1 2Created Mon Sep 8 13:22:16 CEST 2014. All new bugs after this point 3go in here. 4 5=== VEX/amd64 ========================================================== 6 7339218 AVX2 binaries compiled with the Intel compiler fail to 8 run under Valgrind 3.10.0 9 FIXME: should we enable this now? 10 == 338602 11 12338602 AVX2 bit in CPUID missing 13 need to do XSAVE, XRSTOR 14 15339596 vex amd64->IR: 0x8F 0xE8 0x78 0xCD 0xC1 0x4 0xC5 0xF9 16 XOP/FMA -- should take 17 18339820 vex amd64->IR: 0x66 0xF 0x3A 0x63 0xA 0x42 0x74 0x9 19 (pcmpistrix $0x42, (%rdx), %xmm1) 20 Should fix. Reported against 3.8.1 -- still valid? 21 22340469 vex amd64->IR: unhandled instruction bytes: 0x66 0xA7 23 Not clear what's going on here 24 25342571 Valgrind chokes on AVX compare intrinsic with _CMP_GE_QS 26 probably easy to fix 27 28342988 INT 80 not recognised on amd64 29 30=== VEX/arm ============================================================ 31 32335618 mov.w rN, pc/sp (ARM32) 33 HAS PATCH 34 35342680 arm: unhanded instruction 0xF590F000: pldw [r0] 36 37342780 arm + gcc 4.9 produce false positive: Use of uninitialised value of 38 size 4 and segfault on stack extention 39 Has patch 40 41342783 arm: unhandled instruction 0xEEFE1ACA = "vcvt.s32.f32 s3, s3, #12" 42 43344802 disInstr(arm): unhandled instruction: 0xEC510F1E 44 maybe easy fix? 45 but doesn't run natively either 46 47=== VEX/arm64 ========================================================== 48 49=== VEX/x86 ============================================================ 50 51339416 unhandled instruction bytes 0xC5 0xF1 0xEF 0xC9 52 (possible AVX on 32-bit, can't tell) 53 54340849 unsupported AVX instruction on 32 bit 55 we should WONTFIX this 56 57342192 Unhandled instruction on OS X 10.9 in libsystem_m.dylib 58 possibly a wontfix 59 60=== VEX/mips =========================================================== 61 62340777 Illegal instruction on mips (ar71xx) 63 64341997 MIPS64: Cavium OCTEON instructions - immediate operand handled 65 incorrectly 66 67=== VEX/ppc ============================================================ 68 69342610 disInstr(ppc): declined to decode an AltiVec insn. 70 71=== VEX/s390x ========================================================== 72 73342841 s390x unrecognized instruction fiebra 74 -> Florian? 75 76=== VEX general ======================================================== 77 78339778 Linux/TileGx platform support to Valgrind 79 Should action 80 81=== Syscalls/ioctls ==================================================== 82 83339424 Full support of KVM ioctl 84 In progress -- maybe can be closed? 85 86339563 The DVB demux DMX_STOP ioctl doesn't have a wrapper 87 has patch 88 89341537 POST(sys_io_getevents) shouldn't check iocb or iovec 90 unclear whether really valid or not 91 92342009 Error with valgrind on setcap file on 32bits (only) system 93 == 335143 94 possibly unfixable 95 96342040 Valgrind mishandles clone with CLONE_VFORK | CLONE_VM that clones 97 to a different stack 98 99342809 mremap fails when protection is enabled 100 101343099 Linux setns syscall wrapper missing, unhandled syscall: 308 102 HAS PATCH. Check that mayBlock is enabled. 103 104=== Debuginfo reader =================================================== 105 106327427 ifunc crash when symbols are discarded 107 unclear what to do 108 109330321 debuginfo readers complain about DT_AT_signature 110 no action so far 111 112339744 warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xf2 113 no action so far 114 115340854 valgrind-di-server: allow specification of multiple paths 116 has patch, unclear if Right Thing To Do 117 118=== Tools/Memcheck ===================================================== 119 120339499 memcheck does not always detect access beyond mmaped page 121 probably wontfix 122 123339762 missing replacement alias for glibc __strchr_sse42, __strcpy_sse2, 124 __strcpy_ssse3 125 obscure? 126 127340392 Incorrect "Conditional jump..." message 128 Optimised code a la Clang 129 PROBABLE CANTFIX 130 131342683 memory past the brk limit is not initially inaccessible/unaddressable 132 133=== Tools/DRD ========================================================== 134=== Tools/Helgrind ===================================================== 135 136339330 Feature request: Add support for C++11's std::atomic 137 Helgrind 138 139340905 hg_main.c:291 (lockN_acquire_reader): Assertion 'lk->kind 140 == LK_rdwr' failed. 141 against 3.8.1 142 143=== Tools/SGCheck ====================================================== 144 145340021 sgcheck tool does not work properly on ARM 146 == 296166 147 148=== Tools/Massif ======================================================= 149 150342353 Allow dumping full massif output while valgrind is still running 151 Has patch 152 153=== Tools/Cachegrind =================================================== 154 155=== Tools/Lackey ======================================================= 156 157343385 Improve lackey's address traces 158 w/ several patches 159 160=== other/mips ========================================================= 161 162341036 dumping core and Assertion 'sizeof(*regs) == sizeof(prs->pr_reg) 163 possibly a dup of 328670 ? 164 165341481 MIPS64: Iop_CmpNE32 triggers false warning on MIPS64 platforms 166 (memcheck) 167 168342356 MIPS: determine page size at runtime 169 Has patch; could possibly land? 170 171=== other/arm ========================================================== 172 173=== other/s390 ========================================================= 174=== other/Android ====================================================== 175 176339605 vgpreload_helgrind-arm-linux.so is not liked to libc.so 177 and can't find sched_yield, pthread_self and so on 178 recent Android 179 180339861 Remove Elf32_Nhdr definition for Android 181 uh, do we want this? 182 183339862 v6intThumb.c does not compile when -pie options is specified 184 should possible takes; causes build failures on newer Android 185 Might be important 186 187339945 Patch: replace #if !defined (VGPV_*_linux_android) with 188 #ifndef AT_FPUCW (probably WONTFIX) 189 190341704 [patch] Remove unnecessary ifdef android for 'sh' 191 do we want this? 192 193341726 [patch] Remove uneccessary ifdef _android_ from launcher-linux.c 194 do we want this? 195 196=== other/MacOS ======================================================== 197 198338781 OSX: unable to read debug info 199 200339745 Valgrind crash when check Marmalade app (OSX 10.?) 201 Rhys partially fixed 202 203341419 Signal handler ucontext_t not filled out correctly on OS X 204 Has patch 205 206342635 OS X 10.10 (Yosemite) - missing system calls and fcntl code 207 Has patch; -> Rhys? 208 209343306 OS X 10.10: UNKNOWN mach_msg unhandled MACH_SEND_TRAILER option 210 Partially fixed 211 212343523 OS X mach_ports_register: UNKNOWN task message 213 [id 3403, to mach_task_self(), reply 0x30f] 214 215343525 OS X host_get_special_port: UNKNOWN host message 216 [id 412, to mach_host_self(), reply 0x........] 217 218343649 OS X host_create_mach_voucher: UNKNOWN host message 219 [id 222, to mach_host_self(), reply 0x........] 220 221343663 [OSX Yosemite 10.10.1] The memcheck tool always reports a 222 leak regardless of the simplicity of the program. 223 224=== other/Win32 ======================================================= 225 226211529 valgrind doesn't show proper call stacks for programs compiled 227 by newer versions of visual c++ 228 229342117 Valgrind hangs after loading PDB file for MSVC compiled Firefox 230 under Wine 231 232=== GDB server ========================================================= 233=== Output ============================================================= 234 235339405 Adds ability to invoke a script in order to determine a 236 log-file name for a child being forked 237 Has plausible patch, should take, but needs test case 238 239342423 Log files should have the history 'execve' calls in them 240 Wishlist 241 242=== MPI ================================================================ 243 244330147 libmpiwrap PMPI_Get_count (should take; simple fix) 245 246=== Documentation ====================================================== 247 248340571 x86/Linux should no longer claim "almost complete" 249 250=== Uncategorised/run ================================================== 251 252326487 child of debugged process exits without becoming zombie 253 fork can fail if it receives a signal? 254 255339160 Running signal handler with alternate stack allocated on current 256 stack crashes callgrind 257 m_stacks.c brokenness? 258 259343173 helgrind crash during stack unwind 260 m_stacks.c: need to (de)register stack at munmap? 261 262343357 Please fix semaphore starvation when running threads-per-CPU 263 Unclear if we should do anything 264 265343715 valgrind hangs after vgdb help request 266 267=== Uncategorised/build================================================= 268 269339215 Valgrind 3.10.0 contain 2013 in copyrights notice 270 Update copyright dates! 271 272339542 Experiences with compiling valgrind with ICC 273 possibly closable now 274 275339636 Use fxsave64 and fxrstor64 mnemonics instead of rex64 prefix 276 Should apply 277 278342008 valgrind.h needs type cast for _zzq_default to compile with 279 clang/llvm in 64-bit mode 280 281343553 e500 chip now supports it? and,how about armv5 282 283======================================================================== 284======================================================================== 285======================================================================== 286 287arm64 critical regtest fails: 288 getregset.stderr.out sys 117, ptrace 289 proc-auxv no auxv? 290 memcheck/tests/linux/stack_changes asserts 291 sigaltstack.stderr.out segfaults -- host seems broken too 292 recursive signal delivery (!) 293 294======================================================================== 295======================================================================== 296======================================================================== 297 298Tue 3 Feb 00:14:18 CET 2015 299 300Check: space recovery in aspacemgr segnames array re-implemented? 301 302344019 new tool: trackptr to track pointer usage 303no action so far 304 305344052 Cost of dynamically loaded library is not properly linked to 306 source code lines if the library is unloaded before exit 307 308344139 vex x86->IR: 0x36 0x8A 0x18 0x22 (and many other examples) 309stack segment overrides 310 311344337 OS X 10.10 unhandled syscall 312 313344382 Memcheck has high false error rates on MSVC2013 compiled, 314 optimised, code in progress 315fixed, but preffed off 316 317344401 x32 ABI support 318no action 319 320344416 'make regtest' does not work cleanly on OS X 321meta-bug 322 323344512 unhandled syscal: unix:349 324 325344524 store conditional of guest applications always fail (Octeon3(MIPS) 326discussed, partial fix available 327 328344543 --leak-check=full memleak errors from system libraries in OS X 10.10.2 329can possibly close 330 331344547 vex x86->IR: 0xC5 0xF8 0x77 0xE9 332avx on x86 333 334344560 Stack traces missing penultimate frame 335has plausible patch 336 337Thu 5 Mar 18:08:27 CET 2015 338 339344886 V should not try to extend a sigaltstack 340unclear what the situation is 341 342344936 unhandled syscall: unix:473 (readlinkat) on OS X 10.10 343 344345077 linux syscall execveat support (linux 3.19) 345has patch 346 347345121 helgrind/tests/hg05_race2 fails intermittently 348 349345126 Incorrect handling of VIDIOC_G_AUDIO and G_AUDOUT 350has patch 351 352345177 arm64: prfm (reg) not implemented 353 354345248 add support for Solaris OS in valgrind 355 356345261 Unsupported instruction 0xC5 0xFB 0x10 0x44 (vmovsd 357 0x8(%esp),%xmm0) 358avx on x86 359 360345307 Please suppress warning about "still reachable" memory 361 when using libstdc++ from gcc 5 362 363345338 TIOCGSERIAL and TIOCSSERIAL ioctl support on Linux 364has patch 365 366345414 I get the following when I ran valgrind WARNING: 367 unhandled syscall: 317 368sys_seccomp 369 370345751 OS X: Incorrect result for operator Iop_Add32 and Iop_Add64 371due to mce.useLLVMworkarounds = True 372 373345753 OS X: sanityCheckFail: exiting due to bad IR for Iop_AddF64 374IR type error; should look at it 375 376345763 MIPS N32 ABI support 377various patches, no upstream action so far 378 379345811 annotate also FP for memcheck dirty helpers 380has patch, should commit 381 382345824 aspacem segment mismatch on OS X: seen with none/tests/bigcode 383but also for s390-linux; confusing 384 385345921 Add TileGX instruction set test in non/tests/tilegx/ 386 387345928 amd64: callstack only contains current function for small stacks 388has patch, but needs further investigation 389 390345929 Crash when running with `--gen-suppressions=yes` after 391 entering 'y' to confirm 392probable easy fix 393 394345975 Artificial functions in backtrace 395more Dwarf complexity 396 397345984 disInstr(arm): unhandled instruction: 0xEE193F1E 398 399345985 disInstr(arm): unhandled instruction: 0xF2200150 400 401345987 MIPS64: Implement cavium LHX instruction 402has patches 403 404346023 valgrind crashes when run over mono program 405roundsd, sse 4.1, x86 406 407346031 MIPS: Implement support for the CvmCount register (rhwr %0, 31) 408has patch 409 410346185 Fix typo saving altivec register v24 411has patch 412 413346411 MIPS: SysRes::_valEx handling is incorrect 414 415346476 vex amd64->IR: 0xC5 0xFD 0xC2 0xC1 0x10 0xC5 0xFD 0x29 416 417346562 MIPS64: lwl/lwr instructions are performing 64bit loads 418 and causing spurious "invalid read of size 8" warnings 419has patch 420 421346665 valgrind: the 'impossible' happened: LibVEX called failure_exit(). 422armv6 probable failure 423 424Tue 28 Apr 13:48:40 CEST 2015 425 426346912 callgrind bails out on vpmaskmovd instruction 427 428347198 vex x86->IR: unhandled instruction bytes: 0x66 0xF 0x3A 0x63 429 430347233 Fix memcheck/tests/strchr on OS X 10.10 (Haswell) 431 432347322 Power PC regression test cleanup 433 434Fri 8 May 11:52:14 CEST 2015 435 436347686 Patch set to cleanup PPC64 regtests 437 438347698 3.10.1 doesn't recognize cmpsd generated by DynASM 439 440347885 unhandled instruction bytes: 0xD6 0xF 0x84 0x1B (SALC amd64) 441 442Tue 19 May 11:22:52 CEST 2015 443