From bugzilla-daemon at cs.uiuc.edu Sat Dec 1 00:30:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 1 Dec 2007 00:30:55 -0600 Subject: [LLVMbugs] [Bug 817] [scev trip-count computation] performance regression on nested-loop In-Reply-To: Message-ID: <200712010630.lB16UtCa020104@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=817 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |DUPLICATE --- Comment #13 from Nick Lewycky 2007-12-01 00:30:54 --- As filed, nothing. The run-time performance regression on nested-loop has been fixed. The bug transmuted a bit, and the patch affects entirely different code than the change from PR1614. PR1614 removed extra work from HowManyLessThans while this patch changes ComputeIterationCount so that it will use HowManyLessThans more often, instead of falling back to exhaustive computation. I'm going to close this bug and not refile it, given that we don't have a code example that would necessitate modifying ComputeIterationCount this way. *** This bug has been marked as a duplicate of bug 1614 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Dec 1 11:29:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 1 Dec 2007 11:29:03 -0600 Subject: [LLVMbugs] [Bug 1835] New: Register pressure reduction list schedulers are extremely slow on very long basic blocks Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1835 Summary: Register pressure reduction list schedulers are extremely slow on very long basic blocks 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: romixlev at yahoo.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1275) --> (http://llvm.org/bugs/attachment.cgi?id=1275) A BC file containing a very long basic block During the work on BigBlock regalloc, Duraid provided me with some test *.bc files for testing. They contain very, very long basic blocks. Before recent changes done by Evan in ScheduleDAGRRList.cpp, those files could be compiled by LLC is a reasonable time, when RR schedulers were used. But since a while, there is a huge compile-time regression. After running under a profiler, it turned out that it is the isReachable() function in this file, that comsumes up-to 98% of the LLC run-time on those input files with very long basic blocks. Most likely, a more efficient data structure than a SmallPtrSet should be used in isReachable() function to improve the performance. Attached are a few basic blocks for your use, provided by Duraid. They are all of the following simple form: - load some 128-bit vector values from an array in memory - perform various logical operations using these values as input - store the results to another array in memory There is no flow control at all (other than a "ret void" at the end of each function!), so there's just one basic block in each of the files attached. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 3 02:37:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 3 Dec 2007 02:37:05 -0600 Subject: [LLVMbugs] [Bug 1836] New: lli interpreter crashed for integer type whose bitwidth > 64 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1836 Summary: lli interpreter crashed for integer type whose bitwidth > 64 Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Interpreter AssignedTo: unassignedbugs at nondot.org ReportedBy: zhousheng00 at gmail.com CC: baldrick at free.fr, llvmbugs at cs.uiuc.edu The lli interpreter crashed for the following case: ; ModuleID = 'x.c' target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32" target triple = "i686-pc-linux-gnu" define i32 @main() { entry: %retval = alloca i32 ; [#uses=2] %tmp = alloca i32 ; [#uses=2] %x = alloca i75, align 16 ; [#uses=1] %"alloca point" = bitcast i32 0 to i32 ; [#uses=0] store i75 999, i75* %x, align 16 store i32 0, i32* %tmp, align 4 %tmp1 = load i32* %tmp, align 4 ; [#uses=1] store i32 %tmp1, i32* %retval, align 4 br label %return return: ; preds = %entry %retval2 = load i32* %retval ; [#uses=1] ret i32 %retval2 } The output is : *** glibc detected *** lli: free(): invalid pointer: 0x088cb1d8 *** ======= Backtrace: ========= /lib/libc.so.6[0x479efd] /lib/libc.so.6(cfree+0x90)[0x47d550] lli(_ZN4llvm12AllocaHolderD1Ev+0x2b)[0x84bcb4d] lli(_ZN4llvm18AllocaHolderHandleD1Ev+0x37)[0x84bcbab] lli(_ZN4llvm16ExecutionContextD1Ev+0x14)[0x84bd780] lli(_ZN9__gnu_cxx13new_allocatorIN4llvm16ExecutionContextEE7destroyEPS2_+0x11)[0x84bd7af] ... ... -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 3 06:04:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 3 Dec 2007 06:04:07 -0600 Subject: [LLVMbugs] [Bug 1837] New: lli: free(): invalid next size (fast) Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1837 Summary: lli: free(): invalid next size (fast) Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sanxiyn at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1277) --> (http://llvm.org/bugs/attachment.cgi?id=1277) Cheap sed compiled with Clang SVN r44532. Debian GNU/Linux Sid (up-to-date). In the attached archive, sed.bc is Cheap sed 030913 by Laurent Vogel compiled with Clang. http://lvogel.free.fr/sed.htm sedopt.bc was produced by "opt -std-compile-opts sed.bc -o sedopt.bc". "lli sed.bc -f factor.sed input" produces output. "lli sedopt.bc -f factor.sed input" crashes with the message "lli: free(): invalid next size (fast)". After extraction, run "make" to reproduce. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 3 13:18:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 3 Dec 2007 13:18:14 -0600 Subject: [LLVMbugs] [Bug 1828] loop-index-split produce bogus code In-Reply-To: Message-ID: <200712031918.lB3JIEq7021672@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1828 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Devang Patel 2007-12-03 13:18:13 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071203/055939.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 3 16:10:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 3 Dec 2007 16:10:23 -0600 Subject: [LLVMbugs] [Bug 1838] New: x86 backend doesn't insert emms when it is needed Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1838 Summary: x86 backend doesn't insert emms when it is needed Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu target datalayout = "e-p:32:32-f64:32:64-i64:32:64-f80:32:32" declare i32 @printf(i8*, ...) @G = global double 4.0 @H = global <2 x i32> zeroinitializer @L = global [5 x i8] c"%.2f\0A" define i32 @main() { %c = load double* @G %a = load <2 x i32>* @H %cc = add <2 x i32> %a, %a store <2 x i32> %cc, <2 x i32>* @H %l = getelementptr [5 x i8]* @L, i64 0, i64 0 call i32 (i8 *, ...)* @printf(i8* %l, double %c) ret i32 0 This testcase, when compiled with llc, then gcc, and run on my x86-linux computer, prints out "nan"; it should print out "4.0". The cause is that the necessary "emms" is not generated after using the MMX registers (the generated asm works fine once I manually add the emms). This testcase, when compiled with llc, then gcc, and run on my x86-linux computer, prints out "nan"; it should print out "4.0". The cause is that the necessary "emms" is not generated after using the MMX registers (the generated asm works fine once I manually add the emms). Front-ends that generate vector code shouldn't be expected to know that on X86 specifically, for vectors 64 bits wide, a special intrinsic has to be called after vector operations; also, the front-end has no real way to guarantee code movement won't screw up such a fix. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 4 15:37:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 4 Dec 2007 15:37:46 -0600 Subject: [LLVMbugs] [Bug 1839] New: __builtin_trap doesn't trap Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1839 Summary: __builtin_trap doesn't trap Product: tools Version: 2.1 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: djg at cray.com CC: llvmbugs at cs.uiuc.edu This C testcase: void foo(void) { __builtin_trap(); } is compiled by llvm-gcc into define void @foo() { entry: ret void } In otherwords, __builtin_trap() is treated as a no-op. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 02:25:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 02:25:39 -0600 Subject: [LLVMbugs] [Bug 1840] New: asm calls should be marked nounwind Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1840 Summary: asm calls should be marked nounwind 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 CC: llvmbugs at cs.uiuc.edu The inliner knows that asm calls cannot unwind. If asm calls were marked nounwind this logic would not be needed, and all the places that reason about nounwind calls would automatically do the same for asm. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 03:01:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 03:01:08 -0600 Subject: [LLVMbugs] [Bug 1841] New: opt tool will not work on Windows for pass libraries Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1841 Summary: opt tool will not work on Windows for pass libraries Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows 2000 Status: NEW Severity: major Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: humeafo at gmail.com CC: llvmbugs at cs.uiuc.edu opt tool will not work on Windows for pass dlls. this is because pass manager dynamic load the dll but the static pass targetar variable is not exported and not visible to the dll on windows. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 04:09:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 04:09:06 -0600 Subject: [LLVMbugs] [Bug 1841] opt tool will not work on Windows for pass libraries In-Reply-To: Message-ID: <200712051009.lB5A96ZP012720@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1841 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asl at math.spbu.ru Status|NEW |RESOLVED Resolution| |INVALID --- Comment #1 from Anton Korobeynikov 2007-12-05 04:08:50 --- It won't work at all because of lack of dynamic linking. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 16:48:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 16:48:15 -0600 Subject: [LLVMbugs] [Bug 1842] New: dag combiner changes select into and incorrectly Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1842 Summary: dag combiner changes select into and incorrectly Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1279) --> (http://llvm.org/bugs/attachment.cgi?id=1279) triggers bug in the attached code, a setcc -> select i8 is transformed into a setcc (which yields 0 or 1) -> and which would work if the setcc returned ~0, but it doesn't (and AlphaTargetLowering.cpp says so). The selection dag is correct before the dag combiner is run. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 18:03:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 18:03:06 -0600 Subject: [LLVMbugs] [Bug 1831] CodeGen/ARM/lsr-code-insertion.ll failing In-Reply-To: Message-ID: <200712060003.lB6036VH031441@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1831 Evan Cheng changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #2 from Evan Cheng 2007-12-05 18:03:05 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071203/056003.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 5 19:45:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 5 Dec 2007 19:45:09 -0600 Subject: [LLVMbugs] [Bug 1842] dag combiner changes select into and incorrectly In-Reply-To: Message-ID: <200712060145.lB61j9bT004561@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1842 Dale Johannesen changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Dale Johannesen 2007-12-05 19:45:08 --- Fixed. (I can't verify the Alpha code is correct, but the optimization is no longer being done.) http://llvm.org/viewvc/llvm-project?view=rev&revision=44649 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 7 08:39:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 7 Dec 2007 08:39:55 -0600 Subject: [LLVMbugs] [Bug 1843] New: compiler segfaults when compiling ruby 1.9 ( latest svn version) Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1843 Summary: compiler segfaults when compiling ruby 1.9 (latest svn version) Product: tools Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: v.konrad at lse.ac.uk CC: llvmbugs at cs.uiuc.edu llvm-gcc -g -O2 -I. -I.ext/include/i686-linux -I./include -I. -DRUBY_EXPORT -D_GNU_SOURCE=1 -c vm.c vm.c: In function ???vm_call0???: vm.c:483: warning: passing argument 7 of ???vm_call_bmethod??? discards qualifiers from pointer target type vm.c: At top level: vm.c:1901: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See for instructions. make: *** [vm.o] Error 1 -------------------------------- llvm-gcc --version: llvm-gcc (GCC) 4.2.1(llvm) (Based on Apple Inc. build 5530) Copyright (C) 2007 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. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 7 14:44:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 7 Dec 2007 14:44:51 -0600 Subject: [LLVMbugs] [Bug 1843] compiler segfaults when compiling ruby 1.9 (latest svn version) In-Reply-To: Message-ID: <200712072044.lB7Kipvw022847@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1843 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |DUPLICATE --- Comment #3 from Anton Korobeynikov 2007-12-07 14:44:50 --- Yes, looks like dup of 1805, but now with C frontend also. The assertion is triggered with optimization level > 0 *** This bug has been marked as a duplicate of bug 1805 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 7 20:02:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 7 Dec 2007 20:02:25 -0600 Subject: [LLVMbugs] [Bug 1844] New: Build problem on Fedora core 7 x86_64 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1844 Summary: Build problem on Fedora core 7 x86_64 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: mcmillan at cadence.com CC: llvmbugs at cs.uiuc.edu LLVM won't build on this platform. I am using the RedHat binary of the llvm-gcc frontend and gcc 4.1.2. Build error is: llvm[2]: Linking Release executable tblgen (without symbols) /usr/lib/libltdl.so: could not read symbols: File in wrong format collect2: ld returned 1 exit status When I configure with CC="gcc34 -m32" and CXX="g++34 -m32" the build completes. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 7 20:03:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 7 Dec 2007 20:03:11 -0600 Subject: [LLVMbugs] [Bug 1845] New: <4 x i1> & <4 x i8> can break in compiler in TargetData:: getAlignmentInfo Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1845 Summary: <4 x i1> & <4 x i8> can break in compiler in TargetData::getAlignmentInfo Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: major Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: cfr at adobe.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1280) --> (http://llvm.org/bugs/attachment.cgi?id=1280) Modification of the fibonacci example The following function will fail to compile: ; ModuleID = 'test' define void @boolVectorSelect(<4 x i1>* %boolVectorPtr) { Body: %castPtr = bitcast <4 x i1>* %boolVectorPtr to <4 x i1>* ; <<4 x i1>*> [#uses=1] %someBools = load <4 x i1>* %castPtr, align 1 ; <<4 x i1>> [#u ses=1] %internal = alloca <4 x i1>, align 16 ; <<4 x i1>*> [#uses=1] store <4 x i1> %someBools, <4 x i1>* %internal, align 1 ret void } verifying... Assertion failed: BestMatchIdx != -1 && "Didn't find alignment info for this datatype!", file ..\..\lib\Target\TargetData.cpp, line 303 Which can be repro'd with the attached program. (based on fibonacci llvm example) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Dec 8 22:15:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 8 Dec 2007 22:15:39 -0600 Subject: [LLVMbugs] [Bug 1846] New: error compiling llvm-gcc4.2 for iphone (arm) darwin Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1846 Summary: error compiling llvm-gcc4.2 for iphone (arm) darwin Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: blocker Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: K3DavidK3 at gmail.com CC: llvmbugs at cs.uiuc.edu configure opts: ./configure --enable-llvm=`llvm-config --obj-root` \ --enable-languages=c,c++,objc,obj-c++ --target=arm-apple-darwin --enable-sjlj-exceptions \ --with-heavenly=$HEAVENLY --with-as=/usr/local/bin/arm-apple-darwin-as \ --with-ld=/usr/local/bin/arm-apple-darwin-ld ... make ... make[2]: Entering directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/g cc' test -d po || mkdir po : --statistics -o po/be.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/be.po test -d po || mkdir po : --statistics -o po/ca.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/ca.po test -d po || mkdir po : --statistics -o po/da.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/da.po test -d po || mkdir po : --statistics -o po/de.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/de.po test -d po || mkdir po : --statistics -o po/el.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/el.po test -d po || mkdir po : --statistics -o po/es.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/es.po test -d po || mkdir po : --statistics -o po/fr.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/fr.po test -d po || mkdir po : --statistics -o po/ja.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/ja.po test -d po || mkdir po : --statistics -o po/nl.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/nl.po test -d po || mkdir po : --statistics -o po/ru.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/ru.po test -d po || mkdir po : --statistics -o po/rw.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/rw.po test -d po || mkdir po : --statistics -o po/sr.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/sr.po test -d po || mkdir po : --statistics -o po/sv.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/sv.po test -d po || mkdir po : --statistics -o po/tr.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/tr.po test -d po || mkdir po : --statistics -o po/zh_CN.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/zh_CN.po test -d po || mkdir po : --statistics -o po/zh_TW.gmo ../../../llvm-gcc4.2-2.1.source/gcc/po/zh_TW.po make \ CFLAGS="-g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing- prototypes -Wold-style-definition -Wmissing-format-attribute " \ CONFIG_H="config.h auto-host.h ../../../llvm-gcc4.2-2.1.source/gcc/.. /include/ansidecl.h ../../../llvm-gcc4.2-2.1.source/gcc/config/i386/xm-cygwin.h" \ MAKEOVERRIDES= \ -f libgcc.mk all make[3]: Entering directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/g cc' make GCC_FOR_TARGET="/home/Administrator/build/llvm-gcc-4.2-2.1.source/./gcc/xgc c -B/home/Administrator/build/llvm-gcc-4.2-2.1.source/./gcc/ -B/usr/local/arm-ap ple-darwin/bin/ -B/usr/local/arm-apple-darwin/lib/ -isystem /usr/local/arm-apple -darwin/include -isystem /usr/local/arm-apple-darwin/sys-include" \ AR_FOR_TARGET="arm-apple-darwin-ar" \ AR_CREATE_FOR_TARGET="arm-apple-darwin-ar rc" \ AR_EXTRACT_FOR_TARGET="arm-apple-darwin-ar x" \ AR_FLAGS_FOR_TARGET="" \ CC="gcc" CFLAGS="-g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes - Wmissing-prototypes -Wold-style-definition -Wmissing-format-attribute " \ BUILD_PREFIX="" \ BUILD_PREFIX_1="" \ LANGUAGES="" \ LIBGCC2_CFLAGS="-O2 -O2 -g -O2 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style- definition -isystem ./include -fPIC -pipe -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc " \ MULTILIB_CFLAGS="" T= crt3.o make[4]: Entering directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/g cc' # APPLE LOCAL use -mlongcall for large text support /home/Administrator/build/llvm-gcc-4.2-2.1.source/./gcc/xgcc -B/home/Administrat or/build/llvm-gcc-4.2-2.1.source/./gcc/ -B/usr/local/arm-apple-darwin/bin/ -B/us r/local/arm-apple-darwin/lib/ -isystem /usr/local/arm-apple-darwin/include -isys tem /usr/local/arm-apple-darwin/sys-include -O2 -g -O2 -DIN_GCC -DCROSS_DIRECTO RY_STRUCTURE -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -I. -I. -I../../../llvm-gcc4.2-2.1. source/gcc -I../../../llvm-gcc4.2-2.1.source/gcc/. -I../../../llvm-gcc4.2-2.1.so urce/gcc/../include -I../../../llvm-gcc4.2-2.1.source/gcc/../libcpp/include -I. ./../../llvm-gcc4.2-2.1.source/gcc/../libdecnumber -I../libdecnumber -I/tmp/llvm -2.1/include -I/tmp/llvm-2.1/.build/include -mlongcall \ -fno-tree-dominator-opts \ \ -c ../../../llvm-gcc4.2-2.1.source/gcc/config/darwin-crt3.c -o crt3.o cc1: internal error: #pragma pack is already registered make[4]: *** [crt3.o] Error 1 make[4]: Leaving directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/gc c' make[3]: *** [extra] Error 2 make[3]: Leaving directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/gc c' make[2]: *** [stmp-multilib] Error 2 make[2]: Leaving directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source/gc c' make[1]: *** [all-gcc] Error 2 make[1]: Leaving directory `/home/Administrator/build/llvm-gcc-4.2-2.1.source' make: *** [all] Error 2 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 01:32:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 01:32:55 -0600 Subject: [LLVMbugs] [Bug 1846] error compiling llvm-gcc4.2 for iphone (arm) darwin In-Reply-To: Message-ID: <200712090732.lB97WtaY012001@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1846 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #1 from Chris Lattner 2007-12-09 01:32:54 --- sorry, but this is not a supported configuration. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 04:35:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 04:35:38 -0600 Subject: [LLVMbugs] [Bug 1847] New: clang hangs with -warn-uninit-values Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1847 Summary: clang hangs with -warn-uninit-values Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu $ clang -warn-uninit-values str3.i ... doesn't finish even after minutes ... $ cat str3.i char *cli_strrcpy(char *dest, const char *source) { if(!dest || !source) { return ((void *)0); } while((*dest++ = *source++)); return --dest; } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 09:24:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 09:24:46 -0600 Subject: [LLVMbugs] [Bug 1848] New: bad spacing with -E preprocessed output Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1848 Summary: bad spacing with -E preprocessed output Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: minor Priority: P2 Component: preprocessor AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu #define t(x) x t(8 __LINE__) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 12:04:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 12:04:17 -0600 Subject: [LLVMbugs] [Bug 1847] clang hangs with -warn-uninit-values In-Reply-To: Message-ID: <200712091804.lB9I4Huq017531@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1847 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #1 from Ted Kremenek 2007-12-09 12:04:16 --- I cannot reproduce this bug. T??r??k: What revision of clang are you using (svn info)? Mine is 44739. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 13:44:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 13:44:37 -0600 Subject: [LLVMbugs] [Bug 1849] New: llvm-gcc -emit-llvm does not generate exe Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1849 Summary: llvm-gcc -emit-llvm does not generate exe Product: tools Version: 2.1 Platform: PC OS/Version: Windows XP Status: NEW Severity: major Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: anteusz at freemail.hu CC: llvmbugs at cs.uiuc.edu llvm-gcc -emit-llvm does not generate exe because of an error... Something like temp\.o has wrong format... It was explained to me that this is because a bug in llvm-gcc.. creates llvm coded object file and it passes them to gcc linker which -of course- does not expect such a format.. In short, a step is missing... linking the object files with its llvm linker... -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 14:16:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 14:16:59 -0600 Subject: [LLVMbugs] [Bug 1849] llvm-gcc lacks proper linker support for llvm IR files. In-Reply-To: Message-ID: <200712092016.lB9KGxdP026274@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1849 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #3 from Chris Lattner 2007-12-09 14:16:59 --- *** This bug has been marked as a duplicate of bug 948 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 9 15:11:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 9 Dec 2007 15:11:40 -0600 Subject: [LLVMbugs] [Bug 1848] bad spacing with -E preprocessed output In-Reply-To: Message-ID: <200712092111.lB9LBexI029658@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1848 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Chris Lattner 2007-12-09 15:11:40 --- The second half is fixed here, thanks! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20071203/003232.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 04:59:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 04:59:49 -0600 Subject: [LLVMbugs] [Bug 1850] New: comparing pointer constants gives surprising results Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1850 Summary: comparing pointer constants gives surprising results Product: libraries Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1281) --> (http://llvm.org/bugs/attachment.cgi?id=1281) test case The attached test case is based on gcc.c-torture/execute/20010329-1.c in the GCC testsuite. If I compile it with "llvm-gcc -O1 -S -emit-llvm pre.c", I get this: define i32 @main() { entry: tail call void @abort( ) unreachable } This is a bit surprising, because the test looks like it should obviously pass (i.e. call exit(0) instead of abort()). You could probably argue that the results of the pointer comparisons are undefined, because the pointers don't point into any valid object, but I still think it would be nice if the code behaved in the obvious way. I've investigated a bit, and the GCC front end is generating code a bit like this: %x = inttoptr i64 2147483649 to i8* ; 0x80000001 %x1 = getelementptr i8* %x, i64 2147483648 ; 0x80000000 %cond = icmp ule i8* %x1, inttoptr (i64 5 to i8*) It's a bit strange that it adds 0x80000000 to x instead of subtracting it, but if pointers are 32 bit then the result will be the same. Then, after -instcombine, this is folded to false, presumably because the operations have been evaluated using 64-bit arithmetic. So maybe -instcombine is wrong to assume that pointers are 64 bit. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 07:58:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 07:58:01 -0600 Subject: [LLVMbugs] [Bug 1851] New: instcombine removes stacksave/stackrestore around alloca Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1851 Summary: instcombine removes stacksave/stackrestore around alloca Product: libraries Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1282) --> (http://llvm.org/bugs/attachment.cgi?id=1282) test case The attached test case is based on gcc.c-torture/execute/20040811-1.c in the GCC testsuite. If I run it through "opt -f -o post.o pre.o -instcombine" I get this: loop: ; preds = %cleanup, %entry %n = phi i32 [ 1, %entry ], [ %phitmp, %cleanup ] ; [#uses=2] %mem1 = alloca [1000 x i32] ; <[1000 x i32]*> [#uses=1] %mem1.sub = getelementptr [1000 x i32]* %mem1, i32 0, i32 0 ; [#uses=1] call void @foo( i32* %mem1.sub ) %cond = icmp slt i32 %n, 1000000 ; [#uses=1] br i1 %cond, label %cleanup, label %return cleanup: ; preds = %loop %phitmp = add i32 %n, 1 ; [#uses=1] br label %loop It has completely removed the calls to llvm.stacksave and llvm.stackrestore around the large alloca! I think the problem is that InstructionCombining.cpp's TryToSinkInstruction() thinks it's OK to sink the call to llvm.stacksave past the alloca. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 08:43:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 08:43:41 -0600 Subject: [LLVMbugs] [Bug 1852] New: GCC 4.0 front end won't build for a target that doesn' t support DWARF exception handling Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1852 Summary: GCC 4.0 front end won't build for a target that doesn't support DWARF exception handling Product: libraries Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1283) --> (http://llvm.org/bugs/attachment.cgi?id=1283) patch (I couldn't find a bugzilla product/component for the GCC front end.) When building the GCC 4.0 front end for LLVM 2.1, configured for a proprietary target that doesn't support DWARF debug info or exception handling, I got this: .../gcc/llvm-convert.cpp: In member function ???bool TreeToLLVM::EmitBuiltinEHReturnDataRegno(tree_node*, llvm::Value*&)???: .../gcc/llvm-convert.cpp:4749: error: ???EH_RETURN_DATA_REGNO??? was not declared in this scope .../gcc/llvm-convert.cpp: In member function ???bool TreeToLLVM::EmitBuiltinInitDwarfRegSizes(tree_node*, llvm::Value*&)???: .../gcc/llvm-convert.cpp:4811: error: ???DWARF_FRAME_RETURN_COLUMN??? was not declared in this scope .../gcc/llvm-convert.cpp:4828: error: ???DWARF_FRAME_RETURN_COLUMN??? was not declared in this scope make: *** [llvm-convert.o] Error 1 I've attached an obvious patch to fix this. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 11:44:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 11:44:24 -0600 Subject: [LLVMbugs] [Bug 1836] lli interpreter crashed for integer type whose bitwidth > 64 In-Reply-To: Message-ID: <200712101744.lBAHiOKM027368@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1836 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Duncan Sands 2007-12-10 11:44:24 --- Fixed by: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056137.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 12:06:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 12:06:56 -0600 Subject: [LLVMbugs] [Bug 1851] instcombine removes stacksave/stackrestore around alloca In-Reply-To: Message-ID: <200712101806.lBAI6uRG028408@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1851 Dale Johannesen changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dalej at apple.com Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #1 from Dale Johannesen 2007-12-10 12:06:55 --- I reported this as PR 1745 and it appears to be fixed in top of tree. *** This bug has been marked as a duplicate of bug 1745 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 12:38:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 12:38:39 -0600 Subject: [LLVMbugs] [Bug 1847] clang hangs with -warn-uninit-values In-Reply-To: Message-ID: <200712101838.lBAIcdm8029922@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1847 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|VERIFIED |REOPENED Resolution|WORKSFORME | -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 14:15:48 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 14:15:48 -0600 Subject: [LLVMbugs] [Bug 1853] New: Assert being triggered on valid ExtractElementInst Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1853 Summary: Assert being triggered on valid ExtractElementInst Product: new-bugs Version: unspecified Platform: PC OS/Version: other Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: cfr at adobe.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1284) --> (http://llvm.org/bugs/attachment.cgi?id=1284) Modification of the fibonacci example to repro the bug. Here is the LLVM IR. Basic idea is that you have a pointer to a <4 x float> and a i32 index. You load the vector, extract the specified element and return it. ; ModuleID = 'test' define float @floatVectorSelect(<4 x float>* %floatVectorPtr, i32 %index) { Body: %floatVector = load <4 x float>* %floatVectorPtr, align 1 ; <<4 x float>> [#uses=1] %selectedFloat = extractelement <4 x float> %floatVector, i32 %index ; [#uses=1] ret float %selectedFloat } verifying... Assertion failed: N2C && "Bad EXTRACT_VECTOR_ELT!", file ..\..\lib\ CodeGen\SelectionDAG\SelectionDAG.cpp, line 2111 I found this under the x86 JIT using the attached modification to the fibonacci program. I am synced to the mainline as of 11am 10-Dec-2007. Thanks, Chuck. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 16:04:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 16:04:38 -0600 Subject: [LLVMbugs] [Bug 1847] clang hangs with -warn-uninit-values In-Reply-To: Message-ID: <200712102204.lBAM4cu9008499@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1847 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #7 from Ted Kremenek 2007-12-10 16:04:38 --- Fixed with patch 44813: http://llvm.org/viewvc/llvm-project?rev=44813&view=rev -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 10 16:53:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 10 Dec 2007 16:53:43 -0600 Subject: [LLVMbugs] [Bug 1850] instcombine miscompiles pointer comparison In-Reply-To: Message-ID: <200712102253.lBAMrh2X011265@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1850 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |miscompilation Resolution| |FIXED Target Milestone|--- |2.2 --- Comment #5 from Chris Lattner 2007-12-10 16:53:43 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056149.html Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 05:26:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 05:26:45 -0600 Subject: [LLVMbugs] [Bug 1850] instcombine miscompiles pointer comparison In-Reply-To: Message-ID: <200712111126.lBBBQjJk001770@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1850 Jay Foad changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #7 from Jay Foad 2007-12-11 05:26:38 --- This bit of the fix looks obviously wrong, because an IntToPtr only has one operand: + if ((CE0->getOpcode() == Instruction::IntToPtr && + CE0->getOperand(0)->getType() == IntPtrTy && + CE0->getOperand(1)->getType() == IntPtrTy) || -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 17:18:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 17:18:38 -0600 Subject: [LLVMbugs] [Bug 1805] llvm-g++ emits trees for computed labels too late In-Reply-To: Message-ID: <200712112318.lBBNIcV2002656@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1805 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #8 from Devang Patel 2007-12-11 17:18:37 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056203.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 19:18:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 19:18:43 -0600 Subject: [LLVMbugs] [Bug 1854] New: nightly tester is not recording users who check out from svn Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1854 Summary: nightly tester is not recording users who check out from svn 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: tonic at nondot.org CC: llvmbugs at cs.uiuc.edu It appears that svn does not log users who checkout from svn. Therefore, the nightly tester fails to track this information like we did with cvs. If someone knows how to get this information out of svn, please comment. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 20:32:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 20:32:14 -0600 Subject: [LLVMbugs] [Bug 1855] New: __sync_synchronize() Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1855 Summary: __sync_synchronize() Product: tools Version: trunk Platform: PC OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: dpatel at apple.com CC: llvmbugs at cs.uiuc.edu llvm-gcc is not handling __sync_synchronize() builtin. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 20:34:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 20:34:40 -0600 Subject: [LLVMbugs] [Bug 1856] New: atomic builtins Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1856 Summary: atomic builtins Product: tools Version: trunk Platform: PC OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: dpatel at apple.com CC: llvmbugs at cs.uiuc.edu llvm-gcc does not support atomic builtins, e.g. __sync_fetch_and_add -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 11 21:57:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 11 Dec 2007 21:57:19 -0600 Subject: [LLVMbugs] [Bug 1850] instcombine miscompiles pointer comparison In-Reply-To: Message-ID: <200712120357.lBC3vJo1028679@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1850 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #9 from Chris Lattner 2007-12-11 21:57:19 --- Excellent catch, you're exactly right: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056218.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 00:37:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 00:37:55 -0600 Subject: [LLVMbugs] [Bug 1629] lli with -force-interpreter will cause assertion when having static class object in function In-Reply-To: Message-ID: <200712120637.lBC6btf7011137@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1629 Sheng Zhou changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 12:37:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 12:37:42 -0600 Subject: [LLVMbugs] [Bug 1857] New: Patch for Edge Profiling Correctness Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1857 Summary: Patch for Edge Profiling Correctness Product: libraries Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: Miscellaneous Instrumentation passes AssignedTo: unassignedbugs at nondot.org ReportedBy: mdekruijf at gmail.com CC: llvmbugs at cs.uiuc.edu The diff below makes edge profiling work correctly. Previously, function and basic block counts using edge profiling were inconsistent with those found doing block/function profiling directly. The code now also matches the comment. $ svn diff lib/Transforms/Instrumentation/EdgeProfiling.cpp Index: EdgeProfiling.cpp =================================================================== --- EdgeProfiling.cpp (revision 4) +++ EdgeProfiling.cpp (working copy) @@ -84,7 +84,7 @@ // Okay, we are guaranteed that the edge is no longer critical. If we // only have a single successor, insert the counter in this block, // otherwise insert it in the successor block. - if (TI->getNumSuccessors() == 0) { + if (TI->getNumSuccessors() == 1) { // Insert counter at the start of the block IncrementCounterInBlock(BB, i++, Counters); } else { -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 12:47:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 12:47:59 -0600 Subject: [LLVMbugs] [Bug 1858] New: -warn-uninit-values sizeof() handling wrong Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1858 Summary: -warn-uninit-values sizeof() handling wrong Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu I warnings like these: ../../../libclamav/rtf.c:212:52: warning: use of uninitialized variable struct rtf_object_data* data = cli_malloc(sizeof(*data)); ^ This way of mallocing is my favorite, and I use it in lots of places, hence I get lots of these warnings. I think it should be safe to exclude sizeof() from uninit-var checks. sizeof() is evaluated at compile-time, so if its argument would be "uninitialized", the compiler wouldn't be able to compute it, hence I'd get a compile-error, right? -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 12:53:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 12:53:36 -0600 Subject: [LLVMbugs] [Bug 1859] New: -warn-uninit-values and "output" parameters Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1859 Summary: -warn-uninit-values and "output" parameters Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu I get warnings when I have pointer parameters that are written to by called function, and its contents is uninitialized prior to call ../../../libclamav/others.c:460:12: warning: use of uninitialized variable memcpy(salt, name_salt, 16); ^ This is the code: unsigned char salt[16 + 32]; memcpy(salt, name_salt, 16); for(i = 16; i < 48; i++) salt[i] = cli_rndnum(256); Hmm, there's nothing in the prototype of a function that tells "this parameter is only written to", so yes, clang has no way to know this generally. Maybe it could have "built-in knowledge" of common libc functions. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 18:05:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 18:05:04 -0600 Subject: [LLVMbugs] [Bug 1857] Patch for Edge Profiling Correctness In-Reply-To: Message-ID: <200712130005.lBD054I2021322@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1857 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2007-12-12 18:05:04 --- Thanks! http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056276.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 22:49:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 22:49:30 -0600 Subject: [LLVMbugs] [Bug 1858] -warn-uninit-values sizeof() handling wrong In-Reply-To: Message-ID: <200712130449.lBD4nUcn013636@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1858 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Ted Kremenek 2007-12-12 22:49:30 --- Fixed in r44982: http://llvm.org/viewvc/llvm-project?rev=44982&view=rev -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 12 23:19:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 12 Dec 2007 23:19:17 -0600 Subject: [LLVMbugs] [Bug 1859] -warn-uninit-values and "output" parameters In-Reply-To: Message-ID: <200712130519.lBD5JHbI015138@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1859 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Ted Kremenek 2007-12-12 23:19:17 --- Fixed in r44984 (http://llvm.org/viewvc/llvm-project?rev=44984&view=rev). The problem here is that "salt" really refers to a memory block on the stack. This is valid, allocated memory, and so expressions such as "(char *) salt" really do refer to a valid (i.e. initialized) pointer address. What is uninitialized is the contents of "salt", but these values are not being read anyway (memcpy writes to them). Further, the dataflow analysis does not reason about the contents of arrays, as the precision needed for this requires more machinery than what is employed by this simple dataflow analysis. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 11:38:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 11:38:27 -0600 Subject: [LLVMbugs] [Bug 1837] lli: free(): invalid next size (fast) In-Reply-To: Message-ID: <200712131738.lBDHcRoJ013857@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1837 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |LLVM Codegen Product|new-bugs |clang Resolution| |FIXED --- Comment #8 from Chris Lattner 2007-12-13 11:38:26 --- Excellent reduction, fix as r45003, see llvmdev for details. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 13:18:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 13:18:12 -0600 Subject: [LLVMbugs] [Bug 1860] New: linker infinate loops in when linking types Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1860 Summary: linker infinate loops in when linking types Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu CC: llvmbugs at cs.uiuc.edu RecursiveResolveTypesI in LinkModules.cpp infinately loops when linking the two attached files with: llvm-ld -disable-opt -v -r abc.o d.opt.bc -o abcd.o -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 16:07:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 16:07:18 -0600 Subject: [LLVMbugs] [Bug 1861] New: "LLVM type size doesn't match GCC type size!" Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1861 Summary: "LLVM type size doesn't match GCC type size!" Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1291) --> (http://llvm.org/bugs/attachment.cgi?id=1291) reduction assertion in head with attached file (x86/linux/32bit). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 16:11:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 16:11:10 -0600 Subject: [LLVMbugs] [Bug 1861] "LLVM type size doesn't match GCC type size!" In-Reply-To: Message-ID: <200712132211.lBDMBAd0027318@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1861 Andrew Lenharth changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #1 from Andrew Lenharth 2007-12-13 16:11:09 --- reduction is not valid. will try again. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 17:18:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 17:18:13 -0600 Subject: [LLVMbugs] [Bug 1861] "LLVM type size doesn't match GCC type size!" In-Reply-To: Message-ID: <200712132318.lBDNIDYM031600@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1861 Andrew Lenharth changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|enhancement |normal Status|RESOLVED |REOPENED Resolution|INVALID | --- Comment #3 from Andrew Lenharth 2007-12-13 17:18:12 --- here is a reduction that triggers the bug in llvm-gcc but passes through gcc just fine. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 18:39:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 18:39:23 -0600 Subject: [LLVMbugs] [Bug 1782] Assertion `Offset1 != Offset2 && " There is at least one different constant here!"' failed In-Reply-To: Message-ID: <200712140039.lBE0dNO0002285@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1782 Wojciech Matyjewicz changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Global Analyses Keywords| |compile-fail Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.2 Version|unspecified |trunk --- Comment #6 from Wojciech Matyjewicz 2007-12-13 18:39:23 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071203/056099.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Dec 13 18:49:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 13 Dec 2007 18:49:40 -0600 Subject: [LLVMbugs] [Bug 1858] -warn-uninit-values sizeof() handling wrong In-Reply-To: Message-ID: <200712140049.lBE0nexm002748@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1858 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|CLOSED |REOPENED Resolution|FIXED | --- Comment #3 from Ted Kremenek 2007-12-13 18:49:39 --- Special care must be taken for VLAs where the sizeof can actual cause its subexpression to be evaluated. For example: int bar() { printf("World\n"); return 5; } int f() { printf("Hello\n"); return 3; } main() { typedef int a[f()]; return sizeof (a[bar()]); } The output of this program is: Hello World -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 14 12:30:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 14 Dec 2007 12:30:41 -0600 Subject: [LLVMbugs] [Bug 1862] New: verify fails based on ordering even with -disable-opt Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1862 Summary: verify fails based on ordering even with -disable-opt Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu CC: llvmbugs at cs.uiuc.edu llvm-ld -disable-opt -r -o foo.o a.bc b.bc c.bc d.bc e.bc GEP is not of right type for indices! %tmp157 = getelementptr %struct.inode* %tmp27, i32 0, i32 23 ; <%struct.inode_operations**> [#uses=1] { i32 (%struct.inode*, %struct.dentry*, i32, %struct.nameidata*)*, %struct.dentry* (%struct.inode*, %struct.dentry*, %struct.nameidata*)*, i32 (%struct.dentry*, %struct.inode*, %struct.dentry*)*, i32 (%struct.inode*, %struct.dentry*)*, i32 (%struct.inode*, %struct.dentry*, i8*)*, i32 (%struct.inode*, %struct.dentry*, i32)*, i32 (%struct.inode*, %struct.dentry*)*, i32 (%struct.inode*, %struct.dentry*, i32, i32)*, i32 (%struct.inode*, %struct.dentry*, %struct.inode*, %struct.dentry*)*, i32 (%struct.dentry*, i8*, i32)*, i8* (%struct.dentry*, %struct.nameidata*)*, void (%struct.dentry*, %struct.nameidata*, i8*)*, void (%struct.inode*)*, i32 (%struct.inode*, i32, %struct.nameidata*)*, i32 (%struct.dentry*, %struct.iattr*)*, i32 (%struct.vfsmount*, %struct.dentry*, %struct.kstat*)*, i32 (%struct.dentry*, i8*, i8*, i32, i32)*, i32 (%struct.dentry*, i8*, i8*, i32)*, i32 (%struct.dentry*, i8*, i32)*, i32 (%struct.dentry*, i8*)*, void (%struct.inode*, i64, i64)*, i32 (%struct.inode*, i32, i64, i64)* }*Instruction does not dominate all uses! %tmp157 = getelementptr %struct.inode* %tmp27, i32 0, i32 23 ; <%struct.inode_operations**> [#uses=1] store %struct.inode_operations* %tmp148, %struct.inode_operations** %tmp157, align 4 Broken module found, compilation aborted! however, opt -preverify -verify passes each individual module also llvm-ld -disable-opt -r -o bar.o a.bc b.bc c.bc d.bc llvm-ld -disable-opt -r -o foo.o bar.o f.bc works (and so does swaping a and f, I haven't tried other combinations). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 14 15:30:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 14 Dec 2007 15:30:20 -0600 Subject: [LLVMbugs] [Bug 1863] New: parser crash involving isConstantExpr() on incorrect code Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1863 Summary: parser crash involving isConstantExpr() on incorrect code Product: clang Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: parser AssignedTo: unassignedbugs at nondot.org ReportedBy: kremenek at cs.stanford.edu CC: llvmbugs at cs.uiuc.edu The following (incorrect) code breaks the parser: int foo(); int test() {} typedef int x[foo()]; int y = sizeof(x); int z = sizeof(int[foo()]); return y + z; } from gdb: (gdb) r Starting program: /Volumes/Data/Users/kremenek/llvm/Debug/bin/clang /tmp/t4.c Reading symbols for shared libraries ++++. done /tmp/t4.c:4:15: error: variable length array declared outside of any function typedef int x[foo()]; ^ ~~~~~ Assertion failed: (loc.isValid() && "isConstantExpr didn't return a loc!"), function CheckSingleInitializer, file SemaDecl.cpp, line 362. Program received signal SIGABRT, Aborted. Notice the extra "}" on the line "int test() {}". I'm not certain what the correct behavior should be in this case, but it certainly shouldn't be a crash. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 14 17:10:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 14 Dec 2007 17:10:51 -0600 Subject: [LLVMbugs] [Bug 1858] -warn-uninit-values sizeof() handling wrong In-Reply-To: Message-ID: <200712142310.lBENApTC007756@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1858 Ted Kremenek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #5 from Ted Kremenek 2007-12-14 17:10:50 --- sizeof handling hopefully now fixed for VLA types in r45043: http://llvm.org/viewvc/llvm-project?rev=45043&view=rev -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Dec 16 16:42:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 16 Dec 2007 16:42:40 -0600 Subject: [LLVMbugs] [Bug 1864] New: Excess newlines in -E output Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1864 Summary: Excess newlines in -E output Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: minor Priority: P2 Component: preprocessor AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu #define str(x) #x str() str() -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 00:55:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 00:55:45 -0600 Subject: [LLVMbugs] [Bug 1865] New: GEPs aren't amenable to further analysis Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1865 Summary: GEPs aren't amenable to further analysis Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Loop Optimizer AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: llvmbugs at cs.uiuc.edu This example is taken from the SLP paper as an example of something that isn't vectorizable: void f(int *dst, int *end, int *src1, int *src2) { do { dst[0] = (src1[0] + src2[0]) >> 1; dst[1] = (src1[1] + src2[1]) >> 1; dst[2] = (src1[2] + src2[2]) >> 1; dst[3] = (src1[3] + src2[3]) >> 1; dst += 4; src1 += 4; src2 += 4; } while (dst != end); return; } The problem is that dst[1] is calculated using a variable GEP. Looking just at the instructions relevant to the computation of dst[1], we see: %indvar = phi i32 [ 0, %entry ], [ %indvar.next, %bb ] ; [#uses=2] %src1_addr.0.rec = shl i32 %indvar, 2 ; [#uses=7] %src1_addr.0.sum5657 = or i32 %src1_addr.0.rec, 1 ; [#uses=3] %tmp19 = getelementptr i32* %dst, i32 %src1_addr.0.sum5657 ; [#uses=1] It would be much nicer if we could say that %tmp19 = GEP %dst, 1. With this change, I think the code from the example would be vectorizable. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 13:26:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 13:26:04 -0600 Subject: [LLVMbugs] [Bug 1866] New: clang says "byte" is not assignable Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1866 Summary: clang says "byte" is not assignable Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu testcase-min.i:10:12: error: expression is not assignable byte <<= 4; ~~~~ ^ 1 diagnostic generated. Testcase: static char * rfc2231(const char *in) { const char *ptr; enum { LANGUAGE, CHARSET, CONTENTS } field; while(*ptr) { switch(field) { case CONTENTS: if(*ptr == '%') { unsigned char byte; byte <<= 4; } } } } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 13:38:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 13:38:13 -0600 Subject: [LLVMbugs] [Bug 1867] New: typedef-ing the same type twice in different headers Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1867 Summary: typedef-ing the same type twice in different headers Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: parser AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu GCC allows this, and compiles the code. Clang gives an error. Different headers declare the same type. GCC only allows it if the headers are different, if I remove the # 328, then I get an error from gcc too. As long as the types are compatible, is this a bug in the source code, or in clang? $ gcc -c headers.i $ clang headers.i headers.i:4:23: error: redefinition of 'Byte' typedef unsigned char Byte; ^ headers.i:2:23: error: previous definition is here typedef unsigned char Byte; ^ 2 diagnostics generated. headers.i: # 13 "../../../libclamav/lzma/LzmaTypes.h" typedef unsigned char Byte; # 328 "/usr/include/zconf.h" 3 4 typedef unsigned char Byte; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 14:02:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 14:02:01 -0600 Subject: [LLVMbugs] [Bug 1868] New: PassManager does not reschedule invalidated passes Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1868 Summary: PassManager does not reschedule invalidated passes 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: criswell at uiuc.edu CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1300) --> (http://llvm.org/bugs/attachment.cgi?id=1300) LLVM Source file that will trigger the bug I get the following assertion failure in my tool that schedules several LLVM passes: test: /home/vadve/criswell/src/llvm22/lib/VMCore/PassManager.cpp:1285: virtual void llvm::MPPassManager::addLowerLevelRequiredPass(llvm::Pass*, llvm::Pass*): Assertion `(P->getPotentialPassManagerType() < RequiredPass->getPotentialPassManagerType()) && "Unable to handle Pass that requires lower level Analysis pass"' failed. My investigation currently leads me to believe that when a pass is invalidated, the Pass Manager does not properly reschedule it. In the attached code, I have three passes: BottomPass, Pass1, and Pass2. Pass1 preserves all analyses; Pass2 preserves none. BottomPass requires both Pass1 and Pass2. When BottomPass is added to the PassManager, PassManager will first schedule Pass1 and Pass2. However, Pass2 invalidates Pass1, so Pass1 is taken off the AvailableAnalysis list in PMDataManager. Later on, PassManager goes looking for Pass1 and, not finding it, assumes it's a lower level pass. However, since it's a ModulePass (like BottomPass), it hits the assertion and aborts. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 15:09:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 15:09:21 -0600 Subject: [LLVMbugs] [Bug 1866] clang says "byte" is not assignable In-Reply-To: Message-ID: <200712172109.lBHL9LXC025311@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1866 T??r??k Edwin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #3 from T??r??k Edwin 2007-12-17 15:09:20 --- (In reply to comment #2) > > I am svn up-ing now, and compiling. > It no longer gives this error. Sorry for the noise. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 15:31:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 15:31:31 -0600 Subject: [LLVMbugs] [Bug 1868] PassManager does not reschedule invalidated passes In-Reply-To: Message-ID: <200712172131.lBHLVVU6027301@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1868 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dpatel at apple.com Status|NEW |RESOLVED Resolution| |WONTFIX --- Comment #1 from Devang Patel 2007-12-17 15:31:30 --- PassManager schedules Pass1 for BottomPass. However it is invalidated when PassManager scheduled Pass2. PassManager does not dependency analysis of required passes to ensure that they do not step on each other's foot. You have two workarounds: 1) Pass2 preserves Pass1 or 2) BottomPass requires Pass2 before Pass1. At the moment this is IMO not to be fixed because it may add lots of unnecessary complexity in PassManager implementation for little gain. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 19:05:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 19:05:20 -0600 Subject: [LLVMbugs] [Bug 1869] New: C++ Frontend ICE Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1869 Summary: C++ Frontend ICE Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: isanbard at gmail.com CC: llvmbugs at cs.uiuc.edu Top-of-tree LLVM-GCC 4.2 fails the regression tests with this error: $ make RUNTESTFLAGS=C++Frontend/dg.exp ( ulimit -t 600 ; ulimit -d 512000 ; \ PATH="/Volumes/Gir/devel/llvm/llvm.obj/Debug/bin:/Volumes/Gir/devel/llvm/llvm.src/test/Scripts:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin:/usr/local/bin:/usr/local/sbin:/Users/void/install/bin:/sw/bin:/sw/sbin:/opt/intel/cc/9.1.037/bin:/opt/intel/idb/9.1.037/bin:/Applications/Graphviz.app/Contents/MacOS:/usr/local/mysql/bin:/Applications/Doxygen.app/Contents/Resources:/Volumes/Gaz/Applications/Graphviz.app/Contents/MacOS:/Volumes/Dib/Applications/Graphviz.app/Contents/MacOS:/Volumes/Zim/Applications/Graphviz.app/Contents/MacOS" \ /usr/bin/runtest C++Frontend/dg.exp ) WARNING: Couldn't find the global config file. WARNING: No tool specified Test Run By void on Mon Dec 17 17:00:37 2007 Native configuration is i686-apple-darwin9.1.0 === tests === Schedule of variations: unix Running target unix Using /usr/share/dejagnu/baseboards/unix.exp as board description file for target. Using /usr/share/dejagnu/config/unix.exp as generic interface file for target. WARNING: Couldn't find tool config file for unix, using default. Running /Volumes/Gir/devel/llvm/llvm.src/test/C++Frontend/dg.exp ... FAIL: /Volumes/Gir/devel/llvm/llvm.src/test/C++Frontend/2007-03-27-FunctionVarRename.cpp Failed with exit(1) at line 1 while running: /Volumes/Gir/devel/llvm/llvm-gcc-4.2.install/bin/llvm-g++ -I/usr/lib/gcc//i686-apple-darwin9/4.2.1/include -I/usr/include/c++/4.0.0 -I/usr/include/c++/4.0.0/i686-apple-darwin9 -emit-llvm /Volumes/Gir/devel/llvm/llvm.src/test/C++Frontend/2007-03-27-FunctionVarRename.cpp -emit-llvm -S -o - | not grep eprintf1 Assertion failed: (FnEntry->getName() == Name &&"Preexisting fn with the same name!"), function make_decl_llvm, file ../../llvm-gcc-4.2.src/gcc/llvm-backend.cpp, line 1058. /Volumes/Gir/devel/llvm/llvm.src/test/C++Frontend/2007-03-27-FunctionVarRename.cpp: In function ???void foo()???: /Volumes/Gir/devel/llvm/llvm.src/test/C++Frontend/2007-03-27-FunctionVarRename.cpp:9: internal compiler error: Abort trap Please submit a full bug report, with preprocessed source if appropriate. See for instructions. === Summary === # of expected passes 81 # of unexpected failures 1 # of expected failures 1 make: *** [check-local] Error 1 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 17 19:32:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 17 Dec 2007 19:32:17 -0600 Subject: [LLVMbugs] [Bug 1864] Excess newlines in -E output In-Reply-To: Message-ID: <200712180132.lBI1WHXa007820@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1864 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #2 from Chris Lattner 2007-12-17 19:32:16 --- We currently produce: "" "" Which seems pretty reasonable, and it matches what GCC produces. I'm closing this as invalid, but if you think its a bug in both GCC and clang, please include the expected output and the GCC PR #. Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 01:04:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 01:04:58 -0600 Subject: [LLVMbugs] [Bug 1863] parser crash involving isConstantExpr() on incorrect code In-Reply-To: Message-ID: <200712180704.lBI74w0W025386@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1863 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2007-12-18 01:04:57 --- Fixed, we now emit: t3.c:4:14: error: variable length array declared outside of any function typedef int x[foo()]; ^ ~~~~~ t3.c:4:16: error: initializer element is not constant typedef int x[foo()]; ^ t3.c:6:21: error: initializer element is not constant int z = sizeof(int[foo()]); ~~~~~~~~~~~^~~~~~~ t3.c:7:2: error: expected identifier or '(' return y + z; ^ t3.c:8:1: error: expected identifier or '(' } ^ The second error looks bogus though. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 09:39:31 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 09:39:31 -0600 Subject: [LLVMbugs] [Bug 1840] asm calls should be marked nounwind In-Reply-To: Message-ID: <200712181539.lBIFdVKL026317@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1840 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #5 from Duncan Sands 2007-12-18 09:39:31 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071210/056373.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056401.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 10:54:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 10:54:51 -0600 Subject: [LLVMbugs] [Bug 1813] llvm-gcc thinks that A*abs(B) == abs(A*B) In-Reply-To: Message-ID: <200712181654.lBIGspUW030175@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1813 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #5 from Chris Lattner 2007-12-18 10:54:50 --- This is fixed in the apple 4.2 tree, next time we do a merge we'll get the fix. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 11:06:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 11:06:58 -0600 Subject: [LLVMbugs] [Bug 1865] GEPs aren't amenable to further analysis In-Reply-To: Message-ID: <200712181706.lBIH6wx4031411@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1865 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #1 from Chris Lattner 2007-12-18 11:06:57 --- That wouldn't be legal: %tmp19 = GEP %dst, 1 Would produce a constant offset from dst for every iteration of the loop. What you really want is get dst, ((iv << 2) + 1) instead of |1. I don't think this is a useful simplification. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 12:51:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 12:51:44 -0600 Subject: [LLVMbugs] [Bug 1870] New: format string checker does not employ simple constant propagation Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1870 Summary: format string checker does not employ simple constant propagation Product: clang Version: unspecified Platform: All OS/Version: All Status: ASSIGNED Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: kremenek at cs.stanford.edu ReportedBy: kremenek at cs.stanford.edu CC: llvmbugs at cs.uiuc.edu The following is a false positive (reported by Nico Weber) emitted by the format string checker: s0539:src nico$ cat test.c #include "stdio.h" int main() { char buf[1]; sprintf(buf, 1 ? "" : ""); } s0539:src nico$ ./clang test.c running "/Users/nico/src/llvm-svn/Debug/bin/clang -fsyntax-only test.c" test.c:5:16: warning: format string is not a string literal (potentially insecure) sprintf(buf, 1 ? "" : ""); ~~~~~~~ ^ 1 diagnostic generated. The false positive is a result of the checker not performing simple constant propagation of the string literal. Such expressions may be the result of macro expansion, and thus should be handled in order to suppress spurious warnings on correct code. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 13:11:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 13:11:16 -0600 Subject: [LLVMbugs] [Bug 1871] New: Need to remove IncludeFile Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1871 Summary: Need to remove IncludeFile Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Support Libraries AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu IncludeFile was added as a hack to allow building stuff as .a files and ensuring that stuff gets linked in appropriately. It was then extended to support abuse when trying to dynamically load plugins. This whole idea is flawed: symbols should be resolved the normal way, the simple act of #including a header should not pollute the .o file. Unfortunately, fixing this may require some work in the passmanager. Currently, I don't think that getAnalysis() adds a symbol dependency between the .o file that does that and the .o file that implements somepass. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 13:16:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 13:16:21 -0600 Subject: [LLVMbugs] [Bug 1872] New: SelectionDAG load / store CSE bug Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1872 Summary: SelectionDAG load / store CSE 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: evan.cheng at apple.com CC: llvmbugs at cs.uiuc.edu See PR1799. Using SrcValue to compute node id's means we can end up with multiple identical load / store nodes with the only difference being the SrcValue. Test case: target triple = "i686-pc-linux-gnu" %struct.c34007g__designated___XUB = type { i32, i32, i32, i32 } %struct.c34007g__pkg__parent = type { i32*, %struct.c34007g__designated___XUB* } define void @_ada_c34007g() { entry: %x8 = alloca %struct.c34007g__pkg__parent, align 8 ; <%struct.c34007g__pkg__parent*> [#uses=2] %tmp1272 = getelementptr %struct.c34007g__pkg__parent* %x8, i32 0, i32 0 ; [#uses=1] %x82167 = bitcast %struct.c34007g__pkg__parent* %x8 to i64* ; [#uses=1] br i1 true, label %bb4668, label %bb848 bb4668: ; preds = %bb4648 %tmp5464 = load i64* %x82167, align 8 ; [#uses=1] %tmp5467 = icmp ne i64 0, %tmp5464 ; [#uses=1] %tmp5470 = load i32** %tmp1272, align 8 ; [#uses=1] %tmp5471 = icmp eq i32* %tmp5470, null ; [#uses=1] %tmp5475 = or i1 %tmp5471, %tmp5467 ; [#uses=1] %tmp5497 = or i1 %tmp5475, false ; [#uses=1] br i1 %tmp5497, label %bb848, label %bb5507 bb848: ; preds = %entry ret void bb5507: ; preds = %bb4668 ret void } This compiles to: .text .align 4,0x90 .globl __ada_c34007g __ada_c34007g: subl $12, %esp LBB1_1: # bb4668 movl 4(%esp), %eax movl (%esp), %ecx cmpl $0, (%esp) je LBB1_3 # bb848 LBB1_2: # bb4668 orl %eax, %ecx testl %ecx, %ecx LBB1_3: # bb848 addl $12, %esp ret Note it issues two loads from (%esp) (the second one is folded into the cmpl). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 13:21:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 13:21:16 -0600 Subject: [LLVMbugs] [Bug 1799] "Node emitted out of order - early" llc failure In-Reply-To: Message-ID: <200712181921.lBIJLG9F007059@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1799 Evan Cheng changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #7 from Evan Cheng 2007-12-18 13:21:15 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056445.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 13:39:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 13:39:42 -0600 Subject: [LLVMbugs] [Bug 1872] SelectionDAG load / store CSE bug In-Reply-To: Message-ID: <200712181939.lBIJdgCk008189@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1872 Evan Cheng changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Evan Cheng 2007-12-18 13:39:41 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056459.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 18 15:29:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 18 Dec 2007 15:29:00 -0600 Subject: [LLVMbugs] [Bug 1873] New: ERROR: Program used external function 'stat' which could not be resolved! Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1873 Summary: ERROR: Program used external function 'stat' which could not be resolved! Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu I get this when running the JIT: /home/edwin/llvm-svn/llvm/Release/bin/lli -force-interpreter=false --disable-core-files Output/clamscan.llvm.bc --debug --no-summary -ddbdir -r inputs/* ERROR: Program used external function 'stat' which could not be resolved! /home/edwin/llvm-svn/llvm/Release/bin/lli[0x8cc51b] /lib/libc.so.6[0x3508432040] /lib/libc.so.6(gsignal+0x35)[0x3508431fd5] /lib/libc.so.6(abort+0x110)[0x3508433a30] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::JIT::getPointerToNamedFunction(std::basic_string, std::allocator > const&)+0xf1)[0x5e0c71] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::JIT::getPointerToFunction(llvm::Function*)+0xe1)[0x5e1fb1] /home/edwin/llvm-svn/llvm/Release/bin/lli[0x5e37b0] /home/edwin/llvm-svn/llvm/Release/bin/lli[0x5e43f8] /home/edwin/llvm-svn/llvm/Release/bin/lli[0x4af0be] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::MachineFunctionPass::runOnFunction(llvm::Function&)+0x2c)[0x5c06dc] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::FPPassManager::runOnFunction(llvm::Function&)+0x231)[0x87a351] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::FunctionPassManagerImpl::run(llvm::Function&)+0xdc)[0x87a83c] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::FunctionPassManager::run(llvm::Function&)+0x30)[0x87a990] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::JIT::runJITOnFunction(llvm::Function*)+0x38)[0x5e0d68] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::JIT::getPointerToFunction(llvm::Function*)+0x12d)[0x5e1ffd] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::JIT::runFunction(llvm::Function*, std::vector > const&)+0x4c)[0x5e104c] /home/edwin/llvm-svn/llvm/Release/bin/lli(llvm::ExecutionEngine::runFunctionAsMain(llvm::Function*, std::vector, std::allocator >, std::allocator, std::allocator > > > const&, char const* const*)+0x382)[0x5eaf92] /home/edwin/llvm-svn/llvm/Release/bin/lli(main+0x233)[0x4a5463] /lib/libc.so.6(__libc_start_main+0xf4)[0x350841e1c4] /home/edwin/llvm-svn/llvm/Release/bin/lli(sinh+0x49)[0x4a4919] Aborted Trying to get a bugpoint reduced bytecode ... -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 19 01:55:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 19 Dec 2007 01:55:18 -0600 Subject: [LLVMbugs] [Bug 1874] New: Multiplication codegen issues Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1874 Summary: Multiplication codegen issues Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: major Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: evan.cheng at apple.com CC: llvmbugs at cs.uiuc.edu Chris noticed this: We're missing an obvious fold of a load into imul: int test(long a, long b) { return a * b; } LLVM produces: _test: movl 4(%esp), %ecx movl 8(%esp), %eax imull %ecx, %eax ret vs: _test: movl 8(%esp), %eax imull 4(%esp), %eax ret I've looked into this and noticed a bigger issue with multiplication codegen in general. On x86, ISD::MUL i32 is marked expand. Legalizer expands this to a ISD::SMUL_LOHI. DAG combiner failed to turn it back into a ISD::MUL (it's high part isn't needed) because ISD::MUL is not legal. X86 isel then sees it's a ISD::SMUL_LOHI whose high value is not needed and create a new ISD::MUL node and select it to an IMULL node. However since the operands (which are loads) have multiple uses, they are not folded into the IMULL node. 1. ISD::MUL i32 should not have been marked expand. It's a perfectly selectable node. Only i64 ISD::MUL (unless it's x86-64) should be marked expand. 2. It makes sense for the legalizer to issue {S|U}MUL_LOHI when expanding a MUL. But we also need DAG combiner to combine MULH{S|U} + MUL to a {S|U}MUL_LOHI. 3. X86 isel should not need any hack to handle {S|U}MUL_LOHI nodes whose high value is not used. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 19 13:46:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 19 Dec 2007 13:46:41 -0600 Subject: [LLVMbugs] [Bug 1875] New: Found a case where llvm-g++ 2.1 (or TOT) produces worse code than 2.0 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1875 Summary: Found a case where llvm-g++ 2.1 (or TOT) produces worse code than 2.0 Product: new-bugs Version: unspecified Platform: PC OS/Version: FreeBSD Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: niklasso at gmail.com CC: llvmbugs at cs.uiuc.edu The source in question is a version of the open-source SAT solver MiniSat (http://minisat.se). I'm compiling it using "llvm-g++ --emit-llvm", and then using llvm-ld to do IPO, and finally llc to produce an x86 binary. If I compile with llvm-g++ based on version 2.0 I get better code than llvm-g++ based on 2.1 or current svn version (roughly 3-5% faster). If I use 2.0 or 2.1 versions of llvm-ld and llc does not seem to matter much however. See the README in the attached archive for instructions on how I built the program and how to run the testcases. If you want to include this in your test framework that is fine. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 19 20:26:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 19 Dec 2007 20:26:17 -0600 Subject: [LLVMbugs] [Bug 1829] CodeGen/ARM/2007-03-13-InstrSched.ll failing In-Reply-To: Message-ID: <200712200226.lBK2QH5m015685@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1829 Evan Cheng changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #4 from Evan Cheng 2007-12-19 20:26:17 --- One more lsr patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056518.html Scheduling issue is fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056519.html Finally, a coalescer issue: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056520.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 21 21:49:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 21 Dec 2007 21:49:05 -0600 Subject: [LLVMbugs] [Bug 1876] New: llvm-ar: corrupted double-linked list Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1876 Summary: llvm-ar: corrupted double-linked list Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sanxiyn at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1303) --> (http://llvm.org/bugs/attachment.cgi?id=1303) Tcl 6.4 with Makefile for clang SVN r45307. Debian GNU/Linux Sid (up-to-date). Attached is lightly patched version of Tcl 6.4 with a modified Makefile to compile with clang. Extract and run "make" to reproduce. "llvm-ar cr libtcl.a tclGlob.bc tclParse.bc tclVar.bc" crashes with the message "llvm-ar: corrupted double-linked list". This is a minimal set. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From tom.primozic at gmail.com Sat Dec 22 15:44:22 2007 From: tom.primozic at gmail.com (=?UTF-8?Q?Tom_Primo=C5=BEi=C4=8D?=) Date: Sat, 22 Dec 2007 22:44:22 +0100 Subject: [LLVMbugs] Compiling to Win32 Message-ID: Hello! I didn't know where exactly to post this question, so I hope you will redirect me if this is the wrong list. I am interested in LLVM, particularly for the implementation of a programming language, and have been looking at the LLVM documentation recently. However, I was wondering what exactly do these two lines, extracted from the Portability and Supported Platforms (http://llvm.org/docs/ReleaseNotes.html#portability): - Intel and AMD machines running on Win32 using MinGW libraries (native) - Intel and AMD machines running on Win32 with the Cygwin libraries (limited support is available for native builds with Visual C++). Does this mean that LLVM cannot produce Win32 application independently? So that I can not use it (for example, from an IDE) without installing third-party libraries? Do these libraries need to be distributed along with the generated executables? Please help me. I would like to understand exactly what I need to (i) compile a native, complete, optimized program executable (from source code to LLVM internal representation using my front-end, then optimizing, emmiting machine code, and linkindg with LLVM), ready to be distributed to end users, (ii) compile to some custom bytecode, than make a runtime, on top on LLVM, that would JIT-compile that bytecode and run it (in client's computer - like Flash, Java, ...), and (iii) to provide the end user with the option to optimize and recompile the program. - Tom -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20071222/6ae55e22/attachment.html From bugzilla-daemon at cs.uiuc.edu Sun Dec 23 03:04:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 23 Dec 2007 03:04:08 -0600 Subject: [LLVMbugs] [Bug 1877] New: Backedge copy not coalesced in Freebench/neural Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1877 Summary: Backedge copy not coalesced in Freebench/neural Product: libraries Version: 1.0 Platform: PC OS/Version: All Status: NEW Keywords: code-quality Severity: minor Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: evan.cheng at apple.com, llvmbugs at cs.uiuc.edu The significant (negative) performance delta on Freebench/neural on x86 is due to a backedge copy not getting coalesced. The backedge critical edge is then split to hold the copy, and the critical edge block is put into a very bad place. This makes the code run significantly slower than the GCC code, whcih doesn't make this mistake. Here's a reduced testcase: unsigned NNTOT; float **Tmatrix; volatile float G; int runcont (signed int source[], signed int dest[]) { int row = 0, neuron; // for(row=0; row http://llvm.org/bugs/show_bug.cgi?id=1878 Summary: opt -std-compile-opts drops Product: libraries Version: trunk Platform: All OS/Version: All Status: ASSIGNED Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: gordonhenriksen at mac.com ReportedBy: gordonhenriksen at mac.com CC: llvmbugs at cs.uiuc.edu Armin Rigo reports from the PyPy project: > Here is what appears to be another bug of 'opt': the "gc" annotation is > sometimes lost, or so it seems. With the following input (140k gzip): > > http://wyvern.cs.uni-duesseldorf.de/~arigo/new_entrypoint.ll.gz > > and the following command: > > llvm-as < new_entrypoint.ll | opt -std-compile-opts > new_entrypoint.bc > > then llvm-dis shows that in the .bc file, the function > @pypy_SemiSpaceGC_malloc_varsize_clear has lost its gc "gcrootsingle" > attribute. I'll need to do some further investigation to reduce the input and discover which pass is at fault. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 24 15:01:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 24 Dec 2007 15:01:25 -0600 Subject: [LLVMbugs] [Bug 1879] New: Branch folding deficiency Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1879 Summary: Branch folding deficiency Product: libraries Version: 1.0 Platform: PC OS/Version: All Status: NEW Keywords: code-quality Severity: minor Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu Ptrdist/ks contains this amusing code in its inner loop: movss _cost(,%edx,4), %xmm1 LBB2_3: # bb je LBB2_7 # bb41 LBB2_4: # bb32 movaps %xmm1, %xmm2 addss %xmm0, %xmm2 cmpl %ecx, 4(%esi) je LBB2_6 # bb32 LBB2_5: # bb32 movaps %xmm0, %xmm2 LBB2_6: # bb32 movaps %xmm2, %xmm0 movl (%esi), %esi testl %esi, %esi jmp LBB2_3 # bb The inner lop is BB3/4/5/6. Note that BB6 contains an uncond branch *just* to a cond branch. It seems branch folding should duplicate the conditional branch into the end of BB6. The bigger issue with this code is that it contains a conditional add. It is turning: if ((*modNode).module == module) gain = gain + netCost; Into a select: cond ? (gain+netCost) : gain; And then the select is turned into a conditional branch, and the add isn't sunk into the conditional block. To see this, plop this into the top of KS-2.c float CAiBj(ModuleRecPtr mrA, ModuleRecPtr mrB) __attribute__((noinline)); -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Dec 25 20:48:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 25 Dec 2007 20:48:35 -0600 Subject: [LLVMbugs] [Bug 1878] opt -std-compile-opts drops "gc" attribute In-Reply-To: Message-ID: <200712260248.lBQ2mZ0J026184@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1878 Gordon Henriksen changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #2 from Gordon Henriksen 2007-12-25 20:48:34 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071224/056598.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071224/056599.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Dec 26 19:52:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 26 Dec 2007 19:52:16 -0600 Subject: [LLVMbugs] [Bug 1880] New: Building LLVM with llvm-gcc-4.2: duplicate absolute symbols Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1880 Summary: Building LLVM with llvm-gcc-4.2: duplicate absolute symbols Product: new-bugs Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: blocker Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org CC: llvmbugs at cs.uiuc.edu Using the latest llvm trunk and llvm-gcc-4.2, attempted to rebuild llvm in a separate directory with the resulting llvm-gcc-4.2 binaries, installed as llvm-gcc and llvm-g++, on x86 Darwin (10.4.11) and xcode 2.5. The compile does not succeed when linking tblgen ("make VERBOSE=1" output attached below). I double checked my installation paths and configuration, cannot find anything detectably faulty. I can't reasonably tell if this is a llvm code generation bug or a llvm-gcc-4.2 misconfiguration. This was apparently fixed for another platform -- what was the corresponding SVN revision to llvm or llvm-gcc-4.2? llvm-c++ -I/Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen -I/Users/scottm/play/llvm/trunk/utils/TableGen -I/Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/include -I/Users/scottm/play/llvm/trunk/include -I/Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/include -I/Users/scottm/play/llvm/trunk/include -D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O3 -Woverloaded-virtual -pedantic -Wall -W -Wwrite-strings -Wno-long-long -Wunused -Wno-unused-parameter -fstrict-aliasing -O3 -o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/Release/bin/tblgen /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/AsmWriterEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CallingConvEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/IntrinsicEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/RegisterInfoEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/SubtargetEmitter.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGLexer.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TableGen.o /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TableGenBackend.o -L/Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/Release/lib -lLLVMSupport -lLLVMSystem -lpthread /usr/lib/libltdl.3.1.0.dylib -ldl -lm /usr/bin/ld: multiple definitions of symbol __ZN4llvm15TableGenBackendD1Ev.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/AsmWriterEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CallingConvEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm4InitD1Ev.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZN4llvm4InitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZN4llvm4InitD1Ev.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm4Init10isCompleteEv.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZNK4llvm4Init10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZNK4llvm4Init10isCompleteEv.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm4Init12getFieldTypeERKSs.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeEmitterGen.o definition of absolute __ZNK4llvm4Init12getFieldTypeERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/CodeGenTarget.o definition of absolute __ZNK4llvm4Init12getFieldTypeERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm4InitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm4Init10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm4Init12getFieldTypeERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_10StringInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10StringInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10StringInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_10VarBitInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10VarBitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10VarBitInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_7BitInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7BitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7BitInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_7DagInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DagInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DagInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_7DefInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DefInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DefInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_7IntInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7IntInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7IntInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_8BitsInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8BitsInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8BitsInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_8CodeInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8CodeInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8CodeInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_8ListInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8ListInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8ListInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_9BinOpInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9BinOpInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9BinOpInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_9TypedInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9TypedInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9TypedInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTy12convertValueEPNS_9UnsetInitE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9UnsetInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9UnsetInitE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm5RecTyD1Ev.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZN4llvm5RecTyD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZN4llvm5RecTyD1Ev.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_11RecordRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11RecordRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11RecordRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_11StringRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11StringRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11StringRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_8BitRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8BitRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8BitRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_8DagRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8DagRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8DagRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_8IntRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8IntRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8IntRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_9BitsRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9BitsRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9BitsRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_9CodeRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9CodeRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9CodeRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm5RecTy11baseClassOfEPKNS_9ListRecTyE.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/DAGISelEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9ListRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/InstrInfoEmitter.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9ListRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/IntrinsicEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm4InitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10StringInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10VarBitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7BitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DagInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DefInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7IntInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8BitsInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8CodeInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8ListInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9BinOpInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9TypedInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9UnsetInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm5RecTyD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm4Init10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm4Init12getFieldTypeERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11RecordRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11StringRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8BitRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8DagRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8IntRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9BitsRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9CodeRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9ListRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/RegisterInfoEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/SubtargetEmitter.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm4Init17resolveReferencesERNS_6RecordEPKNS_9RecordValE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm4Init20convertInitListSliceERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm4Init26convertInitializerBitRangeERKSt6vectorIjSaIjEE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm4InitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10StringInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_10VarBitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7BitInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DagInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7DefInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_7IntInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8BitsInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8CodeInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_8ListInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9BinOpInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9TypedInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTy12convertValueEPNS_9UnsetInitE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm5RecTyD1Ev.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZN4llvm9UnsetInitD1Ev.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZN4llvm9UnsetInitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZN4llvm9UnsetInitD1Ev.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm4Init10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm4Init12getFieldInitERNS_6RecordERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm4Init12getFieldTypeERKSs.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11RecordRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_11StringRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8BitRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8DagRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_8IntRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9BitsRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9CodeRecTyE.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm5RecTy11baseClassOfEPKNS_9ListRecTyE.eh (value 0x0) /usr/bin/ld: multiple definitions of symbol __ZNK4llvm9UnsetInit10isCompleteEv.eh /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/Record.o definition of absolute __ZNK4llvm9UnsetInit10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TGParser.o definition of absolute __ZNK4llvm9UnsetInit10isCompleteEv.eh (value 0x0) /Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/utils/TableGen/Release/TableGen.o definition of absolute __ZN4llvm15TableGenBackendD1Ev.eh (value 0x0) collect2: ld returned 1 exit status make[2]: *** [/Users/scottm/play/llvm/trunk/obj/i686-apple-darwin-llvm/Release/bin/tblgen] Error 1 make[1]: *** [TableGen/.makeall] Error 2 make: *** [all] Error 1 llvm-gcc -v: Using built-in specs. Target: i686-apple-darwin8 Configured with: ../../configure --enable-languages=c,c++,objc,obj-c++ --with-gxx-include-dir=/usr/include/c++/4.0.0 --build=i686-apple-darwin8 --host=i686-apple-darwin8 --target=i686-apple-darwin8 --with-arch=nocona --with-tune=generic --disable-intl --enable-llvm=/../../../llvm/trunk/obj/i686-apple-darwin --prefix=/../../i686-apple-darwin8 --program-prefix=llvm- Thread model: posix gcc version 4.2.1(llvm) (Based on Apple Inc. build 5546) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 08:38:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 08:38:00 -0600 Subject: [LLVMbugs] [Bug 1881] New: llvm-ar: Segmentation fault when archiving bitcode library Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1881 Summary: llvm-ar: Segmentation fault when archiving bitcode library Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: llvm-ar AssignedTo: unassignedbugs at nondot.org ReportedBy: shackan at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1306) --> (http://llvm.org/bugs/attachment.cgi?id=1306) Bitcode files, to reproduce do `llvm-ar rcsf libcrypto_rsa.bca *.bc` in the Debug/ directory make[2]: Entering directory `/build-x86/projects/rsa/openssl-0.9.8g/crypto/rsa' llvm[2]: Compiling rsa_asn1.c for Debug build (bytecode) llvm[2]: Compiling rsa_asn1.ll to rsa_asn1.bc for Debug build (bytecode) llvm[2]: Compiling rsa_chk.c for Debug build (bytecode) llvm[2]: Compiling rsa_chk.ll to rsa_chk.bc for Debug build (bytecode) llvm[2]: Compiling rsa_depr.c for Debug build (bytecode) llvm[2]: Compiling rsa_depr.ll to rsa_depr.bc for Debug build (bytecode) llvm[2]: Compiling rsa_eay.c for Debug build (bytecode) llvm[2]: Compiling rsa_eay.ll to rsa_eay.bc for Debug build (bytecode) llvm[2]: Compiling rsa_err.c for Debug build (bytecode) llvm[2]: Compiling rsa_err.ll to rsa_err.bc for Debug build (bytecode) llvm[2]: Compiling rsa_gen.c for Debug build (bytecode) llvm[2]: Compiling rsa_gen.ll to rsa_gen.bc for Debug build (bytecode) llvm[2]: Compiling rsa_lib.c for Debug build (bytecode) llvm[2]: Compiling rsa_lib.ll to rsa_lib.bc for Debug build (bytecode) llvm[2]: Compiling rsa_none.c for Debug build (bytecode) llvm[2]: Compiling rsa_none.ll to rsa_none.bc for Debug build (bytecode) llvm[2]: Compiling rsa_null.c for Debug build (bytecode) llvm[2]: Compiling rsa_null.ll to rsa_null.bc for Debug build (bytecode) llvm[2]: Compiling rsa_oaep.c for Debug build (bytecode) llvm[2]: Compiling rsa_oaep.ll to rsa_oaep.bc for Debug build (bytecode) llvm[2]: Compiling rsa_pk1.c for Debug build (bytecode) llvm[2]: Compiling rsa_pk1.ll to rsa_pk1.bc for Debug build (bytecode) llvm[2]: Compiling rsa_pss.c for Debug build (bytecode) llvm[2]: Compiling rsa_pss.ll to rsa_pss.bc for Debug build (bytecode) llvm[2]: Compiling rsa_saos.c for Debug build (bytecode) llvm[2]: Compiling rsa_saos.ll to rsa_saos.bc for Debug build (bytecode) llvm[2]: Compiling rsa_sign.c for Debug build (bytecode) llvm[2]: Compiling rsa_sign.ll to rsa_sign.bc for Debug build (bytecode) llvm[2]: Compiling rsa_ssl.c for Debug build (bytecode) llvm[2]: Compiling rsa_ssl.ll to rsa_ssl.bc for Debug build (bytecode) llvm[2]: Compiling rsa_test.c for Debug build (bytecode) llvm[2]: Compiling rsa_test.ll to rsa_test.bc for Debug build (bytecode) llvm[2]: Compiling rsa_x931.c for Debug build (bytecode) llvm[2]: Compiling rsa_x931.ll to rsa_x931.bc for Debug build (bytecode) llvm[2]: Building Debug Bytecode Archive libcrypto_rsa.bca /build-x86/Debug/bin/llvm-ar((anonymous namespace)::PrintStackTrace()+0x22)[0x82496c0] /build-x86/Debug/bin/llvm-ar((anonymous namespace)::SignalHandler(int)+0x112)[0x8249986] [0xffffe420] /build-x86/Debug/bin/llvm-ar(llvm::ParamAttrsList::~ParamAttrsList()+0x22)[0x81c0108] /build-x86/Debug/bin/llvm-ar(llvm::ParamAttrsList::dropRef() const+0x63)[0x81c2f57] /build-x86/Debug/bin/llvm-ar(llvm::Function::~Function()+0x66)[0x81c0e7a] /build-x86/Debug/bin/llvm-ar(llvm::iplist >::erase(llvm::ilist_iterator)+0x33)[0x81c2631] /build-x86/Debug/bin/llvm-ar(llvm::iplist >::erase(llvm::ilist_iterator, llvm::ilist_iterator)+0x21)[0x81de765] /build-x86/Debug/bin/llvm-ar(llvm::iplist >::clear()+0x5c)[0x81de7f0] /build-x86/Debug/bin/llvm-ar(llvm::Module::~Module()+0x2a)[0x81dd710] /build-x86/Debug/bin/llvm-ar(llvm::ModuleProvider::~ModuleProvider()+0x2a)[0x81dea02] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReader::~BitcodeReader()+0xc4)[0x81810f6] /build-x86/Debug/bin/llvm-ar(llvm::Archive::writeMember(llvm::ArchiveMember const&, std::basic_ofstream >&, bool, bool, bool, std::basic_string, std::allocator >*)+0x442)[0x81643cc] /build-x86/Debug/bin/llvm-ar(llvm::Archive::writeToDisk(bool, bool, bool, std::basic_string, std::allocator >*)+0x344)[0x81649ee] /build-x86/Debug/bin/llvm-ar(doReplaceOrInsert(std::basic_string, std::allocator >*)+0x74d)[0x8155ddd] /build-x86/Debug/bin/llvm-ar(main+0x3be)[0x8157998] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0)[0x40199050] /build-x86/Debug/bin/llvm-ar(__gxx_personality_v0+0x2a5)[0x81546c1] make[2]: *** [/build-x86/projects/rsa/Debug/lib/libcrypto_rsa.bca] Segmentation fault (core dumped) make[2]: Leaving directory `/build-x86/projects/rsa/openssl-0.9.8g/crypto/rsa' make[1]: *** [all] Error 1 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 08:51:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 08:51:38 -0600 Subject: [LLVMbugs] [Bug 1882] New: llvm-ar: Segmentation fault when archiving bitcode library Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1882 Summary: llvm-ar: Segmentation fault when archiving bitcode library Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: llvm-ar AssignedTo: unassignedbugs at nondot.org ReportedBy: shackan at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1307) --> (http://llvm.org/bugs/attachment.cgi?id=1307) Bitcode files, to reproduce do `llvm-ar rcsf libavcodec.bca *.bc` maybe a duplicate of #1881, but from the stack traces they seem two different beasts llvm[2]: Building Debug Bytecode Archive libavcodec.bca llvm-ar: /lib/Bitcode/Reader/BitcodeReader.cpp:162: llvm::Value* llvm::BitcodeReaderValueList::getValueFwdRef(unsigned int, const llvm::Type*): Assertion `(Ty == 0 || Ty == V->getType()) && "Type mismatch in value table!"' failed. /build-x86/Debug/bin/llvm-ar((anonymous namespace)::PrintStackTrace()+0x22)[0x82496c0] /build-x86/Debug/bin/llvm-ar((anonymous namespace)::SignalHandler(int)+0x112)[0x8249986] [0xffffe420] /lib/tls/i686/cmov/libc.so.6(abort+0x101)[0x401af201] /lib/tls/i686/cmov/libc.so.6(__assert_fail+0xee)[0x401a6b6e] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReaderValueList::getValueFwdRef(unsigned int, llvm::Type const*)+0xe7)[0x816afb7] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReader::getFnValueByID(unsigned int, llvm::Type const*)+0x25)[0x8180bc5] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReader::ParseFunctionBody(llvm::Function*)+0x2deb)[0x816de5b] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReader::materializeFunction(llvm::Function*, std::basic_string, std::allocator >*)+0xea)[0x816e704] /build-x86/Debug/bin/llvm-ar(llvm::BitcodeReader::materializeModule(std::basic_string, std::allocator >*)+0x97)[0x81655b5] /build-x86/Debug/bin/llvm-ar(llvm::GetBitcodeSymbols(unsigned char const*, unsigned int, std::basic_string, std::allocator > const&, std::vector, std::allocator >, std::allocator, std::allocator > > >&, std::basic_string, std::allocator >*)+0x7b)[0x815d425] /build-x86/Debug/bin/llvm-ar(llvm::Archive::writeMember(llvm::ArchiveMember const&, std::basic_ofstream >&, bool, bool, bool, std::basic_string, std::allocator >*)+0x27b)[0x8164205] /build-x86/Debug/bin/llvm-ar(llvm::Archive::writeToDisk(bool, bool, bool, std::basic_string, std::allocator >*)+0x344)[0x81649ee] /build-x86/Debug/bin/llvm-ar(doReplaceOrInsert(std::basic_string, std::allocator >*)+0x74d)[0x8155ddd] /build-x86/Debug/bin/llvm-ar(main+0x3be)[0x8157998] /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0)[0x40199050] /build-x86/Debug/bin/llvm-ar(__gxx_personality_v0+0x2a5)[0x81546c1] make[2]: *** [build-x86/projects/h264/Debug/lib/libavcodec.bca] Aborted (core dumped) make[2]: Leaving directory `build-x86/projects/h264/ffmpeg/libavcodec' -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 14:12:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 14:12:00 -0600 Subject: [LLVMbugs] [Bug 1883] New: Overly strict error with initializer + cast Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1883 Summary: Overly strict error with initializer + cast Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Not sure if I have the right component. Compiling the following with clang -fsyntax-only: typedef struct Test (int a;int b;} Test; static Test t = (Test) {0,0}; gives inittest.c:2:24: error: initializer element is not constant static Test t = (Test) {0,0}; ^~~~~ gcc -fsyntax-only accepts this code, although it does warn if -pedantic is used, and gives exactly the same error if -std=c99 is used. So this is apparently valid C90 code, but not valid C99? Also, gcc accepts this code with -std=gnu99. Found by running clang over ffmpeg code. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 14:51:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 14:51:10 -0600 Subject: [LLVMbugs] [Bug 1884] New: Incorrect redeclaration error with enum+int? Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1884 Summary: Incorrect redeclaration error with enum+int? Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Not sure if this is the right component. gcc accepts the following code (in all modes, as far as I can tell): char *func(int format); enum Test {A=-1}; char *func(enum Test x); However, clang does not accept this. (I think clang is right here, but I don't know the C standard that well.) Found in the ffmpeg source code. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 15:41:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 15:41:06 -0600 Subject: [LLVMbugs] [Bug 1885] New: Cannot forward-declare static array Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1885 Summary: Cannot forward-declare static array Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Not sure if this is the right component. gcc -fsyntax-only accepts the following, but clang does not: static int array[]; Note that gcc also rejects this code with the -pedantic flag enabled. Found in the ffmpeg source code. On a side note, the error "variable has incomplete type 'int []'" is a bit obscure. It would be nice if the error actually gave a hint as to why this is an error. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 23:36:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 23:36:19 -0600 Subject: [LLVMbugs] [Bug 1534] Testcase MultiSource/Application/oggenc fails (llc) with -relocation-model=pic -mcpu=i486. In-Reply-To: Message-ID: <200712290536.lBT5aJD0026542@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1534 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #1 from Chris Lattner 2007-12-28 23:36:18 --- Works for me now. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Dec 28 23:46:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 28 Dec 2007 23:46:53 -0600 Subject: [LLVMbugs] [Bug 1534] Testcase MultiSource/Application/oggenc fails (llc) with -relocation-model=pic -mcpu=i486. In-Reply-To: Message-ID: <200712290546.lBT5kr82027120@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1534 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|WORKSFORME | --- Comment #2 from Chris Lattner 2007-12-28 23:46:53 --- Actually, you're right. It fails with -mcpu=pentium2 and lower, and passes with -mcpu=pentium3 and above. Presumably the difference is that SSE regs are being used for f32 values. It may be instability in the program. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Dec 29 02:04:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 29 Dec 2007 02:04:13 -0600 Subject: [LLVMbugs] [Bug 1425] gvn is slow on this testcase in debug mode In-Reply-To: Message-ID: <200712290804.lBT84DpT004381@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1425 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #12 from Chris Lattner 2007-12-29 02:04:12 --- Ok, the original issue is fixed, I'm closing this as done. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Dec 29 18:01:53 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 29 Dec 2007 18:01:53 -0600 Subject: [LLVMbugs] [Bug 1886] New: Crash with incomplete struct initializer Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1886 Summary: Crash with incomplete struct initializer Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu using clang -emit-llvm, the following crashes with an assertion failure: struct a {int b; int c}; struct a c = {5}; The assertion: clang: /home/eli/llvm/lib/VMCore/Constants.cpp:367: llvm::ConstantStruct::ConstantStruct(const llvm::StructType*, const std::vector >&): Assertion `V.size() == T->getNumElements() && "Invalid initializer vector for constant structure"' failed. Looks like just a missing case in codegen, since both "struct a c = {};" and "struct a c = {5,3};" currently work correctly. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Dec 29 19:28:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 29 Dec 2007 19:28:20 -0600 Subject: [LLVMbugs] [Bug 1887] New: Error with bitfield codegen Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1887 Summary: Error with bitfield codegen Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu struct Test { int a:1; int e; }; static struct Test input_files; crashes with clang -emit-llvm with the assertion: clang: /home/eli/llvm/tools/clang/CodeGen/CodeGenTypes.cpp:479: void::RecordOrganizer::addPaddingFields(unsigned int): Assertion `(RequiredBits % 8) == 0 && "FIXME Invalid struct layout"' failed. I saw the bit on the mailing list about this assertion (http://lists.cs.uiuc.edu/pipermail/cfe-dev/2007-December/000561.html), but in this case, non-integral padding is actually necessary. I guess this might fall into the category of "bitfields aren't implemented", but I figured I might as well file it. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 03:11:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 03:11:04 -0600 Subject: [LLVMbugs] [Bug 1620] dead argument elimination loses parameter attributes In-Reply-To: Message-ID: <200712310911.lBV9B4fA021767@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1620 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #11 from Duncan Sands 2007-12-31 03:10:56 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056546.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 03:21:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 03:21:41 -0600 Subject: [LLVMbugs] [Bug 1443] Should propagate nothrow attribute to LLVM IR In-Reply-To: Message-ID: <200712310921.lBV9LfXH022214@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1443 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Duncan Sands 2007-12-31 03:21:39 --- This is done in llvm-gcc-4.2. A call cannot throw for basically two reasons: (1) the callee is a 'nothrow' function; (2) the call occurs in a 'nothrow region' [these are generated by C++, which doesn't allow a new exception to be thrown when finalizing objects during unwinding of an exception]. In case (1) we mark the LLVM function 'nounwind', and have done for some time. In case (2) we mark the call 'nounwind'. This was done here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071126/055898.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 03:24:27 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 03:24:27 -0600 Subject: [LLVMbugs] [Bug 870] llvm-gcc4 lacks support for C++ exceptions In-Reply-To: Message-ID: <200712310924.lBV9ORI0022376@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=870 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |baldrick at free.fr Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Duncan Sands 2007-12-31 03:24:26 --- As far as I know llvm-gcc-4.2 fully supports C++ exceptions, while llvm-gcc-4.0 mostly supports them (it doesn't handle some corner cases correctly). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 12:37:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 12:37:05 -0600 Subject: [LLVMbugs] [Bug 1833] llc -enable-eh crash in LegalizeDAG In-Reply-To: Message-ID: <200712311837.lBVIb5KO016518@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1833 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Duncan Sands 2007-12-31 12:37:04 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056701.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 13:13:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 13:13:49 -0600 Subject: [LLVMbugs] [Bug 592] Need zero cost exceptions in the native code generators In-Reply-To: Message-ID: <200712311913.lBVJDndW018181@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=592 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Depends on|1414 | Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Chris Lattner 2007-12-31 13:13:48 --- This seems done to me. Any remaining issues are target-specific bugs, not lack of the feature. Thanks! I'm also removing Bug 1414 as a blocking dependency. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Dec 31 20:41:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 31 Dec 2007 20:41:35 -0600 Subject: [LLVMbugs] [Bug 1888] New: Compilier errors in Building LLVM using Windows MSVC Projects Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1888 Summary: Compilier errors in Building LLVM using Windows MSVC Projects Product: Build scripts Version: trunk Platform: PC OS/Version: Windows XP Status: NEW Keywords: build-problem Severity: enhancement Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: willi.t1 at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1312) --> (http://llvm.org/bugs/attachment.cgi?id=1312) Patch for the dobison script and codegen project The MSVC project under llvm/win32 and some source files has to be adjusted in order to be compiled. For the Project, ths CodeGen project is missing MachineRegisterInfo.cpp so when llc links to codegen.lib, gets linker error about MachineRegisterInfo. Simply including it into the project will do. As for the files, the main problem is that Window's float.h defines a macro named FLT_ROUNDS which conflicts with ISD:FLT_ROUNDS defined in SelectionDagNodes.h there are many solutions to this but what I did is that I undef'ed FLT_ROUNDS in the following files: SelectionDAGNodes.h SelectionDAGISel.cpp SelectionDAG.cpp LegalizeDAG.cpp Another issue is the bison script (I'm not so sure about the flex script) basically, the fallback method copies the pregenerated files to the wrong directory instead to the target directory. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.