From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 02:44:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 02:44:41 -0600 Subject: [LLVMbugs] [Bug 1144] bad codegen Message-ID: <200702010844.l118ifdV023315@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1144 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-02-01 02:44 ------- This was the result of "sret" patch. Codegen really thought, that memmove is struct return function :) This was due to uninitialized stuff inside LegalizedDAG.cpp. Fortunately, the only affected part is memmove call inside x86 backend (since memcpy & memset are custom lowered and X86ISelLowering.cpp was ok). Sorry for the breakage. Fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043666.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043667.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043668.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 07:27:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 07:27:45 -0600 Subject: [LLVMbugs] [Bug 1147] test/ExecutionEngine/test-fp.ll sometimes fails (FoldingSet.cpp) Message-ID: <200702011327.l11DRjX2029437@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1147 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-01 07:27 ------- This no longer fails after Anton fixed his uninitialized data issue: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043667.html I can't make the test case fail any more and valgrind produces 0 ERRORS on the test case. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 07:35:16 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 07:35:16 -0600 Subject: [LLVMbugs] [Bug 1155] NEW: stage2 compiles c++ files with the host c++ compiler Message-ID: <200702011335.l11DZG0L029578@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1155 Summary: stage2 compiles c++ files with the host c++ compiler Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com Stage2 uses c++ instead of /home/laurov/llvm/llvm-gcc/build/prev-gcc/cpp. make[2]: Entering directory `/home/laurov/llvm/llvm-gcc/build/stage2-gcc' c++ -c -O2 -g -fomit-frame-pointer -DIN_GCC -W -Wall -Wwrite-strings -pedantic -Wno-long-long -Wno-variadic-macros -Werror -fno-common -DHAVE_CONFIG_H -Wno-unused -DTARGET_NAME=\"i686-pc-linux-gnu\" -DENABLE_LLVM -D__STDC_LIMIT_MACROS -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include -I/home/laurov/llvm/llvm/include -I/home/laurov/llvm/llvm/build//include ../../gcc/llvm-main.cpp -o llvm-main.o /home/laurov/llvm/llvm-gcc/build/prev-gcc/xgcc -B/home/laurov/llvm/llvm-gcc/build/prev-gcc/ -B/usr/local/i686-pc-linux-gnu/bin/ -O2 -g -fomit-frame-pointer -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wold-style-definition -Werror -fno-common -DHAVE_CONFIG_H -DENABLE_LLVM -D__STDC_LIMIT_MACROS -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../include -I../../gcc/../libcpp/include -I/home/laurov/llvm/llvm/include -I/home/laurov/llvm/llvm/build//include -c insn-attrtab.c \ -o insn-attrtab.o ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 07:36:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 07:36:24 -0600 Subject: [LLVMbugs] [Bug 1155] stage2 compiles c++ files with the host c++ compiler Message-ID: <200702011336.l11DaOGs029608@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1155 lauro.venancio at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From lauro.venancio at gmail.com 2007-02-01 07:36 ------- *** This bug has been marked as a duplicate of 1150 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 12:22:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 12:22:44 -0600 Subject: [LLVMbugs] [Bug 1156] NEW: llvm-upgrade folds constant expressions Message-ID: <200702011822.l11IMiWa003804@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1156 Summary: llvm-upgrade folds constant expressions Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Because llvm-upgrade is now using VMCore for its representation, it gets all constants that it creates folded automatically. This doesn't cause any semantic problems but also doesn't lead to an apples-to-apples translation. For example: %D = ashr int %X, ubyte trunc ( int shl (int 3, ubyte 1) to ubyte ) becomes %D = ashr int %X, ubyte 6 It's unclear whether this should be fixed or not. Its also unclear how to fix it unless we add a "-disable-constant-folding" switch to vmcore. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 15:34:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 15:34:35 -0600 Subject: [LLVMbugs] [Bug 1147] test/ExecutionEngine/test-fp.ll sometimes fails (FoldingSet.cpp) Message-ID: <200702012134.l11LYZcH011956@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1147 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From rspencer at x10sys.com 2007-02-01 15:34 ------- This just failed again for me. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 1 16:05:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 1 Feb 2007 16:05:25 -0600 Subject: [LLVMbugs] [Bug 1147] test/ExecutionEngine/test-fp.ll sometimes fails (FoldingSet.cpp) Message-ID: <200702012205.l11M5Pou013295@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1147 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-01 16:05 ------- The test case wasn't initializing the data passed to the function. This was fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043687.html Fixed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 15:22:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 15:22:20 -0600 Subject: [LLVMbugs] [Bug 1157] NEW: gccas -time-passes -stats doesn't work Message-ID: <200702022122.l12LMKth023934@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1157 Summary: gccas -time-passes -stats doesn't work Product: tools Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: gccas AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org gccas.sh should pass unknown options down to opt. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 15:51:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 15:51:30 -0600 Subject: [LLVMbugs] [Bug 1157] gccas -time-passes -stats doesn't work Message-ID: <200702022151.l12LpU6f024474@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1157 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-02 15:51 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043753.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 16:11:07 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 16:11:07 -0600 Subject: [LLVMbugs] [Bug 1158] NEW: passmgr bug: Analysis used but not available! Message-ID: <200702022211.l12MB7Tq024894@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1158 Summary: passmgr bug: Analysis used but not available! Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org $ echo "define i8 @test() { unreachable }" | llvm-as | opt -domtree -idom -gcse -etforest -constmerge - disable-output PassManager.cpp:663: failed assertion `0 && "Analysis used but not available!"' Abort -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 16:25:40 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 16:25:40 -0600 Subject: [LLVMbugs] [Bug 51] [passmgr] Inliner should only inline functions after global optimization Message-ID: <200702022225.l12MPeND025318@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=51 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-02-02 16:25 ------- Yup. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 17:41:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 17:41:39 -0600 Subject: [LLVMbugs] [Bug 1159] NEW: I'm impatient, please make llvm-test faster Message-ID: <200702022341.l12NfdkZ027246@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1159 Summary: I'm impatient, please make llvm-test faster Product: Test Suite Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Nightly Tester AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Ideas: 1. Instead of using llvm-gcc -S -emit-llvm + llvm-as to get to the .rbc file for each.c file, just use llvm- gcc -c -emit-llvm. 2. To get from linked.rbc to linked.bc, we currently invoke llvm-dis, then 'gccas' (a shell script which invokes llvm-dis then opt). It would be much better to just run opt! Both of these would significantly reduce the disk I/O, disk space required, and cpu time requirements for running llvm-test. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 22:32:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 22:32:55 -0600 Subject: [LLVMbugs] [Bug 1159] I'm impatient, please make llvm-test faster Message-ID: <200702030432.l134Wt5D001827@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1159 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-02 22:32 ------- This is now implemented, with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043789.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043791.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043792.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043793.html This significantly speeds up llvm-test, especially on the larger test cases. Even the smaller test suites are sped up by a second or two. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 2 22:41:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 2 Feb 2007 22:41:52 -0600 Subject: [LLVMbugs] [Bug 1160] NEW: loop code badly optimized Message-ID: <200702030441.l134fqU6002152@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1160 Summary: loop code badly optimized Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca This small program produces bad x86 assembly code: unsigned int huge[2048]; unsigned int target[2048]; extern void foo(); int main(int argc, char *argv[]) { foo(); for (unsigned int i = 0; i < 2048; i++) { target[i] |= huge[i]; } unsigned int x = 0; for (unsigned int i = 0; i < 2048; i++) { x += target[i]; } return x; } There are two problems with the resulting x86 assembly. For the body of the first loop, LLVM produces: movl huge(,%eax,4), %ecx orl target(,%eax,4), %ecx movl %ecx, target(,%eax,4) when it should really be loading from huge, then or'ing into target. In the second loop, LLVM mis-predicts the exit branch, which is inexcusable seeing as the loop is countable: .LBB1_5: #bb23 cmpl $2048, %ecx jne .LBB1_3 #bb15 <--- should be reversed .LBB1_6: #bb27 addl $4, %esp ret .LBB1_3: #bb15 addl target(,%ecx,4), %eax incl %ecx jmp .LBB1_5 #bb23 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 11:57:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 11:57:20 -0600 Subject: [LLVMbugs] [Bug 1156] llvm-upgrade folds constant expressions Message-ID: <200702031757.l13HvK14025133@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1156 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From rspencer at x10sys.com 2007-02-03 11:57 ------- Not worth fixing. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 13:13:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 13:13:34 -0600 Subject: [LLVMbugs] [Bug 909] Assertion failure in DSAA Message-ID: <200702031913.l13JDYtg026737@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=909 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |LATER ------- Additional Comments From sabre at nondot.org 2007-02-03 13:13 ------- Andrew, if you'd like, feel free to reopen this. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 14:25:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 14:25:02 -0600 Subject: [LLVMbugs] [Bug 1161] NEW: build system ignores many user set directories Message-ID: <200702032025.l13KP2c0028266@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1161 Summary: build system ignores many user set directories Product: Build scripts Version: 1.9 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: ismail at pardus.org.tr configure --help mentions these: --bindir=DIR user executables [EPREFIX/bin] --sbindir=DIR system admin executables [EPREFIX/sbin] --libexecdir=DIR program executables [EPREFIX/libexec] --sysconfdir=DIR read-only single-machine data [PREFIX/etc] --sharedstatedir=DIR modifiable architecture-independent data [PREFIX/com] --localstatedir=DIR modifiable single-machine data [PREFIX/var] --libdir=DIR object code libraries [EPREFIX/lib] --includedir=DIR C header files [PREFIX/include] --oldincludedir=DIR C header files for non-gcc [/usr/include] --datarootdir=DIR read-only arch.-independent data root [PREFIX/share] --datadir=DIR read-only architecture-independent data [DATAROOTDIR] --infodir=DIR info documentation [DATAROOTDIR/info] --localedir=DIR locale-dependent data [DATAROOTDIR/locale] --mandir=DIR man documentation [DATAROOTDIR/man] --docdir=DIR documentation root [DATAROOTDIR/doc/-llvm-] --htmldir=DIR html documentation [DOCDIR] --dvidir=DIR dvi documentation [DOCDIR] --pdfdir=DIR pdf documentation [DOCDIR] --psdir=DIR ps documentation [DOCDIR] But none of these have any affect because Makefile.config.in has: PROJ_bindir := $(DESTDIR)$(PROJ_prefix)/bin PROJ_libdir := $(DESTDIR)$(PROJ_prefix)/lib PROJ_datadir := $(DESTDIR)$(PROJ_prefix)/share PROJ_docsdir := $(DESTDIR)$(PROJ_prefix)/docs/llvm PROJ_etcdir := $(DESTDIR)$(PROJ_prefix)/etc/llvm PROJ_includedir := $(DESTDIR)$(PROJ_prefix)/include PROJ_infodir := $(DESTDIR)$(PROJ_prefix)/info PROJ_mandir := $(DESTDIR)$(PROJ_prefix)/man so all of these directories are hardcoded. This makes it hard/ugly to package llvm. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 15:16:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 15:16:19 -0600 Subject: [LLVMbugs] [Bug 1162] NEW: MachineModuleInfo should migrate away from Uniquevector. Message-ID: <200702032116.l13LGJ0G029325@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1162 Summary: MachineModuleInfo should migrate away from Uniquevector. 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: jlaskey at apple.com UniqueVector is very expensive: its cost is the sum of the cost of maintaining both the map and vector, it has high complexity, high constant factors, and produces a lot of malloc traffic. It should be avoided. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 16:06:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 16:06:30 -0600 Subject: [LLVMbugs] [Bug 1163] NEW: Module should not use SetVector of libraries Message-ID: <200702032206.l13M6UfI030253@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1163 Summary: Module should not use SetVector of libraries Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org It should just contain a vector of libraries, and clients that manipulate it should ensure that duplicates do not get entered. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 17:19:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 17:19:15 -0600 Subject: [LLVMbugs] [Bug 1072] Investigate elimination of the -raise pass Message-ID: <200702032319.l13NJF8w031868@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1072 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-03 17:19 ------- The -raise pass was remvoed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043817.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043818.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043819.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043820.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043821.html Resolved. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 17:33:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 17:33:39 -0600 Subject: [LLVMbugs] [Bug 1164] NEW: CBE can generate name conflicts Message-ID: <200702032333.l13NXdOW032224@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1164 Summary: CBE can generate name conflicts Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider: @G = global i32 123 @ltmp_0_1 = global i32 123 define i32 @test(i32 *%G) { %A = load i32* %G %B = load i32* @ltmp_0_1 %C = add i32 %A, %B ret i32 %C } The CBE generated code is: unsigned int test(unsigned int *ltmp_0_1) { ... ltmp_1_2 = *ltmp_0_1; // local ltmp_2_2 = *(<mp_0_1); // global return (ltmp_1_2 + ltmp_2_2); ... This is a serious bug. A not-as-serious bug: the CBE ignores the local names of symbols, turning %A into ltmp_1_2? -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 18:43:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 18:43:25 -0600 Subject: [LLVMbugs] [Bug 1163] Module should not use SetVector of libraries Message-ID: <200702040043.l140hP2G001706@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1163 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-03 18:43 ------- Fixed with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043837.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043838.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043839.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043840.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043841.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 3 19:08:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 3 Feb 2007 19:08:51 -0600 Subject: [LLVMbugs] [Bug 1151] llvm-upgrade deletes Type instances !! Message-ID: <200702040108.l1418pOO002644@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1151 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-03 19:08 ------- works for me now, thanks Reid! ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 08:40:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 08:40:47 -0600 Subject: [LLVMbugs] [Bug 963] x86 JIT miscompilation of ldecod Message-ID: <200702041440.l14Eel1W019757@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=963 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-02-04 08:40 ------- Fixed with http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043880.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043881.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043883.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043882.html Reopen bug in case of any problems. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 12:57:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 12:57:46 -0600 Subject: [LLVMbugs] [Bug 1165] NEW: 'make bugpoint-gccas' broken in externals dirs Message-ID: <200702041857.l14IvkjB024538@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1165 Summary: 'make bugpoint-gccas' broken in externals dirs Product: Test Suite Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Programs Tests AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org e.g. SPEC. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 13:12:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 13:12:25 -0600 Subject: [LLVMbugs] [Bug 1166] NEW: llvm-gcc4 should use gcc testsuite Message-ID: <200702041912.l14JCPPt024856@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1166 Summary: llvm-gcc4 should use gcc testsuite Product: Test Suite Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Programs Tests AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru We should investigate the usage of gcc testsuite for testing llvm-gcc4. At least some part of it can be used to test llvm-gcc. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 13:40:41 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 13:40:41 -0600 Subject: [LLVMbugs] [Bug 1165] 'make bugpoint-gccas' broken in externals dirs Message-ID: <200702041940.l14JefWL025353@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1165 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-04 13:40 ------- Fixed with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043884.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 14:21:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 14:21:57 -0600 Subject: [LLVMbugs] [Bug 1167] NEW: all 'gccas' tests fail due to recent gccas changes Message-ID: <200702042021.l14KLvlv026150@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1167 Summary: all 'gccas' tests fail due to recent gccas changes Product: Test Suite Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Nightly Tester AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The new 'gccas' doesn't output timing info and stats in the right place, causing the nightly tester to report gccas failing every test. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 15:13:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 15:13:00 -0600 Subject: [LLVMbugs] [Bug 1168] NEW: Assertion failed in llvm-gcc Message-ID: <200702042113.l14LD0FW027098@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1168 Summary: Assertion failed in llvm-gcc Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Let's start our gcc.c-torture series. llvm-gcc 20000804-1.c cc1: /home/asl/proj/llvm/src/lib/VMCore/Value.cpp:39: llvm::Value::Value(const llvm::Type*, unsigned int, const std::string&): Assertion `(Ty->isFirstClassType() || Ty == Type::VoidTy || isa(ty)) && "Cannot create non-first-class values except for constants!"' failed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 15:16:19 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 15:16:19 -0600 Subject: [LLVMbugs] [Bug 1169] NEW: AsmPrinter assertion Message-ID: <200702042116.l14LGJPw027200@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1169 Summary: AsmPrinter assertion Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru llvm-gcc 20010327-1.c -S cc1: /home/asl/proj/llvm/src/lib/CodeGen/AsmPrinter.cpp:676: void llvm::AsmPrinter::EmitConstantValueOnly(const llvm::Constant*): Assertion `0 && "Unsupported operator!"' failed. 20010327-1.c:9: internal compiler error: Aborted Please submit a full bug report, with preprocessed source if appropriate. See for instructions. .c source & .ll attached ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 15:28:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 15:28:34 -0600 Subject: [LLVMbugs] [Bug 1170] NEW: llvm-gcc cycles on pretty simple testcase Message-ID: <200702042128.l14LSYrc027470@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1170 Summary: llvm-gcc cycles on pretty simple testcase Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached testcase causes llvm-gcc to eat all memory & die. Testcase actually was from: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5615 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 16:02:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 16:02:09 -0600 Subject: [LLVMbugs] [Bug 1171] NEW: domset runs out of memory Message-ID: <200702042202.l14M29Gi028046@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1171 Summary: domset runs out of memory Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca This attachment causes -domset to eat up more than 512MB of RAM. Test with "ulimit -v $((512*1024))" ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 16:27:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 16:27:17 -0600 Subject: [LLVMbugs] [Bug 1172] NEW: Assertion in LegalizeDAG for vector stuff Message-ID: <200702042227.l14MRH0C028615@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1172 Summary: Assertion in LegalizeDAG for vector stuff Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru NODE: 0x8a24030: i64 = vbit_convert 0x8a22cc8 cc1: /home/asl/proj/llvm/src/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4602: void ::SelectionDAGLegalize::ExpandOp(llvm::SDOperand, llvm::SDOperand&, llvm::SDOperand&): Assertion `0 && "Do not know how to expand this operator!"' failed. simd-1.c:7: internal compiler error: Aborted Please submit a full bug report, with preprocessed source if appropriate. See for instructions. Attached are two sources from gcc.c-torture ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 16:41:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 16:41:33 -0600 Subject: [LLVMbugs] [Bug 1173] NEW: Assertion failed in EmitLV() Message-ID: <200702042241.l14MfXsU028960@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1173 Summary: Assertion failed in EmitLV() Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru I think attached two sources are triggering the same bug: 20050516-1.c:12: internal compiler error: in EmitLV, at llvm-convert.cpp:5037 Please submit a full bug report, with preprocessed source if appropriate. See for instructions. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 16:45:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 16:45:52 -0600 Subject: [LLVMbugs] [Bug 1174] NEW: Another assertion in EmitLV Message-ID: <200702042245.l14MjqhS029089@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1174 Summary: Another assertion in EmitLV Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru 20030224-1.c: In function 'zzz': 20030224-1.c:2: internal compiler error: in EmitLV, at llvm-convert.cpp:675 Please submit a full bug report, with preprocessed source if appropriate. See for instructions. C source attached ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 16:47:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 16:47:15 -0600 Subject: [LLVMbugs] [Bug 1175] NEW: Floating point exception in FE Message-ID: <200702042247.l14MlFNw029167@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1175 Summary: Floating point exception in FE Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached source causes: 20021110.c: In function 'foo': 20021110.c:8: internal compiler error: Floating point exception Please submit a full bug report, with preprocessed source if appropriate. See for instructions. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 17:28:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 17:28:55 -0600 Subject: [LLVMbugs] [Bug 1169] AsmPrinter assertion Message-ID: <200702042328.l14NStmp029938@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1169 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 17:56:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 17:56:04 -0600 Subject: [LLVMbugs] [Bug 1175] Floating point exception in FE Message-ID: <200702042356.l14Nu4gf030640@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1175 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-04 17:56 ------- Fixed (by Anton), patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043901.html testcase here: test/CFrontend/2007-02-04-EmptyStruct.c -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 19:03:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 19:03:56 -0600 Subject: [LLVMbugs] [Bug 1176] NEW: llvm-gcc should support compound expressions for constant lvalues Message-ID: <200702050103.l1513uPu031968@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1176 Summary: llvm-gcc should support compound expressions for constant lvalues Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru This is second part of PR1173. llvm-gcc should support emitting of compound expressions as lvalues. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 20:19:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 20:19:57 -0600 Subject: [LLVMbugs] [Bug 1173] Assertion failed in EmitLV() Message-ID: <200702050219.l152JvSx000959@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1173 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-04 20:19 ------- Anton fixed this, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043904.html Testcase here: test/CFrontend/2007-02-04-AddrLValue.c -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 21:51:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 21:51:43 -0600 Subject: [LLVMbugs] [Bug 1176] llvm-gcc should support compound expressions for constant lvalues Message-ID: <200702050351.l153ph34002578@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1176 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-04 21:51 ------- Fixed, testcase here: test/CFrontend/2007-02-04-AddrLValue.c Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043909.html Thanks Anton! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 23:03:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 23:03:13 -0600 Subject: [LLVMbugs] [Bug 1177] NEW: DenseMap::grow causes memory corruption Message-ID: <200702050503.l1553DcB003883@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1177 Summary: DenseMap::grow causes memory corruption Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Support Libraries AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com I'm getting this glibc detected failure: /proj/llvm/llvm-3/Debug/bin/opt: malloc(): memory corruption: 0x08de9f1c Occurring at: llvm::DenseMap::grow()+0x34 llvm::DenseMap::InsertIntoBucket(llvm::DominatorTreeBase::Node* const&, llvm::Value* const&, std::pair*) llvm::DenseMap::operator[](llvm::DominatorTreeBase::Node* const&) (anonymous namespace)::LCSSA::GetValueForBlock(llvm::DominatorTreeBase::Node*,llvm::Instruction*, llvm::DenseMap&) (anonymous namespace)::LCSSA::ProcessInstruction(llvm::Instruction*,std::vector > const&) (anonymous namespace)::LCSSA::visitSubloop(llvm::Loop*) (anonymous namespace)::LCSSA::visitSubloop(llvm::Loop*) (anonymous namespace)::LCSSA::runOnFunction(llvm::Function&) llvm::FPPassManager::runOnFunction(llvm::Function&) CGPassManager::runOnModule(llvm::Module&) llvm::MPPassManager::runOnModule(llvm::Module&) llvm::PassManagerImpl::run(llvm::Module&) llvm::PassManager::run(llvm::Module&) main ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 23:25:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 23:25:30 -0600 Subject: [LLVMbugs] [Bug 1177] LCSSA is broken due to use of DenseMap Message-ID: <200702050525.l155PUSE004386@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1177 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|normal |major Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From rspencer at x10sys.com 2007-02-04 23:25 ------- Reverting this patch fixed the problem: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070129/043914.html Resolved. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 4 23:26:38 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 4 Feb 2007 23:26:38 -0600 Subject: [LLVMbugs] [Bug 1167] all 'gccas' tests fail due to recent gccas changes Message-ID: <200702050526.l155QcBU004421@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1167 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From sabre at nondot.org 2007-02-04 23:26 ------- I will update my llvm-gcc, if it doesn't fix this, I'll reopen. Thanks! ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 10:52:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 10:52:59 -0600 Subject: [LLVMbugs] [Bug 1178] NEW: Sometimes AsmPrinter::PrintSpecial assigns the same uid for two Machine Instructions Message-ID: <200702051652.l15GqxBq025390@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1178 Summary: Sometimes AsmPrinter::PrintSpecial assigns the same uid for two Machine Instructions Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com The code: void AsmPrinter::PrintSpecial(const MachineInstr *MI, const char *Code) { [...] } else if (!strcmp(Code, "uid")) { // Assign a unique ID to this machine instruction. static const MachineInstr *LastMI = 0; static unsigned Counter = 0U-1; // If this is a new machine instruction, bump the counter. if (LastMI != MI) { ++Counter; LastMI = MI; } O << Counter; } else { [...] } ************************** Comparison (LastMI != MI) is incorrect. Machine Instructions of different Machine Functions can have the same address. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 11:41:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 11:41:22 -0600 Subject: [LLVMbugs] [Bug 1179] NEW: -indvars pass fails to replace exit value of loop when profitable Message-ID: <200702051741.l15HfMw3027834@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1179 Summary: -indvars pass fails to replace exit value of loop when profitable Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Loop Optimizer AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The -indvars pass is not computing the exit value of j in either of these cases: int ltst(int x) { int i, j; j=0; for(i=0; i http://llvm.org/bugs/show_bug.cgi?id=1152 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-05 12:41 ------- This conversion has been completed. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 13:35:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 13:35:25 -0600 Subject: [LLVMbugs] [Bug 1158] passmgr bug: Analysis used but not available! Message-ID: <200702051935.l15JZPhT031540@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1158 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-02-05 13:35 ------- Fixed.. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/043936.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 15:22:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 15:22:15 -0600 Subject: [LLVMbugs] [Bug 1180] NEW: Linker is slower than it needs to be when linking 1 file Message-ID: <200702052122.l15LMF6D002359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1180 Summary: Linker is slower than it needs to be when linking 1 file Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Linker AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com In situations where the linker is processing a single bytecode file (as in many test cases in llvm-test), the processing moves the bytecode from the original module to an empty destination module. This is wasteful and takes significant time on large bytecode files. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 15:23:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 15:23:21 -0600 Subject: [LLVMbugs] [Bug 411] [SymbolTable] Reconsider one symbol table for each type Message-ID: <200702052123.l15LNLSZ002418@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=411 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-05 15:23 ------- w00t! This is finally done with a series of commits today. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 15:24:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 15:24:57 -0600 Subject: [LLVMbugs] [Bug 1178] Sometimes AsmPrinter::PrintSpecial assigns the same uid for two Machine Instructions Message-ID: <200702052124.l15LOvae002523@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1178 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |miscompilation OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-05 15:24 ------- This patch should fix this: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/043978.html With no testcase, I can't tell for sure though. Please reopen if this isn't enough to fix the issue. Thanks! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 15:37:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 15:37:04 -0600 Subject: [LLVMbugs] [Bug 1181] NEW: C backend assertion: (castIsValid(getOpcode(), S, Ty) && "Illegal Trunc") Message-ID: <200702052137.l15Lb4YI002835@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1181 Summary: C backend assertion: (castIsValid(getOpcode(), S, Ty) && "Illegal Trunc") Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: evan.cheng at apple.com llc -march=c triggers this assertion. Test case is: target datalayout = "e-p:64:64" target triple = "x86_64-apple-darwin8" implementation ; Functions: declare void @llvm.memset.i64(i8*, i8, i64, i32) define fastcc void @InitUser_data_unregistered() { entry: tail call void @llvm.memset.i64( i8* null, i8 0, i64 65496, i32 1 ) ret void } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 22:13:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 22:13:51 -0600 Subject: [LLVMbugs] [Bug 1182] NEW: predsimplify makes lamba 10% slower Message-ID: <200702060413.l164Dp8m019008@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1182 Summary: predsimplify makes lamba 10% slower Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca Adding -predsimplify to the compilation of MultiSource/Applications/lambda-0.1.3/lambda causes it to take 10% longer. It appears that the optimization is actually triggering bad behaviour in the backend. I've extracted just "main" to illustrate the problem. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 5 23:59:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 5 Feb 2007 23:59:20 -0600 Subject: [LLVMbugs] [Bug 1149] stage1 fails to build the stage2 on x86 Message-ID: <200702060559.l165xKI0020849@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1149 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From sabre at nondot.org 2007-02-05 23:59 ------- This is a bug, but I don't think it is worth fixing. Performing a top-level bootstrap like this turns warnings into errors. While getting GCC warning free would be nice, I don't think it's worth the maintenance burden (for future merges) to do this. Just using 'make bootstrap' will test exactly the same thing as a top-level configured bootstrap. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 6 00:08:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 6 Feb 2007 00:08:42 -0600 Subject: [LLVMbugs] [Bug 1181] C backend assertion: (castIsValid(getOpcode(), S, Ty) && "Illegal Trunc") Message-ID: <200702060608.l1668g71021164@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1181 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-06 00:08 ------- Fixed, testcase here: CodeGen/CBackend/2007-02-05-memset.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044016.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 6 12:51:30 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 6 Feb 2007 12:51:30 -0600 Subject: [LLVMbugs] [Bug 1183] NEW: CBE generated code often provokes GCC ICE's Message-ID: <200702061851.l16IpUP5030736@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1183 Summary: CBE generated code often provokes GCC ICE's Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: sabre at nondot.org ReportedBy: sabre at nondot.org We're generating code like this now: ltmp_2731_1 = ((unsigned char * (*) (unsigned char *, unsigned char , unsigned int ))(void*)memset) (ltmp_2730_1, ((unsigned char )0), ltmp_34_2); Which is causing tons of GCC ice's. :( -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 6 13:07:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 6 Feb 2007 13:07:33 -0600 Subject: [LLVMbugs] [Bug 1183] CBE generated code often provokes GCC ICE's Message-ID: <200702061907.l16J7X4j031518@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1183 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-06 13:07 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044039.html testcase: most of llvm-test. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 6 19:24:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 6 Feb 2007 19:24:39 -0600 Subject: [LLVMbugs] [Bug 1184] NEW: ScheduleDAG.cpp patch: error message more meaningful. Message-ID: <200702070124.l171OdII011889@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1184 Summary: ScheduleDAG.cpp patch: error message more meaningful. Product: libraries Version: trunk Platform: All OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org Small patch to make tracking down (self-inflicted) bugs easier when the scheduler barfs. The assert() wasn't really helpful at all. At least this patch tells you which instruction and operand to go look at. (Note: In the end, yes, it was my fault, but this patch actually helped.) Index: lib/CodeGen/SelectionDAG/ScheduleDAG.cpp =================================================================== --- lib/CodeGen/SelectionDAG/ScheduleDAG.cpp (revision 522) +++ lib/CodeGen/SelectionDAG/ScheduleDAG.cpp (working copy) @@ -301,9 +301,20 @@ if (II) { const TargetRegisterClass *RC = getInstrOperandRegClass(MRI, TII, II, IIOpNum); + const TargetRegisterClass *VRC = RegMap->getRegClass(VReg); assert(RC && "Don't have operand info for this instruction!"); - assert(RegMap->getRegClass(VReg) == RC && - "Register class of operand and regclass of use don't agree!"); + if (VRC != RC) { + cerr << "Register class of operand and regclass of use don't agree!\n"; + cerr << "Operand = " << IIOpNum << "\n"; + cerr << "Op->Val = "; Op.Val->dump(0); cerr << "\n"; + cerr << "MI = "; MI->print(cerr); + cerr << "VReg = " << VReg << "\n"; + cerr << "VReg RegClass size = " << VRC->getSize() + << ", align = " << VRC->getAlignment() << "\n"; + cerr << "Expected RegClass size = " << RC->getSize() + << ", align = " << RC->getAlignment() << "\n"; + abort(); + } } } else if (ConstantSDNode *C = dyn_cast(Op)) { ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 6 22:48:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 6 Feb 2007 22:48:10 -0600 Subject: [LLVMbugs] [Bug 230] Automate llvm-gcc build process as a "project" Message-ID: <200702070448.l174mAZe015631@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=230 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From sabre at nondot.org 2007-02-06 22:48 ------- > NB: llvm-gcc depends on tools. The "llvm bootstrap" process involves building > the tools first, then llvm-gcc, and then any other runtime libraries, projects, etc. This is no longer true. I don't think this is worth worrying about now. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 13:40:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 13:40:26 -0600 Subject: [LLVMbugs] [Bug 1185] NEW: [llvm-gcc] Remove EmittedGlobalVars and EmittedFunctions maps from Message-ID: <200702071940.l17JeQca015440@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1185 Summary: [llvm-gcc] Remove EmittedGlobalVars and EmittedFunctions maps from Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Now that PR411 has landed, these are fully redundant with the module symbol table. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 14:22:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 14:22:58 -0600 Subject: [LLVMbugs] [Bug 1186] NEW: Assertion failure in PassManager.cpp:643 Message-ID: <200702072022.l17KMwxi016724@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1186 Summary: Assertion failure in PassManager.cpp:643 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: babic.domagoj at gmail.com Build the attached example as loadable module, run -tst3. Used to work with the old PassManager [last time I updated was Dec 30, 06']. Domagoj opt: /work/projects/llvm/lib/VMCore/PassManager.cpp:643: void llvm::PMDataManager::collectRequiredAnalysisPasses(std::vector >&, llvm::Pass*): Assertion `AnalysisPass && "Analysis pass is not available"' failed. opt((anonymous namespace)::PrintStackTrace()+0x17)[0xa1f13b] opt((anonymous namespace)::SignalHandler(int)+0x10e)[0xa1f40a] /lib/libc.so.6[0x2aaaab39f1b0] /lib/libc.so.6(gsignal+0x3d)[0x2aaaab39f11d] /lib/libc.so.6(abort+0xfe)[0x2aaaab3a084e] /lib/libc.so.6(__assert_fail+0xf1)[0x2aaaab3988f1] opt(llvm::PMDataManager::collectRequiredAnalysisPasses(std::vector >&, llvm::Pass*)+0xbc)[0x9b1502] opt(llvm::PMDataManager::add(llvm::Pass*, bool)+0xc1)[0x9b2d31] opt(llvm::FunctionPass::assignPassManager(llvm::PMStack&, llvm::PassManagerType)+0x1bb)[0x9b33b9] opt(llvm::PassManagerImpl::addTopLevelPass(llvm::Pass*)+0xdf)[0x9bb8bf] opt(llvm::PMTopLevelManager::schedulePass(llvm::Pass*)+0x115)[0x9b0d8b] opt(llvm::PassManagerImpl::add(llvm::Pass*)+0x24)[0x9b6998] opt(llvm::PassManager::add(llvm::Pass*)+0x20)[0x9b0db6] opt((anonymous namespace)::addPass(llvm::PassManager&, llvm::Pass*)+0x1d)[0x761f05] opt(main+0x5e1)[0x76036b] /lib/libc.so.6(__libc_start_main+0xdb)[0x2aaaab38c49b] opt(__gxx_personality_v0+0x12a)[0x75278a] Aborted ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 16:27:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 16:27:26 -0600 Subject: [LLVMbugs] [Bug 1187] NEW: Invalid definition error in llvm-upgrade Message-ID: <200702072227.l17MRQKu026273@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1187 Summary: Invalid definition error in llvm-upgrade Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-upgrade AssignedTo: unassignedbugs at nondot.org ReportedBy: babic.domagoj at gmail.com llvm-upgrade -emit-bytecode file.ll -o tst.bc terminates with the following error: llvm-upgrade: file.ll:258527: error: Reference to an invalid definition: 'hash' of type 'i32 (i8 *, i32) *' ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 18:08:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 18:08:43 -0600 Subject: [LLVMbugs] [Bug 1188] NEW: BitMask error Message-ID: <200702080008.l1808hCA017010@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1188 Summary: BitMask error Product: libraries Version: 1.8 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Interpreter AssignedTo: unassignedbugs at nondot.org ReportedBy: wenwenti at hotmail.com I find there is a potential bitmask error with long long type. In the INTEGER_ASSIGN macro, the program uses (1ull << BitWidth) -1 to get the bitmask. If BitWidth = 64, the value is 0. The correct way is to use ~(uint64_t)(0ull) >> (BitWidth-1). Please review the following patch. Index: lib/ExecutionEngine/Interpreter/Execution.cpp =================================================================== RCS file: /var/cvs/llvm/llvm/lib/ExecutionEngine/Interpreter/Execution.cpp,v retrieving revision 1.167 diff -t -d -u -p -5 -r1.167 Execution.cpp --- lib/ExecutionEngine/Interpreter/Execution.cpp 2 Feb 2007 02:16:22 -0000 1.167 +++ lib/ExecutionEngine/Interpreter/Execution.cpp 7 Feb 2007 23:58:56 -0000 @@ -1305,11 +1305,11 @@ void Interpreter::visitAShr(BinaryOperat SetValue(&I, Dest, SF); } #define INTEGER_ASSIGN(DEST, BITWIDTH, VAL) \ { \ - uint64_t Mask = (1ull << BITWIDTH) - 1; \ + uint64_t Mask = ~(uint64_t)(0ull) >> (64-BITWIDTH); \ if (BITWIDTH == 1) { \ Dest.Int1Val = (bool) (VAL & Mask); \ } else if (BITWIDTH <= 8) { \ Dest.Int8Val = (uint8_t) (VAL & Mask); \ } else if (BITWIDTH <= 16) { \ Index: lib/ExecutionEngine/Interpreter/Interpreter.h =================================================================== RCS file: /var/cvs/llvm/llvm/lib/ExecutionEngine/Interpreter/Interpreter.h,v retrieving revision 1.82 diff -t -d -u -p -5 -r1.82 Interpreter.h --- lib/ExecutionEngine/Interpreter/Interpreter.h 2 Feb 2007 02:16:22 -0000 1.82 +++ lib/ExecutionEngine/Interpreter/Interpreter.h 7 Feb 2007 23:58:56 -0000 @@ -234,11 +234,11 @@ private: // Helper functions void popStackAndReturnValueToCaller(const Type *RetTy, GenericValue Result); }; inline void maskToBitWidth(GenericValue& GV, unsigned BitWidth) { - uint64_t BitMask = (1ull << BitWidth) - 1; + uint64_t BitMask = ~(uint64_t)(0ull) >> (64-BitWidth); if (BitWidth <= 8) GV.Int8Val &= BitMask; else if (BitWidth <= 16) GV.Int16Val &= BitMask; else if (BitWidth <= 32) ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 18:30:47 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 18:30:47 -0600 Subject: [LLVMbugs] [Bug 1188] BitMask error Message-ID: <200702080030.l180UlA0018446@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1188 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-07 18:30 ------- Looks right to me. Patches applied. Thanks, Leo. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 7 22:36:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 7 Feb 2007 22:36:10 -0600 Subject: [LLVMbugs] [Bug 947] llvm-gcc4 doesn't support GCC address of label extension Message-ID: <200702080436.l184aAGl022917@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=947 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 Version|trunk |1.7 ------- Additional Comments From sabre at nondot.org 2007-02-07 22:36 ------- Implemented. Patches here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044158.html Testcase here: test/CFrontend/2007-02-07-AddrLabel.c I verified that the other cases attached to this PR work as well. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 02:39:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 02:39:11 -0600 Subject: [LLVMbugs] [Bug 1189] NEW: llvm-convert.cpp - intrinsic getOrInsertFunction calls should be replaced by Intrinsic::getDeclaration Message-ID: <200702080839.l188dBjm003517@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1189 Summary: llvm-convert.cpp - intrinsic getOrInsertFunction calls should be replaced by Intrinsic::getDeclaration Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: jlaskey at apple.com Purely to simplify the code in the FE. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 03:15:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 03:15:49 -0600 Subject: [LLVMbugs] [Bug 1187] Invalid definition error in llvm-upgrade Message-ID: <200702080915.l189FnNW004400@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1187 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-08 03:15 ------- Resolved with these patches: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044130.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044131.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044166.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044173.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044174.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044175.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 10:12:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 10:12:56 -0600 Subject: [LLVMbugs] [Bug 1190] NEW: llvm-gcc does not support pointers to nested functions (trampolines) Message-ID: <200702081612.l18GCuDY012778@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1190 Summary: llvm-gcc does not support pointers to nested functions (trampolines) Product: new-bugs Version: unspecified Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: baldrick at free.fr ReportedBy: baldrick at free.fr This PR is to track progress of, and discussion on, support for pointers to nested functions in llvm-gcc, aka trampolines. Implementing this inevitably needs help from the llvm code generators. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 10:15:37 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 10:15:37 -0600 Subject: [LLVMbugs] [Bug 915] llvm-gcc doesn't support nested functions Message-ID: <200702081615.l18GFbMS012836@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=915 baldrick at free.fr changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From baldrick at free.fr 2007-02-08 10:15 ------- Nested subroutines as used in the examples in this PR are now supported. However there is no support yet for trampolines, i.e. taking the address of a nested subroutine, nor for non-local gotos. I've opened PR1190 for tracking progress on trampolines. I didn't bother for non-local gotos because it's less clear that anyone is interested in that. By the way, the testcase in attachment 625 still fails to compile, because tree-nested.c goes into an infinite loop, i.e. this seems to be a gcc bug not an LLVM one. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 14:22:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 14:22:57 -0600 Subject: [LLVMbugs] [Bug 1191] NEW: Factor standard pass manager construction to separate library Message-ID: <200702082022.l18KMvsn019782@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1191 Summary: Factor standard pass manager construction to separate library Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Transformation Utilities AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Right now, we have opt, llvmc and llvm-ld providing various options for controlling which passes get run to perform compile time and link time optimization, stripping names, stripping debug info, disabling inlining, and a variety of other things. We also want opt, llvmc, and llvm-ld to support the -O options to control the level of optimization applied to a module. It seems that it would be useful to these tools, and perhaps other software using llvm, if all this logic for building a pass manager with these various options was built into a library. The library would provide these options: -O0 Disable optimization -O1 Optimize for link/compile speed, not execution speed -O2 Perform only standard compilation optimizations -O3 Perform aggressive compilation optimizations -O4 Perform standard link time optimizations -O5 Perform aggressive link time optimizations -disable-opt Alias for -O0 -disable-inlining Don't run the inliner -disable-internalize Don't run the internalizer, if it would have been -disable-verify Don't verify the final result -std-compile-opts Alias for -O2 -std-link-opts Alias for -O4 -strip-all Strip debug info and symbols -strip-debug Strip debug info but leave symbols -verify-each Run the verifier after each pass and possibly others. Its interface would consist of a single function that filled in a PassManager (provided as an argument) with the requested set of passes. This approach would allow us to have consistent options and results across the llvm tools and make this logic available for other (non-llvm) tools as well. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 8 16:25:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 8 Feb 2007 16:25:46 -0600 Subject: [LLVMbugs] [Bug 1192] NEW: using a ll file as source for debuging info Message-ID: <200702082225.l18MPktD028493@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1192 Summary: using a ll file as source for debuging info Product: tools Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-as AssignedTo: unassignedbugs at nondot.org ReportedBy: alenhar2 at uiuc.edu either a pass, llc, or llvm-as should be able to insert debugging info relative the .ll file/llvm representation. I want to take a bytecode (without debugging symbols), and with some tool invocation magic add debugging intrinsics. Obviously I can't get source language stuff, but I'd like to be able to treat the .ll file as the source file, inspecting virtual registers in gdb, having sane backtraces, etc. I don't see a reason this couldn't be generated from the bytecode. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 9 10:12:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 9 Feb 2007 10:12:05 -0600 Subject: [LLVMbugs] [Bug 1153] Move gccld optimization functionality to llvm-ld Message-ID: <200702091612.l19GC5cu024628@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1153 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-09 10:12 ------- gccld has been replaced with a script: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044210.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044211.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044212.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044231.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044232.html llvm-ld has been made to accept same args as gccld: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044189.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044192.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044209.html llvm-test suite has been updated: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044214.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044215.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044216.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044217.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044218.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044219.html documentation has been updated: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044220.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044221.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044234.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044236.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044237.html Sounds done to me. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 9 20:45:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 9 Feb 2007 20:45:09 -0600 Subject: [LLVMbugs] [Bug 1193] NEW: incomplete optimization Message-ID: <200702100245.l1A2j9HU032084@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1193 Summary: incomplete optimization Product: libraries Version: 1.8 Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: wenwenti at hotmail.com I find a problem in my progrom with bit-accurate types. The C program is as follows. typedef unsigned int __attribute__((bitwidth(11))) uint11; void test_add_sub(uint11 a, uint11 b, uint11* r1, uint11* r2) { *r1 = a + b; *r2 = a -b; return ; } The generated llvm code looks like this ; ModuleID = '' target datalayout = "e-p:32:32" target triple = "i686-pc-linux-gnu" implementation ; Functions: define void @test_add_sub(i11 sext %a, i11 sext %b, i11* %r1, i11* %r2) { entry: %tmp45 = add i11 %b, %a ; [#uses=1] store i11 %tmp45, i11* %r1 %tmp78 = sext i11 %a to i16 ; [#uses=1] %tmp910 = sext i11 %b to i16 ; [#uses=1] %tmp11 = sub i16 %tmp78, %tmp910 ; [#uses=1] %tmp1112 = trunc i16 %tmp11 to i11 ; [#uses=1] store i11 %tmp1112, i11* %r2 ret void } The add instruction using 11-bit is fine. However, the sub instruction uses 16-bit. Then the value is truncated back to 11-bit. The optimization to the add instruction should also be applied to sub instruction. I don't know why it does not. Thanks. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 9 21:00:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 9 Feb 2007 21:00:04 -0600 Subject: [LLVMbugs] [Bug 982] bugpoint crash Message-ID: <200702100300.l1A304Me032432@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=982 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-09 21:00 ------- This was almost certainly fixed on 2006/11/09, by this patch to bugpoint: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20061106/039715.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 10 05:43:24 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 10 Feb 2007 05:43:24 -0600 Subject: [LLVMbugs] [Bug 1194] NEW: bcwriter turns into Message-ID: <200702101143.l1ABhOfH015461@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1194 Summary: bcwriter turns into Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Bytecode Writer AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com Causes this test case to fail: test/Integer/packed_bt.ll with this stack trace: 3 0x005c4859 in __assert_fail () from /lib/libc.so.6 #4 0x081d9c39 in PackedType (this=0x9e3fe70, ElType=0x9e3fed0, NumEl=2) at Type.cpp:429 #5 0x081dcd53 in llvm::PackedType::get (ElementType=0x9e3fed0, NumElements=2) at Type.cpp:1165 #6 0x0815a10a in llvm::BytecodeReader::ParseType (this=0x9e3f788) at Reader.cpp:1019 #7 0x0815a417 in llvm::BytecodeReader::ParseTypes (this=0x9e3f788, Tab=@0x9e3f84c, NumEntries=12) at Reader.cpp:1091 #8 0x0815a537 in llvm::BytecodeReader::ParseGlobalTypes (this=0x9e3f788) at Reader.cpp:1668 #9 0x0815aa0d in llvm::BytecodeReader::ParseModule (this=0x9e3f788) at Reader.cpp:1923 #10 0x0815b3c2 in llvm::BytecodeReader::ParseBytecode (this=0x9e3f788, Buf=0x9e3f910 "llvm\001", Length=178, ModuleID=@0xbff89814, Decompressor=0x820a54e , ErrMsg=0xbff89918) at Reader.cpp:2078 #11 0x0814d5a6 in (anonymous namespace)::BytecodeStdinReader::read (this=0x9e3f788, ErrMsg=0xbff89918) at ReaderWrappers.cpp:185 #12 0x0814d6ce in llvm::getBytecodeModuleProvider (Filename=@0x82554fc, BCDC=0x820a54e , ErrMsg=0xbff89918, H=0x0) at ReaderWrappers.cpp:241 #13 0x0814d711 in llvm::ParseBytecodeFile (Filename=@0x82554fc, BCDC=0x820a54e , ErrMsg=0xbff89918) at ReaderWrappers.cpp:251 #14 0x0814a9e5 in main (argc=1, argv=0xbff89c14) at llvm-dis.cpp:57 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 10 08:19:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 10 Feb 2007 08:19:22 -0600 Subject: [LLVMbugs] [Bug 1194] bcwriter turns <2 x i10> into <2 x opaque> Message-ID: <200702101419.l1AEJM7D004456@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1194 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-10 08:19 ------- This had to do with a combination of things: 1. IntegerType is no longer "primitive" 2. Not all IntegerType instances are inserted early. 3. BCReader in some cases needs IntegerType sub-types to be inserted before (for example) a packed type uses it. 4. Recent slot calculator changes caused non-standard IntegerType instances to get slot calculator slots higher than their users. The above caused a type slot number arrangement that forced the IntegerType sub-type of a PackedType to have a higher slot number than the PackedType. Consequently, bcreader created an OpaqueType for the IntegerType and then tried to instantiate a PackedType with that. The resulting assertion ensued. The fix for this is to just make sure that we insert things in a sane order. This patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044290.html looks at a type's sub-types and inserts any that cannot recurse back before the main type is inserted. For example, this type: <2 x i10>* was previously inserted like this: #9: opaque#10 * #10: <2 x opaque#11> #11: i10 Now, it is inserted like this: #9: i10 #10: <2 x i10> #11: <2 x i10>* ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 10 13:09:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 10 Feb 2007 13:09:45 -0600 Subject: [LLVMbugs] [Bug 1194] bcwriter turns <2 x i10> into <2 x opaque> Message-ID: <200702101909.l1AJ9jsN009398@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1194 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From sabre at nondot.org 2007-02-10 13:09 ------- I'm reopening this, as Reid is reevaluating this approach. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 10 13:09:51 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 10 Feb 2007 13:09:51 -0600 Subject: [LLVMbugs] [Bug 1195] NEW: Rename PackedType as VectorType Message-ID: <200702101909.l1AJ9pkI009422@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1195 Summary: Rename PackedType as VectorType 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: rspencer at x10sys.com Vector is more technically accurate for the purposes of this type. Its also less confusing with PackedStructType. Code, comments, doc must all be udpated. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From asl at math.spbu.ru Sat Feb 10 17:46:41 2007 From: asl at math.spbu.ru (Anton Korobeynikov) Date: Sun, 11 Feb 2007 02:46:41 +0300 (MSK) Subject: [LLVMbugs] =?koi8-r?b?bGx2bS1nY2M0ICYgUENI?= Message-ID: <200702102346.l1ANkfJb019201@star.math.spbu.ru> Hello, Everyone. Are PCH broken with TOT? I'm just having junk in the end of generated .s when sources are compiled with PCH enabled. This "junk" is just bytecode for blank LLVM module. Any ideas, where it can come from? PS: My llvm-gcc4 sourcez here can be little bit out of date (~week). -- WBR, Anton Korobeynikov From sabre at nondot.org Sat Feb 10 18:40:10 2007 From: sabre at nondot.org (Chris Lattner) Date: Sat, 10 Feb 2007 16:40:10 -0800 (PST) Subject: [LLVMbugs] =?koi8-r?b?bGx2bS1nY2M0ICYgUENI?= In-Reply-To: <200702102346.l1ANkfJb019201@star.math.spbu.ru> References: <200702102346.l1ANkfJb019201@star.math.spbu.ru> Message-ID: On Sun, 11 Feb 2007, Anton Korobeynikov wrote: > Are PCH broken with TOT? I'm just having junk in the end of generated .s > when sources are compiled with PCH enabled. This "junk" is just bytecode > for blank LLVM module. Any ideas, where it can come from? > > PS: My llvm-gcc4 sourcez here can be little bit out of date (~week). Yes it is. It was subtly broken until recently. Devang is actively working on fixing it right. Until then, it will probably be obviously broken :). It'll probably take a week, maybe two on the outside to get resolved. -Chris -- http://nondot.org/sabre/ http://llvm.org/ From asl at math.spbu.ru Sun Feb 11 01:58:45 2007 From: asl at math.spbu.ru (Anton Korobeynikov) Date: Sun, 11 Feb 2007 10:58:45 +0300 (MSK) Subject: [LLVMbugs] =?koi8-r?b?bGx2bS1nY2M0ICYgUENI?= Message-ID: <200702110758.l1B7wjiO011399@star.math.spbu.ru> Hello, Chris. > Yes it is. It was subtly broken until recently. Devang is actively > working on fixing it right. Until then, it will probably be obviously > broken :). It'll probably take a week, maybe two on the outside to get > resolved. Well, ok. So, my 5 cents for PCH fixing. Attached patch will allow correct path to PCH file to be printed into preprocessed source. Currently on systems using backward slashes as path separator (e.g. mingw) PCH path is printed with mixed "\" and "/", even more, "\"'s are not esacaped, so later everything got confused :) This patch uses standard gcc way to do platform dependent filename printing via output_quoted_string() call. -- WBR, Anton Korobeynikov -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/octet-stream Size: 752 bytes Desc: not available Url : http://lists.cs.uiuc.edu/pipermail/llvmbugs/attachments/20070211/86a3b3f8/attachment.obj From bugzilla-daemon at cs.uiuc.edu Sun Feb 11 12:23:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 11 Feb 2007 12:23:26 -0600 Subject: [LLVMbugs] [Bug 1196] NEW: ValueSymTab::insert can loop infinitely Message-ID: <200702111823.l1BINQqm024299@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1196 Summary: ValueSymTab::insert can loop infinitely 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: rspencer at x10sys.com When llvm-upgrade calls Value::setName from setValueName it ends up attempting to insert a value in the symbol table. This can recurse infinitely in this loop: // Otherwise, there is a naming conflict. Rename this value. std::string UniqueName = V->getName(); unsigned BaseSize = UniqueName.size(); do { // Trim any suffix off. UniqueName.resize(BaseSize); UniqueName += utostr(++LastUnique); // Try insert the vmap entry with this suffix. } while (!vmap.insert(make_pair(UniqueName, V)).second); Last night's nightly test was halted by this. It looped 22,821,681,606 times before I broke into it with the debugger. Its unclear why the map is refusing to insert the new name. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 11 13:36:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 11 Feb 2007 13:36:29 -0600 Subject: [LLVMbugs] [Bug 1196] ValueSymTab::insert can loop infinitely Message-ID: <200702111936.l1BJaTtf025513@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1196 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-11 13:36 ------- Doh. Uninitialized variable: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070205/044347.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From dpatel at apple.com Mon Feb 12 10:48:36 2007 From: dpatel at apple.com (Devang Patel) Date: Mon, 12 Feb 2007 08:48:36 -0800 Subject: [LLVMbugs] llvm-gcc4 & PCH In-Reply-To: <200702110758.l1B7wjiO011399@star.math.spbu.ru> References: <200702110758.l1B7wjiO011399@star.math.spbu.ru> Message-ID: <72F13AC7-93ED-44B4-ADD4-434EF49CB41A@apple.com> On Feb 10, 2007, at 11:58 PM, Anton Korobeynikov wrote: > Hello, Chris. > >> Yes it is. It was subtly broken until recently. Devang is actively >> working on fixing it right. Until then, it will probably be >> obviously >> broken :). It'll probably take a week, maybe two on the outside to >> get >> resolved. > Well, ok. So, my 5 cents for PCH fixing. Attached patch will allow > correct path to PCH file to be printed into preprocessed source. > Currently on systems using backward slashes as path separator (e.g. > mingw) PCH path is printed with mixed "\" and "/", even more, "\"'s > are > not esacaped, so later everything got confused :) > > This patch uses standard gcc way to do platform dependent filename > printing via output_quoted_string() call. Thanks Anton. I'll apply this patch. - Devang From bugzilla-daemon at cs.uiuc.edu Mon Feb 12 16:03:36 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 12 Feb 2007 16:03:36 -0600 Subject: [LLVMbugs] [Bug 1197] NEW: bytecode causing llc to eat up vast amounts of memory Message-ID: <200702122203.l1CM3a8g030815@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1197 Summary: bytecode causing llc to eat up vast amounts of memory Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr The following llvm bytecode file causes llc to eat up all memory on my machine then die. To reproduce, run as follows: llc exp_attr.bc This is not x86 specific, since llc -march=ppc32 exp_attr.bc also uses up vast (maybe infinite) amounts of memory. However the C backend works ok. I'll attach the bytecode file in a second. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 12 19:06:58 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 12 Feb 2007 19:06:58 -0600 Subject: [LLVMbugs] [Bug 1197] bytecode causing llc to eat up vast amounts of memory Message-ID: <200702130106.l1D16wf5002112@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1197 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Common Code Generator Code OS/Version|Linux |All Product|new-bugs |libraries Platform|Other |All Resolution| |FIXED Target Milestone|--- |2.0 Version|unspecified |1.9 ------- Additional Comments From sabre at nondot.org 2007-02-12 19:06 ------- Fixed. Testcase here: test/CodeGen/Generic/switch-lower.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044404.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 12 20:22:12 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 12 Feb 2007 20:22:12 -0600 Subject: [LLVMbugs] [Bug 1161] build system ignores many user set directories Message-ID: <200702130222.l1D2MCQj004202@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1161 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX ------- Additional Comments From rspencer at x10sys.com 2007-02-12 20:22 ------- I now remember why we did this .. because autoconf gets it wrong. Using the autoconf names for these variables, we get the following expansions: PROJ_bindir := $(DESTDIR)/proj/llvm/install-3/progs PROJ_libdir := $(DESTDIR)${exec_prefix}/lib PROJ_datadir := $(DESTDIR)${datarootdir} PROJ_docsdir := $(DESTDIR)${datarootdir}/doc/${PACKAGE_TARNAME} PROJ_etcdir := $(DESTDIR)${sysconfigdir} PROJ_includedir := $(DESTDIR)${prefix}/include PROJ_infodir := $(DESTDIR)${datarootdir}/info PROJ_mandir := $(DESTDIR)${datarootdir}/man PROJ_datarootdir:= $(DESTDIR)${prefix}/share This is invalid syntax for makefiles. I don't know why autoconf can't resolve its variable names on its own. Consequently, I'm not going to "fix" this. If someone wants to patch autoconf to get this right then I might consider making the change. Perhaps when we upgrade our autoconf version it will just magically fix itself. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 12 20:26:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 12 Feb 2007 20:26:14 -0600 Subject: [LLVMbugs] [Bug 598] Need Support For Building Tools From Bytecode Message-ID: <200702130226.l1D2QEY8004325@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=598 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |LATER ------- Additional Comments From rspencer at x10sys.com 2007-02-12 20:26 ------- This used to be an issue when we were building the runtime library and Stacker as part of LLVM. However, we're not any more. Consequently, I'm deferring this until bug 731 gets implemented. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 13 09:15:15 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 13 Feb 2007 09:15:15 -0600 Subject: [LLVMbugs] [Bug 1198] NEW: assertion failure handling 64-bit umulh sequence Message-ID: <200702131515.l1DFFF17029524@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1198 Summary: assertion failure handling 64-bit umulh sequence 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: djg at cray.com This test case: define i64 @foo(i64 %x, i64 %y) { %tmp0 = zext i64 %x to i128 %tmp1 = zext i64 %y to i128 %tmp2 = mul i128 %tmp0, %tmp1 %tmp7 = zext i32 64 to i128 %tmp3 = lshr i128 %tmp2, %tmp7 %tmp4 = trunc i128 %tmp3 to i64 ret i64 %tmp4 } compiled with -march=x86-64 triggers the following assertion failure: lib/CodeGen/SelectionDAG/TargetLowering.cpp:2197: llvm::MVT::ValueType llvm::TargetLowering::getValueType(const llvm::Type*) const: Assertion `0 && "Invalid width for value type"' failed. The following patch fixes it. Index: TargetLowering.cpp =================================================================== RCS file: /var/cvs/llvm/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp,v retrieving revision 1.89 diff -u -r1.89 TargetLowering.cpp --- TargetLowering.cpp +++ TargetLowering.cpp @@ -2200,6 +2200,7 @@ case 16: return MVT::i16; case 32: return MVT::i32; case 64: return MVT::i64; + case 128: return MVT::i128; } break; case Type::FloatTyID: return MVT::f32; ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 13 11:39:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 13 Feb 2007 11:39:20 -0600 Subject: [LLVMbugs] [Bug 1199] NEW: nightly test fails many tests Message-ID: <200702131739.l1DHdKv8011047@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1199 Summary: nightly test fails many tests Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com On MultisSource/Applications/Burg, bugpoint gives: *** The following function is being miscompiled: main_bb_2E_i_2E_i12_2E_i_bb_2E_i_2E_i12_2E_i_2E_ce You can reproduce the problem with the command line: llc -f bugpoint.test.bc -o bugpoint.test.bc.s gcc ./bugpoint.safe.bc.cbe.c.so bugpoint.test.bc.s -o bugpoint.test.bc.exe -Wl,-R. bugpoint.test.bc.exe The shared object was created with: llc -march=c bugpoint.safe.bc -o temporary.c gcc -xc temporary.c -O2 -o ./bugpoint.safe.bc.cbe.c.so -shared -fno-strict-aliasing Attachments will be forthcoming. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 13 17:43:46 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 13 Feb 2007 17:43:46 -0600 Subject: [LLVMbugs] [Bug 1198] assertion failure handling 64-bit umulh sequence Message-ID: <200702132343.l1DNhkEt004133@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1198 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-13 17:43 ------- Very nice catch. I applied your patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044465.html and checked in your testcase here: test/CodeGen/X86/i128-mul.ll Note that 128-bit integer support in the LLVM level is desired, but not really ready yet. Things like 128- bit immediates are not handled right. Fortunately, the APInt stuff is great steps in the right direction, so we should have it for LLVM 2.0. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 14 01:13:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 14 Feb 2007 01:13:04 -0600 Subject: [LLVMbugs] [Bug 1200] NEW: [x86] unconditional branch not removed in codegen Message-ID: <200702140713.l1E7D4fK026417@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1200 Summary: [x86] unconditional branch not removed in codegen Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The attached LLVM code has an uncond branch that isn't removed by branch folding: _main_bb_2E_i9_2E_i_2E_i932_2E_ce: subl $28, %esp movl %esi, 24(%esp) movl %edi, 20(%esp) movl 36(%esp), %esi movl 32(%esp), %eax jmp LBB1_6 #bb.i9.i.i932.ce ... LBB1_6: #bb.i9.i.i932.ce movl (%eax), %edi movl L_outfile$non_lazy_ptr, %eax movl (%eax), %eax The entry block has a single successor, and LBB1_6 has a single predecessor, so branch folding should merge them. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 14 01:19:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 14 Feb 2007 01:19:28 -0600 Subject: [LLVMbugs] [Bug 1199] nightly test fails many tests Message-ID: <200702140719.l1E7JSEv026765@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1199 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-14 01:19 ------- Bug fixed for real: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044494.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 14 05:40:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 14 Feb 2007 05:40:00 -0600 Subject: [LLVMbugs] [Bug 790] Portability issues for Win32 Message-ID: <200702141140.l1EBe0S3017277@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=790 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From asl at math.spbu.ru 2007-02-14 05:39 ------- Well, I have to reopen this bug. Qt build in shared mode (on windows) is broken due to bad handling of "dllexport" attribute. I'm investigating. Also, it seems we can add PIC mode for mingw, see, e.g. http://gcc.gnu.org/ml/gcc-patches/2007-01/msg02262.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 14 19:23:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 14 Feb 2007 19:23:23 -0600 Subject: [LLVMbugs] [Bug 1194] bcwriter turns <2 x i10> into <2 x opaque> Message-ID: <200702150123.l1F1NNi8011080@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1194 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-14 19:23 ------- The fix for this was to allow the PackedType constructor to allow opaque types just like ArrayType does. The change to the slot calculator was reverted. Problem resolved. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 14 20:58:05 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 14 Feb 2007 20:58:05 -0600 Subject: [LLVMbugs] [Bug 1195] Rename PackedType as VectorType Message-ID: <200702150258.l1F2w5Kj012943@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1195 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-14 20:58 ------- This conversion is now done, with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044525.html through this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044547.html Reid. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 07:59:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 07:59:43 -0600 Subject: [LLVMbugs] [Bug 1201] NEW: Unused malloc/free don't get optimized Message-ID: <200702151359.l1FDxhr9001393@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1201 Summary: Unused malloc/free don't get optimized Product: libraries Version: 1.9 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Miscellaneous Instrumentation passes AssignedTo: unassignedbugs at nondot.org ReportedBy: nicola.lugato at gmail.com The optimizer doesn't remove malloc/free instructions when the allocated memory is not used. My example (but it does the same with simply a couple of malloc/free instructions) Before optimization: int %main(int %argc, ubyte** %argv) { %c_19 = alloca ubyte* %malloc_206 = malloc ubyte, uint 10 store ubyte* %malloc_206, ubyte** %c_19 %tmp_207 = load ubyte** %c_19 free ubyte* %tmp_207 ret int 0 } After optimization: int %main(int %argc, ubyte** %argv) { %malloc_206 = malloc [10 x ubyte] %malloc_206.sub = getelementptr [10 x ubyte]* %malloc_206, int 0, int 0 free ubyte* %malloc_206.sub ret int 0 } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 10:51:02 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 10:51:02 -0600 Subject: [LLVMbugs] [Bug 1202] NEW: PMTopLevelMgr stalls on destructor deadlock via memory corruption. Message-ID: <200702151651.l1FGp2db004668@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1202 Summary: PMTopLevelMgr stalls on destructor deadlock via memory corruption. 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: rspencer at x10sys.com My nightly test last night was stalled by the PassManager incurring an exception in operator delete that caused an abort that went into a deadlock. The test case is llvm/test/CFrontend/2004-02-13-IllegalVararg.c.tr which is: // RUN: %llvmgcc -xc %s -c -o - | llc #include float test(int X, ...) { va_list ap; float F; va_start(ap, X); F = va_arg(ap, float); return F; } During compilation of this, llvm-gcc warned about the following: /proj/llvm/nightly/build/llvm/test/CFrontend/2004-02-13-IllegalVararg.c.tr: In function 'test': /proj/llvm/nightly/build/llvm/test/CFrontend/2004-02-13-IllegalVararg.c.tr:9: warning: 'float' is promoted to 'double' when passed through '...' /proj/llvm/nightly/build/llvm/test/CFrontend/2004-02-13-IllegalVararg.c.tr:9: warning: (so you should pass 'double' not 'float' to 'va_arg') /proj/llvm/nightly/build/llvm/test/CFrontend/2004-02-13-IllegalVararg.c.tr:9: note: if this code is reached, the program will abort When llc tried to compile llvm-gcc's output, the program goes into a deadlock when PMTopLevelMgr's map is destructed. During destruction of one of the map's nodes the libc free() method generated this error: *** glibc detected *** llc: double free or corruption (!prev): 0x098e3938 *** The stack looks like this: #0 0x00bfd402 in __kernel_vsyscall () #1 0x003cecce in __lll_mutex_lock_wait () from /lib/libc.so.6 #2 0x003614e8 in _L_lock_43 () from /lib/libc.so.6 #3 0x0035b35d in ptmalloc_lock_all () from /lib/libc.so.6 #4 0x003849e0 in fork () from /lib/libc.so.6 #5 0x0059bad0 in fork () from /lib/libpthread.so.0 #6 0x0888b952 in (anonymous namespace)::PrintStackTrace () at Unix/Signals.inc:78 #7 0x0888bc3e in (anonymous namespace)::SignalHandler (Sig=6) at Unix/Signals.inc:139 #8 #9 0x00bfd402 in __kernel_vsyscall () #10 0x00321ee9 in raise () from /lib/libc.so.6 #11 0x003234f1 in abort () from /lib/libc.so.6 #12 0x0035653b in __libc_message () from /lib/libc.so.6 #13 0x0035da68 in _int_free () from /lib/libc.so.6 #14 0x00360f6f in free () from /lib/libc.so.6 #15 0x001bf3e1 in operator delete (ptr=0x98e3938) at ../../../../svn-4.0.3/libstdc++-v3/libsupc++/del_op.cc:49 #16 0x088276eb in __gnu_cxx::new_allocator > >::deallocate (this=0x98e2550, __p=0x98e3938) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/ext/new_allocator.h:94 #17 0x08827710 in std::_Rb_tree, std::_Select1st >, std::less, std::allocator > >::_M_put_node (this=0x98e2550, __p=0x98e3938) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_tree.h:360 #18 0x08827760 in std::_Rb_tree, std::_Select1st >, std::less, std::allocator > >::destroy_node (this=0x98e2550, __p=0x98e3938) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_tree.h:390 #19 0x088277a8 in std::_Rb_tree, std::_Select1st >, std::less, std::allocator > >::_M_erase (this=0x98e2550, __x=0x98e3938) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_tree.h:1097 #20 0x08827788 in std::_Rb_tree, std::_Select1st >, std::less, std::allocator > >::_M_erase (this=0x98e2550, __x=0x98e2120) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_tree.h:1095 #21 0x088277d6 in ~_Rb_tree (this=0x98e2550) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_tree.h:571 #22 0x088277f5 in ~map (this=0x98e2550) at /proj/install/lib/gcc/i686-pc-linux-gnu/4.0.3/../../../../include/c++/4.0.3/bits/stl_map.h:92 #23 0x0882158e in ~PMTopLevelManager (this=0x98e250c) at PassManager.cpp:498 #24 0x0882926e in ~FunctionPassManagerImpl (this=0x98e24b8) at PassManager.cpp:116 #25 0x0881f7ff in ~FunctionPassManager (this=0xbff8e224) at PassManager.cpp:867 #26 0x083b3083 in main (argc=1, argv=0xbff8e314) at llc.cpp:292 As this occurs on destruction, llc managed to produce this output anyway: .text .align 16 .globl test .type test, at function test: subl $24, %esp movl 28(%esp), %eax movl %eax, 20(%esp) leal 32(%esp), %eax movl %eax, 8(%esp) movss 0, %xmm0 movss %xmm0, 4(%esp) movss %xmm0, 12(%esp) movss %xmm0, 16(%esp) .LBB1_1: #return movss 16(%esp), %xmm0 movss %xmm0, (%esp) flds (%esp) addl $24, %esp ret .size test, .-test ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 12:18:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 12:18:56 -0600 Subject: [LLVMbugs] [Bug 1184] ScheduleDAG.cpp patch: error message more meaningful. Message-ID: <200702151818.l1FIIu1b014236@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1184 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-15 12:18 ------- I applied your patch. Please make sure not to use tabs though :) http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044578.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 12:39:52 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 12:39:52 -0600 Subject: [LLVMbugs] [Bug 1203] NEW: memory corruption in new BitVector class Message-ID: <200702151839.l1FIdqYd016013@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1203 Summary: memory corruption in new BitVector class Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Support Libraries AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com BitVector is causing llc to fail because of glibc detected memory corruption. Details to be attached shortly. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 12:58:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 12:58:59 -0600 Subject: [LLVMbugs] [Bug 1203] memory corruption in new BitVector class Message-ID: <200702151858.l1FIwx7g017732@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1203 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-15 12:58 ------- Inverting the comparison logic fixed it, with this patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044581.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 12:59:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 12:59:57 -0600 Subject: [LLVMbugs] [Bug 1202] llc deadlock via memory corruption from TargetData::setAlignment Message-ID: <200702151859.l1FIxv51017864@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1202 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From rspencer at x10sys.com 2007-02-15 12:59 ------- Bug 1203 has now been fixed and it resolves this one too. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 15:46:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 15:46:29 -0600 Subject: [LLVMbugs] [Bug 1204] NEW: LocalSpiller not handling virtual register map correctly Message-ID: <200702152146.l1FLkTwR030993@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1204 Summary: LocalSpiller not handling virtual register map correctly Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com When building llvm-gcc, while compiling libiberty.c, I get an assertion failed: 742 // Otherwise, reload it and remember that we have it. 743 PhysReg = VRM.getPhys(VirtReg); 744 assert(PhysReg && "Must map virtreg to physreg!"); In the failure case, VirtReg is 1041 and getPhys returns 0. This is likely more fallout from BitVector. The stack trace is: #0 0x0026b402 in __kernel_vsyscall () #1 0x00321ee9 in raise () from /lib/libc.so.6 #2 0x003234f1 in abort () from /lib/libc.so.6 #3 0x0031b859 in __assert_fail () from /lib/libc.so.6 #4 0x08726741 in (anonymous namespace)::LocalSpiller::RewriteMBB (this=0x9d39c58, MBB=@0x9dfced0, VRM=@0x9e91118) at VirtRegMap.cpp:744 #5 0x0872e848 in (anonymous namespace)::LocalSpiller::runOnMachineFunction (this=0x9d39c58, MF=@0x9e307a8, VRM=@0x9e91118) at VirtRegMap.cpp:231 #6 0x0871b953 in (anonymous namespace)::RA::runOnMachineFunction (this=0x99c80e0, fn=@0x9e307a8) at RegAllocLinearScan.cpp:209 #7 0x0856e426 in llvm::MachineFunctionPass::runOnFunction (this=0x99c80e0, F=@0x99ec680) at /proj/llvm/llvm-2/include/llvm/CodeGen/MachineFunctionPass.h:37 #8 0x088c57e8 in llvm::FPPassManager::runOnFunction (this=0x99c6e50, F=@0x99ec680) at PassManager.cpp:989 #9 0x088c5a3c in llvm::FunctionPassManagerImpl::run (this=0x99c6b88, F=@0x99ec680) at PassManager.cpp:944 #10 0x088c5afe in llvm::FunctionPassManager::run (this=0x99c51b8, F=@0x99ec680) at PassManager.cpp:889 #11 0x0833b9c8 in llvm_asm_file_end () at ../../src-2/gcc/llvm-backend.cpp:471 #12 0x0830c459 in toplev_main (argc=0, argv=0xbf862a94) at ../../src-2/gcc/toplev.c:1167 #13 0x0030f4e4 in __libc_start_main () from /lib/libc.so.6 #14 0x0804e171 in _start () ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 17:33:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 17:33:45 -0600 Subject: [LLVMbugs] [Bug 1204] LocalSpiller not handling virtual register map correctly Message-ID: <200702152333.l1FNXjDK005764@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1204 rspencer at x10sys.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From rspencer at x10sys.com 2007-02-15 17:33 ------- Either I had a local error or Evan's last patch to BitVector.h fixed it. The boostrap is okay now and llvm-gcc can correctly compile Olden, amongst other things. Dejagnu is 100% again. Closing this as INVALID. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 15 21:06:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 15 Feb 2007 21:06:57 -0600 Subject: [LLVMbugs] [Bug 1205] NEW: llvm-dis failed with unsigned 128-bit integer Message-ID: <200702160306.l1G36vMg016304@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1205 Summary: llvm-dis failed with unsigned 128-bit integer Product: new-bugs Version: unspecified Platform: PC OS/Version: Windows XP Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: wjiang at cs.ucla.edu The llvm-dis in TOT crashed with an example temp.c, which uses 128-bit unsigned integers. I used the following command: llvm-gcc -c -emit-llvm -O0 temp.c -o temp.bc opt -instcombine temp.bc -o combine.bc llvm-dis combine.bc -f ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 16 19:08:43 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 16 Feb 2007 19:08:43 -0600 Subject: [LLVMbugs] [Bug 1200] [x86] unconditional branch not removed in codegen Message-ID: <200702170108.l1H18hU8006914@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1200 dalej at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From dalej at apple.com 2007-02-16 19:08 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044660.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044659.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044661.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044662.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 16 19:20:29 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 16 Feb 2007 19:20:29 -0600 Subject: [LLVMbugs] [Bug 1185] [llvm-gcc] Remove EmittedGlobalVars and EmittedFunctions maps from Message-ID: <200702170120.l1H1KT1q007434@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1185 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-02-16 19:20 ------- Fixed with this: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044665.html -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 16 23:45:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 16 Feb 2007 23:45:50 -0600 Subject: [LLVMbugs] [Bug 1170] llvm-gcc crashes on variable sized struct argument Message-ID: <200702170545.l1H5joHF014953@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1170 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|llvm-gcc cycles on pretty |llvm-gcc crashes on variable |simple testcase |sized struct argument Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-16 23:45 ------- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070212/044674.html Testcase here: test/CFrontend/2007-02-16-VariableSizeStructArg.c -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 17 13:23:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 17 Feb 2007 13:23:28 -0600 Subject: [LLVMbugs] [Bug 1206] NEW: llvm::Module::getDataLayout() returns an incorrect layout. Message-ID: <200702171923.l1HJNS1i011663@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1206 Summary: llvm::Module::getDataLayout() returns an incorrect layout. Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: chuckbob17 at earthlink.net I'm using the LLVM 1.9 API. I have an LLVM IR file (in ASCII) that contains the following header: ; ModuleID = 'hello.c' target datalayout = "e-p:32:32" target endian = little target pointersize = 32 target triple = "i686-pc-linux-gnu" I then have some code that reads in the IR file and queries for the data layout: llvm::Module* mod = llvm::ParseAssemblyFile("file.ll"); std::string layout = llvm::Module::getDataLayout(); std::cout << layout; The value I get back from getDataLayout() is "e-p:32:32-e-p:32:32". I think it should return "e-p:32:32" instead. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 17 17:01:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 17 Feb 2007 17:01:57 -0600 Subject: [LLVMbugs] [Bug 1207] NEW: LiveVariables assertion prevents llvm-gcc build Message-ID: <200702172301.l1HN1ueI017780@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1207 Summary: LiveVariables assertion prevents llvm-gcc build Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: rspencer at x10sys.com After updating LLVM and applying recent patches to llvm-gcc (CUMMULATIVE-276.patch sent to llvm-commits today), the llvm-gcc build now fails while compiling libgcc, with: cc1: LiveInterval.cpp:227: void llvm::LiveInterval::removeRange(unsigned int, unsigned int): Assertion `I->contains(Start) && I->contains(End-1) && "Range is not entirely in interval!"' failed. Stack trace: #0 0x00826402 in __kernel_vsyscall () #1 0x00321ee9 in raise () from /lib/libc.so.6 #2 0x003234f1 in abort () from /lib/libc.so.6 #3 0x0031b859 in __assert_fail () from /lib/libc.so.6 #4 0x08735631 in llvm::LiveInterval::removeRange (this=0xaa2f1f0, Start=14, End=115) at LiveInterval.cpp:226 #5 0x0873cd73 in llvm::LiveInterval::removeRange (this=0xaa2f1f0, LR={start = 14, end = 115, ValId = 0}) at /proj/llvm/llvm-1/include/llvm/CodeGen/LiveInterval.h:253 #6 0x0873bbce in llvm::LiveIntervals::runOnMachineFunction (this=0xaa410e0, fn= #7 0x0856a4aa in llvm::MachineFunctionPass::runOnFunction (this=0xaa410e0, F= #8 0x088c13c8 in llvm::FPPassManager::runOnFunction (this=0xaa3fda8, F=@0xaa55978) at PassManager.cpp:998 #9 0x088c162a in llvm::FunctionPassManagerImpl::run (this=0xaa3fac8, F=@0xaa55978) at PassManager.cpp:953 #10 0x088c16ec in llvm::FunctionPassManager::run (this=0xaa3e0f8, F=@0xaa55978) at PassManager.cpp:895 #11 0x0833b8d8 in llvm_asm_file_end () at ../../src-1/gcc/llvm-backend.cpp:473 #12 0x0830bdc9 in toplev_main (argc=0, argv=0xbfed1fc4) at ../../src-1/gcc/toplev.c:1167 #13 0x0030f4e4 in __libc_start_main () from /lib/libc.so.6 #14 0x0804e171 in _start () ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 17 20:53:55 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 17 Feb 2007 20:53:55 -0600 Subject: [LLVMbugs] [Bug 1208] NEW: examples get install Message-ID: <200702180253.l1I2rtLB022890@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1208 Summary: examples get install Product: Build scripts Version: cvs Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Makefiles AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org 'make install' shouldn't install the examples (e.g. fibonacci), it didn't used to, but it does now with recent makefile changes. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 18 15:28:39 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 18 Feb 2007 15:28:39 -0600 Subject: [LLVMbugs] [Bug 1209] NEW: Eliminate std::vector from Type Message-ID: <200702182128.l1ILSdaV016967@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1209 Summary: Eliminate std::vector from Type Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Type::ContainedTys is a vector of PATypeHandle's. It would be much more efficient to coallocate the array on the end of the type object. This is a little tricky due to the fact that Type has subclasses. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 18 15:30:34 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 18 Feb 2007 15:30:34 -0600 Subject: [LLVMbugs] [Bug 1206] llvm::Module::getDataLayout() returns an incorrect layout. Message-ID: <200702182130.l1ILUYHZ017047@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1206 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Target Description Classes OS/Version|Linux |All Product|new-bugs |libraries Platform|PC |All Resolution| |FIXED Target Milestone|--- |2.0 Version|unspecified |1.9 ------- Additional Comments From sabre at nondot.org 2007-02-18 15:30 ------- This is fixed in CVS (2.0). ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 01:49:08 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 01:49:08 -0600 Subject: [LLVMbugs] [Bug 1210] NEW: Constant and Type uniquing should use FoldingSet Message-ID: <200702190749.l1J7n8FU028003@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1210 Summary: Constant and Type uniquing should use FoldingSet Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The current way we unique (complex) constants and types is incredibly inefficient. These should use FoldingSet instead, which will significantly speed up things like bcreading and general IR operations. Once this is done, we can migrate to an API that doesn't require a temporary vector for creating things like a StructType or a ConstantArray. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 14:19:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 14:19:45 -0600 Subject: [LLVMbugs] [Bug 1174] WITH_SIZE_EXPR not implemented Message-ID: <200702192019.l1JKJjc6018807@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1174 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Summary|Another assertion in EmitLV |WITH_SIZE_EXPR not | |implemented Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-19 14:19 ------- This works now with recent changes to pass variable sized stuff by reference. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 16:16:20 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 16:16:20 -0600 Subject: [LLVMbugs] [Bug 1211] NEW: "Register class of operand and regclass of use don't agree!" on x86 glibc atan Message-ID: <200702192216.l1JMGKFP021217@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1211 Summary: "Register class of operand and regclass of use don't agree!" on x86 glibc atan 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: baldrick at free.fr The following is simplified from the x86 atan2 code in glibc's mathinline.h: double atan (double x) { register double value; __asm ("fld1; fpatan" : "=t" (value) : "0" (x) : "st(1)"); return value; } A quite recent llvm-gcc4 barfs on it: $ gcc -c atan.c cc1: llvm/lib/CodeGen/SelectionDAG/ScheduleDAG.cpp:364: void llvm::ScheduleDAG::AddOperand(llvm::MachineInstr*, llvm::SDOperand, unsigned int, const llvm::TargetInstrDescriptor*, llvm::DenseMap&): Assertion `RegMap->getRegClass(VReg) == RC && "Register class of operand and regclass of use don't agree!"' failed. atan.c:6: internal compiler error: Aborted Please submit a full bug report, with preprocessed source if appropriate. See for instructions. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 16:18:56 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 16:18:56 -0600 Subject: [LLVMbugs] [Bug 1211] "Register class of operand and regclass of use don't agree!" on x86 glibc atan Message-ID: <200702192218.l1JMIuHu021275@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1211 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-02-19 16:18 ------- Yep, we don't support the X87 FP stack yet. There should be a -D option you can build with to avoid these asms (-DNO_MATH_INLINES ?) -Chris *** This bug has been marked as a duplicate of 879 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 16:23:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 16:23:50 -0600 Subject: [LLVMbugs] [Bug 1212] NEW: CBE doesn't strip {} delimiters from inline asm Message-ID: <200702192223.l1JMNo9j021402@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1212 Summary: CBE doesn't strip {} delimiters from inline asm Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Consider the following (essentially from x86 glibc): double atan (double x) { register double value; __asm__ ("fld1; fpatan" : "=t" (value) : "0" (x) : "st(1)"); return value; } $ gcc -c atan.c -emit-llvm $ llc -march=c atan.o $ gcc -c atan.o.cbe.c atan.o.cbe.c:107: warning: conflicting types for built-in function ???malloc??? atan.o.cbe.c: In function ???atan???: atan.o.cbe.c:147: error: invalid punctuation ???{??? in constraint atan.o.cbe.c:144: error: invalid punctuation ???{??? in constraint Indeed, there are extra curly brackets in the C output (they come from the LLVM assembly): __asm__ volatile ("fld1; fpatan" :"={st}"(ltmp_7_1) :"0"(ltmp_6_1) :"{st(1)}"); ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 19 16:27:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 19 Feb 2007 16:27:04 -0600 Subject: [LLVMbugs] [Bug 1212] CBE doesn't strip {} delimiters from inline asm Message-ID: <200702192227.l1JMR4dv021468@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1212 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-02-19 16:27 ------- Sorry to keep closing your bugs, but we know that the CBE doesn't really support inline asm. *** This bug has been marked as a duplicate of 802 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 20 01:59:44 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 20 Feb 2007 01:59:44 -0600 Subject: [LLVMbugs] [Bug 1213] NEW: llc -debug is broken Message-ID: <200702200759.l1K7xinl031805@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1213 Summary: llc -debug is broken Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached bytecode causes llc -debug to segfault. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 20 15:08:00 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 20 Feb 2007 15:08:00 -0600 Subject: [LLVMbugs] [Bug 1214] NEW: llvm-gcc4 barfs on simple packed bit-fields Message-ID: <200702202108.l1KL80n3022016@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1214 Summary: llvm-gcc4 barfs on simple packed bit-fields Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr Compiling the following code gives llvm-convert.cpp:4299: LValue TreeToLLVM::EmitLV_COMPONENT_REF(tree_node*): Assertion `NumAlignmentUnits && "Not adjusting pointer?"' failed. struct X { char A : 6; char B : 3; } __attribute__((packed)); int main (void) { struct X x; x.A = 1; x.B = 2; } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 20 18:48:42 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 20 Feb 2007 18:48:42 -0600 Subject: [LLVMbugs] [Bug 1189] llvm-convert.cpp - intrinsic getOrInsertFunction calls should be replaced by Intrinsic::getDeclaration Message-ID: <200702210048.l1L0mgKI025997@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1189 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-02-20 18:48 ------- Fixed http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044823.html -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 20 20:38:21 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 20 Feb 2007 20:38:21 -0600 Subject: [LLVMbugs] [Bug 775] Memory leak in ETNode::setFather Message-ID: <200702210238.l1L2cL9o028042@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=775 dpatel at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From dpatel at apple.com 2007-02-20 20:38 ------- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044826.html ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 21 00:24:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 21 Feb 2007 00:24:23 -0600 Subject: [LLVMbugs] [Bug 1208] examples get installed Message-ID: <200702210624.l1L6ONUk031724@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1208 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED OS/Version|MacOS X |All Platform|Macintosh |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-21 00:24 ------- Fixed by reverting r1.68 of llvm/Makefile: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044830.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 21 05:31:11 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 21 Feb 2007 05:31:11 -0600 Subject: [LLVMbugs] [Bug 1197] switch lowering infinitely loops, chews memory. Message-ID: <200702211131.l1LBVBP2013586@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1197 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | ------- Additional Comments From asl at math.spbu.ru 2007-02-21 05:31 ------- switch lowering is still broken somehow. Consider the folowing code: define i32 @main(i32 %argc, i8** %argv) { old_entry1: br label %cond_next cond_next: %not.tmp14 = icmp slt i32 %argc, 11 %hvar2 = zext i1 %not.tmp14 to i32 switch i32 %hvar2, label %old_entry [ i32 -2, label %cond_true i32 -1, label %cond_next i32 0, label %cond_true15 i32 1, label %cond_next22 ] cond_true15: ret i32 1 cond_true: ret i32 2 cond_next22: ret i32 3 old_entry: ret i32 4 } (Yes, I know, that cases -1 and -2 are really dead). Resulting binary will always produce result code 4. Commenting out "-1" and "-2" cases will fix the problem. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 21 16:44:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 21 Feb 2007 16:44:32 -0600 Subject: [LLVMbugs] [Bug 1207] LiveVariables assertion prevents llvm-gcc build Message-ID: <200702212244.l1LMiWbo025359@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1207 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-02-21 16:44 ------- Fixed, thanks, Evan. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 21 22:28:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 21 Feb 2007 22:28:28 -0600 Subject: [LLVMbugs] [Bug 1215] NEW: can't constantfold sdiv 0, -1 Message-ID: <200702220428.l1M4SSHs032024@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1215 Summary: can't constantfold sdiv 0, -1 Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca I encountered an assertion failure in scalar evolutions: opt: ScalarEvolution.cpp:2362: llvm::SCEVHandle llvm::SCEVAddRecExpr::getNumIterationsInRange(llvm::ConstantRange, bool) const: Assertion `isa(ExitValue) && "Constant folding of integers not implemented?"' failed. The assertion has found a real problem. The line in question is "ConstantExpr::getSDiv(ExitValue, A);" where ExitValue is i32 0 and A is i32 -1. getSDiv eventually gets into this case in ConstantFoldBinaryInstruction (ConstantFolding.cpp): 00572 case Instruction::SDiv: 00573 if (CI2->isNullValue()) return 0; // X / 0 -> can't fold 00574 if (CI2->isAllOnesValue() && 00575 (((CI1->getType()->getPrimitiveSizeInBits() == 64) && 00576 (CI1->getSExtValue() == INT64_MIN)) || 00577 (CI1->getSExtValue() == -CI1->getSExtValue()))) 00578 return 0; // MIN_INT / -1 -> overflow 00579 return ConstantInt::get(C1->getType(), 00580 CI1->getSExtValue() / CI2->getSExtValue()); Given that CI1 is the left hand side, CI1->getSExtValue() == -CI1->getSExtValue() when the left is 0. This code also fails to properly fold arbitrary precision integers, so it deserves to be rewritten. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 22 00:58:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 22 Feb 2007 00:58:49 -0600 Subject: [LLVMbugs] [Bug 1216] NEW: JITResolver shouldn't outlive its ExecutionEngine Message-ID: <200702220658.l1M6wnLK002107@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1216 Summary: JITResolver shouldn't outlive its ExecutionEngine Product: libraries Version: trunk Platform: PC URL: http://xzrq.net/reee/reee.cpp OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Target-Independent JIT AssignedTo: unassignedbugs at nondot.org ReportedBy: ppergame at gmail.com JIT uses the instance of JITResolver stored in a static variable in getJITResolver(). This means that JITResolver::ExternalFnToStubMap is not cleared between consecutive, independent JIT invocations. When the ExecutionEngine is deleted, the stub goes away via JITMemoryManager, but the reference remains. The next ExecutionEngine tries to reuse the invalid stubs and segfaults. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Thu Feb 22 06:45:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 22 Feb 2007 06:45:13 -0600 Subject: [LLVMbugs] [Bug 1217] NEW: infinite loop in instcombine Message-ID: <200702221245.l1MCjDdA017524@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1217 Summary: infinite loop in instcombine Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Core LLVM classes AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca This file causes instcombine to take a very long time. The stacktrace shows FoldOpIntoPhi and visitAdd. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 14:22:35 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 14:22:35 -0600 Subject: [LLVMbugs] [Bug 1218] NEW: Patch for strict alias warning in Support/Allocator.cpp Message-ID: <200702232022.l1NKMZ6V004960@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1218 Summary: Patch for strict alias warning in Support/Allocator.cpp Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: minor Priority: P2 Component: Support Libraries AssignedTo: unassignedbugs at nondot.org ReportedBy: me22.ca at gmail.com I only saw 3 warnings compiling llvm, 2 of which looked like flex's fault, so it made me want to fix the other one: Allocator.cpp: In member function 'void* llvm::BumpPtrAllocator::Allocate(unsigned int, unsigned int)': Allocator.cpp:96: warning: dereferencing type-punned pointer will break strict-aliasing rules I'm not convinced it actually broke aliasing rules, but it's a fairly trivial fix that simplifies the code, so it seems worth it. Only possible drawback is that it (opaquely) exposes MemRegion, but that identifier is used nowhere else in llvm at the moment. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 17:12:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 17:12:17 -0600 Subject: [LLVMbugs] [Bug 1164] CBE can generate name conflicts Message-ID: <200702232312.l1NNCHNm017605@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1164 isanbard at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From isanbard at gmail.com 2007-02-23 17:12 ------- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044959.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044960.html -bw ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 17:13:49 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 17:13:49 -0600 Subject: [LLVMbugs] [Bug 1218] Patch for strict alias warning in Support/Allocator.cpp Message-ID: <200702232313.l1NNDnrn017642@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1218 me22.ca at gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From me22.ca at gmail.com 2007-02-23 17:13 ------- Yes it does. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 19:04:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 19:04:17 -0600 Subject: [LLVMbugs] [Bug 1217] infinite loop in instcombine Message-ID: <200702240104.l1O14Hk3019560@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1217 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|Core LLVM classes |Scalar Optimizations Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-23 19:04 ------- Fixed, testcase here: InstCombine/2007-02-23-PhiFoldInfLoop.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044967.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 19:17:28 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 19:17:28 -0600 Subject: [LLVMbugs] [Bug 1215] can't constantfold sdiv 0, -1 Message-ID: <200702240117.l1O1HSbF019865@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1215 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |compile-fail Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-23 19:17 ------- Fixed, testcase here: Transforms/ConstProp/2007-02-23-sdiv.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044968.html -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 19:28:50 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 19:28:50 -0600 Subject: [LLVMbugs] [Bug 1219] NEW: switch lowering miscompiles a switch stmt Message-ID: <200702240128.l1O1So6p020296@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1219 Summary: switch lowering miscompiles a switch stmt Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org >From Anton: switch lowering is still broken somehow. Consider the folowing code: define i32 @main(i32 %argc, i8** %argv) { old_entry1: br label %cond_next cond_next: %not.tmp14 = icmp slt i32 %argc, 11 %hvar2 = zext i1 %not.tmp14 to i32 switch i32 %hvar2, label %old_entry [ i32 -2, label %cond_true i32 -1, label %cond_next i32 0, label %cond_true15 i32 1, label %cond_next22 ] cond_true15: ret i32 1 cond_true: ret i32 2 cond_next22: ret i32 3 old_entry: ret i32 4 } (Yes, I know, that cases -1 and -2 are really dead). Resulting binary will always produce result code 4. Commenting out "-1" and "-2" cases will fix the problem. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 19:29:18 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 19:29:18 -0600 Subject: [LLVMbugs] [Bug 1197] switch lowering infinitely loops, chews memory. Message-ID: <200702240129.l1O1TIrv020340@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1197 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-23 19:29 ------- Moving new bug to a new bugzilla: Bug 1219. The original issue is fixed. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 20:10:13 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 20:10:13 -0600 Subject: [LLVMbugs] [Bug 1219] dag combine miscompilation of setlt Message-ID: <200702240210.l1O2ADJ5021709@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1219 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |miscompilation OS/Version|MacOS X |All Platform|Macintosh |All Resolution| |FIXED Summary|switch lowering miscompiles |dag combine miscompilation |a switch stmt |of setlt Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-23 20:10 ------- Testcase here: CodeGen/Generic/2007-02-23-DAGCombine-Miscompile.ll Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044972.html Thanks for tracking this down Anton! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 20:58:09 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 20:58:09 -0600 Subject: [LLVMbugs] [Bug 1216] JITResolver shouldn't outlive its ExecutionEngine Message-ID: <200702240258.l1O2w9ks022886@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1216 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |missing-feature OS/Version|Linux |All Platform|PC |All Resolution| |FIXED Target Milestone|--- |2.0 ------- Additional Comments From sabre at nondot.org 2007-02-23 20:58 ------- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044973.html Thanks! -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Fri Feb 23 21:40:32 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 23 Feb 2007 21:40:32 -0600 Subject: [LLVMbugs] [Bug 1220] NEW: projects/sample/AutoRegen.sh patches configure.ac incorrectly Message-ID: <200702240340.l1O3eWMk024560@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1220 Summary: projects/sample/AutoRegen.sh patches configure.ac incorrectly 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: me22.ca at gmail.com AutoRegen.sh patches configure.ac with a path relative to the autoconf directory, not the project's root directory. Relevant discussion at http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-February/007995.html and follow-ups. As Reid specified autoconf 2.60 ( http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-February/008021.html ), the patch also makes the script accept autoconf 2.[56][0-9] instead of the old 2.5[0-9], and displays the actual version used instead of 2.5x. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 24 14:06:14 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 24 Feb 2007 14:06:14 -0600 Subject: [LLVMbugs] [Bug 1221] NEW: Read from object already freed. Message-ID: <200702242006.l1OK6EOq017964@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1221 Summary: Read from object already freed. Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Consider the attached .i file. Valgrind output: ==16630== Invalid read of size 1 ==16630== at 0x828BC96: default_strip_name_encoding (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== Address 0x47D6F1C is 12 bytes inside a block of size 21 free'd ==16630== at 0x402167A: operator delete(void*) (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x40F48A0: std::string::_Rep::_M_destroy(std::allocator const&) (in /usr/lib/gcc/i686-pc-linux-gnu/3.4.6/libstdc++.so.6.0.3) ==16630== by 0x829655E: llvm_get_decl_name (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== ==16630== Invalid read of size 1 ==16630== at 0x4022593: strlen (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x88114AA: xstrdup (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== Address 0x47D6F1C is 12 bytes inside a block of size 21 free'd ==16630== at 0x402167A: operator delete(void*) (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x40F48A0: std::string::_Rep::_M_destroy(std::allocator const&) (in /usr/lib/gcc/i686-pc-linux-gnu/3.4.6/libstdc++.so.6.0.3) ==16630== by 0x829655E: llvm_get_decl_name (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== ==16630== Invalid read of size 1 ==16630== at 0x402259B: strlen (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x88114AA: xstrdup (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== Address 0x47D6F1D is 13 bytes inside a block of size 21 free'd ==16630== at 0x402167A: operator delete(void*) (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x40F48A0: std::string::_Rep::_M_destroy(std::allocator const&) (in /usr/lib/gcc/i686-pc-linux-gnu/3.4.6/libstdc++.so.6.0.3) ==16630== by 0x829655E: llvm_get_decl_name (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== ==16630== Invalid read of size 1 ==16630== at 0x40229B7: memcpy (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x88114C7: xstrdup (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== Address 0x47D6F24 is 20 bytes inside a block of size 21 free'd ==16630== at 0x402167A: operator delete(void*) (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==16630== by 0x40F48A0: std::string::_Rep::_M_destroy(std::allocator const&) (in /usr/lib/gcc/i686-pc-linux-gnu/3.4.6/libstdc++.so.6.0.3) ==16630== by 0x829655E: llvm_get_decl_name (in /home/asl/proj/llvm/gccbuild/gcc/cc1) ==16630== Steps to reproduce: valgrind cc1 libgcc2.i ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 24 19:26:04 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 24 Feb 2007 19:26:04 -0600 Subject: [LLVMbugs] [Bug 1221] llvm-gcc reads from memory object already freed. Message-ID: <200702250126.l1P1Q4Ur024482@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1221 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From sabre at nondot.org 2007-02-24 19:26 ------- Perfect, that was just what I needed. Here's a fix: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070219/044990.html Please let me know if there is any other valgrind related issues, -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 24 19:46:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 24 Feb 2007 19:46:57 -0600 Subject: [LLVMbugs] [Bug 1222] NEW: X86 backend doesn't support MMX Message-ID: <200702250146.l1P1kvHr025101@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1222 Summary: X86 backend doesn't support MMX Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org The X86 backend currently has no support for MMX registers. This can be seen when using MMX intrinsic or when using MMX inline asm. Here's an example that uses inline asm: void foo() { const int Aligned64data_last_mult = 4; asm { movd mm1, [-1][eax][esi] movd mm0, [-1][eax] movd mm2, [-1][eax][esi*2] movq [edi][Aligned64data_last_mult], mm0 lea eax, [esi][ecx*2] movd mm4, [esi] movd mm5, [esi][ecx] movd [-4][edi], mm3 } } llvm-gcc -fasm-blocks -msse3 should successfully compile this, but doesn't due to lack of support for mmx registers. -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sat Feb 24 22:16:33 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 24 Feb 2007 22:16:33 -0600 Subject: [LLVMbugs] [Bug 1223] NEW: verifier produces bytecode that causes a bytecode reader assertion failure Message-ID: <200702250416.l1P4GXLV028073@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1223 Summary: verifier produces bytecode that causes a bytecode reader assertion failure 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: lefever at crhc.uiuc.edu The verifier (called by the -verify option to opt) permits bytecode to be written that causes the bytecode reader to abort due to the following assertion failure: ----- llvm::BytecodeReader::ParseConstantPoolValue(unsigned int): Assertion `(!isa(Result) || !cast(Result)->isNullValue()) || !hasImplicitNull(TypeID) && "Cannot read null values from bytecode!"' failed. ----- The following information was obtained by running gdb on opt when reading in faulty bytecode permitted by the verifier. ----- (gdb) print Result $7 = (class llvm::Constant *) 0x96ad9d0 (gdb) call Result->dump() [41 x sbyte *] zeroinitializer (gdb) print hasImplicitNull(TypeID) $5 = true ----- After further investigation, the bytecode (that passed the verifier but not the bytecode reader) was produced in the following way. We start with a module M that contains GlobalVariables X and Y. In the module's original form, X is initialized using Y. We then apply a transform T to M that retains X but replaces all uses of Y with a NULL value by executing the following code: Y->replaceAllUsesWith(Constant::getNullValue(Y->getType())); T then deletes Y by executing the following code: M.getGlobalList().erase(Y); T then finishes and the bytecode it produces, passes the verifier. When opt is called on the just written bytecode, the bytecode reader fails. Assuming that the bytecode reader is correct, the verifier needs to be updated to warn/abort when such bytecode is produced. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 08:28:59 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 08:28:59 -0600 Subject: [LLVMbugs] [Bug 1224] NEW: llvm-gcc build is broken Message-ID: <200702251428.l1PESxsV009024@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1224 Summary: llvm-gcc build is broken Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: blocker Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Even after PR1221 was fixed, I'm getting assertions during _divsc3.o compilation. Unfortunately, valgrind doesn't show anything wrong and the problem isn't reproducible with llc. Even more, if I compile _divsc3.o with llc by hands and continue llvm-gcc build, I'll get the folowing assertion during guard.cc compilation (from libsupc++): cc1plus: /home/asl/proj/llvm/src/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp:4100: llvm::SDOperand llvm::SelectionDAGLowering::CopyValueToVirtualRegister(llvm::Value*, unsigned int): Assertion `(Op.getOpcode() != ISD::CopyFromReg || cast(Op.getOperand(1))->getReg() != Reg) && "Copy from a reg to the same reg!"' failed. And again no valgrind output... I'm investigating ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 09:26:57 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 09:26:57 -0600 Subject: [LLVMbugs] [Bug 1213] llc -debug is broken Message-ID: <200702251526.l1PFQvwR010209@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1213 asl at math.spbu.ru changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From asl at math.spbu.ru 2007-02-25 09:26 ------- This seems to be fixed somehow. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 11:54:17 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 11:54:17 -0600 Subject: [LLVMbugs] [Bug 1225] NEW: installed llvm-config tries to enter build directory Message-ID: <200702251754.l1PHsHw6012346@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1225 Summary: installed llvm-config tries to enter build directory 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: me22.ca at gmail.com This seems like another bug having to do with tools assuming that llvm is left in its build directory and never installed. ( Like http://llvm.org/bugs/show_bug.cgi?id=1106 ) Even after install, llvm-config references the build directory, and not the prefix. For example: $ llvm-config --src-root sh: line 0: cd: /var/tmp/portage/dev-lang/llvm-9999-r20070224/work/llvm/: No such file or directory This seems like it prevents llvm-gcc from building: config.status: executing default commands sh: line 0: cd: /var/tmp/portage/dev-lang/llvm-9999-r20070224/work/llvm/: No such file or directory llvm-config: unknown component name: transforms echo '/*WARNING: This file is automatically generated!*/' >tmp-c-parse.y TARGET_CPU_DEFAULT="" \ HEADERS="auto-host.h ansidecl.h" DEFINES="" \ /bin/sh ../llvm-gcc4-1.9.source/gcc/mkconfig.sh config.h sed -e "/^@@ifobjc.*/,/^@@end_ifobjc.*/d" \ -e "/^@@ifc.*/d" -e "/^@@end_ifc.*/d" ../llvm-gcc4-1.9.source/gcc/c-parse.in >>tmp-c-parse.y /bin/sh ../llvm-gcc4-1.9.source/gcc/../move-if-change tmp-c-parse.y c-parse.y # APPLE LOCAL mainline rewrite touch c-parse.y TARGET_CPU_DEFAULT="" \ HEADERS="config/i386/i386.h config/i386/unix.h config/i386/att.h config/dbxelf.h config/elfos.h config/svr4.h config/linux.h config/i386/linux.h defaults.h" DEFINES="" \ /bin/sh ../llvm-gcc4-1.9.source/gcc/mkconfig.sh tm.h TARGET_CPU_DEFAULT="" \ HEADERS="auto-host.h ansidecl.h" DEFINES="" \ /bin/sh ../llvm-gcc4-1.9.source/gcc/mkconfig.sh bconfig.h make: *** No rule to make target `../build-i686-pc-linux-gnu/libiberty/libiberty.a', needed by `build/genmodes'. Stop. make: *** Waiting for unfinished jobs.... This will probably also cause problems for resolving http://llvm.org/bugs/show_bug.cgi?id=1106 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 14:35:26 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 14:35:26 -0600 Subject: [LLVMbugs] [Bug 1226] NEW: scalarrepl should be able to scalarrepl aggregates with memcpy uses Message-ID: <200702252035.l1PKZQK2015424@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1226 Summary: scalarrepl should be able to scalarrepl aggregates with memcpy uses Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org Consider: #include #include void assign( long* variable, long v) { std::transform( variable, variable + 1, variable, std::tr1::bind( std::plus< long >(), 0L, v ) ); } This compiles to a single store on x86, but a whole ton of code on x86-64. This is because the temporary structs are larger on x86-64, so EmitAggregateCopy in llvm-gcc emits them as a memcpy instead of scalar transfers. The problem is that this later blocks scalarrepl from promoting the structs, causing much worse codegen: __Z6assignRll: # x86-32 movl 8(%esp), %eax movl 4(%esp), %ecx movl %eax, (%ecx) ret __Z6assignRll: # x86-64 subq $88, %rsp movb $0, 64(%rsp) movq $0, 72(%rsp) movq %rsi, 80(%rsp) movq %rsi, 48(%rsp) movq 72(%rsp), %rax movq %rax, 40(%rsp) movq 64(%rsp), %rax movq %rax, 32(%rsp) movq 40(%rsp), %rax movq %rax, 8(%rsp) movq 48(%rsp), %rax movq %rax, 16(%rsp) movq 32(%rsp), %rax movq %rax, (%rsp) movq 16(%rsp), %rax addq 8(%rsp), %rax movq %rax, (%rdi) addq $88, %rsp ret -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 14:45:45 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 14:45:45 -0600 Subject: [LLVMbugs] [Bug 1227] NEW: Can't compute trip count of simple loop Message-ID: <200702252045.l1PKjjkF015656@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1227 Summary: Can't compute trip count of simple loop Product: libraries Version: 1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Global Analyses AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: dpatel at apple.com This simple loop should be getting unrolled: target datalayout = "e-p:32:32" target triple = "i686-apple-darwin8" implementation ; Functions: define void @_Z6assignPll(i32* %variable, i32 %v) { entry: br label %bb63 bb63: ; preds = %bb63, %entry %__result.0203.0.rec = phi i32 [ 0, %entry ], [ %indvar.next, %bb63 ] ; [#uses=3] %__result.0203.0 = getelementptr i32* %variable, i32 %__result.0203.0.rec ; [#uses=1] store i32 %v, i32* %__result.0203.0 icmp eq i32 %__result.0203.0.rec, 2 ; :0 [#uses=1] %indvar.next = add i32 %__result.0203.0.rec, 1 ; [#uses=1] br i1 %0, label %bb193, label %bb63 bb193: ; preds = %bb63 ret void } However, it doesn't currently because the icmp is using the PHI node, not the increment. Because of this, this trivial loop doesn't get unrolled. This comes from this simple example: #include #include void assign( long* variable, long v ) { std::transform( variable, variable + 3, variable, std::tr1::bind( std::plus< long >(), 0L, v ) ); } -Chris ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 15:47:23 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 15:47:23 -0600 Subject: [LLVMbugs] [Bug 1224] Invoke lowering broken with -enable-eh Message-ID: <200702252147.l1PLlNnI016756@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1224 jlaskey at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From jlaskey at apple.com 2007-02-25 15:47 ------- Index: lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp =============================================================== ==== RCS file: /var/cvs/llvm/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp,v retrieving revision 1.374 diff -u -1 -0 -d -r1.374 SelectionDAGISel.cpp --- lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp 24 Feb 2007 09:45:44 -0000 1.374 +++ lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp 25 Feb 2007 21:15:16 -0000 @@ -493,20 +493,21 @@ void visitBr(BranchInst &I); void visitSwitch(SwitchInst &I); void visitUnreachable(UnreachableInst &I) { /* noop */ } // Helper for visitSwitch void visitSwitchCase(SelectionDAGISel::CaseBlock &CB); void visitJumpTable(SelectionDAGISel::JumpTable &JT); // These all get lowered before this pass. void visitInvoke(InvokeInst &I); + void visitInvoke(InvokeInst &I, bool AsTerminator); void visitUnwind(UnwindInst &I); void visitScalarBinary(User &I, unsigned OpCode); void visitVectorBinary(User &I, unsigned OpCode); void visitEitherBinary(User &I, unsigned ScalarOp, unsigned VectorOp); void visitShift(User &I, unsigned Opcode); void visitAdd(User &I) { if (isa(I.getType())) visitVectorBinary(I, ISD::VADD); else if (I.getType()->isFloatingPoint()) @@ -1098,58 +1099,63 @@ // Emit the code for the jump table MVT::ValueType PTy = TLI.getPointerTy(); SDOperand Index = DAG.getCopyFromReg(getRoot(), JT.Reg, PTy); SDOperand Table = DAG.getJumpTable(JT.JTI, PTy); DAG.setRoot(DAG.getNode(ISD::BR_JT, MVT::Other, Index.getValue(1), Table, Index)); return; } void SelectionDAGLowering::visitInvoke(InvokeInst &I) { + assert(0 && "Should never be visited directly"); +} +void SelectionDAGLowering::visitInvoke(InvokeInst &I, bool AsTerminator) { // Retrieve successors. MachineBasicBlock *Return = FuncInfo.MBBMap[I.getSuccessor(0)]; MachineBasicBlock *LandingPad = FuncInfo.MBBMap[I.getSuccessor(1)]; - // Mark landing pad so that it doesn't get deleted in branch folding. - LandingPad->setIsLandingPad(); - - // Insert a label before the invoke call to mark the try range. - // This can be used to detect deletion of the invoke via the - // MachineModuleInfo. - MachineModuleInfo *MMI = DAG.getMachineModuleInfo(); - unsigned BeginLabel = MMI->NextLabelID(); - DAG.setRoot(DAG.getNode(ISD::LABEL, MVT::Other, getRoot(), - DAG.getConstant(BeginLabel, MVT::i32))); - - LowerCallTo(I, I.getCalledValue()->getType(), - I.getCallingConv(), - false, - getValue(I.getOperand(0)), - 3); + if (!AsTerminator) { + // Mark landing pad so that it doesn't get deleted in branch folding. + LandingPad->setIsLandingPad(); + + // Insert a label before the invoke call to mark the try range. + // This can be used to detect deletion of the invoke via the + // MachineModuleInfo. + MachineModuleInfo *MMI = DAG.getMachineModuleInfo(); + unsigned BeginLabel = MMI->NextLabelID(); + DAG.setRoot(DAG.getNode(ISD::LABEL, MVT::Other, getRoot(), + DAG.getConstant(BeginLabel, MVT::i32))); - // Insert a label before the invoke call to mark the try range. - // This can be used to detect deletion of the invoke via the - // MachineModuleInfo. - unsigned EndLabel = MMI->NextLabelID(); - DAG.setRoot(DAG.getNode(ISD::LABEL, MVT::Other, getRoot(), - DAG.getConstant(EndLabel, MVT::i32))); - - // Inform MachineModuleInfo of range. - MMI->addInvoke(LandingPad, BeginLabel, EndLabel); + LowerCallTo(I, I.getCalledValue()->getType(), + I.getCallingConv(), + false, + getValue(I.getOperand(0)), + 3); - // Drop into normal successor. - DAG.setRoot(DAG.getNode(ISD::BR, MVT::Other, getRoot(), - DAG.getBasicBlock(Return))); - - // Update successor info - CurMBB->addSuccessor(Return); - CurMBB->addSuccessor(LandingPad); + // Insert a label before the invoke call to mark the try range. + // This can be used to detect deletion of the invoke via the + // MachineModuleInfo. + unsigned EndLabel = MMI->NextLabelID(); + DAG.setRoot(DAG.getNode(ISD::LABEL, MVT::Other, getRoot(), + DAG.getConstant(EndLabel, MVT::i32))); + + // Inform MachineModuleInfo of range. + MMI->addInvoke(LandingPad, BeginLabel, EndLabel); + + // Update successor info + CurMBB->addSuccessor(Return); + CurMBB->addSuccessor(LandingPad); + } else { + // Drop into normal successor. + DAG.setRoot(DAG.getNode(ISD::BR, MVT::Other, getRoot(), + DAG.getBasicBlock(Return))); + } } void SelectionDAGLowering::visitUnwind(UnwindInst &I) { } void SelectionDAGLowering::visitSwitch(SwitchInst &I) { // Figure out which block is immediately after the current one. MachineBasicBlock *NextBlock = 0; MachineFunction::iterator BBI = CurMBB; @@ -4212,20 +4218,24 @@ if (LLVMBB == &LLVMBB->getParent()->front()) LowerArguments(LLVMBB, SDL, UnorderedChains); BB = FuncInfo.MBBMap[LLVMBB]; SDL.setCurrentBasicBlock(BB); // Lower all of the non-terminator instructions. for (BasicBlock::iterator I = LLVMBB->begin(), E = --LLVMBB->end(); I != E; ++I) SDL.visit(*I); + + // Lower call part of invoke. + InvokeInst *Invoke = dyn_cast(LLVMBB->getTerminator()); + if (Invoke) SDL.visitInvoke(*Invoke, false); // Ensure that all instructions which are used outside of their defining // blocks are available as virtual registers. for (BasicBlock::iterator I = LLVMBB->begin(), E = LLVMBB->end(); I != E;++I) if (!I->use_empty() && !isa(I)) { DenseMap::iterator VMI =FuncInfo.ValueMap.find(I); if (VMI != FuncInfo.ValueMap.end()) UnorderedChains.push_back( SDL.CopyValueToVirtualRegister(I, VMI->second)); } @@ -4324,21 +4334,26 @@ } if (i == e) UnorderedChains.push_back(Root); } DAG.setRoot(DAG.getNode(ISD::TokenFactor, MVT::Other, &UnorderedChains[0], UnorderedChains.size())); } // Lower the terminator after the copies are emitted. - SDL.visit(*LLVMBB->getTerminator()); + if (Invoke) { + // Just the branch part of invoke. + SDL.visitInvoke(*Invoke, true); + } else { + SDL.visit(*LLVMBB->getTerminator()); + } // Copy over any CaseBlock records that may now exist due to SwitchInst // lowering, as well as any jump table information. SwitchCases.clear(); SwitchCases = SDL.SwitchCases; JT = SDL.JT; // Make sure the root of the DAG is up-to-date. DAG.setRoot(SDL.getRoot()); } ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Sun Feb 25 19:00:03 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 25 Feb 2007 19:00:03 -0600 Subject: [LLVMbugs] [Bug 1228] NEW: Segfault with EH's enabled Message-ID: <200702260100.l1Q103s9021024@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1228 Summary: Segfault with EH's enabled Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: asl at math.spbu.ru Attached bytecode (probably) cycles and consumes all available memory. Reproduce with: ulimit -v 512000 llc -enable-eh locale.bc ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Mon Feb 26 21:22:06 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 26 Feb 2007 21:22:06 -0600 Subject: [LLVMbugs] [Bug 1229] NEW: convert the LLVM repository to subversion Message-ID: <200702270322.l1R3M6ac030205@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1229 Summary: convert the LLVM repository to subversion Product: new-bugs Version: unspecified Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CVS doesn't have good semantics when renaming files. Switch to a source-code repository that does, such as subversion. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From isanbard at gmail.com Tue Feb 27 00:56:21 2007 From: isanbard at gmail.com (Bill Wendling) Date: Mon, 26 Feb 2007 22:56:21 -0800 Subject: [LLVMbugs] Problems with TableGen Message-ID: <2FDCF0C7-AC27-46B3-BDC5-988E8357D128@gmail.com> I'm getting this error when doing a compile: Included from X86.td:130: Parsing X86CallingConv.td:20: Couldn't find class 'CallingConv'! -bw From sabre at nondot.org Tue Feb 27 01:31:07 2007 From: sabre at nondot.org (Chris Lattner) Date: Mon, 26 Feb 2007 23:31:07 -0800 (PST) Subject: [LLVMbugs] Problems with TableGen In-Reply-To: <2FDCF0C7-AC27-46B3-BDC5-988E8357D128@gmail.com> References: <2FDCF0C7-AC27-46B3-BDC5-988E8357D128@gmail.com> Message-ID: On Mon, 26 Feb 2007, Bill Wendling wrote: > I'm getting this error when doing a compile: > Included from X86.td:130: > Parsing X86CallingConv.td:20: Couldn't find class 'CallingConv'! CVS up, it should work now. -Chris -- http://nondot.org/sabre/ http://llvm.org/ From bugzilla-daemon at cs.uiuc.edu Tue Feb 27 12:55:22 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 27 Feb 2007 12:55:22 -0600 Subject: [LLVMbugs] [Bug 1230] NEW: llvm-gcc generates machine dependent LLVM bytecode Message-ID: <200702271855.l1RItMAG029217@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1230 Summary: llvm-gcc generates machine dependent LLVM bytecode Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com *** C code *** struct ss { int x; long long y; }; void f(int a, struct ss b); *** bytecode for i386-linux-gnu *** %struct.ss = type { i32, i64 } declare void @f(i32, i64, i32) *** bytecode for arm-linux-gnueabi *** %struct.ss = type { i32, i64 } declare void @f(i32, i64, i64) This occurs because struct ss size is 12 bytes on i386-linux-gnu and 16 bytes on arm-linux-gnueabi. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Tue Feb 27 13:36:10 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 27 Feb 2007 13:36:10 -0600 Subject: [LLVMbugs] [Bug 1231] NEW: bugpoint help doesn't Message-ID: <200702271936.l1RJaA7n030271@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1231 Summary: bugpoint help doesn't 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: dalej at apple.com The --help option for bugpoint produces USAGE: bugpoint [options] --args ... --tool-args ... followed by a lengthy list of options, most of which, I think, don't apply to bugpoint. This is completely wrong, and produces mysterious errors when you attempt to do things this way, for example: bugpoint -run-llc --tool-args="-o lencod.llc.s -stats -march=thumb -debug-only=arm-cp-islands - mcpu=arm1136jf-s -relocation-model=pic -f" lencod.llvm.bc bugpoint: Not enough positional command line arguments specified! Not real helpful, the reverse in fact. I looked briefly at fixing this but it appears command line processing is centralized for the various tools in a way that makes it nontrivial. Also, the -o and -f options cannot be passed to llc, as bugpoint inserts these itself and llc will not tolerate duplication. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From isanbard at gmail.com Tue Feb 27 15:36:07 2007 From: isanbard at gmail.com (Bill Wendling) Date: Tue, 27 Feb 2007 13:36:07 -0800 Subject: [LLVMbugs] Problems in Build Message-ID: <16e5fdf90702271336s3d135d20p4fefd9186ee92ac8@mail.gmail.com> Here's an error I saw with current TOT: g++ -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/utils/TableGen -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/include -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/include -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/include -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -D_DEBUG -DLLVM_VERSION_INFO='" Apple Build #1033-01"' -Woverloaded-virtual -Wall -W -Wwrite-strings -Wunused -Wno-unused-parameter -fstrict-aliasing -arch ppc -arch i386 -arch ppc64 -arch x86_64 -c /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp -o /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/utils/TableGen/Release/TableGen.o /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:23:32: error: CallingConvEmitter.h: No such file or directory /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:23:32: error: CallingConvEmitter.h: No such file or directory /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:23:32: error: CallingConvEmitter.h: No such file or directory /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:23:32: error: CallingConvEmitter.h: No such file or directory /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp: In function 'int main(int, char**)': /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:146: error: 'CallingConvEmitter' was not declared in this scope /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp: In function 'int main(int, char**)': /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:146: error: 'CallingConvEmitter' was not declared in this scope/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp: In function 'int main(int, char**)': /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:146: error: 'CallingConvEmitter' was not declared in this scope /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp: In function 'int main(int, char**)': /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/utils/TableGen/TableGen.cpp:146: error: 'CallingConvEmitter' was not declared in this scope lipo: can't figure out the architecture type of: /var/tmp//ccmc7QJ1.out make[3]: *** [/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/utils/TableGen/Release/TableGen.o] Error 1 make[2]: *** [TableGen/.makeall] Error 2 make[1]: *** [all] Error 1 -bw From isanbard at gmail.com Tue Feb 27 15:53:38 2007 From: isanbard at gmail.com (Bill Wendling) Date: Tue, 27 Feb 2007 13:53:38 -0800 Subject: [LLVMbugs] Problems in Build In-Reply-To: <7E62243F-7E76-4D4B-AC5F-8DDF3B433EA8@apple.com> References: <16e5fdf90702271336s3d135d20p4fefd9186ee92ac8@mail.gmail.com> <7E62243F-7E76-4D4B-AC5F-8DDF3B433EA8@apple.com> Message-ID: <16e5fdf90702271353x6860b319g7a37456aaa7dfd1c@mail.gmail.com> Thanks! :-) -bw On 2/27/07, Evan Cheng wrote: > I've temporarily backed out the TableGen changes. Should build now. > > Evan > > On Feb 27, 2007, at 1:36 PM, Bill Wendling wrote: > > > Here's an error I saw with current TOT: > > > > g++ -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/ > > obj-llvm/utils/TableGen > > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > > llvm/utils/TableGen > > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > > llvm/include > > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > > llvm/include > > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > > llvm/include > > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > > llvm/include > > -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -D_DEBUG > > -DLLVM_VERSION_INFO='" Apple Build #1033-01"' -Woverloaded-virtual > > -Wall -W -Wwrite-strings -Wunused -Wno-unused-parameter > > -fstrict-aliasing -arch ppc -arch i386 -arch ppc64 -arch x86_64 -c > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp > > -o /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > > llvm/utils/TableGen/Release/TableGen.o > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:23:32: > > error: CallingConvEmitter.h: No such file or directory > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:23:32: > > error: CallingConvEmitter.h: No such file or directory > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:23:32: > > error: CallingConvEmitter.h: No such file or directory > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:23:32: > > error: CallingConvEmitter.h: No such file or directory > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp: > > In function 'int main(int, char**)': > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:146: > > error: 'CallingConvEmitter' was not declared in this scope > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp: > > In function 'int main(int, char**)': > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:146: > > error: 'CallingConvEmitter' was not declared in this > > scope/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > > llvm/utils/TableGen/TableGen.cpp: > > In function 'int main(int, char**)': > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:146: > > error: 'CallingConvEmitter' was not declared in this scope > > > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp: > > In function 'int main(int, char**)': > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > > utils/TableGen/TableGen.cpp:146: > > error: 'CallingConvEmitter' was not declared in this scope > > lipo: can't figure out the architecture type of: /var/tmp// > > ccmc7QJ1.out > > make[3]: *** [/Volumes/SandBox/Clean/llvm-1033-01.roots/ > > llvm-1033-01~obj/obj-llvm/utils/TableGen/Release/TableGen.o] > > Error 1 > > make[2]: *** [TableGen/.makeall] Error 2 > > make[1]: *** [all] Error 1 > > > > > > -bw > > _______________________________________________ > > LLVMbugs mailing list > > LLVMbugs at cs.uiuc.edu > > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs > > From evan.cheng at apple.com Tue Feb 27 15:48:53 2007 From: evan.cheng at apple.com (Evan Cheng) Date: Tue, 27 Feb 2007 13:48:53 -0800 Subject: [LLVMbugs] Problems in Build In-Reply-To: <16e5fdf90702271336s3d135d20p4fefd9186ee92ac8@mail.gmail.com> References: <16e5fdf90702271336s3d135d20p4fefd9186ee92ac8@mail.gmail.com> Message-ID: <7E62243F-7E76-4D4B-AC5F-8DDF3B433EA8@apple.com> I've temporarily backed out the TableGen changes. Should build now. Evan On Feb 27, 2007, at 1:36 PM, Bill Wendling wrote: > Here's an error I saw with current TOT: > > g++ -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/ > obj-llvm/utils/TableGen > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > llvm/utils/TableGen > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > llvm/include > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > llvm/include > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > llvm/include > -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > llvm/include > -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -D_DEBUG > -DLLVM_VERSION_INFO='" Apple Build #1033-01"' -Woverloaded-virtual > -Wall -W -Wwrite-strings -Wunused -Wno-unused-parameter > -fstrict-aliasing -arch ppc -arch i386 -arch ppc64 -arch x86_64 -c > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp > -o /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- > llvm/utils/TableGen/Release/TableGen.o > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:23:32: > error: CallingConvEmitter.h: No such file or directory > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:23:32: > error: CallingConvEmitter.h: No such file or directory > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:23:32: > error: CallingConvEmitter.h: No such file or directory > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:23:32: > error: CallingConvEmitter.h: No such file or directory > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp: > In function 'int main(int, char**)': > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:146: > error: 'CallingConvEmitter' was not declared in this scope > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp: > In function 'int main(int, char**)': > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:146: > error: 'CallingConvEmitter' was not declared in this > scope/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ > llvm/utils/TableGen/TableGen.cpp: > In function 'int main(int, char**)': > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:146: > error: 'CallingConvEmitter' was not declared in this scope > > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp: > In function 'int main(int, char**)': > /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ > utils/TableGen/TableGen.cpp:146: > error: 'CallingConvEmitter' was not declared in this scope > lipo: can't figure out the architecture type of: /var/tmp// > ccmc7QJ1.out > make[3]: *** [/Volumes/SandBox/Clean/llvm-1033-01.roots/ > llvm-1033-01~obj/obj-llvm/utils/TableGen/Release/TableGen.o] > Error 1 > make[2]: *** [TableGen/.makeall] Error 2 > make[1]: *** [all] Error 1 > > > -bw > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs From sabre at nondot.org Tue Feb 27 16:38:02 2007 From: sabre at nondot.org (Chris Lattner) Date: Tue, 27 Feb 2007 14:38:02 -0800 (PST) Subject: [LLVMbugs] Problems in Build In-Reply-To: <16e5fdf90702271353x6860b319g7a37456aaa7dfd1c@mail.gmail.com> References: <16e5fdf90702271336s3d135d20p4fefd9186ee92ac8@mail.gmail.com> <7E62243F-7E76-4D4B-AC5F-8DDF3B433EA8@apple.com> <16e5fdf90702271353x6860b319g7a37456aaa7dfd1c@mail.gmail.com> Message-ID: On Tue, 27 Feb 2007, Bill Wendling wrote: > Thanks! :-) This is my fault, I'm working on it, sorry :( -Chris > On 2/27/07, Evan Cheng wrote: >> I've temporarily backed out the TableGen changes. Should build now. >> >> Evan >> >> On Feb 27, 2007, at 1:36 PM, Bill Wendling wrote: >> >>> Here's an error I saw with current TOT: >>> >>> g++ -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/ >>> obj-llvm/utils/TableGen >>> -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ >>> llvm/utils/TableGen >>> -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- >>> llvm/include >>> -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ >>> llvm/include >>> -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- >>> llvm/include >>> -I/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ >>> llvm/include >>> -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -O2 -D_DEBUG >>> -DLLVM_VERSION_INFO='" Apple Build #1033-01"' -Woverloaded-virtual >>> -Wall -W -Wwrite-strings -Wunused -Wno-unused-parameter >>> -fstrict-aliasing -arch ppc -arch i386 -arch ppc64 -arch x86_64 -c >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp >>> -o /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj- >>> llvm/utils/TableGen/Release/TableGen.o >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:23:32: >>> error: CallingConvEmitter.h: No such file or directory >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:23:32: >>> error: CallingConvEmitter.h: No such file or directory >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:23:32: >>> error: CallingConvEmitter.h: No such file or directory >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:23:32: >>> error: CallingConvEmitter.h: No such file or directory >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp: >>> In function 'int main(int, char**)': >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:146: >>> error: 'CallingConvEmitter' was not declared in this scope >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp: >>> In function 'int main(int, char**)': >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:146: >>> error: 'CallingConvEmitter' was not declared in this >>> scope/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/ >>> llvm/utils/TableGen/TableGen.cpp: >>> In function 'int main(int, char**)': >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:146: >>> error: 'CallingConvEmitter' was not declared in this scope >>> >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp: >>> In function 'int main(int, char**)': >>> /Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/src/llvm/ >>> utils/TableGen/TableGen.cpp:146: >>> error: 'CallingConvEmitter' was not declared in this scope >>> lipo: can't figure out the architecture type of: /var/tmp// >>> ccmc7QJ1.out >>> make[3]: *** [/Volumes/SandBox/Clean/llvm-1033-01.roots/ >>> llvm-1033-01~obj/obj-llvm/utils/TableGen/Release/TableGen.o] >>> Error 1 >>> make[2]: *** [TableGen/.makeall] Error 2 >>> make[1]: *** [all] Error 1 >>> >>> >>> -bw >>> _______________________________________________ >>> LLVMbugs mailing list >>> LLVMbugs at cs.uiuc.edu >>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs >> >> > _______________________________________________ > LLVMbugs mailing list > LLVMbugs at cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs > -Chris -- http://nondot.org/sabre/ http://llvm.org/ From isanbard at gmail.com Tue Feb 27 16:53:58 2007 From: isanbard at gmail.com (Bill Wendling) Date: Tue, 27 Feb 2007 14:53:58 -0800 Subject: [LLVMbugs] Another Boog Message-ID: <16e5fdf90702271453l773a767fu566c24a93a22ac45@mail.gmail.com> Here's another failure. Undefined symbols for architecture ppc64: llvm::RegScavenger::init(), referenced from: (anonymous namespace)::PEI::replaceFrameIndices(llvm::MachineFunction&) in libLLVMCodeGen.a(PrologEpilogInserter.o) ld64-69.4 failed: symbol(s) not found for architecture ppc64 collect2: ld returned 1 exit status lipo: can't open input file: /var/tmp//ccaLkrMR.out (No such file or directory) make[3]: *** [/Volumes/SandBox/Clean/llvm-1033-01.roots/llvm-1033-01~obj/obj-llvm/Release/bin/llc] Error 1 make[2]: *** [llc/.makeall] Error 2 make[1]: *** [all] Error 1 -bw From bugzilla-daemon at cs.uiuc.edu Wed Feb 28 19:46:01 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 28 Feb 2007 19:46:01 -0600 Subject: [LLVMbugs] [Bug 1232] NEW: constant propagation doesn't work through unrolled loops Message-ID: <200703010146.l211k1uY015445@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1232 Summary: constant propagation doesn't work through unrolled loops 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: loufoque at gmail.com With the following C code, the loop is completely eliminated, thanks to loop unrolling and constant propagation. The resulting LLVM code just calls printf with 20. #include #include int main() { int v = 0; size_t i; for(i=0; i<5; i++) { v += i*2; } printf("%d\n", v); } However, if 5 is replaced by 100, loop unrolling doesn't occur, and for some reason constant propagation doesn't work anymore. The resulting code just needs to call printf with 9900, there is no need for any loop. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. From bugzilla-daemon at cs.uiuc.edu Wed Feb 28 19:48:25 2007 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 28 Feb 2007 19:48:25 -0600 Subject: [LLVMbugs] [Bug 1232] constant propagation doesn't work through unrolled loops Message-ID: <200703010148.l211mP08015540@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1232 sabre at nondot.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE ------- Additional Comments From sabre at nondot.org 2007-02-28 19:48 ------- Yep, known. *** This bug has been marked as a duplicate of 1179 *** ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.