From bugzilla-daemon at cs.uiuc.edu Mon Jan 1 06:26:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 1 Jan 2007 06:26:08 -0600 Subject: [LLVMbugs] [Bug 1068] ICE with "-g" option specified Message-ID: <200701011226.l01CQ8cU022996@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1068 jlaskey at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From jlaskey at apple.com 2007-01-01 06:26 ------- Unbounded array needed further check. Index: gcc/llvm-debug.cpp =============================================================== ==== --- gcc/llvm-debug.cpp (revision 121787) +++ gcc/llvm-debug.cpp (working copy) @@ -593,8 +593,10 @@ tree MinValue = TYPE_MIN_VALUE(Domain); tree MaxValue = TYPE_MAX_VALUE(Domain); if (MinValue && MaxValue && - TREE_CODE(MinValue) == INTEGER_CST && - TREE_CODE(MaxValue) == INTEGER_CST) { + TREE_CODE(MinValue) == INTEGER_CST && + TREE_CODE(MaxValue) == INTEGER_CST && + host_integerp(MinValue, 0) && + host_integerp(MaxValue, 0)) { Subrange->setLo(tree_low_cst(MinValue, 0)); Subrange->setHi(tree_low_cst(MaxValue, 0)); } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 2 22:52:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 2 Jan 2007 22:52:03 -0600 Subject: [LLVMbugs] [Bug 1071] NEW: X86 can generate movb al, eax ? Message-ID: <200701030452.l034q3Ia006964@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1071 Summary: X86 can generate movb al, eax ? Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com While testing this function: define bool i32() { ret bool true; } I noticed that it is possible for the X86 backend to generate: movb al, eax which doesn't assemble. To make this happen, go to CodeGen/SelectionISel.cpp in the visitRetInst function and make it *not* do an extend (of any kind) for boolean. This will cause the Legalizer to apply an ANY_EXTEND which, presumably, means that not doing an extend is okay. While that change is not correct, the X86 backend shouldn't generate incorrect instructions either. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 2 23:59:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 2 Jan 2007 23:59:45 -0600 Subject: [LLVMbugs] [Bug 1071] X86 can generate movb al, eax ? Message-ID: <200701030559.l035xj9W008049@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1071 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From evan.cheng at apple.com 2007-01-02 23:59 ------- Looks like legalizer isn't inserting an anyext to i32 so we ended up with a movb with a 32-bit register destination. If you let visitRet insert the anyext node, then it will not generate this illegal instruction. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 3 19:39:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 3 Jan 2007 19:39:38 -0600 Subject: [LLVMbugs] [Bug 1072] NEW: Investigate elimination of the -raise pass Message-ID: <200701040139.l041dc0A005522@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1072 Summary: Investigate elimination of the -raise pass Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: dpatel at apple.com The -raise pass was mainly around to eliminate casts from int -> uint etc through some tricky transformations. With the recent signless work, this pass should be mostly useless. Eliminating it is good as it reduces compile time (one fewer pass), eliminates a bunch of ugly code (the raise pass is very nasty), and potentially eliminating the LLVMTransforms library (ExprTypeConvert.cpp, LevelRaise.cpp, and TransformInternals.cpp). We should do performance analysis (e.g. spec2k and 2k6) to see whether this is completely dead, and if not, reimplement the useful pieces in other passes. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 00:11:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 00:11:28 -0600 Subject: [LLVMbugs] [Bug 1063] test/Regression/Transforms/InstCombine/2006-04-01-InfLoop fails on amd64 Message-ID: <200701040611.l046BSDR010161@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1063 chandlerc at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From chandlerc at gmail.com 2007-01-04 00:11 ------- This is fixed by compiling LLVM and LLVM-GCC with GCC 4.0.3, rather than 4.1.1, so likely a miscompilation... ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 00:50:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 00:50:36 -0600 Subject: [LLVMbugs] [Bug 1073] NEW: When destructing a module in the LTO, ilist assertion fails. Message-ID: <200701040650.l046oaVl010865@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1073 Summary: When destructing a module in the LTO, ilist assertion fails. Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: chandlerc at gmail.com When destroying the LTO, and therefore its various modules, an assertion fails during the clearing of the global variable ilist in the module. A backtrace is as follows: ld: /home/chandlerc/code/compilers/llvm/include/llvm/ADT/ilist:120: typename bidirectional_iterator::reference llvm::ilist_iterator::operator*() const [with NodeTy = llvm::GlobalVariable]: Assertion `Traits::getNext(NodePtr) != 0 && "Dereferencing end()!"' failed. Program received signal SIGABRT, Aborted. [Switching to Thread 47687191382816 (LWP 4590)] _______________________________________________________________________________ Error while running hook_stop: Value can't be converted to integer. 0x00002b5f09f997b5 in raise () from /lib/libc.so.6 gdb> bt #0 0x00002b5f09f997b5 in raise () from /lib/libc.so.6 #1 0x00002b5f09f9aa4e in abort () from /lib/libc.so.6 #2 0x00002b5f09f93156 in __assert_fail () from /lib/libc.so.6 #3 0x00000000005e33b7 in llvm::ilist_iterator::operator* (this=0x7fffa13e3c00) at /home/chandlerc/code/compilers/llvm/include/llvm/ADT/ilist:120 #4 0x00000000005e33d3 in llvm::ilist_iterator::operator-> (this=0x7fffa13e3c00) at /home/chandlerc/code/compilers/llvm/include/llvm/ADT/ilist:123 #5 0x00000000009aba7a in llvm::SymbolTableListTraits >::setParent (this=0xefcfb0, STO=0x0) at /home/chandlerc/code/compilers/llvm/lib/VMCore/SymbolTableListTraitsImpl.h:35 #6 0x00000000009aa8fe in ~Module (this=0xefcfb0) at /home/chandlerc/code/compilers/llvm/lib/VMCore/Module.cpp:77 #7 0x000000000069eaee in ~LTO (this=0xedb430) at /home/chandlerc/code/compilers/llvm/tools/lto/lto.cpp:471 #8 0x000000000069ab88 in llvm_destroy_optimizer (lto=0xedb430) at /home/chandlerc/code/compilers/llvm/tools/lto/lto_capi.cpp:35 #9 0x0000000000418ab5 in remove_llvm_lto () at ../../../binutils/ld/ldmain.c:195 #10 0x000000000069a6d9 in xatexit_cleanup () at ../../../binutils/libiberty/xatexit.c:98 #11 0x000000000069a701 in xexit (code=0x0) at ../../../binutils/libiberty/xexit.c:50 #12 0x000000000041926b in c_main (argc=0x18, argv=0x7fffa13e3e98) at ../../../binutils/ld/ldmain.c:601 #13 0x00002b5f09f87394 in __libc_start_main () from /lib/libc.so.6 #14 0x0000000000407a09 in _start () The only idea I have is that somewhere in the optimization passes, the list became corrupted, as when setParent was called the first time, the same loop succeeds to add the symbols _into_ the symbol table. Removing them causing an error doesn't make sense unless the list were corrupted in-between. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 07:55:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 07:55:14 -0600 Subject: [LLVMbugs] [Bug 1074] NEW: Stacker configure doesn't run automatically Message-ID: <200701041355.l04DtEst027666@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1074 Summary: Stacker configure doesn't run automatically Product: Build scripts Version: cvs Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: autoconf AssignedTo: unassignedbugs at nondot.org ReportedBy: jlaskey at apple.com Top level make does not automatically run the stacker configure script. It has to be run manually. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 10:35:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 10:35:07 -0600 Subject: [LLVMbugs] [Bug 1075] NEW: bottom-up register-reduction scheduling actually increases register pressure Message-ID: <200701041635.l04GZ73N030528@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1075 Summary: bottom-up register-reduction scheduling actually increases register pressure 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: djg at cray.com A high Sethi-Ullman number for an operation indicates it should be scheduled earlier rather than later -- from a top-down perspective. However, in bottom-up scheduling currently Seth-Ullman numbers are translated directly to priority values, so such operations end up being scheduled earlier from a *bottom-up* perspective. The result is that register pressure is actually increased instead of reduced. The effect is somewhat hidden because the special case values are also inverted -- for example, a store is given a low Sethi-Ullman number, but it ends up having the desired effect. In larger test cases, there are fewer nodes that use the special cases, and the effect is much more visible. I've experimented a little with having the priority function take the inverse of the SethiUllman value, and inverting the special-case values, and I've seen spill counts drop significantly in large test cases on x86-64. Here's a simple case that shows the problem: float %foo(float %x) { %tmp1 = mul float %x, 3.000000e+00 %tmp3 = mul float %x, 5.000000e+00 %tmp5 = mul float %x, 7.000000e+00 %tmp7 = mul float %x, 1.100000e+01 %tmp10 = add float %tmp1, %tmp3 %tmp12 = add float %tmp10, %tmp5 %tmp14 = add float %tmp12, %tmp7 ret float %tmp14 } With -march=x86-64, the list-burr (default) schedule places all the multiplies before all the adds. With the change mentioned above, the multiplies are intermixed with the adds, so the register pressure is lower. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 13:04:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 13:04:03 -0600 Subject: [LLVMbugs] [Bug 950] Signless Types For LLVM Message-ID: <200701041904.l04J43To001038@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=950 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-04 13:04 ------- This has been implemented. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 13:34:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 13:34:51 -0600 Subject: [LLVMbugs] [Bug 1076] NEW: llvm-upgrade changes names of global variables Message-ID: <200701041934.l04JYpec001554@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1076 Summary: llvm-upgrade changes names of global variables Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Currently, llvm-upgrade changes the names of global variables in an attempt to disambiguate type plane collapsing. For example: %ptr = external global int* ; [#uses=1] %dst = external global int ; [#uses=2] %src = external global int ; [#uses=1] gets translated to: %ptr.s = external global i32* %dst.s = external global i32 %src.s = external global i32 This is incorrect for external globals. The same applies for external functions. The names cannot be changed or linking errors will result. This bug requests llvm-upgrade to not rename global values of any external linkage type that do not have type plane collapsing conflicts. If there is a type plane collapse conflict, only one variable should be renamed and a warning should be emitted indicating the potential for linkage errors. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 4 17:02:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 4 Jan 2007 17:02:37 -0600 Subject: [LLVMbugs] [Bug 1077] NEW: Duplicate global vars allowed by llvm-as Message-ID: <200701042302.l04N2beu005640@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1077 Summary: Duplicate global vars allowed by llvm-as Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-as AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com The test/Feature/constexpr.ll test case is incorrect. It is specifically allowing this: %t4 = global int** cast (uint** %t3 to int**) %t4 = global int** cast (uint** %t3 to int**) This should produce an error as it is incorrect to allow redefinition of global vars, even if they have exactly the same type and initializer. Such redefinition needs to be made illegal in llvm-as. The test case needs to be modified to not test for allowing redefinition. A new test case should check that llvm-as generates an error on redefinition. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 10:49:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 10:49:52 -0600 Subject: [LLVMbugs] [Bug 1078] NEW: Move Stacker out of the LLVM repository to its own, like Java Message-ID: <200701051649.l05Gnqnm000616@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1078 Summary: Move Stacker out of the LLVM repository to its own, like Java Product: Projects Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: Stacker AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Stacker currently has a dependency on llvm-gcc. It is the last one that remains in LLVM. By moving Stacker to its own repository we can remove this last dependency and simplify building llvm and llvm-gcc for the 2.0 release. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 14:01:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 14:01:00 -0600 Subject: [LLVMbugs] [Bug 1079] NEW: Remove old pass manager from llvm sources Message-ID: <200701052001.l05K10Rs005262@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1079 Summary: Remove old pass manager from llvm sources Product: new-bugs Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: dpatel at apple.com Remove old pass manager from llvm sources. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 14:20:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 14:20:06 -0600 Subject: [LLVMbugs] [Bug 1079] Remove old pass manager from llvm sources Message-ID: <200701052020.l05KK6xZ005735@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1079 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 15:43:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 15:43:53 -0600 Subject: [LLVMbugs] [Bug 1076] llvm-upgrade changes names of global variables Message-ID: <200701052143.l05Lhr9A007307@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1076 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-05 15:43 ------- Fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070101/042063.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 15:45:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 15:45:36 -0600 Subject: [LLVMbugs] [Bug 1070] llvm-upgrade can cause redefinition errors Message-ID: <200701052145.l05LjaQa007358@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1070 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From rspencer at x10sys.com 2007-01-05 15:45 ------- This is a duplicate and now fixed. *** This bug has been marked as a duplicate of 1076 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 15:52:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 15:52:45 -0600 Subject: [LLVMbugs] [Bug 1077] Duplicate global vars allowed by llvm-as Message-ID: <200701052152.l05LqjKm007553@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1077 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-05 15:52 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070101/042099.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 17:46:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 17:46:55 -0600 Subject: [LLVMbugs] [Bug 1080] NEW: Invalid source produced during upgrade Message-ID: <200701052346.l05NktI1009832@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1080 Summary: Invalid source produced during upgrade Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru llvm-as rejected to compile converted source: $ ./llvm-as lpbench.ll ./llvm-as: lpbench.ll:573,0: syntax error, unexpected '(' lpbench.ll:573: while reading token: '(' ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 18:13:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 18:13:17 -0600 Subject: [LLVMbugs] [Bug 1080] Invalid source produced during upgrade Message-ID: <200701060013.l060DHQ5010366@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1080 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-05 18:13 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070101/042109.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 19:31:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 19:31:36 -0600 Subject: [LLVMbugs] [Bug 1081] NEW: Add a C API to the LLVM LTO library Message-ID: <200701060131.l061Va8j011758@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1081 Summary: Add a C API to the LLVM LTO library Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: chandlerc at gmail.com In trying to integrate LLVM into GNU ld, it became necessary to have a C API for the LTO. Rather than place C++, LLVM-specific code into the GNU linker, it seemed more appropriate to include a C interface within LLVM itself. The files and patches here will hopefully accomplish this in a smooth and unobtrusive manner. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 5 23:14:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 5 Jan 2007 23:14:53 -0600 Subject: [LLVMbugs] [Bug 1082] NEW: llvm-upgrade can't upgrade this Message-ID: <200701060514.l065ErjZ015705@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1082 Summary: llvm-upgrade can't upgrade this Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Unfortunately, the current algorithm in llvm-upgrade depends on each value being preceded by its type. However, with signedness elimination, the following is quite possible: %tmp7360 = bitcast ubyte %tmp7354 to sbyte %tmp7361 = sub ubyte 0, %tmp7360 llvm-upgrade current translates this to: %tmp7360.s = bitcast i8 %x.u to i8 %tmp7361.u = sub i8 0, %tmp7360.u %tmp7360.u in the sub instruction should be %tmp7360.s, the intended value. However, since the first argument to sub is ubyte, it gets the .u suffix. llvm-as rightly declares that %tmp7360.u is undefined in the i8 type plane. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 14:38:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 14:38:38 -0600 Subject: [LLVMbugs] [Bug 1083] NEW: Assertion failure in llvm-convert.cpp Message-ID: <200701062038.l06KccGl007584@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1083 Summary: Assertion failure in llvm-convert.cpp Product: tools Version: 1.9 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: babic.domagoj at gmail.com $:~/tmp$ llvm-gcc -emit-llvm -c svc.c -o svc.o cc1: ../../dst-directory/gcc/llvm-convert.cpp:216: void::FunctionPrologArgumentConversion::HandleScalarArgument(const llvm::Type*): Assertion `ArgVal->getType()->isIntegral() && LLVMTy->isIntegral() && "Lowerings don't match?"' failed. svc.c: In function ???svc_register???: svc.c:11: internal compiler error: Aborted Please submit a full bug report, Source: [svc.c from dietlib v0.30] ------------------------------------------------ #include static SVCXPRT **xports; bool_t svc_register(xprt, prog, vers, dispatch, protocol) SVCXPRT *xprt; unsigned long prog; unsigned long vers; void (*dispatch) (); rpcprot_t protocol; { return (TRUE); } ------------------------------------------------ Can be reproduced with: llvm-gcc [-emit-llvm] -c svc.c -o svc.o Domagoj ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 17:43:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 17:43:51 -0600 Subject: [LLVMbugs] [Bug 1083] K&R C compilation failure Message-ID: <200701062343.l06NhpNu011166@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1083 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-06 17:43 ------- Fixed. Testcase here: test/Regression/CFrontend/2007-01-06-KNR-Proto.c Patch sent to llvm-commits, but the mailing lists are apparently blocked, so no URL provided. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 17:54:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 17:54:32 -0600 Subject: [LLVMbugs] [Bug 1081] Add a C API to the LLVM LTO library Message-ID: <200701062354.l06NsWT6011511@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1081 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Linker |lto Keywords| |new-feature Product|libraries |tools Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-06 17:54 ------- Applied! I made a couple of minor changes (brace placement, wrap to 80 lines etc), but nothing functional. Nice work, -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 18:05:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 18:05:25 -0600 Subject: [LLVMbugs] [Bug 1084] Crash building mozilla: mishandling bogus trees for static ptrs to methods Message-ID: <200701070005.l0705PR0011768@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1084 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|Assertion failed |Crash building mozilla: | |mishandling bogus trees for | |static ptrs to methods Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-06 18:05 ------- Testcase here: test/Regression/C++Frontend/2007-01-06-PtrMethodInit.cpp Patch sent to llvm-commits. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 15:56:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 15:56:14 -0600 Subject: [LLVMbugs] [Bug 1085] NEW: Section printed into .ll file Message-ID: <200701062156.l06LuE0v008932@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1085 Summary: Section printed into .ll file Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached source (from Qt) will cause sections of form "gnu.linkonce.t" to be printed into resulting LLVM bytecode (e.g. when compiling with -emit-llvm -S). This causes misc linker problems. A quick workaround was introduced: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070101/042142.html It should be reverted after this bug fixed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 16:03:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 16:03:47 -0600 Subject: [LLVMbugs] [Bug 1086] NEW: CBE doesn't handle signlessness right Message-ID: <200701062203.l06M3lSY009193@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1086 Summary: CBE doesn't handle signlessness right Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider: int test(unsigned short X, short Y) { return X+Y; } $ llvm-gcc t.c -c -emit-llvm -o - -O3 | llc -march=c signed int test(signed short ltmp_0_1, signed short ltmp_1_1) { return ((((signed int )(signed short )ltmp_1_1)) + (((unsigned int )(unsigned short )ltmp_0_1))); } The first argument should be unsigned. This is a serious bug, because it means that bugpoint will get confused more often than it currently does. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 15:48:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 15:48:50 -0600 Subject: [LLVMbugs] [Bug 1084] NEW: Assertion failed Message-ID: <200701062148.l06LmoZe008758@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1084 Summary: Assertion failed Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached source (from mozilla, delta-reduced) causes the folowing assertion: $ /home/asl/proj/llvm/install/bin/llvm-g++ test.ii cc1plus: /home/asl/proj/llvm/src/lib/VMCore/Instructions.cpp:1685: static llvm::Instruction::CastOps llvm::CastInst::getCastOpcode(const llvm::Value*, bool, const llvm::Type*, bool): Assertion `!"Casting to type that is not first-class"' failed. test.ii:73: internal compiler error: Aborted Please submit a full bug report, with preprocessed source if appropriate. See for instructions. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 18:33:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 18:33:57 -0600 Subject: [LLVMbugs] [Bug 1085] Section printed into .ll file Message-ID: <200701070033.l070XvYR012372@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1085 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-06 18:33 ------- This is a ELF-specific bug in the C++ FE. Patch sent to llvm-commits, testcase here: Regression/C++Frontend/2007-01-06-ELF-Thunk-Sections.cpp -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 20:19:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 20:19:04 -0600 Subject: [LLVMbugs] [Bug 1087] NEW: Assertion failure in llvm-internal.h Message-ID: <200701070219.l072J4Gh014667@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1087 Summary: Assertion failure in llvm-internal.h Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: babic.domagoj at gmail.com $:~/tmp$ llvm-g++ -I/work/benchmarks/SOURCES/llvm-1.9/utils/TableGen -I/work/benchmarks/SOURCES/llvm-1.9/include -D__STDC_LIMIT_MACROS CodeEmitterGen.cpp -o /dev/null cc1plus: ../../dst-directory/gcc/llvm-internal.h:305: llvm::Value* TreeToLLVM::NOOPCastToType(llvm::Value*, const llvm::Type*): Assertion `isNoopCast(V, Ty) && "This is not a noop cast!"' failed. CodeEmitterGen.cpp: In member function ???int llvm::CodeEmitterGen::getVariableBit(const std::string&, llvm::BitsInit*, int)???: CodeEmitterGen.cpp:5: internal compiler error: Aborted Please submit a full bug report, Reduced testcase: -------------------------------------------------------------- #include "CodeEmitterGen.h" #include "Record.h" using namespace llvm; int CodeEmitterGen::getVariableBit(const std::string &VarName, BitsInit *BI, int bit) { if (VarBitInit *VBI = dynamic_cast(BI->getBit(bit))) { } } -------------------------------------------------------------- Reproducible with: llvm-g++ -I/llvm-1.9/utils/TableGen -I//llvm-1.9/include -D__STDC_LIMIT_MACROS CodeEmitterGen.cpp -o /dev/null My version of llvm-internal.h: -------------------------------------------------------------- Path: llvm-internal.h Name: llvm-internal.h URL: svn://anonsvn.opensource.apple.com/svn/llvm/trunk/gcc/llvm-internal.h Repository Root: svn://anonsvn.opensource.apple.com/svn/llvm Repository UUID: 51af6f4a-8115-0410-a7cd-c0d7a57f2c62 Revision: 224 Node Kind: file Schedule: normal Last Changed Author: jlaskey Last Changed Rev: 223 Last Changed Date: 2006-12-07 01:03:52 -0800 (Thu, 07 Dec 2006) Text Last Updated: 2006-12-09 13:16:41 -0800 (Sat, 09 Dec 2006) Properties Last Updated: 2006-12-09 13:16:18 -0800 (Sat, 09 Dec 2006) Checksum: b161d41397e225913a087e5b75ec1335 Domagoj ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 20:24:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 20:24:57 -0600 Subject: [LLVMbugs] [Bug 1015] -indvars miscompilation of Qt Message-ID: <200701070224.l072Ov8N014869@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1015 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Loop Optimizer |Global Analyses OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|-indvars miscompilation |-indvars miscompilation of | |Qt Target Milestone|--- |2.0 Version|trunk |1.5 ------- Additional Comments From sabre at nondot.org 2007-01-06 20:24 ------- Fixed. Testcase here: Transforms/IndVarsSimplify/2007-01-06-TripCount.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070101/042170.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 6 21:27:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 6 Jan 2007 21:27:12 -0600 Subject: [LLVMbugs] [Bug 1086] CBE doesn't handle signlessness right Message-ID: <200701070327.l073RCEX015981@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1086 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-06 21:27 ------- Fixed. No llvm-commits email so no patch :( Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 02:17:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 02:17:46 -0600 Subject: [LLVMbugs] [Bug 1092] NEW: Assertion failure in GlobalVariable.h Message-ID: <200701070817.l078Hk8T028923@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1092 Summary: Assertion failure in GlobalVariable.h Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: babic.domagoj at gmail.com $:~/tmp$ llvm-g++ -I/work/benchmarks/SOURCES/llvm-1.9/utils/TableGen -I/work/benchmarks/SOURCES/llvm-1.9/include -D__STDC_LIMIT_MACROS CodeEmitterGen2.ii -o /dev/null CodeEmitterGen.cpp:4: error: expected unqualified-id before ???namespace??? CodeEmitterGen.cpp:8: error: expected unqualified-id before string constant CodeEmitterGen.cpp:16: error: expected unqualified-id before ???namespace??? CodeEmitterGen.cpp:18: error: expected unqualified-id before ???namespace??? CodeEmitterGen.cpp:29: error: two or more data types in declaration of ???__jmp_buf??? CodeEmitterGen.cpp:29: error: duplicate ???typedef??? CodeEmitterGen.cpp:29: error: long, short, signed or unsigned invalid for ???__jmp_buf??? CodeEmitterGen.cpp:30: warning: ???pthread_once??? initialized and declared ???extern??? CodeEmitterGen.cpp:30: error: ???pthread_once_t??? was not declared in this scope CodeEmitterGen.cpp:30: error: ???__once_control??? was not declared in this scope CodeEmitterGen.cpp:30: error: ???__init_routine??? was not declared in this scope CodeEmitterGen.cpp:30: error: expected primary-expression before ???void??? CodeEmitterGen.cpp:30: error: initializer expression list treated as compound expression CodeEmitterGen.cpp:31: error: expected unqualified-id before ???public??? cc1plus: /work/projects/llvm/include/llvm/GlobalVariable.h:78: llvm::Constant* llvm::GlobalVariable::getInitializer() const: Assertion `hasInitializer() && "GV doesn't have initializer!"' failed. CodeEmitterGen.cpp:32: internal compiler error: Aborted Please submit a full bug report, ------------------------------------------------------ To reproduce run: llvm-g++ CodeEmitterGen2.ii -o /dev/null [Testcase attached in the following message.] Domagoj ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 01:56:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 01:56:45 -0600 Subject: [LLVMbugs] [Bug 1091] NEW: CBE generates new warnings since the signless work Message-ID: <200701070756.l077uj21020950@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1091 Summary: CBE generates new warnings since the signless work Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org f.e. on ldecod: Output/ldecod.cbe.c:71646: warning: this decimal constant is unsigned only in ISO C90 Output/ldecod.cbe.c:71657:50: warning: integer constant is so large that it is unsigned Output/ldecod.cbe.c:71657: warning: this decimal constant is unsigned only in ISO C90 Output/ldecod.cbe.c:71668:50: warning: integer constant is so large that it is unsigned Output/ldecod.cbe.c:71668: warning: this decimal constant is unsigned only in ISO C90 Output/ldecod.cbe.c:71688:50: warning: integer constant is so large that it is unsigned Output/ldecod.cbe.c:71688: warning: this decimal constant is unsigned only in ISO C90 Output/ldecod.cbe.c:73136:68: warning: integer constant is so large that it is unsigned This occurs on lines like: ltmp_19690_125__PHI_TEMPORARY = ((long long)-9223372036854775808ll) -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 01:32:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 01:32:12 -0600 Subject: [LLVMbugs] [Bug 1089] NEW: Assertion failure in llvm-internal.h Message-ID: <200701070732.l077WCDs020335@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1089 Summary: Assertion failure in llvm-internal.h Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: babic.domagoj at gmail.com $:~/tmp$ llvm-g++ -I/work/benchmarks/SOURCES/llvm-1.9/utils/TableGen -I/work/benchmarks/SOURCES/llvm-1.9/include -D__STDC_LIMIT_MACROS CodeEmitterGen.cpp -o /dev/null cc1plus: ../../dst-directory/gcc/llvm-internal.h:305: llvm::Value* TreeToLLVM::NOOPCastToType(llvm::Value*, const llvm::Type*): Assertion `isNoopCast(V, Ty) && "This is not a noop cast!"' failed. CodeEmitterGen.cpp: In member function ???int llvm::CodeEmitterGen::getVariableBit(const std::string&, llvm::BitsInit*, int)???: CodeEmitterGen.cpp:5: internal compiler error: Aborted Please submit a full bug report, Reduced testcase: -------------------------------------------------------------- #include "CodeEmitterGen.h" #include "Record.h" using namespace llvm; int CodeEmitterGen::getVariableBit(const std::string &VarName, BitsInit *BI, int bit) { if (VarBitInit *VBI = dynamic_cast(BI->getBit(bit))) { } } -------------------------------------------------------------- Reproducible with: llvm-g++ -I/llvm-1.9/utils/TableGen -I//llvm-1.9/include -D__STDC_LIMIT_MACROS CodeEmitterGen.cpp -o /dev/null My version of llvm-internal.h: -------------------------------------------------------------- Path: llvm-internal.h Name: llvm-internal.h URL: svn://anonsvn.opensource.apple.com/svn/llvm/trunk/gcc/llvm-internal.h Repository Root: svn://anonsvn.opensource.apple.com/svn/llvm Repository UUID: 51af6f4a-8115-0410-a7cd-c0d7a57f2c62 Revision: 224 Node Kind: file Schedule: normal Last Changed Author: jlaskey Last Changed Rev: 223 Last Changed Date: 2006-12-07 01:03:52 -0800 (Thu, 07 Dec 2006) Text Last Updated: 2006-12-09 13:16:41 -0800 (Sat, 09 Dec 2006) Properties Last Updated: 2006-12-09 13:16:18 -0800 (Sat, 09 Dec 2006) Checksum: b161d41397e225913a087e5b75ec1335 Domagoj ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 01:48:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 01:48:56 -0600 Subject: [LLVMbugs] [Bug 1090] NEW: CBE doesn't support exotic fcmp variants Message-ID: <200701070748.l077muL7020658@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1090 Summary: CBE doesn't support exotic fcmp variants Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org This is incorrect: case Instruction::FCmp: switch (CE->getPredicate()) { case FCmpInst::FCMP_ORD: case FCmpInst::FCMP_UEQ: case FCmpInst::FCMP_OEQ: Out << " == "; break; case FCmpInst::FCMP_UNO: case FCmpInst::FCMP_UNE: case FCmpInst::FCMP_ONE: Out << " != "; break; case FCmpInst::FCMP_OLT: case FCmpInst::FCMP_ULT: Out << " < "; break; case FCmpInst::FCMP_OLE: case FCmpInst::FCMP_ULE: Out << " <= "; break; case FCmpInst::FCMP_OGT: case FCmpInst::FCMP_UGT: Out << " > "; break; case FCmpInst::FCMP_OGE: case FCmpInst::FCMP_UGE: Out << " >= "; break; default: assert(0 && "Illegal FCmp predicate"); } -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 00:18:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 00:18:37 -0600 Subject: [LLVMbugs] [Bug 1088] NEW: getOrInsertFunction is unsafe and broken Message-ID: <200701070618.l076IbI4018941@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1088 Summary: getOrInsertFunction is unsafe and broken Product: tools Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: lli AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider this program: #include static void exit(int X) { printf("myexit called\n"); } int main() { exit(1); } This should print 'myexit called' once. However: $ llvm-gcc -c -emit-llvm t.c $ lli t.o myexit called myexit called ERROR: exit(-1073744072) returned! Abort This is due to the failures of the Module::getOrInsertFunction method. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 02:24:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 02:24:24 -0600 Subject: [LLVMbugs] [Bug 1088] Module::getOrInsertFunction is unsafe and broken Message-ID: <200701070824.l078OOhw030235@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1088 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Summary|getOrInsertFunction is |Module::getOrInsertFunction |unsafe and broken |is unsafe and broken Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-07 02:24 ------- getOrInsertFunction fixed. lli fixed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 02:30:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 02:30:57 -0600 Subject: [LLVMbugs] [Bug 1093] NEW: llvm.va_* intrinsics vary in type Message-ID: <200701070830.l078UvAa030352@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1093 Summary: llvm.va_* intrinsics vary in type Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The CFE currently produces llvm.va_* intrinsics where the va_list type varies according to the target. The LangRef document describes this behavior. Both should be updated to change the pointer to always be 'i8*' so that there is always a consistent type. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 12:26:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 12:26:20 -0600 Subject: [LLVMbugs] [Bug 1089] Assertion failure in llvm-internal.h Message-ID: <200701071826.l07IQKrw000830@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1089 babic.domagoj at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From babic.domagoj at gmail.com 2007-01-07 12:26 ------- I guess I hit "commit" twice... sorry. *** This bug has been marked as a duplicate of 1087 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 14:39:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 14:39:58 -0600 Subject: [LLVMbugs] [Bug 1044] LLVM Streams do not support stream manipulators Message-ID: <200701072039.l07KdwR4003273@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1044 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-01-07 14:39 ------- This should be solved by the new way streams are implemented. If it becomes a problem in the future, we can reopen this. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 16:15:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 16:15:56 -0600 Subject: [LLVMbugs] [Bug 1094] NEW: SetVector iterators aren't really mutable Message-ID: <200701072215.l07MFur1004881@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1094 Summary: SetVector iterators aren't really mutable 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: nicholas at mxc.ca Currently the implementation of SetVector::begin and end return iterators from the contained vector. If someone used that iterator to modify the interior of the vector (ala. "*MySetVector.begin() = value") then the SetVector's internal set would be out of sync making further inserts fail or permit the insertion of duplicates. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 16:45:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 16:45:03 -0600 Subject: [LLVMbugs] [Bug 1095] NEW: Miscompilation of asm("{cntlz|cntlzw} %0, %1" ...) on PowerPC Message-ID: <200701072245.l07Mj3o4005341@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1095 Summary: Miscompilation of asm("{cntlz|cntlzw} %0, %1" ...) on PowerPC Product: new-bugs Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: isanbard at gmail.com On my PowerBook G4, compiling this program: #include int main() { unsigned long int ctz_x = 0x0000f000; unsigned long int ctz_c; __asm__("{cntlz|cntlzw} %0,%1" : "=r" (ctz_c) : "r" (ctz_x & -ctz_x)); printf("ctz_c == %d\n", ctz_c); return 0; } results in the use of the mnemonic "cntlz" instead of "cntlzw". This produces an error when trying to assemble the program: cntlzw_miscompile.s:25:Invalid mnemonic 'cntlz' GCC outputs "cntlzw" instead. -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 19:39:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 19:39:02 -0600 Subject: [LLVMbugs] [Bug 1092] Assertion failure in GlobalVariable.h Message-ID: <200701080139.l081d2pr008158@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1092 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-01-07 19:39 ------- Your testcase certainly is not legal C++ code. Perhaps it was misreduced. The start of the file is: typedef union { struct { } namespace __gnu_cxx { -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 19:41:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 19:41:52 -0600 Subject: [LLVMbugs] [Bug 1087] Assertion failure in llvm-internal.h Message-ID: <200701080141.l081fqdV008223@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1087 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From sabre at nondot.org 2007-01-07 19:41 ------- This works fine for me with LLVM CVS head and llvm-gcc svn head. It is a likely dup of Bug 1084. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 19:44:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 19:44:39 -0600 Subject: [LLVMbugs] [Bug 1035] stage1 compiler miscompiles the stage2 on amd64 Message-ID: <200701080144.l081id5v008312@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1035 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From sabre at nondot.org 2007-01-07 19:44 ------- Ok, I will close this. If the problem still exists, we can file a new bug. Since this bug was filed, many fixes have been made. Thanks! ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 7 23:06:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 7 Jan 2007 23:06:51 -0600 Subject: [LLVMbugs] [Bug 1085] Section printed into .ll file Message-ID: <200701080506.l0856pMl011731@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1085 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|VERIFIED |REOPENED Resolution|FIXED | ------- Additional Comments From rspencer at x10sys.com 2007-01-07 23:06 ------- This is not fixed. The Regression/C++Frontend/2007-01-06-ELF-Thunk-Sections.cpp test case fails on Linux at least. The "not grep" part of it is failing. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 00:28:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 00:28:27 -0600 Subject: [LLVMbugs] [Bug 1055] LLVMlto library does not build on Linux (amd64) Message-ID: <200701080628.l086SRME013087@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1055 chandlerc at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From chandlerc at gmail.com 2007-01-08 00:28 ------- Committed patch. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 01:00:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 01:00:53 -0600 Subject: [LLVMbugs] [Bug 1090] CBE doesn't support exotic fcmp variants Message-ID: <200701080700.l0870rnG013615@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1090 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED OS/Version|MacOS X |All Platform|Macintosh |All Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-08 01:00 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042222.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 01:29:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 01:29:57 -0600 Subject: [LLVMbugs] [Bug 1096] NEW: interpreter regression with signless patch Message-ID: <200701080729.l087TvB2014069@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1096 Summary: interpreter regression with signless patch Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Interpreter AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org $ cd llvm/test/Regression/ExecutionEngine $ llvm-upgrade < test-cast.ll | llvm-as | lli -force-interpreter Unhandled type for ICMP_NE predicate: bool Abort ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 02:00:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 02:00:59 -0600 Subject: [LLVMbugs] [Bug 1091] CBE generates new warnings since the signless work Message-ID: <200701080800.l0880xJX014800@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1091 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED OS/Version|MacOS X |All Platform|Macintosh |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-08 02:00 ------- Warnings are gone with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042227.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 10:04:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 10:04:12 -0600 Subject: [LLVMbugs] [Bug 1097] NEW: Strength reduction misses IV reuse opportunities on 64-bit targets Message-ID: <200701081604.l08G4ChO031802@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1097 Summary: Strength reduction misses IV reuse opportunities on 64- bit targets Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Loop Optimizer AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com The strength reduction pass is causing the code generator to miss opportunities for using complex addressing modes. For this test case: define void %foo(double* %y) { entry: br label %bb bb: %i = phi i64 [ 0, %entry ], [ %k, %bb ] %j = getelementptr double* %y, i64 %i store double 0.000000e+00, double* %j %k = add i64 %i, 1 %n = icmp eq i64 %k, 0 br bool %n, label %return, label %bb return: ret void } The strength reduction pass changes the output on x86-64 to have two induction variables: %iv..inc = add i64 %iv., 1 ... %iv.1.inc = add i64 %iv.1, 8 which causes the codegen to produce this: .LBB1_1: #bb movq $0, (%rdi) addq $8, %rdi incq %rax cmpq $0, %rax jne .LBB1_1 #bb instead of this: .LBB1_1: #bb movq $0, (%rdi,%rax,8) incq %rax cmpq $0, %rax jne .LBB1_1 #bb Here's a patch that fixes this. Index: LoopStrengthReduce.cpp =================================================================== RCS file: /var/cvs/llvm/llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp,v retrieving revision 1.104 diff -u -r1.104 LoopStrengthReduce.cpp --- LoopStrengthReduce.cpp +++ LoopStrengthReduce.cpp @@ -893,7 +893,7 @@ if (unsigned(abs(SInt)) < Scale || (SInt % Scale) != 0) continue; std::map::iterator SI = - IVsByStride.find(SCEVUnknown::getIntegerSCEV(SInt/Scale, Type::Int32Ty)); + IVsByStride.find(SCEVUnknown::getIntegerSCEV(SInt/Scale, UIntPtrTy)); if (SI == IVsByStride.end()) continue; for (std::vector::iterator II = SI->second.IVs.begin(), ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 10:20:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 10:20:39 -0600 Subject: [LLVMbugs] [Bug 1097] Strength reduction misses IV reuse opportunities on 64-bit targets Message-ID: <200701081620.l08GKdmi032215@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1097 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-08 10:20 ------- Patch applied: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042228.html No apparent regressions on 32-bit (didn't expect any :) Thanks, Dan! ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 10:44:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 10:44:04 -0600 Subject: [LLVMbugs] [Bug 1098] NEW: Incorrect post-dominators with multiple returns Message-ID: <200701081644.l08Gi4en000423@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1098 Summary: Incorrect post-dominators with multiple returns Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Global Analyses AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com In some test cases with multiple return statements, the post-dominator passes return incorrect information. For this test case: define void %foo(bool %x) { entry: br bool %x, label %bb1, label %bb0 bb0: ret void bb1: br label %bb2 bb2: ret void } The `entry' block's immediate post-dominator should be the exit node, but the immediate post-dominator pass's print() output is Immediate Dominator For Basic Block: %entry is: %bb0 Immediate Dominator For Basic Block: %bb0 is: <> Immediate Dominator For Basic Block: %bb1 is: %bb2 Immediate Dominator For Basic Block: %bb2 is: <> The post-dominator tree and post-dominator set passes exhibit similar behavior. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 15:01:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 15:01:36 -0600 Subject: [LLVMbugs] [Bug 1085] Section printed into .ll file Message-ID: <200701082101.l08L1aiP005854@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1085 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-08 15:01 ------- This fixes the fix. Reid, plz verify and remove the xfail from the testcase if ok now. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042246.html Thanks, -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 15:29:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 15:29:38 -0600 Subject: [LLVMbugs] [Bug 1069] Incorrect Post-Dominance Frontier Calculation Message-ID: <200701082129.l08LTcCq006570@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1069 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-01-08 15:29 ------- *** This bug has been marked as a duplicate of 1098 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 17:56:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 17:56:27 -0600 Subject: [LLVMbugs] [Bug 1075] bottom-up register-reduction scheduling actually increases register pressure Message-ID: <200701082356.l08NuRvX009853@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1075 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-01-08 17:56 ------- Fixed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 8 23:51:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 8 Jan 2007 23:51:46 -0600 Subject: [LLVMbugs] [Bug 1099] NEW: CBE miscompilation of 197.parser Message-ID: <200701090551.l095pkwM016159@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1099 Summary: CBE miscompilation of 197.parser Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Reduced by bugpoint. The issue is that this testcase: target datalayout = "e-p:32:32" target endian = little target pointersize = 32 target triple = "i686-apple-darwin8" %struct.Connector = type { i16, i16, i8, i8, %struct.Connector*, i8* } implementation ; Functions: define bool %prune_match_entry_2E_ce(%struct.Connector* %a, i16 %b.0.0.val) { newFuncRoot: br label %entry.ce cond_next.exitStub: ; preds = %entry.ce ret bool true entry.return_crit_edge.exitStub: ; preds = %entry.ce ret bool false entry.ce: ; preds = %newFuncRoot %tmp = getelementptr %struct.Connector* %a, i32 0, i32 0 ; [#uses=1] %tmp = load i16* %tmp ; [#uses=1] %tmp = icmp eq i16 %tmp, %b.0.0.val ; [#uses=1] br bool %tmp, label %cond_next.exitStub, label %entry.return_crit_edge.exitStub } is compiled to: struct l_struct_2E_Connector { signed short field0; signed short field1; signed char field2; signed char field3; struct l_struct_2E_Connector *field4; signed char *field5; }; ... bool prune_match_entry_2E_ce(struct l_struct_2E_Connector *ltmp_0_1, unsigned short ltmp_1_2) { signed short ltmp_2_2; ltmp_2_2 = *(<mp_0_1->field0); return ((((ltmp_2_2 == ltmp_1_2)) ? (1) : (0))); } Note that the == is comparing a signed short and an unsigned short. These both get extended (differently) to int before the comparison is done. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 9 00:51:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 9 Jan 2007 00:51:46 -0600 Subject: [LLVMbugs] [Bug 1073] When destructing a module in the LTO, ilist assertion fails. Message-ID: <200701090651.l096pk5a017322@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1073 chandlerc at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From chandlerc at gmail.com 2007-01-09 00:51 ------- Devang did commit the patch, just closing the bug. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 9 04:57:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 9 Jan 2007 04:57:17 -0600 Subject: [LLVMbugs] [Bug 1100] NEW: Could not resolve external global address: __dso_handle Message-ID: <200701091057.l09AvHvR030492@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1100 Summary: Could not resolve external global address: __dso_handle Product: libraries Version: 1.9 Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: hp48gx at gmail.com I just compiled llvm1.9 using pre-packaged llvm-gcc4 (I simply fixed the symbolic links in /bin). llvm-gcc works fine on C files, but does not work on c++ Native output is fine, but bytecode does not run, for example powerbookg4:/llvm/test mycroft$ cat hello.cpp #include int main() { std::cout << "hello world\n"; return 0; } powerbookg4:/llvm/test mycroft$ llvm-g++ hello.cpp -emit-llvm -c -o hello.bc powerbookg4:/llvm/test mycroft$ lli hello.bc Could not resolve external global address: __dso_handle Abort trap The bytecode itself seems fine (llvm-dis produces a nice output) I remember I already had this problem once... I vaguely remember llvm-g++ is linking against the wrong libraries, but I don't remember how I fixed it... Can anybody help? ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 9 11:11:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 9 Jan 2007 11:11:42 -0600 Subject: [LLVMbugs] [Bug 1099] CBE miscompilation of 197.parser Message-ID: <200701091711.l09HBgOq004402@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1099 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-09 11:11 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042272.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 9 18:28:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 9 Jan 2007 18:28:05 -0600 Subject: [LLVMbugs] [Bug 1101] NEW: ScalarEvolution can't find iteration count without tail duplication Message-ID: <200701100028.l0A0S5b9028163@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1101 Summary: ScalarEvolution can't find iteration count without tail duplication 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: wjiang at cs.ucla.edu CC: rspencer at x10sys.com Hi, the TOT llvm can't find the iteration count without tail duplication, I guess it's caused by the following code in ScalaEvoluation.cpp: // Currently we check for this by checking to see if the Exit branch goes to // the loop header. If so, we know it will always execute the same number of // times as the loop. More extensive analysis could be done to handle more // cases here. if (ExitBr->getSuccessor(0) != L->getHeader() && ExitBr->getSuccessor(1) != L->getHeader()) return UnknownValue; I need to disable tail duplication because it changes a while loop to a do- while loop by moving the exiting branch to the end of the loop body. I attached two files: (1) foo.bc : a simple loop without tail duplication (2) foo.taildup.bc: a simple loop with tail duplication. I ran "opt -indvar" on these two files, llvm failed to find trip count on the first example; but if I comment those three lines, it works fine and report the correct iteration count. Based on my understanding, those checking are not necessary since scalarevoluation works normally even the checking fails. Thanks. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 07:51:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 07:51:48 -0600 Subject: [LLVMbugs] [Bug 1092] Assertion failure in GlobalVariable.h Message-ID: <200701101351.l0ADpm1v031188@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1092 nicholas at mxc.ca changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Keywords| |crash-on-invalid Resolution|INVALID | ------- Additional Comments From nicholas at mxc.ca 2007-01-10 07:51 ------- Reopening because it's an ICE. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 08:02:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 08:02:59 -0600 Subject: [LLVMbugs] [Bug 1102] NEW: Should support different visibility levels. Message-ID: <200701101402.l0AE2xZQ031454@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1102 Summary: Should support different visibility levels. 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 We should handle __attribute__(__visibility__("hidden"))) and so on... ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 11:47:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 11:47:04 -0600 Subject: [LLVMbugs] [Bug 1103] NEW: %rsp permitted as an index register Message-ID: <200701101747.l0AHl4Mg008578@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1103 Summary: %rsp permitted as an index register Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com The x86 backend allows %rsp to be used as an index register in an address operand. Some assemblers accept this and automatically rewrite it so that %rsp is the base register to make it legal, but some don't, including GAS. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 15:17:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 15:17:43 -0600 Subject: [LLVMbugs] [Bug 1104] NEW: Bad struct argument "packing/conversion" in llvm-gcc4 Message-ID: <200701102117.l0ALHhYb012421@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1104 Summary: Bad struct argument "packing/conversion" in llvm-gcc4 Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: zhiruz at gmail.com When I was playing with the new (TOT) llvm-gcc4 and LLVM opt, I found a pretty serious problem due to a weird struct argument transformation by llvm-gcc4, which could potentially lead to code size blow-up. I have a simple example here. Basically, the C code decribes an addition routine for two complex numbers. //===============complex.c================// typedef struct { short r; short i; } complex; complex add_complex(complex a, complex b) { complex c; c.r = a.r + b.r; c.i = a.i + b.i; return c; } After "llvm-makegcc -O3", the struct parameters "a" and "b" are converted into two "int32" scalars. Since two 16-bit fields are packed together, many extra truncates and shifts are introduced. //========complex.ll (by llvm-makegcc -O3) ================// define csretcc void %add_complex(%struct.complex* %agg.result, i32 %a.0, i32 %b.0) { entry: %tmp30 = trunc i32 %a.0 to i16 ; [#uses=1] %tmp427 = trunc i32 %b.0 to i16 ; [#uses=1] %tmp5 = add i16 %tmp427, %tmp30 ; [#uses=1] %tmp828 = lshr i32 %a.0, i8 16 ; [#uses=1] %tmp829 = trunc i32 %tmp828 to i16 ; [#uses=1] %tmp1025 = lshr i32 %b.0, i8 16 ; [#uses=1] %tmp1026 = trunc i32 %tmp1025 to i16 ; [#uses=1] %tmp11 = add i16 %tmp1026, %tmp829 ; [#uses=1] %tmp13 = getelementptr %struct.complex* %agg.result, i32 0, i32 0 ; [#uses=1] store i16 %tmp5, i16* %tmp13 %tmp16 = getelementptr %struct.complex* %agg.result, i32 0, i32 1 ; [#uses=1] store i16 %tmp11, i16* %tmp16 ret void } I was confused as I thought ScalarReplAggregates pass should convert it to two i16 scalars. To confirm this is not LLVM opt's problem, I tried "llvm-makegcc -O0" option and saw the same conversion for the structure arguments. So it looks like a llvm-gcc4 bug. To me, this conversion does not make much sense. If the purpose is to do scalar promotion, why not split the "complex" struct into two "int16"s? With the hope to get a little bit more insights, I tested more cases and got the following not-so-encouraging results: * struct {int, int, int} ==> int64, int32 (understandable, pack to the same size, but introduces unneccessary truncates and shifts) * struct {double, double} ==> int64, int64 (bad conversion, requires a lot of redundant bitcast/load/store) * struct {double, int a[1000]} ==> int64, plus FIVE HUNDRED separate int64 (WOW, I am stunned ...) The C code and the output .bc are attached. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 15:32:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 15:32:25 -0600 Subject: [LLVMbugs] [Bug 1105] NEW: Provide options in LTO to build a PIC object file for linking into a shared library. Message-ID: <200701102132.l0ALWP5P012735@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1105 Summary: Provide options in LTO to build a PIC object file for linking into a shared library. Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: lto AssignedTo: chandlerc at gmail.com ReportedBy: chandlerc at gmail.com One of the desirable things for a linker integrated with LLVM's LTO library to provide is linking of shared libraries from bytecode. This requires (on some platforms/OSes) building PIC in the object file with a special flag in the assembly step of the LTO. Once PIC can be requested from the LTO library, GNU ld (for example) should be able to check when it is generating a shared library, and pass along the correct options. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 17:47:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 17:47:14 -0600 Subject: [LLVMbugs] [Bug 1106] NEW: [llvm-gcc] "--enable-llvm" only works with a build-location, not an installed version. Message-ID: <200701102347.l0ANlEg8015037@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1106 Summary: [llvm-gcc] "--enable-llvm" only works with a build- location, not an installed version. Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: llvm-gcc AssignedTo: chandlerc at gmail.com ReportedBy: chandlerc at gmail.com When installing LLVM and the LLVM-GCC C/C++ Front End to a system, the build directories of LLVM are often not available at the time of installing the front end. However, the '--enable-llvm' flag in the front end configure scripts does not recognize a prefix for an installed set of LLVM libraries, or default to the install prefix that it is itself configured with, or the existing system path to locate suitable LLVM libraries. There are 3 goals, tied to the same section of code: 1) Recognize the directory passed to '--enable-llvm' as either a build directory, or an install prefix for LLVM. (appending /bin or /lib as necessary) 2) If no directory is specified assume the configured prefix, and search as in 1. 3) (Optional) If still no installation found, search for llvm-config in PATH, and use it to determine the location of all libraries. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 21:15:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 21:15:18 -0600 Subject: [LLVMbugs] [Bug 763] [llvmgcc4] needs transparent IPO support Message-ID: <200701110315.l0B3FIVG018397@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=763 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-10 21:15 ------- Devang implemented this with liblto and integrated support into the Apple linker. Chandler is working on integration with binutils. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 10 23:09:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 10 Jan 2007 23:09:18 -0600 Subject: [LLVMbugs] [Bug 1007] predsimplify rips out constant, replaces it with variable. Message-ID: <200701110509.l0B59IGx020505@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1007 nicholas at mxc.ca changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From nicholas at mxc.ca 2007-01-10 23:09 ------- Fixed in rewrite. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 12 00:55:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 12 Jan 2007 00:55:24 -0600 Subject: [LLVMbugs] [Bug 1107] NEW: -instcombine failure for MiBench/consumer-typeset Message-ID: <200701120655.l0C6tOkf021801@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1107 Summary: -instcombine failure for MiBench/consumer-typeset Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com I have been trying to solve a little mystery. After resolving various issues with the test case itsellf (MultisSource/Benchmarks/MiBench/consumer-typeset), I have determined that there is a codegen bug that this test triggers when -instcombine is run. Here's how I know this: 1. If you run the Output/consumer-typeset.linked.rbc file through lli in JIT mode it works fine. 2. All three backends (llc,jit,cbe) fail in the same way. They all produce the same incorrect output. This indicates a misoptimization rather than a code gen bug. 3. I used findmisopt to find the first optimization that caused the output to differ. It reported this sequence of optimizations as the first set that produces a difference in the output: -lowersetjmp -funcresolve -raiseallocs -simplifycfg -mem2reg -globalopt -globaldce -ipconstprop -deadargelim -instcombine 4. bugpoint produces a non-sensical reduction (two branches and a return) I'm wondernig if someone can bugpoint this on Darwin because I'm starting to think that bugpoint doesn't work so well on Linux. In the last month, a Darwin run of bugpoint was able to reduce a test case that a run on Linux, with the same inputs, could not reduce. If you can reduce this on Darwin, I'll file a bug against bugpoint. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 12 12:32:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 12 Jan 2007 12:32:06 -0600 Subject: [LLVMbugs] [Bug 1096] interpreter regression with signless patch Message-ID: <200701121832.l0CIW6WO011076@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1096 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Keywords| |compile-fail Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-12 12:32 ------- This was fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042516.html amongst several other interpreter bugs. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 12 12:41:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 12 Jan 2007 12:41:12 -0600 Subject: [LLVMbugs] [Bug 1108] NEW: CodeGen/Generic/vector-identity-shuffle.ll regression fails on x86_64-pc-linux-gnu Message-ID: <200701121841.l0CIfCu2011420@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1108 Summary: CodeGen/Generic/vector-identity-shuffle.ll regression fails on x86_64-pc-linux-gnu Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: chandlerc at gmail.com On x86_64-pc-linux-gnu the following regression test is failing: FAIL: /home/chandlerc/code/compilers/llvm/test/Regression/CodeGen/Generic/vector-identity-shuffle.ll: child process exited abnormally test: llc: /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:1143: llvm::SDOperand llvm::SelectionDAG::getNode(unsigned int, llvm::MVT::ValueType, llvm::SDOperand): Assertion `MVT::isVector(VT) && !MVT::isVector(Operand.getValueType()) && MVT::getVectorBaseType(VT) == Operand.getValueType() && "Illegal SCALAR_TO_VECTOR node!"' failed. When run by hand in GDB, I get the following backtrace: gdb> set args vector-identity-shuffle.bc gdb> run [Thread debugging using libthread_db enabled] [New Thread 47416736631584 (LWP 7527)] llc: /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:1143: llvm::SDOperand llvm::SelectionDAG::getNode(unsigned int, llvm::MVT::ValueType, llvm::SDOperand): Assertion `MVT::isVector(VT) && !MVT::isVector(Operand.getValueType()) && MVT::getVectorBaseType(VT) == Operand.getValueType() && "Illegal SCALAR_TO_VECTOR node!"' failed. Program received signal SIGABRT, Aborted. [Switching to Thread 47416736631584 (LWP 7527)] _______________________________________________________________________________ Error while running hook_stop: Value can't be converted to integer. 0x00002b20119d97b5 in raise () from /lib/libc.so.6 gdb> bt #0 0x00002b20119d97b5 in raise () from /lib/libc.so.6 #1 0x00002b20119daa4e in abort () from /lib/libc.so.6 #2 0x00002b20119d3156 in __assert_fail () from /lib/libc.so.6 #3 0x0000000000a195e1 in llvm::SelectionDAG::getNode (this=0x7fff999a3f90, Opcode=0x47, VT=llvm::MVT::v4i32, Operand={Val = 0x104cda0, ResNo = 0x0}) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:1141 #4 0x00000000008ce63e in llvm::X86TargetLowering::LowerBUILD_VECTOR (this=0x1046400, Op={Val = 0x104e380, ResNo = 0x0}, DAG=@0x7fff999a3f90) at /home/chandlerc/code/compilers/llvm/lib/Target/X86/X86ISelLowering.cpp:3407 #5 0x00000000008e2916 in llvm::X86TargetLowering::LowerOperation (this=0x1046400, Op={Val = 0x104e380, ResNo = 0x0}, DAG=@0x7fff999a3f90) at /home/chandlerc/code/compilers/llvm/lib/Target/X86/X86ISelLowering.cpp:5020 #6 0x0000000000a6aca7 in (anonymous namespace)::SelectionDAGLegalize::LegalizeOp (this=0x7fff999a3c80, Op={Val = 0x104e380, ResNo = 0x0}) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:901 #7 0x0000000000a83890 in (anonymous namespace)::SelectionDAGLegalize::PackVectorOp (this=0x7fff999a3c80, Op={Val = 0x104dee0, ResNo = 0x0}, NewVT=llvm::MVT::v4i32) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:5523 #8 0x0000000000a971e4 in (anonymous namespace)::SelectionDAGLegalize::HandleOp (this=0x7fff999a3c80, Op={Val = 0x104dee0, ResNo = 0x0}) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:476 #9 0x0000000000a97708 in (anonymous namespace)::SelectionDAGLegalize::LegalizeDAG (this=0x7fff999a3c80) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:340 #10 0x0000000000a9788d in llvm::SelectionDAG::Legalize (this=0x7fff999a3f90) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:5537 #11 0x0000000000a28a87 in llvm::SelectionDAGISel::CodeGenAndEmitDAG (this=0x1046350, DAG=@0x7fff999a3f90) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:4110 #12 0x0000000000a3c8ee in llvm::SelectionDAGISel::SelectBasicBlock (this=0x1046350, LLVMBB=0x1040880, MF=@0x104c250, FuncInfo=@0x7fff999a4210) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:4140 #13 0x0000000000a3d699 in llvm::SelectionDAGISel::runOnFunction (this=0x1046350, Fn=@0x103c230) at /home/chandlerc/code/compilers/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:3842 #14 0x00000000008f320b in (anonymous namespace)::X86DAGToDAGISel::runOnFunction (this=0x1046350, Fn=@0x103c230) at /home/chandlerc/code/compilers/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp:121 #15 0x0000000000c2d106 in llvm::FPPassManager::runOnFunction (this=0x1045420, F=@0x103c230) at /home/chandlerc/code/compilers/llvm/lib/VMCore/PassManager.cpp:1252 #16 0x0000000000c2d338 in llvm::FunctionPassManagerImpl::run (this=0x1044c60, F=@0x103c230) at /home/chandlerc/code/compilers/llvm/lib/VMCore/PassManager.cpp:1218 #17 0x0000000000c2d3fb in llvm::FunctionPassManager::run (this=0x7fff999a45d0, F=@0x103c230) at /home/chandlerc/code/compilers/llvm/lib/VMCore/PassManager.cpp:1163 #18 0x00000000007d13dc in main (argc=0x2, argv=0x7fff999a47b8) at /home/chandlerc/code/compilers/llvm/tools/llc/llc.cpp:264 Is there any other information I can get that would help? Currently its the only (non-debuginfo) test failing. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 12 16:26:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 12 Jan 2007 16:26:07 -0600 Subject: [LLVMbugs] [Bug 1102] Should support different visibility levels. Message-ID: <200701122226.l0CMQ7Y9015641@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1102 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-01-12 16:26 ------- Fixed with these series of patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042551.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042548.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042549.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042550.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042553.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042552.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042557.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042556.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042555.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042554.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042558.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 12 18:24:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 12 Jan 2007 18:24:32 -0600 Subject: [LLVMbugs] [Bug 1093] llvm.va_* intrinsics vary in type Message-ID: <200701130024.l0D0OWoi018382@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1093 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-12 18:24 ------- Fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042600.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042601.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042608.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 13 18:14:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 13 Jan 2007 18:14:27 -0600 Subject: [LLVMbugs] [Bug 1103] %rsp emitted as an index register Message-ID: <200701140014.l0E0ERwi018071@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1103 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Summary|%rsp permitted as an index |%rsp emitted as an index |register |register Target Milestone|--- |2.0 Version|trunk |1.9 ------- Additional Comments From sabre at nondot.org 2007-01-13 18:14 ------- The most straight-forward way to fix this is in the asmprinter. Addressed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042639.html Testcase here: Regression/CodeGen/X86/2007-01-13-StackPtrIndex.ll -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 13 18:56:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 13 Jan 2007 18:56:46 -0600 Subject: [LLVMbugs] [Bug 1107] -instcombine miscompilation of MiBench/consumer-typeset Message-ID: <200701140056.l0E0ukf6018974@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1107 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-13 18:56 ------- Here is one instcombine bugfix: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042635.html Testcase here: Transforms/InstCombine/2007-01-13-ExtCompareMiscompile.ll The next bug isn't in instcombine. Closing this bug to keep things simple. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 13 19:25:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 13 Jan 2007 19:25:45 -0600 Subject: [LLVMbugs] [Bug 1101] ScalarEvolution can't find iteration count without tail duplication Message-ID: <200701140125.l0E1PjY8019598@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1101 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 Version|1.0 |trunk ------- Additional Comments From sabre at nondot.org 2007-01-13 19:25 ------- I figured out my own testcase. Testcase here: Analysis/ScalarEvolution/trip-count.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042645.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 13 23:46:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 13 Jan 2007 23:46:04 -0600 Subject: [LLVMbugs] [Bug 1109] NEW: basicaa returns noalias for aliasing pointers, miscompiles MiBench/consumer-typeset Message-ID: <200701140546.l0E5k45s023615@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1109 Summary: basicaa returns noalias for aliasing pointers, miscompiles MiBench/consumer-typeset Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Global Analyses AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org In the attached heavily reduced testcase, basicaa returns noaa for two pointers that overlap. This causes miscompilation of consumer-typeset. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 13 23:58:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 13 Jan 2007 23:58:49 -0600 Subject: [LLVMbugs] [Bug 1109] basicaa returns noalias for aliasing pointers, miscompiles MiBench/consumer-typeset Message-ID: <200701140558.l0E5wn01023939@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1109 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 Version|trunk |1.0 ------- Additional Comments From sabre at nondot.org 2007-01-13 23:58 ------- Fixed. Testcase here: Analysis/BasicAA/2007-01-13-BasePointerBadNoAlias.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042647.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 08:32:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 08:32:33 -0600 Subject: [LLVMbugs] [Bug 1110] NEW: crash in domtree after break-crit-edges Message-ID: <200701141432.l0EEWXAL001812@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1110 Summary: crash in domtree after break-crit-edges 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: nicholas at mxc.ca Steps to reproduce: $ llvm/Debug/bin/opt -etforest -break-crit-edges -domtree x.bc -disable-output Segmentation fault The backtrace shows an infinite loop #123147 0x085b20b7 in llvm::DominatorTree::getNodeForBlock (this=0x87191f8, BB=0x0) at Dominators.cpp:380 but before that, it seems that llvm::DominatorTree::calculate's "ImmDom", which is gets from ImmediateDominator is out of date: #174577 0x085b225f in llvm::DominatorTree::calculate (this=0x87191f8, ID=@0x871a1a8) at Dominators.cpp:399 399 Node *IDomNode = getNodeForBlock(ImmDom); (gdb) p ImmDom $2 = (class llvm::BasicBlock *) 0x8719838 (gdb) p ImmDom->getName() Cannot access memory at address 0xff75effc (gdb) down #174576 0x085b20b7 in llvm::DominatorTree::getNodeForBlock (this=0x87191f8, BB=0x8719838) at Dominators.cpp:380 380 Node *IDomNode = getNodeForBlock(IDom); (gdb) p IDom $3 = (class llvm::BasicBlock *) 0x0 and that's the start of the infinite loop. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 09:43:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 09:43:14 -0600 Subject: [LLVMbugs] [Bug 788] Instructions don't point out that Makefile is needed in object tree as well as source tree Message-ID: <200701141543.l0EFhEYG002973@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=788 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From rspencer at x10sys.com 2007-01-14 09:43 ------- The makefiles are copies and configured correctly. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 12:35:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 12:35:01 -0600 Subject: [LLVMbugs] [Bug 1110] break-crit-edges incorrectly updates idoms for unreachable code Message-ID: <200701141835.l0EIZ1nO006248@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1110 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Core LLVM classes |Transformation Utilities Keywords| |compile-fail OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|crash in domtree after |break-crit-edges incorrectly |break-crit-edges |updates idoms for | |unreachable code Target Milestone|--- |2.0 Version|trunk |1.0 ------- Additional Comments From sabre at nondot.org 2007-01-14 12:35 ------- Fixed, testcase here: Analysis/Dominators/2007-01-14-BreakCritEdges.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042657.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 12:54:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 12:54:42 -0600 Subject: [LLVMbugs] [Bug 1100] Could not resolve external global address: __dso_handle Message-ID: <200701141854.l0EIsgZ1006620@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1100 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-14 12:54 ------- This should be fixed in mainline. __dso_handle is required to have hidden visibility, which Anton just recently implemented. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:02:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:02:20 -0600 Subject: [LLVMbugs] [Bug 1026] InstVisit duplicates instruction visitors Message-ID: <200701141902.l0EJ2K53006868@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1026 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From sabre at nondot.org 2007-01-14 13:02 ------- as above, not a problem ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:03:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:03:55 -0600 Subject: [LLVMbugs] [Bug 989] gcc4 front-end, darwin-crt2.o Message-ID: <200701141903.l0EJ3tNq006928@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=989 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-14 13:03 ------- This has almost certainly been fixed. If you run into this in the future, please reopen. Thanks, -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:09:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:09:02 -0600 Subject: [LLVMbugs] [Bug 1067] PIC codegen for Linux/X86 Message-ID: <200701141909.l0EJ92ne007081@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1067 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-01-14 13:09 ------- Well, actually I have Qt crashed here (as I've described). Probably, this is due to lack aliases. However, I'm not sure. Ok. I'll close the bug and reopen, if there will still be some problems. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:24:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:24:51 -0600 Subject: [LLVMbugs] [Bug 372] [llvmgcc/libc] isnan(x) function should turn into fcmp une(X, 0) Message-ID: <200701141924.l0EJOpAP007457@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=372 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-14 13:24 ------- This is implemented by llvm-gcc4. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:26:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:26:13 -0600 Subject: [LLVMbugs] [Bug 1111] NEW: instcombine miscompiles fp X == X Message-ID: <200701141926.l0EJQDL7007493@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1111 Summary: instcombine miscompiles fp X == X Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org instcombine isn't handling nan's right when folding fcmp X, X. For example, it turns this into false: define i1 %test(double %X) { %tmp = fcmp une double %X, %X ; [#uses=1] ret i1 %tmp } -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 13:42:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 13:42:52 -0600 Subject: [LLVMbugs] [Bug 1111] instcombine miscompiles fp X == X Message-ID: <200701141942.l0EJgqKK007942@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1111 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-14 13:42 ------- Fixed, testcase here: Transforms/InstCombine/2007-01-14-FcmpSelf.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042661.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 17:39:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 17:39:51 -0600 Subject: [LLVMbugs] [Bug 1112] NEW: llvm-objdump needed? Message-ID: <200701142339.l0ENdpGf011884@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1112 Summary: llvm-objdump needed? Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com While configuring llvm-gcc, I noticed this error: Checking linker read-only and read-write section mixing... /proj/llvm/cfe/src-2/gcc/configure: line 15320: llvm-objdump: command not found This probably results from use of the --program-prefix=llvm- option during configure. However, without using that option it is very difficult to separate regular gcc from llvm-gcc. Should we create an llvm-objdump program for bytecode files? ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 19:46:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 19:46:57 -0600 Subject: [LLVMbugs] [Bug 1113] NEW: llvm-upgrade generating undefined operand Message-ID: <200701150146.l0F1kvl8014563@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1113 Summary: llvm-upgrade generating undefined operand Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca test/Regression/Transforms/ConstProp/float-to-ptr-cast.ll is failing. Here's the error shown: FAIL: /home/nicholas/llvm/test/Regression/Transforms/ConstProp/float-to-ptr-cast.ll: child process exited abnormally llvm-as: :3,0: Reference to an invalid definition: 'cast_upgrade2' of type 'i64' opt: Standard Input is empty! llvm-dis: Standard Input is empty! It's producing the following output: define i32* %test1() { %cast_upgrade1 = fptoui float 0.0 to i64 %X.pntr.s1.u0 = inttoptr i64 %cast_upgrade2 to i32* ret i32* %X.pntr.s1.u0 } define i32* %test2() { ret i32* inttoptr( i64 fptoui(float 0.0 to i64) to i32*) } The bug is undefined %cast_upgrade2 when it should be %cast_upgrade1. This is a very recent change (today). ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 14 20:41:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 14 Jan 2007 20:41:35 -0600 Subject: [LLVMbugs] [Bug 1113] llvm-upgrade generating undefined operand Message-ID: <200701150241.l0F2fZ65016308@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1113 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rspencer at x10sys.com Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-14 20:41 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070108/042705.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 15 16:43:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 15 Jan 2007 16:43:18 -0600 Subject: [LLVMbugs] [Bug 1114] NEW: Assertion failed: DAG is cyclic. Message-ID: <200701152243.l0FMhI9R018940@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1114 Summary: Assertion failed: DAG is cyclic. 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: asl at math.spbu.ru ./llc bugpoint-reduced-simplified.bc llc: /home/asl/proj/llvm/src/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:336: void ::SelectionDAGLegalize::LegalizeDAG(): Assertion `Order.size() == Visited.size() && Order.size() == (unsigned)std::distance(DAG.allnodes_begin(), DAG.allnodes_end()) && "Error: DAG is cyclic!"' failed. This is reduction from Mozilla. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 15 21:46:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 15 Jan 2007 21:46:29 -0600 Subject: [LLVMbugs] [Bug 1095] Miscompilation of asm("{cntlz|cntlzw} %0, %1" ...) on PowerPC Message-ID: <200701160346.l0G3kTFt024685@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1095 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Core LLVM classes Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.0 Version|unspecified |trunk ------- Additional Comments From isanbard at gmail.com 2007-01-15 21:46 ------- Fixed with: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042747.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042748.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042749.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042750.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042751.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042752.html -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 00:01:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 00:01:11 -0600 Subject: [LLVMbugs] [Bug 1114] dag combine induces cycle when folding c ? (load p) : (load q) Message-ID: <200701160601.l0G61BwZ027490@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1114 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|Assertion failed: DAG is |dag combine induces cycle |cyclic. |when folding c ? (load p) : | |(load q) Target Milestone|--- |2.0 Version|trunk |1.7 ------- Additional Comments From sabre at nondot.org 2007-01-16 00:01 ------- Fixed. Testcase here: CodeGen/Generic/2007-01-15-LoadSelectCycle.ll Patch here: -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 01:24:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 01:24:20 -0600 Subject: [LLVMbugs] [Bug 918] c backend with array types can produce code rejected by gcc 4.0.1 Message-ID: <200701160724.l0G7OKws029330@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=918 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-16 01:24 ------- Nice work, I applied Gordon's testcase here: Regression/CodeGen/CBackend/2007-01-15-NamedArrayType.ll and here's a slightly adapted version of his patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042762.html Thanks Gordon! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 03:50:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 03:50:12 -0600 Subject: [LLVMbugs] [Bug 1115] NEW: Cannot handle multiple ASM constraints Message-ID: <200701160950.l0G9oCkw007982@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1115 Summary: Cannot handle multiple ASM constraints Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: isanbard at gmail.com This program: unsigned long int foo(unsigned long int in1, unsigned long int in2) { unsigned long int result; __asm__("{a%I5|add%I5c} %0,%1,%2" : "=&r"(result) : "r"(in1), "rI"(in2)); return result; } Gives this ICE: $ llvm-gcc.install/bin/gcc -c bad.c bad.c: In function 'foo': bad.c:5: error: operand number out of range /Users/wendling/llvm/llvm.src/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:2459: failed assertion `Constraints[i].Codes.size() == 1 && "Only handles one code so far!"' bad.c: At top level: bad.c:9: internal compiler error: Abort trap Please submit a full bug report, with preprocessed source if appropriate. See for instructions. This is more fall-out from the GMP compilation. -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 11:25:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 11:25:14 -0600 Subject: [LLVMbugs] [Bug 1116] NEW: Remove test/Regression Message-ID: <200701161725.l0GHPEUT016337@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1116 Summary: Remove test/Regression Product: new-bugs Version: unspecified Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com There's an extra level of encapsulation in the test directory that isn't needed. Most of the tests are in the test/Regression directory yet many of those tests are feature tests. Consequently this is a misnomer and just adds confusion and bulk. This enhancement requests asks that test/Regression/* be moved to test and the test/Regression directory deleted entirely. In order to retain history this involves copying some ,v files. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 11:25:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 11:25:57 -0600 Subject: [LLVMbugs] [Bug 1115] Cannot handle multiple ASM constraints Message-ID: <200701161725.l0GHPvdO016366@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1115 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-01-16 11:25 ------- Yep, known issue. In the meantime you can s/rI/r/ to work around this. *** This bug has been marked as a duplicate of 839 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 11:52:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 11:52:54 -0600 Subject: [LLVMbugs] [Bug 1117] NEW: Assembler doesn't catch bad casts. Message-ID: <200701161752.l0GHqsmO016935@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1117 Summary: Assembler doesn't catch bad casts. Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: LLVM assembly language parser AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com The AsmParser isn't catching illegal casts and reporting them as errors. This causes llvm-as to fail with an assertion when the attempt is made to instantiate the illegal cast instruction. For example: %X = trunc i32 %Y to i8* The instruction needed here is inttoptr, not trunc. This causes an assertion instead of a nice error message. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 15:19:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 15:19:47 -0600 Subject: [LLVMbugs] [Bug 1118] NEW: Enh: add stack alignment to llvm::TargetData Message-ID: <200701162119.l0GLJl1k021092@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1118 Summary: Enh: add stack alignment to llvm::TargetData Product: libraries Version: trunk Platform: All OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org There exist platforms for which the stack alignment is different than the usual memory layout alignments, specifically the IBM Cell SPU, in my case. Added support so that stack alignment of types can be specified via the TargetData::init() string. Patched up instances of llvm::MachineFrameInfo::CreateStackObject() so that the type's stack alignment is used versus the memory alignment. Please see the "NOTE:" in SelectionDAGISel.cpp, as this patch might be a solution to the "FIXME:" above the note. Also, this is a sequence of two patches. The first patch updates documentation in TargetData.h, updates AutoRegen.sh to recognize newer versions of the autoconf tools. The second patch contains the actual stack alignment patch. (N.B.: I work out of my own different repo, hence the two patches. Sorry.) ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 16:12:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 16:12:33 -0600 Subject: [LLVMbugs] [Bug 1119] NEW: Enh: add PatLeaf to test for zero valued operands Message-ID: <200701162212.l0GMCXPD022216@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1119 Summary: Enh: add PatLeaf to test for zero valued operands Product: libraries Version: trunk Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Target Description Classes AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org We have immAllZerosV but no immediate test for a zero value. Suggest the attached patch to TargetSelectionDAG.td. Adds a PatLeaf called "immIsZero" that calls Constant::isNullValue(). ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 18:34:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 18:34:29 -0600 Subject: [LLVMbugs] [Bug 1120] NEW: ConstantInt should overload getType() to return an IntegerType* Message-ID: <200701170034.l0H0YTBe024621@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1120 Summary: ConstantInt should overload getType() to return an IntegerType* Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca If a ConstantInt user wants the number of bits in the int, they have to call getType, then cast the Type* to an IntegerType*. Instead, it should be possible to retrieve the IntegerType* directly from ConstantInt. This is difficult because ConstantInt doesn't actually keep the IntegerType* internally, nor does it require one on construction, etc. This will probably cause residual effects throughout LLVM, but it shouldn't be very hard to clean up. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 20:19:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 20:19:13 -0600 Subject: [LLVMbugs] [Bug 1074] Remove Stacker from llvm main repository Message-ID: <200701170219.l0H2JDps006696@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1074 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |build-problem Resolution| |FIXED Summary|Stacker configure doesn't |Remove Stacker from llvm |run automatically |main repository ------- Additional Comments From rspencer at x10sys.com 2007-01-16 20:19 ------- Stacke was the last part of LLVM that required llvm-gcc. By making it a separate project, we have eliminated that dependency. This is now done. Stacker is now known as llvm-stacker and lives it its own repository. The configuration has been updated to reflect the new name. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042795.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 20:24:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 20:24:13 -0600 Subject: [LLVMbugs] [Bug 1094] SetVector iterators aren't really mutable Message-ID: <200701170224.l0H2ODn4007004@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1094 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |code-cleanup Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-16 20:24 ------- Patch applied. Thanks, Gordon. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 20:48:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 20:48:31 -0600 Subject: [LLVMbugs] [Bug 1117] Assembler doesn't catch bad casts. Message-ID: <200701170248.l0H2mVde007582@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1117 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-16 20:48 ------- Fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042809.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042810.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042811.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 21:40:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 21:40:08 -0600 Subject: [LLVMbugs] [Bug 821] llvm-gcc4 doesn't build: pthread_once not found Message-ID: <200701170340.l0H3e86u008577@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=821 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-16 21:40 ------- This appears to have been fixed somewhere along the line. I can now build with --enable-threads (the default) without any trouble. It even runs the dejagnu suite properly. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 21:42:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 21:42:15 -0600 Subject: [LLVMbugs] [Bug 822] llvm needs to support external weak linkage Message-ID: <200701170342.l0H3gFq9008670@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=822 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asl at math.spbu.ru Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-16 21:42 ------- This was fixed and hit the mirror a long time ago. Resolved. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 16 23:28:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 16 Jan 2007 23:28:05 -0600 Subject: [LLVMbugs] [Bug 1078] Move Stacker out of the LLVM repository to its own, like Java Message-ID: <200701170528.l0H5S5aQ017815@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1078 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-01-16 23:28 ------- *** This bug has been marked as a duplicate of 1074 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 01:21:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 01:21:20 -0600 Subject: [LLVMbugs] [Bug 1116] Remove test/Regression Message-ID: <200701170721.l0H7LKo0020028@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1116 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-17 01:21 ------- done ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 01:24:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 01:24:52 -0600 Subject: [LLVMbugs] [Bug 1119] Enh: add PatLeaf to test for zero valued operands Message-ID: <200701170724.l0H7OqUT020124@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1119 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-01-17 01:24 ------- You should just be able to use a literal 0 in your pattern. For example, ineg is defined as: def ineg : PatFrag<(ops node:$in), (sub 0, node:$in)>; We need a node for immAllOnes when defining patterns that are independent of the width of the datatype (e.g. 255 or 65535 or 4294967295, etc), but 0 doesn't have this issue. If 0 doesn't work for some reason, you hit a bug. Please reopen this PR if so. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 01:52:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 01:52:44 -0600 Subject: [LLVMbugs] [Bug 1116] Remove test/Regression Message-ID: <200701170752.l0H7qiRl020813@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1116 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From rspencer at x10sys.com 2007-01-17 01:52 ------- You resolved this, but apparently the Regression directory hasn't been removed? ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 02:02:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 02:02:30 -0600 Subject: [LLVMbugs] [Bug 1116] Remove test/Regression Message-ID: <200701170802.l0H82U2R021075@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1116 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-17 02:02 ------- I guess the repository is okay. People will just have to manually remove their Regression directory. I was able to get this to pass "make check", so I guess its okay. Re-resolving. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 11:30:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 11:30:52 -0600 Subject: [LLVMbugs] [Bug 1121] NEW: Should split libcodegen into libcodegen and libmachine Message-ID: <200701171730.l0HHUq28027435@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1121 Summary: Should split libcodegen into libcodegen and libmachine Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org lib/Codegen/* contains two separate things: implementations of the Machine* classes (e.g. MachineBasicBlock) and a collection of code generator-related passes. We should split this into two pieces: lib/CodeGen/Machine and lib/Codegen. The reason for this is that individual targets (e.g. the PPC target) should depend (in a library dependence sense) on the machine bits, but not on the Codegen bits. Currently, the various targets do depend on various machine code passes (e.g. the asmprinter) which causes large amounts of code to be linked in to apps that don't need it (e.g. lli gets asmprinter, machowriter, elfwriter, etc). Once lib/CodeGen is split up, we can chase these dependencies down more easily and eliminate them. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 11:47:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 11:47:10 -0600 Subject: [LLVMbugs] [Bug 1092] Assertion failure in GlobalVariable.h Message-ID: <200701171747.l0HHlAna003043@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1092 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-17 11:47 ------- Okay, it sounds like this is fixed. Thanks Anton. Domagoj, if you see this again with TOT, please reopen. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 12:40:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 12:40:32 -0600 Subject: [LLVMbugs] [Bug 1030] __APPLE_CC__ is incorrect defined on Linux Message-ID: <200701171840.l0HIeWIC010921@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1030 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-01-17 12:40 ------- Looks good. Applied. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 14:12:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 14:12:27 -0600 Subject: [LLVMbugs] [Bug 570] [gccld] native libraries not passed to the backend Message-ID: <200701172012.l0HKCRWM013043@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=570 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From rspencer at x10sys.com 2007-01-17 14:12 ------- Markus, I'm marking this bug as WORKSFORME because we've dropped llvm-gcc3 support. llvm-gcc4 no longer requires gccld and thus there's no need to build paths to pass to or from gccld. If you still have an issue with this with llvm-gcc4, please reopen. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 14:25:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 14:25:53 -0600 Subject: [LLVMbugs] [Bug 739] Can't `make install prefix=/tmp/llvm'. DESTDIR not supported either. Message-ID: <200701172025.l0HKPrVc013318@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=739 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-17 14:25 ------- Tested. Works. Patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042849.html Resolved. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 17 17:36:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 17 Jan 2007 17:36:36 -0600 Subject: [LLVMbugs] [Bug 1028] CBE should support llvm.stacksave/stackrestore Message-ID: <200701172336.l0HNaawH017171@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1028 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-17 17:36 ------- Nice patch, Gordon. This fixes 176.gcc/CBE. Its been a long outstanding issue. Well done. Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:30:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:30:35 -0600 Subject: [LLVMbugs] [Bug 1118] Enh: add stack alignment to llvm::TargetData Message-ID: <200701180730.l0I7UZ0F016331@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1118 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-01-18 01:30 ------- Ah, there is actually already a bug for this. I'm closing this as a dup of the older bug, I'll review your patch tomorrow. -Chris *** This bug has been marked as a duplicate of 468 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:32:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:32:15 -0600 Subject: [LLVMbugs] [Bug 869] llvm-gcc4 lacks link-time-optimization support Message-ID: <200701180732.l0I7WFCh016397@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=869 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-18 01:32 ------- liblto implements this. Chandler's recent work on binutils makes this available on GNU platforms. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:35:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:35:33 -0600 Subject: [LLVMbugs] [Bug 587] [ppc] Reference to a constant pool in simple functions shouldn't spill LR reg to the stack Message-ID: <200701180735.l0I7ZXM5016540@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=587 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |LATER ------- Additional Comments From sabre at nondot.org 2007-01-18 01:35 ------- I moved the contents of this PR to the PPC readme where it is more visible to those interested in PPC work. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:39:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:39:28 -0600 Subject: [LLVMbugs] [Bug 702] [codegen] Compile copysign more optimally Message-ID: <200701180739.l0I7dSO1016683@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=702 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-18 01:39 ------- we have codegen support for copysign now. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:41:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:41:26 -0600 Subject: [LLVMbugs] [Bug 703] [vmcore] Split InternalLinkage into Private and Local Linkage Message-ID: <200701180741.l0I7fQVF016737@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=703 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-01-18 01:41 ------- As it turns out, we have another way to do this with 'asm' names. This is what the objc front-end does, so we don't need this anymore. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:44:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:44:18 -0600 Subject: [LLVMbugs] [Bug 706] [loops] Simple for-loop does not get Message-ID: <200701180744.l0I7iI1i016789@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=706 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-18 01:44 ------- We generate reasonable, but not optimal code for this loop now. The remaining problems are all documented in the PPC readme. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 01:53:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 01:53:12 -0600 Subject: [LLVMbugs] [Bug 755] [executionengine] Would be nice to be able to execute from multiple modules at a time Message-ID: <200701180753.l0I7rCZU017013@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=755 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |1.9 ------- Additional Comments From sabre at nondot.org 2007-01-18 01:53 ------- This was in llvm 1.9 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 02:03:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 02:03:00 -0600 Subject: [LLVMbugs] [Bug 756] --with-externals change broke --with-spec2000 and friends Message-ID: <200701180803.l0I830R0017857@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=756 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From sabre at nondot.org 2007-01-18 02:02 ------- --with-spec2000 is deprecated. When bug 919 is done, it will be removed. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 02:03:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 02:03:26 -0600 Subject: [LLVMbugs] [Bug 790] Portability issues for Win32 Message-ID: <200701180803.l0I83QpJ017896@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=790 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-01-18 02:03 ------- Yes, it seems so. Only 331 remains. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 02:03:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 02:03:33 -0600 Subject: [LLVMbugs] [Bug 778] Doxygen Documentation Is Lacking Message-ID: <200701180803.l0I83XnV017920@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=778 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |LATER ------- Additional Comments From rspencer at x10sys.com 2007-01-18 02:03 ------- There are several other things needing to be done to: 1. Turn on scavenging of .cpp files not just .h files (if it isn't already). 2. Make sure that every class and every public method of every class in the llvm/include/llvm directory hierarchy has a comment. 3. Check consistency between header comments and actual functionality. Things change over time and comments get stale. However, all of this bug's requests comes under the realm of "continuous improvement". Consequently, we're closing it. The items here are taken as a personal "to do" item for Reid and it is recommended that all developers pitch in to aid in this effort. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 07:13:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 07:13:25 -0600 Subject: [LLVMbugs] [Bug 1122] NEW: getelementptr gets confused whith nested type Message-ID: <200701181313.l0IDDPgO030230@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1122 Summary: getelementptr gets confused whith nested type Product: tools Version: trunk Platform: Macintosh OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: eric at vanrietpaap.nl llvm-upgrade gives this error when upgrading the attached .ll file. llvm-upgrade: testme.ll:9: error: Invalid type for index while reading token: '%source1' ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 18 16:13:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 18 Jan 2007 16:13:53 -0600 Subject: [LLVMbugs] [Bug 919] Eliminate llvm-test --with-program=xxx options for externals Message-ID: <200701182213.l0IMDrBG014568@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=919 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-18 16:13 ------- This is fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042907.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070115/042908.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 03:01:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 03:01:05 -0600 Subject: [LLVMbugs] [Bug 623] request for -malloc2alloca optimization pass Message-ID: <200701190901.l0J9156O003703@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=623 eric at vanrietpaap.nl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From eric at vanrietpaap.nl 2007-01-19 03:01 ------- PyPy now does some escape analysis of its own so it's unclear if this would be still useful to do. Please feel free to 'remove' this bugreport ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 03:02:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 03:02:07 -0600 Subject: [LLVMbugs] [Bug 942] analyze missing-optimization-opportunity Message-ID: <200701190902.l0J927wo003735@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=942 eric at vanrietpaap.nl changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |LATER ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 12:42:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 12:42:36 -0600 Subject: [LLVMbugs] [Bug 942] analyze missing-optimization-opportunity Message-ID: <200701191842.l0JIgaWd017145@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=942 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|LATER | ------- Additional Comments From sabre at nondot.org 2007-01-19 12:42 ------- I'd prefer to keep this bug open, so that others interested in tackling it will see it. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 12:56:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 12:56:46 -0600 Subject: [LLVMbugs] [Bug 1123] NEW: Convert DejaGNU Tests To Use Tcl Fully Message-ID: <200701191856.l0JIuksH017409@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1123 Summary: Convert DejaGNU Tests To Use Tcl Fully Product: Test Suite Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: DejaGNU AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Currently, the RUN: lines in the llvm/test suite depend on Unix utilities. This makes them not portable to other systems and also makes them slower than they might be. If the RUN: lines were interpreted by Tcl instead of a Unix shell, it would be portable and faster (fewer processes created). Tcl is able to emulate grep which is the main tool used in the scripts. It is also extensible so we could create a library of Tcl functions for executing things like "not grep" or "notcast" or "prcontext", etc. Piping of data could also be handled by Tcl. Tcl is also flexible on its syntax so it might be possible to come up with Tcl definitions that allow it to interpret the the shell syntax. If not, the RUN: lines would need to be translated to Tcl. A phased approach is suggested to implement this: 1. Scan the RUN: lines in use and find out what Unix type things need to be supported in Tcl. 2. Implement the things needed from the list in #1 in llvm/test/lib as Tcl function definitions (e.g. notcast). 3. Implement the "TestRunner.sh" functionality in Tcl so that it can interpret the RUN:, XFAIL: and other lines in a script. 4. Write documentation on how to use the Tcl-based testing facilities and encourage developers to write new test cases using it. 5. Convert test cases one by one to use the new Tcl facilities. 6. When all old syntax test cases are fixed to use the new syntax, remove the support scripts and etc. This approach allows the feature to be implemented incrementally by keeping both systems alive until the conversion is completed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 15:16:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 15:16:35 -0600 Subject: [LLVMbugs] [Bug 1043] Arbitrary bitwidth integers Message-ID: <200701192116.l0JLGZro020042@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1043 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-19 15:16 ------- All tasks for this PR have been completed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 15:16:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 15:16:51 -0600 Subject: [LLVMbugs] [Bug 1120] ConstantInt should overload getType() to return an IntegerType* Message-ID: <200701192116.l0JLGpTY020069@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1120 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From isanbard at gmail.com Fri Jan 19 16:35:26 2007 From: isanbard at gmail.com (Bill Wendling) Date: Fri, 19 Jan 2007 14:35:26 -0800 Subject: [LLVMbugs] Compilation Errors Message-ID: <16e5fdf90701191435p36c7f0c6q8062330593ebbb05@mail.gmail.com> Hi all, I'm getting these errors. Anyone have a suggestion? g++ -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/include -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/include -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -fno-exceptions -D_DEBUG -DLLVM_VERSION_INFO='" Apple Build #1031-08"' -Woverloaded-virtual -pedantic -Wall -W -Wwrite-strings -Wno-long-long -Wunused -Wno-unused-parameter -fstrict-aliasing -arch ppc -arch i386 -c /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp -o /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT/Release/HowToUseJIT.o /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp: In function 'int main()': /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:57: error: no matching function for call to 'llvm::Module::getOrInsertFunction(const char [5], const llvm::IntegerType*&, const llvm::IntegerType*&, llvm::Type*)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: note: candidates are: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::FunctionType*) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: note: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::Type*, ...) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:64: error: no matching function for call to 'llvm::ConstantInt::get(const llvm::IntegerType*&, int)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: note: candidates are: static llvm::ConstantInt* llvm::ConstantInt::get(const llvm::Type*, int64_t) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:83: error: no matching function for call to 'llvm::Module::getOrInsertFunction(const char [4], const llvm::IntegerType*&, llvm::Type*)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: note: candidates are: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::FunctionType*) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: note: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::Type*, ...) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:89: error: no matching function for call to 'llvm::ConstantInt::get(const llvm::IntegerType*&, int)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: note: candidates are: static llvm::ConstantInt* llvm::ConstantInt::get(const llvm::Type*, int64_t) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp: In function 'int main()': /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:57: error: no matching function for call to 'llvm::Module::getOrInsertFunction(const char [5], const llvm::IntegerType*&, const llvm::IntegerType*&, llvm::Type*)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: note: candidates are: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::FunctionType*) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: note: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::Type*, ...) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:64: error: no matching function for call to 'llvm::ConstantInt::get(const llvm::IntegerType*&, int)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: note: candidates are: static llvm::ConstantInt* llvm::ConstantInt::get(const llvm::Type*, int64_t) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:83: error: no matching function for call to 'llvm::Module::getOrInsertFunction(const char [4], const llvm::IntegerType*&, llvm::Type*)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: note: candidates are: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::FunctionType*) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: note: llvm::Constant* llvm::Module::getOrInsertFunction(const std::string&, const llvm::Type*, ...) /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:89: error: no matching function for call to 'llvm::ConstantInt::get(const llvm::IntegerType*&, int)' /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: note: candidates are: static llvm::ConstantInt* llvm::ConstantInt::get(const llvm::Type*, int64_t) lipo: can't figure out the architecture type of: /var/tmp//cckeom7f.out make[3]: *** [/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT/Release/HowToUseJIT.o] Error 1 make[2]: *** [HowToUseJIT/.makeall] Error 2 make[1]: *** [all] Error 1 -bw From rspencer at reidspencer.com Fri Jan 19 16:46:46 2007 From: rspencer at reidspencer.com (Reid Spencer) Date: Fri, 19 Jan 2007 14:46:46 -0800 Subject: [LLVMbugs] Compilation Errors In-Reply-To: <16e5fdf90701191435p36c7f0c6q8062330593ebbb05@mail.gmail.com> References: <16e5fdf90701191435p36c7f0c6q8062330593ebbb05@mail.gmail.com> Message-ID: <1169246806.23360.288.camel@bashful.x10sys.com> Hi Bill, On Fri, 2007-01-19 at 14:35 -0800, Bill Wendling wrote: > Hi all, > > I'm getting these errors. Anyone have a suggestion? Sorry, I forgot to check in some files. Please update your exmaples directory, should be fixed now. Reid. > > g++ -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT > -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT > -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/include > -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include > -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/include > -I/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include > -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -fno-exceptions -D_DEBUG > -DLLVM_VERSION_INFO='" Apple Build #1031-08"' -Woverloaded-virtual > -pedantic -Wall -W -Wwrite-strings -Wno-long-long -Wunused > -Wno-unused-parameter -fstrict-aliasing -arch ppc -arch i386 -c > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp > -o /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT/Release/HowToUseJIT.o > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp: > In function 'int main()': > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:57: > error: no matching function for call to > 'llvm::Module::getOrInsertFunction(const char [5], const > llvm::IntegerType*&, const llvm::IntegerType*&, llvm::Type*)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: > note: candidates are: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::FunctionType*) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: > note: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::Type*, ...) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:64: > error: no matching function for call to 'llvm::ConstantInt::get(const > llvm::IntegerType*&, int)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: > note: candidates are: static llvm::ConstantInt* > llvm::ConstantInt::get(const llvm::Type*, int64_t) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:83: > error: no matching function for call to > 'llvm::Module::getOrInsertFunction(const char [4], const > llvm::IntegerType*&, llvm::Type*)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: > note: candidates are: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::FunctionType*) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: > note: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::Type*, ...) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:89: > error: no matching function for call to 'llvm::ConstantInt::get(const > llvm::IntegerType*&, int)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: > note: candidates are: static llvm::ConstantInt* > llvm::ConstantInt::get(const llvm::Type*, int64_t) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp: > In function 'int main()': > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:57: > error: no matching function for call to > 'llvm::Module::getOrInsertFunction(const char [5], const > llvm::IntegerType*&, const llvm::IntegerType*&, llvm::Type*)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: > note: candidates are: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::FunctionType*) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: > note: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::Type*, ...) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:64: > error: no matching function for call to 'llvm::ConstantInt::get(const > llvm::IntegerType*&, int)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: > note: candidates are: static llvm::ConstantInt* > llvm::ConstantInt::get(const llvm::Type*, int64_t) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:83: > error: no matching function for call to > 'llvm::Module::getOrInsertFunction(const char [4], const > llvm::IntegerType*&, llvm::Type*)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:177: > note: candidates are: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::FunctionType*) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Module.h:184: > note: llvm::Constant* > llvm::Module::getOrInsertFunction(const std::string&, const > llvm::Type*, ...) > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/examples/HowToUseJIT/HowToUseJIT.cpp:89: > error: no matching function for call to 'llvm::ConstantInt::get(const > llvm::IntegerType*&, int)' > /Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/src/llvm/include/llvm/Constants.h:93: > note: candidates are: static llvm::ConstantInt* > llvm::ConstantInt::get(const llvm::Type*, int64_t) > lipo: can't figure out the architecture type of: /var/tmp//cckeom7f.out > make[3]: *** [/Volumes/Gir/devel/llvm/SUBMISSIONS/llvm-nightly.roots/llvm-nightly~obj/obj-llvm/examples/HowToUseJIT/Release/HowToUseJIT.o] > Error 1 > make[2]: *** [HowToUseJIT/.makeall] Error 2 > make[1]: *** [all] Error 1 > > > -bw > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From isanbard at gmail.com Fri Jan 19 17:09:44 2007 From: isanbard at gmail.com (Bill Wendling) Date: Fri, 19 Jan 2007 15:09:44 -0800 Subject: [LLVMbugs] Compilation Errors In-Reply-To: <1169246806.23360.288.camel@bashful.x10sys.com> References: <16e5fdf90701191435p36c7f0c6q8062330593ebbb05@mail.gmail.com> <1169246806.23360.288.camel@bashful.x10sys.com> Message-ID: <16e5fdf90701191509m4ba25438g87000b6f906ee891@mail.gmail.com> On 1/19/07, Reid Spencer wrote: > Hi Bill, > > On Fri, 2007-01-19 at 14:35 -0800, Bill Wendling wrote: > > Hi all, > > > > I'm getting these errors. Anyone have a suggestion? > > Sorry, I forgot to check in some files. Please update your exmaples > directory, should be fixed now. > Hi Reid, That solved the problem. thank you! -bw From bugzilla-daemon at cs.uiuc.edu Fri Jan 19 19:15:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 19 Jan 2007 19:15:41 -0600 Subject: [LLVMbugs] [Bug 1124] NEW: CBE Fails On Linux if __sigsetjmp used Message-ID: <200701200115.l0K1FfYW024526@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1124 Summary: CBE Fails On Linux if __sigsetjmp used Product: libraries Version: 1.0 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com gcc Output/400.perlbench.cbe.c -lm -O3 -fno-strict-aliasing -O2 -fno-inline -o Output/400.perlbench.cbe Output/400.perlbench.cbe.c:2274: error: conflicting types for ???__sigsetjmp??? /usr/include/setjmp.h:58: error: previous declaration of ???__sigsetjmp??? was here Output/400.perlbench.cbe.c:2291: error: conflicting types for ???siglongjmp??? /usr/include/setjmp.h:109: error: previous declaration of ???siglongjmp??? was here This is occurring because the program defines __sigsetjmp one way but it is also defined in setjmp.h, differently. The CBE needs to include setjmp.h for its own uses. Perhaps the thing to do is have the CBE just emit the declarations it needs and not include setjmp.h. This appears to be a Linux specific problem. This is currently causing failures for: External/SPEC/CINT2006/400.perlbench External/SPEC/CINT2006/403.gcc ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From isanbard at gmail.com Fri Jan 19 19:47:00 2007 From: isanbard at gmail.com (Bill Wendling) Date: Fri, 19 Jan 2007 17:47:00 -0800 Subject: [LLVMbugs] Error during compilation Message-ID: <16e5fdf90701191746j799ab707kbd6cf21b77c412a6@mail.gmail.com> Hi all, You're going to hate me, but I'm kind of at a loss as to how to reduce this testcase. There's a failure when I compile a program. It uses PCHs, so I couldn't get a .i file out of it. Attached are the .ll and .s files. The error when doing the assembly is: bad.s:3899:Parameter syntax error (parameter 1) bad.s:3900:FATAL:Symbol L46$pb already defined. My hope is that someone kind of knows what this is doing. Also, any hints on how to use bugpoint to reduce these types of cases would be appreciated. -bw -------------- next part -------------- A non-text attachment was scrubbed... Name: bad.s Type: application/octet-stream Size: 177911 bytes Desc: not available Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070119/4a5720de/attachment-0002.obj -------------- next part -------------- A non-text attachment was scrubbed... Name: bad.ll Type: application/octet-stream Size: 346814 bytes Desc: not available Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070119/4a5720de/attachment-0003.obj From bugzilla-daemon at cs.uiuc.edu Sat Jan 20 01:43:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 20 Jan 2007 01:43:08 -0600 Subject: [LLVMbugs] [Bug 1125] NEW: Update autoconf and libtool Message-ID: <200701200743.l0K7h8XH030766@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1125 Summary: Update autoconf and libtool Product: Build scripts Version: cvs Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: autoconf AssignedTo: rspencer at x10sys.com ReportedBy: rspencer at x10sys.com We've had a couple requests for updating to the latest autoconf tools. This is just a reminder to do it sometime before 2.0 is released. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 20 04:25:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 20 Jan 2007 04:25:54 -0600 Subject: [LLVMbugs] [Bug 1108] CodeGen/Generic/vector-identity-shuffle.ll regression fails on x86_64 Message-ID: <200701201025.l0KAPsKV010108@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1108 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-01-20 04:25 ------- Ok. Fixed DAGCombiner to use iPtr instead. My concern with this fix is even though i64 is a legal type, v4i64 isn't. If DAG combiner creates a VBUILD_VECTOR with 4 i64 elements (as it does here), would it cause some troubles downstream? Fix appears to work for now. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 20 11:14:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 20 Jan 2007 11:14:37 -0600 Subject: [LLVMbugs] [Bug 1126] NEW: CBE miscompiles or crashes compiling programs with vectors Message-ID: <200701201714.l0KHEbma016376@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1126 Summary: CBE miscompiles or crashes compiling programs with vectors Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: gordonhenriksen at mac.com Noticed through inspection. % cat cbev.ll define <4 x i32> %foo(<4 x i32> %a, <4 x i32> %b) { %c = sub <4 x i32> %a, %b %d = add <4 x i32> %b, %c %e = mul <4 x i32> %c, %d ; %f = udiv <4 x i32> %d, %e ; Fails an assertion in CBackend.cpp ; %g = sdiv <4 x i32> %e, %f ; Same ret <4 x i32> %e } % llvm-as -o cbev.bc cbev.ll % llc -march=c -o cbev.c cbev.bc % tail -n 7 cbev.c unsigned int foo(unsigned int ltmp_0_1[4], unsigned int ltmp_1_1[4])[4] { unsigned int ltmp_2_1[4]; ltmp_2_1 = ltmp_0_1 - ltmp_1_1; return (ltmp_2_1*(ltmp_1_1 + ltmp_2_1)); } % gcc -o cbev.o -c cbev.c cbev.c:106: error: ???foo??? declared as function returning an array cbev.c:107: warning: conflicting types for built-in function ???malloc??? cbev.c:129: error: ???foo??? declared as function returning an array cbev.c: In function ???foo???: cbev.c:132: error: incompatible types in assignment cbev.c:133: error: invalid operands to binary + An error message, inserting a vector lowering pass before the CBE, or fixing the codegen would all be improvements. Emitting Altivec 'vector' types instead of arrays would make fixing the codegen relatively easy when targetting PowerPC, since Altivec provides operator overloads for vector ops. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 20 14:47:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 20 Jan 2007 14:47:20 -0600 Subject: [LLVMbugs] [Bug 1127] NEW: Redesign LLVM 2.0 bytecode format Message-ID: <200701202047.l0KKlKTS020199@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1127 Summary: Redesign LLVM 2.0 bytecode format Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Bytecode Writer AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org We want to scrap the current way we handle .bc files. Here are some ideas/desires for the new format: 1. I'd like to get rid of libbzip2. 2. I think it makes sense to use a completely variable bitwidth format, not trying to keep things at byte alignments. 3. We should probably keep type planes, as they allow us to use the type constraints in llvm to make the encoding more dense. 4. We should consider a packetized format like dwarf for things like the moduleinfoblock and functioninfoblocks, which makes it easier to extend the format without a new revision #. Any other ideas? -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 20 18:33:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 20 Jan 2007 18:33:18 -0600 Subject: [LLVMbugs] [Bug 970] Should audit LLVM for incorrect uses of isFloatingPoint() Message-ID: <200701210033.l0L0XIUG024981@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=970 gordonhenriksen at mac.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From gordonhenriksen at mac.com 2007-01-20 18:33 ------- Thanks, Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 21 00:03:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 21 Jan 2007 00:03:29 -0600 Subject: [LLVMbugs] [Bug 1128] NEW: configure script doesn't know about arm backend Message-ID: <200701210603.l0L63TiA030158@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1128 Summary: configure script doesn't know about arm backend Product: Build scripts Version: cvs Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: autoconf AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The configure.ac script doesn't know about arm, thus --enable-targets=arm doesn't work. It looks like ARM entries should be added everywhere x86 is handled in configure.ac. This was reported by ramana on irc. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 21 00:33:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 21 Jan 2007 00:33:02 -0600 Subject: [LLVMbugs] [Bug 1128] configure script doesn't know about arm backend Message-ID: <200701210633.l0L6X2v4030942@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1128 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-21 00:33 ------- Should be fixed now, but please test it. Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 21 13:46:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 21 Jan 2007 13:46:18 -0600 Subject: [LLVMbugs] [Bug 927] Get Rid Of Annotations Message-ID: <200701211946.l0LJkI5K013863@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=927 gordonhenriksen at mac.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From gordonhenriksen at mac.com 2007-01-21 13:46 ------- *** This bug has been marked as a duplicate of 746 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 21 19:12:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 21 Jan 2007 19:12:53 -0600 Subject: [LLVMbugs] [Bug 1129] NEW: Debugging support for JIT code Message-ID: <200701220112.l0M1CrFH019530@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1129 Summary: Debugging support for JIT code Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Target-Independent JIT AssignedTo: unassignedbugs at nondot.org ReportedBy: contact at leonard-ritter.com When debugging JIT code, I'd like to have some means to either step through code using GDB or see a trace of JITed code. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 22 13:06:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 22 Jan 2007 13:06:09 -0600 Subject: [LLVMbugs] [Bug 468] [target description] Need a way to describe *preferred* target alignments Message-ID: <200701221906.l0MJ69CE013264@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=468 scottm at aero.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From scottm at aero.org 2007-01-22 13:06 ------- Patch from bug 1118 applied, fixes this bug too. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From dpatel at apple.com Mon Jan 22 13:25:48 2007 From: dpatel at apple.com (Devang Patel) Date: Mon, 22 Jan 2007 11:25:48 -0800 Subject: [LLVMbugs] Error during compilation In-Reply-To: <16e5fdf90701191746j799ab707kbd6cf21b77c412a6@mail.gmail.com> References: <16e5fdf90701191746j799ab707kbd6cf21b77c412a6@mail.gmail.com> Message-ID: <7C1B4227-12BD-41C6-821F-89141B88A330@apple.com> On Jan 19, 2007, at 5:47 PM, Bill Wendling wrote: > Hi all, > > You're going to hate me, but I'm kind of at a loss as to how to reduce > this testcase. There's a failure when I compile a program. It uses > PCHs, so I couldn't get a .i file out of it. Follow these steps : 1) Find the command line that is used to build PCH file. Note down the include prefix header files used to build PCH. 2) If there is a -include switch on failing command line pointing to a PCH file (or directory) then remove it. (Sometimes developers #include ) 3) Add -include command line switch to include prefix header file from step 1) 4) Reproduce the crash. Now, PCH is out of equation. - Devang > Attached are the .ll and > .s files. The error when doing the assembly is: > > bad.s:3899:Parameter syntax error (parameter 1) > bad.s:3900:FATAL:Symbol L46$pb already defined. > > My hope is that someone kind of knows what this is doing. Also, any > hints on how to use bugpoint to reduce these types of cases would be > appreciated. > > -bw_______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From dpatel at apple.com Mon Jan 22 13:28:49 2007 From: dpatel at apple.com (Devang Patel) Date: Mon, 22 Jan 2007 11:28:49 -0800 Subject: [LLVMbugs] [Bug 1127] NEW: Redesign LLVM 2.0 bytecode format In-Reply-To: <200701202047.l0KKlKTS020199@zion.cs.uiuc.edu> References: <200701202047.l0KKlKTS020199@zion.cs.uiuc.edu> Message-ID: On Jan 20, 2007, at 12:47 PM, bugzilla-daemon at cs.uiuc.edu wrote: > 2. I think it makes sense to use a completely variable bitwidth > format, not trying to keep things at byte > alignments However, if variable bitwidth format noticeably increase BC read time then it may not be a good idea. - Devang -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070122/91a5064d/attachment.html From bugzilla-daemon at cs.uiuc.edu Wed Jan 24 04:09:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 24 Jan 2007 04:09:49 -0600 Subject: [LLVMbugs] [Bug 1130] NEW: JIT (x86) produces wrong exception for i8 srem Message-ID: <200701241009.l0OA9nCM011876@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1130 Summary: JIT (x86) produces wrong exception for i8 srem Product: tools Version: trunk Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: lli AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr $ lli f.bc lli((anonymous namespace)::PrintStackTrace()+0x1a)[0x878977a] lli((anonymous namespace)::SignalHandler(int)+0x112)[0x8789a40] [0xb7f04420] lli(llvm::ExecutionEngine::runFunctionAsMain(llvm::Function*, std::vector, std::allocator >, std::allocator, std::allocator > > > const&, char const* const*)+0x23a)[0x84cc066] lli(main+0x2ba)[0x83b6056] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc)[0xb7c8cebc] lli[0x83b58e1] Floating point exception vs $ lli -force-interpreter f.bc -128 rd 1 = 0 The C back-end agrees with lli -force-interpreter (and so do I!). Running "bugpoint --run-jit f.bc" reduces it a bit. I send the byte-code in an attachment in a moment. PS: Probably I should report this for one of the libraries, but it is not clear where x86 JIT bugs should go. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 24 12:31:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 24 Jan 2007 12:31:57 -0600 Subject: [LLVMbugs] [Bug 1130] JIT (x86) produces wrong exception for i8 srem Message-ID: <200701241831.l0OIVvBc020459@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1130 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED OS/Version|Linux |All Platform|Other |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-24 12:31 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043184.html Very very nice catch, thanks! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 25 01:37:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 25 Jan 2007 01:37:55 -0600 Subject: [LLVMbugs] [Bug 1131] NEW: SimplfyCfg deletes too much probably due to instcombine fault Message-ID: <200701250737.l0P7btiv003372@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1131 Summary: SimplfyCfg deletes too much probably due to instcombine fault Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Consider the attached bytecode. ./opt -instcombine -simplifycfg nsSubstring.bc strips all code from %_ZN11nsSubstring7ReplaceEjjRK18nsAString_internal function (the same for "EjjRK19" one). Without instcombine - everything is ok. This breaks Mozilla builds. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 25 09:49:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 25 Jan 2007 09:49:55 -0600 Subject: [LLVMbugs] [Bug 1132] NEW: Missing Promote case in legalize Message-ID: <200701251549.l0PFntng021843@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1132 Summary: Missing Promote case in legalize Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu NODE: 0x120b12b80: i16 = Register #1024 llc: LegalizeDAG.cpp:3089: llvm::SDOperand::SelectionDAGLegalize::PromoteOp(llvm::SDOperand): Assertion `0 && "Do not know how to promote this operator!"' failed. For ; ModuleID = 'bugpoint-reduced-simplified.bc' target datalayout = "e-p:64:64" target endian = little target pointersize = 64 target triple = "alphaev6-unknown-linux-gnu" implementation ; Functions: define void %main() { entry: br label %bb bb: ; preds = %bb, %entry br i1 false, label %bb26, label %bb bb19: ; preds = %bb26 ret void bb26: ; preds = %bb br i1 false, label %bb30, label %bb19 bb30: ; preds = %bb26 br label %bb45 bb45: ; preds = %bb45, %bb30 %V.0 = phi <8 x i16> [ %tmp42, %bb45 ], [ zeroinitializer, %bb30 ] ; <<8 x i16>> [#uses=1] %tmp42 = mul <8 x i16> zeroinitializer, %V.0 ; <<8 x i16>> [#uses=1] br label %bb45 } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Jan 25 18:54:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 25 Jan 2007 18:54:43 -0600 Subject: [LLVMbugs] [Bug 1133] NEW: x86 inline asm crash (llc) Message-ID: <200701260054.l0Q0shqn013900@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1133 Summary: x86 inline asm crash (llc) Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu the attched reduction is from code that passes regular gcc. llc: SelectionDAGISel.cpp:2622: void llvm::SelectionDAGLowering::visitInlineAsm(llvm::CallInst&): Assertion `(NumOps & 7) == 2 && "Skipped past definitions?"' failed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 02:32:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 02:32:56 -0600 Subject: [LLVMbugs] [Bug 645] ambiguity parsing local vs global symbols Message-ID: <200701260832.l0Q8WutS032190@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=645 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 02:34:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 02:34:09 -0600 Subject: [LLVMbugs] [Bug 761] TargetData needs to be parameterized better in the .bc/.ll files Message-ID: <200701260834.l0Q8Y9lf032243@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=761 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-26 02:34 ------- This has now been implemented. The patches for the code have been applied. The patch for llvm-gcc4 now needs to be committed. The patch for the testsuite was partially applied. In some cases it was preferable to leave the old endian/pointersize specifications in and let llvm-upgrade upgrade them. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 02:35:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 02:35:49 -0600 Subject: [LLVMbugs] [Bug 1082] llvm-upgrade can fail on names with unspecified or defaulted type Message-ID: <200701260835.l0Q8Zn3T032297@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1082 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-26 02:35 ------- With the llvm-upgrade, this is now fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043254.html The XFAIL from the test case has been removed and it passes. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 02:38:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 02:38:05 -0600 Subject: [LLVMbugs] [Bug 1122] getelementptr gets confused with nested type Message-ID: <200701260838.l0Q8c5Wl032359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1122 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-26 02:38 ------- The recent re-write of llvm-upgrade should have fixed this but. It no longer gives the error and produces this output for the attached test case: ; ModuleID = '' %RPyString = type { i32, %arraytype_Char_1 } %arraytype_Char_1 = type { i32, [0 x i8] } %structtype_rpy_string = type { i32, %arraytype_Char_1 } implementation ; Functions: define i8* @RPyString_AsString(%RPyString* %structstring) { %source1ptr = getelementptr %RPyString* %structstring, i32 0, i32 1, i32 1 ; <[0 x i8]*> [#uses=1] %source1 = bitcast [0 x i8]* %source1ptr to i8* ; [#uses=1] ret i8* %source1 } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 03:27:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 03:27:00 -0600 Subject: [LLVMbugs] [Bug 808] configure fails on NetBSD Message-ID: <200701260927.l0Q9R0Mm000737@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=808 ich at christoph-bauer.net changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From ich at christoph-bauer.net 2007-01-26 03:26 ------- Hi Reid, sorry, i'm quite busy at the moment. I did a quick check this morning and everything looks good. Thanks a lot for your work. Christoph ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 11:36:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 11:36:30 -0600 Subject: [LLVMbugs] [Bug 761] TargetData needs to be parameterized better in the .bc/.ll files Message-ID: <200701261736.l0QHaUBZ008941@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=761 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From rspencer at x10sys.com 2007-01-26 11:36 ------- Owen, You're still on the hook for the bytecode format changes in the documentation. Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 18:17:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 18:17:19 -0600 Subject: [LLVMbugs] [Bug 1134] NEW: We should merge ARMConstantInslands and PPCBranchSelect into one generic pass Message-ID: <200701270017.l0R0HJ6U008836@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1134 Summary: We should merge ARMConstantInslands and PPCBranchSelect into one generic pass Product: libraries Version: 1.0 Platform: Macintosh OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: natebegeman at mac.com ARMConstantInslands and the PPC branch selector pass both do branch shortening and more. It would be nice for this to be target independent. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Jan 26 19:29:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 26 Jan 2007 19:29:50 -0600 Subject: [LLVMbugs] [Bug 1135] NEW: Change "+" to "|" may cause problem in loop dependence analysis Message-ID: <200701270129.l0R1TofE010083@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1135 Summary: Change "+" to "|" may cause problem in loop dependence analysis Product: new-bugs Version: unspecified Platform: All OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: wjiang at cs.ucla.edu This is actually not a bug, but it may bring problem to further optimizations. The TOT llvm now can change the addition(+) to or(|) if one operation is a constant and satisfies some condition ( around InstCombiner.cpp: 1107) . But if somebody want to analyze the dependence inside one loop (most of the dependency techniques don't work on bit operations), it may cause some problems. For exmaple: for ( i=0; i http://llvm.org/bugs/show_bug.cgi?id=761 resistor at mac.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From resistor at mac.com 2007-01-27 13:24 ------- This is done. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043374.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 27 14:58:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 27 Jan 2007 14:58:12 -0600 Subject: [LLVMbugs] [Bug 1136] NEW: Rename Function::isExternal() -> isDeclaration() Message-ID: <200701272058.l0RKwC63005309@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1136 Summary: Rename Function::isExternal() -> isDeclaration() Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com For people learning LLVM, the isExternal() method on Function can be confusing. It gets confused with isExternalLinkage(). To make things more clear in the code, I suggest we rename this to isDeclaration() or isPrototype(). This keeps it clear from using the word External which can cause the confusion. There's no functional change implied by this, just a name change. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 27 16:57:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 27 Jan 2007 16:57:09 -0600 Subject: [LLVMbugs] [Bug 1137] NEW: llvm-upgrade can't do uint %x and int %x together. Message-ID: <200701272257.l0RMv9Ji007101@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1137 Summary: llvm-upgrade can't do uint %x and int %x together. Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: patil.nikhil at gmail.com llvm-upgrade cannot upgrade the following program: ; ModuleID = 'b.c' target datalayout = "e-p:32:32" target endian = little target pointersize = 32 target triple = "i686-pc-linux-gnu" implementation ; Functions: void %main() { entry: %tmp = alloca uint, align 4 ; [#uses=1] %tmp = alloca int, align 4 ; [#uses=1] "alloca point" = cast int 0 to int ; [#uses=0] store uint 1, uint* %tmp store int 2, int* %tmp br label %return return: ; preds = %entry ret void } It fails with: Redefinition of value named 'tmp' in the 'i32 *' type plane (The above program was produced by compiling the following with the llvm-gcc4- 1.9-x86-FC5 binary: void main() { { unsigned int tmp = 1; } { int tmp = 2; } } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Jan 27 18:53:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 27 Jan 2007 18:53:26 -0600 Subject: [LLVMbugs] [Bug 1137] llvm-upgrade can't do uint %x and int %x together. Message-ID: <200701280053.l0S0rQDw009895@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1137 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-27 18:53 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043382.html Test case placed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043381.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 02:24:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 02:24:55 -0600 Subject: [LLVMbugs] [Bug 1138] NEW: CBE can't handle programs with llvm.memcpy.i64 and llvm.memcpy.i32 Message-ID: <200701280824.l0S8OtJY025677@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1138 Summary: CBE can't handle programs with llvm.memcpy.i64 and llvm.memcpy.i32 Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The CBE is emitting: unsigned char *memcpy(unsigned char *, unsigned char *, unsigned int ); unsigned char *memcpy(unsigned char *, unsigned char *, unsigned long long ); which causes the C compiler to barf. This prevents burg from working with the CBE and also prevents bugpoint from working, which doesn't allow us to track down why jit/llc also fail. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 14:11:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 14:11:55 -0600 Subject: [LLVMbugs] [Bug 1139] NEW: scalarrepl miscompile Message-ID: <200701282011.l0SKBtbD030481@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1139 Summary: scalarrepl miscompile Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca Bugpoint reduced this out of McCat/09-vor. I'm not sure where the miscompile is yet. Attachment to come ... ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 14:47:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 14:47:36 -0600 Subject: [LLVMbugs] [Bug 1140] NEW: "LLC Compile" reports incorrectly Message-ID: <200701282047.l0SKlaF4031144@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1140 Summary: "LLC Compile" reports incorrectly Product: Test Suite Version: trunk Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: Nightly Tester AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com If the run of the llc generated program (Output/*.llc) fails then the nightly tester reports the "LLC Compile" time as "*". This is incorrect. If the compile finished correctly, it should report the compile time, not "*", even if the resulting compiled program failed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 17:35:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 17:35:13 -0600 Subject: [LLVMbugs] [Bug 1141] NEW: Instcombine makes wrong array size calculation Message-ID: <200701282335.l0SNZDvR001488@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1141 Summary: Instcombine makes wrong array size calculation Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com It seems the recent changes to TargetData have affected the computation of array sizes in PromoteCastOfAllocation. This is the problem that is killing the recent nightly tests. I'm filing this because I didn't have time to sufficiently debug it. The error is somewhere in this code: // Get the type really allocated and the type casted to. const Type *AllocElTy = AI.getAllocatedType(); const Type *CastElTy = PTy->getElementType(); if (!AllocElTy->isSized() || !CastElTy->isSized()) return 0; unsigned AllocElTyAlign = TD->getTypeAlignmentABI(AllocElTy); unsigned CastElTyAlign = TD->getTypeAlignmentABI(CastElTy); if (CastElTyAlign < AllocElTyAlign) return 0; // If the allocation has multiple uses, only promote it if we are strictly // increasing the alignment of the resultant allocation. If we keep it the // same, we open the door to infinite loops of various kinds. if (!AI.hasOneUse() && CastElTyAlign == AllocElTyAlign) return 0; uint64_t AllocElTySize = TD->getTypeSize(AllocElTy); uint64_t CastElTySize = TD->getTypeSize(CastElTy); if (CastElTySize == 0 || AllocElTySize == 0) return 0; // See if we can satisfy the modulus by pulling a scale out of the array // size argument. unsigned ArraySizeScale, ArrayOffset; Value *NumElements = // See if the array size is a decomposable linear expr. DecomposeSimpleLinearExpr(AI.getOperand(0), ArraySizeScale, ArrayOffset); // If we can now satisfy the modulus, by using a non-1 scale, we really can // do the xform. if ((AllocElTySize*ArraySizeScale) % CastElTySize != 0 || (AllocElTySize*ArrayOffset ) % CastElTySize != 0) return 0; unsigned Scale = (AllocElTySize*ArraySizeScale)/CastElTySize; ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 18:10:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 18:10:28 -0600 Subject: [LLVMbugs] [Bug 1141] Instcombine makes wrong array size calculation Message-ID: <200701290010.l0T0ASMN002279@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1141 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-01-28 18:10 ------- This isn't a bug. Because your module doesn't have a target data line, LLVM is defaulting to 8-byte pointers. We need to figure out why you don't have a targetdata line. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Jan 28 18:22:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 28 Jan 2007 18:22:18 -0600 Subject: [LLVMbugs] [Bug 1139] linker drops target data info Message-ID: <200701290022.l0T0MIh9002576@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1139 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Scalar Optimizations |Linker Keywords| |miscompilation, regression Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-28 18:22 ------- This is fallout from the recent target data changes. Fix here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070122/043423.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 11:59:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 11:59:57 -0600 Subject: [LLVMbugs] [Bug 1142] NEW: llvm-upgrade broken on test/Transforms/SimplifyCFG/2006-10-29-InvokeCrash.ll Message-ID: <200701291759.l0THxvkP001027@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1142 Summary: llvm-upgrade broken on test/Transforms/SimplifyCFG/2006- 10-29-InvokeCrash.ll Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org $ llvm-upgrade < 2006-10-29-InvokeCrash.ll > /dev/null llvm-upgrade: :222: error: Reference to an invalid definition: '_ZNK10QTextFrame11frameFormatEv' of type 'void (sret { { { opaque * }, i32 } } *, { { { i32 (...) * *, { i32 (...) * *, \4 *, \4 *, { { { { { i32 }, i32, i32, i32, i8, [1 x i8 *] } * } } }, i8, [3 x i8], i32, i32 } * } } } *) *' ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 12:12:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 12:12:31 -0600 Subject: [LLVMbugs] [Bug 1143] NEW: Missed LICM load/store hoisting opportunities Message-ID: <200701291812.l0TICVU0001226@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1143 Summary: Missed LICM load/store hoisting opportunities Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Global Analyses AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider: struct a { int X, Y; }; struct a arr[100]; void foo (unsigned b) { unsigned i; for (i = 0; i < 100; i++) { arr[b+8].X += i; arr[b+9].X += i; } } void foo2(struct a *A, unsigned b) { unsigned i; for (i = 0; i < 100; i++) { A[b].X += i; A[b+1].Y += i; } } When I run: $ llvm-gcc t.c -S -emit-llvm -O3 -o - I see that the load/store accesses in the first loop are not hoisted. This is due to BasicAA not determining that the pointer can't alias. In the second loop, some sort of phase ordering issue prevents one of the accesses from being hoisted. If I run -licm again on the output, both are hoisted. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 17:11:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 17:11:17 -0600 Subject: [LLVMbugs] [Bug 1138] CBE can't handle programs with llvm.memcpy.i64 and llvm.memcpy.i32 Message-ID: <200701292311.l0TNBHXx013139@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1138 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-01-29 17:11 ------- A proper fix for this was made with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043434.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043435.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043436.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043437.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043438.html This endows IntrinsicLowering with TargetData and uses the IntPtrTy to get the right prototype for memcpy/memmove/memset given the intended target. It also adjusts CBE and LLI to pass the TargetData to IntrinsicLowering. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 17:18:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 17:18:58 -0600 Subject: [LLVMbugs] [Bug 680] Simplify the TargetLowering interface Message-ID: <200701292318.l0TNIwOi013326@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=680 natebegeman at mac.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From natebegeman at mac.com 2007-01-29 17:18 ------- http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043469.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 17:38:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 17:38:33 -0600 Subject: [LLVMbugs] [Bug 1144] NEW: bad codegen Message-ID: <200701292338.l0TNcXbQ013973@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1144 Summary: bad codegen Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca bugpoint reduced this out of Shootout-C++/hash2. It works if I compile it with the CBE, but fails if I compile it with the x86 backend. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Jan 29 17:59:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 29 Jan 2007 17:59:05 -0600 Subject: [LLVMbugs] [Bug 839] inline asm doesn't support multiple alternatives Message-ID: <200701292359.l0TNx5WG014678@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=839 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-01-29 17:59 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043487.html Testcases here: test/CodeGen/X86/2007-01-29-InlineAsm-ir.ll test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 30 11:23:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 30 Jan 2007 11:23:14 -0600 Subject: [LLVMbugs] [Bug 1145] NEW: Support nothrow attribute on function calls Message-ID: <200701301723.l0UHNEiX010874@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1145 Summary: Support nothrow attribute on function calls Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org With exception table support under development, it will soon be useful to know if a function can throw or not. The front-end has this info for some functions (those marked throw()) and the prune-eh pass computes this info, but there is no way to record this info in the IR for the code generator to use. It should be straight-forward to add an attribute to functions to indicate that they cannot throw. If the front-end sets this, it will make the prune-eh pass more effective (deleting more landing pads) and the code generator can then start using it. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 30 11:25:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 30 Jan 2007 11:25:42 -0600 Subject: [LLVMbugs] [Bug 1146] NEW: Move function attributes out of FunctionType into calls and functions Message-ID: <200701301725.l0UHPgEZ010938@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1146 Summary: Move function attributes out of FunctionType into calls and functions 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: sabre at nondot.org Function attributes are currently a property of the FunctionType. This makes it very painful to modify the attributes of a function after the function is created (because the type is immutable). It would be much better to put the attributes on the call/invoke instruction and on the Function object itself. The trick to this is to find a memory-efficient way to represent this. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 30 14:12:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 30 Jan 2007 14:12:22 -0600 Subject: [LLVMbugs] [Bug 1136] Rename Function::isExternal() -> isDeclaration() Message-ID: <200701302012.l0UKCMV7019547@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1136 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-30 14:12 ------- Fixed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 30 16:14:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 30 Jan 2007 16:14:27 -0600 Subject: [LLVMbugs] [Bug 1147] NEW: test/ExecutionEngine/test-fp.ll sometimes fails Message-ID: <200701302214.l0UMER6w025133@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1147 Summary: test/ExecutionEngine/test-fp.ll sometimes fails Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Interpreter AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com There's some kind of non-determinism going on with this test case. Sometimes it passes and sometimes it fails. Attempting to debug it yields a stack like: (gdb) where #0 0x00cb40b7 in ?? () #1 0x0935d190 in ?? () #2 0x00000004 in ?? () #3 0x08496fe8 in llvm::ExecutionEngine::runFunctionAsMain (this=0x935d188, Fn=0x935eea0, argv=@0x881ab3c, envp=0xbf9dbe60) at ExecutionEngine.cpp:247 #4 0x0837d5f7 in main (argc=2, argv=0xbf9dbe54, envp=0xbf9dbe60) at lli.cpp:116 But attempting to step through it from ExecutionEngine.cpp allows it to succeed. I'll try valgrind next. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Jan 30 23:02:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 30 Jan 2007 23:02:18 -0600 Subject: [LLVMbugs] [Bug 1148] NEW: Code quality issue with multiple inheritance Message-ID: <200701310502.l0V52IIk001975@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1148 Summary: Code quality issue with multiple inheritance Product: tools Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-g++ AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider: struct base { virtual ~base(); int X; }; struct A { virtual ~A(); int B; }; struct Derived : public A, public base { int field; }; int foo(base *B) { return static_cast(B)->field; } The static_cast in foo is actually compiled to the equivalent of "B ? B-8 : 0", producing this llvm code: define i32 @_Z3fooP4base(%struct.A* %B) { entry: %tmp = icmp eq %struct.A* %B, null ; [#uses=1] br i1 %tmp, label %cond_next, label %cond_true cond_true: ; preds = %entry %B = bitcast %struct.A* %B to i8* ; [#uses=1] %ctg2 = getelementptr i8* %B, i32 -8 ; [#uses=1] %tmp = bitcast i8* %ctg2 to %struct.Derived* ; <%struct.Derived*> [#uses=1] %tmp36 = getelementptr %struct.Derived* %tmp, i32 0, i32 2 ; [#uses=1] %tmp47 = load i32* %tmp36 ; [#uses=1] ret i32 %tmp47 cond_next: ; preds = %entry %tmp4 = load i32* inttoptr (i32 16 to i32*) ; [#uses=1] ret i32 %tmp4 } However, because the pointer is loaded, we know that B could not have been a null pointer. It's unclear how to handle this in LLVM, loading from a constant address is legal. This has to be handled on the tree level somehow. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 01:27:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 01:27:56 -0600 Subject: [LLVMbugs] [Bug 1142] llvm-upgrade fails to upgrade csretcc calls to functions declared without csretcc Message-ID: <200701310727.l0V7RuLs005086@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1142 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-01-31 01:27 ------- This was fixed yesterday. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 08:03:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 08:03:20 -0600 Subject: [LLVMbugs] [Bug 1149] NEW: stage1 fails to build the stage2 on x86 Message-ID: <200701311403.l0VE3KGp022255@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1149 Summary: stage1 fails to build the stage2 on x86 Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com The error: /home/laurov/llvm/llvm-gcc/build/prev-gcc/xgcc -B/home/laurov/llvm/llvm-gcc/build/prev-gcc/ -B/usr/local/i686-pc-linux-gnu/bin/ -c -O2 -g -fomit-frame-pointer -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wold-style-definition -Werror -fno-common -DHAVE_CONFIG_H -DENABLE_LLVM -D__STDC_LIMIT_MACROS -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include -I/home/laurov/llvm/llvm/include -I/home/laurov/llvm/llvm/build//include ../../gcc/c-common.c -o c-common.o cc1: warnings being treated as errors ../../gcc/c-common.c: In function ???iasm_extra_clobbers???: ../../gcc/c-common.c:6931: warning: comparison of unsigned expression < 0 is always false make[2]: *** [c-common.o] Error 1 make[2]: Leaving directory `/home/laurov/llvm/llvm-gcc/build/stage2-gcc' make[1]: *** [all-stage2-gcc] Error 2 make[1]: Leaving directory `/home/laurov/llvm/llvm-gcc/build' make: *** [stage2-bubble] Error 2 ****************************************** The line 6931 of c-common.c: for (i=0; i < sizeof (db) / sizeof(db[0]) - 1; ++i) **************************************** Configure: ../configure --enable-llvm=/home/laurov/llvm/llvm/build/ --enable-bootstrap --enable-checking --enable-languages=c,c++ --program-prefix=llvm- --disable-threads --disable-shared ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 08:22:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 08:22:10 -0600 Subject: [LLVMbugs] [Bug 1150] NEW: stage2 compiles c++ files with the host c++ compiler Message-ID: <200701311422.l0VEMAZk022672@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1150 Summary: stage2 compiles c++ files with the host c++ compiler Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com Stage2 uses c++ instead of /home/laurov/llvm/llvm-gcc/build/prev-gcc/cpp. make[2]: Entering directory `/home/laurov/llvm/llvm-gcc/build/stage2-gcc' c++ -c -O2 -g -fomit-frame-pointer -DIN_GCC -W -Wall -Wwrite-strings -pedantic -Wno-long-long -Wno-variadic-macros -Werror -fno-common -DHAVE_CONFIG_H -Wno-unused -DTARGET_NAME=\"i686-pc-linux-gnu\" -DENABLE_LLVM -D__STDC_LIMIT_MACROS -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include -I/home/laurov/llvm/llvm/include -I/home/laurov/llvm/llvm/build//include ../../gcc/llvm-main.cpp -o llvm-main.o /home/laurov/llvm/llvm-gcc/build/prev-gcc/xgcc -B/home/laurov/llvm/llvm-gcc/build/prev-gcc/ -B/usr/local/i686-pc-linux-gnu/bin/ -O2 -g -fomit-frame-pointer -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wold-style-definition -Werror -fno-common -DHAVE_CONFIG_H -DENABLE_LLVM -D__STDC_LIMIT_MACROS -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include -I/home/laurov/llvm/llvm/include -I/home/laurov/llvm/llvm/build//include -c insn-attrtab.c \ -o insn-attrtab.o ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 12:32:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 12:32:48 -0600 Subject: [LLVMbugs] [Bug 1131] llvm-gcc doesn't handle struct-returning stdcall functions right Message-ID: <200701311832.l0VIWmEF021102@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1131 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-01-31 12:32 ------- Reid recently committed the llvm-upgrade pieces, this is done. Thanks guys! ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 13:00:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 13:00:54 -0600 Subject: [LLVMbugs] [Bug 1151] NEW: llvm-upgrade fails test/Assembler/2007-01-22-UpgradeTypeMapInvalidMemory.ll Message-ID: <200701311900.l0VJ0so1023386@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1151 Summary: llvm-upgrade fails test/Assembler/2007-01-22- UpgradeTypeMapInvalidMemory.ll Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org test/Assembler/2007-01-22-UpgradeTypeMapInvalidMemory.ll is failing on mainline ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 16:38:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 16:38:24 -0600 Subject: [LLVMbugs] [Bug 1152] NEW: Move gccas/gccld functionality to opt Message-ID: <200701312238.l0VMcOC9002398@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1152 Summary: Move gccas/gccld functionality to opt Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: gccas AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com We currently have multiple ways of invoking optimizations in llvm, via gccas, gccld, and opt. If we moved gccas/gccld functionality into opt we could save some footprint size and a significant amount of link time during builds (gccas and gccld are among the largest tools, on the same scale as opt, lli and llc). To accomplish this we would need to: 1. Add an option such as -std-compile-opts to run the gccas "set" of optimizations. 2. Add an option such as -std-link-opts to run the gccld set of optimizations. 3. Add a -verify-each option to opt to enabled the "verify after each pass" feature of gccas/gccld. 4. Adjust llvm-test to use opt and llvm-link with these new options instead of running gccas and gccld. Anyone see any roadblocks on doing this? Anyone think its not a good idea? ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 19:44:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 19:44:58 -0600 Subject: [LLVMbugs] [Bug 1153] NEW: Move gccld optimization functionality to opt Message-ID: <200702010144.l111iwSw006288@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1153 Summary: Move gccld optimization functionality to opt Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: gccld AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com This is a sister bug to bug 1152. The optimizations that gccld performs should be available with a single command line option such as -std-lto-passes. The linking capability of gccld will not be moved as it is redundant with llvm-link and llvm-ld. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Jan 31 22:14:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 31 Jan 2007 22:14:01 -0600 Subject: [LLVMbugs] [Bug 1154] NEW: LICM missing hoist of calls do to missing loop rotation Message-ID: <200702010414.l114E1Pc009204@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1154 Summary: LICM missing hoist of calls do to missing loop rotation Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: enhancement Priority: P2 Component: Loop Optimizer AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org In this example code: #include #include #define TESTSTRING2 "Dot. date. datum. 123. Some more doubtful demonstration dummy data." int stringSearch_Clib (long count) { int i, c = 0;long j;char * p; char b[sizeof(TESTSTRING2)]; strcpy (b, TESTSTRING2); for (j=0; j < count; j++) { for (c=i=0; i < 250; i++) { if (NULL != (p = strstr (b, "ummy"))) c += (int) (p - b); if (NULL != (p = strstr (b, " data"))) c += (int) (p - b); c += (int) strcspn (b, "by"); } } return c; } The calls to strstr and strcspn are loop invariant and should be hoisted. They aren't though, because the inner loop isn't getting rotated. This can easily be seen with: $ llvm-gcc t.c -emit-llvm -O3 -o - -S | llvm-as | opt -print-cfg and then viewing cfg.stringSearch_Clib.dot. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.