ecx, (%edx)
        popl    %esi
        ret
.LBB1_3:                                                    #
%entry.while.end_crit_edge
        movl    (%edx), %ecx
        jmp     .LBB1_2
        .size   foo, .-foo

        .section        .note.GNU-stack,"", at progbits


Notice that the value of %ecx (which corresponds to 'y' after the loop) always
becomes 0 if the while loop body is ever entered. This is incorrect.


The following is my initial assess of the problem:

Hello, I've encountered a problem similar to 'lost-copy' when using
the StrongPhiElimination and wonder whether it is a incompatibility
issue between the two different algorithms used in
StrongPhiElimination.cpp. The StrongPhiElimination is mostly based on
the algorithm in Zoran Budimilic et al's "Fast Copy Coalescing and
Live-Range Identification" ([1]), while the 'Copy Insertion' part is
from Preston Briggs et al's "Practical Improvements to the
Construction and Destruction of Static Single Assignment Form" ([2]).

The test case I have is as follows,

Block 1
 r1 = ...
 if (...) goto Block 2
 else goto Block 3

Block 2
 r2 = ...
 goto Block 4

Block 3
 r3 = phi(r1, r4)
 r4 = ...
 if (...) goto Block 3
 else goto Block 4

Block 4
 r5 = phi(r2, r3)
 ... = r5


Since the live range of r2, r3, and r5 do not interference,
StrongPhiElimination decides, before InsertCopies(), to rename both r2
and r3 to r5 based onthe algorithm in [1]

During InsertCopies(), because 'r3' is live after Block 3 and this is
the typical case where 'lost-copy' could happen, StrongPhiEilimination
inserts a copy from r3 to a temp register r99 based on the algorithm
in [2]. Block 3 then becomes

Block 3
  r3 = phi(r1, r4)
  r99 = r3
  r4 = ...
  r3 = r4
  if (...) goto Block 3
  else (...) goto Block 4

The Insert-copy algorithm in [2] would normally rename r3 in Block 4
to r99. But StrongPhiEliminination decides unify 'r5', 'r2' and 'r3'
instead of inserting copies for 'r5=phi(r2, r3)' in Block 4, therefore
the 'waiting' set for Block 4 is empty. 'r3' is not renamed to 'r99'
in StrongPhiElimination.

Then StrongPhiElimination proceeds to the renaming phase and rename
both r2 and r3 to r5. The final code becomes

Block 1
 r1 = ...
 r5 = r1
 if (...) goto Block 2
 else goto Block 3

Block 2
 r5 = ...
 goto Block 4

Block 3
  r99 = r5
  r4 = ...
  r5 = r4
  if (...) goto Block 3
  else (...) goto Block 4

Block 4
 ... = r5

The value of r5 in Block 4 is wrong when the path Block 1->Block
3->Block 4 is taken.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 21 16:12:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 16:12:36 -0500
Subject: [LLVMbugs] [Bug 3284] APFloat doesn't fully support fp128
In-Reply-To: <bug-3284-206@http.llvm.org/bugs/>
Message-ID: <200908212112.n7LLCapf004277@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3284


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #3 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-21 16:12:35 ---


*** This bug has been marked as a duplicate of bug 2555 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 21 17:19:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 17:19:24 -0500
Subject: [LLVMbugs] [Bug 2555] APFloat does not support IEEEquad
In-Reply-To: <bug-2555-206@http.llvm.org/bugs/>
Message-ID: <200908212219.n7LMJOOF011095@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2555


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

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




--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-21 17:18:53 ---
Implemented in r79676. Neil, please review!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 21 18:32:47 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Aug 2009 18:32:47 -0500
Subject: [LLVMbugs] [Bug 4744] error compiling multilib for llvm-gcc as a
 cross-compiler for arm
In-Reply-To: <bug-4744-206@http.llvm.org/bugs/>
Message-ID: <200908212332.n7LNWluR020344@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4744


Timm Hannon <timm at apple.com> changed:

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




--- Comment #3 from Timm Hannon <timm at apple.com>  2009-08-21 18:32:45 ---
setting it to an architecture that supported those instructions solved the
problem


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug 22 04:19:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 04:19:22 -0500
Subject: [LLVMbugs] [Bug 4750] New: MC/MachO tests fail on Solaris.
Message-ID: <bug-4750-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4750

           Summary: MC/MachO tests fail on Solaris.
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: blocker
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


FAIL: /export/home2/edward/lab/llvm/llvm/test/MC/MachO/data.s
Failed with exit(1) at line 1
while running: llvm-mc /export/home2/edward/lab/llvm/llvm/test/MC/MachO/data.s
-filetype=obj -o - | macho-dump | FileCheck
/export/home2/edward/lab/llvm/llvm/test/MC/MachO/data.s
/export/home2/edward/lab/llvm/llvm/test/MC/MachO/data.s:17:11: error: expected
string not found in input
// CHECK: ('cputype', 7)
          ^
<stdin>:1:1: note: scanning from here
('cputype', 7L)
^

FAIL: /export/home2/edward/lab/llvm/llvm/test/MC/MachO/sections.s
Failed with exit(1) at line 1
while running: llvm-mc
/export/home2/edward/lab/llvm/llvm/test/MC/MachO/sections.s -filetype=obj -o -
| macho-dump | FileCheck
/export/home2/edward/lab/llvm/llvm/test/MC/MachO/sections.s
/export/home2/edward/lab/llvm/llvm/test/MC/MachO/sections.s:46:11: error:
expected string not found in input
// CHECK: ('cputype', 7)
          ^
<stdin>:1:1: note: scanning from here
('cputype', 7L)
^

Should block llvm 2.6 release.

Cheers,
Edward.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Aug 22 04:29:15 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 04:29:15 -0500
Subject: [LLVMbugs] [Bug 4750] MC/MachO tests fail on Solaris.
In-Reply-To: <bug-4750-206@http.llvm.org/bugs/>
Message-ID: <200908220929.n7M9TFcQ006350@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4750


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-22 04:29:11 ---
Fixed in r79736.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 22 08:53:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 08:53:42 -0500
Subject: [LLVMbugs] [Bug 4751] New: [inline asm] `%a' template: LLVM ERROR:
 Invalid operand found in inline asm
Message-ID: <bug-4751-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4751

           Summary: [inline asm] `%a' template: LLVM ERROR: Invalid operand
                    found in inline asm
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3353)
 --> (http://llvm.org/bugs/attachment.cgi?id=3353)
C and BC files


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 22 09:14:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 09:14:22 -0500
Subject: [LLVMbugs] [Bug 4752] New: [inline asm] Assertion failed: "
 Interval does not exist for register" / LLVM ERROR: Ran out of registers
 during register allocation!
Message-ID: <bug-4752-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4752

           Summary: [inline asm] Assertion failed: "Interval does not exist
                    for register" / LLVM ERROR: Ran out of registers during
                    register allocation!
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3354)
 --> (http://llvm.org/bugs/attachment.cgi?id=3354)
ZIP archive with C and BC files

Two errors, depending on optimization level. LLVM r79531.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 22 16:00:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 16:00:25 -0500
Subject: [LLVMbugs] [Bug 3240] Add *AsmPrinterForceLink to the targets which
 don't have them
In-Reply-To: <bug-3240-206@http.llvm.org/bugs/>
Message-ID: <200908222100.n7ML0P4w027664@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3240


Dan Villiom Podlaski Christiansen <danchr at gmail.com> changed:

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




--- Comment #5 from Dan Villiom Podlaski Christiansen <danchr at gmail.com>  2009-08-22 16:00:24 ---
This bug is no longer relevant.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 22 21:52:04 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Aug 2009 21:52:04 -0500
Subject: [LLVMbugs] [Bug 4753] New: ice: Can't find corresponding
	super-register!
Message-ID: <bug-4753-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4753

           Summary: ice: Can't find corresponding super-register!
           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


Seen on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp185$ llvm-gcc -O -c -w small.c
cc1: VirtRegRewriter.cpp:2274:
void<unnamed>::LocalRewriter::RewriteMBB(llvm::MachineBasicBlock&,
llvm::VirtRegMap&, llvm::LiveIntervals*, <unnamed>::AvailableSpills&,
llvm::BitVector&, std::vector<llvm::MachineOperand*,
std::allocator<llvm::MachineOperand*> >&): Assertion `SuperReg &&
TRI->getSubReg(SuperReg, SubIdx) == PhysReg && "Can't find corresponding
super-register!"' failed.
small.c:35: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp185$ llvm-gcc -v

Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/z/tmp/llvm-gcc-r79741-install
--program-prefix=llvm-r79741- --enable-languages=c,c++
--enable-llvm=/home/regehr/z/tmp/llvm-r79741 --enable-checking=release
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)

regehr at john-home:~/volatile/tmp185$ cat small.c

char foo (int _si1, char _si2)
{
}

char bar (int _ui1, char _ui2)
{
}

volatile uint8;
char g_21;
char g_80;

int func_24 (unsigned short p_26, char p_27);

char func_53 (char p_54, char p_56, char p_57, char p_59);

int safe (int);

uint80 (char p_52)
{
  char l_55;
  char l_89;
  if (func_53
      (safe (~1), 1,
       func_24 (bar, 1 && p_52) != (func_24 (l_89, 1)
                                                           && g_80),
       (1 | func_24 (p_52, foo)) ^ p_52 ^ func_24 (l_55,
                                                                         safe
                                                                         (1))))
    uint8;
  uint8;
  if (func_24 (g_21, 1))
    for (0; 1; safefuncts (p_52, 1))
      uint8;
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 01:58:25 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 01:58:25 -0500
Subject: [LLVMbugs] [Bug 4754] New: Internal compiler error. Varargs
 function with fastcall calling convention.
Message-ID: <bug-4754-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4754

           Summary: Internal compiler error. Varargs function with fastcall
                    calling convention.
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3359)
 --> (http://llvm.org/bugs/attachment.cgi?id=3359)
C and LL files

Both LLVM-GCC and CLANG are affected. "__attribute__((__stdcall__))" does not
cause errors. FSF GCC silently ignores "__attribute__((__fastcall__))".

LLVM r79531.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 02:04:56 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 02:04:56 -0500
Subject: [LLVMbugs] [Bug 4754] Internal compiler error. Varargs function
 with fastcall calling convention.
In-Reply-To: <bug-4754-206@http.llvm.org/bugs/>
Message-ID: <200908230704.n7N74ugD005916@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4754


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-23 02:04:56 ---
I think it's FSF GCC bug, we're doing correct thing rejecting it. Varargs
fastcall function usually indicates user bug.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 04:48:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 04:48:39 -0500
Subject: [LLVMbugs] [Bug 4672] CodeGen/SystemZ ,
 CodeGen/Thumb & CodeGen/  Thumb2 fail test on Solaris.
In-Reply-To: <bug-4672-206@http.llvm.org/bugs/>
Message-ID: <200908230948.n7N9md8l028298@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4672


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-08-23 04:48:39 ---
Fixed in r79846.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 04:48:42 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 04:48:42 -0500
Subject: [LLVMbugs] [Bug 4711] Thumb2 tests fail horrifyingly on Solaris.
In-Reply-To: <bug-4711-206@http.llvm.org/bugs/>
Message-ID: <200908230948.n7N9mg6t028335@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4711


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-08-23 04:48:42 ---
Fixed in r79846.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 07:32:29 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 07:32:29 -0500
Subject: [LLVMbugs] [Bug 4756] New: "Undefined reference" errors when
 linking a C++ program with static objects
Message-ID: <bug-4756-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4756

           Summary: "Undefined reference" errors when linking a C++ program
                    with static objects
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3362)
 --> (http://llvm.org/bugs/attachment.cgi?id=3362)
C file

Cross LLVM-GCC (host i686-pc-linux-gnu, target i686-pc-mingw32), r79531

undefined reference to `___dso_handle'
undefined reference to `___cxa_atexit'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 07:45:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 07:45:00 -0500
Subject: [LLVMbugs] [Bug 4756] "Undefined reference" errors when linking a
 C++ program with static objects
In-Reply-To: <bug-4756-206@http.llvm.org/bugs/>
Message-ID: <200908231245.n7NCj059017770@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4756


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-23 07:44:59 ---
cxa_atexit-style destructors are not supported on mingw (well, they actually
are, but in non-standard way - we cannot backport these patches from mainline
gcc due to GPLv3). Unfortunately cross-builds by default assumes that they are
available (this is default gcc behaviour).

You need to configure with --enable-__cxa_atexit=no


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug 23 08:03:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 08:03:59 -0500
Subject: [LLVMbugs] [Bug 4753] ice: Can't find corresponding super-register!
In-Reply-To: <bug-4753-206@http.llvm.org/bugs/>
Message-ID: <200908231303.n7ND3xHG020110@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4753


Jakob Stoklund Olesen <stoklund at 2pi.dk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |stoklund at 2pi.dk
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Jakob Stoklund Olesen <stoklund at 2pi.dk>  2009-08-23 08:03:58 ---
Fixed here: http://llvm.org/viewvc/llvm-project?rev=79855&view=rev

ReuseInfo::GetRegForReload() was using the wrong sub-register index when
undoing a reuse. A MachineOperand with a physreg has no sub-register index - it
is only used for virtual registers. This meant that %AL was illegally replaced
with %ECX when it should have been replaced with %CL.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 18:46:10 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 18:46:10 -0500
Subject: [LLVMbugs] [Bug 4758] New: cmake out of tree build.
Message-ID: <bug-4758-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4758

           Summary: cmake out of tree build.
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: Solaris
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: eocallaghan at auroraux.org
                CC: llvmbugs at cs.uiuc.edu


Good day,

CMake out of tree build system for llvm is completely broken on Solaris.
For some reason, which I have only a idea what the problem is, a whole range of
functions can not be found, for example:

/export/home2/edward/lab/llvm/cmake/llvm/lib/System/Unix/Process.inc:79:2:
warning: #warning Cannot get malloc info on this platform

Can someone with Linux please confirm that it is indeed broken on Linux as
well. This could allow me to figure out exactly what is going on between the
autohell and cmake versions of config.h

Thanks,
Edward.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug 23 19:42:46 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 19:42:46 -0500
Subject: [LLVMbugs] [Bug 4758] cmake out of tree build.
In-Reply-To: <bug-4758-206@http.llvm.org/bugs/>
Message-ID: <200908240042.n7O0gkRF009034@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4758


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-23 19:42:46 ---
cmake makefiles work fine on linux. We need more information to reproduce or
fix this problem; you will have to diagnose why the cmake check for this
feature is failing on Solaris.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 20:44:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 20:44:06 -0500
Subject: [LLVMbugs] [Bug 4758] cmake out of tree build.
In-Reply-To: <bug-4758-206@http.llvm.org/bugs/>
Message-ID: <200908240144.n7O1i6r2016683@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4758


Edward O'Callaghan <eocallaghan at auroraux.org> changed:

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




--- Comment #2 from Edward O'Callaghan <eocallaghan at auroraux.org>  2009-08-23 20:44:06 ---
How can you close it, if it does not work under Solaris and you did not test
under Solaris !?

Re-opening.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 22:53:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 22:53:01 -0500
Subject: [LLVMbugs] [Bug 4760] New: postfix doesn't get priority to prefix
 and unary maths
Message-ID: <bug-4760-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4760

           Summary: postfix doesn't get priority to prefix and unary maths
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zach at drayer.name
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3366)
 --> (http://llvm.org/bugs/attachment.cgi?id=3366)
Basic testcase that shows the bug

The C standard says that postfix should take place before unary plus or prefix
in the order of operations, however this doesn't happen with Clang.

In the attached test, Clang improperly outputs "21", while compiling with GCC
gives the proper "22".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 23:34:40 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 23:34:40 -0500
Subject: [LLVMbugs] [Bug 4762] New: double free in Kaleidoscope example
Message-ID: <bug-4762-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4762

           Summary: double free in Kaleidoscope example
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


examples/Kaleidoscope/toy.cpp creates a Module and does

  TheExecutionEngine = EngineBuilder(TheModule).create();

and also

    ExistingModuleProvider OurModuleProvider(TheModule);
    FunctionPassManager OurFPM(&OurModuleProvider);

Both of these APIs claim in their comments to take ownership of
the Module, so it appears both are attempting to delete the
same Module when the program is finished.

In an asserts build, this is diagnosed by AssertingVH with
this message:

While deleting: double (double, double)* %binary:
An asserting value handle still pointed to this value!
UNREACHABLE executed at Value.cpp:492!
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  Sun Aug 23 23:49:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 23:49:37 -0500
Subject: [LLVMbugs] [Bug 746] Make Function not annotatable and get rid of
	Annotations
In-Reply-To: <bug-746-206@http.llvm.org/bugs/>
Message-ID: <200908240449.n7O4nbuD007426@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=746


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #8 from Chris Lattner <clattner at apple.com>  2009-08-23 23:49:36 ---
This is done!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 23 23:56:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Aug 2009 23:56:33 -0500
Subject: [LLVMbugs] [Bug 4764] New: rewrite sys::Path
Message-ID: <bug-4764-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4764

           Summary: rewrite sys::Path
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-cleanup
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


sys::Path has some useful functionality but its current implementation is bad
in a number of ways.  As with many of the "sys" classes, it conflates "system
abstraction" with "helpful utility class".  I think that sys::Path should be
distilled down to a collection of a bunch of simple static functions (e.g.
getLast) that take a StringRef and return a StringRef.

Right now sys::Path contains an std::string, which forces all clients to deal
with std::strings really poor performance.  Many clients are using sys::Path
just to remove an extension from a filename: this API forces a bunch of nasty
memory traffic.

The API should also be thinned down.  Path.h shouldn't pull in <set> and
ideally not <vector> or <string> either.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 00:05:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 00:05:11 -0500
Subject: [LLVMbugs] [Bug 760] Passes should use
 getAnalysisToUpdate<TargetData>(), not getAnalysis<TargetData>()
In-Reply-To: <bug-760-206@http.llvm.org/bugs/>
Message-ID: <200908240505.n7O55Bu3009474@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=760


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #6 from Dan Gohman <gohman at apple.com>  2009-08-24 00:05:11 ---
This is now done. PR4542 is tracking the issue of being able
to make use of this from opt, etc.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 00:29:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 00:29:53 -0500
Subject: [LLVMbugs] [Bug 4760] postfix doesn't get priority to prefix and
	unary maths
In-Reply-To: <bug-4760-206@http.llvm.org/bugs/>
Message-ID: <200908240529.n7O5TrX4012561@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4760


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2009-08-24 00:29:52 ---
See http://c-faq.com/expr/seqpoints.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 Aug 24 00:43:00 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 00:43:00 -0500
Subject: [LLVMbugs] [Bug 4762] [2.6] double free in Kaleidoscope example
In-Reply-To: <bug-4762-206@http.llvm.org/bugs/>
Message-ID: <200908240543.n7O5h0DQ014170@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4762


reid.kleckner at gmail.com <reid.kleckner at gmail.com> changed:

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




--- Comment #2 from reid.kleckner at gmail.com  2009-08-24 00:43:00 ---
I think the bug has been present since the original.  It's creating two
ExistingModuleProviders, both of which delete the module in their dtors.  That
said, it's easy to fix, just pass the ModuleProvider to the EngineBuilder.  I
didn't find any other instances of this bug, and I committed the fix as r79896.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 01:43:53 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 01:43:53 -0500
Subject: [LLVMbugs] [Bug 2162] Migrate away from Value::getName()
In-Reply-To: <bug-2162-206@http.llvm.org/bugs/>
Message-ID: <200908240643.n7O6hrYo022158@zion.cs.uiuc.edu>

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


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #4 from Daniel Dunbar <daniel at zuster.org>  2009-08-24 01:43:53 ---
Totally! The magic behavior to propagate null names isn't used, but this is
only a nice-to-have, for a rainy day.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 03:31:37 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 03:31:37 -0500
Subject: [LLVMbugs] [Bug 4764] rewrite sys::Path
In-Reply-To: <bug-4764-206@http.llvm.org/bugs/>
Message-ID: <200908240831.n7O8Vb2E015988@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4764


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-24 03:31:37 ---


*** This bug has been marked as a duplicate of bug 3837 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 08:33:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 08:33:09 -0500
Subject: [LLVMbugs] [Bug 4762] [2.6] AssertingVH abort in Kaleidoscope
	example
In-Reply-To: <bug-4762-206@http.llvm.org/bugs/>
Message-ID: <200908241333.n7ODX9tL017995@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4762


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |
            Summary|[2.6] double free in        |[2.6] AssertingVH abort in
                   |Kaleidoscope example        |Kaleidoscope example




--- Comment #4 from Dan Gohman <gohman at apple.com>  2009-08-24 08:33:08 ---
Kaleidoscope still aborts with an AssertingVH abort. At a brief glance,
it now appears to be due to the ExecutionEngine holding several
AssertingVHs when the Module is deleted.

Attached is a testcase. Run it as "Kaleidoscope < test.kaleidoscope".
It should print out a few fibonacci numbers and a few mandelbrot
renderings and then exit. Currently it aborts with an AssertingVH
abort before exiting.

Also, the tutorial pages themselves have a copy of the toy.cpp code,
so they still have any bugs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 11:08:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 11:08:06 -0500
Subject: [LLVMbugs] [Bug 4754] Invalid code is silently generated for
	varargs function
In-Reply-To: <bug-4754-206@http.llvm.org/bugs/>
Message-ID: <200908241608.n7OG86i8005179@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4754


Dmitry Gorbachev <d.g.gorbachev at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |
            Summary|Internal compiler error.    |Invalid code is silently
                   |Varargs function with       |generated for varargs
                   |fastcall calling convention.|function




--- Comment #5 from Dmitry Gorbachev <d.g.gorbachev at gmail.com>  2009-08-24 11:08:06 ---
Disregard everything before comment #4...


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 11:08:18 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 11:08:18 -0500
Subject: [LLVMbugs] [Bug 4765] New: Dllexport function declaration,
	llvm-gcc aborted
Message-ID: <bug-4765-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4765

           Summary: Dllexport function declaration, llvm-gcc aborted
           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: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3369)
 --> (http://llvm.org/bugs/attachment.cgi?id=3369)
C file

"Global is external, but doesn't have external or dllimport or weak linkage!"


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 16:00:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 16:00:57 -0500
Subject: [LLVMbugs] [Bug 4767] New: Cannot print this instruction.
Message-ID: <bug-4767-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4767

           Summary: Cannot print this instruction.
           Product: libraries
           Version: 2.6
          Platform: PC
               URL: http://svn.freebsd.org/base/head/lib/msun/src/s_fdim.c
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3696


Compiling libm on FreeBSD end with the following assert. What is interesting
about this bug is that if I run s_fdim.c with clang -E and then compile the
pre-processed source it does not crash. So because of this I'm not sure how to
generate a useful testcase.

The source can be found here:
http://svn.freebsd.org/base/head/lib/msun/src/s_fdim.c

$ clang -O0 -c s_fdim.c
Assertion failed: (Bits != 0 && "Cannot print this instruction."), function
printInstruction, file
/usr/src-local/llvm/lib/Target/X86/AsmPrinter/../X86GenAsmWriter.inc, line
2230.
Stack dump:
0.      Program arguments: /usr/opt/llvm/bin/../libexec/clang-cc -triple
x86_64-unknown-freebsd8.0 -S -disable-free -main-file-name s_fdim.c
--relocation-model static --disable-fp-elim --unwind-tables=1 --mcpu=x86-64
--fmath-errno=1 -O0 -fdiagnostics-show-option -o /tmp/cc-Gbl72b.s -x c s_fdim.c
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 AT&T-Style Assembly Printer' on function '@fdiml'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

ps. This happens with 2.6 too and would be nice if 2.6 could build the FreeBSD
world.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 19:18:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 19:18:30 -0500
Subject: [LLVMbugs] [Bug 4751] [inline asm] `%a' template: LLVM ERROR:
 Invalid operand found in inline asm
In-Reply-To: <bug-4751-206@http.llvm.org/bugs/>
Message-ID: <200908250018.n7P0IU1e002453@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4751


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>  2009-08-24 19:18:29 ---
Fixed here.
http://llvm.org/viewvc/llvm-project?rev=79961&view=rev
I have trouble testing these because there are bugs in Darwin's gcc for this
construct, so I'm not entirely sure what it's supposed to be doing (apparently
no one here uses it).  If you report any more please include the Linux code
that you're supposed to be getting.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 24 22:43:45 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 22:43:45 -0500
Subject: [LLVMbugs] [Bug 4769] New: msp430 backend: Cannot yet select
Message-ID: <bug-4769-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4769

           Summary: msp430 backend: Cannot yet select
           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


Seen on Ubuntu Hardy.  No idea if anyone cares to fix this but since I have a
nice reduced test input I figured I'd submit it...

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp2$ clang
-ccc-host-triple msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O3
small.c
fatal error: error in backend: Cannot yet select: 0x9807d08: i16,ch =
      dynamic_stackalloc 0x98074e0, 0x9807ea0, 0x9807ae8

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp2$ clang -v

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 79920)
Target: i386-pc-linux-gnu
Thread model: posix

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp2$ cat small.c

int foo (void)
{
  volatile int result = 0;
  return result;
}

int bar (void);

int main (void)
{
  while (bar ());
  foo ();
  return 0;
}


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 24 22:45:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Aug 2009 22:45:03 -0500
Subject: [LLVMbugs] [Bug 4770] New: msp430 backend: Assertion
 `MO.getOffset() == 0 && " No offsets allowed!"' failed.
Message-ID: <bug-4770-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4770

           Summary: msp430 backend: Assertion `MO.getOffset() == 0 && "No
                    offsets allowed!"' 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: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen on Ubuntu Hardy.

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp1$ clang
-ccc-host-triple msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O2
small.c
clang-cc: MSP430AsmPrinter.cpp:170:
void<unnamed>::MSP430AsmPrinter::printOperand(const llvm::MachineInstr*, int,
const char*): Assertion `MO.getOffset() == 0 && "No offsets allowed!"' failed.
0   clang-cc 0x0933f738
Stack dump:
0.      Program arguments:
/home/regehr/z/tmp/llvm-gcc-r79920-install/bin/../libexec/clang-cc -triple
msp430-generic-generic -S -disable-free -main-file-name small.c
--relocation-model static --disable-fp-elim --unwind-tables=0 --fmath-errno=1
-O2 -w -fdiagnostics-show-option -o small.s -x c small.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'MSP430 Assembly Printer' on function '@bux'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp1$ clang -v

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 79920)
Target: i386-pc-linux-gnu
Thread model: posix

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp1$ cat small.c

struct s
{
  int t0;
};

struct s arr[3];

void bar (char);

void bux (int dt)
{
  bar (1);
}

void baz (int num, int t0)
{
  struct s *timer = &arr[num];
  timer->t0 = t0;
}

void bar (char num)
{
  baz (num, 0);
}


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 25 02:50:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 02:50:35 -0500
Subject: [LLVMbugs] [Bug 4226] ARM does not lower all vshifts
In-Reply-To: <bug-4226-206@http.llvm.org/bugs/>
Message-ID: <200908250750.n7P7oZCH001261@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4226


Xerxes R??nby <xerxes at zafena.se> changed:

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




--- Comment #14 from Xerxes R??nby <xerxes at zafena.se>  2009-08-25 02:50:32 ---
(In reply to comment #13)
> (In reply to comment #12)
> > Also, it's got to be said that TLI.getOperationAction
> > is overloaded: it's used for results and operands and
> > that can cause confusion like here.
> 
> Mmm... have any suggestions for improvements?  It'll be significantly easier to
> make non-trivial changes now that LegalizeDAG is mostly cleaned up.
> 

After your changes in llvm rev 79719 the ARM target now passes the CodeGen
test!
http://llvm.org/viewvc/llvm-project?view=rev&revision=79719

xerxes at sheeva:/usr/src/build/buildbot/llvm-arm-linux/build$ ./Debug/bin/llvm-as
< test/CodeGen/ARM/vshift_split.ll | ./Debug/bin/llc -march=arm
        .file   "<stdin>"
        .eabi_attribute 20, 1
        .eabi_attribute 21, 1
        .eabi_attribute 23, 3
        .eabi_attribute 24, 1
        .eabi_attribute 25, 1


        .text
        .globl  update
        .align  2
update:
        mov r0, r0, lsr #2
        orr r0, r0, r1, lsl #30
        mov r2, r2, lsr #2
        orr r2, r2, r3, lsl #30
        mov r1, r1, lsr #2
        mov r3, r3, lsr #2
        bx lr
        .size update, .-update


Thanks!
Bug closed


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 03:13:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 03:13:36 -0500
Subject: [LLVMbugs] [Bug 4772] New: ARM unittest JIT. GlobalInFunction fail
 during ARM Machine Code Emitter
Message-ID: <bug-4772-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4772

           Summary: ARM unittest JIT.GlobalInFunction fail during ARM
                    Machine Code Emitter
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


(gdb) run
Starting program:
/usr/src/build/buildbot/llvm-arm-linux/build/unittests/ExecutionEngine/JIT/Debug/JITTests 
[Thread debugging using libthread_db enabled]
[==========] Running 10 tests from 3 test cases.
[----------] Global test environment set-up.
[----------] 3 tests from JITEventListenerTest
[ RUN      ] JITEventListenerTest.Simple
[       OK ] JITEventListenerTest.Simple
[ RUN      ] JITEventListenerTest.MultipleListenersDontInterfere
[       OK ] JITEventListenerTest.MultipleListenersDontInterfere
[ RUN      ] JITEventListenerTest.MatchesMachineCodeInfo
[       OK ] JITEventListenerTest.MatchesMachineCodeInfo
[----------] 6 tests from JITMemoryManagerTest
[ RUN      ] JITMemoryManagerTest.NoAllocations
[       OK ] JITMemoryManagerTest.NoAllocations
[ RUN      ] JITMemoryManagerTest.TestCodeAllocation
[       OK ] JITMemoryManagerTest.TestCodeAllocation
[ RUN      ] JITMemoryManagerTest.TestSmallGlobalInts
[       OK ] JITMemoryManagerTest.TestSmallGlobalInts
[ RUN      ] JITMemoryManagerTest.TestLargeGlobalArray
[       OK ] JITMemoryManagerTest.TestLargeGlobalArray
[ RUN      ] JITMemoryManagerTest.TestManyGlobals
[       OK ] JITMemoryManagerTest.TestManyGlobals
[ RUN      ] JITMemoryManagerTest.TestManyStubs
[       OK ] JITMemoryManagerTest.TestManyStubs
[----------] 1 test from JIT
[ RUN      ] JIT.GlobalInFunction
[New Thread 0x4001f590 (LWP 17670)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x4001f590 (LWP 17670)]
0x00401850 in std::vector<llvm::MachineMove, std::allocator<llvm::MachineMove>
>::clear (this=0x5a5a5a72) at /usr/include/c++/4.3/bits/stl_vector.h:892
892           { _M_erase_at_end(this->_M_impl._M_start); }
(gdb) bt
#0  0x00401850 in std::vector<llvm::MachineMove,
std::allocator<llvm::MachineMove> >::clear (this=0x5a5a5a72) at
/usr/include/c++/4.3/bits/stl_vector.h:892
#1  0x003ffc70 in llvm::MachineModuleInfo::EndFunction (this=0x5a5a5a5a) at
MachineModuleInfo.cpp:76
#2  0x00376780 in finishFunction (this=0x9528c0, F=@0x938210) at
JITEmitter.cpp:1151
#3  0x00066e70 in runOnMachineFunction (this=0x950100, MF=@0x938210) at
ARMCodeEmitter.cpp:206
#4  0x003e8ed0 in llvm::MachineFunctionPass::runOnFunction (this=0x950100,
F=@0x9508d0) at MachineFunctionPass.cpp:31
#5  0x006f227c in llvm::FPPassManager::runOnFunction (this=0x9527e8,
F=@0x9508d0) at PassManager.cpp:1372
#6  0x006f30a0 in llvm::FunctionPassManagerImpl::run (this=0x945748,
F=@0x9508d0) at PassManager.cpp:1324
#7  0x006f3340 in llvm::FunctionPassManager::run (this=0x9512e0, F=@0x9508d0)
at PassManager.cpp:1254
#8  0x00368718 in llvm::JIT::runJITOnFunctionUnlocked (this=0x951250,
F=0x9508d0, locked=@0xbec79304) at JIT.cpp:598
#9  0x00368cb4 in llvm::JIT::getPointerToFunction (this=0x951250, F=0x9508d0)
at JIT.cpp:670
#10 0x0002bb08 in TestBody (this=0x9342b8) at JITTest.cpp:88
#11 0x007a5fa0 in testing::Test::Run (this=0x9342b8) at gtest.cc:2035
#12 0x007ac130 in testing::internal::TestInfoImpl::Run (this=0x931ae0) at
gtest.cc:2257
#13 0x007b4dd0 in testing::internal::TestInfoImpl::RunTest (test_info=0x931ad0)
    at
/usr/src/build/buildbot/llvm-arm-linux/build/utils/unittest/googletest/include/gtest/internal/gtest-internal-inl.h:611
#14 0x007ad878 in testing::internal::List<testing::TestInfo*>::ForEach<void
(*)(testing::TestInfo*)> (this=0x931bd0, 
    functor=0x7b4da8
<testing::internal::TestInfoImpl::RunTest(testing::TestInfo*)>)
    at
/usr/src/build/buildbot/llvm-arm-linux/build/utils/unittest/googletest/include/gtest/internal/gtest-internal-inl.h:365
#15 0x007abf9c in testing::TestCase::Run (this=0x931ba0) at gtest.cc:2353
#16 0x007b4da0 in testing::TestCase::RunTestCase (test_case=0x931ba0)
    at
/usr/src/build/buildbot/llvm-arm-linux/build/utils/unittest/googletest/include/gtest/internal/gtest-internal-inl.h:730
#17 0x007ad480 in testing::internal::List<testing::TestCase*>::ForEach<void
(*)(testing::TestCase*)> (this=0x931118, 
    functor=0x7b4d84 <testing::TestCase::RunTestCase(testing::TestCase*)>)
    at
/usr/src/build/buildbot/llvm-arm-linux/build/utils/unittest/googletest/include/gtest/internal/gtest-internal-inl.h:365
#18 0x007abd20 in testing::internal::UnitTestImpl::RunAllTests (this=0x9310d0)
at gtest.cc:3550
#19 0x007abea4 in testing::UnitTest::Run (this=0x9300cc) at gtest.cc:3299
#20 0x007bbe34 in main (argc=1, argv=0xbec797c4) at TestMain.cpp:14
(gdb) 


For some reason the FrameMoves object pointer got the funny inaccessible memory
address 0x5a5a5a72  at frame 1 and makes the ExecutionEngine crash during
clear.

/// EndFunction - Discard function meta information.
///
void MachineModuleInfo::EndFunction() {
  // Clean up frame info.
  FrameMoves.clear();


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 05:16:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 05:16:52 -0500
Subject: [LLVMbugs] [Bug 4772] ARM unittest JIT. GlobalInFunction fail
 during ARM Machine Code Emitter
In-Reply-To: <bug-4772-206@http.llvm.org/bugs/>
Message-ID: <200908251016.n7PAGq77025674@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4772


Xerxes R??nby <xerxes at zafena.se> changed:

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




--- Comment #3 from Xerxes R??nby <xerxes at zafena.se>  2009-08-25 05:16:51 ---
Fixed in llvm rev 79982
http://llvm.org/viewvc/llvm-project?view=rev&revision=79982


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 10:34:32 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 10:34:32 -0500
Subject: [LLVMbugs] [Bug 4773] New: simple float[3] vector usage leads to
	llc failure
Message-ID: <bug-4773-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4773

           Summary: simple float[3] vector usage leads to llc failure
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: brian.e.paul at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following code (vec3.c):

typedef __attribute__(( ext_vector_type(3) )) float vec3;

vec3 sum(vec3 a, vec3 b)
{
   return a + b;
}


Causes clang/llc to fail when compiled with:

$ clang -S vec3.c -o vec3.s
Return operand #2 has unhandled type f32UNREACHABLE executed at
CallingConvLower.cpp:93!
0   clang-cc  0x0913070c
1   clang-cc  0x09130cd8
2             0x00110400 __kernel_sigreturn + 0
3   libc.so.6 0x0036ae18 abort + 392
4   clang-cc  0x091111e4
5   clang-cc  0x08c1ff43
6   clang-cc  0x08be174d
7   clang-cc  0x08ca7e17
8   clang-cc  0x08ca7f53
9   clang-cc  0x08ca8456
10  clang-cc  0x08cbe3ca
11  clang-cc  0x08cbeed4
12  clang-cc  0x08cbfd56
13  clang-cc  0x08da8b9e
14  clang-cc  0x090a5d61
15  clang-cc  0x090a6864
16  clang-cc  0x090a6a31
17  clang-cc  0x080955ec
18  clang-cc  0x080956ea
19  clang-cc  0x082fcebe
20  clang-cc  0x080717a1
21  clang-cc  0x080734d4 main + 3320
22  libc.so.6 0x003546d5 __libc_start_main + 229
23  clang-cc  0x08069701
Stack dump:
0.      Program arguments: /home/local/bin/../libexec/clang-cc -triple
i386-pc-linux-gnu -S -disable-free -main-file-name vec3.c --relocation-model
static --disable-fp-elim --unwind-tables=0 --mcpu=pentium4 --fmath-errno=1
-fdiagnostics-show-option -o vec3.s -x c vec3.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@vec3_sum'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)


I'm using llvm/clang from SVN on Aug 26, 2009, debug build on x86-32.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 11:12:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 11:12:30 -0500
Subject: [LLVMbugs] [Bug 4463] ARM AAPCS-VFP ABI does not support v2f64
	arguments
In-Reply-To: <bug-4463-206@http.llvm.org/bugs/>
Message-ID: <200908251612.n7PGCUqD006302@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4463


Bob Wilson <bob.wilson at apple.com> changed:

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




--- Comment #3 from Bob Wilson <bob.wilson at apple.com>  2009-08-25 11:12:30 ---
I never had a specific testcase and Sandeep and Anton took care of this
already.  I'll close this now.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 25 11:34:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 11:34:33 -0500
Subject: [LLVMbugs] [Bug 4773] simple float[3] vector usage leads to llc
	failure
In-Reply-To: <bug-4773-206@http.llvm.org/bugs/>
Message-ID: <200908251634.n7PGYXVM009041@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4773


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-08-25 11:34:33 ---
Unfortunately, this is a current limitation of the code generator.  It would be
nice for codegen to promote a <3 x float> to <4 x float> for return purposes
though.

*** This bug has been marked as a duplicate of bug 2660 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 11:49:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 11:49:39 -0500
Subject: [LLVMbugs] [Bug 4774] New: delete the gccas/gccld tools
Message-ID: <bug-4774-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4774

           Summary: delete the gccas/gccld tools
           Product: tools
           Version: 2.2
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: gccas
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


the gccas/gccld tools have outlived their usefulness, being replaced with opt
-std-compile-opts and llvm-ld respectively, they should be gracefully 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 Aug 25 12:00:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:00:54 -0500
Subject: [LLVMbugs] [Bug 4769] msp430 backend: Cannot yet select
In-Reply-To: <bug-4769-206@http.llvm.org/bugs/>
Message-ID: <200908251700.n7PH0suZ012467@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4769


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

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




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-25 12:00:54 ---
Implemented in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085522.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 Aug 25 12:16:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:16:50 -0500
Subject: [LLVMbugs] [Bug 4775] New: -indvars pass turning static alloca into
 dynamic alloca
Message-ID: <bug-4775-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4775

           Summary: -indvars pass turning static alloca into dynamic alloca
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-quality, regression
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: gohman at apple.com, llvmbugs at cs.uiuc.edu


This seems pretty bad: opt -indvars turns this:

; ModuleID = '<stdin>'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i386-apple-darwin10.0"

@llvm.used = appending global [1 x i8*] [i8* bitcast (i32 ()* @main to i8*)],
section "llvm.metadata" ; <[1 x i8*]*> [#uses=0]

define i32 @main() nounwind ssp {
entry:
  %result.i = alloca i32, align 4                 ; <i32*> [#uses=2]
  br label %while.cond

while.cond:                                       ; preds = %while.cond, %entry
  %call = call i32 @bar() nounwind ssp            ; <i32> [#uses=1]
  %tobool = icmp eq i32 %call, 0                  ; <i1> [#uses=1]
  br i1 %tobool, label %while.end, label %while.cond

while.end:                                        ; preds = %while.cond
  volatile store i32 0, i32* %result.i
  %tmp.i = volatile load i32* %result.i           ; <i32> [#uses=0]
  ret i32 0
}

declare i32 @bar() ssp

into:

...
  br i1 %tobool, label %while.end, label %while.cond

while.end:                                        ; preds = %while.cond
  %result.i = alloca i32, align 4                 ; <i32*> [#uses=2]
  volatile store i32 0, i32* %result.i
  %tmp.i = volatile load i32* %result.i           ; <i32> [#uses=0]
  ret i32 0

Dynamic allocas are *much* more expensive than static ones.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 12:27:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:27:54 -0500
Subject: [LLVMbugs] [Bug 4776] New: msp430 backend: placing a variable at a
 specific address doesn't seem to work
Message-ID: <bug-4776-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4776

           Summary: msp430 backend: placing a variable at a specific address
                    doesn't seem to work
           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: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


Consider this code:

volatile unsigned char P1OUT __asm ("0x0021");

unsigned char foo (unsigned char x)
{
  unsigned char tmp = P1OUT;
  P1OUT = x;
  return tmp;
}

Using mspgcc...

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ msp430-gcc -O p1.c -S
-o -
foo:
        mov.b   &0x0021, r14
        mov.b   r15, &0x0021
        mov.b   r14, r15
        ret

However using clang...

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ clang -ccc-host-triple
msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O2 p1.c -o -
foo:
        push.w  r4
        mov.w   r1, r4
        mov.b   r15, r12
        mov.b   &_30_x0021, r15
        mov.b   r12, &_30_x0021
        pop.w   r4
        ret

Perhaps this is just a typo in the emitted code?  It looks almost correct.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 12:32:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:32:39 -0500
Subject: [LLVMbugs] [Bug 4777] New: msp430 backend: static function-scope
 variable at specific address doesn't work
Message-ID: <bug-4777-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4777

           Summary: msp430 backend: static function-scope variable at
                    specific address doesn't work
           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


regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ cat p1.c

unsigned char foo (unsigned char x)
{
  static volatile unsigned char P1OUT __asm ("0x0021");
  unsigned char tmp = P1OUT;
  P1OUT = x;
  return tmp;
}

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ clang -ccc-host-triple
msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O2 p1.c

p1.c:3:26: error: cannot compile this __asm__ yet
  volatile unsigned char P1OUT __asm ("0x0021");
                         ^
1 diagnostic generated.



The emitted code should be identical to placing the register in file scope:

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ msp430-gcc -O p1.c -S
-o -
foo:
        mov.b   &0x0021, r14
        mov.b   r15, &0x0021
        mov.b   r14, r15
        ret


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

From bugzilla-daemon at cs.uiuc.edu  Tue Aug 25 12:43:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:43:44 -0500
Subject: [LLVMbugs] [Bug 4775] [2.6] -indvars pass turning static alloca
 into dynamic alloca
In-Reply-To: <bug-4775-206@http.llvm.org/bugs/>
Message-ID: <200908251743.n7PHhi6e018362@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4775


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-25 12:43:43 ---
Fixed on trunk here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085532.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 Aug 25 12:43:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:43:44 -0500
Subject: [LLVMbugs] [Bug 4778] New: msp430 specific: problem with constraint
 'r' in inline asm
Message-ID: <bug-4778-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4778

           Summary: msp430 specific: problem with constraint 'r' in inline
                    asm
           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: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


For this code:

char __nesc_atomic_start(void )
{
  char result = (({
        unsigned int __x;
        __asm volatile ("mov    r2, %0" : "=r"((unsigned int)__x));__x;
      }
      )
    & 0x0008) != 0;
  __asm volatile ("dint");
  __asm volatile ("nop");
  __asm volatile ("" :  :  : "memory");
  return result;
}

We want something like this:

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ msp430-gcc small.c -Os
-S -o -
__nesc_atomic_start:
        mov     r2, r15
        clrc
        rrc     r15
        rra     r15
        rra     r15
        and.b   #llo(1), r15
        dint
        nop
        sxt     r15
        ret

But we get:

regehr at john-home:~/z/tinyos-2.x/apps/Null/build/telosb$ clang -ccc-host-triple
msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O2 small.c
-fheinous-gnu-extensions
fatal error: error in backend: Couldn't allocate output reg for constraint 'r'!


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 12:52:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 12:52:06 -0500
Subject: [LLVMbugs] [Bug 4779] New: msp430 backend ICE: "#operands for dag
 node doesn't match . td file!"
Message-ID: <bug-4779-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4779

           Summary: msp430 backend ICE: "#operands for dag node doesn't
                    match .td file!"
           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


regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp3$ clang
-ccc-host-triple msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O3
small.c
clang-cc: ScheduleDAGSDNodesEmit.cpp:506: void
llvm::ScheduleDAGSDNodes::EmitNode(llvm::SDNode*, bool, bool,
llvm::DenseMap<llvm::SDValue, unsigned int, llvm::DenseMapInfo<llvm::SDValue>,
llvm::DenseMapInfo<unsigned int> >&): Assertion `(II.getNumOperands() ==
NumMIOperands || HasPhysRegOuts || II.isVariadic()) && "#operands for dag node
doesn't match .td file!"' failed.
0   clang-cc 0x09340218
Stack dump:
0.      Program arguments:
/home/regehr/z/tmp/llvm-gcc-r80002-install/bin/../libexec/clang-cc -triple
msp430-generic-generic -S -disable-free -main-file-name small.c
--relocation-model static --disable-fp-elim --unwind-tables=0 --fmath-errno=1
-O3 -w -fdiagnostics-show-option -o small.s -x c small.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'MSP430 DAG->DAG Pattern Instruction Selection' on
function '@foo'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp3$ cat small.c

void foo (void)
{
  volatile char r;
  r |= 1;
}

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb/tmp3$ clang -v

clang version 1.0 (http://llvm.org/svn/llvm-project/cfe/trunk 80002)
Target: i386-pc-linux-gnu
Thread model: posix


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 15:23:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 15:23:43 -0500
Subject: [LLVMbugs] [Bug 4774] delete the gccas/gccld tools
In-Reply-To: <bug-4774-206@http.llvm.org/bugs/>
Message-ID: <200908252023.n7PKNh3p006666@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4774


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #3 from Daniel Dunbar <daniel at zuster.org>  2009-08-25 15:23:43 ---
Fixed in 80023.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 17:08:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 17:08:50 -0500
Subject: [LLVMbugs] [Bug 4780] New: Tutorial PNGs don't get installed
Message-ID: <bug-4780-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4780

           Summary: Tutorial PNGs don't get installed
           Product: Documentation
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: markleone at gmail.com
                CC: llvmbugs at cs.uiuc.edu


"make install" doesn't seem to install the PNGs for the tutorial docs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 19:05:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 19:05:27 -0500
Subject: [LLVMbugs] [Bug 4783] New: Circular dependency in ICC build
Message-ID: <bug-4783-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4783

           Summary: Circular dependency in ICC build
           Product: Build scripts
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: markleone at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I'm having trouble building llvm-trunk for OS X with icc (11.1 20090624).

I used "configure CC=icc CXX=icpc --disable-optimized --enable-threads=no"
(I've had trouble in the past with CXX=icc, oddly, and also had trouble with
pthreads vs. libthr.)

ICC insists on  a virtual destructor to basic_parser in Support/CommandLine.h.

The last straw seems to be a circular dependency:

find-cycles.pl: Circular dependency between *.a files:
find-cycles.pl:   libLLVMAsmPrinter.a libLLVMCodeGen.a libLLVMScalarOpts.a
libLLVMSelectionDAG.a

A workaround would be appreciated if there's no easy fix.  (Do circular
dependencies even matter on OSX?)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 21:26:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 21:26:54 -0500
Subject: [LLVMbugs] [Bug 4785] New: msp430 backend: interrupt handing
	attributes ignored
Message-ID: <bug-4785-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4785

           Summary: msp430 backend: interrupt handing attributes ignored
           Product: new-bugs
           Version: trunk
          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: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


Given this sort of code:

extern void foo (void);

__attribute((wakeup)) __attribute((interrupt(10))) void sig_TIMERA1_VECTOR
(void)
{
  foo();
}

int main (void)
{
  while (1);
}

LLVM gives us this:

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb$ clang -ccc-host-triple
msp430-generic-generic -ccc-clang-archs msp430 -x c -S -w -O irq.c -o -
        .file   "irq.c"


        .text
        .align  4
        .globl  sig_TIMERA1_VECTOR
        .type   sig_TIMERA1_VECTOR, at function
sig_TIMERA1_VECTOR:
        push.w  r4
        mov.w   r1, r4
        call    #foo
        pop.w   r4
        ret
        .size   sig_TIMERA1_VECTOR, .-sig_TIMERA1_VECTOR


        .align  4
        .globl  main
        .type   main, at function
main:
        push.w  r4
        mov.w   r1, r4
.BB2_1:
        jmp     .BB2_1                                      # %while.body
        .size   main, .-main

But it should be more like this:

regehr at john-home:~/z/tinyos-2.x/apps/Blink/build/telosb$ msp430-gcc -O irq.c -S
-o -
vector_ffea:
sig_TIMERA1_VECTOR:
        push    r15
        push    r14
        push    r13
        push    r12
        call    #foo
        pop     r12
        pop     r13
        pop     r14
        pop     r15
        bic     #0xf0,0(r1)
        reti
main:
        mov     #(__stack-0), r1
.L3:
        jmp     .L3
        br      #__stop_progExec__


A quick and easy solution would be to generate canned IRQ stub code that pushes
the caller-saved part of the context, then use the standard calling convention
to get into the user-supplied interrupt handler.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 25 23:27:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Aug 2009 23:27:33 -0500
Subject: [LLVMbugs] [Bug 4270] getelementptr is underspecified
In-Reply-To: <bug-4270-206@http.llvm.org/bugs/>
Message-ID: <200908260427.n7Q4RXrA004510@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4270


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #6 from Dan Gohman <gohman at apple.com>  2009-08-25 23:27:33 ---
Thanks Chris. I've now formalized a set of rules for getelementptr
in LangRef.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 Aug 26 08:50:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 08:50:48 -0500
Subject: [LLVMbugs] [Bug 4778] msp430 specific: problem with constraint 'r'
 in inline asm
In-Reply-To: <bug-4778-206@http.llvm.org/bugs/>
Message-ID: <200908261350.n7QDomhi023529@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4778


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

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




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2009-08-26 08:50:46 ---
Implemented in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085603.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 Aug 26 09:33:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 09:33:59 -0500
Subject: [LLVMbugs] [Bug 4787] New: ICE: Assertion
 `(unsigned)FI-LowSpillSlot < SpillSlotToUsesMap.size() && "Invalid spill
 slot"' failed.
Message-ID: <bug-4787-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4787

           Summary: ICE: Assertion `(unsigned)FI-LowSpillSlot <
                    SpillSlotToUsesMap.size() && "Invalid spill slot"'
                    failed.
           Product: new-bugs
           Version: trunk
          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


Seen using r80083 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp187$ clang -w -O small.c
clang-cc: VirtRegMap.cpp:181: void llvm::VirtRegMap::addSpillSlotUse(int,
llvm::MachineInstr*): Assertion `(unsigned)FI-LowSpillSlot <
SpillSlotToUsesMap.size() && "Invalid spill slot"' failed.
0   clang-cc 0x09334248
Stack dump:
0.      Program arguments:
/home/regehr/z/tmp/llvm-gcc-r80083-install/bin/../libexec/clang-cc -triple
i386-pc-linux-gnu -S -disable-free -main-file-name small.c --relocation-model
static --disable-fp-elim --unwind-tables=0 --mcpu=pentium4 --fmath-errno=1 -O1
-w -fdiagnostics-show-option -o /tmp/cc-UAgyYy.s -x c small.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Linear Scan Register Allocator' on function '@uint86'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/volatile/tmp187$ cat small.c

char foo (int si1, char si2)
{
  return si2 ? : 1;
}

char bar (char ui1, char ui2)
{
  return ui2 ? : 1;
}

static char baz (char left, int right)
{
  return right ? : left;
}

char bux (int ui1, char ui2)
{
  return ui2 ? : 1;
}

static char buzz (int ui1, int uint8ui2)
{
  return ui1;
}

char qixx (int uint32ui2)
{
}

long long int zizz (long long int ui1, char ui2)
{
  return ui2;
}

static char aaa (long long int left, int right)
{
  return right ? : left;
}

long long int g4;
long long int g16;
char g27;

void uint86 (char p28)
{
  char l58;
  func (safe
        (zizz (bar (g16, 1), qixx) != g4 == aaa (g27, g16)
         && safe (g4, 1)), safe (1), safe (foo (safe (~l58),
                                                 bux (g4, g4)),
                                            safe (baz
                                                  (g16 < buzz (p28, 1), 1),
                                                  func (p28 != g4))), !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  Wed Aug 26 10:43:52 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 10:43:52 -0500
Subject: [LLVMbugs] [Bug 4788] New: selected processor does not support `blx
	r*'
Message-ID: <bug-4788-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4788

           Summary: selected processor does not support `blx r*'
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: wjjeon at gmail.com
                CC: llvmbugs at cs.uiuc.edu


While I was trying to compile a Linux kernel for ARM processor, I got this
error. I compiled LLVM-GCC with '--with-arch=armv6' option.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 26 11:34:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 11:34:54 -0500
Subject: [LLVMbugs] [Bug 4789] New: Assertion in Thumb2 code
Message-ID: <bug-4789-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4789

           Summary: Assertion in Thumb2 code
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3379)
 --> (http://llvm.org/bugs/attachment.cgi?id=3379)
Reduced .ll

Consider the attached testcase. We're having currently:

./llc bugpoint-reduced-simplified.bc -mattr=+neon
llc: /home/asl/proj/llvm/src/lib/Target/ARM/Thumb2InstrInfo.cpp:417: int
llvm::rewriteT2FrameIndex(llvm::MachineInstr&, unsigned int, unsigned int, int,
const llvm::ARMBaseInstrInfo&): Assertion `AddrMode == ARMII::AddrMode5'
failed.

Also, the same file yields infinite loop in the codegen in ARM mode
(-march=armv7-elf).


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 26 14:11:35 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 14:11:35 -0500
Subject: [LLVMbugs] [Bug 4790] New: Missing diagnostic for address space
 qualifier in type name of compound literal within function
Message-ID: <bug-4790-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4790

           Summary: Missing diagnostic for address space qualifier in type
                    name of compound literal within function
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ken.dyck at onsemi.com
                CC: llvmbugs at cs.uiuc.edu


Clang accepts address space qualifiers on the type named by compound literals
that appear in functions. This is inconsistent with other objects with
automatic storage duration (and in violation of Embedded C ISO/IEC TR 18037,
FWIW). Other objects with automatic storage are prohibited from having address
specifiers, reporting an error when given. Compound literals in functions
should likewise be restricted and report a similar 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  Wed Aug 26 16:03:11 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 16:03:11 -0500
Subject: [LLVMbugs] [Bug 4762] [2.6] AssertingVH abort in Kaleidoscope
	example
In-Reply-To: <bug-4762-206@http.llvm.org/bugs/>
Message-ID: <200908262103.n7QL3B21013339@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4762


reid.kleckner at gmail.com <reid.kleckner at gmail.com> changed:

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




--- Comment #5 from reid.kleckner at gmail.com  2009-08-26 16:03:10 ---
The module provider needed to be allocated on the heap, not the stack, so it
could be properly deleted, as identified by Chris.  I committed the fix as
r80138.

I think the double-delete bug has been in the code since before it was added to
the repository, but when Jeff added the AssertingVH in the JIT it became a
problem.  Now that the EE owns the MP, the JIT can properly free the machine
code before it deletes the module and all associated functions.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 26 17:19:59 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 17:19:59 -0500
Subject: [LLVMbugs] [Bug 4791] New: PassMananger can't use alternate
 AliasAnalysis with GVN
Message-ID: <bug-4791-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4791

           Summary: PassMananger can't use alternate AliasAnalysis with GVN
           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: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3380)
 --> (http://llvm.org/bugs/attachment.cgi?id=3380)
a testcase in which GVN makes some alias queries

Running a custom alias analysis doesn't work with GVN, even when PR3323 isn't
apparently involved.

Running opt on the attached testcase like this:

  opt -scev-aa -gvn -debug-pass=Structure test.bc

(-scev-aa is a new alias analysis pass in svn) results in this output:

Pass Arguments:  -domtree -loops -scalar-evolution -scev-aa -memdep -gvn
-preverify -verify
Target Data Layout
Basic Alias Analysis (default AA impl)
  ModulePass Manager
    FunctionPass Manager
      Dominator Tree Construction
      Natural Loop Information
      Scalar Evolution Analysis
      ScalarEvolution-based Alias Analysis
      Memory Dependence Analysis
      Global Value Numbering
      Preliminary module verification
      Module Verifier

>From this, it would seem that alias queries in Memory Dependence Analysis
should go to ScalarEvolution-based Alias Analysis, and the problem in
PR3323 isn't present. However, in a debugger, I find that the queries are
going straight to the BasicAliasAnalysis.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 26 17:31:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 17:31:24 -0500
Subject: [LLVMbugs] [Bug 4791] PassMananger can't use alternate
	AliasAnalysis with GVN
In-Reply-To: <bug-4791-206@http.llvm.org/bugs/>
Message-ID: <200908262231.n7QMVOx3025026@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4791


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-26 17:31:24 ---
Oops. Using either -p to print the output or -count-aa to get statistics
introduces a ModulePass, which changes the pass structure and triggers
PR3323.


*** This bug has been marked as a duplicate of bug 3323 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 26 20:24:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Aug 2009 20:24:21 -0500
Subject: [LLVMbugs] [Bug 4789] Assertion in Thumb2 code
In-Reply-To: <bug-4789-206@http.llvm.org/bugs/>
Message-ID: <200908270124.n7R1OLpV014672@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4789


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

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




--- Comment #2 from Evan Cheng <evan.cheng at apple.com>  2009-08-26 20:24:21 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085679.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 Aug 27 00:42:54 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 00:42:54 -0500
Subject: [LLVMbugs] [Bug 4792] New:
 2008-07-22-bitfield-init-after-zero-len-array. c Fails on PPC
Message-ID: <bug-4792-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4792

           Summary: 2008-07-22-bitfield-init-after-zero-len-array.c Fails on
                    PPC
           Product: clang
           Version: 2.6
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: wendling at apple.com
                CC: llvmbugs at cs.uiuc.edu


This error occurred during 2.6 test. When I run this command on Mac OS X PPC G5
running Leopard:

$ clang-cc --emit-llvm -o t.ll 2008-07-22-bitfield-init-after-zero-len-array.c 

I get the following output:

$ more t.ll
; ModuleID = '2008-07-22-bitfield-init-after-zero-len-array.c'
target datalayout =
"E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64"
target triple = "powerpc-apple-darwin9"

%0 = type { [0 x float], i8, [3 x i8] }

@yv7 = global %0 { [0 x float] zeroinitializer, i8 -48, [3 x i8]
zeroinitializer }, align 4 ; <%0*> [#uses=0]

The test is checking for "i8 52".


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 27 00:54:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 00:54:41 -0500
Subject: [LLVMbugs] [Bug 4793] New: MC: Doesn't support assembly parsing on
	PPC
Message-ID: <bug-4793-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4793

           Summary: MC: Doesn't support assembly parsing on PPC
           Product: libraries
           Version: 2.6
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: wendling at apple.com
                CC: llvmbugs at cs.uiuc.edu


For lack of a better place to put these failures, here are the outputs of
running llvm-mc on the indicated tests on my Mac OS X PPC G5 box running
Leopard:

Running /Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/dg.exp ...
FAIL: /Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/data.s
Failed with exit(1) at line 1
while running: llvm-mc /Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/data.s
-filetype=obj -o - | macho-dump | FileCheck
/Volumes/SandBox/2.6/llvm-2.6/test/MC/Mach\
O/data.s
llvm-mc: error: this target does not support assembly parsing.
Traceback (most recent call last):
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 162, in
<module>
    main()
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 159, in
main
    dumpmacho(arg, opts)
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 37, in
dumpmacho
    magic = f.read(4)
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 25, in
read
    raise ValueError,"Out of data!"
ValueError: Out of data!
/Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/data.s:17:11: error: expected
string not found in input
// CHECK: ('cputype', 7)
          ^
<stdin>:1:1: note: scanning from here

^

FAIL: /Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/sections.s
Failed with exit(1) at line 1
while running: llvm-mc /Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/sections.s
-filetype=obj -o - | macho-dump | FileCheck
/Volumes/SandBox/2.6/llvm-2.6/test/MC/\
MachO/sections.s
llvm-mc: error: this target does not support assembly parsing.
Traceback (most recent call last):
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 162, in
<module>
    main()
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 159, in
main
    dumpmacho(arg, opts)
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 37, in
dumpmacho
    magic = f.read(4)
  File "/Volumes/SandBox/2.6/llvm-2.6/test/Scripts/macho-dump", line 25, in
read
    raise ValueError,"Out of data!"
ValueError: Out of data!
/Volumes/SandBox/2.6/llvm-2.6/test/MC/MachO/sections.s:46:11: error: expected
string not found in input
// CHECK: ('cputype', 7)
          ^
<stdin>:1:1: note: scanning from here

^


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

From bugzilla-daemon at cs.uiuc.edu  Thu Aug 27 06:37:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 06:37:06 -0500
Subject: [LLVMbugs] [Bug 4795] New: TableGen assertion
Message-ID: <bug-4795-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4795

           Summary: TableGen assertion
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: TableGen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3381)
 --> (http://llvm.org/bugs/attachment.cgi?id=3381)
Test pattern

Consider the attached easy patch to ARMInstrNEON.td. If one apply it then
TableGen will fail the the following scary assertion:

$ /home/asl/proj/llvm/build_debug/Debug/bin/tblgen -I
/home/asl/proj/llvm/src/lib/Target/ARM -I /home/asl/proj/llvm/src/include -I
/home/asl/proj/llvm/src/include -I /home/asl/proj/llvm/src/lib/Target
-gen-dag-isel -o
/home/asl/proj/llvm/build_debug/lib/Target/ARM/Debug/ARMGenDAGISel.inc.tmp
/home/asl/proj/llvm/src/lib/Target/ARM/ARM.td
vtInt:  (vt:Other)<<P:Predicate_vtInt>>
tblgen: /home/asl/proj/llvm/src/utils/TableGen/DAGISelEmitter.cpp:1501: void
llvm::DAGISelEmitter::GenerateCodeForPattern(const llvm::PatternToMatch&,
std::vector<std::pair<unsigned int, std::basic_string<char,
std::char_traits<char>, std::allocator<char> > >,
std::allocator<std::pair<unsigned int, std::basic_string<char,
std::char_traits<char>, std::allocator<char> > > > >&,
std::set<std::basic_string<char, std::char_traits<char>, std::allocator<char>
>, std::less<std::basic_string<char, std::char_traits<char>,
std::allocator<char> > >, std::allocator<std::basic_string<char,
std::char_traits<char>, std::allocator<char> > > >&,
std::vector<std::basic_string<char, std::char_traits<char>,
std::allocator<char> >, std::allocator<std::basic_string<char,
std::char_traits<char>, std::allocator<char> > > >&,
std::vector<std::basic_string<char, std::char_traits<char>,
std::allocator<char> >, std::allocator<std::basic_string<char,
std::char_traits<char>, std::allocator<char> > > >&, bool&, unsigned int&):
Assertion `0 && "Error: could not find consistent types for something we" "
already decided was ok!"' failed.

removing either fmul or vdumplane node removes the assertion


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 27 07:16:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 07:16:50 -0500
Subject: [LLVMbugs] [Bug 4772] ARM unittest JIT. GlobalInFunction fail
 during ARM Machine Code Emitter
In-Reply-To: <bug-4772-206@http.llvm.org/bugs/>
Message-ID: <200908271216.n7RCGoac014878@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4772


Xerxes R??nby <xerxes at zafena.se> changed:

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




--- Comment #4 from Xerxes R??nby <xerxes at zafena.se>  2009-08-27 07:16:50 ---
After some testing it turned out that the JIT.GlobalInFunction test can still
crash but only in about 1 of 10 runs. So I am reopening this bug.
The new crash seems to happen when running the jitted code so the cause for the
crash are different, and of course the crash message are different.

...
[----------] 1 test from JIT
[ RUN      ] JIT.GlobalInFunction
[New Thread 0x4001f590 (LWP 15227)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x4001f590 (LWP 15227)]
0x4034f014 in ?? ()
(gdb) bt
#0  0x4034f014 in ?? ()
(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  Thu Aug 27 09:03:30 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 09:03:30 -0500
Subject: [LLVMbugs] [Bug 4797] New: null ptr dereference in
 SimpleRegisterCoalescing:: JoinIntervals
Message-ID: <bug-4797-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4797

           Summary: null ptr dereference in
                    SimpleRegisterCoalescing::JoinIntervals
           Product: new-bugs
           Version: trunk
          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


Seen on Ubuntu Hardy.

regehr at john-home:~/volatile/tmp188$ valgrind --trace-children=yes llvm-gcc -O2
small.c
==15437== Memcheck, a memory error detector.
==15437== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.
==15437== Using LibVEX rev 1804, a library for dynamic binary translation.
==15437== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.
==15437== Using valgrind-3.3.0, a dynamic binary instrumentation framework.
==15437== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.
==15437== For more details, rerun with: -v
==15437== 
==15438== Memcheck, a memory error detector.
==15438== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.
==15438== Using LibVEX rev 1804, a library for dynamic binary translation.
==15438== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.
==15438== Using valgrind-3.3.0, a dynamic binary instrumentation framework.
==15438== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.
==15438== For more details, rerun with: -v
==15438== 
small.c: In function ???uint82???:
small.c:23: warning: overflow in implicit constant conversion
small.c:29: warning: passing argument 2 of ???bar??? makes integer from pointer
without a cast
==15438== Invalid read of size 4
==15438==    at 0x89000D2:
llvm::SimpleRegisterCoalescing::JoinIntervals(llvm::LiveInterval&,
llvm::LiveInterval&, bool&) (in
/home/regehr/z/tmp/llvm-gcc-r80167-install/libexec/gcc/i686-pc-linux-gnu/4.2.1/cc1)
==15438==  Address 0x8 is not stack'd, malloc'd or (recently) free'd
small.c: At top level:
small.c:33: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.
==15438== 
==15438== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 21 from 1)
==15438== malloc/free: in use at exit: 609,875 bytes in 2,011 blocks.
==15438== malloc/free: 8,509 allocs, 6,498 frees, 1,651,899 bytes allocated.
==15438== For counts of detected errors, rerun with: -v
==15438== searching for pointers to 2,011 not-freed blocks.
==15438== checked 1,849,460 bytes.
==15438== 
==15438== LEAK SUMMARY:
==15438==    definitely lost: 0 bytes in 0 blocks.
==15438==      possibly lost: 1,743 bytes in 41 blocks.
==15438==    still reachable: 608,132 bytes in 1,970 blocks.
==15438==         suppressed: 0 bytes in 0 blocks.
==15438== Rerun with --leak-check=full to see details of leaked memory.
==15437== 
==15437== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 21 from 1)
==15437== malloc/free: in use at exit: 18,964 bytes in 73 blocks.
==15437== malloc/free: 491 allocs, 418 frees, 42,016 bytes allocated.
==15437== For counts of detected errors, rerun with: -v
==15437== searching for pointers to 73 not-freed blocks.
==15437== checked 146,056 bytes.
==15437== 
==15437== LEAK SUMMARY:
==15437==    definitely lost: 4,314 bytes in 32 blocks.
==15437==      possibly lost: 0 bytes in 0 blocks.
==15437==    still reachable: 14,650 bytes in 41 blocks.
==15437==         suppressed: 0 bytes in 0 blocks.
==15437== Rerun with --leak-check=full to see details of leaked memory.

regehr at john-home:~/volatile/tmp188$ llvm-gcc -v

Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/z/tmp/llvm-gcc-r80167-install
--program-prefix=llvm-r80167- --enable-languages=c,c++
--enable-llvm=/home/regehr/z/tmp/llvm-r80167 --enable-checking=release
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)

regehr at john-home:~/volatile/tmp188$ cat small.c

char
foo (char _si1, char _si2)
{
  return _si1 && _si2 == 1 ? : (_si1 % _si2);
}

short int
bar (short int _si1, char _si2)
{
  return _si1 && 1 + _si1 && _si1 || _si1 && _si2 && _si1
    && _si2 + _si1 ? : 1;
}

char
baz (unsigned short int _ui1, char _ui2)
{
  return 0 ? : (_ui1 % (unsigned short int) _ui2);
}

char g_51;
uint82 (int p_93, char p_94)
{
  char l_95 = 0xC9230FEFL;
  char l_97 = -10L;
  if (func_86
      (foo
       (baz (g_51 >= 1, l_95),
        l_97)) >= bar (func_52 (1),
                       foo))
    for (0; 1; 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  Thu Aug 27 18:19:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 18:19:03 -0500
Subject: [LLVMbugs] [Bug 4793] [2.6] MC test failures on PPC host
In-Reply-To: <bug-4793-206@http.llvm.org/bugs/>
Message-ID: <200908272319.n7RNJ3v2002029@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4793


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #5 from Chris Lattner <clattner at apple.com>  2009-08-27 18:19:03 ---
Ok, this should be fixed.  Please make sure that the relevant patches are
forwarded to tanya to pull into 2.6.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 27 18:22:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 18:22:41 -0500
Subject: [LLVMbugs] [Bug 4792] [2.6]
 2008-07-22-bitfield-init-after-zero-len-array. c Fails on PPC
In-Reply-To: <bug-4792-206@http.llvm.org/bugs/>
Message-ID: <200908272322.n7RNMfBU002515@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4792


Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|2008-07-22-bitfield-init-   |[2.6] 2008-07-22-bitfield-
                   |after-zero-len-array.c Fails|init-after-zero-len-array.c
                   |on PPC                      |Fails on PPC




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-08-27 18:22:41 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20090824/020741.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 Aug 27 19:44:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Aug 2009 19:44:09 -0500
Subject: [LLVMbugs] [Bug 3913] another miscompilation with -loop-index-split
In-Reply-To: <bug-3913-206@http.llvm.org/bugs/>
Message-ID: <200908280044.n7S0i9FJ013267@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3913


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2009-08-27 19:44:09 ---
Fixed by Jakub here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085779.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 Aug 28 00:33:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 00:33:57 -0500
Subject: [LLVMbugs] [Bug 4810] New: [2.6] ExecutionEngine test failures on
 mixed 64/32 system
Message-ID: <bug-4810-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4810

           Summary: [2.6] ExecutionEngine test failures on mixed 64/32
                    system
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Generic Execution Engine Support
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


My test system is a Debian 32-bit userspace running in a 64-bit kernel. LLVM
detects this as an x86-64 because that's what uname says, even though gcc/as
are 32-bit (supporting the -m64 flag). This system is capable of running both
32-bit and 64-bit programs.

Sample failure:

FAIL: /home/nicholas/2.6/llvm-2.6/test/ExecutionEngine/2003-01-04-LoopTest.ll
Failed with exit(255) at line 2
while running:  lli 2003-01-04-LoopTest.ll.tmp.bc > /dev/null
child process exited abnormally

The *exact* same failure happens on:

  test/ExecutionEngine/2003-01-15-AlignmentTest.ll
  test/ExecutionEngine/2003-05-07-ArgumentTest.ll
  test/ExecutionEngine/hello.ll
  test/ExecutionEngine/hello2.ll
  test/ExecutionEngine/simpletest.ll
  test/ExecutionEngine/test-call.ll
  test/ExecutionEngine/test-fp.ll

It repros on the command-line too, but I'm not sure what to do to debug it.

$ head -2 test/ExecutionEngine/2003-01-04-LoopTest.ll
; RUN: llvm-as %s -o %t.bc -f
; RUN: lli %t.bc > /dev/null
$ Release/bin/llvm-as test/ExecutionEngine/2003-01-04-LoopTest.ll -o x.bc -f
$ Release/bin/lli x.bc
$ echo $?
255
$ valgrind --quiet Release/bin/lli x.bc
==13477== Conditional jump or move depends on uninitialised value(s)
==13477==    at 0x83C2632: (anonymous
namespace)::DefaultJITMemoryManager::allocateNewSlab(unsigned int) (in
/home/nicholas/2.6/llvm-2.6/Release/bin/lli)
==13477==    by 0xFFFF: ???
==13477== 
==13477== Syscall param write(buf) points to uninitialised byte(s)
==13477==    at 0x74EFF03: __write_nocancel (in /lib/i686/cmov/libc-2.9.so)
==13477==    by 0x7489206: _IO_do_write@@GLIBC_2.1 (fileops.c:529)
==13477==    by 0x7488AA9: _IO_file_sync@@GLIBC_2.1 (fileops.c:904)
==13477==    by 0x7482989: fflush_unlocked (iofflush_u.c:40)
==13477==    by 0x70C6D88: op_write_native_code (in
/usr/lib/oprofile/libopagent.so.1.0.0)
==13477==    by 0x83C4CF3: (anonymous
namespace)::OProfileJITEventListener::NotifyFunctionEmitted(llvm::Function
const&, void*, unsigned int, llvm::JITEvent_EmittedFunctionDetails const&) (in
/home/nicholas/2.6/llvm-2.6/Release/bin/lli)
==13477==  Address 0x5203024 is not stack'd, malloc'd or (recently) free'd


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 01:31:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 01:31:21 -0500
Subject: [LLVMbugs] [Bug 4811] New: Make asm-verbose show encoding
Message-ID: <bug-4811-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4811

           Summary: Make asm-verbose show encoding
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


The MCCodeEmitter interface makes it easy to show the encoding of instructions,
we should turn this on for -asm-verbose. It would be nice if this would show
other interesting bits, like the how the relocation fits into the encoding.

This is easy to add, I'm just waiting for the code emitter interface to get a
bit more tested / flushed out.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 03:32:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 03:32:03 -0500
Subject: [LLVMbugs] [Bug 4812] New: ARM large JIT regression between
 2009-07-08 rev 74992 and 2009-07-10 rev 75243
Message-ID: <bug-4812-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4812

           Summary: ARM large JIT regression between 2009-07-08 rev 74992
                    and 2009-07-10 rev 75243
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


by analysing the nightly test results for shark-sheeva-armv5tel tester a large
regression can be detected between 2009-07-08 rev 74992 and 2009-07-10 rev
75243.

==> 2009-07-06-CVS-Log.txt <==
Checked out revision 74823.
==> 2009-07-06-Dejagnu-testrun.sum <==
                ===  Summary ===

# of expected passes            3769
# of unexpected failures        46
# of expected failures          21

==> 2009-07-08-CVS-Log.txt <==
Checked out revision 74992.
==> 2009-07-08-Dejagnu-testrun.sum <==
                ===  Summary ===

# of expected passes            3822
# of unexpected failures        6
# of expected failures          21

==> 2009-07-10-CVS-Log.txt <==
Checked out revision 75243.
==> 2009-07-10-Dejagnu-testrun.sum <==
                ===  Summary ===

# of expected passes            3805
# of unexpected failures        32
# of expected failures          22


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 07:04:19 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 07:04:19 -0500
Subject: [LLVMbugs] [Bug 4813] New: llvm-ld does not work with bytecode
	archives (bca)
Message-ID: <bug-4813-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4813

           Summary: llvm-ld does not work with bytecode archives (bca)
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: e0325716 at student.tuwien.ac.at
                CC: llvmbugs at cs.uiuc.edu, e0325716 at student.tuwien.ac.at


Created an attachment (id=3384)
 --> (http://llvm.org/bugs/attachment.cgi?id=3384)
The bytecode archive that presents this problem.

The following does not work:

>$ opt -insert-edge-profiling gzip.1.o -f -o gzip.2.o
>$ llvm-ld -link-as-library -disable-opt -o gzip.3.o gzip.2.o \
   /localtmp/astifter/llvm/llvm-van-install/lib/libprofile_rt.bca
>$ llc  gzip.3.o -f -o gzip.5.s
>$ gcc -g  gzip.5.s -o gzip
/tmp/cchEXazr.o: In function `main':
/nfs/a5/astifter/astifter/spec_cpu_2000_img/benchspec/CINT2000/164.gzip/run/00000001//spec.c:253:
undefined reference to `llvm_start_edge_profiling'
collect2: ld returned 1 exit status

The function llvm_start_edge_profiling is in libprofile_rt.bca but it seems
that it is not linked in the second step.

This bug was discussed at the llvm-commits list starting with
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090817/084957.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 Aug 28 08:46:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 08:46:36 -0500
Subject: [LLVMbugs] [Bug 3416] ice: Assertion `vi.AliveBlocks.none() &&
 "Shouldn' t be alive across any blocks!"' failed.
In-Reply-To: <bug-3416-206@http.llvm.org/bugs/>
Message-ID: <200908281346.n7SDkaA2024560@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3416


John Regehr <regehr at cs.utah.edu> changed:

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




--- Comment #2 from John Regehr <regehr at cs.utah.edu>  2009-08-28 08:46:35 ---
Works now.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Aug 28 10:35:22 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 10:35:22 -0500
Subject: [LLVMbugs] [Bug 4814] New: possible integer wrong code bug
Message-ID: <bug-4814-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4814

           Summary: possible integer wrong code bug
           Product: new-bugs
           Version: trunk
          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


Seen on Ubuntu Hardy.  

g_96 is initialized to 1 and perhaps at -O3 we're forgetting to zero it in the
loop?

regehr at john-home:~/volatile/tmp190$ llvm-gcc -O2 small.c -o small -w
regehr at john-home:~/volatile/tmp190$ ./small
checksum = 0
regehr at john-home:~/volatile/tmp190$ llvm-gcc -O3 small.c -o small -w
regehr at john-home:~/volatile/tmp190$ ./small
checksum = 1
regehr at john-home:~/volatile/tmp190$ llvm-gcc -v
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/z/tmp/llvm-gcc-r80328-install
--program-prefix=llvm-r80328- --enable-languages=c,c++
--enable-llvm=/home/regehr/z/tmp/llvm-r80328 --enable-checking=release
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)
regehr at john-home:~/volatile/tmp190$ cat small.c
#include <stdint.h>
#include <limits.h>
#include <stdio.h>

#define safe_rshift_macro_int8_t_s_u(left,right) \
  (((((int8_t)(left)) < ((int8_t)0)) \
                         || (((unsigned int)(right)) >=
sizeof(int8_t)*CHAR_BIT)) \
                        ? ((int8_t)(left)) \
                        : (((int8_t)(left)) >> ((unsigned int)(right))))

static int8_t
safe_rshift_func_int8_t_s_u(int8_t _left, unsigned int _right)
{
  return safe_rshift_macro_int8_t_s_u(_left,_right);
}

struct S0
{
  uint8_t f0;
  uint8_t f1;
  uint8_t f2;
};

uint8_t g_3 = 0xB4E1265AL;
uint8_t g_96 = 1;
volatile uint8_t g_100;

struct S0 func_47 (void);
struct S0 func_47 (void)
{
  struct S0 l_74 = {
    0xB9CEB518L, 0x4A9CL, 1
  };
  return l_74;
}

int32_t func_18 (struct S0 p_19);
int32_t func_18 (struct S0 p_19)
{
  if (g_96 >= (1 || p_19.f1))
    g_100;
  return 1;
}

int32_t func_4 (int64_t p_5, uint32_t p_6);
int32_t func_4 (int64_t p_5, uint32_t p_6)
{
  struct S0 l_116 = {
    0xA76FC585L, 7L, 1
  };
  func_18 (l_116);
  return l_116.f2;
}

void func_1 (void);
void func_1 (void)
{
  uint8_t l_2;
  uint8_t l_120 = 0;
  for (l_2 = 0; l_2 <= 0; l_2++)
    if (safe_rshift_func_int8_t_s_u (g_3, func_4 (0, 1)))
      for (g_96 = 0; 0; 1)
        {
        }
    else if (func_18 (func_47 ()))
      for (1; l_120; l_120 = 1)
        {
        }
}

int main (void)
{
  func_1 ();
  printf ("checksum = %d\n", g_96);
  return 0;
}


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

From bugzilla-daemon at cs.uiuc.edu  Fri Aug 28 10:42:33 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 10:42:33 -0500
Subject: [LLVMbugs] [Bug 4815] New: Assertion `NodePtr && "--'d off the
 beginning of an ilist!" ' failed.
Message-ID: <bug-4815-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4815

           Summary: Assertion `NodePtr && "--'d off the beginning of an
                    ilist!"' failed.
           Product: new-bugs
           Version: trunk
          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


Seen on Ubuntu Hardy.

regehr at john-home:~/volatile/tmp191$ llvm-gcc -O small.c
cc1: /home/regehr/z/tmp/llvm-r80328/include/llvm/ADT/ilist.h:212:
llvm::ilist_iterator<NodeTy>& llvm::ilist_iterator<NodeTy>::operator--() [with
NodeTy = llvm::Instruction]: Assertion `NodePtr && "--'d off the beginning of
an ilist!"' failed.
small.c:40: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.

regehr at john-home:~/volatile/tmp191$ llvm-gcc -v

Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../configure
--prefix=/home/regehr/z/tmp/llvm-gcc-r80328-install
--program-prefix=llvm-r80328- --enable-languages=c,c++
--enable-llvm=/home/regehr/z/tmp/llvm-r80328 --enable-checking=release
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)

regehr at john-home:~/volatile/tmp191$ cat small.c

char foo (char ui1, char ui2)
{
  return 1 + ui2;
}

int bar (int ui1, int ui2)
{
  return 0;
}

char g9;
char g80;
int g90;

char func10 (int p12, char p13, char p14)
{
  return 0;
}

int func68 (int p69)
{
  if (1 < bar(0,0))
    p69 = 1;
  if (func10 (1, 1, g9))
    return p69;
  return 0;
}

void uint83 (int32p54)
{
  int l60;
  for (g80 = 1;; g80 = foo (g80, 1)) {
    l60=1;
    if (l60)
      {
        g80 = func68 (1);
        g80 = g90;
      }
  }
}


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 11:15:17 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 11:15:17 -0500
Subject: [LLVMbugs] [Bug 4813] llvm-ld does not work with bytecode archives
	(bca)
In-Reply-To: <bug-4813-206@http.llvm.org/bugs/>
Message-ID: <200908281615.n7SGFHo2010626@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4813


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2009-08-28 11:15:16 ---
Ah! The makefile rules for building libprofile_rt.bca were broken, the archive
didn't actually have any bitcode in it.

Fixed in r80373.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 11:36:48 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 11:36:48 -0500
Subject: [LLVMbugs] [Bug 4816] New: JIT lazy compilation broken on Mac/PPC
Message-ID: <bug-4816-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4816

           Summary: JIT lazy compilation broken on Mac/PPC
           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=3386)
 --> (http://llvm.org/bugs/attachment.cgi?id=3386)
x.bc

Disabling lazy compilation fails on PPC:

Release/bin/lli -disable-lazy-compilation  ~/x.bc
Assertion failed: (0 && "This target doesn't implement
emitFunctionStubAtAddr!"), function emitFunctionStubAtAddr, file
/Users/acab/clamav-devel/libclamav/llvm/llvm/include/llvm/Target/TargetJITInfo.h,
line 65.
0   lli               0x0065d50c llvm::sys::RWMutexImpl::writer_release() + 476
1   lli               0x0065dddc llvm::sys::RemoveFileOnSignal(llvm::sys::Path
const&, std::string*) + 1244
2   libSystem.B.dylib 0x92f639fc _sigtramp + 68
Stack dump:
0.      Program arguments: Release/bin/lli -disable-lazy-compilation
/Users/acab/x.bc
Abort trap


It seems that X86 is the only target that implements that method.

In this case it is simply a matter of ordering the functions bottom-up
according to the callgraph, and then non-lazy compilation should succeeed on
the Mac too in this case, BUT it would fail for mutually recursive functions.

Is there a workaround that doesn't need emitFunctinStubAtAddr, or should I just
always use lazy compilation on Mac/PPC? (rather anything non-X86?)


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 17:32:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 17:32:01 -0500
Subject: [LLVMbugs] [Bug 4820] New: [2.6] make clean removes llvm-config
 before it' s done using it
Message-ID: <bug-4820-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4820

           Summary: [2.6] make clean removes llvm-config before it's done
                    using it
           Product: Build scripts
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3387)
 --> (http://llvm.org/bugs/attachment.cgi?id=3387)
full transcript

Steps:
  1. From clean, build LLVM 2.6.
  2. Run "make clean".

I get output like this:

make[1]: Entering directory
`/usr/local/google/home/nlewycky/2.6/llvm-2.6/bindings'
make[2]: Entering directory
`/usr/local/google/home/nlewycky/2.6/llvm-2.6/bindings/ocaml'
make[3]: /usr/local/google/home/nlewycky/2.6/llvm-2.6/Release/bin/llvm-config:
Command not found
make[3]: Entering directory
`/usr/local/google/home/nlewycky/2.6/llvm-2.6/bindings/ocaml/llvm'
make[3]: /usr/local/google/home/nlewycky/2.6/llvm-2.6/Release/bin/llvm-config:
Command not found
make[3]: /usr/local/google/home/nlewycky/2.6/llvm-2.6/Release/bin/llvm-config:
Command not found
[...]

partway through the build. To help clarify the order of operations, I've
attached the full transcript of the make clean run.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Aug 28 19:29:55 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 19:29:55 -0500
Subject: [LLVMbugs] [Bug 4821] New: Vector splat not selected
Message-ID: <bug-4821-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4821

           Summary: Vector splat not selected
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu, bob.wilson at apple.com


Created an attachment (id=3388)
 --> (http://llvm.org/bugs/attachment.cgi?id=3388)
bugpoint-reduced-simplified.ll

Consider the attached testcase. Here we have:

$ ./llc -mattr=+neon bugpoint-reduced-simplified.bc
LLVM ERROR: Cannot yet select: 0x187e7a0: v4i32 = BUILD_VECTOR 0x187e6a0,
0x187e6a0, 0x187e6a0, 0x187e6a0


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 28 20:47:24 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Aug 2009 20:47:24 -0500
Subject: [LLVMbugs] [Bug 4822] New: Unreachable_internal error occurs when
 module is deleted on Win64
Message-ID: <bug-4822-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4822

           Summary: Unreachable_internal error occurs when module is deleted
                    on Win64
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gareth_morgan77 at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3389)
 --> (http://llvm.org/bugs/attachment.cgi?id=3389)
IR that will cause an assert on module delete

The attached IR causes a crash on module delete.  If the IR is compiled and
read as bitcode, then getPointerToFunction is called, and then the function and
module are deleted, (as shown in snippet below) then an assert will occur in
the destructor.  This only occurs on 64-bit builds.

        memBuf = llvm::MemoryBuffer::getFile("c:\\temp\\shutdownassert.bc");
        std::string errStr;
        mainModule = llvm::ParseBitcodeFile(memBuf, LLVMGetContext(), &errStr);
        if(NULL == mainModule)
        {
                printf("Module errStr: %s\n", errStr.c_str());
                exit(1);
        }

        for(llvm::Module::iterator moditer = mainModule->begin(), modend =
mainModule->end(); moditer != modend; moditer++)
        {
                llvm::Function * func = moditer;
                std::cout<<func->getName().str()<<" "<<func<<std::endl;
        }

        Function *func=mainModule->getFunction("main");
        LLVMGetExecutionEngine()->getPointerToFunction(func);

        for(llvm::Module::iterator moditer = mainModule->begin(), modend =
mainModule->end(); moditer != modend; moditer++)
        {
                LLVMGetExecutionEngine()->freeMachineCodeForFunction(func);
        }
        delete mainModule;


The error message is:
While deleting: float (opaque*, i32, <4 x float>, <4 x float>, <4 x float>)*
%computeVal
An asserting value handle still pointed to this value!
UNREACHABLE executed at ..\..\..\lib\VMCore\Value.cpp:500!

The callstack of the assert is:
        LLVMTest.exe!llvm::llvm_unreachable_internal(const char *
msg=0x0000000140625948, const char * file=0x0000000140625928, unsigned int
line=0x000001f4)  Line 71  C++
>	LLVMTest.exe!llvm::ValueHandleBase::ValueIsDeleted(llvm::Value * V=0x0000000000877cc0)  Line 503	C++
        LLVMTest.exe!llvm::Value::~Value()  Line 74     C++
        LLVMTest.exe!llvm::User::~User()  Line 79 + 0xa bytes   C++
        LLVMTest.exe!llvm::Constant::~Constant()  + 0x2d bytes  C++
        LLVMTest.exe!llvm::GlobalValue::~GlobalValue()  Line 75 + 0x1a bytes   
C++
        LLVMTest.exe!llvm::Function::~Function()  Line 179 + 0x40 bytes C++
        LLVMTest.exe!llvm::Function::`scalar deleting destructor'()  + 0x31
bytes       C++
       
LLVMTest.exe!llvm::ilist_node_traits<llvm::Function>::deleteNode(llvm::Function
* V=0x0000000000877cc0)  Line 110 + 0x53 bytes  C++
       
LLVMTest.exe!llvm::iplist<llvm::Function,llvm::ilist_traits<llvm::Function>
>::erase(llvm::ilist_iterator<llvm::Function> where={...})  Line 462        C++
       
LLVMTest.exe!llvm::iplist<llvm::Function,llvm::ilist_traits<llvm::Function>
>::erase(llvm::ilist_iterator<llvm::Function> first={...},
llvm::ilist_iterator<llvm::Function> last={...})  Line 526 + 0x14 bytes  C++
       
LLVMTest.exe!llvm::iplist<llvm::Function,llvm::ilist_traits<llvm::Function>
>::clear()  Line 530 + 0x66 bytes   C++
        LLVMTest.exe!llvm::Module::~Module()  Line 69   C++
        LLVMTest.exe!llvm::Module::`scalar deleting destructor'()  + 0x31 bytes
C++


The last place ValueHandleBase::AddToUseList is called before the assert is
here, inside of JITResolver::getFunctionStub:

        CausticGLES.dll!llvm::ValueHandleBase::AddToUseList()  Line 425 C++
       
CausticGLES.dll!llvm::ValueHandleBase::ValueHandleBase(llvm::ValueHandleBase::HandleBaseKind
Kind=Assert, llvm::Value * V=0x0000000005dbe0f0)  Line 63  C++
       
CausticGLES.dll!llvm::AssertingVH<llvm::Function>::AssertingVH<llvm::Function>(llvm::Function
* P=0x0000000005dbe0f0)  Line 190 + 0x41 bytes    C++
        CausticGLES.dll!`anonymous
namespace'::JITResolver::getFunctionStub(llvm::Function * F=0x0000000005dbe0f0)
 Line 194 + 0x12 bytes       C++
        CausticGLES.dll!`anonymous
namespace'::JITEmitter::getPointerToGlobal(llvm::GlobalValue *
V=0x0000000005dbe0f0, void * Reference=0x00000000027103b6, bool
DoesntNeedStub=false)  Line 662 + 0x16 bytes  C++
        CausticGLES.dll!`anonymous
namespace'::JITEmitter::finishFunction(llvm::MachineFunction & F={...})  Line
1016 + 0x47 bytes      C++
        CausticGLES.dll!`anonymous
namespace'::Emitter<llvm::JITCodeEmitter>::runOnMachineFunction(llvm::MachineFunction
& MF={...})  Line 151 + 0x26 bytes     C++
        CausticGLES.dll!llvm::MachineFunctionPass::runOnFunction(llvm::Function
& F={...})  Line 32     C++
        CausticGLES.dll!llvm::FPPassManager::runOnFunction(llvm::Function &
F={...})  Line 1372 + 0x1d bytes    C++
        CausticGLES.dll!llvm::FunctionPassManagerImpl::run(llvm::Function &
F={...})  Line 1324 + 0x20 bytes    C++
        CausticGLES.dll!llvm::FunctionPassManager::run(llvm::Function &
F={...})  Line 1254 + 0x19 bytes        C++
        CausticGLES.dll!llvm::JIT::runJITOnFunctionUnlocked(llvm::Function *
F=0x0000000005d803d0, const llvm::MutexGuard & locked={...})  Line 595     C++
        CausticGLES.dll!llvm::JIT::getPointerToFunction(llvm::Function *
F=0x0000000005d803d0)  Line 668        C++


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 29 00:50:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Aug 2009 00:50:09 -0500
Subject: [LLVMbugs] [Bug 4825] New: [2.6] xpass test/FrontendC/
 2009-08-11-AsmBlocksComplexJumpTarget.c
Message-ID: <bug-4825-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4825

           Summary: [2.6] xpass test/FrontendC/2009-08-11-
                    AsmBlocksComplexJumpTarget.c
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3390)
 --> (http://llvm.org/bugs/attachment.cgi?id=3390)
.s output

On x86-64 GNU/Linux, test/FrontendC/2009-08-11-AsmBlocksComplexJumpTarget.c
unexpectedly passes. I've attached the entire assembly 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  Sat Aug 29 00:54:26 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Aug 2009 00:54:26 -0500
Subject: [LLVMbugs] [Bug 4795] TableGen assertion
In-Reply-To: <bug-4795-206@http.llvm.org/bugs/>
Message-ID: <200908290554.n7T5sQSq016868@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4795


Bob Wilson <bob.wilson at apple.com> changed:

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




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-08-29 00:54:26 ---
Fixed in svn r80423.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 29 17:26:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Aug 2009 17:26:58 -0500
Subject: [LLVMbugs] [Bug 4814] possible integer wrong code bug
In-Reply-To: <bug-4814-206@http.llvm.org/bugs/>
Message-ID: <200908292226.n7TMQwCT023173@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4814


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #1 from Dan Gohman <gohman at apple.com>  2009-08-29 17:26:57 ---
This was introduced by my recent change to the handling of
8-bit conditional moves, and is fixed here:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085877.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 Aug 30 00:59:58 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 00:59:58 -0500
Subject: [LLVMbugs] [Bug 4831] New: Empty non-void noreturn function gives
 wrong diagnostic
Message-ID: <bug-4831-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4831

           Summary: Empty non-void noreturn function gives wrong diagnostic
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rideau3 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following code:

__attribute__((noreturn)) int foo() { }

Gives an error for "function declared 'noreturn' should not return", but the
function cannot return at all. The diagnostic should be the same as without the
noreturn attribute.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 30 01:24:57 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 01:24:57 -0500
Subject: [LLVMbugs] [Bug 4831] Empty non-void noreturn function gives wrong
	diagnostic
In-Reply-To: <bug-4831-206@http.llvm.org/bugs/>
Message-ID: <200908300624.n7U6Ov8P017430@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4831


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2009-08-30 01:24:57 ---
The warning is correct, the function can return when it should not. Perhaps the
phrasing could be better?

gcc says "does return", instead.

--
ddunbar at giles:tmp$ cat t.c
__attribute__((noreturn)) int foo() { }

ddunbar at giles:tmp$ gcc -c t.c
t.c: In function 'foo':
t.c:1: warning: 'noreturn' function does return
ddunbar at giles:tmp$ clang -c t.c
t.c:1:39: error: function declared 'noreturn' should not return
[-Winvalid-noreturn]
__attribute__((noreturn)) int foo() { }
                                      ^
1 diagnostic generated.
--


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug 30 01:44:39 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 01:44:39 -0500
Subject: [LLVMbugs] [Bug 3355] llvm handles inline functions in C99 poorly
In-Reply-To: <bug-3355-206@http.llvm.org/bugs/>
Message-ID: <200908300644.n7U6idPt019845@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3355


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #3 from Chris Lattner <clattner at apple.com>  2009-08-30 01:44:39 ---
An IR extension was added for this, available_externally linkage.  This is done
now.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Aug 30 12:19:06 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 12:19:06 -0500
Subject: [LLVMbugs] [Bug 4833] New: Information about the sign is lost for
 some NEON intrinsics
Message-ID: <bug-4833-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4833

           Summary: Information about the sign is lost for some NEON
                    intrinsics
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru
                CC: llvmbugs at cs.uiuc.edu, bob.wilson at apple.com


Consider the following code:

#include "arm_neon.h"

void test_vget_lanes8 (void)
{
  int8_t out_int8_t;
  int8x8_t arg0_int8x8_t;

  out_int8_t = vget_lane_s8 (arg0_int8x8_t, 1);
}

We're emitting invalid vmov.u8 here (in -O0) since the information about
zero/sign extension is lost during frontend expansion.

Out current approach of deciding whether we need to do sign/zero extension is
pretty fragile (based on DAG combiner hook), we really need to use an intrinsic
to catch this behaviour for i8/i16.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 30 15:33:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 15:33:36 -0500
Subject: [LLVMbugs] [Bug 4834] New: opt and jit crashes at -O2 for simple
	code.
Message-ID: <bug-4834-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4834

           Summary: opt and jit crashes at -O2 for simple code.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ofv at wanadoo.es
                CC: llvmbugs at cs.uiuc.edu


The attached bitecode file causes opt to crash at -O2.

0.      Running pass 'CallGraph Pass Manager' on module 'code.bc'.
segmentation fault

If I run my jit with assert-enabled LLVM it outputs (I have no access to an
assert-enabled opt):

Assertion failed: Begin + idx < End, file
D:\dev\llvm\head\include/llvm/ADT/SmallVector.h, line 129


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 30 15:39:03 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 15:39:03 -0500
Subject: [LLVMbugs] [Bug 4748] Instcombine transform confuse basicaa
In-Reply-To: <bug-4748-206@http.llvm.org/bugs/>
Message-ID: <200908302039.n7UKd3I1030950@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4748


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #11 from Chris Lattner <clattner at apple.com>  2009-08-30 15:39:02 ---
Fixed here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085937.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 Aug 30 20:40:09 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 20:40:09 -0500
Subject: [LLVMbugs] [Bug 4747] [sjlj EH] assertion failed: setOperand() out
	of range!
In-Reply-To: <bug-4747-206@http.llvm.org/bugs/>
Message-ID: <200908310140.n7V1e9DP003632@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4747


Jim Grosbach <grosbach at apple.com> changed:

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




--- Comment #4 from Jim Grosbach <grosbach at apple.com>  2009-08-30 20:40:08 ---
Shared landing pads run into trouble with SJLJ, as the dispatch table is
mapped to call sites, and merging the pads will throw that off. There needs
to be a one-to-one mapping of landing pad exception table entries to invoke
call points.

Detecting the shared pad during lowering of SJLJ info insn't sufficient, as
the dispatch function may still need separate destinations to properly
handle phi-nodes.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 30 21:16:01 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 21:16:01 -0500
Subject: [LLVMbugs] [Bug 4835] New: Callgraph should use ValueHandle APIs
Message-ID: <bug-4835-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4835

           Summary: Callgraph should use ValueHandle APIs
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Analyses
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 3601


Created an attachment (id=3395)
 --> (http://llvm.org/bugs/attachment.cgi?id=3395)
patch that uses AssertingVH

Here's a patch that converts the callgraph to using AssertingVH.  It works
great except (and caught a bunch of bugs in (e.g.) argpromotion), except that
scalar and loop passes aren't (and shouldn't be) callgraph aware.  Instead, the
callgraph should use CallbackVH's in each CallGraphNode to keep a smart
reference to a call site.  When (e.g.) simplifycfg deletes a dead block that
has a call site in it, it would automatically cause the entry from the
callgraph node to be removed.

I don't have time to work on this right now, so I'm posting this patch in case
anyone else is interested.  This may help fix Bug 3601.

-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 Aug 30 22:16:36 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Aug 2009 22:16:36 -0500
Subject: [LLVMbugs] [Bug 4834] opt and jit crashes at -O2 for simple code.
In-Reply-To: <bug-4834-206@http.llvm.org/bugs/>
Message-ID: <200908310316.n7V3GaXj015538@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4834


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #6 from Chris Lattner <clattner at apple.com>  2009-08-30 22:16:35 ---
After going back and forth on this with several different solutions, I finally
converged.  Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090824/085963.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 Aug 31 01:01:49 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 01:01:49 -0500
Subject: [LLVMbugs] [Bug 4835] Callgraph should use ValueHandle APIs
In-Reply-To: <bug-4835-206@http.llvm.org/bugs/>
Message-ID: <200908310601.n7V61nRn003415@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4835


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2009-08-31 01:01:48 ---
This seems to be essentially the same issue as PR4029, though the
suggested solutions differ.  That PR also contains a patch to make
the callgraph use AssertingVH.

*** This bug has been marked as a duplicate of bug 4029 ***


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 08:15:44 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 08:15:44 -0500
Subject: [LLVMbugs] [Bug 4839] New: SimplifyCFG does not eliminate phi nodes
 with one predecessor
Message-ID: <bug-4839-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4839

           Summary: SimplifyCFG does not eliminate phi nodes with one
                    predecessor
           Product: new-bugs
           Version: 2.5
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: trivial
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: andrelct at dcc.ufmg.br
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3396)
 --> (http://llvm.org/bugs/attachment.cgi?id=3396)
Test case

in the description of SimplifyCFG, it says that it can "Eliminates PHI nodes
for basic blocks with a single predecessor."

I tested this pass with a program that has phi nodes with a single predecessor
and it did not remove it. I also looked in the code and found nothing to remove
it.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 31 12:58:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 12:58:43 -0500
Subject: [LLVMbugs] [Bug 4825] [2.6] xpass
 test/FrontendC/2009-08-11-AsmBlocksComplexJumpTarget .c
In-Reply-To: <bug-4825-206@http.llvm.org/bugs/>
Message-ID: <200908311758.n7VHwhkM010225@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4825


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>  2009-08-31 12:58:42 ---
Marked as passing on x86.
http://llvm.org/viewvc/llvm-project?rev=80596&view=rev


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 13:17:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 13:17:50 -0500
Subject: [LLVMbugs] [Bug 4833] Information about the sign is lost for some
 NEON intrinsics
In-Reply-To: <bug-4833-206@http.llvm.org/bugs/>
Message-ID: <200908311817.n7VIHoBP012717@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4833


Bob Wilson <bob.wilson at apple.com> changed:

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




--- Comment #1 from Bob Wilson <bob.wilson at apple.com>  2009-08-31 13:17:47 ---
This code (from the gcc testsuite) is a bad test.  Only the low byte of the
result is relevant (and not even that since the value is dead), so it doesn't
matter whether llvm selects vmov.s8 or vmov.u8 in this case.  I tried changing
the code to return the value as a signed int, and we correctly generate
vmov.s8.  I don't think there is a bug here.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Aug 31 14:27:41 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 14:27:41 -0500
Subject: [LLVMbugs] [Bug 4841] New: I686 instruction is emitted regardless
 of -mcpu option
Message-ID: <bug-4841-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4841

           Summary: I686 instruction is emitted regardless of -mcpu option
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3398)
 --> (http://llvm.org/bugs/attachment.cgi?id=3398)
BC file

LLVM r80583. `llc -mcpu=pentium pentium.bc' generates `cmovne' 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  Mon Aug 31 15:32:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 15:32:43 -0500
Subject: [LLVMbugs] [Bug 4842] New: C++ exceptions,
 invalid code generated (i386-pc-mingw32)
Message-ID: <bug-4842-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4842

           Summary: C++ exceptions, invalid code generated (i386-pc-mingw32)
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: d.g.gorbachev at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=3399)
 --> (http://llvm.org/bugs/attachment.cgi?id=3399)
ZIP archive with C, BC, and S files

LLVM-GCC r80583, configured with `--enable-sjlj-exceptions'


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 18:06:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 18:06:27 -0500
Subject: [LLVMbugs] [Bug 4843] New: TwoAddressInstructionPass crashes with
 multiple tied destination registers
Message-ID: <bug-4843-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4843

           Summary: TwoAddressInstructionPass crashes with multiple tied
                    destination registers
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


The TwoAddressInstructionPass has problems with instructions that define
multiple registers, where each destination register is tied to a different
source register.  There are several problems here, but most notably when the
tied source operands are the same register, the TwoAddressInstructionPass may
have to insert multiple copies to get that register copied to the tied
destinations.

The following testcase will demonstrate the problem:

define <4 x i16> @v2regbug(<4 x i16>* %B) nounwind {
        %tmp1 = load <4 x i16>* %B
        %tmp2 = shufflevector <4 x i16> %tmp1, <4 x i16> undef, <4 x i32><i32
0, i32 0, i32 1, i32 1>
        ret <4 x i16> %tmp2
}

Assertion failed: (Removed && "Register is not used by this instruction!"),
function removeVirtualRegisterKilled, file
/Users/bwilson/local/llvm/llvm/include/llvm/CodeGen/LiveVariables.h, line 211.
0   llc               0x00b01fdc PrintStackTrace(void*) + 45
1   llc               0x00b025e4 SignalHandler(int) + 410
2   libSystem.B.dylib 0x98115bfb _sigtramp + 43
3   libSystem.B.dylib 0xffffffff _sigtramp + 1743692847
4   libSystem.B.dylib 0x981a3ba5 raise + 26
5   libSystem.B.dylib 0x981b9c5c abort + 93
6   libSystem.B.dylib 0x981a6804 __pthread_markcancel + 0
7   llc               0x0090452b
llvm::LiveVariables::removeVirtualRegisterKilled(unsigned int,
llvm::MachineInstr*) + 277
8   llc               0x00903b2f (anonymous
namespace)::TwoAddressInstructionPass::runOnMachineFunction(llvm::MachineFunction&)
+ 5591
9   llc               0x00850fe7
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 77
10  llc               0x00a8377b
llvm::FPPassManager::runOnFunction(llvm::Function&) + 297
11  llc               0x00a83cae
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 112
12  llc               0x00a83e7a
llvm::FunctionPassManager::run(llvm::Function&) + 130
13  llc               0x0002f710 main + 3579
14  llc               0x0002e289 start + 53
15  llc               0x00000003 start + 4294778287


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 18:40:27 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 18:40:27 -0500
Subject: [LLVMbugs] [Bug 4844] New: instcombine form still confuses basicaa
Message-ID: <bug-4844-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4844

           Summary: instcombine form still confuses basicaa
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan at fallingsnow.net
                CC: llvmbugs at cs.uiuc.edu


Instcombine continues to put code in forms that confuse basicaa.

Attached are 3 files

primary.ll is quite big, but easily shows instcombine creating the bad form:

llvm-as < primary.ll | opt -instcombine | llvm-dis > primary-bad.ll
diff -u primary.ll primary-bad.ll

It's easy to see instcombine flipping BC(GEP)'s to strange GEP(BC)'s. It's
these GEP(BC) forms that confuse basicaa. Which brings me to the other two
files.

gep_bc.ll shows a reduced form of the code in primary.ll, using the GEP(BC)
forms that instcombine creates. Running this through opt will result in no
change to the output.


bc_gep.ll is the same as gep_bc.ll, but the GEP(BC) forms are undone back to
BC(GEP) forms, the same as before instcombine changed them.

Running bc_gep.ll through opt -std-compile-opts, you'll see significant
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  Mon Aug 31 22:03:43 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 22:03:43 -0500
Subject: [LLVMbugs] [Bug 4845] New: crashing inside JITed code when running
	on x86_64
Message-ID: <bug-4845-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4845

           Summary: crashing inside JITed code when running on x86_64
           Product: libraries
           Version: 2.6
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: major
          Priority: P2
         Component: Target-Independent JIT
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lsansonetti at apple.com
                CC: llvmbugs at cs.uiuc.edu


It looks like the JIT is generating strange code when ran in x86_64 mode. My
project (MacRuby) gets a lot of crashes at runtime. This was working before
with an older revision of trunk.

To investigate, I did a simple 64-bit build of LLVM.

$ make EXTRA_OPTIONS="-m64"

Then, I ran the HowToUseJIT example and it crashed. gdb says the crash occurred
inside the JIT'ed code.

(gdb) r
Starting program: /Users/lrz/src/llvm-26-fresh/Release/examples/HowToUseJIT 
Reading symbols for shared libraries +++. done
We just constructed this LLVM module:

; ModuleID = 'test'

define i32 @add1(i32 %AnArg) {
EntryBlock:
  %addresult = add i32 1, %AnArg                  ; <i32> [#uses=1]
  ret i32 %addresult
}

define i32 @foo() {
EntryBlock:
  %add1 = tail call i32 @add1(i32 10)             ; <i32> [#uses=1]
  ret i32 %add1
}


Running foo: 
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x000000005fbff49c
0x0000000100c00013 in ?? ()
(gdb) bt
#0  0x0000000100c00013 in ?? ()
#1  0x000000010000dbb5 in main ()

The same problem appears with Debug and Release builds.

Reproduced on: 2.6 branch, trunk.
Environment: Core 2 Duo, Mac OS 10.6.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 23:19:21 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 23:19:21 -0500
Subject: [LLVMbugs] [Bug 4843] TwoAddressInstructionPass crashes with
 multiple tied destination registers
In-Reply-To: <bug-4843-206@http.llvm.org/bugs/>
Message-ID: <200909010419.n814JLTi024373@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=4843


Bob Wilson <bob.wilson at apple.com> changed:

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




--- Comment #2 from Bob Wilson <bob.wilson at apple.com>  2009-08-31 23:19:21 ---
Fixed in r80654.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?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 Aug 31 23:54:50 2009
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 31 Aug 2009 23:54:50 -0500
Subject: [LLVMbugs] [Bug 4846] New: Diagnostics for declaration typos can be
 terrible, and are in any case inconsistent
Message-ID: <bug-4846-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=4846

           Summary: Diagnostics for declaration typos can be terrible, and
                    are in any case inconsistent
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rideau3 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


At global scope, the following declaration (a common typo) gives a nasty error:

sturct foo { };

If the same declaration is made in function scope:

void bar () { sturct foo { }; }

then the second error disappears. This is inconsistent. Furthermore, the same
error is encountered if the function argument parentheses are accidentally
omitted at global scope, or with "int foo :", which could be a label or (in
C++) a base specifier.

The error message should be consistent between the two scopes, and, if the
"invalid token" message is retained, it should probably disambiguate for some
of the common misspellings.


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