From bugzilla-daemon at cs.uiuc.edu Tue Jan 1 02:03:16 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 Jan 2008 02:03:16 -0600 Subject: [LLVMbugs] [Bug 1889] New: clang allows allocating array that is too large Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1889 Summary: clang allows allocating array that is too large Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: AST AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu void c(int n){ #define ARR_SIZE 0x7FFFFFFF int b[ARR_SIZE]; int c[sizeof(b)-sizeof(int)*(long long)ARR_SIZE]; } For this code, there should either be an error on the declaration of b (if size_t isn't large enough to contain the size of b) or no error (if size_t is large enough to contain the size of b). Currently, there is an error claiming that c has a negative width, which should be impossible since "sizeof(b)" should be equal to "sizeof(int)*(long long)ARR_SIZE". Apparently, the computation of sizeof(b) overflows. For this code, gcc says "error: size of array ???b??? is too large". clang should say something similar. Not sure if this is the right component. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 1 02:13:04 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 1 Jan 2008 02:13:04 -0600 Subject: [LLVMbugs] [Bug 1890] New: clang allows illegal redeclaration of function? Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1890 Summary: clang allows illegal redeclaration of function? Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: AST AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu int b(int c) {return 1;} int f(float b) { return 0; } void a(void){ int f(int (int)); f(b); } I think this code should give a compile error. Both gcc and the comeau c compiler give an error on this code, but clang does not; instead, clang emits code with an rather dangerous function cast. Not sure if this is the right component. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 01:39:02 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 01:39:02 -0600 Subject: [LLVMbugs] [Bug 1891] New: Addres of unnamed array not recognized as constant Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1891 Summary: Addres of unnamed array not recognized as constant Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu The following file scope definition: int *p = (int []){2, 4}; gives this error: inittest.c:5:18: error: initializer element is not constant int *p = (int []){2, 4}; ^~~~~~ using clang -fsyntax-only. However, it is in fact correct per C99. Note that this example is straight out of the standard. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 02:31:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 02:31:35 -0600 Subject: [LLVMbugs] [Bug 1892] New: Redeclaration error with compatible declarations Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1892 Summary: Redeclaration error with compatible declarations Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: AST AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu void f(double (* restrict a)[5]); void f(double a[restrict][5]); should not give any error, but clang complains about the redefinition of f. (This testcase is from the C99 standard.) >From the AST dump, it looks like clang is actually losing the restrict qualifier in the second definition of f somehow. (The second declaration gets dumped as "void f(double (*)[5]);", whereas the first is dumped as "void f(double (*restrict)[5]);".) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 03:51:17 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 03:51:17 -0600 Subject: [LLVMbugs] [Bug 1893] New: Incomplete array "int i[]; " should implicitly have one element Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1893 Summary: Incomplete array "int i[];" should implicitly have one element Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu The declaration: int i[]; currently results in "@i = global [0 x i32] zeroinitializer" It should result in an array with one element, per C99 section 6.9.2. (It should probably also warn, like gcc does.) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 04:24:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 04:24:35 -0600 Subject: [LLVMbugs] [Bug 1895] New: Crash taking size of function Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1895 Summary: Crash taking size of function Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu int zxcv(void); int x=sizeof(zxcv); crashes with Assertion `0 && "Incomplete types have no size!"' failed. /home/eli/llvmbin/Release/bin/clang((anonymous namespace)::PrintStackTrace()+0x19)[0x82aea19] using clang -emit-llvm. It should fail more gracefully. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 04:11:13 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 04:11:13 -0600 Subject: [LLVMbugs] [Bug 1894] New: Landing pad label misplaced Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1894 Summary: Landing pad label misplaced Product: libraries Version: trunk Platform: Macintosh OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: nicolas.geoffray at lip6.fr CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1313) --> (http://llvm.org/bugs/attachment.cgi?id=1313) Bug test case This bug only happens when generating ppc code (by default it uses the list-td scheduler. The other schedulers don't have this bug) The landing pad label is placed after a few instructions that belong to a landing pad basic block. I use the attached test case, which is not generated by llvm-gcc. Just run: ./Release/bin/llvm-as -f lp.ll ./Release/bin/llc -enable-eh -f -march=ppc32 lp.bc For the unwind block, the code generated is: BB1_2: # unwind lis 3, printf at ha lis 4, .str at ha creqv 6, 6, 6 la 5, printf at l(3) la 6, .str at l(4) label3: li 4, 5 li 30, 2 ... It should be: BB1_2: # unwind label3: lis 3, printf at ha lis 4, .str at ha creqv 6, 6, 6 la 5, printf at l(3) la 6, .str at l(4) li 4, 5 li 30, 2 ... -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 15:32:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 15:32:35 -0600 Subject: [LLVMbugs] [Bug 1888] Compilier errors in Building LLVM using Windows MSVC Projects In-Reply-To: Message-ID: <200801022132.m02LWZfj030966@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1888 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-02 15:32:34 --- I applied your patch, thanks! For FLT_ROUNDS, please rename the FLT_ROUNDS enum in SelectionDAGNodes.h to something like FP_GETROUNDINGMODE, which should fix the conflict. Thanks! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 15:54:36 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 15:54:36 -0600 Subject: [LLVMbugs] [Bug 1895] Crash taking size of function In-Reply-To: Message-ID: <200801022154.m02Lsa6N032072@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1895 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|enhancement |normal Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-02 15:54:36 --- This is actually a horrible GCC extension. Fixed. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 16:51:30 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 16:51:30 -0600 Subject: [LLVMbugs] [Bug 1892] clang function merging not fully implemented In-Reply-To: Message-ID: <200801022251.m02MpU9u003242@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1892 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Chris Lattner 2008-01-02 16:51:29 --- Fixed, thanks! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20071231/003617.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 19:21:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 19:21:03 -0600 Subject: [LLVMbugs] [Bug 1882] llvm-ar: Segmentation fault when archiving bitcode library In-Reply-To: Message-ID: <200801030121.m031L3KK011172@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1882 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #1 from Chris Lattner 2008-01-02 19:21:03 --- I think this is a dup of 1881, please verify. *** This bug has been marked as a duplicate of bug 1881 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 19:21:32 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 19:21:32 -0600 Subject: [LLVMbugs] [Bug 1881] llvm-ar: Segmentation fault when archiving bitcode library In-Reply-To: Message-ID: <200801030121.m031LWuF011231@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1881 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Chris Lattner 2008-01-02 19:21:32 --- Fixed, thanks! http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056747.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 2 23:31:47 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 2 Jan 2008 23:31:47 -0600 Subject: [LLVMbugs] [Bug 1862] verify fails based on ordering even with -disable-opt In-Reply-To: Message-ID: <200801030531.m035Vlhk023817@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1862 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #12 from Nick Lewycky 2008-01-02 23:31:46 --- llvm-ld depending on the order of its inputs is not necessarily a bug. Given three files, a.ll: %struct.foo = type { i32 } declare void @f1(%struct.foo* %x) b.ll: %struct.foo = type { i16 } declare void @f2(%struct.foo* %x) c.ll: %struct.foo = type opaque declare void @f3(%struct.foo* %x) the result will differ depending on which of a.ll or b.ll comes first on the command line. In your case, you're getting different structs and trying to index into one with GEP. That works with one definition of %struct.inode, but not the other. The real question, then, is how did you get two different types when compiling the kernel? -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 3 01:07:52 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 Jan 2008 01:07:52 -0600 Subject: [LLVMbugs] [Bug 1876] llvm-ar: corrupted double-linked list In-Reply-To: Message-ID: <200801030707.m0377qNl028823@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1876 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #3 from Chris Lattner 2008-01-03 01:07:52 --- Please verify. *** This bug has been marked as a duplicate of bug 1881 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 3 01:11:49 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 Jan 2008 01:11:49 -0600 Subject: [LLVMbugs] [Bug 1873] ERROR: Program used external function 'stat' which could not be resolved! In-Reply-To: Message-ID: <200801030711.m037BnR4029109@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1873 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #7 from Chris Lattner 2008-01-03 01:11:48 --- Patch looks great to me, applied: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056752.html Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 3 07:37:56 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 Jan 2008 07:37:56 -0600 Subject: [LLVMbugs] [Bug 1896] New: Global Variable Optimizer fails assertion in OptimizeAwayTrappingUsesOfLoads Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1896 Summary: Global Variable Optimizer fails assertion in OptimizeAwayTrappingUsesOfLoads Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: torvald at se.inf.tu-dresden.de CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1316) --> (http://llvm.org/bugs/attachment.cgi?id=1316) input .bc file globalopt fails with an assertion, it cannot handle (or didn't detect) a bitcast from one pointer type to another (see below). llvm-ld -disable-opt works well, and created the input .bc (see attachment and backtrace/ instruction dump below). LLVM 2.1 llvm-ld -native -lpthread -o foo calls-combined.bc -debug-pass=Executions Pass Arguments: -internalize -ipsccp -globalopt -constmerge -deadargelim -instcombine -basiccg -inline -prune-eh -globalopt -globaldce -basiccg -argpromotion -instcombine -domtree -domfrontier -scalarrepl -globalsmodref-aa -domtree -loops -loopsimplify -domfrontier -scalar-evolution -licm -memdep -gvn -dse -instcombine -simplifycfg -globaldce -instcombine -simplifycfg -dce -globaldce -domtree -verify [snip] 0x855ab48 Executing Pass 'Global Variable Optimizer' on Module 'foo'... llvm-ld: /home/torvald/STM/code/cpp/llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:639: bool OptimizeAwayTrappingUsesOfLoads(llvm::GlobalVariable*, llvm::Constant*): Assertion `(isa(*GUI) || isa(*GUI) || isa(*GUI)) && "Only expect load and stores!"' failed. llvm-ld[0x84a96c6] llvm-ld[0x84a998c] [0xffffe420] /lib/i686/cmov/libc.so.6(abort+0x101)[0xb7d26831] /lib/i686/cmov/libc.so.6(__assert_fail+0xee)[0xb7d1e08e] llvm-ld[0x8350a91] llvm-ld[0x8350db7] llvm-ld[0x8351de0] llvm-ld[0x8351fed] llvm-ld[0x83520ae] llvm-ld(llvm::MPPassManager::runOnModule(llvm::Module&)+0x100)[0x843f410] llvm-ld(llvm::PassManagerImpl::run(llvm::Module&)+0x6e)[0x843f5c8] llvm-ld(llvm::PassManager::run(llvm::Module&)+0x1a)[0x843f61a] llvm-ld(llvm::Optimize(llvm::Module*)+0x3bf)[0x8273723] llvm-ld(main+0x689)[0x827b52d] /lib/i686/cmov/libc.so.6(__libc_start_main+0xe0)[0xb7d10450] llvm-ld(realloc+0x79)[0x8272ae1] Aborted (gdb) bt #0 0xffffe410 in __kernel_vsyscall () #1 0xb7caaeb5 in raise () from /lib/i686/cmov/libc.so.6 #2 0xb7cac831 in abort () from /lib/i686/cmov/libc.so.6 #3 0xb7ca408e in __assert_fail () from /lib/i686/cmov/libc.so.6 #4 0x08350a91 in OptimizeAwayTrappingUsesOfLoads (GV=0x8543d20, LV=0x8544d18) at /llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:637 #5 0x08350db7 in OptimizeOnceStoredGlobal (GV=0x8543d20, StoredOnceVal=0x8544d18, GVI=@0xbf859364, TD=@0x8557548) at /llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:1193 #6 0x08351de0 in ProcessInternalGlobal (this=0x85448c0, GV=0x8543d20, GVI=@0xbf859364) at /llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:1455 #7 0x08351fed in OptimizeGlobalVars (this=0x85448c0, M=@0x8543cb8) at /llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:1536 #8 0x083520ae in runOnModule (this=0x85448c0, M=@0x8543cb8) at /llvm-2.1/lib/Transforms/IPO/GlobalOpt.cpp:2078 (gdb) call (*GUI)->dump() i32 * bitcast (void (i32)** @indirect1 to i32*) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 3 23:05:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 3 Jan 2008 23:05:43 -0600 Subject: [LLVMbugs] [Bug 1896] Global Variable Optimizer fails assertion in OptimizeAwayTrappingUsesOfLoads In-Reply-To: Message-ID: <200801040505.m0455hRu012688@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1896 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-03 23:05:42 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056772.html Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 04:42:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 04:42:55 -0600 Subject: [LLVMbugs] [Bug 1897] New: clang driver crashes if no arguments are given. Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1897 Summary: clang driver crashes if no arguments are given. Product: clang Version: unspecified Platform: All OS/Version: All Status: NEW Severity: minor Priority: P2 Component: Basic AssignedTo: unassignedbugs at nondot.org ReportedBy: willi.t1 at gmail.com CC: llvmbugs at cs.uiuc.edu Basically, it continues on looking for an extension even if string::find('.') returns string::npos. The patch should fix it (just return langkind_c immediately when '.' cannt be found). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 08:17:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 08:17:01 -0600 Subject: [LLVMbugs] [Bug 1386] crash on 64 bit bitfield that doesn't start on a byte boundary In-Reply-To: Message-ID: <200801041417.m04EH1Fg012903@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1386 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Duncan Sands 2008-01-04 08:17:00 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056792.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056788.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 13:13:39 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 13:13:39 -0600 Subject: [LLVMbugs] [Bug 1897] clang driver crashes if no arguments are given. In-Reply-To: Message-ID: <200801041913.m04JDd2u027026@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1897 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Chris Lattner 2008-01-04 13:13:38 --- Applied, thanks! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20071231/003647.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 16:11:46 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 16:11:46 -0600 Subject: [LLVMbugs] [Bug 1898] New: Runtime failure from optimization Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1898 Summary: Runtime failure from optimization Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: major Priority: P2 Component: Backend: X86 AssignedTo: unassignedbugs at nondot.org ReportedBy: dalej at apple.com CC: llvmbugs at cs.uiuc.edu Attached .ll file works when run directly through llc (on Darwin x86). Fails when run through opt -std-compile-opts first. This is a regression from 12/13/07. This is from gcc.c-torture/execute/20000403-1.c in the gcc testsuite (with llvm-gcc-4.2). There are 3 similar failures (runtime failure at optimization other than -O0); I'm hopeful they're all the same: gcc.c-torture/execute/920711-1.c gcc.c-torture/execute/cmpsi-1.c gcc.c-torture/execute/loop-3b.c -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 19:20:15 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 19:20:15 -0600 Subject: [LLVMbugs] [Bug 1898] instcombine miscompile In-Reply-To: Message-ID: <200801050120.m051KFmN014020@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1898 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Chris Lattner 2008-01-04 19:20:15 --- Fix here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056817.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 4 21:46:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 4 Jan 2008 21:46:43 -0600 Subject: [LLVMbugs] [Bug 1899] New: Assertion `!Subtarget->isTargetDarwin()' failed on x86_64 target Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1899 Summary: Assertion `!Subtarget->isTargetDarwin()' failed on x86_64 target Product: tools Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: major Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: laguest at archangeli.co.uk CC: llvmbugs at cs.uiuc.edu Configured on the following Ubuntu Feisty machine: $ uname -a Linux rogue 2.6.20-16-generic #2 SMP Sun Sep 23 18:31:23 UTC 2007 x86_64 GNU/Linux $ gcc -v Using built-in specs. Target: x86_64-linux-gnu Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --program-suffix=-4.1 --enable-__cxa_atexit --enable-clocale=gnu --enable-libstdcxx-debug --enable-mpfr --enable-checking=release x86_64-linux-gnu Thread model: posix gcc version 4.1.2 (Ubuntu 4.1.2-0ubuntu4) Configured and installed LLVM (debug and release) fine. Configured llvm-gcc-4.2 with the following: ../../llvm-gcc-4.2/configure --prefix=/home/laguest/opt/gnat-llvm --enable-llvm=/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm --enable-checking --enable-threads=posix --disable-bootstrap --disable-shared --disable-multilib --target=x86_64-pc-linux --host=x86_64-pc-linux --enable-languages=c,ada,c++ and then when making get this: SHLIB_INSTALL='$(mkinstalldirs) $(DESTDIR)$(slibdir)@shlib_slibdir_qual@; /usr/bin/install -c -m 644 @multilib_dir@/@shlib_base_name at .so.1 $(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so.1; rm -f $(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so; ln -s @shlib_base_name at .so.1 $(DESTDIR)$(slibdir)@shlib_slibdir_qual@/@shlib_base_name at .so' \ SHLIB_EXT='.so' \ SHLIB_MULTILIB='' \ SHLIB_MKMAP='../../../llvm-gcc-4.2/gcc/mkmap-symver.awk' \ SHLIB_MKMAP_OPTS='' \ SHLIB_MAPFILES='../../../llvm-gcc-4.2/gcc/libgcc-std.ver ../../../llvm-gcc-4.2/gcc/config/i386/libgcc-x86_64-glibc.ver' \ SHLIB_NM_FLAGS='-pg' \ MULTILIB_OSDIRNAMES='../lib64 ../lib' \ ASM_HIDDEN_OP='' \ GCC_FOR_TARGET='/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2/./gcc/xgcc -B/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2/./gcc/ -B/home/laguest/opt/gnat-llvm/x86_64-pc-linux/bin/ -B/home/laguest/opt/gnat-llvm/x86_64-pc-linux/lib/ -isystem /home/laguest/opt/gnat-llvm/x86_64-pc-linux/include -isystem /home/laguest/opt/gnat-llvm/x86_64-pc-linux/sys-include' \ mkinstalldirs='/bin/sh ../../../llvm-gcc-4.2/gcc/../mkinstalldirs' \ /bin/sh mklibgcc > tmp-libgcc.mk mv tmp-libgcc.mk libgcc.mk TARGET_CPU_DEFAULT="" \ HEADERS="auto-host.h ansidecl.h" DEFINES="USED_FOR_TARGET " \ /bin/sh ../../../llvm-gcc-4.2/gcc/mkconfig.sh tconfig.h /home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2/./gcc/xgcc -B/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2/./gcc/ -B/home/laguest/opt/gnat-llvm/x86_64-pc-linux/bin/ -B/home/laguest/opt/gnat-llvm/x86_64-pc-linux/lib/ -isystem /home/laguest/opt/gnat-llvm/x86_64-pc-linux/include -isystem /home/laguest/opt/gnat-llvm/x86_64-pc-linux/sys-include -O2 -O2 -g -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -I. -I. -I../../../llvm-gcc-4.2/gcc -I../../../llvm-gcc-4.2/gcc/. -I../../../llvm-gcc-4.2/gcc/../include -I../../../llvm-gcc-4.2/gcc/../libcpp/include -I../../../llvm-gcc-4.2/gcc/../libdecnumber -I../libdecnumber -I/home/laguest/src/others/compilers/llvm-svn/llvm/include -I/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm/include -g0 -finhibit-size-directive -fno-inline-functions -fno-exceptions -fno-zero-initialized-in-bss -fno-toplevel-reorder -fno-omit-frame-pointer -fno-asynchronous-unwind-tables \ -c ../../../llvm-gcc-4.2/gcc/crtstuff.c -DCRT_BEGIN \ -o crtbegin.o cc1: /home/laguest/src/others/compilers/llvm-svn/llvm/lib/Target/X86/X86AsmPrinter.cpp:266: virtual bool llvm::X86SharedAsmPrinter::doFinalization(llvm::Module&): Assertion `!Subtarget->isTargetDarwin()' failed. ../../../llvm-gcc-4.2/gcc/crtstuff.c:378: internal compiler error: Aborted Please submit a full bug report, with preprocessed source if appropriate. See for instructions. make[2]: *** [crtbegin.o] Error 1 make[2]: Leaving directory `/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2/gcc' make[1]: *** [all-gcc] Error 2 make[1]: Leaving directory `/home/laguest/src/others/compilers/llvm-svn/x86_64/build-llvm-gcc-4.2' make: *** [all] Error 2 Thanks, Luke. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 5 23:55:18 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 5 Jan 2008 23:55:18 -0600 Subject: [LLVMbugs] [Bug 1900] New: MI opt is enabled in cases where it's not valid Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1900 Summary: MI opt is enabled in cases where it's not valid Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: normal Priority: P2 Component: preprocessor AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu This .c file: #define MACRO #include "if-macro.h" #undef MACRO #define MACRO || 1 #include "if-macro.h" combined with this if-macro.h file must diagnose a duplicate definition. #if !defined foo MACRO #define foo int x = 2; #endif -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 6 13:34:11 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 Jan 2008 13:34:11 -0600 Subject: [LLVMbugs] [Bug 1901] New: 'llvm.memmove' Intrinsic typo Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1901 Summary: 'llvm.memmove' Intrinsic typo Product: Documentation Version: trunk Platform: All URL: http://llvm.org/docs/LangRef.html OS/Version: All Status: NEW Severity: minor Priority: P2 Component: General docs AssignedTo: unassignedbugs at nondot.org ReportedBy: llvm at pixelperplexity.net CC: llvmbugs at cs.uiuc.edu In the 'llvm.memmove' Intrinsic documentation, there's a clause which reads: The 'llvm.memmove.*' intrinsics move a block of memory from the source location to the destination location. It is similar to the 'llvm.memcmp' intrinsic but allows the two memory locations to overlap. I think the instance of "llvm.memcmp" is meant to be an instance of "llvm.memcpy", since that makes more sense, and "llvm.memcmp" doesn't exist. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 6 13:53:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 6 Jan 2008 13:53:01 -0600 Subject: [LLVMbugs] [Bug 1901] 'llvm.memmove' Intrinsic typo In-Reply-To: Message-ID: <200801061953.m06Jr1q5028739@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1901 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-06 13:53:00 --- Fixed, thanks: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071231/056883.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 13:52:52 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 13:52:52 -0600 Subject: [LLVMbugs] [Bug 1900] MI opt is enabled in cases where it's not valid In-Reply-To: Message-ID: <200801071952.m07JqqsD013476@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1900 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-07 13:52:51 --- Fixed: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080107/003677.html Many thanks Neil! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 17:42:11 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 17:42:11 -0600 Subject: [LLVMbugs] [Bug 1902] New: Assertion failed using llvm-g++ SVN head on OS X leopard. Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1902 Summary: Assertion failed using llvm-g++ SVN head on OS X leopard. Product: new-bugs Version: unspecified Platform: PC OS/Version: All Status: NEW Severity: blocker Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tewk at cs.utah.edu CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1324) --> (http://llvm.org/bugs/attachment.cgi?id=1324) Preprocessed source Assertion failed: ((i >= FTy->getNumParams() || FTy->getParamType(i) == Params[i]->getType()) && "Calling a function with a bad signature!"), function init, file Instructions.cpp, line 253. /Users/tewk/srcs/manta/Core/Math/TrigSSE.cc: In function ???void Manta::sincos4(float __vector__, float __vector__*, float __vector__*)???: /Users/tewk/srcs/manta/Core/Math/TrigSSE.cc:269: internal compiler error: Abort trap Please submit a full bug report, with preprocessed source if appropriate. See for instructions. I've also reproduced this bug on Kubuntu 7.10 (llvm-gcc compiled with gcc 4.1.3 20070929) system gcc -v Using built-in specs. Target: i686-apple-darwin9 Configured with: /var/tmp/gcc/gcc-5465~16/src/configure --disable-checking -enable-werror --prefix=/usr --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.0/ --with-gxx-include-dir=/include/c++/4.0.0 --with-slibdir=/usr/lib --build=i686-apple-darwin9 --with-arch=apple --with-tune=generic --host=i686-apple-darwin9 --target=i686-apple-darwin9 Thread model: posix gcc version 4.0.1 (Apple Inc. build 5465) Path: /Users/tewk/srcs/llvm URL: http://llvm.org/svn/llvm-project/llvm/trunk Repository Root: http://llvm.org/svn/llvm-project Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8 Revision: 45723 Node Kind: directory Schedule: normal Last Changed Author: lattner Last Changed Rev: 45723 Last Changed Date: 2008-01-07 14:59:58 -0700 (Mon, 07 Jan 2008) Path: /Users/tewk/srcs/llvm_gcc/llvm-gcc-4.0 URL: http://llvm.org/svn/llvm-project/llvm-gcc-4.0/trunk Repository Root: http://llvm.org/svn/llvm-project Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8 Revision: 45724 Node Kind: directory Schedule: normal Last Changed Author: baldrick Last Changed Rev: 45586 Last Changed Date: 2008-01-04 07:15:44 -0700 (Fri, 04 Jan 2008) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 22:16:00 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 22:16:00 -0600 Subject: [LLVMbugs] [Bug 317] [testsuite] Allow executing programs multiple times with different args for a single test "run" In-Reply-To: Message-ID: <200801080416.m084G0IB006471@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=317 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX --- Comment #4 from Chris Lattner 2008-01-07 22:15:59 --- Seems that the complexity of the change isn't worth it. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 22:20:14 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 22:20:14 -0600 Subject: [LLVMbugs] [Bug 601] need docs on analyses and optimizations available in LLVM In-Reply-To: Message-ID: <200801080420.m084KEAv006855@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=601 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Chris Lattner 2008-01-07 22:20:13 --- This is done, thanks Gordon! http://llvm.org/docs/Passes.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 22:23:57 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 22:23:57 -0600 Subject: [LLVMbugs] [Bug 1732] llvmc doesn't work with default /usr/local/etc/llvm/c In-Reply-To: Message-ID: <200801080423.m084Nv07007098@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1732 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #2 from Chris Lattner 2008-01-07 22:23:56 --- llvmc needs a rewrite *** This bug has been marked as a duplicate of bug 686 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 22:27:44 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 22:27:44 -0600 Subject: [LLVMbugs] [Bug 1797] BlockExtractor's "blocks to not extract" list doesn' t allow for anon blocks In-Reply-To: Message-ID: <200801080427.m084RiIT007377@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1797 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Chris Lattner 2008-01-07 22:27:44 --- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056951.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 7 23:20:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 7 Jan 2008 23:20:03 -0600 Subject: [LLVMbugs] [Bug 1721] miscompilation of unusual sized GCC integer types (e.g. i33) In-Reply-To: Message-ID: <200801080520.m085K3L6010426@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1721 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Component|new bugs |llvm-gcc Product|new-bugs |tools Resolution| |FIXED Target Milestone|--- |2.2 Version|unspecified |1.0 --- Comment #12 from Chris Lattner 2008-01-07 23:20:00 --- And here's the gcc 4.2 patch: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056955.html Fixed. Testcase here: test/CFrontend/2008-01-07-UnusualIntSize.c Duncan, please let me know if this also fixes the related 36-bit integer Ada failures. We can tell it is right when ada works :) -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 00:47:46 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 00:47:46 -0600 Subject: [LLVMbugs] [Bug 642] [PowerPC] Miscompilation of some ordered comparisons In-Reply-To: Message-ID: <200801080647.m086lkPF015049@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=642 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.2 --- Comment #9 from Chris Lattner 2008-01-08 00:47:46 --- This is (finally!) implemented, Nate, please review: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056957.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 01:25:37 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 01:25:37 -0600 Subject: [LLVMbugs] [Bug 1795] Missed optimization: eliminating loads around ptrtoint/inttoptr In-Reply-To: Message-ID: <200801080725.m087Pbgl017222@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1795 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Scalar Optimizations Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.2 Version|unspecified |1.0 --- Comment #3 from Chris Lattner 2008-01-08 01:25:36 --- Implemented: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056960.html $ llvm-as < pr1795.ll | opt -instcombine -gvn | llvm-dis produces: define void @myfunc(i32 %.val24) { EntryBlock: %tmp1 = inttoptr i32 %.val24 to i32* ; [#uses=1] getelementptr i32* %tmp1, i32 -3 ; :0 [#uses=1] store i32 1, i32* %0 tail call i32 @ZZZ( i32 1 ) ; :1 [#uses=0] ret void } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 04:09:30 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 04:09:30 -0600 Subject: [LLVMbugs] [Bug 1899] Assertion `!Subtarget->isTargetDarwin()' failed on x86_64 target In-Reply-To: Message-ID: <200801081009.m08A9UkZ000695@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1899 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #6 from Duncan Sands 2008-01-08 04:09:01 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056962.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From duncan.sands at math.u-psud.fr Tue Jan 8 07:55:49 2008 From: duncan.sands at math.u-psud.fr (Duncan Sands) Date: Tue, 8 Jan 2008 14:55:49 +0100 Subject: [LLVMbugs] [Bug 1721] miscompilation of unusual sized GCC integer types (e.g. i33) In-Reply-To: <200801080520.m085K3L6010426@zion.cs.uiuc.edu> References: <200801080520.m085K3L6010426@zion.cs.uiuc.edu> Message-ID: <200801081455.49523.duncan.sands@math.u-psud.fr> Hi Chris, > Duncan, please let me know if this also fixes the related 36-bit integer Ada > failures. We can tell it is right when ada works :) thanks for doing this, even if it is an evil hack :) It fixed two of the failing ACATS tests (out of 16 failures), and didn't cause any new failures, so it's an improvement. That said, I was expecting it to fix a few more tests - I will investigate. Ciao, D. From sabre at nondot.org Tue Jan 8 13:59:46 2008 From: sabre at nondot.org (Chris Lattner) Date: Tue, 8 Jan 2008 11:59:46 -0800 (PST) Subject: [LLVMbugs] [Bug 1721] miscompilation of unusual sized GCC integer types (e.g. i33) In-Reply-To: <200801081455.49523.duncan.sands@math.u-psud.fr> References: <200801080520.m085K3L6010426@zion.cs.uiuc.edu> <200801081455.49523.duncan.sands@math.u-psud.fr> Message-ID: On Tue, 8 Jan 2008, Duncan Sands wrote: >> Duncan, please let me know if this also fixes the related 36-bit integer Ada >> failures. We can tell it is right when ada works :) > > thanks for doing this, even if it is an evil hack :) It's the same approach GCC takes :) > It fixed two of the failing ACATS tests (out of 16 failures), and didn't cause any > new failures, so it's an improvement. That said, I was expecting it to fix a few > more tests - I will investigate. Hey improvement is good! Thanks Duncan, -Chris -- http://nondot.org/sabre/ http://llvm.org/ From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 15:02:07 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 15:02:07 -0600 Subject: [LLVMbugs] [Bug 1902] Assertion failed using llvm-g++ SVN head on OS X leopard. In-Reply-To: Message-ID: <200801082102.m08L277W004746@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1902 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |llvm-gcc Product|new-bugs |tools Resolution| |FIXED Target Milestone|--- |2.2 Version|unspecified |trunk --- Comment #4 from Chris Lattner 2008-01-08 15:02:06 --- Fixed, patch by Anders Carlsson: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056967.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/056968.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 15:11:05 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 15:11:05 -0600 Subject: [LLVMbugs] [Bug 1746] llvm-gcc crash with virtual base classes In-Reply-To: Message-ID: <200801082111.m08LB5hI005277@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1746 Dale Johannesen changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #13 from Dale Johannesen 2008-01-08 15:11:05 --- Fixed. http://llvm.org/viewvc/llvm-project?view=rev&revision=45757 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 8 18:06:53 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 8 Jan 2008 18:06:53 -0600 Subject: [LLVMbugs] [Bug 1883] Overly strict error with initializer + cast In-Reply-To: Message-ID: <200801090006.m0906rEq014767@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1883 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from snaroff at apple.com 2008-01-08 18:06:53 --- In C99 mode, clang is consistent with GCC. For example... [snaroff:llvm/tools/clang] snarofflocal% ../../Debug/bin/clang comp-lit.c -pedantic comp-lit.c:4:24: error: initializer element is not constant static struct Test t = (struct Test){0,0}; ^~~~~~~~~~~~~~~~~~ 1 diagnostic generated. [snaroff:llvm/tools/clang] snarofflocal% cc -c comp-lit.c -std=c99 comp-lit.c:4: error: initializer element is not constant In C90 mode, GCC appears to omit the type initializer error (which seems bad)... [snaroff:llvm/tools/clang] snarofflocal% ../../Debug/bin/clang -std=c90 comp-lit.c -pedantic comp-lit.c:4:24: warning: compound literals are a C99-specific feature static struct Test t = (struct Test){0,0}; ^ comp-lit.c:4:24: error: initializer element is not constant static struct Test t = (struct Test){0,0}; ^~~~~~~~~~~~~~~~~~ 2 diagnostics generated. [snaroff:llvm/tools/clang] snarofflocal% cc -c comp-lit.c -pedantic comp-lit.c:4: warning: ISO C90 forbids compound literals btw...even though the behavior was/is correct, this did expose a significant bug. Expr::isConstantExpr() wasn't handling CompoundLiteralExpr's at all:-( Since the default was "false", we were just lucking out. See commit r45764 for more details. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 16:45:17 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 16:45:17 -0600 Subject: [LLVMbugs] [Bug 1884] Incorrect redeclaration error with enum+int? In-Reply-To: Message-ID: <200801092245.m09MjHP9032374@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1884 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from snaroff at apple.com 2008-01-09 16:45:16 --- clang was wrong...in C99, enum and int's are compatible. Fixed in r45784. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 17:37:42 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 17:37:42 -0600 Subject: [LLVMbugs] [Bug 1890] clang allows illegal redeclaration of function? In-Reply-To: Message-ID: <200801092337.m09Nbgr3003216@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1890 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from snaroff at apple.com 2008-01-09 17:37:42 --- You are correct...this is an error. Fixed in r45789. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 18:36:22 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 18:36:22 -0600 Subject: [LLVMbugs] [Bug 1845] <4 x i1> & <4 x i8> can break in compiler in TargetData:: getAlignmentInfo In-Reply-To: Message-ID: <200801100036.m0A0aMqh006930@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1845 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Component|new bugs |Target Description Classes Keywords| |compile-fail Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.2 Version|unspecified |2.1 --- Comment #9 from Chris Lattner 2008-01-09 18:36:22 --- Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/057009.html Testcase here: test/CodeGen/Generic/bool-vector.ll Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 20:05:22 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 20:05:22 -0600 Subject: [LLVMbugs] [Bug 1880] Building LLVM with llvm-gcc-4.2: duplicate absolute symbols In-Reply-To: Message-ID: <200801100205.m0A25MBj013276@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1880 Dale Johannesen changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #12 from Dale Johannesen 2008-01-09 20:05:22 --- Fixed as described above (there doesn't seem to be a good way to do it). http://llvm.org/viewvc/llvm-project?view=rev&revision=45811 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 20:37:47 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 20:37:47 -0600 Subject: [LLVMbugs] [Bug 1903] New: __func__ should be of type char[], not char* Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1903 Summary: __func__ should be of type char[], not char* Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Testcase: int abcdefghi12(void) { return sizeof(__func__); } should return 12 because the name of the function is 11 characters long, but clang incorrectly comes up with 4. Also, void a(void) { static char* s = __func__; } should not be a compile error, because __func__ is a constant array, and void abcdefghi12(void) { const char (*ss)[12] = &__func__; } should not be a compile error because __func__ is an array. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 9 23:00:58 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 9 Jan 2008 23:00:58 -0600 Subject: [LLVMbugs] [Bug 1890] clang allows illegal redeclaration of function? In-Reply-To: Message-ID: <200801100500.m0A50wqs025649@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1890 Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #2 from Eli Friedman 2008-01-09 23:00:57 --- Not really fixed; try the following: int r (int) { int ss(void); } int ss(int); -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 10 03:24:48 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 Jan 2008 03:24:48 -0600 Subject: [LLVMbugs] [Bug 1904] New: Regression: llvm has gcc PR28045 bug Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1904 Summary: Regression: llvm has gcc PR28045 bug Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu After I updated to SVN r45831, I am getting this while ./configure in ClamAV: checking for gcc bug PR28045... configure: error: your compiler has gcc PR28045 bug, use a different compiler, see http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28045 Previously I was not getting this error, and I was very happy that llvm has avoided a gcc bug :) llvm-gcc -O1 or greater exhibits this bug, however doing llvm-gcc -O0, and opt -std-compile-opts doesn't show the bug. Does llvm-gcc -O1 use some (buggy) optimizations from gcc? (gcc 4.0.1 has this bug) Testcase and output below. $ gcc -O3 pr28045.c $ ./a.out --->no aborts here $ gcc --version gcc (GCC) 4.2.3 20080102 (prerelease) (Debian 4.2.2-5) $ llvm-gcc -O3 pr28045.c edwin at lightspeed2:~/clam/svn3/trunk$ ./a.out Aborted $ llvm-gcc -O1 pr28045.c $ ./a.out Aborted $ llvm-gcc -O0 pr28045.c $ ./a.out --> no aborts here $ llvm-gcc -O0 pr28045.c -emit-llvm -c -o pr28045.bc $ opt -std-compile-opts pr28045.bc -o opt.bc $ llvm-ld -native opt.bc $ ./a.out ---> no aborts here $ llvm-gcc -O1 pr28045.c -emit-llvm -c -o pr28045.bc $ llvm-ld -native opt.bc $ ./a.out Aborted /* (C) Andrew Pinski */ extern void abort(void); struct a { unsigned int bits : 1; signed long val : ((sizeof(long) * 8) - 1); }; int Fnegate (struct a b) { if ((-((long)b.val)) <= ((long) ((1UL << ((sizeof(long) * 8) - 2)) -1UL)) && (-((long)b.val)) >= (-(((long) ((1UL << ((sizeof(long) * 8) - 2)) -1UL))) - 1)) return 0 ; abort (); } int main () { struct a b = {1, 1}; Fnegate (b); return 0; } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 10 03:30:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 Jan 2008 03:30:35 -0600 Subject: [LLVMbugs] [Bug 1730] lli miscompilation In-Reply-To: Message-ID: <200801100930.m0A9UZNu015001@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1730 T??r??k Edwin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #19 from T??r??k Edwin 2008-01-10 03:30:35 --- (In reply to comment #18) > This looks like a Linux x86-64 JIT specific bug. I don't have a way to debug > this. Anton, can you help take look please? Thanks. > I tested with today's SVN r45831, and the bug is gone! A big thanks for fixing this! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 10 22:48:46 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 Jan 2008 22:48:46 -0600 Subject: [LLVMbugs] [Bug 1905] New: instcombine crash in targetdata on vector Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1905 Summary: instcombine crash in targetdata on vector 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 CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1328) --> (http://llvm.org/bugs/attachment.cgi?id=1328) testcase $ llvm-commit/Debug/bin/opt -instcombine b.bc -disable-output opt: TargetData.cpp:302: unsigned int llvm::TargetData::getAlignmentInfo(llvm::AlignTypeEnum, uint32_t, bool) const: Assertion `BestMatchIdx != -1 && "Didn't find alignment info for this datatype!"' failed. llvm-commit/Debug/bin/opt[0x85e6b2a] llvm-commit/Debug/bin/opt[0x85e6df0] [0xffffe500] Aborted (gdb) bt #0 0xffffe405 in __kernel_vsyscall () #1 0xf7caceb5 in raise () from /lib/i686/cmov/libc.so.6 #2 0xf7cae831 in abort () from /lib/i686/cmov/libc.so.6 #3 0xf7ca608e in __assert_fail () from /lib/i686/cmov/libc.so.6 #4 0x084ec7ac in llvm::TargetData::getAlignmentInfo (this=0x86dc180, AlignType=llvm::VECTOR_ALIGN, BitWidth=32, ABIInfo=true) at TargetData.cpp:302 #5 0x084ed572 in llvm::TargetData::getAlignment (this=0x86dc180, Ty=0x86dd040, abi_or_pref=true) at TargetData.cpp:508 #6 0x084ed6ee in llvm::TargetData::getABITypeAlignment (this=0x86dc180, Ty=0x86dd040) at TargetData.cpp:512 #7 0x083ae2dd in PromoteCastOfAllocation (this=0x86dec00, CI=@0x86debc0, AI=@0x86decd8) at InstructionCombining.cpp:6386 #8 0x083bcd0c in visitBitCast (this=0x86dec00, CI=@0x86debc0) at InstructionCombining.cpp:7237 #9 0x083c75bc in visit (this=0x86dec00, I=@0x86debc0) at /home/nicholas/llvm-commit/include/llvm/Instruction.def:151 #10 0x083c7ce7 in DoOneIteration (this=0x86dec00, F=@0x86e0240, Iteration=0) at InstructionCombining.cpp:10603 #11 0x083c7fd0 in runOnFunction (this=0x86dec00, F=@0x86e0240) at InstructionCombining.cpp:10681 #12 0x08579052 in llvm::FPPassManager::runOnFunction (this=0x86d9fd0, F=@0x86e0240) at PassManager.cpp:1171 ---Type to continue, or q to quit--- #13 0x085791f4 in llvm::FPPassManager::runOnModule (this=0x86d9fd0, M=@0x86da870) at PassManager.cpp:1191 #14 0x08578d28 in llvm::MPPassManager::runOnModule (this=0x86dc0e8, M=@0x86da870) at PassManager.cpp:1240 #15 0x08578ee0 in llvm::PassManagerImpl::run (this=0x86df9c0, M=@0x86da870) at PassManager.cpp:1313 #16 0x08578f32 in llvm::PassManager::run (this=0xffe93d58, M=@0x86da870) at PassManager.cpp:1345 #17 0x082e425b in main (argc=4, argv=0xffe93f54) at opt.cpp:428 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 10 22:54:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 10 Jan 2008 22:54:01 -0600 Subject: [LLVMbugs] [Bug 1905] instcombine crash in targetdata on vector In-Reply-To: Message-ID: <200801110454.m0B4s1aT026150@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1905 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #1 from Nick Lewycky 2008-01-10 22:54:01 --- Amazing what "svn update" can do, huh? Closing. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 00:24:18 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 00:24:18 -0600 Subject: [LLVMbugs] [Bug 1906] New: Diagnostics can be confused when more than one file is involved Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1906 Summary: Diagnostics can be confused when more than one file is involved Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: minor Priority: P2 Component: Basic AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu Silly example: double a; int *b; void f(void) { int c = a + #include "some.h" ; } where some.h has just one line: b Outputs (note the stray ~ character): /tmp/some.c:1:44: error: invalid operands to binary expression ('double' and 'int *') double a; int *b; void f(void) { int c = a + ~ ~ ^ 1 diagnostic generated. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 00:44:13 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 00:44:13 -0600 Subject: [LLVMbugs] [Bug 1906] Diagnostics can be confused when more than one file is involved In-Reply-To: Message-ID: <200801120644.m0C6iDRs001688@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1906 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |quality-of-implementation Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-12 00:44:12 --- Fixed, thanks. Very nice catch: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080107/003742.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 03:29:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 03:29:35 -0600 Subject: [LLVMbugs] [Bug 1907] New: 176.gcc is miscompiled with LTO turned off Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1907 Summary: 176.gcc is miscompiled with LTO turned off Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: critical Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: evan.cheng at apple.com CC: llvmbugs at cs.uiuc.edu On X86-32 Mac OS X: make ENABLE_OPTIMIZED=1 TEST=nightly DISABLE_LTO=1 clean report Since both cbe and llc are failing, I am going to guess it's an optimizer bug. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 07:22:39 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 07:22:39 -0600 Subject: [LLVMbugs] [Bug 1908] New: variable redefiniton error Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1908 Summary: variable redefiniton error Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: csaba.hruska at gmail.com CC: llvmbugs at cs.uiuc.edu llvm/clang revision: 45908 test case: const int a [1] = {1}; extern const int a[]; gcc accepts it. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 12:26:12 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 12:26:12 -0600 Subject: [LLVMbugs] [Bug 1903] __func__ should be of type char[], not char* In-Reply-To: Message-ID: <200801121826.m0CIQCAg017574@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1903 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-12 12:26:11 --- Patch applied, thanks! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080107/003745.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 13:52:21 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 13:52:21 -0600 Subject: [LLVMbugs] [Bug 1909] New: Serious tail merging pessimization Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1909 Summary: Serious tail merging pessimization Product: libraries Version: 2.1 Platform: PC OS/Version: All Status: NEW Keywords: code-quality Severity: normal Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: dalej at apple.com, llvmbugs at cs.uiuc.edu Consider: #include #include #define MIN(a,b) ((a)<(b)?(a):(b)) #define MAX(a,b) ((a)>(b)?(a):(b)) void minmax(float result[8]) { float min_x, max_x, min_y, max_y; min_x = MIN(result[0], MIN(result[2], MIN(result[4], result[6]))); max_x = MAX(result[0], MAX(result[2], MAX(result[4], result[6]))); min_y = MIN(result[1], MIN(result[3], MIN(result[5], result[7]))); max_y = MAX(result[1], MAX(result[3], MAX(result[5], result[7]))); printf("transformed bounds: (%.2f, %.2f), (%.2f, %.2f)\n", min_x, min_y, max_x, max_y); } Without tail merging, we get ugly code. With tail merging, we get aweful code :) It seems to be merging lots of "single instructions", yielding code like this: LBB1_53: # bb ucomiss %xmm3, %xmm1 jmp LBB1_2 # bb21 LBB1_54: # bb128 ucomiss %xmm1, %xmm3 jmp LBB1_15 # bb136 LBB1_55: # bb243 ucomiss %xmm3, %xmm1 jmp LBB1_28 # bb251 LBB1_56: # bb358 ucomiss %xmm1, %xmm3 jmp LBB1_41 # bb366 LBB1_57: # bb395 ucomiss %xmm0, %xmm2 jmp LBB1_44 # bb385 LBB1_58: # bb425.bb456_crit_edge movl %edi, %eax jmp LBB1_52 # bb456 ... at the end of the function. If this is what is happening, it should be fixed. Saving a single instruction isn't worthwhile because we have to add the jump, and this makes code potentially slower. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 12 22:53:52 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 12 Jan 2008 22:53:52 -0600 Subject: [LLVMbugs] [Bug 1910] New: llvm-ld does not support linking archives with mixed . bc and .o objects Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1910 Summary: llvm-ld does not support linking archives with mixed .bc and .o objects Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tomas.l.olsen at gmail.com CC: llvmbugs at cs.uiuc.edu Given the three attached files, this sequence of commands should show the problem: [tomas at myhost llbug1]$ llvm-as bug1.ll [tomas at myhost llbug1]$ llvm-as bug2.ll [tomas at myhost llbug1]$ gcc bug.c -c [tomas at myhost llbug1]$ llvm-ld -native -o=first bug1.bc bug2.bc bug.o [tomas at myhost llbug1]$ ./first 42 123 [tomas at myhost llbug1]$ llvm-ar rsv bug.a bug2.bc bug.o llvm-ar: creating bug.a [tomas at myhost llbug1]$ llvm-ld -native -o=second bug1.bc bug.a /tmp/ccy317Eb.o: In function `main': (.text+0x5): undefined reference to `getint_c' collect2: ld returned 1 exit status llvm-ld: [tomas at myhost llbug1]$ would be really nice if this worked! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 13 15:05:48 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 Jan 2008 15:05:48 -0600 Subject: [LLVMbugs] [Bug 1907] 176.gcc is miscompiled with LTO turned off In-Reply-To: Message-ID: <200801132105.m0DL5mKt014222@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1907 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.2 --- Comment #7 from Chris Lattner 2008-01-13 15:05:47 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080107/057156.html You need to rebuild llvm and llvm-gcc for this to work. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 13 15:41:14 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 Jan 2008 15:41:14 -0600 Subject: [LLVMbugs] [Bug 1911] New: --enable-sinking crash Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1911 Summary: --enable-sinking crash Product: libraries Version: 2.1 Platform: PC OS/Version: All Status: NEW Severity: minor Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu llc --enable-sinking crashes on this file, it has something to do with fpstack. This comes from 447.dealII in DISABLE_LTO mode. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 13 21:30:16 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 Jan 2008 21:30:16 -0600 Subject: [LLVMbugs] [Bug 1912] New: ClamAV triggers assertion in loop index splitting pass Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1912 Summary: ClamAV triggers assertion in loop index splitting pass Product: libraries Version: trunk Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: major Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: evan.cheng at apple.com CC: llvmbugs at cs.uiuc.edu ClamAV is failing to build on Mac OS X / X86: /Users/echeng/LLVM/llvm/Release/bin/opt -std-compile-opts -time-passes -info-output-file=/Volumes/Muggles/LLVM/llvm/projects/llvm-test/MultiSource/Applications/Cl\amAV/Output/clamscan.linked.bc.info Output/clamscan.linked.rbc -o Output/clamscan.linked.bc -f Assertion failed: (getLoopLatch() && "Loop latch is missing"), function verifyLoop, file /Volumes/Muggles/LLVM/llvm/include/llvm/Analysis/LoopInfo.h, line 523. 0 opt 0x0038db7e _ZN40_GLOBAL__N_Signals.cpp_00000000_D25A86D715PrintStackTraceEv + 46 1 opt 0x0038e122 _ZN40_GLOBAL__N_Signals.cpp_00000000_D25A86D713SignalHandlerEi + 594 2 libSystem.B.dylib 0x93c9197b _sigtramp + 43 3 ??? 0xffffffff 0x0 + 4294967295 4 libSystem.B.dylib 0x93d0a782 raise + 26 5 libSystem.B.dylib 0x93d19d3f abort + 73 6 libSystem.B.dylib 0x93d0b923 __assert_rtn + 101 7 opt 0x001c0e46 _ZNK4llvm8LoopBaseINS_10BasicBlockEE10verifyLoopEv + 438 8 opt 0x001c124c _ZNK4llvm8LoopBaseINS_10BasicBlockEE10verifyLoopEv + 1468 9 opt 0x0031fac9 _ZN4llvm13PMDataManager23verifyPreservedAnalysisEPNS_4PassE + 153 10 opt 0x0023f093 _ZN4llvm13LPPassManager13runOnFunctionERNS_8FunctionE + 867 11 opt 0x00322666 _ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE + 406 12 opt 0x00216902 _ZN13CGPassManager11runOnModuleERN4llvm6ModuleE + 1394 13 opt 0x00323011 _ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE + 369 14 opt 0x0032341a _ZN4llvm15PassManagerImpl3runERNS_6ModuleE + 122 15 opt 0x0032349a _ZN4llvm11PassManager3runERNS_6ModuleE + 26 16 opt 0x000083b7 main + 2583 17 opt 0x000019e6 start + 54 make[1]: *** [Output/clamscan.linked.bc] Abort trap -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 13 23:34:08 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 13 Jan 2008 23:34:08 -0600 Subject: [LLVMbugs] [Bug 1904] Regression: llvm has gcc PR28045 bug In-Reply-To: Message-ID: <200801140534.m0E5Y8Rs008147@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1904 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-13 23:34:08 --- I believe this is fixed. Please update your llvm and llvm-gcc trees and try again. If this is not fixed, please reopen, thanks! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 04:25:33 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 04:25:33 -0600 Subject: [LLVMbugs] [Bug 1904] Regression: llvm has gcc PR28045 bug In-Reply-To: Message-ID: <200801141025.m0EAPXih027439@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1904 T??r??k Edwin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #2 from T??r??k Edwin 2008-01-14 04:23:25 --- (In reply to comment #1) > I believe this is fixed. Please update your llvm and llvm-gcc trees and try > again. If this is not fixed, please reopen, thanks! > It indeed works with llvm-gcc-4.2, but the bug is still there in llvm-gcc-4.0. It is a gcc 4.0.1 bug (in operand_equal_p) [*], but llvm-gcc-4.0 didn't have this bug until recently. The last SVN revision of llvm-gcc-4.0 where this bug didn't occur: 45586 SVN revision of llvm-gcc-4.0 where bug occurs first time: 45737 However applying 45736:45737 in reverse to latest SVN doesn't remove the bug. [*] The gcc patch is here: http://gcc.gnu.org/viewcvs?view=rev&revision=114930 I am using latest SVN: $ llvm-gcc -v Using built-in specs. Target: x86_64-unknown-linux-gnu Configured with: ../llvm-gcc4.0-2.1.source/configure --prefix=/home/edwin/obj/../install --program-prefix=llvm- --enable-llvm=/home/edwin/llvm-svn/llvm/ --enable-languages=c,c++ --disable-shared --disable-multilib Thread model: posix gcc version 4.0.1 (Apple Computer, Inc. build 5449)(LLVM build ) $ uname -a Linux lightspeed2 2.6.24-rc6-ge697789d #3 Wed Jan 2 11:15:05 EET 2008 x86_64 GNU/Linux llvm: Path: . URL: http://llvm.org/svn/llvm-project/llvm/trunk Repository Root: http://llvm.org/svn/llvm-project Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8 Revision: 45958 Node Kind: directory Schedule: normal Last Changed Author: baldrick Last Changed Rev: 45956 Last Changed Date: 2008-01-14 08:53:45 +0200 (Mon, 14 Jan 2008) llvm-gcc-4.0: Path: . URL: http://llvm.org/svn/llvm-project/llvm-gcc-4.0/trunk Repository Root: http://llvm.org/svn/llvm-project Repository UUID: 91177308-0d34-0410-b5e6-96231b3b80d8 Revision: 45958 Node Kind: directory Schedule: normal Last Changed Author: baldrick Last Changed Rev: 45935 Last Changed Date: 2008-01-13 20:42:52 +0200 (Sun, 13 Jan 2008) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 12:06:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 12:06:01 -0600 Subject: [LLVMbugs] [Bug 1913] New: Should turn invoke+resume into call Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1913 Summary: Should turn invoke+resume into call Product: libraries Version: trunk Platform: PC OS/Version: All Status: NEW Keywords: code-quality, slow-compile Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu It is common in C++ code to end up with an invoke whose unwind destination just does the C++ equivalent of an unwind. This should be turned into a call, saving on code size and speeding up unwinding. For example, here is a testcase: #include int f() { std::string s("hola"); return s.length(); } This compiles to: ... invoke void @_ZNSsC1EPKcRKSaIcE(...) to label %bb69 unwind label %lpad ... %tmp116 = invoke i32 @_ZN9__gnu_cxx18__exchange_and_addEPVii( i32* %tmp114, i32 -1 ) to label %invcont115 unwind label %lpad148 ... pad: ; preds = %entry %eh_ptr = call i8* @llvm.eh.exception( ) ; [#uses=2] %eh_select147 = call i32 (i8*, i8*, ...)* @llvm.eh.selector.i32( i8* %eh_ptr, i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*), i8* null ) br label %Unwind lpad148: ; preds = %bb111 %eh_ptr149 = call i8* @llvm.eh.exception( ) ; [#uses=2] %eh_select151 = call i32 (i8*, i8*, ...)* @llvm.eh.selector.i32( i8* %eh_ptr149, i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*), i8* null ) br label %Unwind Unwind: ; preds = %lpad148, %lpad %eh_exception.0 = phi i8* [ %eh_ptr, %lpad ], [ %eh_ptr149, %lpad148 ] call i32 (...)* @_Unwind_Resume_or_Rethrow( i8* %eh_exception.0 ) unreachable Both of these invokes can be turned into (throwing) call instructions, saving code size and reducing compile time. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 12:24:17 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 12:24:17 -0600 Subject: [LLVMbugs] [Bug 1914] New: Should turn invoke of a resume into a branch Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1914 Summary: Should turn invoke of a resume into a branch Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu Depends on: 1508 Due to inlining, we can end up with an invoke of _Unwind_Resume. This should be lowered to a branch. However there is the problem of identifying the resume libcall: gcc generates four different ones: _Unwind_Resume (most platforms) _Unwind_Resume_or_Rethrow (LLVM hack for darwin) _Unwind_SjLj_Resume (when doing sj/lj style eh) __cxa_end_cleanup (LLVM hack for ARM) Since these are target dependent but language independent, we could output a special rewind instruction or intrinsic and have it lowered to the right one in the code generators, rather than trying to recognize all these different variants. On the other hand, since both _Unwind_Resume and _Unwind_Resume_or_Rethrow act the same (the first is an optimized version of the second that can only be used in special circumstances), and the second one presumably underlies cxa_throw, it might be better to just recognize these libcalls. There is also the problem that turning the invoke into a branch can leave eh.exception and eh.selector calls floating in space, see PR1508. Here's an example: compile to bitcode at -O0, then use opt -std-compile-opts to optimize. You end up with an invoke of _Unwind_Resume. #include class A { public: A() {} ~A() {} }; void f() { A a; throw 5.0; } main() { try { f(); } catch(...) { printf("caught\n"); } } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 12:28:48 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 12:28:48 -0600 Subject: [LLVMbugs] [Bug 1190] llvm-gcc does not support pointers to nested functions ( trampolines) In-Reply-To: Message-ID: <200801141828.m0EISmoh020450@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1190 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #4 from Duncan Sands 2008-01-14 12:28:48 --- Closing this because it works for me. If someone has trouble with stack-smashing protectors, then they can open a new bug for that. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 12:40:37 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 12:40:37 -0600 Subject: [LLVMbugs] [Bug 1915] New: Speedup linscan by using register iterators Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1915 Summary: Speedup linscan by using register iterators Product: libraries Version: trunk Platform: PC OS/Version: All Status: NEW Keywords: slow-compile Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu Linear scan spends most of its time (on x86) in rewriteInstructionsForSpills and rewriteInstructionForSpills. This is because it scans all of the intervals in the live range looking for uses of the register being spilled. It would be must more efficient to walk the MachineRegisterInfo::reg_iterator list instead. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 12:46:58 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 12:46:58 -0600 Subject: [LLVMbugs] [Bug 1916] New: LLVM stack unwinding is slow Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1916 Summary: LLVM stack unwinding is slow Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu When unwinding code generated by LLVM, every invoke encountered causes a branch to the corresponding landing pad, even if all the handler is going to do for that exception is reraise it (in extreme cases you get PR1913). This is because we add a catch-all to the dwarf unwind tables in order to enforce the semantics of invoke (an invoke always branches to the landing pad whenever an exception unwinds through it). If we don't enforce them then things can break! [Here's an example: compile at -O0 then use opt -std-compile-opts to cause inlining. #include class A { public: A() {} ~A() {} }; void f() { A a; throw 5.0; } main() { try { f(); } catch(...) { printf("caught\n"); } } You don't get "caught" unless a catch-all was pushed in the original IR]. The result is slower unwinding. When unwinding an exception using code compiled with mainline gcc, you only get a branch to the final landing pad that actually handles the exception, and not to every landing pad encountered as the stack is unwound. I think this can be fixed at codegen time (no inlining is done by codegen): we can drop the semantics of invoke and no longer require all exceptions that unwind through an invoke to branch to the landing pad. I think we can do this without needing to know how to recognise the catch-all for the language: the eh.selector has an array of type infos and returns what amounts to an index into that array. For each possible index value we can analyse the control flow, and determine those for which only trivial code is executed followed by a rewind. It would be nice to say that we could remove all those indices from the eh.selector call, but we can't (see later example) but we can remove such an index if it is the last one, which is the most important case anyway since that is almost always the "catch-all" we added. Here's the example: suppose an exception E matches typeinfo A and also typeinfo B (this is possible with C++). If, in the selector, A occurs before B, and we do something trivial if A matches but something non-trivial if B matches, then removing the A typeinfo would change the behaviour if an E is thrown. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 13:30:52 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 13:30:52 -0600 Subject: [LLVMbugs] [Bug 1917] New: Use address space markers to improve GC Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1917 Summary: Use address space markers to improve GC Product: libraries Version: trunk Platform: PC OS/Version: All Status: NEW Keywords: code-quality Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: gordonhenriksen at mac.com, llvmbugs at cs.uiuc.edu One of the big bad things about the llvm.gcroot intrinsic is that it implicitly pins the variable on the stack, preventing optimization and analysis from being effective. Instead of marking GC'd pointers with llvm.gcroot, it would be really cool to: 1. Mark 'GC pointers' as such, and allow mem2reg to promote them. 2. Have the code generator insert spills and reloads of the registers when they are live across safe points (i.e. calls). I've long resisted having a flag on pointers to do this, but now that we have support for alternate address spaces, it seems like a reasonable thing to define one address space as being the "GC address space" or something, and use this mechanism to mark our pointers. This should lead to significantly better codegen for languages that use GC. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 16:36:04 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 16:36:04 -0600 Subject: [LLVMbugs] [Bug 1869] C++ Frontend ICE In-Reply-To: Message-ID: <200801142236.m0EMa4A2005366@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1869 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #6 from Chris Lattner 2008-01-14 16:36:03 --- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080114/057188.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 14 17:13:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 14 Jan 2008 17:13:03 -0600 Subject: [LLVMbugs] [Bug 1904] Regression: llvm has gcc PR28045 bug In-Reply-To: Message-ID: <200801142313.m0END37c008486@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1904 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #5 from Chris Lattner 2008-01-14 17:13:02 --- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080114/057193.html Thanks! -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 15 04:06:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 Jan 2008 04:06:03 -0600 Subject: [LLVMbugs] [Bug 1918] New: clang asserts when including gcc's C99 math.h header. Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1918 Summary: clang asserts when including gcc's C99 math.h header. Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: critical Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: csaba.hruska at gmail.com CC: llvmbugs at cs.uiuc.edu cat test2.c #include ccc -c test2.c clang -emit-llvm-bc -o test2.o -U__GNUC__ -I/usr/lib/gcc/i486-linux-gnu/4.1.3/include test2.c clang: CodeGenModule.cpp:304: llvm::Constant* GenerateAggregateInit(const clang::InitListExpr*, clang::CodeGen::CodeGenModule&): Assertion `(ILE->getType()->isArrayType() || ILE->getType()->isStructureType()) && "Bad type for init list!"' failed. clang((anonymous namespace)::PrintStackTrace()+0x19)[0x82af049] /lib/tls/i686/cmov/libc.so.6(abort+0x101)[0xb7d14201] /lib/tls/i686/cmov/libc.so.6(__assert_fail+0xee)[0xb7d0bb6e] clang[0x80eedc1] -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 15 16:13:16 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 Jan 2008 16:13:16 -0600 Subject: [LLVMbugs] [Bug 1839] __builtin_trap doesn't trap In-Reply-To: Message-ID: <200801152213.m0FMDGZf015268@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1839 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED Target Milestone|--- |2.2 --- Comment #7 from Chris Lattner 2008-01-15 16:13:15 --- This works now, generating ud2 on x86 and abort() on other targets. Thanks to Anton for doing most of it. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 15 22:42:11 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 15 Jan 2008 22:42:11 -0600 Subject: [LLVMbugs] [Bug 745] [llvmgcc] Passing large structures by value to functions results in large code. In-Reply-To: Message-ID: <200801160442.m0G4gB6S011448@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=745 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.2 --- Comment #5 from Chris Lattner 2008-01-15 22:42:10 --- This has been fixed in llvm and llvm-gcc, and awaits specific targets to pick it up. You should be good on x86 and x86-64 now. Thanks to Rafael and Evan for making this happen for 2.2 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 17 09:46:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 Jan 2008 09:46:55 -0600 Subject: [LLVMbugs] [Bug 1919] New: Problem with ModulePass depending on FunctionPass depending on AA Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1919 Summary: Problem with ModulePass depending on FunctionPass depending on AA 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: wmatyjewicz at fastmail.fm CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1336) --> (http://llvm.org/bugs/attachment.cgi?id=1336) source code for the passes While trying to run MyPass (shown below) on any input the optimizer hits the following assertion: TargetData.h:114: llvm::TargetData::TargetData(): Assertion `0 && "ERROR: Bad TargetData ctor used. " "Tool did not specify a TargetData to use?" The reduced code for the pass and the analysis it uses: #include "llvm/Pass.h" #include "llvm/Module.h" #include "llvm/Analysis/AliasAnalysis.h" using namespace llvm; namespace { struct MyAnalysis : public FunctionPass { static char ID; MyAnalysis() : FunctionPass((intptr_t) &ID) {} virtual void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequiredTransitive(); } virtual bool runOnFunction(Function &F) { return false; } }; char MyAnalysis::ID = 0; RegisterPass A("my-analysis", "My analysis"); struct MyPass : public ModulePass { static char ID; MyPass() : ModulePass((intptr_t) &ID) {} virtual void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequired(); } virtual bool runOnModule(Module &M) { return false; } }; char MyPass::ID = 0; RegisterPass X("my-pass", "My pass"); } When I changed MyPass to be a FunctionPass the assertion failure disappeared. I suppos, it's some kind of a pass manager bug. The attachment allows to build the loadable module containing the passes. It can be extracted anywhere. Just run configure script passing --with-llvmsrc and --with-llvmobj arguments. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 17 13:47:42 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 Jan 2008 13:47:42 -0600 Subject: [LLVMbugs] [Bug 1920] New: Structures are not Correctly Passed to Variadic Functions Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1920 Summary: Structures are not Correctly Passed to Variadic Functions Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Keywords: miscompilation Severity: major Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tjablin at cs.princeton.edu CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1337) --> (http://llvm.org/bugs/attachment.cgi?id=1337) Test Case When LLVM tries to pass a structure, it breaks the structure up into several parts, for example when passing the structure: struct tiny { char a; char b; char c; } LLVM generates the following code: %tmp4 = bitcast %struct.tiny* %t to { i16, i8 }* ; <{ i16, i8 }*> [#uses=1] %tmp5 = getelementptr { i16, i8 }* %tmp4, i32 0, i32 0 ; [#uses=1] %tmp6 = load i16* %tmp5 ; [#uses=1] %tmp7 = bitcast %struct.tiny* %t to { i16, i8 }* ; <{ i16, i8 }*> [#uses=1] %tmp8 = getelementptr { i16, i8 }* %tmp7, i32 0, i32 1 ; [#uses=1] %tmp9 = load i8* %tmp8 ; [#uses=1] call void (i32, ...)* @foo( i32 0, i16 %tmp6, i8 %tmp9 ) This approach is okay in general, but disastrous for variadic functions. For variadic functions, the callee will attempt to store each of the register passed parameters into a contiguous region of memory. The code generator will copy the %tmp6 and %tmp9 into adjacent machine-words of memory. Thus, on a 64-bit platform the date will appear as follows in memory: a b X X X X X X c. The solution is to layout parameters to variadic functions exactly as they would be laid out in memory, replicating exactly any packing bits. I have included a test case. The correct output is "a b c". GCC handles the issue correctly. I have tested with llvm2.1-gcc4.2 on x86-64 and IA64. On both platforms the output is "a b". Thanks. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 17 14:36:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 Jan 2008 14:36:55 -0600 Subject: [LLVMbugs] [Bug 1920] Structures are not Correctly Passed to Variadic Functions In-Reply-To: Message-ID: <200801172036.m0HKatxP011852@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1920 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asl at math.spbu.ru Status|NEW |RESOLVED Resolution| |DUPLICATE --- Comment #2 from Anton Korobeynikov 2008-01-17 14:36:54 --- I think is the same as PR1249. Please correct me, if I'm wrong. *** This bug has been marked as a duplicate of bug 1249 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 17 19:54:26 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 Jan 2008 19:54:26 -0600 Subject: [LLVMbugs] [Bug 1919] Problem with ModulePass depending on FunctionPass depending on AA In-Reply-To: Message-ID: <200801180154.m0I1sQhb005454@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1919 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |LATER --- Comment #1 from Devang Patel 2008-01-17 19:54:25 --- Usually, a ModulePass depends on a FunctionPass when while processing entire module the ModulePass intends to collect function specific information through a FunctionPass. In the context, the FunctionPass only has function level information to operator on. In your example, FunctionPass uses AA which uses module level TargetData information. This is not supported in current implementation. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 17 20:11:40 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 17 Jan 2008 20:11:40 -0600 Subject: [LLVMbugs] [Bug 1651] ICE building llvm-gcc-4.2 - invalid free in PassManager In-Reply-To: Message-ID: <200801180211.m0I2BeKO006687@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1651 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #7 from Devang Patel 2008-01-17 20:11:39 --- It seems originally reported bug is now fixed. Please open new PR for other issues. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 04:41:08 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 04:41:08 -0600 Subject: [LLVMbugs] [Bug 1921] New: clang fails with: Assertion `Ty && " Invalid GetElementPtrInst indices for type!"' Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1921 Summary: clang fails with: Assertion `Ty && "Invalid GetElementPtrInst indices for type!"' Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1339) --> (http://llvm.org/bugs/attachment.cgi?id=1339) delta reduced testcase Using clang r46160, llvm r46160, codegen fails for attached reduced testcase $ clang -emit-llvm-bc matcher-bm.i clang: /home/edwin/llvm-svn/llvm/include/llvm/Instructions.h:374: const llvm::Type* llvm::checkType(const llvm::Type*): Assertion `Ty && "Invalid GetElementPtrInst indices for type!"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang(llvm::GetElementPtrInst::GetElementPtrInst(llvm::Value*, llvm::Value*, std::basic_string, std::allocator > const&, llvm::Instruction*)+0x180)[0x5d1f90] clang[0x4be233] clang[0x4c23f7] clang[0x4c076c] clang(clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*)+0x58)[0x4c1478] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&)+0x17a)[0x4c66da] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitWhileStmt(clang::WhileStmt const&)+0x187)[0x4c6337] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&)+0x17a)[0x4c66da] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitForStmt(clang::ForStmt const&)+0x211)[0x4c5cd1] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::GenerateCode(clang::FunctionDecl const*)+0x2ed)[0x4ae5ed] clang(clang::CodeGen::CodeGenModule::EmitFunction(clang::FunctionDecl const*)+0x33)[0x4a48d3] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted $ clang --version Low Level Virtual Machine (http://llvm.org/): llvm version 2.3svn Optimized build with assertions. $ uname -a Linux lightspeed2 2.6.24-rc6-ge697789d #3 Wed Jan 2 11:15:05 EET 2008 x86_64 GNU/Linux -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 04:48:35 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 04:48:35 -0600 Subject: [LLVMbugs] [Bug 1922] New: clang can't codegen pthread_mutex_t initializer Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1922 Summary: clang can't codegen pthread_mutex_t initializer Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1340) --> (http://llvm.org/bugs/attachment.cgi?id=1340) delta reduced testcase Using clang r46160, llvm r46160, codegen fails: $ clang -emit-llvm-bc others.i clang: CodeGenModule.cpp:304: llvm::Constant* GenerateAggregateInit(const clang::InitListExpr*, clang::CodeGen::CodeGenModule&): Assertion `(ILE->getType()->isArrayType() || ILE->getType()->isStructureType()) && "Bad type for init list!"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang[0x4a7636] clang[0x4a6835] clang(clang::CodeGen::CodeGenModule::EmitGlobalVar(clang::FileVarDecl const*)+0x80)[0x4a70a0] clang(clang::CodeGen::CodeGenModule::EmitGlobalVarDeclarator(clang::FileVarDecl const*)+0x27)[0x4a72f7] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted The original file contains this C code: # include static pthread_mutex_t cli_gentemp_mutex = PTHREAD_MUTEX_INITIALIZER; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 04:52:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 04:52:43 -0600 Subject: [LLVMbugs] [Bug 1923] New: clang: Assertion `V.size() == T->getNumElements() && " Invalid initializer vector for constant structure"' Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1923 Summary: clang: Assertion `V.size() == T->getNumElements() && "Invalid initializer vector for constant structure"' Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1341) --> (http://llvm.org/bugs/attachment.cgi?id=1341) delta reduced testcase Using clang r46160, codegen fails: $ clang rtf.i -emit-llvm-bc clang: Constants.cpp:367: llvm::ConstantStruct::ConstantStruct(const llvm::StructType*, const std::vector >&): Assertion `V.size() == T->getNumElements() && "Invalid initializer vector for constant structure"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang[0x59b37c] clang(llvm::ConstantStruct::get(llvm::StructType const*, std::vector > const&)+0x1e7)[0x5a5bc7] clang[0x4a75b0] clang[0x4a6835] clang(clang::CodeGen::CodeGenModule::EmitGlobalVar(clang::FileVarDecl const*)+0x80)[0x4a70a0] clang(clang::CodeGen::CodeGenModule::EmitGlobalVarDeclarator(clang::FileVarDecl const*)+0x27)[0x4a72f7] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted The attached testcase gives no warnings with gcc (it could be reduced even further, but then gcc would give warnings) I have used the following script as delta's -test parameter: clang -emit-llvm-bc $1 2>&1 | grep "Invalid initializer vector for constant structure" if ! test $? == 0; then exit 1 fi gcc -Wall -Wfatal-errors -fsyntax-only $1 if ! test $? == 0; then exit 1 fi exit 0 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 14:14:00 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 14:14:00 -0600 Subject: [LLVMbugs] [Bug 1924] New: Assertion failure at ParseStmt.cpp:983 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1924 Summary: Assertion failure at ParseStmt.cpp:983 Product: clang Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: normal Priority: P2 Component: parser AssignedTo: unassignedbugs at nondot.org ReportedBy: brett at python.org CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1342) --> (http://llvm.org/bugs/attachment.cgi?id=1342) Text file with error info as listed in the bug description for easier reading. [All info mentioned below is in the attached text file for possible easier reading] With a checkout of LLVM (r46142) and clang (r46147) I decided to run clang against a svn checkout of Python (r60021). With the command-line arguments of:: find . -name '*.c' -print | xargs /unix/repos/llvm/Debug/bin/clang -I=. -I=Include -I=$CPPFLAGS -fsyntax-only >From the top-level directory of a checkout of the trunk, I got the following error (chopped out the previous warnings so that only the leading reporting error is show below):: ./Modules/_ctypes/libffi/src/ia64/ffi.c:77:52: error: use of undeclared identifier 'value' asm ("stf.spill %0 = %1%P0" : "=m" (*addr) : "f"(value)); ^ Assertion failed: (Names.size() == Constraints.size() && Constraints.size() == Exprs.size() && "Input operand size mismatch!"), function ParseAsmStatement, file ParseStmt.cpp, line 983. 0 clang 0x00228df5 _ZN40_GLOBAL__N_Signals.cpp_00000000_C27D906915PrintStackTraceEv + 45 1 clang 0x0022919b _ZN40_GLOBAL__N_Signals.cpp_00000000_C27D906913SignalHandlerEi + 323 2 libSystem.B.dylib 0x92f2497b _sigtramp + 43 3 ??? 0xffffffff 0x0 + 4294967295 4 libSystem.B.dylib 0x92f9d782 raise + 26 5 libSystem.B.dylib 0x92facd3f abort + 73 6 libSystem.B.dylib 0x92f9e923 __assert_rtn + 101 7 clang 0x00106372 _ZN5clang6Parser17ParseAsmStatementEv + 1000 8 clang 0x00107211 _ZN5clang6Parser27ParseStatementOrDeclarationEb + 1863 9 clang 0x00106610 _ZN5clang6Parser26ParseCompoundStatementBodyEb + 124 10 clang 0x00106980 _ZN5clang6Parser26ParseFunctionStatementBodyEPvNS_14SourceLocationES2_ + 26 11 clang 0x0010a820 _ZN5clang6Parser23ParseFunctionDefinitionERNS_10DeclaratorE + 418 12 clang 0x0010ac90 _ZN5clang6Parser36ParseDeclarationOrFunctionDefinitionEv + 1122 13 clang 0x0010ade9 _ZN5clang6Parser24ParseExternalDeclarationEv + 83 14 clang 0x0010afc5 _ZN5clang6Parser17ParseTopLevelDeclERPv + 61 15 clang 0x00082ff0 _ZN37_GLOBAL__N__ZN5clang11ASTConsumerD2Ev11ASTStreamer16ReadTopLevelDeclEv + 24 16 clang 0x00083134 _ZN5clang8ParseASTERNS_12PreprocessorEPNS_11ASTConsumerEb + 258 17 clang 0x00023134 _ZNSsC1IN9__gnu_cxx17__normal_iteratorIPcSsEEEET_S4_RKSaIcE + 4808 18 clang 0x000251b9 main + 1315 19 clang 0x000020c2 start + 54 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 17:46:08 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 17:46:08 -0600 Subject: [LLVMbugs] [Bug 1925] New: Assertion failed in RegisterScavenging with regalloc=local Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1925 Summary: Assertion failed in RegisterScavenging with regalloc=local 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 CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1343) --> (http://llvm.org/bugs/attachment.cgi?id=1343) bugpoint-reduced-simplified.bc $ llc -regalloc=local < bugpoint-reduced-simplified.bc llc: /home/laurov/tester/llvm/lib/CodeGen/RegisterScavenging.cpp:114: void llvm::RegScavenger::forward(): Assertion `false && "Using an undefined register!"' failed. llc((anonymous namespace)::PrintStackTrace()+0x19)[0x84cfca9] /lib/tls/i686/cmov/libc.so.6(abort+0x101)[0xb7cc6201] /lib/tls/i686/cmov/libc.so.6(__assert_fail+0xee)[0xb7cbdb6e] llc(llvm::RegScavenger::forward()+0x563)[0x8343ae3] Aborted (core dumped) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 18 22:57:37 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 18 Jan 2008 22:57:37 -0600 Subject: [LLVMbugs] [Bug 1926] New: llvm-gcc-4.2 does not use LLVM_VERSION_INFO Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1926 Summary: llvm-gcc-4.2 does not use LLVM_VERSION_INFO Product: new-bugs Version: unspecified Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tonic at nondot.org CC: llvmbugs at cs.uiuc.edu Specifying LLVM_VERSION_INFO=XX when building llvm-gcc-4.2 does not set the version. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 02:02:00 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 02:02:00 -0600 Subject: [LLVMbugs] [Bug 1927] New: make clean gives errors Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1927 Summary: make clean gives errors Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tonic at nondot.org CC: llvmbugs at cs.uiuc.edu On linux, I get the following error when running make clean: gmake[1]: *** No rule to make target `/mounts/zion/disks/0/localhome/tbrethou/2.2/testing/llvm-2.2/Release/bin/tblgen', needed by `/mounts/zion/disks/0/localhome/tbrethou/2.2/testing/llvm-2.2/lib/VMCore/Release/Intrinsics.gen.tmp'. Stop. gmake[1]: Leaving directory `/mounts/zion/disks/0/localhome/tbrethou/2.2/testing/llvm-2.2/lib/VMCore' -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 02:03:31 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 02:03:31 -0600 Subject: [LLVMbugs] [Bug 1928] New: new warnings on darwin8 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1928 Summary: new warnings on darwin8 Product: new-bugs Version: unspecified Platform: Macintosh OS/Version: MacOS X Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: tonic at nondot.org CC: llvmbugs at cs.uiuc.edu llvm[3]: Compiling MipsRegisterInfo.cpp for Release build MipsRegisterInfo.cpp: In member function 'virtual void llvm::MipsRegisterInfo::emitPrologue(llvm::MachineFunction&) const': MipsRegisterInfo.cpp:280: warning: 'RAOffset' may be used uninitialized in this function MipsRegisterInfo.cpp:280: warning: 'FPOffset' may be used uninitialized in this function APFloat.cpp: In constructor 'llvm::APFloat::APFloat(const llvm::fltSemantics&, const char*)': APFloat.cpp:680: warning: control may reach end of non-void function 'llvm::APFloat::opStatus llvm::APFloat::convertFromString(const char*, llvm::APFloat::roundingMode)' being inlined 'llvm[2]: Compiling ConfigLexer.cpp for Release build ConfigLexer.cpp:2615: warning: 'void yy_fatal_error(const char*)' defined but not used -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 07:41:51 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 07:41:51 -0600 Subject: [LLVMbugs] =?utf-8?b?W0J1ZyAxOTI5XSBOZXc6IEFQSW50Lmg6MjE2OiBlcnJv?= =?utf-8?b?cjogIGV4cGVjdGVkIOKAmCAsIOKAmSBvciDigJggLi4uIOKAmSBiZWZvcmUg?= =?utf-8?q?numeric_constant?= Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1929 Summary: APInt.h:216: error: expected ???,??? or ???...??? before numeric constant Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu Build fails for svn r46191. make[2]: Entering directory `/home/edwin/llvm-svn/llvm/utils/TableGen' llvm[2]: Compiling FileParser.cpp for Release build In file included from /home/edwin/llvm-svn/llvm/include/llvm/ADT/APFloat.h:105, from /home/edwin/llvm-svn/llvm/include/llvm/ADT/StringExtras.h:19, from /home/edwin/llvm-svn/llvm/utils/TableGen/FileParser.y:17: /home/edwin/llvm-svn/llvm/include/llvm/ADT/APInt.h:216: error: expected ???,??? or ???...??? before numeric constant make[2]: *** [/home/edwin/llvm-svn/llvm/utils/TableGen/Release/FileParser.o] Error 1 make[2]: Leaving directory `/home/edwin/llvm-svn/llvm/utils/TableGen' make[1]: *** [TableGen/.makeall] Error 2 make[1]: Leaving directory `/home/edwin/llvm-svn/llvm/utils' make: *** [all] Error 1 That line contains: /// Profile - Used to insert APInt objects, or objects that contain APInt /// objects, into FoldingSets. void Profile(FoldingSetNodeID& ID) const; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 11:39:27 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 11:39:27 -0600 Subject: [LLVMbugs] =?utf-8?b?W0J1ZyAxOTI5XSBBUEludC5oOjIxNjogZXJyb3I6ICBl?= =?utf-8?b?eHBlY3RlZCDigJggLCDigJkgb3Ig4oCYIC4uLiDigJkgYmVmb3JlIG51bWVy?= =?utf-8?q?ic_constant?= In-Reply-To: Message-ID: <200801191739.m0JHdRHk029629@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1929 T??r??k Edwin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #6 from T??r??k Edwin 2008-01-19 11:39:27 --- (In reply to comment #5) > Thanks. It looks like ID is an enum value in the generated output of bison. > Does the build work for you now? > Yes, utils/TableGen builds now. The rest is still compiling, but I suppose those will be ok too. Thanks. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 16:50:21 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 16:50:21 -0600 Subject: [LLVMbugs] [Bug 1930] New: cbe doesn't handle insertelement, extractelement and shufflevector Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1930 Summary: cbe doesn't handle insertelement, extractelement and shufflevector Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: llvmbugs at cs.uiuc.edu The C Backend doesn't deal with the vectors. It doesn't have visitors for InsertElement, ExtractElement or ShuffleVector. This is preventing me from bugpointing vectorized code. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 16:57:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 16:57:43 -0600 Subject: [LLVMbugs] [Bug 1930] cbe doesn't handle insertelement, extractelement and shufflevector In-Reply-To: Message-ID: <200801192257.m0JMvhNA014203@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1930 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asl at math.spbu.ru Status|NEW |RESOLVED Resolution| |FIXED -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 16:57:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 16:57:55 -0600 Subject: [LLVMbugs] [Bug 1930] cbe doesn't handle insertelement, extractelement and shufflevector In-Reply-To: Message-ID: <200801192257.m0JMvtKr014234@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1930 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|FIXED |DUPLICATE --- Comment #1 from Anton Korobeynikov 2008-01-19 16:57:53 --- *** This bug has been marked as a duplicate of bug 1126 *** -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 18:23:27 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 18:23:27 -0600 Subject: [LLVMbugs] [Bug 1852] GCC 4.0 front end won't build for a target that doesn' t support DWARF exception handling In-Reply-To: Message-ID: <200801200023.m0K0NRRx020105@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1852 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Anton Korobeynikov 2008-01-19 18:23:27 --- Sorry for the delay. I've commited slightly different patch to 4.0 to be on par with 4.2 (which has this sort of stuff for ages - r40011). http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080114/057381.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 22:49:33 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 22:49:33 -0600 Subject: [LLVMbugs] [Bug 1931] New: Invalid division instcombine Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1931 Summary: Invalid division instcombine Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu define i32 @fold(i32 %a) { %b = sdiv i32 %a, -1 %c = sdiv i32 %b, -2 ret i32 %c } gets combined into define i32 @fold(i32 %a) nounwind { %c = sdiv i32 %a, 2 ; [#uses=1] ret i32 %c } by opt -std-compile-opts. These functions are not equivalent! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 23:10:38 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 23:10:38 -0600 Subject: [LLVMbugs] [Bug 1932] New: Invalid float division instcombine Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1932 Summary: Invalid float division instcombine Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu define double @fold(i1 %a, double %b) { %s = select i1 %a, double 0., double 1. %c = fdiv double %b, %s ret double %c } folds to define double @fold(i1 %a, double %b) nounwind { %c = fdiv double %b, 1.000000e+00 ; [#uses=1] ret double %c } using opt -std-compile-opts. This is wrong! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 19 23:31:24 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 19 Jan 2008 23:31:24 -0600 Subject: [LLVMbugs] [Bug 1933] New: Invalid remainder instcombine Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1933 Summary: Invalid remainder instcombine Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu define i32 @fold(i32 %a) { %s = mul i32 %a, 3 %c = urem i32 %s, 3 ret i32 %c } using opt -std-compile-opts folds to define i32 @fold(i32 %a) nounwind { ret i32 0 } This is wrong! For example, try @fold(i32 2863311531). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 20 00:24:31 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 Jan 2008 00:24:31 -0600 Subject: [LLVMbugs] [Bug 1934] New: Another invalid division instcombine Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1934 Summary: Another invalid division instcombine Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu define i32 @fold(i32 %a) { %b = sdiv i32 %a, -1 %s = icmp eq i32 %b, -2147483648 %c = zext i1 %s to i32 ret i32 %c } combines to define i32 @fold(i32 %a) nounwind { ret i32 0 } This is not correct! Try calling @fold(i32 -2147483648). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 20 03:07:54 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 Jan 2008 03:07:54 -0600 Subject: [LLVMbugs] [Bug 1935] New: Use of freed memory in CodeGenPrepare::OptimizeBlock Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1935 Summary: Use of freed memory in CodeGenPrepare::OptimizeBlock Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu Here's where the freed memory was used: Invalid read of size 2 at 0x83D3016: llvm::Value::getValueID() const (Value.h:207) by 0x83D32D6: llvm::Instruction::getOpcode() const (Instruction.h:104) by 0x8845026: llvm::ZExtInst::classof(llvm::Instruction const*) (Instructions.h:1932) by 0x88454A6: bool llvm::isa_impl(llvm::Instruction const&) (Casting.h:54) by 0x88454BC: llvm::isa_impl_wrap::doit(llvm::Instruction const&) (Casting.h:71) by 0x88454D2: bool llvm::isa_impl_cl::isa(llvm::Instruction const&) (Casting.h:83) by 0x88454E8: bool llvm::isa_impl_cl::isa(llvm::Instruction*) (Casting.h:101) by 0x8845500: bool llvm::isa(llvm::Instruction* const&) (Casting.h:116) by 0x88441BB: (anonymous namespace)::CodeGenPrepare::OptimizeBlock(llvm::BasicBlock&) (CodeGenPrepare.cpp:1046) by 0x8844F60: (anonymous namespace)::CodeGenPrepare::runOnFunction(llvm::Function&) (CodeGenPrepare.cpp:85) by 0x8937C6B: llvm::FPPassManager::runOnFunction(llvm::Function&) (PassManager.cpp:1171) by 0x8937EC9: llvm::FunctionPassManagerImpl::run(llvm::Function&) (PassManager.cpp:1129) Here is where the memory was freed: Address 0x42c321c is 4 bytes inside a block of size 56 free'd at 0x402231C: operator delete(void*) (vg_replace_malloc.c:342) by 0x89249D4: llvm::TruncInst::~TruncInst() (Instructions.h:1864) by 0x87FD900: llvm::iplist >::erase(llvm::ilist_iterator) (ilist:368) by 0x89141A8: llvm::Instruction::eraseFromParent() (Instruction.cpp:68) by 0x8843C2F: OptimizeNoopCopyExpression(llvm::CastInst*, llvm::TargetLowering const&) (CodeGenPrepare.cpp:405) by 0x884417A: (anonymous namespace)::CodeGenPrepare::OptimizeBlock(llvm::BasicBlock&) (CodeGenPrepare.cpp:1042) by 0x8844F60: (anonymous namespace)::CodeGenPrepare::runOnFunction(llvm::Function&) (CodeGenPrepare.cpp:85) by 0x8937C6B: llvm::FPPassManager::runOnFunction(llvm::Function&) (PassManager.cpp:1171) by 0x8937EC9: llvm::FunctionPassManagerImpl::run(llvm::Function&) (PassManager.cpp:1129) by 0x893801F: llvm::FunctionPassManager::run(llvm::Function&) (PassManager.cpp:1074) by 0x83C3692: main (llc.cpp:296) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 20 07:27:26 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 Jan 2008 07:27:26 -0600 Subject: [LLVMbugs] [Bug 1936] New: Segfault for unterminated macro invocation Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1936 Summary: Segfault for unterminated macro invocation Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: minor Priority: P2 Component: preprocessor AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu In the following testcase #define f(x) x #if f(2 #endif the newline at the end of the #if needs to terminate both the macro invocation and the directive, not just one of them. One of those unfortunate things to which there is not a satisfying solution, in my experience :) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 20 10:52:33 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 Jan 2008 10:52:33 -0600 Subject: [LLVMbugs] [Bug 1935] Use of freed memory in CodeGenPrepare::OptimizeBlock In-Reply-To: Message-ID: <200801201652.m0KGqX0g027398@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1935 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Duncan Sands 2008-01-20 10:52:33 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080114/057389.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 20 13:37:57 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 20 Jan 2008 13:37:57 -0600 Subject: [LLVMbugs] [Bug 1937] New: The StructReturn attribute should imply a copy Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1937 Summary: The StructReturn attribute should imply a copy Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu ByVal for parameters and StructReturn for results are closely related. They are both for passing structures by copy: into the function for ByVal, and out of the function for StructReturn. Yet while ByVal implies that a hidden copy is made, this is not the case for StructReturn - whoever generates the StructReturn has to create and use a temporary for the return value and explicitly copy it to the sret parameter at the end, see EmitMODIFY_EXPR in llvm-gcc (if you don't make a copy then you get wrong code). I think it would be better to have the same semantics as for byval: an implied copy is performed. On x86-32 the ABI results in a copy anyway (so here two copies are currently made), while on x86-64 a copy would need to be generated by the code generators. The arguments for an implicit copy are the same as for byval, and since they triumphed there they should triumph for struct-return too! Of course implementing this would mean: (1) modifying the inliner (2) adjust argpromotion (3) generating a copy for those targets that don't do so already for ABI reasons (4) whatever else needed to be done for byval It would also be possible to teach alias analysis to do a better job for sret and for byval. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 09:26:08 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 09:26:08 -0600 Subject: [LLVMbugs] [Bug 1938] New: SCCP mis-optimizes branches with undefined condition Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1938 Summary: SCCP mis-optimizes branches with undefined condition Product: libraries Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1345) --> (http://llvm.org/bugs/attachment.cgi?id=1345) test case The attached source is derived from this GCC test case: http://gcc.gnu.org/ml/gcc-patches/2003-04/msg00321.html If I run it through "opt -sccp", I get this: bb: ; preds = %bb.backedge, %entry %indvar = phi i32 [ 0, %entry ], [ %k, %bb.backedge ] ; %k = add i32 %indvar, 1 ; [#uses=3] br i1 undef, label %cond_true, label %cond_false ... cond_false: ; preds = %bb br i1 undef, label %bb.backedge, label %bb12 At run time, if bb branches to cond_false, and cond_false branches to bb12, then we'll reach bb12 with %k equal to 1, so the program will call abort(). Before SCCP ran there was no way to reach bb12 until %k was 10, so the program would never abort. So SCCP has broken it. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 13:41:53 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 13:41:53 -0600 Subject: [LLVMbugs] [Bug 1937] The StructReturn attribute should imply a copy In-Reply-To: Message-ID: <200801211941.m0LJfrSO022360@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1937 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #1 from Chris Lattner 2008-01-21 13:41:46 --- I don't think this makes sense. The major difference between byval and stret is that with stret the caller *must* allocate or provide the space for callee to use. If there was an implicit copy, the callee would be writing into some memory that the caller could not access. More significantly, exposing the destination buffer to the caller allows us to do various optimizations, including eliminate copy ctors in C++ (which llvm-g++ already does with "TARGET_EXPR"), reuse allocas for return values (we don't do this, but should someday) etc. I'm not sure what you mean by "On x86-32 the ABI results in a copy anyway (so here two copies are currently made)". -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 14:14:22 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 14:14:22 -0600 Subject: [LLVMbugs] [Bug 1926] llvm-gcc-4.2 does not use LLVM_VERSION_INFO In-Reply-To: Message-ID: <200801212014.m0LKEM6W024649@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1926 Tanya Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Tanya Lattner 2008-01-21 14:14:21 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057398.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 14:35:46 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 14:35:46 -0600 Subject: [LLVMbugs] [Bug 1939] New: LLVM type size doesn't match GCC type size! Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1939 Summary: LLVM type size doesn't match GCC type size! Product: tools Version: trunk Platform: PC OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: dpatel at apple.com CC: llvmbugs at cs.uiuc.edu --- b.c --- struct X { long double b; unsigned char c; double __attribute__((packed)) d; }; struct X x = { 3.0L, 5, 3.0 }; --- Assertion failed: ((!TYPE_SIZE(Tr) || !Ty->isSized() || !isInt64(TYPE_SIZE(Tr), true) || getInt64(TYPE_SIZE(Tr), true) == getTargetData().getABITypeSizeInBits(Ty)) && \ "LLVM type size doesn't match GCC type size!"), function llvm_set_type, file ../../llvmgcc42/gcc/llvm-types.cpp, line 82. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 18:29:51 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 18:29:51 -0600 Subject: [LLVMbugs] [Bug 1940] New: instcombine doesn't fold sgt(zext, zext) into ugt Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1940 Summary: instcombine doesn't fold sgt(zext, zext) into ugt Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: llvmbugs at cs.uiuc.edu define i1 @test(i8 %A, i8 %B) { %a = zext i8 %A to i32 %b = zext i8 %B to i32 %c = icmp sgt i32 %a, %b ret i1 %c } should become: define i1 @test(i8 %A, i8 %B) { %c = icmp ugt i32 %A, %B ret i1 %c } Similarly with the other compare-ops. The ugt(sext, sext) case should be optimized by removing the sext's. define i1 @test(i8 %A, i8 %B) { %a = sext i8 %A to i32 %b = sext i8 %B to i32 %c = icmp ugt i32 %a, %b ret i1 %c } should become: define i1 @test(i8 %A, i8 %B) { %c = icmp ugt i32 %A, %B ret i1 %c } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 23:20:40 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 23:20:40 -0600 Subject: [LLVMbugs] [Bug 1927] make clean gives errors In-Reply-To: Message-ID: <200801220520.m0M5Kejh015418@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1927 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-21 23:20:40 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057428.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 21 23:22:10 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 21 Jan 2008 23:22:10 -0600 Subject: [LLVMbugs] [Bug 1928] new warnings on darwin8 In-Reply-To: Message-ID: <200801220522.m0M5MANl015482@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1928 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Chris Lattner 2008-01-21 23:22:10 --- The APFloat issue appears to be a gcc bug or something. Eric says it's fixed (doesn't occur) with 4.2 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 22 01:01:40 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 Jan 2008 01:01:40 -0600 Subject: [LLVMbugs] [Bug 1941] New: instcombine doesn't merge mixed sign-unsign comparisons Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1941 Summary: instcombine doesn't merge mixed sign-unsign comparisons Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: llvmbugs at cs.uiuc.edu Instcombine will merge comparisons like (x >= 10) && (x < 20) by producing (x - 10) u< 10, but only when the comparisons have matching sign. This program: define i1 @test(i8 %x) { %A = icmp uge i8 %x, 5 %B = icmp slt i8 %x, 20 %C = and i1 %A, %B ret i1 %C } could be converted using the same technique. There's a lot of corner cases here that would need to be thought through thoroughly. (This was found by inspection, I doubt if any llvm-gcc generated input would trigger this.) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 22 05:49:20 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 Jan 2008 05:49:20 -0600 Subject: [LLVMbugs] [Bug 1236] vector of ints not supported by vector scalarizer In-Reply-To: Message-ID: <200801221149.m0MBnKEV015532@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1236 Bill Wendling changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Bill Wendling 2008-01-22 05:49:20 --- We now generate: .text .align 4,0x90 .globl _foo _foo: pushl %ebp movl %esp, %ebp movl 12(%ebp), %eax movq (%eax), %mm0 movl 8(%ebp), %eax movq %mm0, (%eax) emms popl %ebp ret .subsections_via_symbols with llvm-gcc 4.2 -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 22 11:45:54 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 Jan 2008 11:45:54 -0600 Subject: [LLVMbugs] [Bug 1731] [llvm-gcc 4.2] libgcc2.c:541: internal compiler error: Segmentation fault In-Reply-To: Message-ID: <200801221745.m0MHjs6m001964@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1731 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #9 from Chris Lattner 2008-01-22 11:45:53 --- Closing this as 'works for me', please reopen if it is still failing for you. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 22 13:36:07 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 Jan 2008 13:36:07 -0600 Subject: [LLVMbugs] [Bug 1936] Segfault for unterminated macro invocation In-Reply-To: Message-ID: <200801221936.m0MJa7ER010306@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1936 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |crash-on-invalid Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-22 13:36:06 --- Fixed, thanks Neil! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080121/003889.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 22 13:38:52 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 22 Jan 2008 13:38:52 -0600 Subject: [LLVMbugs] [Bug 1922] New: clang can't codegen pthread_mutex_t initializer Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1922 Summary: clang can't codegen pthread_mutex_t initializer Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu, snaroff at apple.com Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |snaroff at apple.com Created an attachment (id=1340) --> (http://llvm.org/bugs/attachment.cgi?id=1340) delta reduced testcase Using clang r46160, llvm r46160, codegen fails: $ clang -emit-llvm-bc others.i clang: CodeGenModule.cpp:304: llvm::Constant* GenerateAggregateInit(const clang::InitListExpr*, clang::CodeGen::CodeGenModule&): Assertion `(ILE->getType()->isArrayType() || ILE->getType()->isStructureType()) && "Bad type for init list!"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang[0x4a7636] clang[0x4a6835] clang(clang::CodeGen::CodeGenModule::EmitGlobalVar(clang::FileVarDecl const*)+0x80)[0x4a70a0] clang(clang::CodeGen::CodeGenModule::EmitGlobalVarDeclarator(clang::FileVarDecl const*)+0x27)[0x4a72f7] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted The original file contains this C code: # include static pthread_mutex_t cli_gentemp_mutex = PTHREAD_MUTEX_INITIALIZER; --- Comment #1 from Chris Lattner 2008-01-22 13:38:51 --- This is probably struct initialization related -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 23 06:46:29 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 Jan 2008 06:46:29 -0600 Subject: [LLVMbugs] [Bug 1942] New: problem with assigning returned object to *this Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1942 Summary: problem with assigning returned object to *this Product: tools Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1348) --> (http://llvm.org/bugs/attachment.cgi?id=1348) test case I'm using LLVM 2.1 with the pre-built "llvm-gcc4.0-2.1-x86-linux-RHEL4" GCC front end. The attached program ought to print "4 6", but when I run it I get: foad at debian:~$ ~/llvm/llvm-gcc4.0-2.1-x86-linux-RHEL4/bin/llvm-gcc -o ke ke.cpp foad at debian:~$ ./ke 3 4 I don't see this problem if I compile with a standard GCC 4.0.1 distribution, configured for i686-pc-linux-gnu. >From looking at the generated llvm code, the problem is that when foo::operator+= calls foo::operator+, it passes in "this" directly as the structure return pointer. That doesn't work because foo::operator+ writes to its return value before reading from its arguments. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 23 19:16:08 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 Jan 2008 19:16:08 -0600 Subject: [LLVMbugs] [Bug 1943] New: A potential single buffer overflow in program.inc for win32 Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1943 Summary: A potential single buffer overflow in program.inc for win32 Product: libraries Version: trunk Platform: PC OS/Version: Windows 2000 Status: NEW Severity: enhancement Priority: P2 Component: Archive library AssignedTo: unassignedbugs at nondot.org ReportedBy: humeafo at gmail.com CC: llvmbugs at cs.uiuc.edu // First, determine the length of the command line. unsigned len = 0; for (unsigned i = 0; args[i]; i++) { len += strlen(args[i]) + 1; if (strchr(args[i], ' ')) len += 2; } // Now build the command line. char *command = reinterpret_cast(_alloca(len)); // should use len+1 to fix this char *p = command; for (unsigned i = 0; args[i]; i++) { const char *arg = args[i]; size_t len = strlen(arg); bool needsQuoting = strchr(arg, ' ') != 0; if (needsQuoting) *p++ = '"'; memcpy(p, arg, len); p += len; if (needsQuoting) *p++ = '"'; *p++ = ' '; } *p = 0; // this may write beyond the boundary -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 23 19:21:54 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 23 Jan 2008 19:21:54 -0600 Subject: [LLVMbugs] [Bug 1943] A potential single buffer overflow in program.inc for win32 In-Reply-To: Message-ID: <200801240121.m0O1Ls5N023790@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1943 Anton Korobeynikov changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asl at math.spbu.ru Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Anton Korobeynikov 2008-01-23 19:21:53 --- Fixed, thanks. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057475.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 24 22:11:59 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 24 Jan 2008 22:11:59 -0600 Subject: [LLVMbugs] [Bug 1944] New: Scheduler deficiency on CodeGen/X86/pr1505b.ll Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1944 Summary: Scheduler deficiency on CodeGen/X86/pr1505b.ll Product: libraries Version: 2.2 Platform: PC OS/Version: All Status: NEW Keywords: code-quality Severity: enhancement Priority: P2 Component: Common Code Generator Code AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu Consider the output of llvm-as < pr1505b.ll | llc -mcpu=i486: call L_tan$stub fstpl 24(%esp) ... call L__ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc$stub fldl 24(%esp) fstps 36(%esp) flds 36(%esp) fstpl 4(%esp) 24(%esp) is a f80 vreg that is live across a call, so it is spilled. It's only use is a truncstore (fstps) to the stack, which is used to truncate it. The only use of this stack slot is the flds which is used to reextend it after the truncation is done. This is all a single MBB, so the scheduler should produce this code instead: call L_tan$stub fstps 36(%esp) ... call L__ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc$stub flds 36(%esp) fstpl 4(%esp) By issuing the store early and the load late, the register is not live across the call. This is a very simple store+load to a frame index pattern, the input chain to the store is the entry node, and the load depends on the store. I don't see any reason the scheduler should be prevented from doing this today. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 03:41:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 03:41:01 -0600 Subject: [LLVMbugs] [Bug 1945] New: can't catch exception thrown by destructor Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1945 Summary: can't catch exception thrown by destructor Product: tools Version: 2.1 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: jay.foad at antixlabs.com CC: llvmbugs at cs.uiuc.edu I'm using LLVM 2.1 on Linux/x86 with the pre-built "llvm-gcc4.0-2.1-x86-linux-RHEL4" GCC front end. If I build and run the attached program with LLVM, it terminates: foad at debian:~$ ~/llvm/llvm-gcc4.0-2.1-x86-linux-RHEL4/bin/llvm-g++ -o dd dd.cc foad at debian:~$ ./dd throw terminate Aborted It seems to have failed to catch the exception thrown by D's destructor, even though the destructor is run inside a try ... catch(...) block. If I build and run with a "normal" native GCC, it works: foad at debian:~$ g++ --version g++ (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) Copyright (C) 2006 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. foad at debian:~$ g++ -o dd dd.cc foad at debian:~$ ./dd throw catch return -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 03:55:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 03:55:03 -0600 Subject: [LLVMbugs] [Bug 1945] can't catch exception thrown by destructor In-Reply-To: Message-ID: <200801250955.m0P9t3sx031020@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1945 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |baldrick at free.fr Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Duncan Sands 2008-01-25 03:55:01 --- It works correctly with llvm-gcc-4.2 from svn: $ ./pr1945 throw catch return -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 04:25:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 04:25:43 -0600 Subject: [LLVMbugs] [Bug 1945] can't catch exception thrown by destructor In-Reply-To: Message-ID: <200801251025.m0PAPhGq032430@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1945 Jay Foad changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | Version|2.1 |2.2 --- Comment #3 from Jay Foad 2008-01-25 04:25:40 --- It passes with llvm-gcc-4.2-2.2-x86-linux-RHEL4 but it still fails with llvm-gcc-4.0-2.2-x86-linux-RHEL4 (using the prereleases annouced yesterday). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 04:40:07 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 04:40:07 -0600 Subject: [LLVMbugs] [Bug 1945] can't catch exception thrown by destructor In-Reply-To: Message-ID: <200801251040.m0PAe7Ag000363@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1945 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |WONTFIX --- Comment #4 from Duncan Sands 2008-01-25 04:40:06 --- I'm sorry, but this is not going to be fixed in 4.0 for this release. Also, this is going to be the last release using 4.0, so it is never going to be fixed in 4.0. Please use 4.2 instead - the exception handling support was completely rewritten for it and works much better than in 4.0 (as you have observed). This does need to be added to the release notes for 2.2 though. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 11:46:36 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 11:46:36 -0600 Subject: [LLVMbugs] [Bug 1942] problem with assigning returned object to *this In-Reply-To: Message-ID: <200801251746.m0PHkaDe014936@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1942 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Duncan Sands 2008-01-25 11:46:35 --- Fixed here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057556.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057559.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 13:19:47 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 13:19:47 -0600 Subject: [LLVMbugs] [Bug 1946] New: memcpy should be converted into llvm.memcpy.* Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1946 Summary: memcpy should be converted into llvm.memcpy.* Product: new-bugs Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1353) --> (http://llvm.org/bugs/attachment.cgi?id=1353) Patch Per description... llvm-gcc happens to do the conversion internally, but simplifylibcalls should be able to handle this itself. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 15:21:30 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 15:21:30 -0600 Subject: [LLVMbugs] [Bug 1947] New: LoopUnroll - use after free Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1947 Summary: LoopUnroll - use after free Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu Valgrind shows memory being used after being freed in the testcase test/Transforms/LoopUnroll/2007-11-05-Crash.ll: Invalid read of size 4 at 0x82D551A: llvm::Use::getNext() const (Use.h:65) by 0x82D62D7: llvm::value_use_iterator::operator++() (Use.h:130) by 0x83ED7EA: (anonymous namespace)::LoopUnroll::unrollLoop(llvm::LoopBase*, unsigned, unsigned) (LoopUnroll.cpp:367) by 0x83EE2FD: (anonymous namespace)::LoopUnroll::runOnLoop(llvm::LoopBase*, llvm::LPPassManager&) (LoopUnroll.cpp:182) by 0x84B375C: llvm::LPPassManager::runOnFunction(llvm::Function&) (LoopPass.cpp:225) by 0x857626F: llvm::FPPassManager::runOnFunction(llvm::Function&) (PassManager.cpp:1171) by 0x8576411: llvm::FPPassManager::runOnModule(llvm::Module&) (PassManager.cpp:1191) by 0x8575F45: llvm::MPPassManager::runOnModule(llvm::Module&) (PassManager.cpp:1240) by 0x85760FD: llvm::PassManagerImpl::run(llvm::Module&) (PassManager.cpp:1313) by 0x857614F: llvm::PassManager::run(llvm::Module&) (PassManager.cpp:1345) by 0x82E2C74: main (opt.cpp:426) Address 0x42e0db4 is 20 bytes inside a block of size 68 free'd at 0x4021EFC: operator delete[](void*) (vg_replace_malloc.c:364) by 0x85572E1: llvm::PHINode::resizeOperands(unsigned) (Instructions.cpp:176) by 0x831DA38: llvm::PHINode::addIncoming(llvm::Value*, llvm::BasicBlock*) (Instructions.h:1303) by 0x83ED7DC: (anonymous namespace)::LoopUnroll::unrollLoop(llvm::LoopBase*, unsigned, unsigned) (LoopUnroll.cpp:373) by 0x83EE2FD: (anonymous namespace)::LoopUnroll::runOnLoop(llvm::LoopBase*, llvm::LPPassManager&) (LoopUnroll.cpp:182) by 0x84B375C: llvm::LPPassManager::runOnFunction(llvm::Function&) (LoopPass.cpp:225) by 0x857626F: llvm::FPPassManager::runOnFunction(llvm::Function&) (PassManager.cpp:1171) by 0x8576411: llvm::FPPassManager::runOnModule(llvm::Module&) (PassManager.cpp:1191) by 0x8575F45: llvm::MPPassManager::runOnModule(llvm::Module&) (PassManager.cpp:1240) by 0x85760FD: llvm::PassManagerImpl::run(llvm::Module&) (PassManager.cpp:1313) by 0x857614F: llvm::PassManager::run(llvm::Module&) (PassManager.cpp:1345) by 0x82E2C74: main (opt.cpp:426) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Fri Jan 25 22:58:59 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Fri, 25 Jan 2008 22:58:59 -0600 Subject: [LLVMbugs] [Bug 1948] New: Local aggregate initializers Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1948 Summary: Local aggregate initializers Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: minor Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu Following recent great work on initializers, this is the only issue I could find: struct foo { int z; } w; int bar (void) { struct foo z = { w }; return z.z; } The above should reject the initializer, as w is not an "int". -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sat Jan 26 23:37:33 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sat, 26 Jan 2008 23:37:33 -0600 Subject: [LLVMbugs] [Bug 1949] New: instcombine doesn't fold away add/icmp ult Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1949 Summary: instcombine doesn't fold away add/icmp ult Product: libraries Version: trunk Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: llvmbugs at cs.uiuc.edu >From other optimizations, instcombine will produce this: define i1 @f(i32 %a) { %a.off = add i32 %a, 4 ; [#uses=1] %C = icmp ult i32 %a.off, 4 ; [#uses=1] ret i1 %C } which is actually equivalent to: define i1 @f(i32 %a) { %C = icmp ugt i32 %a.off, -5 ; [#uses=1] ret i1 %C } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 00:14:17 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 00:14:17 -0600 Subject: [LLVMbugs] [Bug 1769] Should remove the -cee pass In-Reply-To: Message-ID: <200801270614.m0R6EHgs026843@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1769 Bill Wendling changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |isanbard at gmail.com Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Bill Wendling 2008-01-27 00:14:17 --- Fixed: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057610.html http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057611.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 11:28:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 11:28:55 -0600 Subject: [LLVMbugs] [Bug 1931] Invalid division instcombine In-Reply-To: Message-ID: <200801271728.m0RHStnX015201@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1931 Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #6 from Eli Friedman 2008-01-27 11:28:54 --- Marking invalid, since the behavior of INT_MIN/-1 is undefined. (But note the doc fix attached to Bug 1933.) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 11:30:18 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 11:30:18 -0600 Subject: [LLVMbugs] [Bug 1934] Another invalid division instcombine In-Reply-To: Message-ID: <200801271730.m0RHUIPU015261@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1934 Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #2 from Eli Friedman 2008-01-27 11:30:18 --- Marking invalid, since the behavior of INT_MIN/-1 is undefined. (But note the doc fix attached to Bug 1933.) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 12:35:50 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 12:35:50 -0600 Subject: [LLVMbugs] [Bug 1947] LoopUnroll - use after free In-Reply-To: Message-ID: <200801271835.m0RIZoTf017300@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1947 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #1 from Nick Lewycky 2008-01-27 12:35:50 --- Fixed in r46417. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057623.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 13:35:41 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 13:35:41 -0600 Subject: [LLVMbugs] [Bug 1950] New: duplicate functions protos and __restrict throw bogus warning Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1950 Summary: duplicate functions protos and __restrict throw bogus warning Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1354) --> (http://llvm.org/bugs/attachment.cgi?id=1354) patch gcc treats pointer and pointer __restrict as the same when handling duplicate function protos. e.g.: int foo (__const char *__path); int foo(__const char *__restrict __file); -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 18:33:34 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 18:33:34 -0600 Subject: [LLVMbugs] [Bug 1938] SCCP mis-optimizes branches with undefined condition In-Reply-To: Message-ID: <200801280033.m0S0XYKV027977@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1938 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Target Milestone|--- |2.3 --- Comment #4 from Chris Lattner 2008-01-27 18:33:34 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057637.html Eli's fix is not quite correct because it would pessimize all undef processing. I think this is the minimal fix. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 18:59:11 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 18:59:11 -0600 Subject: [LLVMbugs] [Bug 1932] Invalid float division instcombine In-Reply-To: Message-ID: <200801280059.m0S0xBtS029016@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1932 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|enhancement |normal Status|NEW |RESOLVED Component|new bugs |Scalar Optimizations Keywords| |miscompilation Product|new-bugs |libraries Resolution| |FIXED Target Milestone|--- |2.3 Version|unspecified |2.2 --- Comment #2 from Chris Lattner 2008-01-27 18:59:10 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057639.html Nice catch! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 20:01:50 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 20:01:50 -0600 Subject: [LLVMbugs] [Bug 1948] Local aggregate initializers In-Reply-To: Message-ID: <200801280201.m0S21oME030973@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1948 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from snaroff at apple.com 2008-01-27 20:01:50 --- Fixed in r46430. Thanks Neil! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 21:49:41 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 21:49:41 -0600 Subject: [LLVMbugs] [Bug 1940] instcombine doesn't fold sgt(zext, zext) into ugt In-Reply-To: Message-ID: <200801280349.m0S3nfH0001501@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1940 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #9 from Nick Lewycky 2008-01-27 21:49:41 --- Fixed in 46431. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057640.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 22:34:47 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 22:34:47 -0600 Subject: [LLVMbugs] [Bug 1942] problem with assigning returned object to *this In-Reply-To: Message-ID: <200801280434.m0S4YlL9002793@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1942 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #4 from Chris Lattner 2008-01-27 22:34:44 --- This fix is not correct. Consider this testcase: class foo { public: int a[1000]; foo(void) {} const foo operator+(const foo& in) const; }; const foo foo::operator+(const foo& in) const { foo Out; Out.a[0] = 1; Out.a[1] = in.a[0]; return Out; } GCC compiles this code to: __ZNK3fooplERKS_: movl 4(%esp), %eax <- out pointer movl 12(%esp), %edx <- in pointer movl $1, (%eax) <- store to out[0] movl (%edx), %edx <- read in[0] movl %edx, 4(%eax) <- store to out[1] ret $4 This means that GCC is implicitly assuming that the return slot and the output cannot alias. I think this needs to be fixed on the caller side, not the callee side. If llvm-gcc is producing code where the caller assumes that it can alias the return slot with an argument, then the call produced will not match the ABI and would produce an incorrect result when GCC compiled the callee. What do you think Duncan? -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Sun Jan 27 22:42:06 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Sun, 27 Jan 2008 22:42:06 -0600 Subject: [LLVMbugs] [Bug 1946] memcpy should be converted into llvm.memcpy.* In-Reply-To: Message-ID: <200801280442.m0S4g6nI003087@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1946 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-27 22:42:05 --- Applied, thanks! http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057644.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 01:12:19 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 01:12:19 -0600 Subject: [LLVMbugs] [Bug 1951] New: Gratuitous renames of variables Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1951 Summary: Gratuitous renames of variables Product: tools Version: trunk Platform: PC OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: TableGen AssignedTo: unassignedbugs at nondot.org ReportedBy: scottm at aero.org CC: llvmbugs at cs.uiuc.edu Ran across this unique problem in tblgen in the CellSPU development. Note the "Tmp2" variable reference in the code: SDNode *Emit_0(const SDOperand &N, unsigned Opc0, unsigned Opc1, MVT::ValueType VT0, MVT::ValueType VT1) DISABLE_INLINE { SDOperand N0 = N.getOperand(0); SDOperand N00 = N0.getOperand(0); SDOperand N01 = N0.getOperand(1); SDOperand N1 = N.getOperand(1); SDOperand N10 = N1.getOperand(0); SDOperand N11 = N1.getOperand(1); SDOperand Tmp3(CurDAG->getTargetNode(Opc0, VT0, N00), 0); return CurDAG->SelectNodeTo(N.Val, Opc1, VT1, Tmp3, Tmp2); } I did manage to run this bug to ground -- tblgen was gratuitously renaming variables when no temporary was produced. Test case also attached as is a patch. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 11:30:12 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 11:30:12 -0600 Subject: [LLVMbugs] [Bug 1952] New: memcpy between locals at end of function not eliminated Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1952 Summary: memcpy between locals at end of function not eliminated Product: new-bugs Version: unspecified Platform: Other OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: new bugs AssignedTo: unassignedbugs at nondot.org ReportedBy: baldrick at free.fr CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1357) --> (http://llvm.org/bugs/attachment.cgi?id=1357) testcase .ll The testcase finishes with this memcpy between local allocas: call void @llvm.memcpy.i32( i8* %tmp161, i8* %tmp183185, i32 %tmp180, i32 4 ) ret void This is keeping the rest of the code alive... It would be nice to eliminate it. This is the result of compiling test/FrontendAda/placeholder.adb at -O4. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 15:21:55 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 15:21:55 -0600 Subject: [LLVMbugs] [Bug 1942] problem with assigning returned object to *this In-Reply-To: Message-ID: <200801282121.m0SLLtQn011120@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1942 Duncan Sands changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #7 from Duncan Sands 2008-01-28 15:21:52 --- Fixed in a gcc compatible way here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057694.html I don't plan to backport this to 4.0. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 18:17:14 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 18:17:14 -0600 Subject: [LLVMbugs] [Bug 1950] duplicate functions protos and __restrict throw bogus warning In-Reply-To: Message-ID: <200801290017.m0T0HEx0016225@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1950 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #7 from snaroff at apple.com 2008-01-28 18:17:13 --- Fixed in r46472. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 18:48:41 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 18:48:41 -0600 Subject: [LLVMbugs] [Bug 1939] LLVM type size doesn't match GCC type size! In-Reply-To: Message-ID: <200801290048.m0T0mfkG017496@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1939 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Devang Patel 2008-01-28 18:48:40 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080121/057407.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 19:10:25 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 19:10:25 -0600 Subject: [LLVMbugs] [Bug 1861] "LLVM type size doesn't match GCC type size!" In-Reply-To: Message-ID: <200801290110.m0T1APJA018407@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1861 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #5 from Devang Patel 2008-01-28 19:10:25 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057704.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 20:21:04 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 20:21:04 -0600 Subject: [LLVMbugs] [Bug 1912] ClamAV triggers assertion in loop index splitting pass In-Reply-To: Message-ID: <200801290221.m0T2L4M5020729@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1912 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #4 from Devang Patel 2008-01-28 20:21:03 --- Fixed. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057707.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 23:16:57 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 23:16:57 -0600 Subject: [LLVMbugs] [Bug 1953] New: Improve llvm-gcc complex code generation Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1953 Summary: Improve llvm-gcc complex code generation Product: tools Version: 2.2 Platform: PC OS/Version: All Status: NEW Severity: enhancement Priority: P2 Component: llvm-gcc AssignedTo: unassignedbugs at nondot.org ReportedBy: sabre at nondot.org CC: llvmbugs at cs.uiuc.edu llvm-gcc currently lowers code like this: void foo(float f, _Complex float *P) { _Complex float x = f; *P += x; } into two adds. It really should only do one add for performance and numerical correctness (as strongly recommended in annex G of C99). One way to do this would be to run the complex lowering pass which 4.2 already has, but this apparently requires "gimple ssa form", so this is not trivial. The other way to handle this is to hack llvm-convert to somehow do this. The problem is that the temporaries introduced by gimplification that are not in SSA form prevent doing correct dataflow analysis to propagate this information through. -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Mon Jan 28 23:53:49 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Mon, 28 Jan 2008 23:53:49 -0600 Subject: [LLVMbugs] [Bug 1951] tblgen gratuitously renames variables In-Reply-To: Message-ID: <200801290553.m0T5rn7U028193@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1951 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #6 from Chris Lattner 2008-01-28 23:53:49 --- I assume your recent commit implemented this, if not please reopen. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 00:40:59 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 00:40:59 -0600 Subject: [LLVMbugs] [Bug 1952] memcpy between locals at end of function not eliminated In-Reply-To: Message-ID: <200801290640.m0T6exch029655@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1952 Owen Anderson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Owen Anderson 2008-01-29 00:40:58 --- Fixed as of r46494. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 07:19:01 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 07:19:01 -0600 Subject: [LLVMbugs] [Bug 1954] New: spurious "declaration does not declare anything" warnings Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1954 Summary: spurious "declaration does not declare anything" warnings Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: enhancement Priority: P2 Component: parser AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu The following code produces a "declaration does not declare anything" warning. Stripped test case from /usr/include/bits/pthreadtypes.h (linux+glibc+NPTL) typedef union { struct __pthread_mutex_s { union { int a; double b; }; } __data; char c; } pthread_mutex_t; the error is generated in ParseDecl.cpp (diag::w_no_declarators) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 07:52:22 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 07:52:22 -0600 Subject: [LLVMbugs] [Bug 1955] New: complex typeof expression produce invalid result type Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1955 Summary: complex typeof expression produce invalid result type Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1360) --> (http://llvm.org/bugs/attachment.cgi?id=1360) test case The test case provided in attach generates the following error: "invalid operands to binary expression". The code comes from the inline glibc version of hypot(). Sorry for the big test case, but I didn't have the time to strip it further. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 11:16:46 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 11:16:46 -0600 Subject: [LLVMbugs] [Bug 1956] New: Sema+Codegen fail to use '((struct foo*)0)->m' as initializer Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1956 Summary: Sema+Codegen fail to use '((struct foo*)0)->m' as initializer Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: enhancement Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu The following code throws some errors in both the Sema and Codegen code: struct x { struct x *m; }; static const struct x a[] = { { (((char *) (&(((struct x*)((void *)0))->m))) - ((char *) 0)) } }; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 13:57:09 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 13:57:09 -0600 Subject: [LLVMbugs] [Bug 1957] New: initialization of a __complex__ number triggers an assert() Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1957 Summary: initialization of a __complex__ number triggers an assert() Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu the following code triggers an assertion in the codegen module (run with -emit-llvm): double __complex__ z = 0.0; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 14:32:32 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 14:32:32 -0600 Subject: [LLVMbugs] [Bug 1958] New: Assertion when Module Pass requires Function Pass Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1958 Summary: Assertion when Module Pass requires Function Pass 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: koyrehme.bugs at gmail.com CC: llvmbugs at cs.uiuc.edu Created an attachment (id=1361) --> (http://llvm.org/bugs/attachment.cgi?id=1361) Simple module pass requiring LoopInfo The attached pass generates the following assertion: opt -disable-output -disable-verify -debug-pass=Details -load ./llvm/Debug/lib/TEST.so -TEST test.bc Pass Arguments: -TEST Target Data Layout ModulePass Manager A TEST PASS Unnamed pass: implement Pass::getPassName() -- A TEST PASS 0x86d00e0 Executing Pass 'A TEST PASS' on Module '/home/koy/research/test/testcases/relationships2.bc'... 0x86cd940 Required Analyses: Natural Loop Construction Pass Arguments: -domtree -loops FunctionPass Manager Dominator Tree Construction Natural Loop Construction opt: {snip}/llvm/include/llvm/PassAnalysisSupport.h:226: AnalysisType& llvm::Pass::getAnalysisID(const llvm::PassInfo*, llvm::Function&) [with AnalysisType = llvm::LoopInfo]: Assertion `ResultPass && "getAnalysis*() called on an analysis that was not " "'required' by pass!"' failed. The documentation (http://llvm.org/docs/WritingAnLLVMPass.html#ModulePass) says that this should be possible. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 16:42:37 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 16:42:37 -0600 Subject: [LLVMbugs] [Bug 1887] Error with bitfield codegen In-Reply-To: Message-ID: <200801292242.m0TMgbRo001875@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1887 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dpatel at apple.com Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Devang Patel 2008-01-29 16:42:37 --- Now clang emits following for last example. %struct.Test = type { i16, i8, i8, i8, i8, i8, i8, i32 } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 17:23:51 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 17:23:51 -0600 Subject: [LLVMbugs] [Bug 1886] Crash with incomplete struct initializer In-Reply-To: Message-ID: <200801292323.m0TNNpx9003247@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1886 Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dpatel at apple.com Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Devang Patel 2008-01-29 17:23:51 --- Fixed. http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080128/003991.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 17:33:33 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 17:33:33 -0600 Subject: [LLVMbugs] [Bug 1923] New: clang: Assertion `V.size() == T->getNumElements() && " Invalid initializer vector for constant structure"' Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1923 Summary: clang: Assertion `V.size() == T->getNumElements() && "Invalid initializer vector for constant structure"' Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: RESOLVED Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: dpatel at apple.com, llvmbugs at cs.uiuc.edu Devang Patel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dpatel at apple.com Status|NEW |RESOLVED Resolution| |FIXED Created an attachment (id=1341) --> (http://llvm.org/bugs/attachment.cgi?id=1341) delta reduced testcase Using clang r46160, codegen fails: $ clang rtf.i -emit-llvm-bc clang: Constants.cpp:367: llvm::ConstantStruct::ConstantStruct(const llvm::StructType*, const std::vector >&): Assertion `V.size() == T->getNumElements() && "Invalid initializer vector for constant structure"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang[0x59b37c] clang(llvm::ConstantStruct::get(llvm::StructType const*, std::vector > const&)+0x1e7)[0x5a5bc7] clang[0x4a75b0] clang[0x4a6835] clang(clang::CodeGen::CodeGenModule::EmitGlobalVar(clang::FileVarDecl const*)+0x80)[0x4a70a0] clang(clang::CodeGen::CodeGenModule::EmitGlobalVarDeclarator(clang::FileVarDecl const*)+0x27)[0x4a72f7] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted The attached testcase gives no warnings with gcc (it could be reduced even further, but then gcc would give warnings) I have used the following script as delta's -test parameter: clang -emit-llvm-bc $1 2>&1 | grep "Invalid initializer vector for constant structure" if ! test $? == 0; then exit 1 fi gcc -Wall -Wfatal-errors -fsyntax-only $1 if ! test $? == 0; then exit 1 fi exit 0 --- Comment #1 from Devang Patel 2008-01-29 17:33:33 --- This is now fixed. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Tue Jan 29 18:47:07 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Tue, 29 Jan 2008 18:47:07 -0600 Subject: [LLVMbugs] [Bug 1908] variable redefiniton error In-Reply-To: Message-ID: <200801300047.m0U0l7ZF006016@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1908 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from snaroff at apple.com 2008-01-29 18:47:07 --- This is fixed by r46540. I decided not to use patch http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-January/000930.html (it was too simplistic). In any event, thanks for the bug! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 01:36:45 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 01:36:45 -0600 Subject: [LLVMbugs] [Bug 1959] New: -dse mysteriously eliminates "free" instructions Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1959 Summary: -dse mysteriously eliminates "free" instructions Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Scalar Optimizations AssignedTo: unassignedbugs at nondot.org ReportedBy: nicholas at mxc.ca CC: resistor at mac.com, llvmbugs at cs.uiuc.edu Created an attachment (id=1362) --> (http://llvm.org/bugs/attachment.cgi?id=1362) testcase Attached is a reduced testcase from a failure in SingleSource/Benchmarks/Shootout/lists. It looks like DSE eliminated the free's that matched up with the malloc's. This lead to a quick out of memory condition when running the resulting program. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:30:37 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:30:37 -0600 Subject: [LLVMbugs] [Bug 1960] New: Complex integer codegen broken Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1960 Summary: Complex integer codegen broken Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu Simple test: void a() { __complex__ long long v = 2; } (from http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-January/000895.html) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:32:50 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:32:50 -0600 Subject: [LLVMbugs] [Bug 1961] New: Test failures with clang objc tests Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1961 Summary: Test failures with clang objc tests Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: Documentation AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu (From http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-January/000916.html) On my machine (Ubuntu Linux), all of the objc tests cause an error like the following: Command: clang -rewrite-test Sema/rewrite-foreach-4.m | clang Output: In file included from :23: /usr/lib/gcc/i486-linux-gnu/4.1.3/include/objc/objc.h:145:3: error: redefinition of 'Protocol' } Protocol; ^ :7:28: error: previous definition is here typedef struct objc_object Protocol; ^ 2 diagnostics generated. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:35:34 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:35:34 -0600 Subject: [LLVMbugs] [Bug 1962] New: clang struct init codegen + padding broken Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1962 Summary: clang struct init codegen + padding broken Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu struct s { short a; int b; } x = {1, 1}; (From http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-January/000941.html; putting here so it doesn't get lost.) -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:42:27 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:42:27 -0600 Subject: [LLVMbugs] [Bug 1918] clang asserts when including gcc's C99 math.h header. In-Reply-To: Message-ID: <200801301642.m0UGgRZG012174@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1918 Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sharparrow1 at yahoo.com Status|NEW |RESOLVED Resolution| |WORKSFORME --- Comment #2 from Eli Friedman 2008-01-30 10:42:27 --- Seems to be working fine; reopen if there are still issues. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:47:44 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:47:44 -0600 Subject: [LLVMbugs] [Bug 1921] New: clang fails with: Assertion `Ty && " Invalid GetElementPtrInst indices for type!"' Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1921 Summary: clang fails with: Assertion `Ty && "Invalid GetElementPtrInst indices for type!"' Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: edwintorok at gmail.com CC: llvmbugs at cs.uiuc.edu, sharparrow1 at yahoo.com Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sharparrow1 at yahoo.com Created an attachment (id=1339) --> (http://llvm.org/bugs/attachment.cgi?id=1339) delta reduced testcase Using clang r46160, llvm r46160, codegen fails for attached reduced testcase $ clang -emit-llvm-bc matcher-bm.i clang: /home/edwin/llvm-svn/llvm/include/llvm/Instructions.h:374: const llvm::Type* llvm::checkType(const llvm::Type*): Assertion `Ty && "Invalid GetElementPtrInst indices for type!"' failed. clang[0x65292b] /lib/libc.so.6[0x363d232090] /lib/libc.so.6(gsignal+0x35)[0x363d232025] /lib/libc.so.6(abort+0x110)[0x363d233a80] /lib/libc.so.6(__assert_fail+0xef)[0x363d22b42f] clang(llvm::GetElementPtrInst::GetElementPtrInst(llvm::Value*, llvm::Value*, std::basic_string, std::allocator > const&, llvm::Instruction*)+0x180)[0x5d1f90] clang[0x4be233] clang[0x4c23f7] clang[0x4c076c] clang(clang::CodeGen::CodeGenFunction::EmitScalarExpr(clang::Expr const*)+0x58)[0x4c1478] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&)+0x17a)[0x4c66da] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitWhileStmt(clang::WhileStmt const&)+0x187)[0x4c6337] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitIfStmt(clang::IfStmt const&)+0x17a)[0x4c66da] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::EmitForStmt(clang::ForStmt const&)+0x211)[0x4c5cd1] clang(clang::CodeGen::CodeGenFunction::EmitCompoundStmt(clang::CompoundStmt const&, bool, llvm::Value*, bool)+0x71)[0x4c69f1] clang(clang::CodeGen::CodeGenFunction::EmitStmt(clang::Stmt const*)+0x2a5)[0x4c4e45] clang(clang::CodeGen::CodeGenFunction::GenerateCode(clang::FunctionDecl const*)+0x2ed)[0x4ae5ed] clang(clang::CodeGen::CodeGenModule::EmitFunction(clang::FunctionDecl const*)+0x33)[0x4a48d3] clang(clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, bool)+0x3f8)[0x4e5fd8] clang[0x49b009] clang(main+0x9fa)[0x4a0cfa] /lib/libc.so.6(__libc_start_main+0xf4)[0x363d21e1c4] clang[0x481589] Aborted $ clang --version Low Level Virtual Machine (http://llvm.org/): llvm version 2.3svn Optimized build with assertions. $ uname -a Linux lightspeed2 2.6.24-rc6-ge697789d #3 Wed Jan 2 11:15:05 EET 2008 x86_64 GNU/Linux --- Comment #1 from Eli Friedman 2008-01-30 10:47:44 --- Here's a really reduced testcase: int cli_bm_scanbuff() { const unsigned char *bp; bp -= (short)1; } -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 10:51:38 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 10:51:38 -0600 Subject: [LLVMbugs] [Bug 1891] Addres of unnamed array not recognized as constant In-Reply-To: Message-ID: <200801301651.m0UGpc8k012482@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1891 Eli Friedman changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #6 from Eli Friedman 2008-01-30 10:51:37 --- Since the original bug as filed is fixed, I might as well mark this fixed. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 11:04:54 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 11:04:54 -0600 Subject: [LLVMbugs] [Bug 1959] -dse mysteriously eliminates "free" instructions In-Reply-To: Message-ID: <200801301704.m0UH4swo012953@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1959 Nick Lewycky changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #2 from Nick Lewycky 2008-01-30 11:04:53 --- Fixed. Patch here: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20080128/057758.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 12:21:14 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 12:21:14 -0600 Subject: [LLVMbugs] [Bug 1922] clang can't codegen pthread_mutex_t initializer In-Reply-To: Message-ID: <200801301821.m0UILEpR015606@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1922 T??r??k Edwin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from T??r??k Edwin 2008-01-30 12:21:14 --- (In reply to comment #2) > Hmm, this doesn't crash for me; does it crash on an unpatched build? > This crash no longer occurs for me either, and the original others.c compiles now. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 16:56:57 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 16:56:57 -0600 Subject: [LLVMbugs] [Bug 1963] New: Addr of functions not considered a constant expr Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1963 Summary: Addr of functions not considered a constant expr Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu I think this is a known issue, but adding a bug just in case. The following is fine C90 and C99. extern void efunc (void); void (*addr_extern_func) (void) = &efunc; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 17:47:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 17:47:43 -0600 Subject: [LLVMbugs] [Bug 1885] Cannot forward-declare static array In-Reply-To: Message-ID: <200801302347.m0UNlhNF027064@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1885 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 17:48:41 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 17:48:41 -0600 Subject: [LLVMbugs] [Bug 1867] typedef-ing the same type twice in different headers In-Reply-To: Message-ID: <200801302348.m0UNmfpR027152@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1867 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #5 from snaroff at apple.com 2008-01-30 17:48:41 --- Fixed in r46583. Thanks! Please test (and let me know if you have any problems). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 19:35:43 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 19:35:43 -0600 Subject: [LLVMbugs] [Bug 1964] New: constant expression error Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1964 Summary: constant expression error Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: csaba.hruska at gmail.com CC: llvmbugs at cs.uiuc.edu clang segfaults from this code: #define CONST 1 struct a {char b;}; struct a c = {CONST+5}; -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 22:13:40 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 22:13:40 -0600 Subject: [LLVMbugs] [Bug 1921] clang codegen crash on ptr-short In-Reply-To: Message-ID: <200801310413.m0V4DeYl002863@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1921 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED Summary|clang fails with: Assertion |clang codegen crash on ptr- |`Ty && "Invalid |short |GetElementPtrInst indices | |for type!"' | --- Comment #2 from Chris Lattner 2008-01-30 22:13:39 --- Nice reduction Eli, fixed! http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080128/004035.html -Chris -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Wed Jan 30 23:03:03 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Wed, 30 Jan 2008 23:03:03 -0600 Subject: [LLVMbugs] [Bug 1965] New: ParseParenDecl not catching bad cases Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1965 Summary: ParseParenDecl not catching bad cases Product: clang Version: unspecified Platform: All OS/Version: All Status: NEW Severity: critical Priority: P2 Component: parser AssignedTo: unassignedbugs at nondot.org ReportedBy: natebegeman at mac.com CC: llvmbugs at cs.uiuc.edu Clang eats all of these just fine: int bar(, ); int foo(, int a); int baz(int a, ); And actually creates "int" arguments for the empty spaces. GCC rejects these. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 00:11:04 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 00:11:04 -0600 Subject: [LLVMbugs] [Bug 1965] missing error for parameters missing type specifiers In-Reply-To: Message-ID: <200801310611.m0V6B4ma006607@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1965 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Keywords| |accepts-invalid Resolution| |FIXED Summary|ParseParenDecl not catching |missing error for parameters |bad cases |missing type specifiers --- Comment #2 from Chris Lattner 2008-01-31 00:11:03 --- Fixed: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080128/004040.html -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 07:23:41 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 07:23:41 -0600 Subject: [LLVMbugs] [Bug 1966] New: -pedantic-errors unusable Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1966 Summary: -pedantic-errors unusable Product: clang Version: unspecified Platform: PC OS/Version: NetBSD Status: NEW Severity: major Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: neil at daikokuya.co.uk CC: llvmbugs at cs.uiuc.edu Clang defines macros indicating it is GCC, so it pulls in GCC-specific declarations in system header files. However, it then rejects them with stuff like: /usr/include/sys/cdefs_aout.h:26:5: error: extension used __asm(".global " _C_LABEL_STRING(#alias) "\n" \ ^ This makes -pedantic-errors not useful for real code. It should either a) not claim to be GCC in pedantic-errors mode, or b) accept this stuff in system headers even with --pedantic-errors Note that in this example there is no need to reject the __asm construct at all as it is an identifier reserved for the implementation. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 10:36:06 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 10:36:06 -0600 Subject: [LLVMbugs] [Bug 1967] New: sema crashes on if() without then Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1967 Summary: sema crashes on if() without then Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Keywords: crash-on-invalid Severity: normal Priority: P2 Component: Semantic Analyzer AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu clang crashes with invalid code (if without then clause): void x() { if (1) { -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 10:57:09 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 10:57:09 -0600 Subject: [LLVMbugs] [Bug 1968] New: clang codegen for static forward declaration broken Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1968 Summary: clang codegen for static forward declaration broken Product: clang Version: unspecified Platform: PC OS/Version: Linux Status: NEW Severity: enhancement Priority: P2 Component: LLVM Codegen AssignedTo: unassignedbugs at nondot.org ReportedBy: sharparrow1 at yahoo.com CC: llvmbugs at cs.uiuc.edu static int array[]; static int array[4]; produces @array = internal global [0 x i32] zeroinitializer ; <[0 x i32]*> [#uses=0] @array1 = internal global [4 x i32] zeroinitializer ; <[4 x i32]*> [#uses=0] with clang -emit-llvm... -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 11:21:09 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 11:21:09 -0600 Subject: [LLVMbugs] [Bug 1969] New: HeaderSearch::LookupSubframeworkHeader() crashes when input comes from stdin Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1969 Summary: HeaderSearch::LookupSubframeworkHeader() crashes when input comes from stdin Product: clang Version: unspecified Platform: PC OS/Version: Windows NT Status: NEW Severity: enhancement Priority: P2 Component: preprocessor AssignedTo: unassignedbugs at nondot.org ReportedBy: nunoplopes at sapo.pt CC: llvmbugs at cs.uiuc.edu clang crashes when a objc file is given by stdin. this make the whole objc rewrite test suite crash on my pc. patch: Index: Lex/HeaderSearch.cpp =================================================================== --- Lex/HeaderSearch.cpp (revision 46612) +++ Lex/HeaderSearch.cpp (working copy) @@ -296,6 +296,9 @@ // Framework names must have a '/' in the filename. Find it. const char *SlashPos = std::find(FilenameStart, FilenameEnd, '/'); if (SlashPos == FilenameEnd) return 0; + + if (ContextFileEnt == 0) + return 0; // Look up the base framework name of the ContextFileEnt. const char *ContextName = ContextFileEnt->getName(); -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 12:29:48 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 12:29:48 -0600 Subject: [LLVMbugs] [Bug 1967] sema crashes on if() without then In-Reply-To: Message-ID: <200801311829.m0VITmDQ004189@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1967 snaroff at apple.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from snaroff at apple.com 2008-01-31 12:29:48 --- Fixed in r46616. -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 14:10:17 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 14:10:17 -0600 Subject: [LLVMbugs] [Bug 1970] New: CBackend doesn't handle properly unaligned load/stores Message-ID: http://llvm.org/bugs/show_bug.cgi?id=1970 Summary: CBackend doesn't handle properly unaligned load/stores Product: libraries Version: trunk Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: Backend: C AssignedTo: unassignedbugs at nondot.org ReportedBy: lauro.venancio at gmail.com CC: llvmbugs at cs.uiuc.edu It makes some CBE tests fail on ARM (for example Applications/ClamAV/clamscan and Benchmarks/VersaBench/dbms/dbms). -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. From bugzilla-daemon at cs.uiuc.edu Thu Jan 31 23:34:36 2008 From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu) Date: Thu, 31 Jan 2008 23:34:36 -0600 Subject: [LLVMbugs] [Bug 1969] HeaderSearch::LookupSubframeworkHeader() crashes when input comes from stdin In-Reply-To: Message-ID: <200802010534.m115YaaK023742@zion.cs.uiuc.edu> http://llvm.org/bugs/show_bug.cgi?id=1969 Chris Lattner changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|enhancement |minor Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from Chris Lattner 2008-01-31 23:34:35 --- Fixed, patch here: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080128/004053.html Please verify, thanks! -- Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.