1 2 Bugs reported after the release of 3.7.0 3 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 4 5 This file lists bugs that were reported after 3.7.0 was released, both 6 in 3.7.0 and trunk, and that have not been closed. Bugs that have 7 been closed as "fixed" are listed in the top level NEWS file instead. 8 In other words they get listed here first and moved to NEWS if they 9 get fixed. 10 11 -- reenable amd64_toIR.c commented out insns 12 -- amd64 toIR memcheck icc -O3 -mAVX probs? 13 -- add arm stack-scan hacks? 14 -- add allocator-mismatch checking? 15 -- 302709, minor tidying needed? 16 -- reconsider https://bugs.kde.org/show_bug.cgi?id=253657 (AMD64 seg regs) 17 -- make sure it works OK on Ubuntu 12.04 and Fedora latest (glibc 2.16) 18 -- carll 2.11 Limitations, what insn sets does ppc now support 19 -- ditto 2.11 Limitations for MIPS 20 21 284540 Memcheck shouldn't count suppressions matching 22 still-reachable allocations 23 wishlist, w/patch, unclear if important 24 25 284904 Avoid cycles by disabling inclusive costs for cycles? 26 wishlist 27 28 285780 Nice to be able to display (animate) movement of variables through caches 29 wishlist 30 31 286769 replace __intel_new_memcpy with valgrind's memcpy 32 not started 33 34 286849 Interceptors for new/delete on Darwin were erroneously commented 35 out in r12043 36 **possible 3.8.0 37 38 286864 strlen function redirection error 39 **probably invalid 40 41 288099 valgrind should print arguments/return values of functions in stack 42 wishlist 43 44 287862 MPI_IN_PLACE not supported for MPI collect 45 libmpiwrap.c problem; low prio 46 47 288274 Aborts at illegal instruction 0x48 0xF 0x5A 0x7 0x48 0xF 48 == 278744 REX.W CVTPS2PD 49 **possible 3.8.0 (redundant REX prefix) 50 51 288363 valgrind: m_libcfile.c:73 (vgPlain_safe_fd): Assertion 52 'newfd >= VG_(fd_hard_limit)' failed. 53 OSX weirdness; QUERIED (maybe invalid) 54 55 288756 add '-x' to strip when stripping dynamic libraries on OSX 56 OSX. low prio 57 58 289360 parse_type_DIE confused by DW_TAG_enumeration_type 59 low prio; --read-var-info=yes problem; OP queried 60 61 289578 Backtraces with unwind tables (=without debug symbols) 62 using offline symbol resolving 63 wishlist 64 65 289836 Unhanded instructions mfatbu and mfatbl 66 ppc e500 insns 67 68 289912 link_tool_exe_darwin should not use relative paths 69 unsupported build configuration; low prio 70 71 290061 pie elf always loaded at 0x108000 72 w/ patch; looks OK; easy fix ? 73 not sure this wouldn't break something else -- needs looking into 74 75 290066 Warning messages from libraries with Valgrind 3.7.0 76 ppc elfery wierdness 77 **revisit -- related to some other bug re ELF problems (296318)? 78 79 290384 Valgrind does not show line numbers 80 on 32 bit FC16 81 **uninvestigated 82 83 290673 wine+valgrind: failed to start tool 'memcheck' for platform 84 'x86-linux': No such file or directory 85 **probably invalid; build config problem? 86 87 291924 vex x86->IR: unhandled instruction bytes: 0x66 0xF 0x38 0x31 88 invalid (SSE4) (WONTFIX) 89 90 292281 Not supported LL cache size of AMD Opteron 6176 91 hmm, more cache size kludgery required 92 **need access to the machine to fix 93 94 293855 vex amd64->IR: 0x2F 0x55 0x73 0x65 0x72 0x73 0x2F 0x6A 95 some problem with interceptors on Helgrind/OSX 96 97 294065 valgrind running wine does not find pdb file because of wrong 98 path to egrep 99 w/ somewhat dodgy patch 100 101 294147 Exception in thread "main" java 102 probably invalid 103 not sure what is going on here 104 105 294285 --partial-loads-ok does not work for 16-byte SSE loads 106 true 107 108 294454 delphi uses 16-bit addressing in the TEB (%fs) 109 obscure (16 bit (0x67 ?) insns in 32 bit mode) 110 111 294861 ./configure fails to locate Boost (located in /usr/local) 112 low prio 113 114 295084 Hard-coded /usr/include 115 OSX, XCode 4.3 problem 116 117 251569 rdtscp not supported: vex amd64->IR: 0xF 0x1 0xF9 0x8B 0x4C 0x24 118 has doesn't-really-look-right patch 119 120 295403 Memory access below SP with some STRD instructions. 121 ARM, fiddly to fix/verify 122 123 295472 "executable has wrong endian-ness" when target executable has 124 correct endianess 125 cross-compiling x86->ppc32 weirdness 126 127 295514 unhandled instruction bytes : 0x66 0x9C (pushfw on x86_64) 128 not high prio 129 **possible 3.8.0 (easy to fix?) 130 131 295808 vex amd64->IR: 0xF3 0xF 0xBC 0xC0 0x48 0x1 0xD0 0x48 (TZCNT) 132 **possible 3.8.0 ? see also 301011 (32 bit) 133 134 295974 vex-x86 omits some SSE4.1 opcodes such as PEXTRD 135 **check before 3.8.0; maybe some 64-bit insns missing? OP queried 136 137 296166 Add support for ARM in ptrcheck 138 **no patch yet 139 140 296308 helgrind tests fail on Linux/x86 due to clone syscall 141 missing unwind info 142 **possible 3.8.0; has low risk patch 143 144 296311 Wrong stack traces due to -fomit-frame-pointer (x86) 145 **not sure what to do about this. Is it serious? 146 147 296480 valgrind hangs early when run on Python 148 OSX. Not started. Queried. 149 (might be == 301281, 284004, but not sure) 150 151 296577 vex x86->IR: 0x66 0xF 0x3A 0x17 152 **wontfix (AES on 32-bit) 153 154 296578 vex amd64->IR: 0x60 0x0 0x0 0x0 0x0 0x0 0x0 0x0 155 **no idea what this is 156 157 296947 vex amd64->IR: 0x1F 0x7 0x0 0x0 0x0 0x0 158 probably invalid 159 160 297407 Failed reading debug information for Fortran codes; invalid cuOff 161 Tromey made a patch for this. Dunno if it should land. Queried. 162 163 298097 on ppc32 (e500mc core) target, valgrind report a app a 164 mem error but the app is good to me. 165 probably closable (may be dup of 296318) 166 167 298546 wishlist: track virtual memory leaks due to mmap without munmap 168 wishlist 169 170 298905 argv / argc to be changed to non valgrind values. 171 wishlist 172 173 300102 memcheck tester 174 unclear what status is 175 176 301011 vex x86->IR: 0xF3 0xF 0xBC 0xDE (TZCNT) 177 **possible 3.8.0; see also 295808 (64 bit) 178 179 301269 optimize Memcheck to use light tracking for trusted code 180 too big a change for 3.8.0; + in development still 181 182 301336 Valgrind vassert(0) in dis_neon_data_2reg_misc() 183 **possible 3.8.0. OP queried; is this the same as 287175 ? 184 185 301830 mtV: improve Valgrind so that it runs multiple threads in parallel 186 not 3.8.0 (alas) 187 188 301922 gcc -O2 problem with SSE optimization? 189 invalid (versioning problem) .. should close 190 191 301967 vex x86->IR: 0xC5 0xF9 0x6E 0x40 (AVX in 32-bit mode) 192 probable wontfix 193 194 302234 fesetround() function does not affect FPU calculations 195 x86 rounding; dup of 136779 (which has a patch!) 196 197 302271 Bug with Callgrind dump-instr and Trace-children 198 unclear what status is; queried JosefW 199 200 302630 Memcheck on multithreaded program fails with Assertion 201 'sizeof(UWord) == sizeof(UInt)' failed in m_syscall.c 202 configuration error, very weird 203 204 303460 Killed by fatal signal at doRegisterAllocation 205 queried OP (gentoo weirdness) 206 207 -- Tues 17 July 2012 11am 208 209 (transitioned to 3_8_BUGSTATUS after this point) 210
