From bugzilla-daemon at cs.uiuc.edu Tue May 1 07:28:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 May 2007 07:28:29 -0500 Subject: [LLVMbugs] [Bug 1375] NEW: llvm.returnaddress intrinsic is not handled correctly Message-ID: <200705011228.l41CSTaV021294@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1375 Summary: llvm.returnaddress intrinsic is not handled correctly Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Consider the following code: #include struct _Unwind_Context { void *ra; }; void bar(struct _Unwind_Context*); void foo(struct _Unwind_Context* context) { void *ra = __builtin_extract_return_addr (__builtin_return_address (0)); memset (context, 0, sizeof (struct _Unwind_Context)); context->ra = ra; bar(context); } llvm-gcc produces: %tmp = call i8* @llvm.returnaddress( i32 0 ) ; [#uses=0] store i8* null, i8** %ra which is definitely wrong. This is one small bit needed for EH support in llvm-gcc. ------- 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 May 1 10:05:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 May 2007 10:05:31 -0500 Subject: [LLVMbugs] [Bug 1376] NEW: Section order should be tweaked Message-ID: <200705011505.l41F5VB3023887@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1376 Summary: Section order should be tweaked 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 Consider we have function with jump tables. Currently we're doing something like this: .text ... function body .. ret .section .rodata ... jump table entries ... .size funcname, .-funcname Note, that .size directive is actually located in rodata section and "." refers to .rodata section too, so size is definitely incorrect. Ok. This doesn't make things too much bad. Consider we have eh enabled. We have: .text eh_func_begin: ... function body ... ret .section .rodata ... jump table entries ... .size funcname, .-funcname eh_func_end: .section eh_frame .... Note, eh_func_end is in *rodata* section. eh_frame should contain misc. offsets, involving eh_func_begin and eh_func_end. Surely this information will be invalid. Even more, as blames, because it cannot compute cross-section offsets. The solution seems to be easy: just emit JT info in the late end, but I'm little bit worried, because I saw the following comment in X86ATTAsmPrinter.cpp: // Print out jump tables referenced by the function. // Mac OS X requires that the jump table follow the function, so that the jump // table is part of the same atom that the function is in. EmitJumpTableInfo(MF.getJumpTableInfo(), MF); if (TAI->hasDotTypeDotSizeDirective()) O << "\t.size " << CurrentFnName << ", .-" << CurrentFnName << "\n"; if (Subtarget->isTargetDarwin() || Subtarget->isTargetELF() || Subtarget->isTargetCygMing()) { // Emit post-function debug information. DW.EndFunction(); } It seems for me, that EmitJumpTableInfo() should be moved to the late end, but what does that comment mean" What is "atom" for MacOS X? We really need someone with darwin knowledge to resolve 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 Tue May 1 16:51:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 May 2007 16:51:13 -0500 Subject: [LLVMbugs] [Bug 1377] NEW: bad '.space' constant emitted building unwind-dw2.c Message-ID: <200705012151.l41LpDmN032753@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1377 Summary: bad '.space' constant emitted building unwind-dw2.c Product: tools Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: blocker Priority: P2 Component: llc AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org Building llvm-gcc on powerpc-apple-darwin8.8.0 with sources synced from llvm head at or about 2pm PDT. Compiler output is: /Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/gcc/xgcc -B/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/gcc/ -B/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/powerpc-apple-darwin8.9.0/bin/ -B/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/powerpc-apple-darwin8.9.0/lib/ -isystem /Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/powerpc-apple-darwin8.9.0/include -isystem /Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/powerpc-apple-darwin8.9.0/sys-include -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -Wa,-force_cpusubtype_ALL -pipe -mmacosx-version-min=10.4 -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../../gcc -I../../../gcc/. -I../../../gcc/../include -I../../../gcc/../libcpp/include -I/Users/scottm/play/llvm-spu/include -I/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../../llvm-spu/obj/powerpc-apple-darwin8.8.0/include -fexceptions -c ../../../gcc/unwind-dw2.c -o libgcc/./unwind-dw2.o {standard input}:2775:Repeat < 0, .space ignored Looking at the assembly source from xgcc, the offending line looks like: .data .align 4 _once_regsizes.6143: ; 'once_regsizes.6143' .long 816954554 .space 16 .space 18446744073709551568 llc produces the following (which is still a bad constant): .data .align 4 _once_regsizes.6143: ; 'once_regsizes.6143' .long 816954554 .space 16 .space 17179869136 ------- 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 May 1 18:45:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 May 2007 18:45:37 -0500 Subject: [LLVMbugs] [Bug 1378] NEW: Vector initialization causes C++ front-end ICE Message-ID: <200705012345.l41NjbfT003288@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1378 Summary: Vector initialization causes C++ front-end ICE Product: new-bugs Version: unspecified Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: christopher.lamb at gmail.com llvm/trunk/lib/VMCore/Constants.cpp:1201: failed assertion `!V.empty() && "Cannot infer type if V is empty"' ../../splat_g++_ice.cpp: In function 'float __vector__ foo(float)': ../../splat_g++_ice.cpp:11: internal compiler error: Abort trap 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 Tue May 1 21:16:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 May 2007 21:16:03 -0500 Subject: [LLVMbugs] [Bug 1306] LiveVariables improperly handles killing aliased registers Message-ID: <200705020216.l422G3mY006185@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1306 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-05-01 21:16 ------- I like to leave bugs open to give the impression that I am busy. ------- 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 May 2 11:03:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 May 2007 11:03:59 -0500 Subject: [LLVMbugs] [Bug 1377] bad '.space' constant emitted building unwind-dw2.c Message-ID: <200705021603.l42G3xHs016921@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1377 scottm at aero.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From scottm at aero.org 2007-05-02 11:03 ------- After this morning's update from HEAD, appears to be 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 Wed May 2 15:15:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 May 2007 15:15:04 -0500 Subject: [LLVMbugs] [Bug 1379] NEW: Need support for eh-related builtins Message-ID: <200705022015.l42KF4gP022090@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1379 Summary: Need support for eh-related builtins Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Runtime used for eh upport (e.g stack unwinding, etc) is full of builtins currently unsupported. We need to implement these builtins (either in FE itself, or in codegenerator) in order to get working exceptions ------- 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 May 2 15:23:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 May 2007 15:23:28 -0500 Subject: [LLVMbugs] [Bug 1380] NEW: EH frames must be emitted in one "atom". Message-ID: <200705022023.l42KNShU022331@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1380 Summary: EH frames must be emitted in one "atom". 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 Currently there are two ways of registering eh function frames during runtime: 1. __register_frame_info call (and friends) in the early beginning (crtbegin.o). 2. Direct eh section inspection ("GLIBC" method). Unfortunately, it seems the __register_frame_info call not always emitted (e.g. crtbegin.o compiled during llvm-gcc compilation without LLVM itself lacks this call). The second approach uses some assumptions. One of the biggest assumption is that entries in eh frame section are sorted according to PC's stored in entries. We can easily achieve this, if we emit information at once in the end (like gcc does). When we're emitting this information in separate "parts" (but in one section, of course), linker can permute "parts" making sort assumption invalid. This is needed for backward compatibility of generated eh information. ------- 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 Wed May 2 16:17:31 2007 From: isanbard at gmail.com (Bill) Date: Wed, 2 May 2007 14:17:31 -0700 Subject: [LLVMbugs] Anyone seeing this? Message-ID: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a@mail.gmail.com> Anyone seeing this failure? FAIL: /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/2007-04-14-EHSelectorCrash.ll for PR1326 Failed with signal(SIGABRT) at line 2 while running: llvm-as < /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/2007-04-14-EHSelectorCrash.ll | llc -enable-eh -march=x86-64 Assertion failed: (false && "Couldn't find the register class"), function getPhysicalRegisterRegClass, file /Volumes/Gir/devel/llvm/llvm.src/lib/CodeGen/SelectionDAG/ScheduleDAG.cpp, line 396. 0 llc 0x0024b5b7 _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cpp_00000000_F63E927415PrintStackTraceEv + 35 1 llc 0x0024b95d _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cpp_00000000_F63E927413SignalHandlerEi + 323 2 libSystem.B.dylib 0x955f0d61 _sigtramp + 55 3 ??? 0xffffffff 0x0 + 4294967295 4 libSystem.B.dylib 0x955f9091 raise + 26 5 libSystem.B.dylib 0x9561fa23 abort + 84 6 libSystem.B.dylib 0x955fe53f __assert_rtn + 101 7 llc 0x0011376d _ZN4llvm11ScheduleDAG13CountOperandsEPNS_6SDNodeE + 609 8 llc 0x00115a59 _ZN4llvm11ScheduleDAG8EmitNodeEPNS_6SDNodeERNS_8DenseMapIS2_jNS_15DenseMapKeyInfoIS2_EEEE + 1967 9 llc 0x0011612e _ZN4llvm11ScheduleDAG12EmitScheduleEv + 594 10 llc 0x00118bd2 _ZNK108_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_CodeGen_SelectionDAG_ScheduleDAGRRList.cpp_00000000_4B0754BF13td_ls_rr_sortclEPKN4llvm5SUnitES4_ + 5278 11 llc 0x0011380a _ZN4llvm11ScheduleDAG3RunEv + 148 12 llc 0x0012d6be _ZN4llvm16SelectionDAGISel18ScheduleAndEmitDAGERNS_12SelectionDAGE + 126 13 llc 0x00088fe2 _ZN4llvm16createX86ISelDagERNS_16X86TargetMachineEb + 16366 14 llc 0x0012df44 _ZN4llvm16SelectionDAGISel17CodeGenAndEmitDAGERNS_12SelectionDAGE + 348 15 llc 0x001423ba _ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockERNS_15MachineFunctionERNS_20FunctionLoweringInfoE + 196 16 llc 0x00143893 _ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE + 275 17 llc 0x00309caf _ZN4llvm17X86TargetLoweringD1Ev + 103 18 llc 0x0020a9a2 _ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE + 328 19 llc 0x0020ac16 _ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE + 110 20 llc 0x0020ad93 _ZN4llvm19FunctionPassManager3runERNS_8FunctionE + 155 21 llc 0x00003359 main + 3227 22 llc 0x00001d47 start + 327 23 llc 0x00001c29 start + 41 It seems new. -bw From evan.cheng at apple.com Wed May 2 16:28:22 2007 From: evan.cheng at apple.com (Evan Cheng) Date: Wed, 2 May 2007 14:28:22 -0700 Subject: [LLVMbugs] Anyone seeing this? In-Reply-To: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a@mail.gmail.com> References: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a@mail.gmail.com> Message-ID: <52BBAF26-800E-46B0-9F3E-D36FBFC4EF7C@apple.com> Not seeing this. It works fine on my X86-64 box. Are you on TOT? Evan On May 2, 2007, at 2:17 PM, Bill wrote: > Anyone seeing this failure? > > FAIL: /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/ > 2007-04-14-EHSelectorCrash.ll > for PR1326 > Failed with signal(SIGABRT) at line 2 > while running: llvm-as < > /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/2007-04-14- > EHSelectorCrash.ll > | llc -enable-eh -march=x86-64 > Assertion failed: (false && "Couldn't find the register class"), > function getPhysicalRegisterRegClass, file > /Volumes/Gir/devel/llvm/llvm.src/lib/CodeGen/SelectionDAG/ > ScheduleDAG.cpp, > line 396. > 0 llc 0x0024b5b7 > _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cp > p_00000000_F63E927415PrintStackTraceEv > + 35 > 1 llc 0x0024b95d > _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cp > p_00000000_F63E927413SignalHandlerEi > + 323 > 2 libSystem.B.dylib 0x955f0d61 _sigtramp + 55 > 3 ??? 0xffffffff 0x0 + 4294967295 > 4 libSystem.B.dylib 0x955f9091 raise + 26 > 5 libSystem.B.dylib 0x9561fa23 abort + 84 > 6 libSystem.B.dylib 0x955fe53f __assert_rtn + 101 > 7 llc 0x0011376d > _ZN4llvm11ScheduleDAG13CountOperandsEPNS_6SDNodeE + 609 > 8 llc 0x00115a59 > _ZN4llvm11ScheduleDAG8EmitNodeEPNS_6SDNodeERNS_8DenseMapIS2_jNS_15Dens > eMapKeyInfoIS2_EEEE > + 1967 > 9 llc 0x0011612e > _ZN4llvm11ScheduleDAG12EmitScheduleEv + 594 > 10 llc 0x00118bd2 > _ZNK108_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_CodeGen_Selecti > onDAG_ScheduleDAGRRList.cpp_00000000_4B0754BF13td_ls_rr_sortclEPKN4llv > m5SUnitES4_ > + 5278 > 11 llc 0x0011380a > _ZN4llvm11ScheduleDAG3RunEv + 148 > 12 llc 0x0012d6be > _ZN4llvm16SelectionDAGISel18ScheduleAndEmitDAGERNS_12SelectionDAGE + > 126 > 13 llc 0x00088fe2 > _ZN4llvm16createX86ISelDagERNS_16X86TargetMachineEb + 16366 > 14 llc 0x0012df44 > _ZN4llvm16SelectionDAGISel17CodeGenAndEmitDAGERNS_12SelectionDAGE + > 348 > 15 llc 0x001423ba > _ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockERNS_15Ma > chineFunctionERNS_20FunctionLoweringInfoE > + 196 > 16 llc 0x00143893 > _ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE + 275 > 17 llc 0x00309caf > _ZN4llvm17X86TargetLoweringD1Ev + 103 > 18 llc 0x0020a9a2 > _ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE + 328 > 19 llc 0x0020ac16 > _ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE + 110 > 20 llc 0x0020ad93 > _ZN4llvm19FunctionPassManager3runERNS_8FunctionE + 155 > 21 llc 0x00003359 main + 3227 > 22 llc 0x00001d47 start + 327 > 23 llc 0x00001c29 start + 41 > > It seems new. > > -bw > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From isanbard at gmail.com Wed May 2 16:38:58 2007 From: isanbard at gmail.com (Bill) Date: Wed, 2 May 2007 14:38:58 -0700 Subject: [LLVMbugs] Anyone seeing this? In-Reply-To: <52BBAF26-800E-46B0-9F3E-D36FBFC4EF7C@apple.com> References: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a@mail.gmail.com> <52BBAF26-800E-46B0-9F3E-D36FBFC4EF7C@apple.com> Message-ID: <16e5fdf90705021438p22b535b5o888ad5a7a6d16222@mail.gmail.com> Yeah. I even got it on a clean build. -bw On 5/2/07, Evan Cheng wrote: > Not seeing this. It works fine on my X86-64 box. Are you on TOT? > > Evan > > On May 2, 2007, at 2:17 PM, Bill wrote: > > > Anyone seeing this failure? > > > > FAIL: /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/ > > 2007-04-14-EHSelectorCrash.ll > > for PR1326 > > Failed with signal(SIGABRT) at line 2 > > while running: llvm-as < > > /Volumes/Gir/devel/llvm/llvm.src/test/CodeGen/Generic/2007-04-14- > > EHSelectorCrash.ll > > | llc -enable-eh -march=x86-64 > > Assertion failed: (false && "Couldn't find the register class"), > > function getPhysicalRegisterRegClass, file > > /Volumes/Gir/devel/llvm/llvm.src/lib/CodeGen/SelectionDAG/ > > ScheduleDAG.cpp, > > line 396. > > 0 llc 0x0024b5b7 > > _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cp > > p_00000000_F63E927415PrintStackTraceEv > > + 35 > > 1 llc 0x0024b95d > > _ZN84_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_System_Signals.cp > > p_00000000_F63E927413SignalHandlerEi > > + 323 > > 2 libSystem.B.dylib 0x955f0d61 _sigtramp + 55 > > 3 ??? 0xffffffff 0x0 + 4294967295 > > 4 libSystem.B.dylib 0x955f9091 raise + 26 > > 5 libSystem.B.dylib 0x9561fa23 abort + 84 > > 6 libSystem.B.dylib 0x955fe53f __assert_rtn + 101 > > 7 llc 0x0011376d > > _ZN4llvm11ScheduleDAG13CountOperandsEPNS_6SDNodeE + 609 > > 8 llc 0x00115a59 > > _ZN4llvm11ScheduleDAG8EmitNodeEPNS_6SDNodeERNS_8DenseMapIS2_jNS_15Dens > > eMapKeyInfoIS2_EEEE > > + 1967 > > 9 llc 0x0011612e > > _ZN4llvm11ScheduleDAG12EmitScheduleEv + 594 > > 10 llc 0x00118bd2 > > _ZNK108_GLOBAL__N__Volumes_Gir_devel_llvm_llvm.src_lib_CodeGen_Selecti > > onDAG_ScheduleDAGRRList.cpp_00000000_4B0754BF13td_ls_rr_sortclEPKN4llv > > m5SUnitES4_ > > + 5278 > > 11 llc 0x0011380a > > _ZN4llvm11ScheduleDAG3RunEv + 148 > > 12 llc 0x0012d6be > > _ZN4llvm16SelectionDAGISel18ScheduleAndEmitDAGERNS_12SelectionDAGE + > > 126 > > 13 llc 0x00088fe2 > > _ZN4llvm16createX86ISelDagERNS_16X86TargetMachineEb + 16366 > > 14 llc 0x0012df44 > > _ZN4llvm16SelectionDAGISel17CodeGenAndEmitDAGERNS_12SelectionDAGE + > > 348 > > 15 llc 0x001423ba > > _ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockERNS_15Ma > > chineFunctionERNS_20FunctionLoweringInfoE > > + 196 > > 16 llc 0x00143893 > > _ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE + 275 > > 17 llc 0x00309caf > > _ZN4llvm17X86TargetLoweringD1Ev + 103 > > 18 llc 0x0020a9a2 > > _ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE + 328 > > 19 llc 0x0020ac16 > > _ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE + 110 > > 20 llc 0x0020ad93 > > _ZN4llvm19FunctionPassManager3runERNS_8FunctionE + 155 > > 21 llc 0x00003359 main + 3227 > > 22 llc 0x00001d47 start + 327 > > 23 llc 0x00001c29 start + 41 > > > > It seems new. > > > > -bw > > _______________________________________________ > > LLVMbugs mailing list > > LLVMbugs at cs.uiuc.edu > > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs > > From isanbard at gmail.com Wed May 2 16:40:22 2007 From: isanbard at gmail.com (Bill) Date: Wed, 2 May 2007 14:40:22 -0700 Subject: [LLVMbugs] Another Failure Message-ID: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> Anyone seeing this? Running /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/dg.exp ... FAIL: /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll for PR1099 Failed with exit(1) at line 1 while running: llvm-as < /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll | llc -march=c | grep {return ((((llvm_cbe_tmp2 == llvm_cbe_b_0_0_val)) ? (1) : (0)))} child process exited abnormally -bw From rspencer at reidspencer.com Wed May 2 16:46:42 2007 From: rspencer at reidspencer.com (Reid Spencer) Date: Wed, 02 May 2007 17:46:42 -0400 Subject: [LLVMbugs] Another Failure In-Reply-To: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> References: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> Message-ID: On Wed, 2 May 2007 14:40:22 -0700 Bill wrote: >Anyone seeing this? > >Running /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/dg.exp ... >FAIL: /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll >for PR1099 >Failed with exit(1) at line 1 >while running: llvm-as < >/Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll >| llc -march=c | grep {return ((((llvm_cbe_tmp2 == >llvm_cbe_b_0_0_val)) ? (1) : (0)))} >child process exited abnormally No, this didn't fail for me. I changed some sign extension stuff yesterday. What platform? Reid. > > >-bw >_______________________________________________ >LLVMbugs mailing list >LLVMbugs at cs.uiuc.edu >http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From dpatel at apple.com Wed May 2 17:03:00 2007 From: dpatel at apple.com (Devang Patel) Date: Wed, 2 May 2007 15:03:00 -0700 Subject: [LLVMbugs] Another Failure In-Reply-To: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> References: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> Message-ID: <09684823-97FB-4DA5-971D-98B5BCF969EC@apple.com> On May 2, 2007, at 2:40 PM, Bill wrote: > Anyone seeing this? Yes, I am seeing this. - Devang > > > Running /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/ > dg.exp ... > FAIL: /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/ > 2007-01-08-ParamAttr-ICmp.ll > for PR1099 > Failed with exit(1) at line 1 > while running: llvm-as < > /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/ > 2007-01-08-ParamAttr-ICmp.ll > | llc -march=c | grep {return ((((llvm_cbe_tmp2 == > llvm_cbe_b_0_0_val)) ? (1) : (0)))} > child process exited abnormally > > > -bw > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From asl at math.spbu.ru Wed May 2 17:17:42 2007 From: asl at math.spbu.ru (Anton Korobeynikov) Date: Thu, 03 May 2007 02:17:42 +0400 Subject: [LLVMbugs] Anyone seeing this? In-Reply-To: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a.SS3211SS@mail.gmail.com> References: <16e5fdf90705021417q1a4b70f2pd24715f407ab34a.SS3211SS@mail.gmail.com> Message-ID: <1178144262.29965.27.camel@asl.dorms.spbu.ru> > Anyone seeing this failure? Fixed. -- With best regards, Anton Korobeynikov. Faculty of Mathematics & Mechanics, Saint Petersburg State University. From isanbard at gmail.com Wed May 2 18:09:42 2007 From: isanbard at gmail.com (Bill) Date: Wed, 2 May 2007 16:09:42 -0700 Subject: [LLVMbugs] Another Failure In-Reply-To: References: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> Message-ID: <16e5fdf90705021609h452b4143iaee45c3ab3058640@mail.gmail.com> On 5/2/07, Reid Spencer wrote: > On Wed, 2 May 2007 14:40:22 -0700 > Bill wrote: > >Anyone seeing this? > > > >Running /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/dg.exp ... > >FAIL: /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll > >for PR1099 > >Failed with exit(1) at line 1 > >while running: llvm-as < > >/Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll > >| llc -march=c | grep {return ((((llvm_cbe_tmp2 == > >llvm_cbe_b_0_0_val)) ? (1) : (0)))} > >child process exited abnormally > > No, this didn't fail for me. I changed some sign extension stuff yesterday. What platform? > iMac core2. -bw From bugzilla-daemon at cs.uiuc.edu Wed May 2 18:45:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 May 2007 18:45:45 -0500 Subject: [LLVMbugs] [Bug 1381] NEW: Implement the -fmerge-constants optimization Message-ID: <200705022345.l42NjjEO028064@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1381 Summary: Implement the -fmerge-constants optimization Product: new-bugs Version: unspecified Platform: All OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com LLVM should support the -fmerge-constants optimization (as described in the GCC manual) The following message has patch to do this for x86/ELF, though it does not do so under the control of an option. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/048796.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From me22.ca at gmail.com Wed May 2 18:02:17 2007 From: me22.ca at gmail.com (me22) Date: Wed, 2 May 2007 19:02:17 -0400 Subject: [LLVMbugs] Another Failure In-Reply-To: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> References: <16e5fdf90705021440t5b5d30a7m8a9bf92d2b79baf@mail.gmail.com> Message-ID: On 02/05/07, Bill wrote: > Anyone seeing this? > > Running /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/dg.exp ... > FAIL: /Volumes/Gir/devel/llvm/clean/llvm.src/test/CodeGen/CBackend/2007-01-08-ParamAttr-ICmp.ll > for PR1099 > I'm seeing it as well. (i686-pc-linux-gnu) ~ Scott McMurray From bugzilla-daemon at cs.uiuc.edu Thu May 3 11:18:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 11:18:19 -0500 Subject: [LLVMbugs] [Bug 1382] NEW: [inlineasm] 's' constraint bug Message-ID: <200705031618.l43GIJEj029912@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1382 Summary: [inlineasm] 's' constraint bug Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org This testcase: int x[2] = { 1, 2 }; static inline void __attribute__((__always_inline__)) insn1 (int x) { asm volatile ("" :: "n" (x), "i" (x)); } static inline void __attribute__((__always_inline__)) insn2 (const void *x) { asm volatile ("%0 %1" :: "s" (x), "i" (x)); } void foo (void) { insn2 (&x[1]); } Should compile to: _foo: _x+4 _x+4 blr We currently produce: Invalid operand for inline asm constraint 's'! -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 May 3 11:55:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 11:55:40 -0500 Subject: [LLVMbugs] [Bug 1382] [inlineasm] 's' constraint bug Message-ID: <200705031655.l43GteNB030973@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1382 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 2007-05-03 11:55 ------- Fixed. This exposed several bugs in the arm, x86 and ppc backends which are now fixed. The patch in common code is here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/048970.html Testcase here: CodeGen/PowerPC/2007-05-03-InlineAsm-S-Constraint.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 Thu May 3 12:51:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 12:51:33 -0500 Subject: [LLVMbugs] [Bug 1381] [x86] Implement the -fmerge-constants optimization Message-ID: <200705031751.l43HpWpJ032454@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1381 djg at cray.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Summary|Implement the -fmerge- |[x86] Implement the -fmerge- |constants optimization |constants optimization ------- Additional Comments From djg at cray.com 2007-05-03 12:51 ------- I don't have a need for it, so I guess if someone does, they can open a new PR. ------- 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 May 3 16:41:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 16:41:02 -0500 Subject: [LLVMbugs] [Bug 1383] NEW: new per-function attributes for pure/const Message-ID: <200705032141.l43Lf2Sd005359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1383 Summary: new per-function attributes for pure/const 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: sabre at nondot.org llvm-gcc knows that some function calls are pure/const based on flags that it gets passed. For example, we noticed that CoyoteBench/almabench goes 2x faster with llvm+cbe+gcc than it does with either GCC or llvm alone. It turns out that LLVM does a lot of memory promotion and other stuff that ends up making an entire chain of computation be used only by a dead call to asin. Then the CBE+GCC are run, they nuke asin (and the large chain of scalar computations that feed it) because GCC on macos knows that asin is pure/const (darwin math libraries don't set errno), but llvm doesn't know this because it's not a safe assumption. It would be preferable for the CFE to set the pure/const flags on external functions, when it knows them, allowing the optimizers to be more aggressive when safe. This would let us handle this case (a huge speedup for almabench, woo? ), but it would also allow us to drop the huge tables in lib/Analysis/ BasicAliasAnalysis.cpp because libm headers already apply these attributes for GCCs use. -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 May 3 16:49:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 16:49:43 -0500 Subject: [LLVMbugs] [Bug 1346] llvm-gcc unable to handle NOT operation Message-ID: <200705032149.l43LnhcE005580@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1346 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-03 16:49 ------- Yup. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070423/048356.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 Thu May 3 19:32:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 19:32:15 -0500 Subject: [LLVMbugs] [Bug 1332] Crash in addNewBitField for zero-width bitfield after padding Message-ID: <200705040032.l440WFFa010877@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1332 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-03 19:32 ------- http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049052.html Duncan let me know if this works for you or not. 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 Thu May 3 19:52:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 May 2007 19:52:58 -0500 Subject: [LLVMbugs] [Bug 1378] Vector initialization causes C++ front-end ICE Message-ID: <200705040052.l440qwsg011360@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1378 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |llvm-gcc Keywords| |compile-fail Product|new-bugs |tools Resolution| |FIXED Target Milestone|--- |2.0 Version|unspecified |1.8 ------- Additional Comments From sabre at nondot.org 2007-05-03 19:52 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049058.html Testcase here: C++Frontend/2007-05-03-VectorInit.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 Fri May 4 11:51:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 11:51:00 -0500 Subject: [LLVMbugs] [Bug 1364] Assertion failure in PassManager Message-ID: <200705041651.l44Gp0Dw012097@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1364 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From dpatel at apple.com 2007-05-04 11:50 ------- PathExpressions does not required DominatorTree and PostDominatorTree hence it is incorrect to use getAnalysis(). This happens because you have typo "getAnalsisUsage" should be "getAnalysisUsage" :) ------- 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 May 4 11:58:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 11:58:37 -0500 Subject: [LLVMbugs] [Bug 1273] CBE fails on SingleSource/Regression/C/2003-05-22-VarSizeArray Message-ID: <200705041658.l44Gwb7k012807@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1273 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-04 11:58 ------- Duncan's patch is already in. ------- 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 May 4 13:42:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 13:42:54 -0500 Subject: [LLVMbugs] [Bug 1250] Invalid conversion of mode attribute. Message-ID: <200705041842.l44IgsPf019176@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1250 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-04 13:42 ------- Applied Duncan's patch. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049091.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 May 4 16:57:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 16:57:07 -0500 Subject: [LLVMbugs] [Bug 1278] [llvm-gcc] struct layout suboptimality Message-ID: <200705042157.l44Lv7OH024654@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1278 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-04 16:57 ------- First part. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049110.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 May 4 18:18:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 18:18:20 -0500 Subject: [LLVMbugs] [Bug 1278] [llvm-gcc] struct layout suboptimality Message-ID: <200705042318.l44NIKRm027359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1278 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From dpatel at apple.com 2007-05-04 18:18 ------- Reverted prevous patch. It breaks --- struct s { char s1; } __attribute__ ((aligned (8))); struct s foo(void) { struct s S = {'a'}; return S; } ---- ------- 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 May 4 19:09:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 19:09:18 -0500 Subject: [LLVMbugs] [Bug 1384] NEW: crash in opt Message-ID: <200705050009.l4509IBc028241@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1384 Summary: crash in opt Product: new-bugs Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: dalej at apple.com Attached file crashes with opt -instcombine. Somebody's not handling 0-length arrays right but I'm not sure who. ------- 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 May 4 21:00:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 May 2007 21:00:12 -0500 Subject: [LLVMbugs] [Bug 1384] instcombine not handling [0 x type] Message-ID: <200705050200.l4520Cx5030523@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1384 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Scalar Optimizations Keywords| |compile-fail Product|new-bugs |libraries Resolution| |FIXED Summary|crash in opt |instcombine not handling [0 | |x type] Target Milestone|--- |2.0 Version|unspecified |trunk ------- Additional Comments From sabre at nondot.org 2007-05-04 21:00 ------- Fixed. Testcase here: Instcombine/2007-05-04-Crash.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049137.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From asl at math.spbu.ru Sat May 5 03:36:26 2007 From: asl at math.spbu.ru (Anton Korobeynikov) Date: Sat, 05 May 2007 12:36:26 +0400 Subject: [LLVMbugs] llvm-gcc broken Message-ID: <1178354187.13114.42.camel@asl.dorms.spbu.ru> Hello, Everyone. Recent update breaks llvm-gcc at least on x86/linux. Solution: revert attached part of that update. -- With best regards, Anton Korobeynikov. Faculty of Mathematics & Mechanics, Saint Petersburg State University. -------------- next part -------------- A non-text attachment was scrubbed... Name: limits-fix.diff Type: text/x-patch Size: 1691 bytes Desc: not available Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070505/f4e7dee2/attachment.bin From bugzilla-daemon at cs.uiuc.edu Sat May 5 04:16:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 04:16:37 -0500 Subject: [LLVMbugs] [Bug 1385] NEW: Miscompilation of t-pow in GMP testsuite Message-ID: <200705050916.l459Gb8G014736@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1385 Summary: Miscompilation of t-pow in GMP testsuite Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Backend: PowerPC AssignedTo: unassignedbugs at nondot.org ReportedBy: isanbard at gmail.com There seems to be a problem with how code is generated for a call after a loop is unrolled. Consider this C code (on a PowerBook G4 machine): typedef unsigned long int mp_limb_t; typedef struct { int _mp_alloc; int _mp_size; mp_limb_t *_mp_d; } __mpz_struct; typedef __mpz_struct mpz_t[1]; typedef const __mpz_struct *mpz_srcptr; typedef __mpz_struct *mpz_ptr; void Foo(mpz_srcptr base) { unsigned i; mpz_t want; __gmpz_init(want); for (i = 0; i < 2; i++) { __gmpz_mul(want, want, base); } __gmpz_clear(want); } void Bar(mpz_srcptr base) { mpz_t want; __gmpz_init(want); __gmpz_mul(want, want, base); __gmpz_mul(want, want, base); __gmpz_clear(want); } If you look at "testcase.llvm.s", you'll notice that, even though both Foo and Bar functions compute the exact same thing, the Foo function, because it's a loop that's been unrolled, has an "implicit def" of R3 before the gmpz_clear call: mr r3, r29 mr r4, r29 mr r5, r30 bl L___gmpz_mul$stub ;IMPLICIT_DEF_GPRC r3 bl L___gmpz_clear$stub while the Bar case has: mr r3, r29 mr r4, r29 mr r5, r30 bl L___gmpz_mul$stub mr r3, r29 bl L___gmpz_clear$stub The Foo case would be fine, if R3 wasn't trashed in the calls, but it appears to be. Though my guess is that it probably should be marked as "clobbered" across calls. The "mpz_clear" function's in a library and looks like this: void mpz_clear (mpz_ptr m) { (*__gmp_free_func) (m->_mp_d, m->_mp_alloc * BYTES_PER_MP_LIMB); } Now, gcc produces code for Foo that looks like this: L2: addi r3,r1,56 mr r5,r29 mr r4,r3 bl L___gmpz_mul$stub addic. r30,r30,-1 bne- cr0,L2 addi r3,r1,56 bl L___gmpz_clear$stub and Bar that looks like this: addi r3,r1,56 mr r5,r29 mr r4,r3 bl L___gmpz_mul$stub addi r3,r1,56 bl L___gmpz_clear$stub So it's doing the correct thing. We should too. BTW, this is the only test which failed in the GMP testsuite. Woo! -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 Sat May 5 04:22:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 04:22:34 -0500 Subject: [LLVMbugs] [Bug 1376] Section order should be tweaked Message-ID: <200705050922.l459MYoH014886@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1376 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-05 04:22 ------- Fixed with: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049144.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049143.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049142.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 May 5 06:52:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 06:52:04 -0500 Subject: [LLVMbugs] [Bug 1386] NEW: crash on 64 bit bitfield that doesn't start on a byte boundary Message-ID: <200705051152.l45Bq4iY017191@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1386 Summary: crash on 64 bit bitfield that doesn't start on a byte boundary Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr $ gcc -S -emit-llvm bitbash1.c bitbash1.c: In function ???main???: bitbash1.c:14: warning: large integer implicitly truncated to unsigned type bitbash1.c:16: warning: large integer implicitly truncated to unsigned type cc1: llvm-convert.cpp:4913: LValue TreeToLLVM::EmitLV_COMPONENT_REF(tree_node*): Assertion `NumAlignmentUnits && "Not adjusting pointer?"' failed. bitbash1.c:11: internal compiler error: Aborted ------- 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 May 5 06:54:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 06:54:55 -0500 Subject: [LLVMbugs] [Bug 1387] NEW: bitfield access performs unaligned load Message-ID: <200705051154.l45Bst23017265@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1387 Summary: bitfield access performs unaligned load Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Accessing the field X.b does a 64 bit load at a multiple of 4 bytes, not a multiple of 8 bytes. This is presumably problematic on some 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 Sat May 5 09:49:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 09:49:40 -0500 Subject: [LLVMbugs] [Bug 1388] NEW: ARM THUMB: cpy instruction mustn't be emitted by default Message-ID: <200705051449.l45EneT3020169@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1388 Summary: ARM THUMB: cpy instruction mustn't be emitted by default Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: critical Priority: P2 Component: Backend: ARM AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com The cpy instruction is only defined in ARMv6. ------- 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 May 5 10:04:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 10:04:45 -0500 Subject: [LLVMbugs] [Bug 1389] NEW: Nightly Tester shouldn't mark not executed tests as failed Message-ID: <200705051504.l45F4jcF020414@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1389 Summary: Nightly Tester shouldn't mark not executed tests as failed Product: Test Suite Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Nightly Tester AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com In this results (http://llvm.org/nightlytest/test.php?machine=121&night=2812), the JIT and LLCBETA tests were not executed, but they were marked as failed. This makes difficult to see which tests failed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From sabre at nondot.org Sat May 5 14:03:34 2007 From: sabre at nondot.org (Chris Lattner) Date: Sat, 5 May 2007 12:03:34 -0700 (PDT) Subject: [LLVMbugs] llvm-gcc broken In-Reply-To: <1178354187.13114.42.camel@asl.dorms.spbu.ru> References: <1178354187.13114.42.camel@asl.dorms.spbu.ru> Message-ID: On Sat, 5 May 2007, Anton Korobeynikov wrote: > Recent update breaks llvm-gcc at least on x86/linux. Solution: revert > attached part of that update. Applied, thanks Anton! -Chris -- http://nondot.org/sabre/ http://llvm.org/ From bugzilla-daemon at cs.uiuc.edu Sat May 5 13:21:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 13:21:23 -0500 Subject: [LLVMbugs] [Bug 1363] protected visibility not supported Message-ID: <200705051821.l45ILNFU023949@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1363 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-05 13:21 ------- Yes. Definitely. ------- 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 May 5 13:51:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 13:51:16 -0500 Subject: [LLVMbugs] [Bug 1385] [loop unroll] mishandling LCSSA phi of value defined above loop Message-ID: <200705051851.l45IpGsD024656@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1385 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Backend: PowerPC |Loop Optimizer Keywords| |miscompilation OS/Version|MacOS X |All Platform|Macintosh |All Resolution| |FIXED Summary|Miscompilation of t-pow in |[loop unroll] mishandling |GMP testsuite |LCSSA phi of value defined | |above loop Target Milestone|--- |2.0 Version|trunk |1.9 ------- Additional Comments From sabre at nondot.org 2007-05-05 13:51 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049154.html testcase here: Transforms/LoopUnroll/2007-05-05-UnrollMiscomp.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 May 5 14:40:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 14:40:02 -0500 Subject: [LLVMbugs] [Bug 1371] Cast from unsupported vector assertion from LegalizeDAG Message-ID: <200705051940.l45Je2at025736@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1371 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Common Code Generator Code Keywords| |compile-fail Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.0 Version|unspecified |1.8 ------- Additional Comments From sabre at nondot.org 2007-05-05 14:40 ------- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049157.html Testcase here: CodeGen/X86/2007-05-05-VecCastExpand.ll 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 Sat May 5 15:35:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 15:35:07 -0500 Subject: [LLVMbugs] [Bug 1390] NEW: ARM THUMB: invalid push/pop instruction when the stack is 8-bytes aligned Message-ID: <200705052035.l45KZ7T8026671@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1390 Summary: ARM THUMB: invalid push/pop instruction when the stack is 8-bytes aligned Product: libraries Version: trunk Platform: All OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: Backend: ARM AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com ARM backend is emitting push/pop with high registers when the stack is 8-bytes aligned. ------- 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 May 5 16:42:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 16:42:28 -0500 Subject: [LLVMbugs] [Bug 1252] Read-after-free bug in llvm-gcc Message-ID: <200705052142.l45LgSEI028297@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1252 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-05 16:42 ------- Fixed with http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049173.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 May 5 16:44:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 16:44:22 -0500 Subject: [LLVMbugs] [Bug 1336] DejaGnu regressions Message-ID: <200705052144.l45LiMZI028360@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1336 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-05-05 16:44 ------- this is 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 Sat May 5 16:46:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 16:46:53 -0500 Subject: [LLVMbugs] [Bug 1391] NEW: [llvm-gcc] non-local gotos not supported Message-ID: <200705052146.l45LkrbG028423@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1391 Summary: [llvm-gcc] non-local gotos not supported Product: tools Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org This is needed for full nested function support. Here is a testcase: main() { __label__ l; void*x() { goto l; } x(); abort(); return; l: exit(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 Sat May 5 17:02:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 17:02:54 -0500 Subject: [LLVMbugs] [Bug 1392] NEW: Need support for true tail calls Message-ID: <200705052202.l45M2suI028940@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1392 Summary: Need support for true tail calls Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Some testcases: ------ ; RUN: llvm-upgrade < %s | llvm-as | llc -march=x86 | grep {jmp \\*ecx} declare int %putchar(int) int %main(){ %f = cast int (int, int*, int)* %check_tail to int* %res = tail call fastcc int %check_tail( int 10, int* %f,int 10) ret int %res } fastcc int %check_tail(int %x, int* %f, int %g) { %tmp1 = setgt int %x, 0 br bool %tmp1, label %if-then, label %if-else if-then: %fun_ptr = cast int* %f to int(int, int*, int)* %arg1 = add int %x, -1 call int %putchar(int 90) %res = tail call fastcc int %fun_ptr( int %arg1, int * %f, int %g) ret int %res if-else: ret int %x } ------ ------ ; RUN: llvm-upgrade < %s | llvm-as | llc -march=x86 -x86-asm-syntax=intel | not grep call fastcc int %bar(int %X, int(double, int) *%FP) { %Y = tail call fastcc int %FP(double 0.0, int %X) ret int %Y } ------ ------- 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 May 5 17:47:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 17:47:15 -0500 Subject: [LLVMbugs] [Bug 1393] NEW: llvm-upgrade fails to upgrade these tests Message-ID: <200705052247.l45MlFeu030645@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1393 Summary: llvm-upgrade fails to upgrade these tests Product: tools Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: rspencer at reidspencer.com These tests are xfailed because of bugs in llvm-upgrade: Assembler/2007-02-07-UpgradeCSRETCC.ll Transforms/SimplifyCFG/2006-10-29-InvokeCrash.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 May 5 18:47:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 May 2007 18:47:42 -0500 Subject: [LLVMbugs] [Bug 1390] ARM THUMB: invalid push/pop instruction when the stack is 8-bytes aligned Message-ID: <200705052347.l45NlgKc001161@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1390 lauro.venancio at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From lauro.venancio at gmail.com 2007-05-05 18:47 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049201.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 May 6 05:28:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 May 2007 05:28:31 -0500 Subject: [LLVMbugs] [Bug 1394] NEW: Let Verifier check function/parameters attributes more accurate Message-ID: <200705061028.l46ASV21010392@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1394 Summary: Let Verifier check function/parameters attributes more accurate 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: asl at math.spbu.ru Currently, Verifier misses some important checks on function/argument attributs. This includes: like "noreturn" on a parameter (instead of on the function) or inreg on the function instead of a parameter or sret anywhere other than the first argument 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 Sun May 6 08:01:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 May 2007 08:01:01 -0500 Subject: [LLVMbugs] [Bug 1395] NEW: Inliner pass' name not being reported correctly Message-ID: <200705061301.l46D11Zo013243@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1395 Summary: Inliner pass' name not being reported correctly Product: libraries Version: trunk Platform: PC OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Interprocedural Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca On any file, try this: $ llvm/Debug/bin/opt -inline -time-passes b2.bc WARNING: You're attempting to print out a bytecode file. This is inadvisable as it may cause display problems. If you REALLY want to taste LLVM bytecode first-hand, you can force output with the `-f' option. ===-------------------------------------------------------------------------=== ... Pass execution timing report ... ===-------------------------------------------------------------------------=== Total Execution Time: 0.0000 seconds (0.0003 wall clock) ---Wall Time--- --- Name --- 0.0001 ( 36.1%) Dominator Tree Construction 0.0000 ( 21.8%) ET Forest Construction 0.0000 ( 15.7%) Module Verifier 0.0000 ( 14.5%) Basic CallGraph Construction 0.0000 ( 11.0%) N45_GLOBAL__N_InlineSimple.cpp_00000000_47DAB73C13SimpleInlinerE 0.0000 ( 0.5%) Target Data Layout 0.0003 (100.0%) TOTAL ------- 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 May 6 11:28:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 May 2007 11:28:14 -0500 Subject: [LLVMbugs] [Bug 1396] NEW: llvm-gcc can't compile limits.h on Linux Message-ID: <200705061628.l46GSEvF016958@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1396 Summary: llvm-gcc can't compile limits.h on Linux Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at reidspencer.com I tried to run the nightly test but get this error: /usr/include/limits.h:125:26: error: no include path in which to search for limits.h from many sources. The limits.h code looks like: /* Get the compiler's limits.h, which defines almost all the ISO constants. We put this #include_next outside the double inclusion check because it should be possible to include this file more than once and still get the definitions from gcc's header. */ #if defined __GNUC__ && !defined _GCC_LIMITS_H_ /* `_GCC_LIMITS_H_' is what GCC's file defines. */ # include_next #endif The #include_next directive should find the compiler's limits.h which is in llvm-gcc's install dir under lib/gcc/i686-pc-linux-gnu/4.0.1/include. Why it doesn't find the file there, I don't know. Is something wrong with fix_includes? ------- 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 May 6 13:28:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 May 2007 13:28:09 -0500 Subject: [LLVMbugs] [Bug 1396] llvm-gcc can't compile limits.h on Linux Message-ID: <200705061828.l46IS9Nj019112@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1396 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-06 13:28 ------- This problem was fixed by Anton's patch and is in rev 328 of the llvm-gcc mirror. ------- 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 May 6 18:15:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 May 2007 18:15:22 -0500 Subject: [LLVMbugs] [Bug 1395] Inliner pass' name not being reported correctly Message-ID: <200705062315.l46NFM1g025888@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1395 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-05-06 18:15 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/049342.html This is a regression from the ID'ification of passes. Thanks for noticing 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 Mon May 7 01:41:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 May 2007 01:41:27 -0500 Subject: [LLVMbugs] [Bug 1397] NEW: K&R style struct returning functions don't get sret attribute Message-ID: <200705070641.l476fRgB000990@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1397 Summary: K&R style struct returning functions don't get sret attribute Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Example: struct X {}; struct X p() { struct X x; return x; } -> define void @p(%struct.X* %agg.result) { ------- 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 May 7 11:02:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 May 2007 11:02:00 -0500 Subject: [LLVMbugs] [Bug 1398] NEW: wrong code (bad esp update) with -enable-llvm on x86 Message-ID: <200705071602.l47G20IG020051@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1398 Summary: wrong code (bad esp update) with -enable-llvm on x86 Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr I've been seeing mysterious segfaults when compiling using -enable-llvm, even though no exceptions are raised. It turns out to be due to a wrong stack pointer update: $ llc ol.bc -o no_eh $ llc ol.bc -enable-eh -o eh $ diff no_eh eh ... 144c153,155 < addl $12, %esp --- > addl $16, %esp If I change the $16 to $12 in the eh .s, then the segfault goes away and all is well. ------- 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 May 7 11:45:09 2007 From: dpatel at apple.com (Devang Patel) Date: Mon, 7 May 2007 09:45:09 -0700 Subject: [LLVMbugs] llvm-gcc broken In-Reply-To: <1178354187.13114.42.camel@asl.dorms.spbu.ru> References: <1178354187.13114.42.camel@asl.dorms.spbu.ru> Message-ID: On May 5, 2007, at 1:36 AM, Anton Korobeynikov wrote: > Hello, Everyone. > > Recent update breaks llvm-gcc at least on x86/linux. Solution: revert > attached part of that update. This was added to handle the case when user uses -iquote and user project has local header named limits.h See, http://gcc.gnu.org/ml/gcc-patches/2007-04/msg00074.html What error are you seeing ? Thanks, - Devang > > > -- > With best regards, Anton Korobeynikov. > > Faculty of Mathematics & Mechanics, Saint Petersburg State University. > -------------- next part -------------- A non-text attachment was scrubbed... Name: limits-fix.diff Type: text/x-patch Size: 1691 bytes Desc: not available Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070507/359d75ce/attachment.bin -------------- next part -------------- > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From bugzilla-daemon at cs.uiuc.edu Mon May 7 15:33:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 May 2007 15:33:31 -0500 Subject: [LLVMbugs] [Bug 1399] NEW: "llvm-gcc -m64" assert in CodeGen/LiveInterval.cpp:224 Message-ID: <200705072033.l47KXVtM025485@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1399 Summary: "llvm-gcc -m64" assert in CodeGen/LiveInterval.cpp:224 Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: critical Priority: P2 Component: Backend: PowerPC AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org Compiling the following program, "llvm-gcc -m64 main.c": #include int main(void) { printf("hello world.\n"); return 0; } generates the following output: $ llvm-gcc -m64 -v main.c Using built-in specs. Target: powerpc-apple-darwin8.9.0 Configured with: ../../configure --prefix=/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0 --enable-llvm=/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../../llvm/trunk/obj/powerpc-apple-darwin8.8.0 --disable-intl --disable-nls --enable-languages=c,c++ : (reconfigured) ../../configure --prefix=/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0 --enable-llvm=/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../../llvm/trunk/obj/powerpc-apple-darwin8.8.0 --disable-intl --disable-nls --with-gcc-version-trigger=/Users/scottm/play/llvm-cfrontend/gcc/version.c --enable-languages=c,c++ --no-create --no-recursion Thread model: posix gcc version 4.0.1 LLVM (Apple Computer, Inc. build 5449) /Users/scottm/play/llvm-cfrontend/powerpc-apple-darwin8.8.0/bin/../libexec/gcc/powerpc-apple-darwin8.9.0/4.0.1/cc1 -quiet -v -iprefix /Users/scottm/play/llvm-cfrontend/powerpc-apple-darwin8.8.0/bin/../lib/gcc/powerpc-apple-darwin8.9.0/4.0.1/ -D__DYNAMIC__ main.c -fPIC -quiet -dumpbase main.c -mmacosx-version-min=10.4 -m64 -auxbase main -version -o /var/tmp//ccXnbSOF.s ignoring nonexistent directory "/Users/scottm/play/llvm-cfrontend/powerpc-apple-darwin8.8.0/bin/../lib/gcc/powerpc-apple-darwin8.9.0/4.0.1/../../../../powerpc-apple-darwin8.9.0/include" ignoring nonexistent directory "/usr/local/include" ignoring duplicate directory "/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/lib/gcc/powerpc-apple-darwin8.9.0/4.0.1/include" ignoring nonexistent directory "/Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/lib/gcc/powerpc-apple-darwin8.9.0/4.0.1/../../../../powerpc-apple-darwin8.9.0/include" #include "..." search starts here: #include <...> search starts here: /Users/scottm/play/llvm-cfrontend/powerpc-apple-darwin8.8.0/bin/../lib/gcc/powerpc-apple-darwin8.9.0/4.0.1/include /Users/scottm/play/llvm-cfrontend/obj/powerpc-apple-darwin8.8.0/../../powerpc-apple-darwin8.8.0/include /usr/include /System/Library/Frameworks (framework directory) /Library/Frameworks (framework directory) End of search list. GNU C version 4.0.1 LLVM (Apple Computer, Inc. build 5449) (powerpc-apple-darwin8.9.0) compiled by GNU C version 4.0.1 (Apple Computer, Inc. build 5367). GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072 Compiler executable checksum: 3fb6b8555bd4abe30e257ef177238c8e /Users/scottm/play/llvm/trunk/lib/CodeGen/LiveInterval.cpp:224: failed assertion `I != ranges.begin() && "Range is not in interval!"' main.c:4: internal compiler error: Abort trap 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 Mon May 7 15:51:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 May 2007 15:51:15 -0500 Subject: [LLVMbugs] [Bug 1398] wrong code (bad esp update) with -enable-llvm on x86 Message-ID: <200705072051.l47KpFHK026513@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1398 baldrick at free.fr changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From baldrick at free.fr 2007-05-07 15:51 ------- Testcase: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049374.html Fix: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049375.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 May 8 01:16:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 May 2007 01:16:27 -0500 Subject: [LLVMbugs] [Bug 1278] [llvm-gcc] struct layout suboptimality Message-ID: <200705080616.l486GRif006714@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1278 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-08 01:16 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049407.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 May 8 10:54:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 May 2007 10:54:24 -0500 Subject: [LLVMbugs] [Bug 1400] NEW: LLVM frontend crashes when compiling a header file with implementation Message-ID: <200705081554.l48FsOZi026784@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1400 Summary: LLVM frontend crashes when compiling a header file with implementation Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: vjaaskel at cs.tut.fi vjaaskel at kizarwexius:~/testdata/c/testi$ cat testi.h int main() { return 0; } vjaaskel at kizarwexius:~/testdata/c/testi$ llvm-gcc -emit-llvm testi.h testi.h:3: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See for instructions. vjaaskel at kizarwexius:~/testdata/c/testi$ gcc testi.h vjaaskel at kizarwexius:~/testdata/c/testi$ llvm-gcc --version llvm-gcc (GCC) 4.0.1 LLVM (Apple Computer, Inc. build 5449) Copyright (C) 2005 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. ------- 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 May 8 13:10:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 May 2007 13:10:10 -0500 Subject: [LLVMbugs] [Bug 1400] LLVM frontend crashes when compiling a header file with implementation Message-ID: <200705081810.l48IAAkG029359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1400 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-08 13:10 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049421.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 May 8 19:03:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 May 2007 19:03:52 -0500 Subject: [LLVMbugs] [Bug 1401] NEW: Targets should model flags explicitly Message-ID: <200705090003.l4903qpC004731@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1401 Summary: Targets should model flags explicitly Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The various targets (x86, arm, ppc, etc) should model their effect on their flags register (EFLAGS, CR0, etc). This will let us optimize these away, and is a dependency for various code motion optimizations at the machine instr level (scheduling, licm, etc). -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 May 9 03:22:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 03:22:09 -0500 Subject: [LLVMbugs] [Bug 1402] NEW: Update LoopUnswitch to preserve dom info Message-ID: <200705090822.l498M9S2027690@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1402 Summary: Update LoopUnswitch to preserve dom info Product: libraries Version: 2.0 Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Loop Optimizer AssignedTo: unassignedbugs at nondot.org ReportedBy: dpatel at apple.com LoopUnswitch pass should preserve dom info. ------- 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 May 9 03:26:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 03:26:48 -0500 Subject: [LLVMbugs] [Bug 1333] LoopPassManager gives incorrect analysis preservation information Message-ID: <200705090826.l498Qmwq027817@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1333 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-09 03:26 ------- Fixed by updating LoopUnswitch pass to preserve loop info after invoking SplitCriticalEdge(). Now, LoopUnswitch does not request SplitCriticalEdge() to update dominator info and loop info. This means, loop unswitch does not preserve dominator info at the moment. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049453.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 Wed May 9 06:34:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 06:34:36 -0500 Subject: [LLVMbugs] [Bug 1403] NEW: SPARC and IA64 backend: assertion fails when compiling switch-case Message-ID: <200705091134.l49BYaqg031450@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1403 Summary: SPARC and IA64 backend: assertion fails when compiling switch-case 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: vjaaskel at cs.tut.fi The following assertion fails with 2.0 CVS branch when compiling a simple switch-case with four cases. At least IA64 and Sparc backends are affected. Compiles fine with x86 and PPC backends. llc: SelectionDAGISel.cpp:1626: bool llvm::SelectionDAGLowering::handleBTSplitSwitchCase(llvm::SelectionDAGLowering::CaseRec&, std::vector >&, llvm::Value*, llvm::MachineBasicBlock*): Assertion `(FMetric > 0) && "Should handle dense range earlier!"' 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 Wed May 9 10:31:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 10:31:13 -0500 Subject: [LLVMbugs] [Bug 1404] NEW: Assertion "New cannot be in the map!" in llvm-gcc Message-ID: <200705091531.l49FVD9H002743@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1404 Summary: Assertion "New cannot be in the map!" in llvm-gcc Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr I hit the assertion assert(!LLVMValuesMap.count(New) && "New cannot be in the map!"); in changeLLVMValue. What's going on is that LLVMValuesMap, which maps numbers to LLVM values, contains an entry for a freed value. The freed memory has been reused when allocating the new value. Then the assertion thinks that the same value is being added twice, because the new pointer to be added is already present, and fires. Needless to say I can't hit this every time, just from time to time. I only hit it while compiling the Ada f-e. Ada does not use unit-at-a-time, maybe this is relevant? Here's where the memory was freed: operator delete(void*) (vg_replace_malloc.c:336) llvm::AllocaInst::~AllocaInst() (Instructions.h:144) llvm::iplist >::erase(llvm::ilist_iterator) (ilist:330) llvm::Instruction::eraseFromParent() (Instruction.cpp:68) (anonymous namespace)::InstCombiner::DoOneIteration(llvm::Function&, unsigned) (InstructionCombining.cpp:9837) (anonymous namespace)::InstCombiner::runOnFunction(llvm::Function&) (InstructionCombining.cpp:9959) llvm::FPPassManager::runOnFunction(llvm::Function&) (PassManager.cpp:1145) llvm::FunctionPassManagerImpl::run(llvm::Function&) (PassManager.cpp:1103) llvm::FunctionPassManager::run(llvm::Function&) (PassManager.cpp:1048) llvm_emit_code_for_current_function (llvm-backend.cpp:524) tree_rest_of_compilation (tree-optimize.c:706) cgraph_expand_function (cgraphunit.c:896) By the way, valgrind doesn't complain. This is normal: freed memory is not being accessed. A quick fix is to add LLVMValuesMap.erase(New); at the start of changeLLVMValue. However that begs the question of whether freed memory should be floating around in this map or not? ------- 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 May 9 11:32:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 11:32:58 -0500 Subject: [LLVMbugs] [Bug 1405] NEW: the -dA gcc option should be mapped to -asm-verbose Message-ID: <200705091632.l49GWwAN031429@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1405 Summary: the -dA gcc option should be mapped to -asm-verbose Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Passing -dA to gcc -S causes the assembler to be annotated with useful info in mainline, very similar to llc's -asm-verbose. In llvm-gcc -dA has no effect, but could be usefully mapped to -asm-verbose. ------- 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 May 9 13:37:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 13:37:48 -0500 Subject: [LLVMbugs] [Bug 1405] the -dA gcc option should be mapped to -asm-verbose Message-ID: <200705091837.l49IbmTT000378@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1405 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |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 Wed May 9 14:21:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 14:21:32 -0500 Subject: [LLVMbugs] [Bug 1399] "llvm-gcc -m64" assert in CodeGen/LiveInterval.cpp:224 Message-ID: <200705091921.l49JLWCE001711@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1399 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-05-09 14:21 ------- Test here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049460.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 Wed May 9 15:08:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 15:08:23 -0500 Subject: [LLVMbugs] [Bug 1403] SPARC and IA64 backend: assertion fails when compiling switch-case Message-ID: <200705092008.l49K8N7d018058@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1403 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-09 15:08 ------- Fixed with: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049462.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049463.html 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 May 9 15:17:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 15:17:41 -0500 Subject: [LLVMbugs] [Bug 1406] NEW: Assertion failed in RegisterScavenging Message-ID: <200705092017.l49KHfSP032566@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1406 Summary: Assertion failed in RegisterScavenging Product: new-bugs Version: unspecified Platform: PC OS/Version: All Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com laurov at laurov-desktop:~/ffmpeg/build_llvm_arm/libavcodec$ llc < bugpoint-reduced-simplified.bc llc: /home/laurov/llvm/llvm/lib/CodeGen/RegisterScavenging.cpp:113: void llvm::RegScavenger::forward(): Assertion `false' failed. llc((anonymous namespace)::PrintStackTrace()+0x1a)[0x8921d16] llc((anonymous namespace)::SignalHandler(int)+0x112)[0x8921fdc] [0xffffe420] /lib/tls/i686/cmov/libc.so.6(abort+0x101)[0xb7d45641] /lib/tls/i686/cmov/libc.so.6(__assert_fail+0xfb)[0xb7d3d43b] llc(llvm::RegScavenger::forward()+0x1e5)[0x879427b] llc(llvm::RegScavenger::forward(llvm::ilist_iterator)+0x13)[0x84962a5] llc[0x877e0b7] llc[0x8780b42] llc(llvm::MachineFunctionPass::runOnFunction(llvm::Function&)+0x28)[0x8443ada] llc(llvm::FPPassManager::runOnFunction(llvm::Function&)+0x13a)[0x88bd4d8] llc(llvm::FunctionPassManagerImpl::run(llvm::Function&)+0x6e)[0x88bd75c] llc(llvm::FunctionPassManager::run(llvm::Function&)+0x88)[0x88bd8aa] llc(main+0x97d)[0x84133dd] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc)[0xb7d2febc] llc[0x8411b31] Aborted (core dumped) ------- 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 May 9 23:26:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 May 2007 23:26:40 -0500 Subject: [LLVMbugs] [Bug 1407] NEW: BitCodeFormat documentation missing/incomplete Message-ID: <200705100426.l4A4QedL007878@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1407 Summary: BitCodeFormat documentation missing/incomplete Product: Documentation Version: trunk Platform: Macintosh URL: http://llvm.org/docs/BitCodeFormat.html OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: devnull1999 at yahoo.com The page http://llvm.org/docs/BitCodeFormat.html contains no description of the bit-code format. ------- 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 May 10 13:40:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 May 2007 13:40:36 -0500 Subject: [LLVMbugs] [Bug 1408] NEW: Simple Analysis Message-ID: <200705101840.l4AIeaqw017272@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1408 Summary: Simple Analysis 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: dpatel at apple.com Alias information, dominator information, loop information etc... are used by various transformation passes. The interface available to the pass to preserve these information is not simple and straight forward. This results in cases of "will do later". For example, Loop Rotate pass :) 1) This can be simplified if analysis pass enforces SimpleAnalysis protocol. class SimpleAnalysis { void addNewBlock(BasicBlock *B) = 0; void addNewInstruction(Instruction *I) = 0; void updateBasicBlock(BasicBlock *B) = 0; void updateInstruction(Instruction *I) = 0; void deleteBasicBlock(BasicBlock *B) = 0; void deleteInstruction(BasicBlock *I) = 0; }; 2) Pass Manager should be updated to keep track of available analysis passes that implements SimpleAnalysis protocol and provide a way to update _all_ such information. Instead of if (ET = getAnalysisToUpdate()) { ... } if (DI = getAnalysisToUpdate()) { ... } if (LI = getAnalysisToUpdate()) { ... } one should be able to do if (SA = getSimpleAnalysis()) { SA.addBasicBlock(NewPreheader); SA.updateBasicBlock(Header1); SA.updateBasicBlock(PreHeader1); ... } ------- 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 May 10 13:52:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 May 2007 13:52:56 -0500 Subject: [LLVMbugs] [Bug 1409] NEW: Should merge etforest into domtree Message-ID: <200705101852.l4AIquY8017492@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1409 Summary: Should merge etforest into domtree Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Global Analyses AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org ETForest and Domtree both build a tree structure that reflects the dominator tree, ETForest just has a cute numbering scheme which makes certain queries really fast. Merging the two would be very good, both by reducing the number of distinct dominator analyses running around, but also by reducing memory usage (one tree instead of two) and it would make domtree the standard basic dominator analysis. This would also fix specific other problems. For example, domfrontier can be *very* slow right now in some cases because it calls DomTreeNode::properlyDominates. This method is O(n) in the height of the domtree, which is extremely bad for compile time in some cases (e.g. 5186590). If it had etforest available, it would be constant time. -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 May 10 14:33:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 May 2007 14:33:42 -0500 Subject: [LLVMbugs] [Bug 1410] NEW: llc -enable-eh moves invoke calls out of the eh region Message-ID: <200705101933.l4AJXgmL004428@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1410 Summary: llc -enable-eh moves invoke calls out of the eh region Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr $ llvm-as eh5.ll $ llc -enable-eh eh5.bc Inspect eh5.s to see that the first two invokes of @__gnat_rcheck_07 have been moved out of line, and out of any eh region! ------- 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 May 11 01:37:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 11 May 2007 01:37:27 -0500 Subject: [LLVMbugs] [Bug 1411] NEW: dejagnu link variable not set correctly. Message-ID: <200705110637.l4B6bRB5009496@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1411 Summary: dejagnu link variable not set correctly. Product: Test Suite Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: DejaGNU AssignedTo: unassignedbugs at nondot.org ReportedBy: tonic at nondot.org Feature/llvm2cpp.ll fails for a release build because the "link" variable is not set correctly by dejagnu. /usr/bin/ld: can't locate file for: -lLLVMCore This should be 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 May 11 02:19:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 11 May 2007 02:19:37 -0500 Subject: [LLVMbugs] [Bug 1412] NEW: x86 regressions on 2.0 release Message-ID: <200705110719.l4B7JbT9017203@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1412 Summary: x86 regressions on 2.0 release Product: libraries Version: 2.0 Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: tonic at nondot.org These are the current x86 regressions on the 2.0 release branch (fails now but not in 1.9): External/FPGrowth/fpgrowth - CBE External/SPEC/CFP2000/177.mesa/177.mesa - LLC, JIT External/SPEC/CFP2000/188.ammp/188.ammp - CBE External/SPEC/CINT2000/175.vpr/175.vpr - LLC, JIT External/SPEC/CINT2000/176.gcc/176.gcc - LLC, JIT External/SPEC/CINT95/126.gcc/126.gcc - CBE External/SPEC/CINT95/134.perl/134.perl - CBE MultiSource/Applications/SPASS/SPASS - CBE SingleSource/CustomChecked/flops - LLC SingleSource/CustomChecked/stepanov_v1p2 - LLC SingleSource/UnitTests/Vector/build - CBE llc: CBackend.cpp:389: std::ostream& ::CWriter::printSimpleType(std::ostream&, const llvm::Type*, bool, const std::string&): Assertion `(Ty->isPrimitiveType() || Ty->isInteger()) && "Invalid type for printSimpl eType"' 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 May 11 10:22:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 11 May 2007 10:22:57 -0500 Subject: [LLVMbugs] [Bug 1411] dejagnu link variable not set correctly. Message-ID: <200705111522.l4BFMvxw028747@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1411 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.1 ------- Additional Comments From rspencer at reidspencer.com 2007-05-11 10:22 ------- I'm pretty certain the patch resolves any issue with the dejagnu variables. Re-open if the problem persists. ------- 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 May 11 13:31:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 11 May 2007 13:31:05 -0500 Subject: [LLVMbugs] [Bug 1411] dejagnu link variable not set correctly. Message-ID: <200705111831.l4BIV53V030843@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1411 tonic at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From tonic at nondot.org 2007-05-11 13:31 ------- I'm stilling getting this with a Release build on x86 (after using the new Makefile): FAIL: /mounts/zion/disks/0/localhome/tbrethou/llvm/test/Feature/llvm2cpp.ll Failed with exit(1) at line 3 while running: g++ -I/mounts/zion/disks/0/localhome/tbrethou/llvm/test -I/mounts/zion/disks/0/localhome/tbrethou/llvm/test -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -g -fno-exceptions -D_DEBUG -Woverloaded-virtual -pedantic -Wall -W -Wwrite-strings -Wno-long-long -Wunused -Wno-unused-parameter -g -L/mounts/zion/disks/0/localhome/tbrethou/llvm/Debug/lib -L/mounts/zion/disks/0/localhome/tbrethou/llvm/Debug/lib -o Output/llvm2cpp.ll.tmp2.exe Output/llvm2cpp.ll.tmp2.cpp -lLLVMCore -lLLVMSupport -lLLVMSystem -lstdc++ Output/llvm2cpp.ll.tmp2.cpp: In function ???llvm::Module* makeLLVMModule()???: Output/llvm2cpp.ll.tmp2.cpp:1991: warning: unused variable ???ptr_Z??? Output/llvm2cpp.ll.tmp2.cpp:2006: warning: unused variable ???ptr_Z_203??? Output/llvm2cpp.ll.tmp2.cpp:2033: warning: unused variable ???int32_217??? Output/llvm2cpp.ll.tmp2.cpp:2035: warning: unused variable ???int1_219??? Output/llvm2cpp.ll.tmp2.cpp:2387: warning: unused variable ???int32_287??? Output/llvm2cpp.ll.tmp2.cpp:2395: warning: unused variable ???int32_289??? Output/llvm2cpp.ll.tmp2.cpp:2414: warning: unused variable ???int32_293??? [cut same type warnings] Output/llvm2cpp.ll.tmp2.cpp:1681: warning: unused variable ???gvar_int32_114??? /usr/bin/ld: cannot find -lLLVMCore collect2: ld returned 1 exit status >From site.exp: set link "g++ -I/mounts/zion/disks/0/localhome/tbrethou/llvm/test -I/mounts/zion/disks/0/localhome/tbrethou/llvm/test -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -I/mounts/zion/disks/0/localhome/tbrethou/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -g -fno-exceptions -D_DEBUG -Woverloaded-virtual -pedantic -Wall -W -Wwrite-strings -Wno-long-long -Wunused -Wno-unused-parameter -g -L/mounts/zion/disks/0/localhome/tbrethou/llvm/Debug/lib -L/mounts/zion/disks/0/localhome/tbrethou/llvm/Debug/lib " Shouldn't it be look in Release/lib? ------- 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 May 11 17:17:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 11 May 2007 17:17:09 -0500 Subject: [LLVMbugs] [Bug 1413] NEW: CBE produces calls to __builtin_stack_restore with C compilers that don't support them Message-ID: <200705112217.l4BMH9Ga015756@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1413 Summary: CBE produces calls to __builtin_stack_restore with C compilers that don't support them Product: libraries Version: 2.0 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 unconditionally turning calls to llvm.stacksave into calls to __builtin_stack_restore. It should only do this if the C compiler supports these. Worst case, it should emit a header that looks like this: #if !gcc or gcc is too old #define __builtin_stack_restore(X) 0 #endif -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 May 12 15:49:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 May 2007 15:49:32 -0500 Subject: [LLVMbugs] [Bug 1414] NEW: Should support multiple personality functions per module Message-ID: <200705122049.l4CKnWiP005269@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1414 Summary: Should support multiple personality functions per module 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 Current LLVM's EH design in some cases is neither consistent nor complete: in one places it assumes only 1 personality function per module (and thus only one common EH frame), but in other places it can theoretically support multiple personality functions (recording them in the landing pads), but ... lacks callee traversal code in order to assign them correct personality function. For now I'm going to do "only one personality function per module". This will allow to emit eh information at least without extra "hacks" and workaround. Later (maybe for 2.1) we will probably do extra cfg traversal code in order to assign correct personality function. However, this seems to be many problems with indirect function calls.... ------- 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 May 12 17:39:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 May 2007 17:39:11 -0500 Subject: [LLVMbugs] [Bug 1380] EH frames must be emitted in one "atom". Message-ID: <200705122239.l4CMdBan007480@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1380 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-12 17:39 ------- Fixed in http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049536.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049537.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049538.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 May 13 12:37:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 May 2007 12:37:54 -0500 Subject: [LLVMbugs] [Bug 1415] NEW: Refactor debug/eh emission code. Message-ID: <200705131737.l4DHbsi3017598@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1415 Summary: Refactor debug/eh emission code. Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Current DwarfWriter is not so accurate during emission of information in the end of functions. Usually it does the following (in the EndFunction method): -Emit some labels (marking end of function) -Emit some information in new section Note, that labels should be in ".text" section and information is located in the new section. Currently it does everything at once, and if, for example, writer wants to emit some additional stuff *after function end* and *before* and other information - it will fail. I suggest to refactor interface: split EndFunction into EndFunctionPrepare and EndFunction. First function should emit everything needed in the function scope (thus in .text section) and second - emit everything in new section, etc. This will allow us to "stack" different infomation emission passes without thoughts about their order, etc. ------- 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 May 13 13:01:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 May 2007 13:01:00 -0500 Subject: [LLVMbugs] [Bug 1410] llc -enable-eh moves invoke calls out of the eh region Message-ID: <200705131801.l4DI100K017995@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1410 baldrick at free.fr changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From baldrick at free.fr 2007-05-13 13:00 ------- The original failing code now works, so let's consider this 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 May 13 17:20:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 May 2007 17:20:53 -0500 Subject: [LLVMbugs] [Bug 1413] CBE produces calls to __builtin_stack_restore with C compilers that don't support them Message-ID: <200705132220.l4DMKrZW023196@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1413 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.1 ------- Additional Comments From sabre at nondot.org 2007-05-13 17:20 ------- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049548.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 May 13 18:20:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 May 2007 18:20:47 -0500 Subject: [LLVMbugs] [Bug 1416] NEW: cbe miscompiles SPASS Message-ID: <200705132320.l4DNKlDJ024213@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1416 Summary: cbe miscompiles SPASS Product: libraries Version: 2.0 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 The CBE miscompiles SPASS. 'make bugpoint-cbe' starts tracking it down, but eventually gets confused and produces a useless testcase. ------- 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 May 14 16:13:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 May 2007 16:13:13 -0500 Subject: [LLVMbugs] [Bug 1406] Assertion failed in RegisterScavenging Message-ID: <200705142113.l4ELDDed024688@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1406 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-05-14 16:13 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049584.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049583.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 May 14 19:18:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 May 2007 19:18:08 -0500 Subject: [LLVMbugs] [Bug 1417] NEW: llvm-gcc packing empty structs Message-ID: <200705150018.l4F0I88e028161@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1417 Summary: llvm-gcc packing empty structs Product: tools Version: 2.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org This C code: struct { } *X; compiles into: %struct.anon = type <{ }> @X = weak global %struct.anon* null ; <%struct.anon**> [#uses=0] I would expect: %struct.anon = type { } @X = weak global %struct.anon* null ; <%struct.anon**> [#uses=0] -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 May 14 19:36:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 May 2007 19:36:37 -0500 Subject: [LLVMbugs] [Bug 1418] NEW: APInt::toString prints extra digits when radix != 10 and bitwidth > 64 Message-ID: <200705150036.l4F0abNQ028534@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1418 Summary: APInt::toString prints extra digits when radix != 10 and bitwidth > 64 Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: Support Libraries AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at reidspencer.com The rather naive implementation of APInt::toString for radix!=10 and bitwidth > 64 causes it to print more digits than the bit width of the APInt specifies. The algorithm prints whole words, even if only a few bits of the highest order word are used. It needs to not do 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 Mon May 14 21:58:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 May 2007 21:58:30 -0500 Subject: [LLVMbugs] [Bug 1419] NEW: llvm-gcc PaddingElement optimization results in valid fields not being copied Message-ID: <200705150258.l4F2wUZX031503@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1419 Summary: llvm-gcc PaddingElement optimization results in valid fields not being copied Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Test case: struct A { short x; long long :0; }; struct B { char a; char b; unsigned char i; }; union X { struct A a; struct B b; }; int check(void) { union X x, y; y.b.i = 0xff; x = y; return (x.b.i == 0xff); } When compiled with -O2: define i32 @check() { entry: ret i32 0 } It should return 1. This is because the X.i field is not being copied by the aggregate copy, since it lives inside the padding of struct A. ------- 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 May 15 00:43:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 00:43:21 -0500 Subject: [LLVMbugs] [Bug 1411] dejagnu link variable not set correctly. Message-ID: <200705150543.l4F5hLpK007276@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1411 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-15 00:43 ------- Even with the Debug directory gone, it still works fine using Release binaries: testcase /proj/llvm/llvm/test/Feature/dg.exp completed in 25 seconds === Feature Summary === # of expected passes 47 runtest completed at Mon May 14 22:38:50 2007 I verified manually that the program generated by llvm2cpp could be compiled, executed and that it produced the same .ll file as the input. I don't know what's going on in the 2.0 release, but this is definitely working for 2.1 ------- 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 May 15 01:48:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 01:48:31 -0500 Subject: [LLVMbugs] [Bug 1420] NEW: llvm-gcc asserts on address of field in bit field structure Message-ID: <200705150648.l4F6mVDg023902@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1420 Summary: llvm-gcc asserts on address of field in bit field structure Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at reidspencer.com The following transcript shows llvm-gcc failing on input that gcc 4.0.4 compiles correctly. [reid at bashful C++Frontend]$ cat 2007-05-15-FieldAccess.cpp #include class bitFieldStruct { public: int i; unsigned char c:7; int s:17; char c2; }; int main() { printf("sizeof(bitFieldStruct) == %d\n", sizeof(bitFieldStruct)); if (sizeof(bitFieldStruct) != 2 * sizeof(int)) printf("bitFieldStruct should be %d but is %d \n", 2 * sizeof(int), sizeof(bitFieldStruct)); bitFieldStruct x; char* xip = (char*) &x.i; char* xc2p = (char*) &x.c2; printf("Offset bitFieldStruct.i = %d\n", xip - xip); printf("Offset bitFieldStruct.c2 = %d\n", xc2p - xip); return 0; } [reid at bashful C++Frontend]$ /proj/llvm/cfe/install-2/bin/llvm-gcc -o doit 2007-05-15-FieldAccess.cpp -lstdc++ cc1plus: ../../src-2/gcc/llvm-convert.cpp:5583: LValue TreeToLLVM::EmitLV_COMPONENT_REF(tree_node*): Assertion `BitStart == 0 && "It's a bitfield reference or we didn't get to the field!"' failed. 2007-05-15-FieldAccess.cpp: In function 'int main()': 2007-05-15-FieldAccess.cpp:18: internal compiler error: Aborted [reid at bashful C++Frontend]$ gcc -o doit 2007-05-15-FieldAccess.cpp -lstdc++ [reid at bashful C++Frontend]$ ./doit sizeof(bitFieldStruct) == 8 Offset bitFieldStruct.i = 0 Offset bitFieldStruct.c2 = 7 ------- 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 May 15 04:27:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 04:27:04 -0500 Subject: [LLVMbugs] [Bug 1421] NEW: CopyAggregate doesn't play nice with unions Message-ID: <200705150927.l4F9R4Su003797@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1421 Summary: CopyAggregate doesn't play nice with unions Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Testcase: struct A { char c; int i; }; struct B { char c; unsigned char x; }; union U { struct A a; struct B b; }; int check (void) { union U u, v; v.b.x = 0xff; u = v; return (u.b.x == 0xff); } Compiled with -O2: ret i32 0 Should return 1. The point is that the union takes its fields from struct A. The second char x in struct B is in the empty space (due to alignment) in struct A between char c and int i, so is not copied when CopyAggregate copies by fields. Copying by fields is completely bogus with respect to unions of structs unless every field in every union member is covered by an LLVM field. This cannot be done by converting only one union member, as is done right 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 Tue May 15 10:10:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 10:10:29 -0500 Subject: [LLVMbugs] [Bug 1422] NEW: Landing pad wrongly removed Message-ID: <200705151510.l4FFATjo009213@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1422 Summary: Landing pad wrongly removed Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr In the attached testcase, consider the following call to __gnat_rcheck_12: invoke void @__gnat_rcheck_12( i8* getelementptr ([13 x i8]* @.str, i32 0, i32 0), i32 103 ) to label %UnifiedUnreachableBlock unwind label %unwind869 [This is the only call taking 103 as last parameter]. After codegen with llc -enable-eh, this invoke has no entry in the eh table and the unwind block %unwind869 seems to have been deleted. This is unfortunate because when run this call is executed and raises an exception which is not caught... ------- 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 May 15 15:40:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 15:40:36 -0500 Subject: [LLVMbugs] [Bug 1274] [llvm-gcc] front-end crash in lower_nested_functions, not llvm-related Message-ID: <200705152040.l4FKeaQY015972@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1274 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED 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 Tue May 15 15:41:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 15:41:36 -0500 Subject: [LLVMbugs] [Bug 1274] [llvm-gcc] front-end crash in lower_nested_functions, not llvm-related Message-ID: <200705152041.l4FKfamf016029@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1274 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.1 ------- Additional Comments From sabre at nondot.org 2007-05-15 15:41 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049634.html testcase here: test/CFrontend/nested-functions.c -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 May 15 16:06:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 16:06:08 -0500 Subject: [LLVMbugs] [Bug 1423] NEW: opt takes more than 20 min to optimize the attached function Message-ID: <200705152106.l4FL68RK017183@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1423 Summary: opt takes more than 20 min to optimize the attached function 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: lauro.venancio at gmail.com This is a very problematic function. - The opt takes more than 20 min. - The llc takes about 10 min. - The code generated for ARM is very poor. ------- 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 May 15 17:06:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 17:06:55 -0500 Subject: [LLVMbugs] [Bug 1423] opt takes more than 20 min to optimize the attached function Message-ID: <200705152206.l4FM6tVe019988@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1423 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From asl at math.spbu.ru 2007-05-15 17:06 ------- *** This bug has been marked as a duplicate of 930 *** ------- 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 May 15 18:47:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 May 2007 18:47:26 -0500 Subject: [LLVMbugs] [Bug 1419] llvm-gcc PaddingElement optimization results in valid fields not being copied Message-ID: <200705152347.l4FNlQw4022081@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1419 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-15 18:47 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049642.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 Wed May 16 01:22:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 01:22:03 -0500 Subject: [LLVMbugs] [Bug 1423] llc is really slow on the attached function Message-ID: <200705160622.l4G6M3oh028702@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1423 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|DUPLICATE | Summary|opt takes more than 20 min |llc is really slow on the |to optimize the attached |attached function |function | ------- Additional Comments From sabre at nondot.org 2007-05-16 01:22 ------- DSE is a known issue, but this actually triggers a horrible thing in sdisel as well. Reopening for that issue. ------- 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 May 16 01:30:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 01:30:57 -0500 Subject: [LLVMbugs] [Bug 1424] NEW: infinite recursion in scavengeregister Message-ID: <200705160630.l4G6UveK028878@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1424 Summary: infinite recursion in scavengeregister Product: libraries Version: 2.0 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 CC: evan.cheng at apple.com Get the example from PR1423 (the optimized bc file), run this command: llc attachment.cgi -o t.s -time-passes -f -fast You get an infinite recursion: #6132 0x0020e575 in llvm::RegScavenger::scavengeRegister () #6133 0x000356f5 in llvm::ARMRegisterInfo::eliminateFrameIndex () #6134 0x0020e575 in llvm::RegScavenger::scavengeRegister () #6135 0x000356f5 in llvm::ARMRegisterInfo::eliminateFrameIndex () #6136 0x0020e575 in llvm::RegScavenger::scavengeRegister () #6137 0x000356f5 in llvm::ARMRegisterInfo::eliminateFrameIndex () This seems to depend on -fast. -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 May 16 01:39:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 01:39:15 -0500 Subject: [LLVMbugs] [Bug 1423] dagcombine is really slow on this input Message-ID: <200705160639.l4G6dFu6029106@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1423 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Common Code Generator Code Keywords| |slow-compile OS/Version|Linux |All Product|new-bugs |libraries Platform|PC |All Resolution| |FIXED Summary|llc is really slow on the |dagcombine is really slow on |attached function |this input Target Milestone|--- |2.1 Version|unspecified |1.0 ------- Additional Comments From sabre at nondot.org 2007-05-16 01:39 ------- Fixed, patch here: This speeds up isel on this function from: ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 60.0536 ( 96.8%) 1.4269 ( 97.9%) 61.4805 ( 96.8%) 61.4842 ( 96.8%) ARM Instruction Selection 1.6086 ( 2.5%) 0.0115 ( 0.7%) 1.6202 ( 2.5%) 1.6203 ( 2.5%) Linear Scan Register Allocator to: ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 3.9081 ( 66.9%) 1.6220 ( 98.1%) 5.5301 ( 73.8%) 5.5876 ( 74.0%) ARM Instruction Selection 1.5635 ( 26.7%) 0.0119 ( 0.7%) 1.5754 ( 21.0%) 1.5757 ( 20.8%) Linear Scan Register Allocator a speedup of over 11x, in a release build. -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 May 16 01:44:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 01:44:14 -0500 Subject: [LLVMbugs] [Bug 1425] NEW: llvm-gcc or later optzn turns struct copy into hundreds of loads/stores Message-ID: <200705160644.l4G6iEgU029207@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1425 Summary: llvm-gcc or later optzn turns struct copy into hundreds of loads/stores Product: tools Version: 2.0 Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: lauro.venancio at gmail.com The iterative_me function in ffmpeg contains hundreds of loads/stores. This is badness. Lauro, please attach a .i file. -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 May 16 02:48:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 02:48:49 -0500 Subject: [LLVMbugs] [Bug 1426] NEW: SingleSource/CustomChecked/flops fails Message-ID: <200705160748.l4G7mnbD006932@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1426 Summary: SingleSource/CustomChecked/flops fails Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at reidspencer.com This test fails llc/jit/cbe. I ran bugpoint-cbe and it reduced this test case: *** The following function is being miscompiled: main_bb1139_2E_ce You can reproduce the problem with the command line: llc -f bugpoint.test.bc -o bugpoint.test.bc.s gcc ./bugpoint.safe.bc.llc.s.so bugpoint.test.bc.s -o bugpoint.test.bc.exe -Wl,-R. bugpoint.test.bc.exe The shared object was created with: llc -march=c bugpoint.safe.bc -o temporary.c gcc -xc temporary.c -O2 -o ./bugpoint.safe.bc.llc.s.so -shared -fno-strict-aliasing I'll attach the various 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 Wed May 16 02:56:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 02:56:27 -0500 Subject: [LLVMbugs] [Bug 1427] NEW: MMX maskmovq is not properly selected. Message-ID: <200705160756.l4G7uRZE007156@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1427 Summary: MMX maskmovq is not properly selected. 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 Consider attached testcase. We have: Cannot yet select: 0x878fe20: v4i16 = vector_shuffle 0x878fb90, 0x878fd28, 0x878fda0 ./llc((anonymous namespace)::PrintStackTrace()+0x1f)[0x866467f] /lib/libc.so.6(abort+0xec)[0xb7c9bc0c] ./llc((anonymous namespace)::X86DAGToDAGISel::Select_ISD_VECTOR_SHUFFLE_v4i16(llvm::SDOperand const&)+0x436)[0x8385016] This breaks Qt 4.3.0rc1 ------- 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 May 16 04:04:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 04:04:27 -0500 Subject: [LLVMbugs] [Bug 1428] NEW: CFP2006/444.namd fails jit test Message-ID: <200705160904.l4G94Raq008957@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1428 Summary: CFP2006/444.namd fails jit test 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: rspencer at reidspencer.com The jit fails on this test on Linux. The difference in output between native and jit is: diff Output/444.namd.out-nat Output/444.namd.out-jit 1,11c1,8 < *** TEST RUN - 1 ITERATIONS *** < writing to output file namd.out < iteration 0: 1 0 0 < iteration 0: 1 1 0 < iteration 0: 1 1 1 < iteration 0: 0 0 0 < iteration 0: 0 1 0 < iteration 0: 0 1 1 < *** TEST RUN - 1 ITERATIONS *** < SUCCESSFUL COMPLETION < 0000 --- > RESULTSET_BEGIN > 92224 1 0 0 > -355359.705797 0.000000 24229.348329 > 229430.745102 5781.466932 -2454.694382 > 5781.466932 226184.623759 -7867.535225 > -2454.694382 -7867.535225 252281.168869 > 0.000000 0.000000 0.000000 > 0.000000 0.000000 0.000000 13c10 < -0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 15c12 < -0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 26c23 < 0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 28c25 < 0.000000 0.000000 0.000000 24.079992 --- > 0.000000 -0.000000 0.000000 24.079992 40,41c37,38 < -0.000000 -0.000000 0.000000 24.079992 < -0.000000 -0.000000 0.000000 24.079992 --- > 0.000000 -0.000000 0.000000 24.079992 > 0.000000 -0.000000 0.000000 24.079992 52c49 < -0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 54c51 < -0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 65c62 < -0.000000 0.000000 0.000000 24.034240 --- > 0.000000 -0.000000 0.000000 24.034240 67c64 < -0.000000 0.000000 0.000000 24.079992 --- > 0.000000 -0.000000 0.000000 24.079992 79,80c76,77 < -0.000000 -0.000000 0.000000 24.079992 < -0.000000 -0.000000 0.000000 24.079992 --- > 0.000000 -0.000000 0.000000 24.079992 > 0.000000 -0.000000 0.000000 24.079992 ------- 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 May 16 04:40:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 04:40:05 -0500 Subject: [LLVMbugs] [Bug 1429] NEW: 403.gcc CBE output causes GCC 4.0.3 to ICE Message-ID: <200705160940.l4G9e5RR009555@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1429 Summary: 403.gcc CBE output causes GCC 4.0.3 to ICE 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: rspencer at reidspencer.com GCC 4.0.3 will ICE on the (rather large) output generated by 403.gcc. While this is probably just a GCC issue, its also possible that we are generating something strange. It would be interesting to know if other versions of GCC have this problem as well. If not, we can mark this INVALID. But, I'd like to do some investigation first and track the progress here. The ICE is: gcc Output/403.gcc.cbe.c -lm -fno-strict-aliasing -O2 -fno-inline -o Output/403.gcc.cbe Output/403.gcc.cbe.c:6645: warning: conflicting types for built-in function ???malloc??? Output/403.gcc.cbe.c: In function ???fold???: Output/403.gcc.cbe.c:100416: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c:100593: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???c_lex???: Output/403.gcc.cbe.c:341034: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???simplify_unary_operation???: Output/403.gcc.cbe.c:505022: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c:505092: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???simplify_binary_operation???: Output/403.gcc.cbe.c:515312: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c:517421: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???simplify_relational_operation???: Output/403.gcc.cbe.c:527657: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???const_binop???: Output/403.gcc.cbe.c:737554: warning: passing argument 1 of ???_setjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???main???: Output/403.gcc.cbe.c:1326738: warning: return type of ???main??? is not ???int??? Output/403.gcc.cbe.c: In function ???float_signal???: Output/403.gcc.cbe.c:1500270: warning: passing argument 1 of ???longjmp??? from incompatible pointer type Output/403.gcc.cbe.c: In function ???output_init_element???: Output/403.gcc.cbe.c:371760: internal compiler error: Segmentation fault 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 Wed May 16 11:53:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 11:53:34 -0500 Subject: [LLVMbugs] [Bug 1417] llvm-gcc packing empty structs Message-ID: <200705161653.l4GGrY2o017379@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1417 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-16 11:53 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049668.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 Wed May 16 12:36:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 12:36:26 -0500 Subject: [LLVMbugs] [Bug 1412] x86 regressions on 2.0 release Message-ID: <200705161736.l4GHaQJd019506@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1412 tonic at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED OS/Version|MacOS X |Linux Platform|Macintosh |PC Resolution| |WONTFIX ------- Additional Comments From tonic at nondot.org 2007-05-16 12:36 ------- CBE regressions can't be fixed for this release (ie. gcc 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 Wed May 16 12:57:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 12:57:52 -0500 Subject: [LLVMbugs] [Bug 1355] TestRunner enhancement request Message-ID: <200705161757.l4GHvqoA020023@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1355 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-16 12:57 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049671.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 Wed May 16 14:02:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 14:02:42 -0500 Subject: [LLVMbugs] [Bug 1429] 403.gcc CBE output causes GCC 4.0.3 to ICE Message-ID: <200705161902.l4GJ2g7L021456@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1429 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-16 14:02 ------- All recent versions of gcc have this bug, a work around is to build with -fno-builtins. This is not an llvm issue, feel free to file a gcc bug :) -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 May 16 14:08:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 14:08:44 -0500 Subject: [LLVMbugs] [Bug 1430] NEW: mem2reg ignores volatile Message-ID: <200705161908.l4GJ8ic4021741@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1430 Summary: mem2reg ignores volatile Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com The mem2reg pass (llvm::isAllocaPromotable, specifically) does not consider whether the loads/stores it examines are volatile and allows them to be eliminated. Here's a test case. LLVM optimization eliminates the volatile store and load. define i32 @foo() { entry: %retval = alloca i32, align 4 ; [#uses=2] %tmp = alloca i32, align 4 ; [#uses=2] %x = alloca i32, align 4 ; [#uses=2] "alloca point" = bitcast i32 0 to i32 ; [#uses=0] volatile store i32 2, i32* %x %tmp1 = volatile load i32* %x ; [#uses=1] store i32 %tmp1, i32* %tmp %tmp2 = load i32* %tmp ; [#uses=1] store i32 %tmp2, i32* %retval br label %return return: ; preds = %entry %retval3 = load i32* %retval ; [#uses=1] ret i32 %retval3 } ------- 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 May 16 16:21:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 16:21:31 -0500 Subject: [LLVMbugs] [Bug 1426] SingleSource/CustomChecked/flops fails Message-ID: <200705162121.l4GLLV8f024569@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1426 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-16 16:21 ------- The solution here is to break the tie between CustomChecked and TestRunner.sh. The test suite should use its own machinery and not depend on the machinery of the thing its testing. These patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049699.html implement a simple custom checked test running capability without TestRunner.sh, Dejagnu, the LLVM makefile system or Tcl. Instead it is self contained and we can extend it in the future if needed. The makefile changes allow each test to specify a GREP pattern for the output and a set of ARGS to run the program with. That is sufficient for the current set of tests in the CustomChecked directory. Not sure if this should be put on the 2.0 branch or not. ------- 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 May 16 19:44:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 19:44:55 -0500 Subject: [LLVMbugs] [Bug 1430] mem2reg ignores volatile Message-ID: <200705170044.l4H0itTd029694@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1430 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-16 19:44 ------- I'm closing this as invalid. If you feel strongly about this, or have an example where this is a miscompilation (not just suprising behavior), 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 Wed May 16 22:30:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 16 May 2007 22:30:30 -0500 Subject: [LLVMbugs] [Bug 1427] MMX shuffle is not properly selected. Message-ID: <200705170330.l4H3UUZI032519@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1427 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.1 ------- Additional Comments From sabre at nondot.org 2007-05-16 22:30 ------- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049726.html Testcase here: test/CodeGen/X86/mmx-shuffle.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 Thu May 17 02:51:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 02:51:16 -0500 Subject: [LLVMbugs] [Bug 1427] MMX shuffle is not properly selected. Message-ID: <200705170751.l4H7pGO7013815@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1427 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From asl at math.spbu.ru 2007-05-17 02:51 ------- This patch breaks almost all vector tests. Some of them just assert, others - cycle, eat all memory and die. I've reverted it in order not to break nightly testers. ------- 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 May 17 05:03:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 05:03:34 -0500 Subject: [LLVMbugs] [Bug 1431] NEW: segfault with -sccp Message-ID: <200705171003.l4HA3YLQ016401@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1431 Summary: segfault with -sccp Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr $ opt -sccp tc.bc opt((anonymous namespace)::PrintStackTrace()+0x1a)[0x8650486] opt((anonymous namespace)::SignalHandler(int)+0x112)[0x865074c] [0xffffe420] opt(llvm::DominatorTree::DFSPass(llvm::BasicBlock*, llvm::DominatorTreeBase::InfoRec&, unsigned int)+0xf5)[0x85b0fdf] opt(llvm::DominatorTree::calculate(llvm::Function&)+0x108)[0x85b11c6] opt(llvm::DominatorTree::runOnFunction(llvm::Function&)+0x46)[0x85b1692] opt(llvm::FPPassManager::runOnFunction(llvm::Function&)+0x13a)[0x85eb2bc] opt(llvm::FPPassManager::runOnModule(llvm::Module&)+0x6e)[0x85eb484] opt(llvm::MPPassManager::runOnModule(llvm::Module&)+0x11d)[0x85eaf5d] opt(llvm::PassManagerImpl::run(llvm::Module&)+0x6e)[0x85eb12a] opt(llvm::PassManager::run(llvm::Module&)+0x1a)[0x85eb17c] opt(main+0xa10)[0x8388296] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc)[0xb7cc0ebc] opt(realloc+0x7d)[0x837a4c1] Segmentation fault (core dumped) ------- 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 May 17 10:13:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 10:13:18 -0500 Subject: [LLVMbugs] [Bug 1432] NEW: Slow compilation time and unefficient code generated Message-ID: <200705171513.l4HFDI3W021673@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1432 Summary: Slow compilation time and unefficient code generated 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: asl at math.spbu.ru Consider attached preprocessed source (bzipped, from recent Qt 4.3). llvm-gcc compilation took ~30 minutes here (I killed llvm-gcc on my notebook after ~4 hours of compilation and recompiled with -O0). Almost all the time is spent in LCSSA. Generated code for "execute" function is ugly: it uses 30k of stack (compare with 3k stack of gcc-compiled version) and is full of useless moves. I've tried to dig into the problem. At the LCSSA entry (well, actually loop-unswitch) bitcode file is ~10M and contains 68k phi nodes. LCSSA transforms it into 50M bitcode file (I haven't counted phi nodes there). After some optimizations the bitcode size is reduced to 9M, but still is full of phi nodes. The resulting uglyness of code is due to such huge amount of phis. Please note, that I wasn't be able to reproduce this bad behaviour just taking -O0 output and running "opt -std-compile-opts". Instead of this, I had to disable "PerModulePasses" in llvm-gcc, grab the output of "PerFunctionPasses" and run "opt -std-compile-opts" after. I can also attach/send bytecode after "PerFunctionPasses" or just before loop-unswitch (~10M bitcode, 4M bzippped .ll). ------- 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 May 17 12:14:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 12:14:08 -0500 Subject: [LLVMbugs] [Bug 1427] MMX shuffle is not properly selected. Message-ID: <200705171714.l4HHE8Ms024083@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1427 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-05-17 12:14 ------- Fixed again, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049750.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 Thu May 17 12:51:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 12:51:00 -0500 Subject: [LLVMbugs] [Bug 1127] Redesign LLVM 2.0 bytecode format Message-ID: <200705171751.l4HHp0Fl025133@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1127 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Version|trunk |1.0 ------- Additional Comments From sabre at nondot.org 2007-05-17 12:51 ------- This is done. Welcome bitcode. ------- 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 May 17 12:55:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 12:55:53 -0500 Subject: [LLVMbugs] [Bug 1370] Make generic CodeGen tests truly generic Message-ID: <200705171755.l4HHtrYx025330@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1370 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-17 12:55 ------- This has been implemented. Only target-independent test cases should go in CodeGen. All test cases that had a -march= option have been moved or duplicated into the target-specific subdirectories of test/CodeGen. From now on, please do not use the -march option in the test/CodeGen/Generic directory. ------- 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 May 17 13:25:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 13:25:44 -0500 Subject: [LLVMbugs] [Bug 1433] NEW: X86 isel problem with flags Message-ID: <200705171825.l4HIPiew026184@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1433 Summary: X86 isel problem with flags Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The graph input to -view-isel-dags looks ok. The dag input to -view-sched-dags does not. It looks like something bad is happening with flags. define void @test(i128* %A, i128* %B) { %C = load i128* %A %D = load i128* %B %E = add i128 %C, %D store i128 %E, i128* %A ret void } $ llc t2.bc -o - -march=x86 ScheduleDAG.cpp:153: failed assertion `OpVT != MVT::Flag && "Flagged nodes should be in same sunit!"' -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 May 17 13:59:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 13:59:06 -0500 Subject: [LLVMbugs] [Bug 711] [linscan] Coallescing physregs with virtregs should allow spilling the virtreg portion Message-ID: <200705171859.l4HIx63M027015@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=711 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-05-17 13:59 ------- Evan fixed this a while back. Qt doesn't require any regparm hacks to build with 2.0. ------- 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 Thu May 17 14:14:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 14:14:20 -0500 Subject: [LLVMbugs] [Bug 1434] NEW: Assertion failed in llvm-ld when linking a library Message-ID: <200705171914.l4HJEKU1027371@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1434 Summary: Assertion failed in llvm-ld when linking a library Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: llvm-ld AssignedTo: unassignedbugs at nondot.org ReportedBy: zhiruz at gmail.com My llvm-ld crashed (assertion failure) when I tried to link a bc library. Will attached the test files later. What I did is: llvm-gcc -c -O0 -emit-llvm libfoo.c -o libfoo.bc llvm-ar cru libfoo.a libfoo.bc llvm-ranlib libfoo.a llvm-gcc -c -O0 -emit-llvm foo.c -o foo.bc llvm-ld foo.bc libfoo.a -o all.bc (and llvm-ld foo.bc -lfoo -o all.bc) Both produce the following error: ========================================================================== llvm-ld: /home/llvm/lib/VMCore/Value.cpp:57: virtual llvm::Value::~Value(): Assertion `use_begin() == use_end() && "Uses remain when a value is destroyed!"' failed. llvm-ld((anonymous namespace)::PrintStackTrace()+0x1a)[0x84e61b6] llvm-ld((anonymous namespace)::SignalHandler(int)+0x110)[0x84e64de] /lib/tls/libc.so.6[0x683898] /lib/tls/libc.so.6(abort+0xe9)[0x685209] /lib/tls/libc.so.6(__assert_fail+0x101)[0x67cd91] llvm-ld(llvm::Value::~Value()+0x1c4)[0x84a2dd6] llvm-ld(llvm::User::~User()+0x1b)[0x82ee97f] llvm-ld(llvm::Constant::~Constant()+0x1b)[0x82eeab7] llvm-ld(llvm::GlobalValue::~GlobalValue()+0x34)[0x845921e] llvm-ld(llvm::Function::~Function()+0x9d)[0x8454cdb] llvm-ld(llvm::iplist >::erase(llvm::ilist_iterator)+0x33)[0x8367a15] llvm-ld(llvm::iplist >::erase(llvm::ilist_iterator, llvm::ilist_iterator)+0x21)[0x8479465] llvm-ld(llvm::iplist >::clear()+0x53)[0x84794e7] llvm-ld(llvm::Module::~Module()+0x2a)[0x84785ea] llvm-ld(std::auto_ptr::~auto_ptr()+0x1f)[0x82af817] llvm-ld(llvm::Linker::LinkInArchive(llvm::sys::Path const&, bool&)+0x618)[0x82de55a] llvm-ld(llvm::Linker::LinkInLibrary(std::basic_string, std::allocator > const&, bool&)+0x239)[0x82d09b3] llvm-ld(llvm::Linker::LinkInItems(std::vector, std::allocator >, bool>, std::allocator, std::allocator >, bool> > > const&, std::vector, std::allocator >, bool>, std::allocator, std::allocator >, bool> > >&)+0x87)[0x82d0c51] llvm-ld(main+0x5e2)[0x82ac518] /lib/tls/libc.so.6(__libc_start_main+0xd3)[0x670de3] llvm-ld(__gxx_personality_v0+0x149)[0x82a4041] Aborted ------- 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 May 17 17:04:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 17:04:15 -0500 Subject: [LLVMbugs] [Bug 1418] APInt::toString prints extra digits when radix != 10 and bitwidth > 64 Message-ID: <200705172204.l4HM4FeN025223@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1418 rspencer at reidspencer.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at reidspencer.com 2007-05-17 17:04 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049767.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 Thu May 17 17:35:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 17:35:50 -0500 Subject: [LLVMbugs] [Bug 1431] segfault with -sccp Message-ID: <200705172235.l4HMZoSi026801@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1431 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |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 Thu May 17 17:47:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 17:47:43 -0500 Subject: [LLVMbugs] [Bug 1435] NEW: licm ignores volatile store Message-ID: <200705172247.l4HMlhOd027271@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1435 Summary: licm ignores volatile store Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: zhiruz at gmail.com I know that there is some recent discussions about the optimizations on volatile loads and stores, and the conclusion is that LLVM is doing the right thing. But this time the story seems different. See below for details. The test case will be posted shortly. ======================================================================= For my first example, LLVM is able to preserve both volatile load and store correctly. void PassThrough(volatile int* DataIn, volatile int* DataOut) { int i; int buffer[64]; for (i = 0; i < 64; i++) buffer[i] = *DataIn; for (i = 0; i < 64; i++) *DataOut = buffer[i]; } ======================================================================= For the second one, however, LICM will move the volatile store outside of the loop, which is apparently wrong. void Transpose(volatile int* DataIn, volatile int* DataOut) { int i, j; int buffer[64]; for (i = 0; i < 64; i++) buffer[i] = *DataIn; for (i = 0; i < 8; i++) for (j = 0; j < 8; j++) *DataOut = buffer[j * 8 + i]; } ========================================================= What I did: llvm-gcc -O0 -c -emit-llvm test_volatile.c -o test.bc opt -mem2reg -licm test.bc -o new.bc -f ------- 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 May 17 23:04:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 May 2007 23:04:06 -0500 Subject: [LLVMbugs] [Bug 1434] Assertion failed in llvm-ld when linking a library Message-ID: <200705180404.l4I446a6006664@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1434 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|llvm-ld |Bitcode Reader Keywords| |regression OS/Version|Linux |All Product|tools |libraries Platform|PC |All Resolution| |FIXED Target Milestone|2.1 |2.0 ------- Additional Comments From sabre at nondot.org 2007-05-17 23:04 ------- This is a serious regression from 2.0. Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070514/049805.html Testcase here: test/Linker/link-archive.ll Many thanks to Reid for helping put the pieces together and create the testcase. -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 May 18 07:42:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 May 2007 07:42:07 -0500 Subject: [LLVMbugs] [Bug 1400] LLVM frontend crashes when compiling a header file with implementation Message-ID: <200705181242.l4ICg7CE027157@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1400 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From asl at math.spbu.ru 2007-05-18 07:42 ------- Recent addendum of eraseLocalLLVMValues() breaks this again (However, release branch is happy!) ------- 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 May 18 07:44:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 May 2007 07:44:47 -0500 Subject: [LLVMbugs] [Bug 1400] LLVM frontend crashes when compiling a header file with implementation Message-ID: <200705181244.l4ICilfN027222@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1400 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-18 07:44 ------- Sorry, eraseLocalLLVMValues() wasn't commited yet. It seems to be local problem of mine and Duncan's tree :) ------- 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 May 19 00:41:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 May 2007 00:41:07 -0500 Subject: [LLVMbugs] [Bug 1436] NEW: Bootstrapping llvm-gcc4 fails at compile-time, not configure-time Message-ID: <200705190541.l4J5f7Or014266@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1436 Summary: Bootstrapping llvm-gcc4 fails at compile-time, not configure-time 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: danchr at daimi.au.dk I forgot to pass --enable-llvm to the configure script. Unfortunately, this lead to a compile-time error which didn't make sense to me. Optimally, the configure script should have detected that llvm-gcc4 was unable to bootstrap in my environment[1], and reported the error to me. I'll post a patch which has llvm-gcc4 check for LLVM and fail if it isn't present. Reid indicated that it might bootstrap in Apple-private environments, but shouldn't configure then detect those? Please say so if you want the patch to detect Darwin 9, or anything else that might cause the bootstrap failure. This is my first bug to LLVM; any constructive criticism would be greatly appreciated :-) [1] A fairly standard setup of Mac OS X 10.4.9. P.S.: For the record, I was attempting a profiledbootstrap, and I'm afraid I've lost the original failure. Still, isn't it a fair assumption that anyone attempting to compile llvm-gcc4 on Tiger would be doing so for LLVM? ------- 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 May 20 16:35:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 May 2007 16:35:32 -0500 Subject: [LLVMbugs] [Bug 1437] NEW: RunSafely.sh doesn't run safely Message-ID: <200705202135.l4KLZWqH025260@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1437 Summary: RunSafely.sh doesn't run safely Product: Build scripts Version: 1.6 Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at reidspencer.com RunSafely.sh attempts to use the ulimit command to limit cpu/file size/core size, etc when test cases run. However, on almost all Unixes the ulimit command is part of the shell and the way RunSafely.sh invokes them, it occurs in a subshell separate from the actual command being tested. Consequently, things are not run 'safely'. Using the "eval" command works on Linux but not on Darwin. Its possible that we could arrange it to run just as: ( ulimit ... ; command ) That should work on all shells. ------- 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 Mon May 21 05:06:38 2007 From: isanbard at gmail.com (Bill Wendling) Date: Mon, 21 May 2007 03:06:38 -0700 Subject: [LLVMbugs] LICM Test Failure? Message-ID: <9E88B41C-B398-4A60-8772-50A4DD1E421C@gmail.com> Anyone seeing this failure in a fresh tree? Running /Users/wendling/llvm/test/Transforms/LICM/dg.exp ... FAIL: /Users/wendling/llvm/test/Transforms/LICM/scalar_promote.ll Failed with signal(SIGPIPE) at line 1 while running: llvm-upgrade < /Users/wendling/llvm/test/Transforms/ LICM/scalar_promote.ll | llvm-as | opt -licm -stats |& grep {memory locations promoted to register} Binary file (standard input) matches child killed: write on pipe with no readers The output I get from running this command is: ===--------------------------------------------------------------------- ----=== ... Statistics Collected ... ===--------------------------------------------------------------------- ----=== 2 licm - Number of instructions hoisted out of loop 2 licm - Number of memory locations promoted to registers This is on a PowerBook G4. -bw From baldrick at free.fr Mon May 21 08:08:02 2007 From: baldrick at free.fr (Duncan Sands) Date: Mon, 21 May 2007 15:08:02 +0200 Subject: [LLVMbugs] LICM Test Failure? In-Reply-To: <9E88B41C-B398-4A60-8772-50A4DD1E421C@gmail.com> References: <9E88B41C-B398-4A60-8772-50A4DD1E421C@gmail.com> Message-ID: <200705211508.05217.baldrick@free.fr> Hi Bill, > Anyone seeing this failure in a fresh tree? > Running /Users/wendling/llvm/test/Transforms/LICM/dg.exp ... > FAIL: /Users/wendling/llvm/test/Transforms/LICM/scalar_promote.ll > Failed with signal(SIGPIPE) at line 1 > while running: llvm-upgrade < /Users/wendling/llvm/test/Transforms/ > LICM/scalar_promote.ll | llvm-as | opt -licm -stats |& grep {memory > locations promoted to register} > Binary file (standard input) matches > child killed: write on pipe with no readers I've been seen failures like this from time to time (not sure if it was for this test). I couldn't work out what was going on. It smells to me like one of the commands in the pipeline has exited before the preceding command finished writing into the pipe, meaning it gets hit with SIGPIPE. The fact that this doesn't always happens makes me suspect a race condition, for example when you do a fork does the parent or child run first? Answer: either, depending on the whim of the OS. On linux, I think it almost always used to be the child that ran first, but in fact this was never reliable; that didn't stop some programs (eg bash) relying on it anyway - though I understand this was fixed in bash some time ago. I wouldn't be surprised if something like this was causing the problem you see. Ciao, Duncan. From rspencer at reidspencer.com Mon May 21 17:21:53 2007 From: rspencer at reidspencer.com (Reid Spencer) Date: Mon, 21 May 2007 15:21:53 -0700 Subject: [LLVMbugs] LICM Test Failure? In-Reply-To: <200705211508.05217.baldrick@free.fr> References: <9E88B41C-B398-4A60-8772-50A4DD1E421C@gmail.com> <200705211508.05217.baldrick@free.fr> Message-ID: <1179786113.9168.36.camel@bashful.x10sys.com> On Mon, 2007-05-21 at 15:08 +0200, Duncan Sands wrote: > Hi Bill, > > > Anyone seeing this failure in a fresh tree? > > > Running /Users/wendling/llvm/test/Transforms/LICM/dg.exp ... > > FAIL: /Users/wendling/llvm/test/Transforms/LICM/scalar_promote.ll > > Failed with signal(SIGPIPE) at line 1 > > while running: llvm-upgrade < /Users/wendling/llvm/test/Transforms/ > > LICM/scalar_promote.ll | llvm-as | opt -licm -stats |& grep {memory > > locations promoted to register} > > Binary file (standard input) matches > > child killed: write on pipe with no readers > > I've been seen failures like this from time to time (not sure if it > was for this test). I couldn't work out what was going on. It smells > to me like one of the commands in the pipeline has exited before the > preceding command finished writing into the pipe, meaning it gets hit > with SIGPIPE. That's exactly what's going on. The new Tcl-based interpreter of the RUN: lines does not tolerate any bad exit codes or signals from any of the commands in the pipeline. > The fact that this doesn't always happens makes me > suspect a race condition, for example when you do a fork does the > parent or child run first? Tcl should be handling this correctly. > Answer: either, depending on the whim of > the OS. On linux, I think it almost always used to be the child that > ran first, but in fact this was never reliable; that didn't stop some > programs (eg bash) relying on it anyway - though I understand this was > fixed in bash some time ago. Bash isn't involved here. > I wouldn't be surprised if something like > this was causing the problem you see. The only thing I can think of is that the new BitcodeReader is somehow sensitive to "no input" and exits early causing the following command in the pipeline to SIGPIPE. I don't know how that could happen without a (very short) I/O timeout, however. > > Ciao, > > Duncan. > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From bugzilla-daemon at cs.uiuc.edu Mon May 21 18:30:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 May 2007 18:30:32 -0500 Subject: [LLVMbugs] [Bug 1259] LiveVars assertion handling inline assembler. Message-ID: <200705212330.l4LNUWC0031571@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1259 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-05-21 18:30 ------- Looks like this has been already been 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 May 22 02:54:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 May 2007 02:54:49 -0500 Subject: [LLVMbugs] [Bug 1438] NEW: Possible SingleSource/UnitTests/Vector/SSE regression Message-ID: <200705220754.l4M7snuU018128@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1438 Summary: Possible SingleSource/UnitTests/Vector/SSE regression Product: Test Suite Version: 2.0 Platform: PC OS/Version: FreeBSD Status: NEW Severity: normal Priority: P2 Component: Programs Tests AssignedTo: unassignedbugs at nondot.org ReportedBy: emil at cs.rmit.edu.au Failing 'sse.expandfft' and 'sse.stepfft' tests: $ cd llvm-test/SingleSource/UnitTests/Vector/SSE $ gmake /usr/home/emil/llvm-test/DiffOutput.sh "/usr/home/emil/objdir/Release/bin/fpcmp -r 0.016" llc sse.expandfft ******************** TEST (llc) 'sse.expandfft' FAILED! ******************** Execution Context Diff: /usr/home/emil/objdir/Release/bin/fpcmp: Compared: 1.076652e-07 and 1.159249e-07 abs. diff = 8.259700e-09 rel.diff = 7.125044e-02 Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 ******************** TEST (llc) 'sse.expandfft' **************************** /usr/home/emil/llvm-test/DiffOutput.sh "/usr/home/emil/objdir/Release/bin/fpcmp -r 0.016" llc sse.stepfft ******************** TEST (llc) 'sse.stepfft' FAILED! ******************** Execution Context Diff: /usr/home/emil/objdir/Release/bin/fpcmp: FP Comparison failed, not a numeric difference between 'f' and 'e' ******************** TEST (llc) 'sse.stepfft' **************************** /usr/home/emil/llvm-test/DiffOutput.sh "/usr/home/emil/objdir/Release/bin/fpcmp -r 0.016" jit sse.expandfft ******************** TEST (jit) 'sse.expandfft' FAILED! ******************** Execution Context Diff: /usr/home/emil/objdir/Release/bin/fpcmp: Compared: 1.076652e-07 and 1.159249e-07 abs. diff = 8.259700e-09 rel.diff = 7.125044e-02 Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 ******************** TEST (jit) 'sse.expandfft' **************************** /usr/home/emil/llvm-test/DiffOutput.sh "/usr/home/emil/objdir/Release/bin/fpcmp -r 0.016" jit sse.stepfft ******************** TEST (jit) 'sse.stepfft' FAILED! ******************** Execution Context Diff: /usr/home/emil/objdir/Release/bin/fpcmp: FP Comparison failed, not a numeric difference between 'f' and 'e' ******************** TEST (jit) 'sse.stepfft' **************************** $ The sse.stepfft test is failing because the -jit and -llc variants are dumping core: $ grep "exit 138" Output/sse.*.out* Output/sse.stepfft.out-jit:exit 138 Output/sse.stepfft.out-jit.time:exit 138 Output/sse.stepfft.out-llc:exit 138 Output/sse.stepfft.out-llc.time:exit 138 Chris said I should file the bug so this can be tracked. He also said this could be related to stack alignment. Perhaps it's related to http://llvm.org/bugs/show_bug.cgi?id=951 ? ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From baldrick at free.fr Tue May 22 10:04:24 2007 From: baldrick at free.fr (Duncan Sands) Date: Tue, 22 May 2007 17:04:24 +0200 Subject: [LLVMbugs] LICM Test Failure? In-Reply-To: <1179786113.9168.36.camel@bashful.x10sys.com> References: <9E88B41C-B398-4A60-8772-50A4DD1E421C@gmail.com> <200705211508.05217.baldrick@free.fr> <1179786113.9168.36.camel@bashful.x10sys.com> Message-ID: <200705221704.25393.baldrick@free.fr> Hi Reid, > That's exactly what's going on. The new Tcl-based interpreter of the > RUN: lines does not tolerate any bad exit codes or signals from any of > the commands in the pipeline. > > > The fact that this doesn't always happens makes me > > suspect a race condition, for example when you do a fork does the > > parent or child run first? > > Tcl should be handling this correctly. let's hope so. > > Answer: either, depending on the whim of > > the OS. On linux, I think it almost always used to be the child that > > ran first, but in fact this was never reliable; that didn't stop some > > programs (eg bash) relying on it anyway - though I understand this was > > fixed in bash some time ago. > > Bash isn't involved here. It was just an example of a widely used program that suffered from this kind of problem :) > > I wouldn't be surprised if something like > > this was causing the problem you see. > > The only thing I can think of is that the new BitcodeReader is somehow > sensitive to "no input" and exits early causing the following command in > the pipeline to SIGPIPE. I don't know how that could happen without a > (very short) I/O timeout, however. It existed before the bitcode reader came into use (you may remember us discussing it on irc a while back). Ciao, Duncan. From bugzilla-daemon at cs.uiuc.edu Tue May 22 16:10:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 May 2007 16:10:31 -0500 Subject: [LLVMbugs] [Bug 1439] NEW: invoke has no landing pad after codegen Message-ID: <200705222110.l4MLAViK005755@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1439 Summary: invoke has no landing pad after codegen Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr In the testcase there are two calls to @__cxa_throw, one of which is an invoke the other a normal call. [Only the invoke one should be called when the program is run]. After codegen llc -enable-eh -asm-verbose e.bc there is only one call to @__cxa_throw but it is the wrong one! First of all, from the assembler comment (#finally) you can see that it is the normal call version that got put in (the other is in finally.i): .LBB2_1: #finally movl $_ZTI1A, 4(%esp) movl $_ZN1AD1Ev, 8(%esp) movl %esi, (%esp) call __cxa_throw .LBB2_2: #unwind Secondly, in the exception handling table there is no landing pad for this code range, there is only .long .Llabel4-.Leh_func_begin2 # Region start .long .Llabel5-.Llabel4 # Region length .long 0x0 # Landing pad .uleb128 0 # Action which does not cover this. My guess is that codegen cleverly saw that it could collapse the two calls to @__cxa_throw into one call, thinking that they were equivalent, without noticing that they differed in that one was an invoke. ------- 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 May 23 00:07:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 00:07:19 -0500 Subject: [LLVMbugs] [Bug 1440] NEW: llvm demo page should be able to invoke llc to show generated machine code Message-ID: <200705230507.l4N57JIT014285@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1440 Summary: llvm demo page should be able to invoke llc to show generated machine code Product: Website Version: unspecified Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: General Website AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org we should allow selection of any target, by invoking llc -march=, with a radio button on the web site that allows selection of the targets we have. Experimental targets (ia64, msil, etc) should be listed as such on the web page. -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 May 23 01:30:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 01:30:37 -0500 Subject: [LLVMbugs] [Bug 1441] NEW: Pass Manager reruns analysis pass if requested explicitly Message-ID: <200705230630.l4N6Ubbf011796@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1441 Summary: Pass Manager reruns analysis pass if requested explicitly 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: dpatel at apple.com $ opt -domtree -domtree Here there is no need to rerun -domtree. However, $ opt -instcombine -instcombine Here -instcombine is run twice, as expected. ------- 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 May 23 01:38:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 01:38:12 -0500 Subject: [LLVMbugs] [Bug 1435] licm breaks volatile store in multiply-nested loops Message-ID: <200705230638.l4N6cCVL015181@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1435 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Scalar Optimizations |Global Analyses OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|licm ignores volatile store |licm breaks volatile store | |in multiply-nested loops Target Milestone|--- |2.1 Version|trunk |1.0 ------- Additional Comments From sabre at nondot.org 2007-05-23 01:38 ------- Nice catch. This impacts nested loops only. Thanks for the excellent testcase reduction. Fixed. Testcase here: Transforms/LICM/2007-05-22-VolatileSink.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/049945.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 Wed May 23 02:27:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 02:27:08 -0500 Subject: [LLVMbugs] [Bug 1442] NEW: Transforms/LICM/scalar_promote.ll randomly fails Message-ID: <200705230727.l4N7R8aD026461@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1442 Summary: Transforms/LICM/scalar_promote.ll randomly fails Product: Test Suite Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: DejaGNU AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org on my machine, running it in a tight loop sees it fail one out of 20 times. This seems likely to be a bug in the dejagnu stuff or something. See some discussion here: http://lists.cs.uiuc.edu/pipermail/llvmbugs/2007-May/003081.html The test itself doesn't appear to have a race condition in its run 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 Wed May 23 05:00:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 05:00:52 -0500 Subject: [LLVMbugs] [Bug 1443] NEW: Should propagate nothrow attribute to LLVM IR Message-ID: <200705231000.l4NA0q7U003125@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1443 Summary: Should propagate nothrow attribute to LLVM IR 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 When exceptions are enabled any call should be thought as "could throw". However, FE knows, that some call couldn't throw. This information should be propagated to LLVM IR. This moderately reduces size of generated 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 Wed May 23 06:09:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 06:09:30 -0500 Subject: [LLVMbugs] [Bug 1439] invoke has no landing pad after codegen Message-ID: <200705231109.l4NB9UJQ004643@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1439 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-05-23 06:09 ------- Fixed with: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/049953.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/049954.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/049955.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 Wed May 23 10:33:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 10:33:53 -0500 Subject: [LLVMbugs] [Bug 1404] [llvm-gcc] dangling pointers in value map Message-ID: <200705231533.l4NFXrlG009538@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1404 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-05-23 10:33 ------- Applied. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/049957.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 Wed May 23 11:07:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 11:07:22 -0500 Subject: [LLVMbugs] [Bug 1444] NEW: Function deleted in an error Message-ID: <200705231607.l4NG7Mqh010137@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1444 Summary: Function deleted in an error 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 Consider attached testcase. Run llc via "./llc -enable-eh". You'll see, that _Z17handle_unexpected function is dropped from the generated assembly. This is horrible bad :(. Everything is ok with "-fast". Almost surely the problem is due to "unreachable" inst in the "invcont16" BB (note, that cxa_throw will always throw, so the code is correct). ------- 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 May 23 14:50:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 May 2007 14:50:09 -0500 Subject: [LLVMbugs] [Bug 1445] NEW: Allow redifinition of REAL_LD_FILE_NAME for changing linker Message-ID: <200705231950.l4NJo9jl018429@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1445 Summary: Allow redifinition of REAL_LD_FILE_NAME for changing linker Product: tools Version: trunk Platform: All OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: abdulras at uiuc.edu As per IRC: < Devang> | aha.. I see. There is a patch in collect2.c from Apple's GCC team to undef REAL_LD_FILE_NAME. I don't understand motivation behind it. If you really need this then file bugzilla PR for llvm-gcc. ------- 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 May 24 08:10:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 24 May 2007 08:10:11 -0500 Subject: [LLVMbugs] [Bug 1446] NEW: scalarrepl cycles on easy input Message-ID: <200705241310.l4ODABB1018978@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1446 Summary: scalarrepl cycles on easy input 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 reduced .ll file. It seems, that scalarrepl is cycles. Steps to reproduce: ./opt -scalarrepl PS: This is reducion from ioquake3, reported by Bram Adams (http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-May/009165.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 Thu May 24 13:44:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 24 May 2007 13:44:22 -0500 Subject: [LLVMbugs] [Bug 1446] scalarrepl decimates giant struct Message-ID: <200705241844.l4OIiMh5030625@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1446 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |code-quality, slow-compile OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|scalarrepl cycles on easy |scalarrepl decimates giant |input |struct Target Milestone|--- |2.1 ------- Additional Comments From sabre at nondot.org 2007-05-24 13:44 ------- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/050016.html Testcase here: Transforms/ScalarRepl/2007-05-24-LargeAggregate.ll 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 May 24 18:11:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 24 May 2007 18:11:44 -0500 Subject: [LLVMbugs] [Bug 1444] Function deleted in an error Message-ID: <200705242311.l4ONBisr004160@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1444 dalej at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From dalej at apple.com 2007-05-24 18:11 ------- Fixed here. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/050010.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070521/050013.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 May 25 06:14:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 May 2007 06:14:18 -0500 Subject: [LLVMbugs] [Bug 1447] NEW: llvm-gcc should pass -fast to llc when compiling at -O0 Message-ID: <200705251114.l4PBEIn3017261@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1447 Summary: llvm-gcc should pass -fast to llc when compiling at -O0 Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr If compiling at -O0 then no optimizations should be performed [*]. However it looks like codegen still performs optimizations - passing -fast to llc would disable these. [*] People compiling at -O0 might want speed, or an unoptimized baseline for debugging purposes etc. ------- 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 May 25 08:42:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 May 2007 08:42:56 -0500 Subject: [LLVMbugs] [Bug 1447] llvm-gcc should pass -fast to llc when compiling at -O0 Message-ID: <200705251342.l4PDgutl008008@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1447 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From sabre at nondot.org 2007-05-25 08:42 ------- You're right, we should! llvm-backend.cpp contains this: if (TheTarget->addPassesToEmitFileFinish(*PM, 0, /*Fast*/optimize == 0)) { cerr << "Error interfacing to target machine!\n"; exit(1); Which does this. :) Perhaps tail merging is enabled at -fast? If so, that seems wrong. -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 May 26 17:24:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 26 May 2007 17:24:30 -0500 Subject: [LLVMbugs] [Bug 1448] NEW: Pointer to invalid memory kept Message-ID: <200705262224.l4QMOUvL013189@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1448 Summary: Pointer to invalid memory kept Product: tools Version: 2.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-ld AssignedTo: unassignedbugs at nondot.org ReportedBy: jlh at gmx.ch In tools/llvm-ld/llvm-ld.cpp, around line 353, a 'const char *' to a temporary std::string is being stored, with the string going out of scope right after, making that pointer invalid. Snippet: std::string lib_name = "-l" + LinkItems[index].first; args.push_back(lib_name.c_str()); I originally reported this problem on the llvm-commit list, because I was told to do so, but maybe I should have opened a bug report anyway. See [1] and its follow-ups for a simple fix. Funnily, in my particular case, things still worked fine as long as I only had one -l option. But with several such option, all pointers in 'args' pointed to the exact same memory location, still containing the last library name processed and gcc got called like "gcc -llib -llib -llib -llib". [1] http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070430/ 048734.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 May 28 10:57:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 May 2007 10:57:06 -0500 Subject: [LLVMbugs] [Bug 1449] NEW: Script Message-ID: <200705281557.l4SFv6kA026897@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1449 Summary: Script Product: Build scripts Version: cvs Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: autoconf AssignedTo: unassignedbugs at nondot.org ReportedBy: humerlude at yahoo.com 1 ------- 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 May 28 11:12:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 May 2007 11:12:57 -0500 Subject: [LLVMbugs] [Bug 1449] Script Message-ID: <200705281612.l4SGCvaJ027253@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1449 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From asl at math.spbu.ru 2007-05-28 11:12 ------- Crap. The bots are clever 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 Mon May 28 12:20:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 May 2007 12:20:40 -0500 Subject: [LLVMbugs] [Bug 1450] NEW: qq Message-ID: <200705281720.l4SHKeEH028451@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1450 Summary: qq Product: Build scripts Version: cvs Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: cosmoball at tiscali.cz ww ------- 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 May 28 12:48:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 May 2007 12:48:11 -0500 Subject: [LLVMbugs] [Bug 1450] qq Message-ID: <200705281748.l4SHmBEZ028945@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1450 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-28 12:48 ------- bugzilla spam :( ------- 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 May 28 13:28:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 May 2007 13:28:12 -0500 Subject: [LLVMbugs] [Bug 1451] NEW: llalalalaa Message-ID: <200705281828.l4SISCfp029620@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1451 Summary: llalalalaa Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: mail1958559 at handymail.org llalalalaa ------- 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 May 29 00:41:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 00:41:57 -0500 Subject: [LLVMbugs] [Bug 1451] llalalalaa Message-ID: <200705290541.l4T5fvB9007452@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1451 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- 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 May 29 02:39:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 02:39:35 -0500 Subject: [LLVMbugs] [Bug 1452] NEW: typo in getelementptr docs Message-ID: <200705290739.l4T7dZWJ019529@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1452 Summary: typo in getelementptr docs Product: Documentation Version: 2.0 Platform: All URL: http://llvm.org/docs/LangRef.html#i_getelementptr OS/Version: All Status: NEW Severity: trivial Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: scott.llvm at h4ck3r.net In the section starting: "For example, let's consider a C code fragment and how it gets compiled to LLVM:" the "C" function has "define" before it, making things minorly confusing. ------- 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 May 29 03:54:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 03:54:35 -0500 Subject: [LLVMbugs] [Bug 1453] NEW: LLVM code Message-ID: <200705290854.l4T8sZnY021278@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1453 Summary: LLVM code Product: tools Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: extract AssignedTo: unassignedbugs at nondot.org ReportedBy: qualertin at yahoo.com LLVM 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 Tue May 29 04:08:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 04:08:33 -0500 Subject: [LLVMbugs] [Bug 1452] typo in getelementptr docs Message-ID: <200705290908.l4T98X7G022255@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1452 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-05-29 04:08 ------- Fixed. Removed the "define" LLVM keyword. -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 May 29 04:03:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 04:03:33 -0500 Subject: [LLVMbugs] [Bug 1453] LLVM code Message-ID: <200705290903.l4T93Xtg022043@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1453 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- 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 May 29 05:45:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 05:45:54 -0500 Subject: [LLVMbugs] [Bug 1454] NEW: 8cLcBzrN Message-ID: <200705291045.l4TAjsbx024273@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1454 Summary: 8cLcBzrN Product: Bugzilla Admin Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Mail AssignedTo: unassignedbugs at nondot.org ReportedBy: tatarins at gmail.com 8cLcBzrN ------- 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 May 29 05:47:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 05:47:03 -0500 Subject: [LLVMbugs] [Bug 1455] NEW: crte library Message-ID: <200705291047.l4TAl3tR024337@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1455 Summary: crte library Product: Runtime Libraries Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: crtend library AssignedTo: unassignedbugs at nondot.org ReportedBy: qualertin at yahoo.com crte library ------- 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 May 29 05:48:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 05:48:01 -0500 Subject: [LLVMbugs] [Bug 1456] NEW: crte library Message-ID: <200705291048.l4TAm10w024440@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1456 Summary: crte library Product: Runtime Libraries Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: crtend library AssignedTo: unassignedbugs at nondot.org ReportedBy: qualertin at yahoo.com crte library ------- 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 May 29 07:14:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 07:14:27 -0500 Subject: [LLVMbugs] [Bug 1457] NEW: sd Message-ID: <200705291214.l4TCERBE026450@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1457 Summary: sd Product: Documentation Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: crucialerror at tiscali.cz sdsd ------- 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 May 29 07:16:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 07:16:23 -0500 Subject: [LLVMbugs] [Bug 1458] NEW: xc Message-ID: <200705291216.l4TCGNVb026548@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1458 Summary: xc Product: libraries Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: Backend: Alpha AssignedTo: unassignedbugs at nondot.org ReportedBy: crucialerror at tiscali.cz xcxc ------- 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 May 29 10:24:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 10:24:18 -0500 Subject: [LLVMbugs] [Bug 1454] 8cLcBzrN Message-ID: <200705291524.l4TFOIxb030671@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1454 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-29 10:24 ------- I made some changes that should stop this spam. ------- 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 May 29 10:24:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 10:24:34 -0500 Subject: [LLVMbugs] [Bug 1455] crte library Message-ID: <200705291524.l4TFOYWE030697@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1455 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-29 10:24 ------- spam ------- 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 May 29 10:24:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 10:24:51 -0500 Subject: [LLVMbugs] [Bug 1456] crte library Message-ID: <200705291524.l4TFOpBN030721@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1456 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-29 10:24 ------- spam ------- 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 May 29 10:25:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 10:25:04 -0500 Subject: [LLVMbugs] [Bug 1457] sd Message-ID: <200705291525.l4TFP4ov030742@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1457 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-29 10:25 ------- spam ------- 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 May 29 10:25:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 10:25:33 -0500 Subject: [LLVMbugs] [Bug 1458] xc Message-ID: <200705291525.l4TFPXeT030770@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1458 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-29 10:25 ------- spam ------- 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 May 29 12:50:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 12:50:58 -0500 Subject: [LLVMbugs] [Bug 1459] NEW: lalalalalla Message-ID: <200705291750.l4THowSX001915@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1459 Summary: lalalalalla Product: Test Suite Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: QMTest AssignedTo: unassignedbugs at nondot.org ReportedBy: mail1958559 at handymail.org lalalalallalalalalallalalalalalla ------- 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 May 29 12:51:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 12:51:56 -0500 Subject: [LLVMbugs] [Bug 1459] lalalalalla Message-ID: <200705291751.l4THpuWr001966@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1459 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From mail1958559 at handymail.org 2007-05-29 12:51 ------- Created an attachment (id=896) --> (http://llvm.org/bugs/attachment.cgi?id=896&action=view) llalalalaa llalalalaa ------- 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 May 29 14:14:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 14:14:53 -0500 Subject: [LLVMbugs] [Bug 1460] NEW: Some Message-ID: <200705291914.l4TJErG1004084@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1460 Summary: Some Product: Test Suite Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: QMTest AssignedTo: unassignedbugs at nondot.org ReportedBy: alexpushnoy at yahoo.com Crashes ------- 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 May 29 14:18:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 14:18:17 -0500 Subject: [LLVMbugs] [Bug 1461] NEW: HTML Documentation Message-ID: <200705291918.l4TJIHM3004237@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1461 Summary: HTML Documentation Product: Documentation Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: alexpushnoy at yahoo.com Where to find? ------- 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 May 29 14:48:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 14:48:34 -0500 Subject: [LLVMbugs] [Bug 1462] NEW: APInt.h assert: "Too many bits for uint64_t" Message-ID: <200705291948.l4TJmYhc004811@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1462 Summary: APInt.h assert: "Too many bits for uint64_t" Product: tools Version: trunk Platform: Other OS/Version: Linux Status: NEW Severity: blocker Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org Linux x86_64 llvm-gcc compile from this morning (all cvs repo updates up to 10am PDT): /work/scottm/llvm-cfrontend/obj/x86_64-unknown-linux-gnu/gcc/xgcc -B/work/scottm/llvm-cfrontend/obj/x86_64-unknown-linux-gnu/gcc/ -B/work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/bin/ -B/work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/lib/ -isystem /work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/include -isystem /work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/sys-include -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../../gcc -I../../../gcc/. -I../../../gcc/../include -I../../../gcc/../libcpp/include -I/work/scottm/llvm/trunk/include -I/work/scottm/llvm/trunk/obj/x86_64-unknown-linux-gnu//include -DL_addvdi3 -fvisibility=hidden -DHIDE_EXPORTS -c ../../../gcc/libgcc2.c -o libgcc/./_addvdi3.o cc1: /work/scottm/llvm/trunk/include/llvm/ADT/APInt.h:859: uint64_t llvm::APInt::getZExtValue() const: Assertion `getActiveBits() <= 64 && "Too many bits for uint64_t"' failed. ../../../gcc/libgcc2.c:117: 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 Tue May 29 15:36:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 15:36:21 -0500 Subject: [LLVMbugs] [Bug 1460] Some Message-ID: <200705292036.l4TKaL8e005664@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1460 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID Summary|Some |Some ------- 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 May 29 15:45:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 15:45:00 -0500 Subject: [LLVMbugs] [Bug 1461] HTML Documentation Message-ID: <200705292045.l4TKj0Q5005949@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1461 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- 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 May 29 16:50:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 16:50:17 -0500 Subject: [LLVMbugs] [Bug 1463] NEW: APInt.h assert: "Too many bits for uint64_t" Message-ID: <200705292150.l4TLoHnP007362@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1463 Summary: APInt.h assert: "Too many bits for uint64_t" Product: tools Version: trunk Platform: Other OS/Version: Linux Status: NEW Severity: blocker Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org Linux x86_64 llvm-gcc compile from this morning (all cvs repo updates up to 10am PDT): /work/scottm/llvm-cfrontend/obj/x86_64-unknown-linux-gnu/gcc/xgcc -B/work/scottm/llvm-cfrontend/obj/x86_64-unknown-linux-gnu/gcc/ -B/work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/bin/ -B/work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/lib/ -isystem /work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/include -isystem /work/scottm/llvm-cfrontend/i686-unknown-linux-gnu/x86_64-unknown-linux-gnu/sys-include -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../../gcc -I../../../gcc/. -I../../../gcc/../include -I../../../gcc/../libcpp/include -I/work/scottm/llvm/trunk/include -I/work/scottm/llvm/trunk/obj/x86_64-unknown-linux-gnu//include -DL_addvdi3 -fvisibility=hidden -DHIDE_EXPORTS -c ../../../gcc/libgcc2.c -o libgcc/./_addvdi3.o cc1: /work/scottm/llvm/trunk/include/llvm/ADT/APInt.h:859: uint64_t llvm::APInt::getZExtValue() const: Assertion `getActiveBits() <= 64 && "Too many bits for uint64_t"' failed. ../../../gcc/libgcc2.c:117: 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 Tue May 29 23:17:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 May 2007 23:17:14 -0500 Subject: [LLVMbugs] [Bug 1463] APInt.h assert: "Too many bits for uint64_t" Message-ID: <200705300417.l4U4HEes015428@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1463 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-05-29 23:17 ------- *** This bug has been marked as a duplicate of 1462 *** ------- 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 May 30 01:13:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 01:13:37 -0500 Subject: [LLVMbugs] [Bug 1421] union aggregate copies broken in some cases Message-ID: <200705300613.l4U6DbYV017618@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1421 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |llvm-gcc Keywords| |miscompilation OS/Version|Linux |All Product|new-bugs |tools Platform|Other |All Resolution| |FIXED Summary|CopyAggregate doesn't play |union aggregate copies |nice with unions |broken in some cases Target Milestone|--- |2.1 Version|unspecified |1.0 ------- Additional Comments From sabre at nondot.org 2007-05-30 01:13 ------- The bug in llvm-gcc is fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050081.html The bug in scalarrepl is fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050083.html Testcases here: test/CFrontend/2007-05-29-UnionCopy.c test/Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.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 Wed May 30 01:16:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 01:16:30 -0500 Subject: [LLVMbugs] [Bug 1421] union aggregate copies broken in some cases Message-ID: <200705300616.l4U6GUKT017692@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1421 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From sabre at nondot.org 2007-05-30 01:16 ------- Urg, reduced testcases work, but the full ones don't. ------- 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 May 30 01:24:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 01:24:37 -0500 Subject: [LLVMbugs] [Bug 1421] union aggregate copies broken in some cases Message-ID: <200705300624.l4U6ObaP017968@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1421 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-05-30 01:24 ------- Here's the final fix: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050084.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 Wed May 30 06:50:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 06:50:20 -0500 Subject: [LLVMbugs] [Bug 1464] NEW: 8cLcBzrN Message-ID: <200705301150.l4UBoKSt000802@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1464 Summary: 8cLcBzrN Product: Build scripts Version: cvs Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: autoconf AssignedTo: unassignedbugs at nondot.org ReportedBy: tatarins at gmail.com 8cLcBzrN ------- 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 May 30 06:51:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 06:51:08 -0500 Subject: [LLVMbugs] [Bug 1465] NEW: qw Message-ID: <200705301151.l4UBp80P000838@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1465 Summary: qw Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: allgema at tiscali.cz qwqw ------- 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 May 30 06:53:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 06:53:29 -0500 Subject: [LLVMbugs] [Bug 1466] NEW: bv Message-ID: <200705301153.l4UBrTTR001068@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1466 Summary: bv Product: Projects Version: unspecified Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: Stacker AssignedTo: unassignedbugs at nondot.org ReportedBy: allgema at tiscali.cz bvbv ------- 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 May 30 08:58:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 08:58:21 -0500 Subject: [LLVMbugs] [Bug 1467] NEW: Java component Message-ID: <200705301358.l4UDwLuv003862@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1467 Summary: Java component Product: Projects Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Java AssignedTo: unassignedbugs at nondot.org ReportedBy: wecunda at yahoo.com ------- 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 May 30 09:00:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 09:00:11 -0500 Subject: [LLVMbugs] [Bug 1468] NEW: Geneal docs issue Message-ID: <200705301400.l4UE0B7v003979@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1468 Summary: Geneal docs issue Product: Documentation Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: wecunda at yahoo.com Geneal docs issue ------- 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 May 30 10:56:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 10:56:02 -0500 Subject: [LLVMbugs] [Bug 1469] NEW: IVM Tools Message-ID: <200705301556.l4UFu2rH006903@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1469 Summary: IVM Tools Product: tools Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: llc AssignedTo: unassignedbugs at nondot.org ReportedBy: iscandero at yahoo.com What`s new? ------- 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 May 30 10:58:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 10:58:34 -0500 Subject: [LLVMbugs] [Bug 1470] NEW: Online profit Message-ID: <200705301558.l4UFwYxQ007154@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1470 Summary: Online profit Product: Website Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Documentation AssignedTo: unassignedbugs at nondot.org ReportedBy: iscandero at yahoo.com About new recommendations ------- 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 May 30 11:01:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:01:11 -0500 Subject: [LLVMbugs] [Bug 1471] NEW: BTHM Libraries Message-ID: <200705301601.l4UG1B39007419@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1471 Summary: BTHM Libraries Product: libraries Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Bitcode Reader AssignedTo: unassignedbugs at nondot.org ReportedBy: iscandero at yahoo.com Not found? ------- 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 May 30 11:02:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:51 -0500 Subject: [LLVMbugs] [Bug 1464] 8cLcBzrN Message-ID: <200705301602.l4UG2pvC007608@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1464 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:52 -0500 Subject: [LLVMbugs] [Bug 1468] Geneal docs issue Message-ID: <200705301602.l4UG2qXl007622@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1468 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:53 -0500 Subject: [LLVMbugs] [Bug 1466] bv Message-ID: <200705301602.l4UG2rvv007636@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1466 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:53 -0500 Subject: [LLVMbugs] [Bug 1471] BTHM Libraries Message-ID: <200705301602.l4UG2rhf007650@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1471 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:54 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:54 -0500 Subject: [LLVMbugs] [Bug 1467] Java component Message-ID: <200705301602.l4UG2sxp007665@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1467 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:55 -0500 Subject: [LLVMbugs] [Bug 1469] IVM Tools Message-ID: <200705301602.l4UG2trN007679@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1469 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:56 -0500 Subject: [LLVMbugs] [Bug 1470] Online profit Message-ID: <200705301602.l4UG2uUB007697@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1470 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:02:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:56 -0500 Subject: [LLVMbugs] [Bug 1472] NEW: QM test Message-ID: <200705301602.l4UG2uIc007710@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1472 Summary: QM test Product: Test Suite Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: QMTest AssignedTo: unassignedbugs at nondot.org ReportedBy: iscandero at yahoo.com Is it correct? ------- 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 May 30 11:02:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:02:56 -0500 Subject: [LLVMbugs] [Bug 1465] qw Message-ID: <200705301602.l4UG2u7w007723@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1465 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 11:02 ------- Spam ------- 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 May 30 11:04:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:04:47 -0500 Subject: [LLVMbugs] [Bug 1472] QM test Message-ID: <200705301604.l4UG4liA007956@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1472 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- 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 May 30 11:16:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:16:35 -0500 Subject: [LLVMbugs] [Bug 1473] NEW: dagcombine miscompilation Message-ID: <200705301616.l4UGGZuf008305@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1473 Summary: dagcombine miscompilation Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org dag combine is miscompiling this on ppc: define i8 @foo(i16 zext %a) zext { entry: %tmp2 = lshr i16 %a, 10 ; [#uses=1] %tmp23 = trunc i16 %tmp2 to i8 ; [#uses=1] %tmp4 = shl i8 %tmp23, 1 ; [#uses=1] %tmp5 = and i8 %tmp4, 2 ; [#uses=1] ret i8 %tmp5 } -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 May 30 11:28:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:28:36 -0500 Subject: [LLVMbugs] [Bug 1474] NEW: sss Message-ID: <200705301628.l4UGSaiX008641@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1474 Summary: sss Product: Runtime Libraries Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: libprofile library AssignedTo: unassignedbugs at nondot.org ReportedBy: quidproquo at tiscali.cz sssss ------- 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 May 30 11:30:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:30:26 -0500 Subject: [LLVMbugs] [Bug 1475] NEW: bbb Message-ID: <200705301630.l4UGUQc9008868@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1475 Summary: bbb Product: Test Suite Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: QMTest AssignedTo: unassignedbugs at nondot.org ReportedBy: quidproquo at tiscali.cz bbbb ------- 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 May 30 11:30:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:30:48 -0500 Subject: [LLVMbugs] [Bug 1473] dagcombine miscompilation of shifts Message-ID: <200705301630.l4UGUmcc008904@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1473 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Summary|dagcombine miscompilation |dagcombine miscompilation of | |shifts Target Milestone|--- |2.1 ------- Additional Comments From sabre at nondot.org 2007-05-30 11:30 ------- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050089.html Testcase here: CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.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 Wed May 30 11:33:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:33:31 -0500 Subject: [LLVMbugs] [Bug 1476] NEW: rt Message-ID: <200705301633.l4UGXVUN009046@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1476 Summary: rt Product: Runtime Libraries Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: normal Priority: P2 Component: other AssignedTo: unassignedbugs at nondot.org ReportedBy: aleed at tiscali.cz rtrt ------- 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 May 30 11:51:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:51:02 -0500 Subject: [LLVMbugs] [Bug 1477] NEW: Website advice Message-ID: <200705301651.l4UGp2ID009556@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1477 Summary: Website advice Product: Website Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Publications AssignedTo: unassignedbugs at nondot.org ReportedBy: evadonta at yahoo.com ------- 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 May 30 11:53:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 11:53:26 -0500 Subject: [LLVMbugs] [Bug 1478] NEW: last try Message-ID: <200705301653.l4UGrQYK009667@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1478 Summary: last try Product: Build scripts Version: cvs Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: evadonta at yahoo.com last try ------- 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 May 30 12:14:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 12:14:05 -0500 Subject: [LLVMbugs] [Bug 1478] last try Message-ID: <200705301714.l4UHE5QE010452@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1478 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 12:14 ------- spam ------- 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 May 30 12:14:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 12:14:06 -0500 Subject: [LLVMbugs] [Bug 1476] rt Message-ID: <200705301714.l4UHE6w2010466@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1476 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 12:14 ------- spam ------- 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 May 30 12:14:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 12:14:07 -0500 Subject: [LLVMbugs] [Bug 1475] bbb Message-ID: <200705301714.l4UHE7JK010480@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1475 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 12:14 ------- spam ------- 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 May 30 12:14:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 12:14:08 -0500 Subject: [LLVMbugs] [Bug 1474] sss Message-ID: <200705301714.l4UHE8Hd010494@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1474 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 12:14 ------- spam ------- 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 May 30 12:14:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 12:14:05 -0500 Subject: [LLVMbugs] [Bug 1477] Website advice Message-ID: <200705301714.l4UHE5fp010438@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1477 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-30 12:14 ------- spam ------- 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 May 30 14:35:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 14:35:44 -0500 Subject: [LLVMbugs] [Bug 1479] NEW: wwessdddd Message-ID: <200705301935.l4UJZi7T018088@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1479 Summary: wwessdddd Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: mail1958559 at handymail.org wwessddddwwessdddd ------- 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 May 30 14:39:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 14:39:08 -0500 Subject: [LLVMbugs] [Bug 1480] NEW: reefrefreergtreg Message-ID: <200705301939.l4UJd8Xo018447@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1480 Summary: reefrefreergtreg Product: Test Suite Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: Programs Tests AssignedTo: unassignedbugs at nondot.org ReportedBy: mail1958559 at handymail.org reefrefreergtregreefrefreergtreg ------- 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 May 30 15:05:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 15:05:15 -0500 Subject: [LLVMbugs] [Bug 1479] wwessdddd Message-ID: <200705302005.l4UK5Fug019342@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1479 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From isanbard at gmail.com 2007-05-30 15:05 ------- spam ------- 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 May 30 15:05:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 15:05:41 -0500 Subject: [LLVMbugs] [Bug 1480] reefrefreergtreg Message-ID: <200705302005.l4UK5fxh019406@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1480 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From isanbard at gmail.com 2007-05-30 15:05 ------- spam ------- 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 May 30 16:30:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 16:30:29 -0500 Subject: [LLVMbugs] [Bug 1481] NEW: Unsupported intrinsics by llc .c generator Message-ID: <200705302130.l4ULUTfQ022457@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1481 Summary: Unsupported intrinsics by llc .c generator Product: new-bugs Version: unspecified Platform: Macintosh URL: http://users.ugent.be/~badams/secret/ioquake3.ppc.bc.tar .bz2 OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: bram.adams at ugent.be Error: Code generator does not support intrinsic function 'llvm.ppc.altivec.lvsl'! See file on http://users.ugent.be/~badams/secret/ioquake3.ppc.bc.tar.bz2. Problem does not occur on Linux X86, only on OSX PPC (version 2.0). ------- 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 Wed May 30 16:40:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 16:40:52 -0500 Subject: [LLVMbugs] [Bug 1446] scalarrepl decimates giant struct Message-ID: <200705302140.l4ULeqOW022814@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1446 bram.adams at ugent.be changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From bram.adams at ugent.be 2007-05-30 16:40 ------- It looks like the fix only works on OSX (PPC), as on Linux X86 the frontend is now stuck inside the for- loop on lines 176-->391 of PromoteMemoryToRegister.cpp when in the control flow of -scalarrepl: #4 0x08764f72 in (anonymous namespace)::PromoteMem2Reg::run (this=0xbff9bfa0) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/Transforms/Utils/ PromoteMemoryToRegister.cpp:271 #5 0x08767021 in llvm::PromoteMemToReg (Allocas=@0xbff9c120, ET=@0x8b14c10, DF=@0x8b14d98, AST=0x0) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/Transforms/Utils/ PromoteMemoryToRegister.cpp:812 #6 0x08744713 in (anonymous namespace)::SROA::performPromotion (this=0x8b14f38, F=@0x8bc8468) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/Transforms/Scalar/ ScalarReplAggregates.cpp:130 #7 0x087445ba in (anonymous namespace)::SROA::runOnFunction (this=0x8b14f38, F=@0x8bc8468) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/Transforms/Scalar/ ScalarReplAggregates.cpp:101 #8 0x08826499 in llvm::FPPassManager::runOnFunction (this=0x8b15d00, F=@0x8bc8468) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/VMCore/PassManager.cpp:1145 #9 0x08826237 in llvm::FunctionPassManagerImpl::run (this=0x8b14e78, F=@0x8bc8468) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/VMCore/PassManager.cpp:1103 #10 0x0882611f in llvm::FunctionPassManager::run (this=0x8b156e8, F=@0x8bc8468) at /home/bram/workspace/svn/aspicere2/trunk/llvm/lib/VMCore/PassManager.cpp:1048 Steps to reproduce (new attached file): llvm-gcc -emit-llvm -Wall -fno-strict-aliasing -Wimplicit -Wstrict-prototypes -pipe -DUSE_OPENAL=1 -DUSE_CURL=1 -DUSE_CURL_DLOPEN=1 -DUSE_SDL_VIDEO=1 -DUSE_SDL_SOUND=1 -I/usr/include/ SDL -D_REENTRANT -m32 -DUSE_LOCAL_HEADERS=1 -DSVN_VERSION=\"1.33_SVN6303:6304\" - DNDEBUG -O3 -march=i586 -fomit-frame-pointer -ffast-math -funroll-loops -falign-loops=2 - falign-jumps=2 -falign-functions=2 -fstrength-reduce -o file.o -c file.c ------- 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 May 30 16:42:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 May 2007 16:42:49 -0500 Subject: [LLVMbugs] [Bug 1482] NEW: Compiling generated .c file (by llc) on Linux X86 fails Message-ID: <200705302142.l4ULgnCo022884@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1482 Summary: Compiling generated .c file (by llc) on Linux X86 fails 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: bram.adams at ugent.be On Linux X86, I get this while compiling a generated .c file (by llc): error: unknown register name 'S' in 'asm' This is the offending line in the attached .c file (notice the "S" on the last line): __asm__ volatile ("\tpushal\t\t\t\t\n\tmovl %0,%%esi\t\t\n\tmovl %1,%%edi\t\t\n\tcall *%2\t\t\t\n \tmovl %%esi,%0\t\t\n\tmo\ vl %%edi,%1\t\t\n\tpopal\t\t\t\t\n" :"=m"(memProgramStack_2E_5872),"=m"(memOpStack_2E_5873) :"m"(memEntryPoint_2E_5874),"m"(memProgramStack_2E_5872),"m"(memOpStack_2E_5873) :"D","S"); Apparently, leaving out "S" helps (compiling the .c file works), but what does it do and why is it here? ------- 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 May 31 05:27:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 05:27:31 -0500 Subject: [LLVMbugs] [Bug 1446] scalarrepl decimates giant struct Message-ID: <200705311027.l4VARVxH020078@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1446 bram.adams at ugent.be changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From bram.adams at ugent.be 2007-05-31 05:27 ------- False alarm. Had only recompiled LLVM, not the frontend. ------- 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 May 31 08:07:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 08:07:55 -0500 Subject: [LLVMbugs] [Bug 1483] NEW: Run time library Message-ID: <200705311307.l4VD7tB3022995@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1483 Summary: Run time library Product: Runtime Libraries Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: libprofile library AssignedTo: unassignedbugs at nondot.org ReportedBy: dulert at yahoo.com Run time library ------- 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 May 31 08:11:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 08:11:36 -0500 Subject: [LLVMbugs] [Bug 1484] NEW: QMTest Message-ID: <200705311311.l4VDBax4023147@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1484 Summary: QMTest Product: Test Suite Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: QMTest AssignedTo: unassignedbugs at nondot.org ReportedBy: dulert at yahoo.com tet ------- 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 May 31 11:21:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 11:21:12 -0500 Subject: [LLVMbugs] [Bug 1483] Run time library Message-ID: <200705311621.l4VGLCP0027585@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1483 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-05-31 11:21 ------- spam ------- 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 May 31 11:21:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 11:21:29 -0500 Subject: [LLVMbugs] [Bug 1484] QMTest Message-ID: <200705311621.l4VGLTrZ027608@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1484 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- 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 May 31 13:43:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 13:43:13 -0500 Subject: [LLVMbugs] [Bug 1424] infinite recursion in scavengeregister Message-ID: <200705311843.l4VIhDKB031423@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1424 lauro.venancio at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From lauro.venancio at gmail.com 2007-05-31 13:43 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050110.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070528/050111.html The proposed idea was not implemented because the PrologEpilogInserter already has code to define the ScavengingFrameIndex offset. This bug was caused by a bug in this 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 Thu May 31 14:13:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 14:13:43 -0500 Subject: [LLVMbugs] [Bug 1485] NEW: BS Message-ID: <200705311913.l4VJDh92032312@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1485 Summary: BS Product: Build scripts Version: cvs Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: autoconf AssignedTo: unassignedbugs at nondot.org ReportedBy: taboliser at yahoo.com BS ------- 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 May 31 15:49:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 15:49:04 -0500 Subject: [LLVMbugs] [Bug 1485] BS Message-ID: <200705312049.l4VKn4v3002384@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1485 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-05-31 15:49 ------- ugh! ------- 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 May 31 16:29:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 16:29:00 -0500 Subject: [LLVMbugs] [Bug 1424] infinite recursion in scavengeregister Message-ID: <200705312129.l4VLT0BS003364@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1424 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From evan.cheng at apple.com 2007-05-31 16:28 ------- This is the wrong fix. It's sweeping the bug under the rug. Even if disabling the use of RS when FP is used is correct (when I am not convinced of), that's up to the target to not instantiate RS when it detects that. PEI shouldn't do this check. Please revert the patch. I've not have a chance to look into this. Please explain, in this case, why is ScavengingFrameIndex not within the displacement limit? If there is a bug in the ARMRegisterInfo.cpp code that causes the miscalculation, then please fix it. If we ever run into situation where it's not possible to reserve a legal frame index, then we need a proper solution. Can we detect that before register allocation? If so, we need to detect that and reserve R12 and make sure RS isn't instantiated. ------- 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 May 31 17:46:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 May 2007 17:46:07 -0500 Subject: [LLVMbugs] [Bug 1424] infinite recursion in scavengeregister Message-ID: <200705312246.l4VMk7J5005497@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1424 evan.cheng at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From evan.cheng at apple.com 2007-05-31 17:46 ------- Now I understand it! :-) ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.