1 Release 3.14.0 (?? ????????? 201?)
2 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~--
3 Release 3.14.0 is under development, not yet released.
5 3.14.0 is a feature release with many improvements and the usual collection of
8 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
9 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
10 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
11 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
12 support for X86/macOS 10.13, AMD64/macOS 10.13.
14 * ==================== CORE CHANGES ===================
16 * The new option --keep-debuginfo=no|yes (default no) can be used to
17 keep symbols etc for unloaded code. This allows saved stack traces
18 (e.g. memory leaks) to include file/line info for code that has been
19 dlclose'd (or similar). See user manual for more information and
22 * ================== PLATFORM CHANGES =================
24 * Preliminary support for macOS 10.13 has been added.
25 * mips: support for MIPS SIMD architecture (MSA) has been added.
27 * ==================== TOOL CHANGES ====================
31 - Addition of a flag --delta-stacktrace=no|yes [yes on linux amd64/x86],
32 specifying how full history stack traces are computed. Using yes can speed
33 up Helgrind by 25% when using --history-level=full.
35 * ==================== OTHER CHANGES ====================
38 * ==================== FIXED BUGS ====================
40 The following bugs have been fixed or resolved. Note that "n-i-bz"
41 stands for "not in bugzilla" -- that is, a bug that was reported to us
42 but never got a bugzilla entry. We encourage you to file bugs in
43 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
44 than mailing the developers (or mailing lists) directly -- bugs that
45 are not entered into bugzilla tend to get forgotten about or ignored.
47 To see details of a given bug, visit
48 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
49 where XXXXXX is the bug number as listed below.
51 79362 Debug info is lost for .so files when they are dlclose'd
52 208052 strlcpy error when n = 0
53 255603 exp-sgcheck Assertion '!already_present' failed
54 376257 helgrind history full speed up using a cached stack
55 379373 Fix syscall param msg->desc.port.name points to uninitialised byte(s)
57 379748 Fix missing pselect syscall (OS X 10.11)
58 380397 s390x: __GI_strcspn() replacemenet needed
59 381162 possible array overrun in VEX register allocator
60 381272 ppc64 doesn't compile test_isa_2_06_partx.c without VSX support
61 381289 epoll_pwait can have a NULL sigmask
62 381274 powerpc too chatty even with --sigill-diagnostics=no
63 381553 VEX register allocator v3
64 381769 Use ucontext_t instead of struct ucontext
65 381805 arm32 needs ld.so index hardwire for new glibc security fixes
66 382256 gz compiler flag test doesn't work for gold
67 382407 vg_perf needs "--terse" command line option
68 382515 "Assertion 'di->have_dinfo' failed." on wine's dlls/mscoree/tests/mscoree.c
69 382563 MIPS MSA ASE support
70 382998 xml-socket doesn't work
71 383275 massif valgrind: m_xarray.c:162 (ensureSpaceXA): Assertion '!xa->arr' failed
72 384096 Mention AddrCheck at Memcheck's command line option --undef-value-errors=no
73 384526 reduce number of spill instructions generated by VEX register allocator v3
74 384584 Callee saved registers listed first for AMD64, X86, and PPC architectures
75 384987 VEX register allocator: allocate caller-save registers for short lived vregs
76 385182 PPC64 is missing support for the DSCR
77 385207 PPC64, generate_store_FPRF() generates too many Iops
78 385208 PPC64, xxperm instruction exhausts temporary memory
79 385210 PPC64, vpermr instruction could exhaust temporary memory
80 385183 PPC64, Add support for xscmpeqdp, xscmpgtdp, xscmpgedp, xsmincdp instructions
81 385334 PPC64, fix vpermr, xxperm, xxpermr mask value.
82 385868 glibc ld.so _dl_runtime_resolve_avx_slow conditional jump warning.
83 385912 none/tests/rlimit_nofile fails on newer glibc/kernel.
84 385939 Optionally exit on the first error
85 386397 PPC64, valgrind truncates powerpc timebase to 32-bits.
86 387712 s390x cgijnl reports Conditional jump depends on uninitialised value
87 387773 .gnu_debugaltlink paths resolve relative to .debug file, not symlink
89 n-i-bz Fix missing workq_ops operations (macOS)
90 n-i-bz fix bug in strspn replacement
92 Release 3.13.0 (15 June 2017)
93 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
95 3.13.0 is a feature release with many improvements and the usual collection of
98 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
99 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
100 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
101 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12.
103 * ==================== CORE CHANGES ===================
105 * The translation cache size has been increased to keep up with the demands of
106 large applications. The maximum number of sectors has increased from 24 to
107 48. The default number of sectors has increased from 16 to 32 on all
108 targets except Android, where the increase is from 6 to 12.
110 * The amount of memory that Valgrind can use has been increased from 64GB to
111 128GB. In particular this means your application can allocate up to about
112 60GB when running on Memcheck.
114 * Valgrind's default load address has been changed from 0x3800'0000 to
115 0x5800'0000, so as to make it possible to load larger executables. This
116 should make it possible to load executables of size at least 1200MB.
118 * A massive spaceleak caused by reading compressed debuginfo files has been
119 fixed. Valgrind should now be entirely usable with gcc-7.0 "-gz" created
122 * The C++ demangler has been updated.
124 * Support for demangling Rust symbols has been added.
126 * A new representation of stack traces, the "XTree", has been added. An XTree
127 is a tree of stacktraces with data associated with the stacktraces. This is
128 used by various tools (Memcheck, Helgrind, Massif) to report on the heap
129 consumption of your program. Reporting is controlled by the new options
130 --xtree-memory=none|allocs|full and --xtree-memory-file=<file>.
132 A report can also be produced on demand using the gdbserver monitor command
133 'xtmemory [<filename>]>'. The XTree can be output in 2 formats: 'callgrind
134 format' and 'massif format. The existing visualisers for these formats (e.g.
135 callgrind_annotate, KCachegrind, ms_print) can be used to visualise and
136 analyse these reports.
138 Memcheck can also produce XTree leak reports using the Callgrind file
139 format. For more details, see the user manual.
141 * ================== PLATFORM CHANGES =================
143 * ppc64: support for ISA 3.0B and various fixes for existing 3.0 support
145 * amd64: fixes for JIT failure problems on long AVX2 code blocks
147 * amd64 and x86: support for CET prefixes has been added
149 * arm32: a few missing ARMv8 instructions have been implemented
151 * arm64, mips64, mips32: an alternative implementation of Load-Linked and
152 Store-Conditional instructions has been added. This is to deal with
153 processor implementations that implement the LL/SC specifications strictly
154 and as a result cause Valgrind to hang in certain situations. The
155 alternative implementation is automatically enabled at startup, as required.
156 You can use the option --sim-hints=fallback-llsc to force-enable it if you
159 * Support for OSX 10.12 has been improved.
161 * On Linux, clone handling has been improved to honour CLONE_VFORK that
162 involves a child stack. Note however that CLONE_VFORK | CLONE_VM is handled
163 like CLONE_VFORK (by removing CLONE_VM), so applications that depend on
164 CLONE_VM exact semantics will (still) not work.
166 * The TileGX/Linux port has been removed because it appears to be both unused
169 * ==================== TOOL CHANGES ====================
173 - Memcheck should give fewer false positives when running optimised
174 Clang/LLVM generated code.
176 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
178 - New command line options --xtree-leak=no|yes and --xtree-leak-file=<file>
179 to produce the end of execution leak report in a xtree callgrind format
182 - New option 'xtleak' in the memcheck leak_check monitor command, to produce
183 the leak report in an xtree file.
187 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
189 - For some workloads (typically, for big applications), Massif memory
190 consumption and CPU consumption has decreased significantly.
194 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
196 - addition of client request VALGRIND_HG_GNAT_DEPENDENT_MASTER_JOIN, useful
197 for Ada gnat compiled applications.
199 * ==================== OTHER CHANGES ====================
201 * For Valgrind developers: in an outer/inner setup, the outer Valgrind will
202 append the inner guest stacktrace to the inner host stacktrace. This helps
203 to investigate the errors reported by the outer, when they are caused by the
204 inner guest program (such as an inner regtest). See README_DEVELOPERS for
207 * To allow fast detection of callgrind files by desktop environments and file
208 managers, the format was extended to have an optional first line that
209 uniquely identifies the format ("# callgrind format"). Callgrind creates
210 this line now, as does the new xtree functionality.
212 * File name template arguments (such as --log-file, --xtree-memory-file, ...)
213 have a new %n format letter that is replaced by a sequence number.
215 * "--version -v" now shows the SVN revision numbers from which Valgrind was
218 * ==================== FIXED BUGS ====================
220 The following bugs have been fixed or resolved. Note that "n-i-bz"
221 stands for "not in bugzilla" -- that is, a bug that was reported to us
222 but never got a bugzilla entry. We encourage you to file bugs in
223 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
224 than mailing the developers (or mailing lists) directly -- bugs that
225 are not entered into bugzilla tend to get forgotten about or ignored.
227 To see details of a given bug, visit
228 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
229 where XXXXXX is the bug number as listed below.
231 162848 --log-file output isn't split when a program forks
232 340777 Illegal instruction on mips (ar71xx)
233 341481 MIPS64: Iop_CmpNE32 triggers false warning on MIPS64 platforms
234 342040 Valgrind mishandles clone with CLONE_VFORK | CLONE_VM that clones
235 to a different stack.
236 344139 x86 stack-seg overrides, needed by the Wine people
237 344524 store conditional of guest applications always fail - observed on
239 348616 Wine/valgrind: noted but unhandled ioctl 0x5390 [..] (DVD_READ_STRUCT)
240 352395 Please provide SVN revision info in --version -v
241 352767 Wine/valgrind: noted but unhandled ioctl 0x5307 [..] (CDROMSTOP)
242 356374 Assertion 'DRD_(g_threadinfo)[tid].pt_threadid !=
243 INVALID_POSIX_THREADID' failed
244 358213 helgrind/drd bar_bad testcase hangs or crashes with new glibc pthread
245 barrier implementation
246 358697 valgrind.h: Some code remains even when defining NVALGRIND
247 359202 Add musl libc configure/compile
248 360415 amd64 instructions ADCX and ADOX are not implemented in VEX
249 == 372828 (vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10)
250 360429 unhandled ioctl 0x530d with no size/direction hints (CDROMREADMODE1)
251 362223 assertion failed when .valgrindrc is a directory instead of a file
252 367543 bt/btc/btr/bts x86/x86_64 instructions are poorly-handled wrt flags
253 367942 Segfault vgPlain_do_sys_sigaction (m_signals.c:1138)
254 368507 can't malloc chunks larger than about 34GB
255 368529 Android arm target link error, missing atexit and pthread_atfork
256 368863 WARNING: unhandled arm64-linux syscall: 100 (get_robust_list)
257 368865 WARNING: unhandled arm64-linux syscall: 272 (kcmp)
258 368868 disInstr(arm64): unhandled instruction 0xD53BE000 = cntfrq_el0 (ARMv8)
259 368917 WARNING: unhandled arm64-linux syscall: 218 (request_key)
260 368918 WARNING: unhandled arm64-linux syscall: 127 (sched_rr_get_interval)
261 368922 WARNING: unhandled arm64-linux syscall: 161 (sethostname)
262 368924 WARNING: unhandled arm64-linux syscall: 84 (sync_file_range)
263 368925 WARNING: unhandled arm64-linux syscall: 130 (tkill)
264 368926 WARNING: unhandled arm64-linux syscall: 97 (unshare)
265 369459 valgrind on arm64 violates the ARMv8 spec (ldxr/stxr)
266 370028 Reduce the number of compiler warnings on MIPS platforms
267 370635 arm64 missing syscall getcpu
268 371225 Fix order of timer_{gettime,getoverrun,settime} syscalls on arm64
269 371227 Clean AArch64 syscall table
270 371412 Rename wrap_sys_shmat to sys_shmat like other wrappers
271 371471 Valgrind complains about non legit memory leaks on placement new (C++)
272 371491 handleAddrOverrides() is [incorrect] when ASO prefix is used
273 371503 disInstr(arm64): unhandled instruction 0xF89F0000
274 371869 support '%' in symbol Z-encoding
275 371916 execution tree xtree concept
276 372120 c++ demangler demangles symbols which are not c++
277 372185 Support of valgrind on ARMv8 with 32 bit executable
278 372188 vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10 0x10 0x48 (PCMPxSTRx $0x10)
279 372195 Power PC, xxsel instruction is not always recognized.
280 372504 Hanging on exit_group
281 372600 process loops forever when fatal signals are arriving quickly
282 372794 LibVEX (arm32 front end): 'Assertion szBlg2 <= 3' failed
283 373046 Stacks registered by core are never deregistered
284 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
285 373086 Implement additional Xen hypercalls
286 373192 Calling posix_spawn in glibc 2.24 completely broken
287 373488 Support for fanotify API on ARM64 architecture
288 == 368864 WARNING: unhandled arm64-linux syscall: 262 (fanotify_init)
289 373555 Rename BBPTR to GSPTR as it denotes guest state pointer only
290 373938 const IRExpr arguments for matchIRExpr()
291 374719 some spelling fixes
292 374963 increase valgrind's load address to prevent mmap failure
293 375514 valgrind_get_tls_addr() does not work in case of static TLS
294 375772 +1 error in get_elf_symbol_info() when computing value of 'hi' address
295 for ML_(find_rx_mapping)()
296 375806 Test helgrind/tests/tc22_exit_w_lock fails with glibc 2.24
297 375839 Temporary storage exhausted, with long sequence of vfmadd231ps insns
298 == 377159 "vex: the `impossible' happened" still present
299 == 375150 Assertion 'tres.status == VexTransOK' failed
300 == 378068 valgrind crashes on AVX2 function in FFmpeg
301 376142 Segfaults on MIPS Cavium Octeon boards
302 376279 disInstr(arm64): unhandled instruction 0xD50320FF
303 376455 Solaris: unhandled syscall lgrpsys(180)
304 376518 Solaris: unhandled fast trap getlgrp(6)
305 376611 ppc64 and arm64 don't know about prlimit64 syscall
306 376729 PPC64, remove R2 from the clobber list
308 376956 syswrap of SNDDRV and DRM_IOCTL_VERSION causing some addresses
309 to be wrongly marked as addressable
310 377066 Some Valgrind unit tests fail to compile on Ubuntu 16.10 with
311 PIE enabled by default
312 377376 memcheck/tests/linux/getregset fails with glibc2.24
313 377427 PPC64, lxv instruction failing on odd destination register
314 377478 PPC64: ISA 3.0 setup fixes
315 377698 Missing memory check for futex() uaddr arg for FUTEX_WAKE
316 and FUTEX_WAKE_BITSET, check only 4 args for FUTEX_WAKE_BITSET,
317 and 2 args for FUTEX_TRYLOCK_PI
318 377717 Fix massive space leak when reading compressed debuginfo sections
319 377891 Update Xen 4.6 domctl wrappers
320 377930 fcntl syscall wrapper is missing flock structure check
321 378524 libvexmultiarch_test regression on s390x and ppc64
322 378535 Valgrind reports INTERNAL ERROR in execve syscall wrapper
323 378673 Update libiberty demangler
324 378931 Add ISA 3.0B additional isnstructions, add OV32, CA32 setting support
325 379039 syscall wrapper for prctl(PR_SET_NAME) must not check more than 16 bytes
326 379094 Valgrind reports INTERNAL ERROR in rt_sigsuspend syscall wrapper
327 379371 UNKNOWN task message [id 3444, to mach_task_self(), reply 0x603]
328 (task_register_dyld_image_infos)
329 379372 UNKNOWN task message [id 3447, to mach_task_self(), reply 0x603]
330 (task_register_dyld_shared_cache_image_info)
331 379390 unhandled syscall: mach:70 (host_create_mach_voucher_trap)
332 379473 MIPS: add support for rdhwr cycle counter register
333 379504 remove TileGX/Linux port
334 379525 Support more x86 nop opcodes
335 379838 disAMode(x86): not an addr!
336 379703 PC ISA 3.0 fixes: stxvx, stxv, xscmpexpdp instructions
337 379890 arm: unhandled instruction: 0xEBAD 0x1B05 (sub.w fp, sp, r5, lsl #4)
338 379895 clock_gettime does not execute POST syscall wrapper
339 379925 PPC64, mtffs does not set the FPCC and C bits in the FPSCR correctly
340 379966 WARNING: unhandled amd64-linux syscall: 313 (finit_module)
341 380200 xtree generated callgrind files refer to files without directory name
342 380202 Assertion failure for cache line size (cls == 64) on aarch64.
343 380397 s390x: __GI_strcspn() replacement needed
344 n-i-bz Fix pub_tool_basics.h build issue with g++ 4.4.7.
346 (3.13.0.RC1: 2 June 2017, vex r3386, valgrind r16434)
347 (3.13.0.RC2: 9 June 2017, vex r3389, valgrind r16443)
348 (3.13.0: 14 June 2017, vex r3396, valgrind r16446)
352 Release 3.12.0 (20 October 2016)
353 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
355 3.12.0 is a feature release with many improvements and the usual
356 collection of bug fixes.
358 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
359 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
360 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
361 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
362 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
363 X86/MacOSX 10.11/12, AMD64/MacOSX 10.11/12 and TILEGX/Linux.
365 * ================== PLATFORM CHANGES =================
367 * POWER: Support for ISA 3.0 has been added
369 * mips: support for O32 FPXX ABI has been added.
370 * mips: improved recognition of different processors
371 * mips: determination of page size now done at run time
373 * amd64: Partial support for AMD FMA4 instructions.
375 * arm, arm64: Support for v8 crypto and CRC instructions.
377 * Improvements and robustification of the Solaris port.
379 * Preliminary support for MacOS 10.12 (Sierra) has been added.
381 Whilst 3.12.0 continues to support the 32-bit x86 instruction set, we
382 would prefer users to migrate to 64-bit x86 (a.k.a amd64 or x86_64)
383 where possible. Valgrind's support for 32-bit x86 has stagnated in
384 recent years and has fallen far behind that for 64-bit x86
385 instructions. By contrast 64-bit x86 is well supported, up to and
388 * ==================== TOOL CHANGES ====================
392 - Added meta mempool support for describing a custom allocator which:
393 - Auto-frees all chunks assuming that destroying a pool destroys all
395 - Uses itself to allocate other memory blocks
397 - New flag --ignore-range-below-sp to ignore memory accesses below
398 the stack pointer, if you really have to. The related flag
399 --workaround-gcc296-bugs=yes is now deprecated. Use
400 --ignore-range-below-sp=1024-1 as a replacement.
404 - Improved thread startup time significantly on non-Linux platforms.
408 - Added collection of the metric "tot-blocks-allocd"
410 * ==================== OTHER CHANGES ====================
412 * Replacement/wrapping of malloc/new related functions is now done not just
413 for system libraries by default, but for any globally defined malloc/new
414 related function (both in shared libraries and statically linked alternative
415 malloc implementations). The dynamic (runtime) linker is excluded, though.
416 To only intercept malloc/new related functions in
417 system libraries use --soname-synonyms=somalloc=nouserintercepts (where
418 "nouserintercepts" can be any non-existing library name).
419 This new functionality is not implemented for MacOS X.
421 * The maximum number of callers in a suppression entry is now equal to
422 the maximum size for --num-callers (500).
423 Note that --gen-suppressions=yes|all similarly generates suppressions
424 containing up to --num-callers frames.
426 * New and modified GDB server monitor features:
428 - Valgrind's gdbserver now accepts the command 'catch syscall'.
429 Note that you must have GDB >= 7.11 to use 'catch syscall' with
432 * New option --run-cxx-freeres=<yes|no> can be used to change whether
433 __gnu_cxx::__freeres() cleanup function is called or not. Default is
436 * Valgrind is able to read compressed debuginfo sections in two formats:
437 - zlib ELF gABI format with SHF_COMPRESSED flag (gcc option -gz=zlib)
438 - zlib GNU format with .zdebug sections (gcc option -gz=zlib-gnu)
440 * Modest JIT-cost improvements: the cost of instrumenting code blocks
441 for the most common use case (x86_64-linux, Memcheck) has been
444 * Improved performance for programs that do a lot of discarding of
445 instruction address ranges of 8KB or less.
447 * The C++ symbol demangler has been updated.
449 * More robustness against invalid syscall parameters on Linux.
451 * ==================== FIXED BUGS ====================
453 The following bugs have been fixed or resolved. Note that "n-i-bz"
454 stands for "not in bugzilla" -- that is, a bug that was reported to us
455 but never got a bugzilla entry. We encourage you to file bugs in
456 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
457 than mailing the developers (or mailing lists) directly -- bugs that
458 are not entered into bugzilla tend to get forgotten about or ignored.
460 To see details of a given bug, visit
461 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
462 where XXXXXX is the bug number as listed below.
464 191069 Exiting due to signal not reported in XML output
465 199468 Suppressions: stack size limited to 25
466 while --num-callers allows more frames
467 212352 vex amd64 unhandled opc_aux = 0x 2, first_opcode == 0xDC (FCOM)
468 278744 cvtps2pd with redundant RexW
469 303877 valgrind doesn't support compressed debuginfo sections.
470 345307 Warning about "still reachable" memory when using libstdc++ from gcc 5
471 348345 Assertion fails for negative lineno
472 348924 MIPS: Load doubles through memory so the code compiles with the FPXX ABI
473 351282 V 3.10.1 MIPS softfloat build broken with GCC 4.9.3 / binutils 2.25.1
474 351692 Dumps created by valgrind are not readable by gdb (mips32 specific)
475 351804 Crash on generating suppressions for "printf" call on OS X 10.10
476 352197 mips: mmap2() not wrapped correctly for page size > 4096
477 353083 arm64 doesn't implement various xattr system calls
478 353084 arm64 doesn't support sigpending system call
479 353137 www: update info for Supported Platforms
480 353138 www: update "The Valgrind Developers" page
481 353370 don't advertise RDRAND in cpuid for Core-i7-4910-like avx2 machine
484 353384 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x62 (pcmpXstrX $0x62)
485 353398 WARNING: unhandled amd64-solaris syscall: 207
486 353660 XML in auxwhat tag not escaping reserved symbols properly
487 353680 s390x: Crash with certain glibc versions due to non-implemented TBEGIN
488 353727 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x72 (pcmpXstrX $0x72)
489 353802 ELF debug info reader confused with multiple .rodata sections
490 353891 Assert 'bad_scanned_addr < VG_ROUNDDN(start+len, sizeof(Addr))' failed
491 353917 unhandled amd64-solaris syscall fchdir(120)
492 353920 unhandled amd64-solaris syscall: 170
493 354274 arm: unhandled instruction: 0xEBAD 0x0AC1 (sub.w sl, sp, r1, lsl #3)
494 354392 unhandled amd64-solaris syscall: 171
495 354797 Vbit test does not include Iops for Power 8 instruction support
496 354883 tst->os_state.pthread - magic_delta assertion failure on OSX 10.11
500 354933 Fix documentation of --kernel-variant=android-no-hw-tls option
501 355188 valgrind should intercept all malloc related global functions
502 355454 do not intercept malloc related symbols from the runtime linker
503 355455 stderr.exp of test cases wrapmalloc and wrapmallocstatic overconstrained
504 356044 Dwarf line info reader misinterprets is_stmt register
505 356112 mips: replace addi with addiu
506 356393 valgrind (vex) crashes because isZeroU happened
509 356676 arm64-linux: unhandled syscalls 125, 126 (sched_get_priority_max/min)
510 356678 arm64-linux: unhandled syscall 232 (mincore)
511 356817 valgrind.h triggers compiler errors on MSVC when defining NVALGRIND
512 356823 Unsupported ARM instruction: stlex
513 357059 x86/amd64: SSE cvtpi2ps with memory source does transition to MMX state
514 357338 Unhandled instruction for SHA instructions libcrypto Boring SSL
515 357673 crash if I try to run valgrind with a binary link with libcurl
516 357833 Setting RLIMIT_DATA to zero breaks with linux 4.5+
517 357871 pthread_spin_destroy not properly wrapped
518 357887 Calls to VG_(fclose) do not close the file descriptor
519 357932 amd64->IR: accept redundant REX prefixes for {minsd,maxsd} m128, xmm.
520 358030 support direct socket calls on x86 32bit (new in linux 4.3)
521 358478 drd/tests/std_thread.cpp doesn't build with GCC6
522 359133 Assertion 'eltSzB <= ddpa->poolSzB' failed
523 359181 Buffer Overflow during Demangling
524 359201 futex syscall "skips" argument 5 if op is FUTEX_WAIT_BITSET
525 359289 s390x: popcnt (B9E1) not implemented
526 359472 The Power PC vsubuqm instruction doesn't always give the correct result
527 359503 Add missing syscalls for aarch64 (arm64)
528 359645 "You need libc6-dbg" help message could be more helpful
529 359703 s390: wire up separate socketcalls system calls
530 359724 getsockname might crash - deref_UInt should call safe_to_deref
531 359733 amd64 implement ld.so strchr/index override like x86
532 359767 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 1/5
533 359829 Power PC test suite none/tests/ppc64/test_isa_2_07.c uses
535 359838 arm64: Unhandled instruction 0xD5033F5F (clrex)
536 359871 Incorrect mask handling in ppoll
537 359952 Unrecognised PCMPESTRM variants (0x70, 0x19)
538 360008 Contents of Power vr registers contents is not printed correctly when
539 the --vgdb-shadow-registers=yes option is used
540 360035 POWER PC instruction bcdadd and bcdsubtract generate result with
542 360378 arm64: Unhandled instruction 0x5E280844 (sha1h s4, s2)
543 360425 arm64 unsupported instruction ldpsw
545 360519 none/tests/arm64/memory.vgtest might fail with newer gcc
546 360571 Error about the Android Runtime reading below the stack pointer on ARM
547 360574 Wrong parameter type for an ashmem ioctl() call on Android and ARM64
548 360749 kludge for multiple .rodata sections on Solaris no longer needed
549 360752 raise the number of reserved fds in m_main.c from 10 to 12
550 361207 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 2/5
551 361226 s390x: risbgn (EC59) not implemented
552 361253 [s390x] ex_clone.c:42: undefined reference to `pthread_create'
553 361354 ppc64[le]: wire up separate socketcalls system calls
554 361615 Inconsistent termination for multithreaded process terminated by signal
555 361926 Unhandled Solaris syscall: sysfs(84)
556 362009 V dumps core on unimplemented functionality before threads are created
557 362329 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 3/5
558 362894 missing (broken) support for wbit field on mtfsfi instruction (ppc64)
559 362935 [AsusWRT] Assertion 'sizeof(TTEntryC) <= 88' failed
560 362953 Request for an update to the Valgrind Developers page
561 363680 add renameat2() support
562 363705 arm64 missing syscall name_to_handle_at and open_by_handle_at
563 363714 ppc64 missing syscalls sync, waitid and name_to/open_by_handle_at
564 363858 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 4/5
565 364058 clarify in manual limitations of array overruns detections
566 364413 pselect sycallwrapper mishandles NULL sigmask
567 364728 Power PC, missing support for several HW registers in
568 get_otrack_shadow_offset_wrk()
569 364948 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 5/5
570 365273 Invalid write to stack location reported after signal handler runs
571 365912 ppc64BE segfault during jm-insns test (RELRO)
572 366079 FPXX Support for MIPS32 Valgrind
573 366138 Fix configure errors out when using Xcode 8 (clang 8.0.0)
574 366344 Multiple unhandled instruction for Aarch64
575 (0x0EE0E020, 0x1AC15800, 0x4E284801, 0x5E040023, 0x5E056060)
576 367995 Integration of memcheck with custom memory allocator
577 368120 x86_linux asm _start functions do not keep 16-byte aligned stack pointer
578 368412 False positive result for altivec capability check
579 368416 Add tc06_two_races_xml.exp output for ppc64
580 368419 Perf Events ioctls not implemented
581 368461 mmapunmap test fails on ppc64
582 368823 run_a_thread_NORETURN assembly code typo for VGP_arm64_linux target
583 369000 AMD64 fma4 instructions unsupported.
584 369169 ppc64 fails jm_int_isa_2_07 test
585 369175 jm_vec_isa_2_07 test crashes on ppc64
586 369209 valgrind loops and eats up all memory if cwd doesn't exist.
587 369356 pre_mem_read_sockaddr syscall wrapper can crash with bad sockaddr
588 369359 msghdr_foreachfield can crash when handling bad iovec
589 369360 Bad sigprocmask old or new sets can crash valgrind
590 369361 vmsplice syscall wrapper crashes on bad iovec
591 369362 Bad sigaction arguments crash valgrind
592 369383 x86 sys_modify_ldt wrapper crashes on bad ptr
593 369402 Bad set/get_thread_area pointer crashes valgrind
594 369441 bad lvec argument crashes process_vm_readv/writev syscall wrappers
595 369446 valgrind crashes on unknown fcntl command
596 369439 S390x: Unhandled insns RISBLG/RISBHG and LDE/LDER
597 369468 Remove quadratic metapool algorithm using VG_(HT_remove_at_Iter)
598 370265 ISA 3.0 HW cap stuff needs updating
599 371128 BCD add and subtract instructions on Power BE in 32-bit mode do not work
600 372195 Power PC, xxsel instruction is not always recognized
602 n-i-bz Fix incorrect (or infinite loop) unwind on RHEL7 x86 and amd64
603 n-i-bz massif --pages-as-heap=yes does not report peak caused by mmap+munmap
604 n-i-bz false positive leaks due to aspacemgr merging heap & non heap segments
605 n-i-bz Fix ppoll_alarm exclusion on OS X
606 n-i-bz Document brk segment limitation, reference manual in limit reached msg.
607 n-i-bz Fix clobber list in none/tests/amd64/xacq_xrel.c [valgrind r15737]
608 n-i-bz Bump allowed shift value for "add.w reg, sp, reg, lsl #N" [vex r3206]
609 n-i-bz amd64: memcheck false positive with shr %edx
610 n-i-bz arm3: Allow early writeback of SP base register in "strd rD, [sp, #-16]"
611 n-i-bz ppc: Fix two cases of PPCAvFpOp vs PPCFpOp enum confusion
612 n-i-bz arm: Fix incorrect register-number constraint check for LDAEX{,B,H,D}
613 n-i-bz DHAT: added collection of the metric "tot-blocks-allocd"
615 (3.12.0.RC1: 20 October 2016, vex r3282, valgrind r16094)
616 (3.12.0.RC2: 20 October 2016, vex r3282, valgrind r16096)
617 (3.12.0: 21 October 2016, vex r3282, valgrind r16098)
621 Release 3.11.0 (22 September 2015)
622 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
624 3.11.0 is a feature release with many improvements and the usual
625 collection of bug fixes.
627 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
628 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
629 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
630 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
631 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
632 X86/MacOSX 10.11, AMD64/MacOSX 10.11 and TILEGX/Linux.
634 * ================== PLATFORM CHANGES =================
636 * Support for Solaris/x86 and Solaris/amd64 has been added.
638 * Preliminary support for Mac OS X 10.11 (El Capitan) has been added.
640 * Preliminary support for the Tilera TileGX architecture has been added.
642 * s390x: It is now required for the host to have the "long displacement"
643 facility. The oldest supported machine model is z990.
645 * x86: on an SSE2 only host, Valgrind in 32 bit mode now claims to be a
646 Pentium 4. 3.10.1 wrongly claimed to be a Core 2, which is SSSE3.
648 * The JIT's register allocator is significantly faster, making the JIT
649 as a whole somewhat faster, so JIT-intensive activities, for example
650 program startup, are modestly faster, around 5%.
652 * There have been changes to the default settings of several command
653 line flags, as detailed below.
655 * Intel AVX2 support is more complete (64 bit targets only). On AVX2
656 capable hosts, the simulated CPUID will now indicate AVX2 support.
658 * ==================== TOOL CHANGES ====================
662 - The default value for --leak-check-heuristics has been changed from
663 "none" to "all". This helps to reduce the number of possibly
664 lost blocks, in particular for C++ applications.
666 - The default value for --keep-stacktraces has been changed from
667 "malloc-then-free" to "malloc-and-free". This has a small cost in
668 memory (one word per malloc-ed block) but allows Memcheck to show the
669 3 stacktraces of a dangling reference: where the block was allocated,
670 where it was freed, and where it is acccessed after being freed.
672 - The default value for --partial-loads-ok has been changed from "no" to
673 "yes", so as to avoid false positive errors resulting from some kinds
676 - A new monitor command 'xb <addr> <len>' shows the validity bits of
677 <len> bytes at <addr>. The monitor command 'xb' is easier to use
678 than get_vbits when you need to associate byte data value with
679 their corresponding validity bits.
681 - The 'block_list' monitor command has been enhanced:
682 o it can print a range of loss records
683 o it now accepts an optional argument 'limited <max_blocks>'
684 to control the number of blocks printed.
685 o if a block has been found using a heuristic, then
686 'block_list' now shows the heuristic after the block size.
687 o the loss records/blocks to print can be limited to the blocks
688 found via specified heuristics.
690 - The C helper functions used to instrument loads on
691 x86-{linux,solaris} and arm-linux (both 32-bit only) have been
692 replaced by handwritten assembly sequences. This gives speedups
693 in the region of 0% to 7% for those targets only.
695 - A new command line option, --expensive-definedness-checks=yes|no,
696 has been added. This is useful for avoiding occasional invalid
697 uninitialised-value errors in optimised code. Watch out for
698 runtime degradation, as this can be up to 25%. As always, though,
699 the slowdown is highly application specific. The default setting
704 - A new monitor command 'all_snapshots <filename>' dumps all
705 snapshots taken so far.
709 - Significant memory reduction and moderate speedups for
710 --history-level=full for applications accessing a lot of memory
711 with many different stacktraces.
713 - The default value for --conflict-cache-size=N has been doubled to
714 2000000. Users that were not using the default value should
715 preferably also double the value they give.
717 The default was changed due to the changes in the "full history"
718 implementation. Doubling the value gives on average a slightly more
719 complete history and uses similar memory (or significantly less memory
720 in the worst case) than the previous implementation.
722 - The Helgrind monitor command 'info locks' now accepts an optional
723 argument 'lock_addr', which shows information about the lock at the
726 - When using --history-level=full, the new Helgrind monitor command
727 'accesshistory <addr> [<len>]' will show the recorded accesses for
728 <len> (or 1) bytes at <addr>.
730 * ==================== OTHER CHANGES ====================
732 * The default value for the --smc-check option has been changed from
733 "stack" to "all-non-file" on targets that provide automatic D-I
734 cache coherence (x86, amd64 and s390x). The result is to provide,
735 by default, transparent support for JIT generated and self-modifying
738 * Mac OS X only: the default value for the --dsymutil option has been
739 changed from "no" to "yes", since any serious usage on Mac OS X
740 always required it to be "yes".
742 * The command line options --db-attach and --db-command have been removed.
743 They were deprecated in 3.10.0.
745 * When a process dies due to a signal, Valgrind now shows the signal
746 and the stacktrace at default verbosity (i.e. verbosity 1).
748 * The address description logic used by Memcheck and Helgrind now
749 describes addresses in anonymous segments, file mmap-ed segments,
750 shared memory segments and the brk data segment.
752 * The new option --error-markers=<begin>,<end> can be used to mark the
753 begin/end of errors in textual output mode, to facilitate
754 searching/extracting errors in output files that mix valgrind errors
757 * The new option --max-threads=<number> can be used to change the number
758 of threads valgrind can handle. The default is 500 threads which
759 should be more than enough for most applications.
761 * The new option --valgrind-stacksize=<number> can be used to change the
762 size of the private thread stacks used by Valgrind. This is useful
763 for reducing memory use or increasing the stack size if Valgrind
764 segfaults due to stack overflow.
766 * The new option --avg-transtab-entry-size=<number> can be used to specify
767 the expected instrumented block size, either to reduce memory use or
768 to avoid excessive retranslation.
770 * Valgrind can be built with Intel's ICC compiler, version 14.0 or later.
772 * New and modified GDB server monitor features:
774 - When a signal is reported in GDB, you can now use the GDB convenience
775 variable $_siginfo to examine detailed signal information.
777 - Valgrind's gdbserver now allows the user to change the signal
778 to deliver to the process. So, use 'signal SIGNAL' to continue execution
779 with SIGNAL instead of the signal reported to GDB. Use 'signal 0' to
780 continue without passing the signal to the process.
782 - With GDB >= 7.10, the command 'target remote'
783 will automatically load the executable file of the process running
784 under Valgrind. This means you do not need to specify the executable
785 file yourself, GDB will discover it itself. See GDB documentation about
786 'qXfer:exec-file:read' packet for more info.
788 * ==================== FIXED BUGS ====================
790 The following bugs have been fixed or resolved. Note that "n-i-bz"
791 stands for "not in bugzilla" -- that is, a bug that was reported to us
792 but never got a bugzilla entry. We encourage you to file bugs in
793 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
794 than mailing the developers (or mailing lists) directly -- bugs that
795 are not entered into bugzilla tend to get forgotten about or ignored.
797 To see details of a given bug, visit
798 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
799 where XXXXXX is the bug number as listed below.
801 116002 VG_(printf): Problems with justification of strings and integers
802 155125 avoid cutting away file:lineno after long function name
803 197259 Unsupported arch_prtctl PR_SET_GS option
804 201152 ppc64: Assertion in ppc32g_dirtyhelper_MFSPR_268_269
805 201216 Fix Valgrind does not support pthread_sigmask() on OS X
806 201435 Fix Darwin: -v does not show kernel version
807 208217 "Warning: noted but unhandled ioctl 0x2000747b" on Mac OS X
808 211256 Fixed an outdated comment regarding the default platform.
809 211529 Incomplete call stacks for code compiled by newer versions of MSVC
810 211926 Avoid compilation warnings in valgrind.h with -pedantic
811 212291 Fix unhandled syscall: unix:132 (mkfifo) on OS X
813 226609 Crediting upstream authors in man page
814 231257 Valgrind omits path when executing script from shebang line
815 254164 OS X task_info: UNKNOWN task message [id 3405, to mach_task_self() [..]
816 294065 Improve the pdb file reader by avoiding hardwired absolute pathnames
817 269360 s390x: Fix addressing mode selection for compare-and-swap
818 302630 Memcheck: Assertion failed: 'sizeof(UWord) == sizeof(UInt)'
820 312989 ioctl handling needs to do POST handling on generic ioctls and [..]
821 319274 Fix unhandled syscall: unix:410 (sigsuspend_nocancel) on OS X
822 324181 mmap does not handle MAP_32BIT (handle it now, rather than fail it)
823 327745 Fix valgrind 3.9.0 build fails on Mac OS X 10.6.8
824 330147 libmpiwrap PMPI_Get_count returns undefined value
825 333051 mmap of huge pages fails due to incorrect alignment
827 334802 valgrind does not always explain why a given option is bad
828 335618 mov.w rN, pc/sp (ARM32)
829 335785 amd64->IR 0xC4 0xE2 0x75 0x2F (vmaskmovpd)
834 335907 segfault when running wine's ddrawex/tests/surface.c under valgrind
835 338602 AVX2 bit in CPUID missing
836 338606 Strange message for scripts with invalid interpreter
837 338731 ppc: Fix testuite build for toolchains not supporting -maltivec
838 338995 shmat with hugepages (SHM_HUGETLB) fails with EINVAL
839 339045 Getting valgrind to compile and run on OS X Yosemite (10.10)
841 339156 gdbsrv not called for fatal signal
842 339215 Valgrind 3.10.0 contain 2013 in copyrights notice
843 339288 support Cavium Octeon MIPS specific BBIT*32 instructions
844 339636 Use fxsave64 and fxrstor64 mnemonics instead of old-school rex64 prefix
845 339442 Fix testsuite build failure on OS X 10.9
846 339542 Enable compilation with Intel's ICC compiler
847 339563 The DVB demux DMX_STOP ioctl doesn't have a wrapper
848 339688 Mac-specific ASM does not support .version directive (cpuid,
849 tronical and pushfpopf tests)
850 339745 Valgrind crash when check Marmalade app (partial fix)
851 339755 Fix known deliberate memory leak in setenv() on Mac OS X 10.9
852 339778 Linux/TileGx platform support for Valgrind
853 339780 Fix known uninitialised read in pthread_rwlock_init() on Mac OS X 10.9
854 339789 Fix none/tests/execve test on Mac OS X 10.9
855 339808 Fix none/tests/rlimit64_nofile test on Mac OS X 10.9
856 339820 vex amd64->IR: 0x66 0xF 0x3A 0x63 0xA 0x42 0x74 0x9 (pcmpistri $0x42)
857 340115 Fix none/tests/cmdline[1|2] tests on systems which define TMPDIR
858 340392 Allow user to select more accurate definedness checking in memcheck
859 to avoid invalid complaints on optimised code
860 340430 Fix some grammatical weirdness in the manual.
861 341238 Recognize GCC5/DWARFv5 DW_LANG constants (Go, C11, C++11, C++14)
862 341419 Signal handler ucontext_t not filled out correctly on OS X
863 341539 VG_(describe_addr) should not describe address as belonging to client
864 segment if it is past the heap end
865 341613 Enable building of manythreads and thread-exits tests on Mac OS X
866 341615 Fix none/tests/darwin/access_extended test on Mac OS X
867 341698 Valgrind's AESKEYGENASSIST gives wrong result in words 0 and 2 [..]
868 341789 aarch64: shmat fails with valgrind on ARMv8
869 341997 MIPS64: Cavium OCTEON insns - immediate operand handled incorrectly
870 342008 valgrind.h needs type cast [..] for clang/llvm in 64-bit mode
871 342038 Unhandled syscalls on aarch64 (mbind/get/set_mempolicy)
872 342063 wrong format specifier for test mcblocklistsearch in gdbserver_tests
873 342117 Hang when loading PDB file for MSVC compiled Firefox under Wine
874 342221 socket connect false positive uninit memory for unknown af family
875 342353 Allow dumping full massif output while valgrind is still running
876 342571 Valgrind chokes on AVX compare intrinsic with _CMP_GE_QS
880 342603 Add I2C_SMBUS ioctl support
881 342635 OS X 10.10 (Yosemite) - missing system calls and fcntl code
882 342683 Mark memory past the initial brk limit as unaddressable
883 342783 arm: unhandled instruction 0xEEFE1ACA = "vcvt.s32.f32 s3, s3, #12"
884 342795 Internal glibc __GI_mempcpy call should be intercepted
885 342841 s390x: Support instructions fiebr(a) and fidbr(a)
886 343012 Unhandled syscall 319 (memfd_create)
887 343069 Patch updating v4l2 API support
888 343173 helgrind crash during stack unwind
889 343219 fix GET_STARTREGS for arm
890 343303 Fix known deliberate memory leak in setenv() on Mac OS X 10.10
891 343306 OS X 10.10: UNKNOWN mach_msg unhandled MACH_SEND_TRAILER option
892 343332 Unhandled instruction 0x9E310021 (fcvtmu) on aarch64
893 343335 unhandled instruction 0x1E638400 (fccmp) aarch64
894 343523 OS X mach_ports_register: UNKNOWN task message [id 3403, to [..]
895 343525 OS X host_get_special_port: UNKNOWN host message [id 412, to [..]
896 343597 ppc64le: incorrect use of offseof macro
897 343649 OS X host_create_mach_voucher: UNKNOWN host message [id 222, to [..]
898 343663 OS X 10.10 Memchecj always reports a leak regardless of [..]
899 343732 Unhandled syscall 144 (setgid) on aarch64
900 343733 Unhandled syscall 187 (msgctl and related) on aarch64
901 343802 s390x: False positive "conditional jump or move depends on [..]
902 343902 --vgdb=yes doesn't break when --xml=yes is used
903 343967 Don't warn about setuid/setgid/setcap executable for directories
904 343978 Recognize DWARF5/GCC5 DW_LANG_Fortran 2003 and 2008 constants
905 344007 accept4 syscall unhandled on arm64 (242) and ppc64 (344)
906 344033 Helgrind on ARM32 loses track of mutex state in pthread_cond_wait
907 344054 www - update info for Solaris/illumos
908 344416 'make regtest' does not work cleanly on OS X
909 344235 Remove duplicate include of pub_core_aspacemgr.h
910 344279 syscall sendmmsg on arm64 (269) and ppc32/64 (349) unhandled
911 344295 syscall recvmmsg on arm64 (243) and ppc32/64 (343) unhandled
912 344307 2 unhandled syscalls on aarch64/arm64: umount2(39), mount (40)
913 344314 callgrind_annotate ... warnings about commands containing newlines
914 344318 socketcall should wrap recvmmsg and sendmmsg
915 344337 Fix unhandled syscall: mach:41 (_kernelrpc_mach_port_guard_trap)
916 344416 Fix 'make regtest' does not work cleanly on OS X
917 344499 Fix compilation for Linux kernel >= 4.0.0
918 344512 OS X: unhandled syscall: unix:348 (__pthread_chdir),
919 unix:349 (__pthread_fchdir)
920 344559 Garbage collection of unused segment names in address space manager
921 344560 Fix stack traces missing penultimate frame on OS X
922 344621 Fix memcheck/tests/err_disable4 test on OS X
923 344686 Fix suppression for pthread_rwlock_init on OS X 10.10
924 344702 Fix missing libobjc suppressions on OS X 10.10
926 344936 Fix unhandled syscall: unix:473 (readlinkat) on OS X 10.10
927 344939 Fix memcheck/tests/xml1 on OS X 10.10
928 345016 helgrind/tests/locked_vs_unlocked2 is failing sometimes
929 345079 Fix build problems in VEX/useful/test_main.c
930 345126 Incorrect handling of VIDIOC_G_AUDIO and G_AUDOUT
931 345177 arm64: prfm (reg) not implemented
932 345215 Performance improvements for the register allocator
933 345248 add support for Solaris OS in valgrind
934 345338 TIOCGSERIAL and TIOCSSERIAL ioctl support on Linux
935 345394 Fix memcheck/tests/strchr on OS X
936 345637 Fix memcheck/tests/sendmsg on OS X
937 345695 Add POWERPC support for AT_DCACHESIZE and HWCAP2
938 345824 Fix aspacem segment mismatch: seen with none/tests/bigcode
939 345887 Fix an assertion in the address space manager
940 345928 amd64: callstack only contains current function for small stacks
941 345984 disInstr(arm): unhandled instruction: 0xEE193F1E
942 345987 MIPS64: Implement cavium LHX instruction
943 346031 MIPS: Implement support for the CvmCount register (rhwr %0, 31)
944 346185 Fix typo saving altivec register v24
945 346267 Compiler warnings for PPC64 code on call to LibVEX_GuestPPC64_get_XER()
946 and LibVEX_GuestPPC64_get_CR()
947 346270 Regression tests none/tests/jm_vec/isa_2_07 and
948 none/tests/test_isa_2_07_part2 have failures on PPC64 little endian
949 346307 fuse filesystem syscall deadlocks
950 346324 PPC64 missing support for lbarx, lharx, stbcx and sthcx instructions
951 346411 MIPS: SysRes::_valEx handling is incorrect
952 346416 Add support for LL_IOC_PATH2FID and LL_IOC_GETPARENT Lustre ioctls
953 346474 PPC64 Power 8, spr TEXASRU register not supported
954 346487 Compiler generates "note" about a future ABI change for PPC64
955 346562 MIPS64: lwl/lwr instructions are performing 64bit loads
956 and causing spurious "invalid read of size 8" warnings
957 346801 Fix link error on OS X: _vgModuleLocal_sf_maybe_extend_stack
958 347151 Fix suppression for pthread_rwlock_init on OS X 10.8
959 347233 Fix memcheck/tests/strchr on OS X 10.10 (Haswell)
960 347322 Power PC regression test cleanup
961 347379 valgrind --leak-check=full leak errors from system libs on OS X 10.8
963 347389 unhandled syscall: 373 (Linux ARM syncfs)
964 347686 Patch set to cleanup PPC64 regtests
965 347978 Remove bash dependencies where not needed
966 347982 OS X: undefined symbols for architecture x86_64: "_global" [..]
967 347988 Memcheck: the 'impossible' happened: unexpected size for Addr (OSX/wine)
969 348102 Patch updating v4l2 API support
970 348247 amd64 front end: jno jumps wrongly when overflow is not set
971 348269 Improve mmap MAP_HUGETLB support.
972 348334 (ppc) valgrind does not simulate dcbfl - then my program terminates
973 348345 Assertion fails for negative lineno
974 348377 Unsupported ARM instruction: yield
975 348565 Fix detection of command line option availability for clang
976 348574 vex amd64->IR pcmpistri SSE4.2 unsupported (pcmpistri $0x18)
977 348728 Fix broken check for VIDIOC_G_ENC_INDEX
978 348748 Fix redundant condition
979 348890 Fix clang warning about unsupported --param inline-unit-growth=900
980 348949 Bogus "ERROR: --ignore-ranges: suspiciously large range"
981 349034 Add Lustre ioctls LL_IOC_GROUP_LOCK and LL_IOC_GROUP_UNLOCK
982 349086 Fix UNKNOWN task message [id 3406, to mach_task_self(), [..]
983 349087 Fix UNKNOWN task message [id 3410, to mach_task_self(), [..]
984 349626 Implemented additional Xen hypercalls
985 349769 Clang/osx: ld: warning: -read_only_relocs cannot be used with x86_64
986 349790 Clean up of the hardware capability checking utilities.
987 349828 memcpy intercepts memmove causing src/dst overlap error (ppc64 ld.so)
988 349874 Fix typos in source code
989 349879 memcheck: add handwritten assembly for helperc_LOADV*
990 349941 di_notify_mmap might create wrong start/size DebugInfoMapping
991 350062 vex x86->IR: 0x66 0xF 0x3A 0xB (ROUNDSD) on OS X
992 350202 Add limited param to 'monitor block_list'
993 350290 s390x: Support instructions fixbr(a)
994 350359 memcheck/tests/x86/fxsave hangs indefinetely on OS X
995 350809 Fix none/tests/async-sigs for Solaris
996 350811 Remove reference to --db-attach which has been removed.
997 350813 Memcheck/x86: enable handwritten assembly helpers for x86/Solaris too
998 350854 hard-to-understand code in VG_(load_ELF)()
999 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
1000 351386 Solaris: Cannot run ld.so.1 under Valgrind
1001 351474 Fix VG_(iseqsigset) as obvious
1002 351531 Typo in /include/vki/vki-xen-physdev.h header guard
1003 351756 Intercept platform_memchr$VARIANT$Haswell on OS X
1004 351858 ldsoexec support on Solaris
1005 351873 Newer gcc doesn't allow __builtin_tabortdc[i] in ppc32 mode
1006 352130 helgrind reports false races for printfs using mempcpy on FILE* state
1007 352284 s390: Conditional jump depends on uninitialised value(s) in vfprintf
1008 352320 arm64 crash on none/tests/nestedfs
1009 352765 Vbit test fails on Power 6
1010 352768 The mbar instruction is missing from the Power PC support
1011 352769 Power PC program priority register (PPR) is not supported
1012 n-i-bz Provide implementations of certain compiler builtins to support
1013 compilers that may not provide those
1014 n-i-bz Old STABS code is still being compiled, but never used. Remove it.
1015 n-i-bz Fix compilation on distros with glibc < 2.5
1016 n-i-bz (vex 3098) Avoid generation of Neon insns on non-Neon hosts
1017 n-i-bz Enable rt_sigpending syscall on ppc64 linux.
1018 n-i-bz mremap did not work properly on shared memory
1019 n-i-bz Fix incorrect sizeof expression in syswrap-xen.c reported by Coverity
1020 n-i-bz In VALGRIND_PRINTF write out thread name, if any, to xml
1022 (3.11.0.TEST1: 8 September 2015, vex r3187, valgrind r15646)
1023 (3.11.0.TEST2: 21 September 2015, vex r3193, valgrind r15667)
1024 (3.11.0: 22 September 2015, vex r3195, valgrind r15674)
1028 Release 3.10.1 (25 November 2014)
1029 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1030 3.10.1 is a bug fix release. It fixes various bugs reported in 3.10.0
1031 and backports fixes for all reported missing AArch64 ARMv8 instructions
1032 and syscalls from the trunk. If you package or deliver 3.10.0 for others
1033 to use, you might want to consider upgrading to 3.10.1 instead.
1035 The following bugs have been fixed or resolved. Note that "n-i-bz"
1036 stands for "not in bugzilla" -- that is, a bug that was reported to us
1037 but never got a bugzilla entry. We encourage you to file bugs in
1038 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1039 than mailing the developers (or mailing lists) directly -- bugs that
1040 are not entered into bugzilla tend to get forgotten about or ignored.
1042 To see details of a given bug, visit
1043 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1044 where XXXXXX is the bug number as listed below.
1046 335440 arm64: ld1 (single structure) is not implemented
1047 335713 arm64: unhanded instruction: prfm (immediate)
1048 339020 ppc64: memcheck/tests/ppc64/power_ISA2_05 failing in nightly build
1049 339182 ppc64: AvSplat ought to load destination vector register with [..]
1050 339336 PPC64 store quad instruction (stq) is not supposed to change [..]
1051 339433 ppc64 lxvw4x instruction uses four 32-byte loads
1052 339645 Use correct tag names in sys_getdents/64 wrappers
1053 339706 Fix false positive for ioctl(TIOCSIG) on linux
1054 339721 assertion 'check_sibling == sibling' failed in readdwarf3.c ...
1055 339853 arm64 times syscall unknown
1056 339855 arm64 unhandled getsid/setsid syscalls
1057 339858 arm64 dmb sy not implemented
1058 339926 Unhandled instruction 0x1E674001 (frintx) on aarm64
1059 339927 Unhandled instruction 0x9E7100C6 (fcvtmu) on aarch64
1060 339938 disInstr(arm64): unhandled instruction 0x4F8010A4 (fmla)
1062 339940 arm64: unhandled syscall: 83 (sys_fdatasync) + patch
1063 340033 arm64: unhandled insn dmb ishld and some other isb-dmb-dsb variants
1064 340028 unhandled syscalls for arm64 (msync, pread64, setreuid and setregid)
1065 340036 arm64: Unhandled instruction ld4 (multiple structures, no offset)
1066 340236 arm64: unhandled syscalls: mknodat, fchdir, chroot, fchownat
1067 340509 arm64: unhandled instruction fcvtas
1068 340630 arm64: fchmod (52) and fchown (55) syscalls not recognized
1069 340632 arm64: unhandled instruction fcvtas
1070 340722 Resolve "UNKNOWN attrlist flags 0:0x10000000"
1071 340725 AVX2: Incorrect decoding of vpbroadcast{b,w} reg,reg forms
1072 340788 warning: unhandled syscall: 318 (getrandom)
1073 340807 disInstr(arm): unhandled instruction: 0xEE989B20
1074 340856 disInstr(arm64): unhandled instruction 0x1E634C45 (fcsel)
1075 340922 arm64: unhandled getgroups/setgroups syscalls
1076 350251 Fix typo in VEX utility program (test_main.c).
1077 350407 arm64: unhandled instruction ucvtf (vector, integer)
1078 350809 none/tests/async-sigs breaks when run under cron on Solaris
1079 350811 update README.solaris after r15445
1080 350813 Use handwritten memcheck assembly helpers on x86/Solaris [..]
1081 350854 strange code in VG_(load_ELF)()
1082 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
1083 n-i-bz DRD and Helgrind: Handle Imbe_CancelReservation (clrex on ARM)
1084 n-i-bz Add missing ]] to terminate CDATA.
1085 n-i-bz Glibc versions prior to 2.5 do not define PTRACE_GETSIGINFO
1086 n-i-bz Enable sys_fadvise64_64 on arm32.
1087 n-i-bz Add test cases for all remaining AArch64 SIMD, FP and memory insns.
1088 n-i-bz Add test cases for all known arm64 load/store instructions.
1089 n-i-bz PRE(sys_openat): when checking whether ARG1 == VKI_AT_FDCWD [..]
1090 n-i-bz Add detection of old ppc32 magic instructions from bug 278808.
1091 n-i-bz exp-dhat: Implement missing function "dh_malloc_usable_size".
1092 n-i-bz arm64: Implement "fcvtpu w, s".
1093 n-i-bz arm64: implement ADDP and various others
1094 n-i-bz arm64: Implement {S,U}CVTF (scalar, fixedpt).
1095 n-i-bz arm64: enable FCVT{A,N}S X,S.
1097 (3.10.1: 25 November 2014, vex r3026, valgrind r14785)
1101 Release 3.10.0 (10 September 2014)
1102 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1104 3.10.0 is a feature release with many improvements and the usual
1105 collection of bug fixes.
1107 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1108 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1109 MIPS64/Linux, ARM/Android, MIPS32/Android, X86/Android, X86/MacOSX 10.9
1110 and AMD64/MacOSX 10.9. Support for MacOSX 10.8 and 10.9 is
1111 significantly improved relative to the 3.9.0 release.
1113 * ================== PLATFORM CHANGES =================
1115 * Support for the 64-bit ARM Architecture (AArch64 ARMv8). This port
1116 is mostly complete, and is usable, but some SIMD instructions are as
1119 * Support for little-endian variant of the 64-bit POWER architecture.
1121 * Support for Android on MIPS32.
1123 * Support for 64bit FPU on MIPS32 platforms.
1125 * Both 32- and 64-bit executables are supported on MacOSX 10.8 and 10.9.
1127 * Configuration for and running on Android targets has changed.
1128 See README.android in the source tree for details.
1130 * ================== DEPRECATED FEATURES =================
1132 * --db-attach is now deprecated and will be removed in the next
1133 valgrind feature release. The built-in GDB server capabilities are
1134 superior and should be used instead. Learn more here:
1135 http://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.gdbserver
1137 * ==================== TOOL CHANGES ====================
1141 - Client code can now selectively disable and re-enable reporting of
1142 invalid address errors in specific ranges using the new client
1143 requests VALGRIND_DISABLE_ADDR_ERROR_REPORTING_IN_RANGE and
1144 VALGRIND_ENABLE_ADDR_ERROR_REPORTING_IN_RANGE.
1146 - Leak checker: there is a new leak check heuristic called
1147 "length64". This is used to detect interior pointers pointing 8
1148 bytes inside a block, on the assumption that the first 8 bytes
1149 holds the value "block size - 8". This is used by
1150 sqlite3MemMalloc, for example.
1152 - Checking of system call parameters: if a syscall parameter
1153 (e.g. bind struct sockaddr, sendmsg struct msghdr, ...) has
1154 several fields not initialised, an error is now reported for each
1155 field. Previously, an error was reported only for the first
1156 uninitialised field.
1158 - Mismatched alloc/free checking: a new flag
1159 --show-mismatched-frees=no|yes [yes] makes it possible to turn off
1160 such checks if necessary.
1164 - Improvements to error messages:
1166 o Race condition error message involving heap allocated blocks also
1167 show the thread number that allocated the raced-on block.
1169 o All locks referenced by an error message are now announced.
1170 Previously, some error messages only showed the lock addresses.
1172 o The message indicating where a lock was first observed now also
1173 describes the address/location of the lock.
1175 - Helgrind now understands the Ada task termination rules and
1176 creates a happens-before relationship between a terminated task
1177 and its master. This avoids some false positives and avoids a big
1178 memory leak when a lot of Ada tasks are created and terminated.
1179 The interceptions are only activated with forthcoming releases of
1180 gnatpro >= 7.3.0w-20140611 and gcc >= 5.0.
1182 - A new GDB server monitor command "info locks" giving the list of
1183 locks, their location, and their status.
1187 - callgrind_control now supports the --vgdb-prefix argument,
1188 which is needed if valgrind was started with this same argument.
1190 * ==================== OTHER CHANGES ====================
1192 * Unwinding through inlined function calls. Stack unwinding can now
1193 make use of Dwarf3 inlined-unwind information if it is available.
1194 The practical effect is that inlined calls become visible in stack
1195 traces. The suppression matching machinery has been adjusted
1196 accordingly. This is controlled by the new option
1197 --read-inline-info=yes|no. Currently this is enabled by default
1198 only on Linux and Android targets and only for the tools Memcheck,
1201 * Valgrind can now read EXIDX unwind information on 32-bit ARM
1202 targets. If an object contains both CFI and EXIDX unwind
1203 information, Valgrind will prefer the CFI over the EXIDX. This
1204 facilitates unwinding through system libraries on arm-android
1207 * Address description logic has been improved and is now common
1208 between Memcheck and Helgrind, resulting in better address
1209 descriptions for some kinds of error messages.
1211 * Error messages about dubious arguments (eg, to malloc or calloc) are
1212 output like other errors. This means that they can be suppressed
1213 and they have a stack trace.
1215 * The C++ demangler has been updated for better C++11 support.
1217 * New and modified GDB server monitor features:
1219 - Thread local variables/storage (__thread) can now be displayed.
1221 - The GDB server monitor command "v.info location <address>"
1222 displays information about an address. The information produced
1223 depends on the tool and on the options given to valgrind.
1224 Possibly, the following are described: global variables, local
1225 (stack) variables, allocated or freed blocks, ...
1227 - The option "--vgdb-stop-at=event1,event2,..." allows the user to
1228 ask the GDB server to stop at the start of program execution, at
1229 the end of the program execution and on Valgrind internal errors.
1231 - A new monitor command "v.info stats" shows various Valgrind core
1232 and tool statistics.
1234 - A new monitor command "v.set hostvisibility" allows the GDB server
1235 to provide access to Valgrind internal host status/memory.
1237 * A new option "--aspace-minaddr=<address>" can in some situations
1238 allow the use of more memory by decreasing the address above which
1239 Valgrind maps memory. It can also be used to solve address
1240 conflicts with system libraries by increasing the default value.
1241 See user manual for details.
1243 * The amount of memory used by Valgrind to store debug info (unwind
1244 info, line number information and symbol data) has been
1245 significantly reduced, even though Valgrind now reads more
1246 information in order to support unwinding of inlined function calls.
1248 * Dwarf3 handling with --read-var-info=yes has been improved:
1250 - Ada and C struct containing VLAs no longer cause a "bad DIE" error
1252 - Code compiled with
1253 -ffunction-sections -fdata-sections -Wl,--gc-sections
1254 no longer causes assertion failures.
1256 * Improved checking for the --sim-hints= and --kernel-variant=
1257 options. Unknown strings are now detected and reported to the user
1260 * The semantics of stack start/end boundaries in the valgrind.h
1261 VALGRIND_STACK_REGISTER client request has been clarified and
1262 documented. The convention is that start and end are respectively
1263 the lowest and highest addressable bytes of the stack.
1265 * ==================== FIXED BUGS ====================
1267 The following bugs have been fixed or resolved. Note that "n-i-bz"
1268 stands for "not in bugzilla" -- that is, a bug that was reported to us
1269 but never got a bugzilla entry. We encourage you to file bugs in
1270 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1271 than mailing the developers (or mailing lists) directly -- bugs that
1272 are not entered into bugzilla tend to get forgotten about or ignored.
1274 To see details of a given bug, visit
1275 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1276 where XXXXXX is the bug number as listed below.
1278 175819 Support for ipv6 socket reporting with --track-fds
1279 232510 make distcheck fails
1280 249435 Analyzing wine programs with callgrind triggers a crash
1281 278972 support for inlined function calls in stacktraces and suppression
1283 291310 FXSAVE instruction marks memory as undefined on amd64
1284 303536 ioctl for SIOCETHTOOL (ethtool(8)) isn't wrapped
1285 308729 vex x86->IR: unhandled instruction bytes 0xf 0x5 (syscall)
1286 315199 vgcore file for threaded app does not show which thread crashed
1287 315952 tun/tap ioctls are not supported
1288 323178 Unhandled instruction: PLDW register (ARM)
1289 323179 Unhandled instruction: PLDW immediate (ARM)
1290 324050 Helgrind: SEGV because of unaligned stack when using movdqa
1291 325110 Add test-cases for Power ISA 2.06 insns: divdo/divdo. and divduo/divduo.
1292 325124 [MIPSEL] Compilation error
1293 325477 Phase 4 support for IBM Power ISA 2.07
1294 325538 cavium octeon mips64, valgrind reported "dumping core" [...]
1295 325628 Phase 5 support for IBM Power ISA 2.07
1296 325714 Empty vgcore but RLIMIT_CORE is big enough (too big)
1297 325751 Missing the two privileged Power PC Transactional Memory Instructions
1298 325816 Phase 6 support for IBM Power ISA 2.07
1299 325856 Make SGCheck fail gracefully on unsupported platforms
1300 326026 Iop names for count leading zeros/sign bits incorrectly imply [..]
1301 326436 DRD: False positive in libstdc++ std::list::push_back
1302 326444 Cavium MIPS Octeon Specific Load Indexed Instructions
1303 326462 Refactor vgdb to isolate invoker stuff into separate module
1304 326469 amd64->IR: 0x66 0xF 0x3A 0x63 0xC1 0xE (pcmpistri 0x0E)
1305 326623 DRD: false positive conflict report in a field assignment
1306 326724 Valgrind does not compile on OSX 1.9 Mavericks
1307 326816 Intercept for __strncpy_sse2_unaligned missing?
1308 326921 coregrind fails to compile m_trampoline.S with MIPS/Linux port of V
1309 326983 Clear direction flag after tests on amd64.
1310 327212 Do not prepend the current directory to absolute path names.
1311 327223 Support for Cavium MIPS Octeon Atomic and Count Instructions
1312 327238 Callgrind Assertion 'passed <= last_bb->cjmp_count' failed
1313 327284 s390x: Fix translation of the risbg instruction
1314 327639 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x34
1315 327837 dwz compressed alternate .debug_info and .debug_str not read correctly
1316 327916 DW_TAG_typedef may have no name
1317 327943 s390x: add a redirection for the 'index' function
1318 328100 XABORT not implemented
1319 328205 Implement additional Xen hypercalls
1320 328454 add support Backtraces with ARM unwind tables (EXIDX)
1321 328455 s390x: SIGILL after emitting wrong register pair for ldxbr
1322 328711 valgrind.1 manpage "memcheck options" section is badly generated
1323 328878 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x14
1324 329612 Incorrect handling of AT_BASE for image execution
1325 329694 clang warns about using uninitialized variable
1326 329956 valgrind crashes when lmw/stmw instructions are used on ppc64
1327 330228 mmap must align to VKI_SHMLBA on mips32
1328 330257 LLVM does not support `-mno-dynamic-no-pic` option
1329 330319 amd64->IR: unhandled instruction bytes: 0xF 0x1 0xD5 (xend)
1330 330459 --track-fds=yes doesn't track eventfds
1331 330469 Add clock_adjtime syscall support
1332 330594 Missing sysalls on PowerPC / uClibc
1333 330622 Add test to regression suite for POWER instruction: dcbzl
1334 330939 Support for AMD's syscall instruction on x86
1336 330941 Typo in PRE(poll) syscall wrapper
1337 331057 unhandled instruction: 0xEEE01B20 (vfma.f64) (has patch)
1338 331254 Fix expected output for memcheck/tests/dw4
1339 331255 Fix race condition in test none/tests/coolo_sigaction
1340 331257 Fix type of jump buffer in test none/tests/faultstatus
1341 331305 configure uses bash specific syntax
1342 331337 s390x WARNING: unhandled syscall: 326 (dup3)
1343 331380 Syscall param timer_create(evp) points to uninitialised byte(s)
1344 331476 Patch to handle ioctl 0x5422 on Linux (x86 and amd64)
1345 331829 Unexpected ioctl opcode sign extension
1346 331830 ppc64: WARNING: unhandled syscall: 96/97
1347 331839 drd/tests/sem_open specifies invalid semaphore name
1348 331847 outcome of drd/tests/thread_name is nondeterministic
1349 332037 Valgrind cannot handle Thumb "add pc, reg"
1350 332055 drd asserts on platforms with VG_STACK_REDZONE_SZB == 0 and
1351 consistency checks enabled
1352 332263 intercepts for pthread_rwlock_timedrdlock and
1353 pthread_rwlock_timedwrlock are incorrect
1354 332265 drd could do with post-rwlock_init and pre-rwlock_destroy
1356 332276 Implement additional Xen hypercalls
1357 332658 ldrd.w r1, r2, [PC, #imm] does not adjust for 32bit alignment
1358 332765 Fix ms_print to create temporary files in a proper directory
1359 333072 drd: Add semaphore annotations
1360 333145 Tests for missaligned PC+#imm access for arm
1361 333228 AAarch64 Missing instruction encoding: mrs %[reg], ctr_el0
1362 333230 AAarch64 missing instruction encodings: dc, ic, dsb.
1363 333248 WARNING: unhandled syscall: unix:443
1364 333428 ldr.w pc [rD, #imm] instruction leads to assertion
1365 333501 cachegrind: assertion: Cache set count is not a power of two.
1368 333666 Recognize MPX instructions and bnd prefix.
1369 333788 Valgrind does not support the CDROM_DISC_STATUS ioctl (has patch)
1370 333817 Valgrind reports the memory areas written to by the SG_IO
1372 334049 lzcnt fails silently (x86_32)
1373 334384 Valgrind does not have support Little Endian support for
1375 334585 recvmmsg unhandled (+patch) (arm)
1376 334705 sendmsg and recvmsg should guard against bogus msghdr fields.
1377 334727 Build fails with -Werror=format-security
1378 334788 clarify doc about --log-file initial program directory
1379 334834 PPC64 Little Endian support, patch 2
1380 334836 PPC64 Little Endian support, patch 3 testcase fixes
1381 334936 patch to fix false positives on alsa SNDRV_CTL_* ioctls
1382 335034 Unhandled ioctl: HCIGETDEVLIST
1383 335155 vgdb, fix error print statement.
1384 335262 arm64: movi 8bit version is not supported
1385 335263 arm64: dmb instruction is not implemented
1386 335441 unhandled ioctl 0x8905 (SIOCATMARK) when running wine under valgrind
1387 335496 arm64: sbc/abc instructions are not implemented
1388 335554 arm64: unhandled instruction: abs
1389 335564 arm64: unhandled instruction: fcvtpu Xn, Sn
1390 335735 arm64: unhandled instruction: cnt
1391 335736 arm64: unhandled instruction: uaddlv
1392 335848 arm64: unhandled instruction: {s,u}cvtf
1393 335902 arm64: unhandled instruction: sli
1394 335903 arm64: unhandled instruction: umull (vector)
1395 336055 arm64: unhandled instruction: mov (element)
1396 336062 arm64: unhandled instruction: shrn{,2}
1397 336139 mip64: [...] valgrind hangs and spins on a single core [...]
1398 336189 arm64: unhandled Instruction: mvn
1399 336435 Valgrind hangs in pthread_spin_lock consuming 100% CPU
1400 336619 valgrind --read-var-info=yes doesn't handle DW_TAG_restrict_type
1401 336772 Make moans about unknown ioctls more informative
1402 336957 Add a section about the Solaris/illumos port on the webpage
1403 337094 ifunc wrapper is broken on ppc64
1404 337285 fcntl commands F_OFD_SETLK, F_OFD_SETLKW, and F_OFD_GETLK not supported
1405 337528 leak check heuristic for block prefixed by length as 64bit number
1406 337740 Implement additional Xen hypercalls
1407 337762 guest_arm64_toIR.c:4166 (dis_ARM64_load_store): Assertion `0' failed.
1408 337766 arm64-linux: unhandled syscalls mlock (228) and mlockall (230)
1409 337871 deprecate --db-attach
1410 338023 Add support for all V4L2/media ioctls
1411 338024 inlined functions are not shown if DW_AT_ranges is used
1412 338106 Add support for 'kcmp' syscall
1413 338115 DRD: computed conflict set differs from actual after fork
1414 338160 implement display of thread local storage in gdbsrv
1415 338205 configure.ac and check for -Wno-tautological-compare
1416 338300 coredumps are missing one byte of every segment
1417 338445 amd64 vbit-test fails with unknown opcodes used by arm64 VEX
1418 338499 --sim-hints parsing broken due to wrong order in tokens
1419 338615 suppress glibc 2.20 optimized strcmp implementation for ARMv7
1420 338681 Unable to unwind through clone thread created on i386-linux
1421 338698 race condition between gdbsrv and vgdb on startup
1422 338703 helgrind on arm-linux gets false positives in dynamic loader
1423 338791 alt dwz files can be relative of debug/main file
1424 338878 on MacOS: assertion 'VG_IS_PAGE_ALIGNED(clstack_end+1)' failed
1425 338932 build V-trunk with gcc-trunk
1426 338974 glibc 2.20 changed size of struct sigaction sa_flags field on s390
1427 345079 Fix build problems in VEX/useful/test_main.c
1428 n-i-bz Fix KVM_CREATE_IRQCHIP ioctl handling
1429 n-i-bz s390x: Fix memory corruption for multithreaded applications
1430 n-i-bz vex arm->IR: allow PC as basereg in some LDRD cases
1431 n-i-bz internal error in Valgrind if vgdb transmit signals when ptrace invoked
1432 n-i-bz Fix mingw64 support in valgrind.h (dev@, 9 May 2014)
1433 n-i-bz drd manual: Document how to C++11 programs that use class "std::thread"
1434 n-i-bz Add command-line option --default-suppressions
1435 n-i-bz Add support for BLKDISCARDZEROES ioctl
1436 n-i-bz ppc32/64: fix a regression with the mtfsb0/mtfsb1 instructions
1437 n-i-bz Add support for sys_pivot_root and sys_unshare
1439 (3.10.0.BETA1: 2 September 2014, vex r2940, valgrind r14428)
1440 (3.10.0.BETA2: 8 September 2014, vex r2950, valgrind r14503)
1441 (3.10.0: 10 September 2014, vex r2950, valgrind r14514)
1445 Release 3.9.0 (31 October 2013)
1446 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1447 3.9.0 is a feature release with many improvements and the usual
1448 collection of bug fixes.
1450 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
1451 PPC64/Linux, S390X/Linux, MIPS32/Linux, MIPS64/Linux, ARM/Android,
1452 X86/Android, X86/MacOSX 10.7 and AMD64/MacOSX 10.7. Support for
1453 MacOSX 10.8 is significantly improved relative to the 3.8.0 release.
1455 * ================== PLATFORM CHANGES =================
1457 * Support for MIPS64 LE and BE running Linux. Valgrind has been
1458 tested on MIPS64 Debian Squeeze and Debian Wheezy distributions.
1460 * Support for MIPS DSP ASE on MIPS32 platforms.
1462 * Support for s390x Decimal Floating Point instructions on hosts that
1463 have the DFP facility installed.
1465 * Support for POWER8 (Power ISA 2.07) instructions
1467 * Support for Intel AVX2 instructions. This is available only on 64
1470 * Initial support for Intel Transactional Synchronization Extensions,
1473 * Initial support for Hardware Transactional Memory on POWER.
1475 * Improved support for MacOSX 10.8 (64-bit only). Memcheck can now
1476 run large GUI apps tolerably well.
1478 * ==================== TOOL CHANGES ====================
1482 - Improvements in handling of vectorised code, leading to
1483 significantly fewer false error reports. You need to use the flag
1484 --partial-loads-ok=yes to get the benefits of these changes.
1486 - Better control over the leak checker. It is now possible to
1487 specify which leak kinds (definite/indirect/possible/reachable)
1488 should be displayed, which should be regarded as errors, and which
1489 should be suppressed by a given leak suppression. This is done
1490 using the options --show-leak-kinds=kind1,kind2,..,
1491 --errors-for-leak-kinds=kind1,kind2,.. and an optional
1492 "match-leak-kinds:" line in suppression entries, respectively.
1494 Note that generated leak suppressions contain this new line and
1495 are therefore more specific than in previous releases. To get the
1496 same behaviour as previous releases, remove the "match-leak-kinds:"
1497 line from generated suppressions before using them.
1499 - Reduced "possible leak" reports from the leak checker by the use
1500 of better heuristics. The available heuristics provide detection
1501 of valid interior pointers to std::stdstring, to new[] allocated
1502 arrays with elements having destructors and to interior pointers
1503 pointing to an inner part of a C++ object using multiple
1504 inheritance. They can be selected individually using the
1505 option --leak-check-heuristics=heur1,heur2,...
1507 - Better control of stacktrace acquisition for heap-allocated
1508 blocks. Using the --keep-stacktraces option, it is possible to
1509 control independently whether a stack trace is acquired for each
1510 allocation and deallocation. This can be used to create better
1511 "use after free" errors or to decrease Valgrind's resource
1512 consumption by recording less information.
1514 - Better reporting of leak suppression usage. The list of used
1515 suppressions (shown when the -v option is given) now shows, for
1516 each leak suppressions, how many blocks and bytes it suppressed
1517 during the last leak search.
1521 - False errors resulting from the use of statically initialised
1522 mutexes and condition variables (PTHREAD_MUTEX_INITIALISER, etc)
1525 - False errors resulting from the use of pthread_cond_waits that
1526 timeout, have been removed.
1528 * ==================== OTHER CHANGES ====================
1530 * Some attempt to tune Valgrind's space requirements to the expected
1531 capabilities of the target:
1533 - The default size of the translation cache has been reduced from 8
1534 sectors to 6 on Android platforms, since each sector occupies
1535 about 40MB when using Memcheck.
1537 - The default size of the translation cache has been increased to 16
1538 sectors on all other platforms, reflecting the fact that large
1539 applications require instrumentation and storage of huge amounts
1540 of code. For similar reasons, the number of memory mapped
1541 segments that can be tracked has been increased by a factor of 6.
1543 - In all cases, the maximum number of sectors in the translation
1544 cache can be controlled by the new flag --num-transtab-sectors.
1546 * Changes in how debug info (line numbers, etc) is read:
1548 - Valgrind no longer temporarily mmaps the entire object to read
1549 from it. Instead, reading is done through a small fixed sized
1550 buffer. This avoids virtual memory usage spikes when Valgrind
1551 reads debuginfo from large shared objects.
1553 - A new experimental remote debug info server. Valgrind can read
1554 debug info from a different machine (typically, a build host)
1555 where debuginfo objects are stored. This can save a lot of time
1556 and hassle when running Valgrind on resource-constrained targets
1557 (phones, tablets) when the full debuginfo objects are stored
1558 somewhere else. This is enabled by the --debuginfo-server=
1561 - Consistency checking between main and debug objects can be
1562 disabled using the --allow-mismatched-debuginfo option.
1564 * Stack unwinding by stack scanning, on ARM. Unwinding by stack
1565 scanning can recover stack traces in some cases when the normal
1566 unwind mechanisms fail. Stack scanning is best described as "a
1567 nasty, dangerous and misleading hack" and so is disabled by default.
1568 Use --unw-stack-scan-thresh and --unw-stack-scan-frames to enable
1571 * Detection and merging of recursive stack frame cycles. When your
1572 program has recursive algorithms, this limits the memory used by
1573 Valgrind for recorded stack traces and avoids recording
1574 uninteresting repeated calls. This is controlled by the command
1575 line option --merge-recursive-frame and by the monitor command
1576 "v.set merge-recursive-frames".
1578 * File name and line numbers for used suppressions. The list of used
1579 suppressions (shown when the -v option is given) now shows, for each
1580 used suppression, the file name and line number where the suppression
1583 * New and modified GDB server monitor features:
1585 - valgrind.h has a new client request, VALGRIND_MONITOR_COMMAND,
1586 that can be used to execute gdbserver monitor commands from the
1589 - A new monitor command, "v.info open_fds", that gives the list of
1590 open file descriptors and additional details.
1592 - An optional message in the "v.info n_errs_found" monitor command,
1593 for example "v.info n_errs_found test 1234 finished", allowing a
1594 comment string to be added to the process output, perhaps for the
1595 purpose of separating errors of different tests or test phases.
1597 - A new monitor command "v.info execontext" that shows information
1598 about the stack traces recorded by Valgrind.
1600 - A new monitor command "v.do expensive_sanity_check_general" to run
1601 some internal consistency checks.
1603 * New flag --sigill-diagnostics to control whether a diagnostic
1604 message is printed when the JIT encounters an instruction it can't
1605 translate. The actual behavior -- delivery of SIGILL to the
1606 application -- is unchanged.
1608 * The maximum amount of memory that Valgrind can use on 64 bit targets
1609 has been increased from 32GB to 64GB. This should make it possible
1610 to run applications on Memcheck that natively require up to about 35GB.
1612 * ==================== FIXED BUGS ====================
1614 The following bugs have been fixed or resolved. Note that "n-i-bz"
1615 stands for "not in bugzilla" -- that is, a bug that was reported to us
1616 but never got a bugzilla entry. We encourage you to file bugs in
1617 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1618 than mailing the developers (or mailing lists) directly -- bugs that
1619 are not entered into bugzilla tend to get forgotten about or ignored.
1621 To see details of a given bug, visit
1622 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1623 where XXXXXX is the bug number as listed below.
1625 123837 system call: 4th argument is optional, depending on cmd
1626 135425 memcheck should tell you where Freed blocks were Mallocd
1627 164485 VG_N_SEGNAMES and VG_N_SEGMENTS are (still) too small
1628 207815 Adds some of the drm ioctls to syswrap-linux.c
1629 251569 vex amd64->IR: 0xF 0x1 0xF9 0xBF 0x90 0xD0 0x3 0x0 (RDTSCP)
1630 252955 Impossible to compile with ccache
1631 253519 Memcheck reports auxv pointer accesses as invalid reads.
1632 263034 Crash when loading some PPC64 binaries
1633 269599 Increase deepest backtrace
1634 274695 s390x: Support "compare to/from logical" instructions (z196)
1635 275800 s390x: Autodetect cache info (part 2)
1636 280271 Valgrind reports possible memory leaks on still-reachable std::string
1637 284540 Memcheck shouldn't count suppressions matching still-reachable [..]
1638 289578 Backtraces with ARM unwind tables (stack scan flags)
1639 296311 Wrong stack traces due to -fomit-frame-pointer (x86)
1640 304832 ppc32: build failure
1641 305431 Use find_buildid shdr fallback for separate .debug files
1642 305728 Add support for AVX2 instructions
1643 305948 ppc64: code generation for ShlD64 / ShrD64 asserts
1644 306035 s390x: Fix IR generation for LAAG and friends
1645 306054 s390x: Condition code computation for convert-to-int/logical
1646 306098 s390x: alternate opcode form for convert to/from fixed
1647 306587 Fix cache line detection from auxiliary vector for PPC.
1648 306783 Mips unhandled syscall : 4025 / 4079 / 4182
1649 307038 DWARF2 CFI reader: unhandled DW_OP_ opcode 0x8 (DW_OP_const1u et al)
1650 307082 HG false positive: pthread_cond_destroy: destruction of unknown CV
1651 307101 sys_capget second argument can be NULL
1652 307103 sys_openat: If pathname is absolute, then dirfd is ignored.
1653 307106 amd64->IR: f0 0f c0 02 (lock xadd byte)
1654 307113 s390x: DFP support
1655 307141 valgrind does't work in mips-linux system
1656 307155 filter_gdb should filter out syscall-template.S T_PSEUDO
1657 307285 x86_amd64 feature test for avx in test suite is wrong
1658 307290 memcheck overlap testcase needs memcpy version filter
1659 307463 Please add "&limit=0" to the "all open bugs" link
1660 307465 --show-possibly-lost=no should reduce the error count / exit code
1661 307557 Leaks on Mac OS X 10.7.5 libraries at ImageLoader::recursiveInit[..]
1662 307729 pkgconfig support broken valgrind.pc
1663 307828 Memcheck false errors SSE optimized wcscpy, wcscmp, wcsrchr, wcschr
1664 307955 Building valgrind 3.7.0-r4 fails in Gentoo AMD64 when using clang
1665 308089 Unhandled syscall on ppc64: prctl
1666 308135 PPC32 MPC8xx has 16 bytes cache size
1667 308321 testsuite memcheck filter interferes with gdb_filter
1669 308341 vgdb should report process exit (or fatal signal)
1670 308427 s390 memcheck reports tsearch cjump/cmove depends on uninit
1671 308495 Remove build dependency on installed Xen headers
1672 308573 Internal error on 64-bit instruction executed in 32-bit mode
1674 308627 pmovmskb validity bit propagation is imprecise
1675 308644 vgdb command for having the info for the track-fds option
1676 308711 give more info about aspacemgr and arenas in out_of_memory
1677 308717 ARM: implement fixed-point VCVT.F64.[SU]32
1678 308718 ARM implement SMLALBB family of instructions
1679 308886 Missing support for PTRACE_SET/GETREGSET
1680 308930 syscall name_to_handle_at (303 on amd64) not handled
1681 309229 V-bit tester does not report number of tests generated
1682 309323 print unrecognized instuction on MIPS
1683 309425 Provide a --sigill-diagnostics flag to suppress illegal [..]
1684 309427 SSE optimized stpncpy trigger uninitialised value [..] errors
1685 309430 Self hosting ppc64 encounters a vassert error on operand type
1686 309600 valgrind is a bit confused about 0-sized sections
1687 309823 Generate errors for still reachable blocks
1688 309921 PCMPISTRI validity bit propagation is imprecise
1689 309922 none/tests/ppc64/test_dfp5 sometimes fails
1690 310169 The Iop_CmpORD class of Iops is not supported by the vbit checker.
1691 310424 --read-var-info does not properly describe static variables
1692 310792 search additional path for debug symbols
1693 310931 s390x: Message-security assist (MSA) instruction extension [..]
1694 311100 PPC DFP implementation of the integer operands is inconsistent [..]
1695 311318 ARM: "128-bit constant is not implemented" error message
1696 311407 ssse3 bcopy (actually converted memcpy) causes invalid read [..]
1697 311690 V crashes because it redirects branches inside of a redirected function
1698 311880 x86_64: make regtest hangs at shell_valid1
1699 311922 WARNING: unhandled syscall: 170
1701 312171 ppc: insn selection for DFP
1702 312571 Rounding mode call wrong for the DFP Iops [..]
1703 312620 Change to Iop_D32toD64 [..] for s390 DFP support broke ppc [..]
1704 312913 Dangling pointers error should also report the alloc stack trace
1705 312980 Building on Mountain Lion generates some compiler warnings
1706 313267 Adding MIPS64/Linux port to Valgrind
1709 313811 Buffer overflow in assert_fail
1710 314099 coverity pointed out error in VEX guest_ppc_toIR.c insn_suffix
1711 314269 ppc: dead code in insn selection
1712 314718 ARM: implement integer divide instruction (sdiv and udiv)
1713 315345 cl-format.xml and callgrind/dump.c don't agree on using cfl= or cfi=
1714 315441 sendmsg syscall should ignore unset msghdr msg_flags
1715 315534 msgrcv inside a thread causes valgrind to hang (block)
1716 315545 Assertion '(UChar*)sec->tt[tteNo].tcptr <= (UChar*)hcode' failed
1717 315689 disInstr(thumb): unhandled instruction: 0xF852 0x0E10 (LDRT)
1718 315738 disInstr(arm): unhandled instruction: 0xEEBE0BEE (vcvt.s32.f64)
1719 315959 valgrind man page has bogus SGCHECK (and no BBV) OPTIONS section
1720 316144 valgrind.1 manpage contains unknown ??? strings [..]
1721 316145 callgrind command line options in manpage reference (unknown) [..]
1722 316145 callgrind command line options in manpage reference [..]
1723 316181 drd: Fixed a 4x slowdown for certain applications
1724 316503 Valgrind does not support SSE4 "movntdqa" instruction
1725 316535 Use of |signed int| instead of |size_t| in valgrind messages
1726 316696 fluidanimate program of parsec 2.1 stuck
1727 316761 syscall open_by_handle_at (304 on amd64, 342 on x86) not handled
1728 317091 Use -Wl,-Ttext-segment when static linking if possible [..]
1729 317186 "Impossible happens" when occurs VCVT instruction on ARM
1730 317318 Support for Threading Building Blocks "scalable_malloc"
1731 317444 amd64->IR: 0xC4 0x41 0x2C 0xC2 0xD2 0x8 (vcmpeq_uqps)
1732 317461 Fix BMI assembler configure check and avx2/bmi/fma vgtest prereqs
1733 317463 bmi testcase IR SANITY CHECK FAILURE
1734 317506 memcheck/tests/vbit-test fails with unknown opcode after [..]
1735 318050 libmpiwrap fails to compile with out-of-source build
1736 318203 setsockopt handling needs to handle SOL_SOCKET/SO_ATTACH_FILTER
1737 318643 annotate_trace_memory tests infinite loop on arm and ppc [..]
1738 318773 amd64->IR: 0xF3 0x48 0x0F 0xBC 0xC2 0xC3 0x66 0x0F
1739 318929 Crash with: disInstr(thumb): 0xF321 0x0001 (ssat16)
1740 318932 Add missing PPC64 and PPC32 system call support
1741 319235 --db-attach=yes is broken with Yama (ptrace scoping) enabled
1742 319395 Crash with unhandled instruction on STRT (Thumb) instructions
1743 319494 VEX Makefile-gcc standalone build update after r2702
1744 319505 [MIPSEL] Crash: unhandled UNRAY operator.
1745 319858 disInstr(thumb): unhandled instruction on instruction STRBT
1746 319932 disInstr(thumb): unhandled instruction on instruction STRHT
1747 320057 Problems when we try to mmap more than 12 memory pages on MIPS32
1748 320063 Memory from PTRACE_GET_THREAD_AREA is reported uninitialised
1749 320083 disInstr(thumb): unhandled instruction on instruction LDRBT
1750 320116 bind on AF_BLUETOOTH produces warnings because of sockaddr_rc padding
1751 320131 WARNING: unhandled syscall: 369 on ARM (prlimit64)
1752 320211 Stack buffer overflow in ./coregrind/m_main.c with huge TMPDIR
1753 320661 vgModuleLocal_read_elf_debug_info(): "Assertion '!di->soname'
1754 320895 add fanotify support (patch included)
1755 320998 vex amd64->IR pcmpestri and pcmpestrm SSE4.2 instruction
1756 321065 Valgrind updates for Xen 4.3
1757 321148 Unhandled instruction: PLI (Thumb 1, 2, 3)
1758 321363 Unhandled instruction: SSAX (ARM + Thumb)
1759 321364 Unhandled instruction: SXTAB16 (ARM + Thumb)
1760 321466 Unhandled instruction: SHASX (ARM + Thumb)
1761 321467 Unhandled instruction: SHSAX (ARM + Thumb)
1762 321468 Unhandled instruction: SHSUB16 (ARM + Thumb)
1763 321619 Unhandled instruction: SHSUB8 (ARM + Thumb)
1764 321620 Unhandled instruction: UASX (ARM + Thumb)
1765 321621 Unhandled instruction: USAX (ARM + Thumb)
1766 321692 Unhandled instruction: UQADD16 (ARM + Thumb)
1767 321693 Unhandled instruction: LDRSBT (Thumb)
1768 321694 Unhandled instruction: UQASX (ARM + Thumb)
1769 321696 Unhandled instruction: UQSAX (Thumb + ARM)
1770 321697 Unhandled instruction: UHASX (ARM + Thumb)
1771 321703 Unhandled instruction: UHSAX (ARM + Thumb)
1772 321704 Unhandled instruction: REVSH (ARM + Thumb)
1773 321730 Add cg_diff and cg_merge man pages
1774 321738 Add vgdb and valgrind-listener man pages
1776 321891 Unhandled instruction: LDRHT (Thumb)
1777 321960 pthread_create() then alloca() causing invalid stack write errors
1778 321969 ppc32 and ppc64 don't support [lf]setxattr
1779 322254 Show threadname together with tid if set by application
1780 322294 Add initial support for IBM Power ISA 2.07
1781 322368 Assertion failure in wqthread_hijack under OS X 10.8
1782 322563 vex mips->IR: 0x70 0x83 0xF0 0x3A
1783 322807 VALGRIND_PRINTF_BACKTRACE writes callstack to xml and text to stderr
1784 322851 0bXXX binary literal syntax is not standard
1785 323035 Unhandled instruction: LDRSHT(Thumb)
1786 323036 Unhandled instruction: SMMLS (ARM and Thumb)
1787 323116 The memcheck/tests/ppc64/power_ISA2_05.c fails to build [..]
1788 323175 Unhandled instruction: SMLALD (ARM + Thumb)
1789 323177 Unhandled instruction: SMLSLD (ARM + Thumb)
1790 323432 Calling pthread_cond_destroy() or pthread_mutex_destroy() [..]
1791 323437 Phase 2 support for IBM Power ISA 2.07
1792 323713 Support mmxext (integer sse) subset on i386 (athlon)
1793 323803 Transactional memory instructions are not supported for Power
1794 323893 SSE3 not available on amd cpus in valgrind
1795 323905 Probable false positive from Valgrind/drd on close()
1796 323912 valgrind.h header isn't compatible for mingw64
1797 324047 Valgrind doesn't support [LDR,ST]{S}[B,H]T ARM instructions
1798 324149 helgrind: When pthread_cond_timedwait returns ETIMEDOUT [..]
1799 324181 mmap does not handle MAP_32BIT
1800 324227 memcheck false positive leak when a thread calls exit+block [..]
1801 324421 Support for fanotify API on ARM architecture
1802 324514 gdbserver monitor cmd output behaviour consistency [..]
1803 324518 ppc64: Emulation of dcbt instructions does not handle [..]
1804 324546 none/tests/ppc32 test_isa_2_07_part2 requests -m64
1805 324582 When access is made to freed memory, report both allocation [..]
1806 324594 Fix overflow computation for Power ISA 2.06 insns: mulldo/mulldo.
1807 324765 ppc64: illegal instruction when executing none/tests/ppc64/jm-misc
1808 324816 Incorrect VEX implementation for xscvspdp/xvcvspdp for SNaN inputs
1809 324834 Unhandled instructions in Microsoft C run-time for x86_64
1810 324894 Phase 3 support for IBM Power ISA 2.07
1811 326091 drd: Avoid false race reports from optimized strlen() impls
1812 326113 valgrind libvex hwcaps error on AMD64
1813 n-i-bz Some wrong command line options could be ignored
1814 n-i-bz patch to allow fair-sched on android
1815 n-i-bz report error for vgdb snapshot requested before execution
1816 n-i-bz same as 303624 (fixed in 3.8.0), but for x86 android
1818 (3.9.0: 31 October 2013, vex r2796, valgrind r13708)
1822 Release 3.8.1 (19 September 2012)
1823 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1824 3.8.1 is a bug fix release. It fixes some assertion failures in 3.8.0
1825 that occur moderately frequently in real use cases, adds support for
1826 some missing instructions on ARM, and fixes a deadlock condition on
1827 MacOSX. If you package or deliver 3.8.0 for others to use, you might
1828 want to consider upgrading to 3.8.1 instead.
1830 The following bugs have been fixed or resolved. Note that "n-i-bz"
1831 stands for "not in bugzilla" -- that is, a bug that was reported to us
1832 but never got a bugzilla entry. We encourage you to file bugs in
1833 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1834 than mailing the developers (or mailing lists) directly -- bugs that
1835 are not entered into bugzilla tend to get forgotten about or ignored.
1837 To see details of a given bug, visit
1838 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1839 where XXXXXX is the bug number as listed below.
1842 289584 Unhandled instruction: 0xF 0x29 0xE5 (MOVAPS)
1843 295808 amd64->IR: 0xF3 0xF 0xBC 0xC0 (TZCNT)
1844 298281 wcslen causes false(?) uninitialised value warnings
1845 301281 valgrind hangs on OS X when the process calls system()
1846 304035 disInstr(arm): unhandled instruction 0xE1023053
1847 304867 implement MOVBE instruction in x86 mode
1848 304980 Assertion 'lo <= hi' failed in vgModuleLocal_find_rx_mapping
1849 305042 amd64: implement 0F 7F encoding of movq between two registers
1850 305199 ARM: implement QDADD and QDSUB
1851 305321 amd64->IR: 0xF 0xD 0xC (prefetchw)
1852 305513 killed by fatal signal: SIGSEGV
1853 305690 DRD reporting invalid semaphore when sem_trywait fails
1854 305926 Invalid alignment checks for some AVX instructions
1855 306297 disInstr(thumb): unhandled instruction 0xE883 0x000C
1856 306310 3.8.0 release tarball missing some files
1857 306612 RHEL 6 glibc-2.X default suppressions need /lib*/libc-*patterns
1858 306664 vex amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x46 0x66 0xF
1859 n-i-bz shmat of a segment > 4Gb does not work
1860 n-i-bz simulate_control_c script wrong USR1 signal number on mips
1861 n-i-bz vgdb ptrace calls wrong on mips [...]
1862 n-i-bz Fixes for more MPI false positives
1863 n-i-bz exp-sgcheck's memcpy causes programs to segfault
1864 n-i-bz OSX build w/ clang: asserts at startup
1865 n-i-bz Incorrect undef'dness prop for Iop_DPBtoBCD and Iop_BCDtoDPB
1866 n-i-bz fix a couple of union tag-vs-field mixups
1867 n-i-bz OSX: use __NR_poll_nocancel rather than __NR_poll
1869 The following bugs were fixed in 3.8.0 but not listed in this NEWS
1872 254088 Valgrind should know about UD2 instruction
1875 304754 NEWS blows TeX's little mind
1877 (3.8.1: 19 September 2012, vex r2537, valgrind r12996)
1881 Release 3.8.0 (10 August 2012)
1882 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1883 3.8.0 is a feature release with many improvements and the usual
1884 collection of bug fixes.
1886 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
1887 PPC64/Linux, S390X/Linux, MIPS/Linux, ARM/Android, X86/Android,
1888 X86/MacOSX 10.6/10.7 and AMD64/MacOSX 10.6/10.7. Support for recent
1889 distros and toolchain components (glibc 2.16, gcc 4.7) has been added.
1890 There is initial support for MacOSX 10.8, but it is not usable for
1891 serious work at present.
1893 * ================== PLATFORM CHANGES =================
1895 * Support for MIPS32 platforms running Linux. Valgrind has been
1896 tested on MIPS32 and MIPS32r2 platforms running different Debian
1897 Squeeze and MeeGo distributions. Both little-endian and big-endian
1898 cores are supported. The tools Memcheck, Massif and Lackey have
1899 been tested and are known to work. See README.mips for more details.
1901 * Preliminary support for Android running on x86.
1903 * Preliminary (as-yet largely unusable) support for MacOSX 10.8.
1905 * Support for Intel AVX instructions and for AES instructions. This
1906 support is available only for 64 bit code.
1908 * Support for POWER Decimal Floating Point instructions.
1910 * ==================== TOOL CHANGES ====================
1912 * Non-libc malloc implementations are now supported. This is useful
1913 for tools that replace malloc (Memcheck, Massif, DRD, Helgrind).
1914 Using the new option --soname-synonyms, such tools can be informed
1915 that the malloc implementation is either linked statically into the
1916 executable, or is present in some other shared library different
1917 from libc.so. This makes it possible to process statically linked
1918 programs, and programs using other malloc libraries, for example
1919 TCMalloc or JEMalloc.
1921 * For tools that provide their own replacement for malloc et al, the
1922 option --redzone-size=<number> allows users to specify the size of
1923 the padding blocks (redzones) added before and after each client
1924 allocated block. Smaller redzones decrease the memory needed by
1925 Valgrind. Bigger redzones increase the chance to detect blocks
1926 overrun or underrun. Prior to this change, the redzone size was
1927 hardwired to 16 bytes in Memcheck.
1931 - The leak_check GDB server monitor command now can
1932 control the maximum nr of loss records to output.
1934 - Reduction of memory use for applications allocating
1935 many blocks and/or having many partially defined bytes.
1937 - Addition of GDB server monitor command 'block_list' that lists
1938 the addresses/sizes of the blocks of a leak search loss record.
1940 - Addition of GDB server monitor command 'who_points_at' that lists
1941 the locations pointing at a block.
1943 - If a redzone size > 0 is given, VALGRIND_MALLOCLIKE_BLOCK now will
1944 detect an invalid access of these redzones, by marking them
1945 noaccess. Similarly, if a redzone size is given for a memory
1946 pool, VALGRIND_MEMPOOL_ALLOC will mark the redzones no access.
1947 This still allows to find some bugs if the user has forgotten to
1948 mark the pool superblock noaccess.
1950 - Performance of memory leak check has been improved, especially in
1951 cases where there are many leaked blocks and/or many suppression
1952 rules used to suppress leak reports.
1954 - Reduced noise (false positive) level on MacOSX 10.6/10.7, due to
1955 more precise analysis, which is important for LLVM/Clang
1956 generated code. This is at the cost of somewhat reduced
1957 performance. Note there is no change to analysis precision or
1958 costs on Linux targets.
1962 - Added even more facilities that can help finding the cause of a data
1963 race, namely the command-line option --ptrace-addr and the macro
1964 DRD_STOP_TRACING_VAR(x). More information can be found in the manual.
1966 - Fixed a subtle bug that could cause false positive data race reports.
1968 * ==================== OTHER CHANGES ====================
1970 * The C++ demangler has been updated so as to work well with C++
1971 compiled by up to at least g++ 4.6.
1973 * Tool developers can make replacement/wrapping more flexible thanks
1974 to the new option --soname-synonyms. This was reported above, but
1975 in fact is very general and applies to all function
1976 replacement/wrapping, not just to malloc-family functions.
1978 * Round-robin scheduling of threads can be selected, using the new
1979 option --fair-sched= yes. Prior to this change, the pipe-based
1980 thread serialisation mechanism (which is still the default) could
1981 give very unfair scheduling. --fair-sched=yes improves
1982 responsiveness of interactive multithreaded applications, and
1983 improves repeatability of results from the thread checkers Helgrind
1986 * For tool developers: support to run Valgrind on Valgrind has been
1987 improved. We can now routinely Valgrind on Helgrind or Memcheck.
1989 * gdbserver now shows the float shadow registers as integer
1990 rather than float values, as the shadow values are mostly
1991 used as bit patterns.
1993 * Increased limit for the --num-callers command line flag to 500.
1995 * Performance improvements for error matching when there are many
1996 suppression records in use.
1998 * Improved support for DWARF4 debugging information (bug 284184).
2000 * Initial support for DWZ compressed Dwarf debug info.
2002 * Improved control over the IR optimiser's handling of the tradeoff
2003 between performance and precision of exceptions. Specifically,
2004 --vex-iropt-precise-memory-exns has been removed and replaced by
2005 --vex-iropt-register-updates, with extended functionality. This
2006 allows the Valgrind gdbserver to always show up to date register
2009 * Modest performance gains through the use of translation chaining for
2012 * ==================== FIXED BUGS ====================
2014 The following bugs have been fixed or resolved. Note that "n-i-bz"
2015 stands for "not in bugzilla" -- that is, a bug that was reported to us
2016 but never got a bugzilla entry. We encourage you to file bugs in
2017 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2018 than mailing the developers (or mailing lists) directly -- bugs that
2019 are not entered into bugzilla tend to get forgotten about or ignored.
2021 To see details of a given bug, visit
2022 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2023 where XXXXXX is the bug number as listed below.
2025 197914 Building valgrind from svn now requires automake-1.10
2026 203877 increase to 16Mb maximum allowed alignment for memalign et al
2027 219156 Handle statically linked malloc or other malloc lib (e.g. tcmalloc)
2028 247386 make perf does not run all performance tests
2029 270006 Valgrind scheduler unfair
2030 270777 Adding MIPS/Linux port to Valgrind
2031 270796 s390x: Removed broken support for the TS insn
2032 271438 Fix configure for proper SSE4.2 detection
2033 273114 s390x: Support TR, TRE, TROO, TROT, TRTO, and TRTT instructions
2034 273475 Add support for AVX instructions
2035 274078 improved configure logic for mpicc
2036 276993 fix mremap 'no thrash checks'
2037 278313 Fedora 15/x64: err read debug info with --read-var-info=yes flag
2038 281482 memcheck incorrect byte allocation count in realloc() for silly argument
2039 282230 group allocator for small fixed size, use it for MC_Chunk/SEc vbit
2040 283413 Fix wrong sanity check
2041 283671 Robustize alignment computation in LibVEX_Alloc
2042 283961 Adding support for some HCI IOCTLs
2043 284124 parse_type_DIE: confused by: DWARF 4
2044 284864 == 273475 (Add support for AVX instructions)
2045 285219 Too-restrictive constraints for Thumb2 "SP plus/minus register"
2046 285662 (MacOSX): Memcheck needs to replace memcpy/memmove
2047 285725 == 273475 (Add support for AVX instructions)
2048 286261 add wrapper for linux I2C_RDWR ioctl
2049 286270 vgpreload is not friendly to 64->32 bit execs, gives ld.so warnings
2050 286374 Running cachegrind with --branch-sim=yes on 64-bit PowerPC program fails
2051 286384 configure fails "checking for a supported version of gcc"
2052 286497 == 273475 (Add support for AVX instructions)
2053 286596 == 273475 (Add support for AVX instructions)
2054 286917 disInstr(arm): unhandled instruction: QADD (also QSUB)
2055 287175 ARM: scalar VFP fixed-point VCVT instructions not handled
2056 287260 Incorrect conditional jump or move depends on uninitialised value(s)
2057 287301 vex amd64->IR: 0x66 0xF 0x38 0x41 0xC0 0xB8 0x0 0x0 (PHMINPOSUW)
2058 287307 == 273475 (Add support for AVX instructions)
2059 287858 VG_(strerror): unknown error
2060 288298 (MacOSX) unhandled syscall shm_unlink
2061 288995 == 273475 (Add support for AVX instructions)
2062 289470 Loading of large Mach-O thin binaries fails.
2063 289656 == 273475 (Add support for AVX instructions)
2064 289699 vgdb connection in relay mode erroneously closed due to buffer overrun
2065 289823 == 293754 (PCMPxSTRx not implemented for 16-bit characters)
2066 289839 s390x: Provide support for unicode conversion instructions
2067 289939 monitor cmd 'leak_check' with details about leaked or reachable blocks
2068 290006 memcheck doesn't mark %xmm as initialized after "pcmpeqw %xmm %xmm"
2069 290655 Add support for AESKEYGENASSIST instruction
2070 290719 valgrind-3.7.0 fails with automake-1.11.2 due to"pkglibdir" usage
2071 290974 vgdb must align pages to VKI_SHMLBA (16KB) on ARM
2072 291253 ES register not initialised in valgrind simulation
2073 291568 Fix 3DNOW-related crashes with baseline x86_64 CPU (w patch)
2074 291865 s390x: Support the "Compare Double and Swap" family of instructions
2075 292300 == 273475 (Add support for AVX instructions)
2076 292430 unrecognized instruction in __intel_get_new_mem_ops_cpuid
2077 292493 == 273475 (Add support for AVX instructions)
2078 292626 Missing fcntl F_SETOWN_EX and F_GETOWN_EX support
2079 292627 Missing support for some SCSI ioctls
2080 292628 none/tests/x86/bug125959-x86.c triggers undefined behavior
2081 292841 == 273475 (Add support for AVX instructions)
2082 292993 implement the getcpu syscall on amd64-linux
2083 292995 Implement the “cross memory attach” syscalls introduced in Linux 3.2
2084 293088 Add some VEX sanity checks for ppc64 unhandled instructions
2085 293751 == 290655 (Add support for AESKEYGENASSIST instruction)
2086 293754 PCMPxSTRx not implemented for 16-bit characters
2087 293755 == 293754 (No tests for PCMPxSTRx on 16-bit characters)
2088 293808 CLFLUSH not supported by latest VEX for amd64
2089 294047 valgrind does not correctly emulate prlimit64(..., RLIMIT_NOFILE, ...)
2090 294048 MPSADBW instruction not implemented
2091 294055 regtest none/tests/shell fails when locale is not set to C
2092 294185 INT 0x44 (and others) not supported on x86 guest, but used by Jikes RVM
2093 294190 --vgdb-error=xxx can be out of sync with errors shown to the user
2094 294191 amd64: fnsave/frstor and 0x66 size prefixes on FP instructions
2095 294260 disInstr_AMD64: disInstr miscalculated next %rip
2096 294523 --partial-loads-ok=yes causes false negatives
2097 294617 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A
2098 294736 vex amd64->IR: 0x48 0xF 0xD7 0xD6 0x48 0x83
2099 294812 patch allowing to run (on x86 at least) helgrind/drd on tool.
2100 295089 can not annotate source for both helgrind and drd
2101 295221 POWER Processor decimal floating point instruction support missing
2102 295427 building for i386 with clang on darwin11 requires "-new_linker linker"
2103 295428 coregrind/m_main.c has incorrect x86 assembly for darwin
2104 295590 Helgrind: Assertion 'cvi->nWaiters > 0' failed
2105 295617 ARM - Add some missing syscalls
2106 295799 Missing \n with get_vbits in gdbserver when line is % 80 [...]
2107 296229 Linux user input device ioctls missing wrappers
2108 296318 ELF Debug info improvements (more than one rx/rw mapping)
2109 296422 Add translation chaining support
2110 296457 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A (dup of AES)
2111 296792 valgrind 3.7.0: add SIOCSHWTSTAMP (0x89B0) ioctl wrapper
2112 296983 Fix build issues on x86_64/ppc64 without 32-bit toolchains
2113 297078 gdbserver signal handling problems [..]
2114 297147 drd false positives on newly allocated memory
2115 297329 disallow decoding of IBM Power DFP insns on some machines
2116 297497 POWER Processor decimal floating point instruction support missing
2117 297701 Another alias for strncasecmp_l in libc-2.13.so
2118 297911 'invalid write' not reported when using APIs for custom mem allocators.
2119 297976 s390x: revisit EX implementation
2120 297991 Valgrind interferes with mmap()+ftell()
2121 297992 Support systems missing WIFCONTINUED (e.g. pre-2.6.10 Linux)
2122 297993 Fix compilation of valgrind with gcc -g3.
2123 298080 POWER Processor DFP support missing, part 3
2124 298227 == 273475 (Add support for AVX instructions)
2125 298335 == 273475 (Add support for AVX instructions)
2126 298354 Unhandled ARM Thumb instruction 0xEB0D 0x0585 (streq)
2127 298394 s390x: Don't bail out on an unknown machine model. [..]
2128 298421 accept4() syscall (366) support is missing for ARM
2129 298718 vex amd64->IR: 0xF 0xB1 0xCB 0x9C 0x8F 0x45
2130 298732 valgrind installation problem in ubuntu with kernel version 3.x
2131 298862 POWER Processor DFP instruction support missing, part 4
2132 298864 DWARF reader mis-parses DW_FORM_ref_addr
2133 298943 massif asserts with --pages-as-heap=yes when brk is changing [..]
2134 299053 Support DWARF4 DW_AT_high_pc constant form
2135 299104 == 273475 (Add support for AVX instructions)
2136 299316 Helgrind: hg_main.c:628 (map_threads_lookup): Assertion 'thr' failed.
2137 299629 dup3() syscall (358) support is missing for ARM
2138 299694 POWER Processor DFP instruction support missing, part 5
2139 299756 Ignore --free-fill for MEMPOOL_FREE and FREELIKE client requests
2140 299803 == 273475 (Add support for AVX instructions)
2141 299804 == 273475 (Add support for AVX instructions)
2142 299805 == 273475 (Add support for AVX instructions)
2143 300140 ARM - Missing (T1) SMMUL
2144 300195 == 296318 (ELF Debug info improvements (more than one rx/rw mapping))
2145 300389 Assertion `are_valid_hwcaps(VexArchAMD64, [..])' failed.
2146 300414 FCOM and FCOMP unimplemented for amd64 guest
2147 301204 infinite loop in canonicaliseSymtab with ifunc symbol
2148 301229 == 203877 (increase to 16Mb maximum allowed alignment for memalign etc)
2149 301265 add x86 support to Android build
2150 301984 configure script doesn't detect certain versions of clang
2151 302205 Fix compiler warnings for POWER VEX code and POWER test cases
2152 302287 Unhandled movbe instruction on Atom processors
2153 302370 PPC: fnmadd, fnmsub, fnmadds, fnmsubs insns always negate the result
2154 302536 Fix for the POWER Valgrind regression test: memcheck-ISA2.0.
2155 302578 Unrecognized isntruction 0xc5 0x32 0xc2 0xca 0x09 vcmpngess
2156 302656 == 273475 (Add support for AVX instructions)
2157 302709 valgrind for ARM needs extra tls support for android emulator [..]
2158 302827 add wrapper for CDROM_GET_CAPABILITY
2159 302901 Valgrind crashes with dwz optimized debuginfo
2160 302918 Enable testing of the vmaddfp and vnsubfp instructions in the testsuite
2161 303116 Add support for the POWER instruction popcntb
2162 303127 Power test suite fixes for frsqrte, vrefp, and vrsqrtefp instructions.
2163 303250 Assertion `instrs_in->arr_used <= 10000' failed w/ OpenSSL code
2164 303466 == 273475 (Add support for AVX instructions)
2165 303624 segmentation fault on Android 4.1 (e.g. on Galaxy Nexus OMAP)
2166 303963 strstr() function produces wrong results under valgrind callgrind
2167 304054 CALL_FN_xx macros need to enforce stack alignment
2168 304561 tee system call not supported
2169 715750 (MacOSX): Incorrect invalid-address errors near 0xFFFFxxxx (mozbug#)
2170 n-i-bz Add missing gdbserver xml files for shadow registers for ppc32
2171 n-i-bz Bypass gcc4.4/4.5 code gen bugs causing out of memory or asserts
2172 n-i-bz Fix assert in gdbserver for watchpoints watching the same address
2173 n-i-bz Fix false positive in sys_clone on amd64 when optional args [..]
2174 n-i-bz s390x: Shadow registers can now be examined using vgdb
2176 (3.8.0-TEST3: 9 August 2012, vex r2465, valgrind r12865)
2177 (3.8.0: 10 August 2012, vex r2465, valgrind r12866)
2181 Release 3.7.0 (5 November 2011)
2182 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2183 3.7.0 is a feature release with many significant improvements and the
2184 usual collection of bug fixes.
2186 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2187 PPC64/Linux, S390X/Linux, ARM/Android, X86/Darwin and AMD64/Darwin.
2188 Support for recent distros and toolchain components (glibc 2.14, gcc
2189 4.6, MacOSX 10.7) has been added.
2191 * ================== PLATFORM CHANGES =================
2193 * Support for IBM z/Architecture (s390x) running Linux. Valgrind can
2194 analyse 64-bit programs running on z/Architecture. Most user space
2195 instructions up to and including z10 are supported. Valgrind has
2196 been tested extensively on z9, z10, and z196 machines running SLES
2197 10/11, RedHat 5/6m, and Fedora. The Memcheck and Massif tools are
2198 known to work well. Callgrind, Helgrind, and DRD work reasonably
2199 well on z9 and later models. See README.s390 for more details.
2201 * Preliminary support for MacOSX 10.7 and XCode 4. Both 32- and
2202 64-bit processes are supported. Some complex threaded applications
2203 (Firefox) are observed to hang when run as 32 bit applications,
2204 whereas 64-bit versions run OK. The cause is unknown. Memcheck
2205 will likely report some false errors. In general, expect some rough
2206 spots. This release also supports MacOSX 10.6, but drops support
2209 * Preliminary support for Android (on ARM). Valgrind can now run
2210 large applications (eg, Firefox) on (eg) a Samsung Nexus S. See
2211 README.android for more details, plus instructions on how to get
2214 * Support for the IBM Power ISA 2.06 (Power7 instructions)
2216 * General correctness and performance improvements for ARM/Linux, and,
2217 by extension, ARM/Android.
2219 * Further solidification of support for SSE 4.2 in 64-bit mode. AVX
2220 instruction set support is under development but is not available in
2223 * Support for AIX5 has been removed.
2225 * ==================== TOOL CHANGES ====================
2227 * Memcheck: some incremental changes:
2229 - reduction of memory use in some circumstances
2231 - improved handling of freed memory, which in some circumstances
2232 can cause detection of use-after-free that would previously have
2235 - fix of a longstanding bug that could cause false negatives (missed
2236 errors) in programs doing vector saturated narrowing instructions.
2238 * Helgrind: performance improvements and major memory use reductions,
2239 particularly for large, long running applications which perform many
2240 synchronisation (lock, unlock, etc) events. Plus many smaller
2243 - display of locksets for both threads involved in a race
2245 - general improvements in formatting/clarity of error messages
2247 - addition of facilities and documentation regarding annotation
2248 of thread safe reference counted C++ classes
2250 - new flag --check-stack-refs=no|yes [yes], to disable race checking
2251 on thread stacks (a performance hack)
2253 - new flag --free-is-write=no|yes [no], to enable detection of races
2254 where one thread accesses heap memory but another one frees it,
2255 without any coordinating synchronisation event
2257 * DRD: enabled XML output; added support for delayed thread deletion
2258 in order to detect races that occur close to the end of a thread
2259 (--join-list-vol); fixed a memory leak triggered by repeated client
2260 memory allocatation and deallocation; improved Darwin support.
2262 * exp-ptrcheck: this tool has been renamed to exp-sgcheck
2264 * exp-sgcheck: this tool has been reduced in scope so as to improve
2265 performance and remove checking that Memcheck does better.
2266 Specifically, the ability to check for overruns for stack and global
2267 arrays is unchanged, but the ability to check for overruns of heap
2268 blocks has been removed. The tool has accordingly been renamed to
2269 exp-sgcheck ("Stack and Global Array Checking").
2271 * ==================== OTHER CHANGES ====================
2273 * GDB server: Valgrind now has an embedded GDB server. That means it
2274 is possible to control a Valgrind run from GDB, doing all the usual
2275 things that GDB can do (single stepping, breakpoints, examining
2276 data, etc). Tool-specific functionality is also available. For
2277 example, it is possible to query the definedness state of variables
2278 or memory from within GDB when running Memcheck; arbitrarily large
2279 memory watchpoints are supported, etc. To use the GDB server, start
2280 Valgrind with the flag --vgdb-error=0 and follow the on-screen
2283 * Improved support for unfriendly self-modifying code: a new option
2284 --smc-check=all-non-file is available. This adds the relevant
2285 consistency checks only to code that originates in non-file-backed
2286 mappings. In effect this confines the consistency checking only to
2287 code that is or might be JIT generated, and avoids checks on code
2288 that must have been compiled ahead of time. This significantly
2289 improves performance on applications that generate code at run time.
2291 * It is now possible to build a working Valgrind using Clang-2.9 on
2294 * new client requests VALGRIND_{DISABLE,ENABLE}_ERROR_REPORTING.
2295 These enable and disable error reporting on a per-thread, and
2296 nestable, basis. This is useful for hiding errors in particularly
2297 troublesome pieces of code. The MPI wrapper library (libmpiwrap.c)
2298 now uses this facility.
2300 * Added the --mod-funcname option to cg_diff.
2302 * ==================== FIXED BUGS ====================
2304 The following bugs have been fixed or resolved. Note that "n-i-bz"
2305 stands for "not in bugzilla" -- that is, a bug that was reported to us
2306 but never got a bugzilla entry. We encourage you to file bugs in
2307 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
2308 mailing the developers (or mailing lists) directly -- bugs that are
2309 not entered into bugzilla tend to get forgotten about or ignored.
2311 To see details of a given bug, visit
2312 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2313 where XXXXXX is the bug number as listed below.
2315 79311 malloc silly arg warning does not give stack trace
2316 210935 port valgrind.h (not valgrind) to win32 to support client requests
2317 214223 valgrind SIGSEGV on startup gcc 4.4.1 ppc32 (G4) Ubuntu 9.10
2318 243404 Port to zSeries
2319 243935 Helgrind: incorrect handling of ANNOTATE_HAPPENS_BEFORE()/AFTER()
2320 247223 non-x86: Suppress warning: 'regparm' attribute directive ignored
2321 250101 huge "free" memory usage due to m_mallocfree.c fragmentation
2322 253206 Some fixes for the faultstatus testcase
2323 255223 capget testcase fails when running as root
2324 256703 xlc_dbl_u32.c testcase broken
2325 256726 Helgrind tests have broken inline asm
2326 259977 == 214223 (Valgrind segfaults doing __builtin_longjmp)
2327 264800 testcase compile failure on zseries
2328 265762 make public VEX headers compilable by G++ 3.x
2329 265771 assertion in jumps.c (r11523) fails with glibc-2.3
2330 266753 configure script does not give the user the option to not use QtCore
2331 266931 gen_insn_test.pl is broken
2332 266961 ld-linux.so.2 i?86-linux strlen issues
2333 266990 setns instruction causes false positive
2334 267020 Make directory for temporary files configurable at run-time.
2335 267342 == 267997 (segmentation fault on Mac OS 10.6)
2336 267383 Assertion 'vgPlain_strlen(dir) + vgPlain_strlen(file) + 1 < 256' failed
2337 267413 Assertion 'DRD_(g_threadinfo)[tid].synchr_nesting >= 1' failed.
2338 267488 regtest: darwin support for 64-bit build
2339 267552 SIGSEGV (misaligned_stack_error) with DRD, but not with other tools
2340 267630 Add support for IBM Power ISA 2.06 -- stage 1
2341 267769 == 267997 (Darwin: memcheck triggers segmentation fault)
2342 267819 Add client request for informing the core about reallocation
2343 267925 laog data structure quadratic for a single sequence of lock
2344 267968 drd: (vgDrd_thread_set_joinable): Assertion '0 <= (int)tid ..' failed
2345 267997 MacOSX: 64-bit V segfaults on launch when built with Xcode 4.0.1
2346 268513 missed optimizations in fold_Expr
2347 268619 s390x: fpr - gpr transfer facility
2348 268620 s390x: reconsider "long displacement" requirement
2349 268621 s390x: improve IR generation for XC
2350 268715 s390x: FLOGR is not universally available
2351 268792 == 267997 (valgrind seg faults on startup when compiled with Xcode 4)
2352 268930 s390x: MHY is not universally available
2353 269078 arm->IR: unhandled instruction SUB (SP minus immediate/register)
2354 269079 Support ptrace system call on ARM
2355 269144 missing "Bad option" error message
2356 269209 conditional load and store facility (z196)
2357 269354 Shift by zero on x86 can incorrectly clobber CC_NDEP
2358 269641 == 267997 (valgrind segfaults immediately (segmentation fault))
2359 269736 s390x: minor code generation tweaks
2360 269778 == 272986 (valgrind.h: swap roles of VALGRIND_DO_CLIENT_REQUEST() ..)
2361 269863 s390x: remove unused function parameters
2362 269864 s390x: tweak s390_emit_load_cc
2363 269884 == 250101 (overhead for huge blocks exhausts space too soon)
2364 270082 s390x: Make sure to point the PSW address to the next address on SIGILL
2365 270115 s390x: rewrite some testcases
2366 270309 == 267997 (valgrind crash on startup)
2367 270320 add support for Linux FIOQSIZE ioctl() call
2368 270326 segfault while trying to sanitize the environment passed to execle
2369 270794 IBM POWER7 support patch causes regression in none/tests
2370 270851 IBM POWER7 fcfidus instruction causes memcheck to fail
2371 270856 IBM POWER7 xsnmaddadp instruction causes memcheck to fail on 32bit app
2372 270925 hyper-optimized strspn() in /lib64/libc-2.13.so needs fix
2373 270959 s390x: invalid use of R0 as base register
2374 271042 VSX configure check fails when it should not
2375 271043 Valgrind build fails with assembler error on ppc64 with binutils 2.21
2376 271259 s390x: fix code confusion
2377 271337 == 267997 (Valgrind segfaults on MacOS X)
2378 271385 s390x: Implement Ist_MBE
2379 271501 s390x: misc cleanups
2380 271504 s390x: promote likely and unlikely
2381 271579 ppc: using wrong enum type
2382 271615 unhandled instruction "popcnt" (arch=amd10h)
2383 271730 Fix bug when checking ioctls: duplicate check
2384 271776 s390x: provide STFLE instruction support
2385 271779 s390x: provide clock instructions like STCK
2386 271799 Darwin: ioctls without an arg report a memory error
2387 271820 arm: fix type confusion
2388 271917 pthread_cond_timedwait failure leads to not-locked false positive
2389 272067 s390x: fix DISP20 macro
2390 272615 A typo in debug output in mc_leakcheck.c
2391 272661 callgrind_annotate chokes when run from paths containing regex chars
2392 272893 amd64->IR: 0x66 0xF 0x38 0x2B 0xC1 0x66 0xF 0x7F == (closed as dup)
2393 272955 Unhandled syscall error for pwrite64 on ppc64 arch
2394 272967 make documentation build-system more robust
2395 272986 Fix gcc-4.6 warnings with valgrind.h
2396 273318 amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38 (missing PCMPxSTRx case)
2397 273318 unhandled PCMPxSTRx case: vex amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38
2398 273431 valgrind segfaults in evalCfiExpr (debuginfo.c:2039)
2399 273465 Callgrind: jumps.c:164 (new_jcc): Assertion '(0 <= jmp) && ...'
2400 273536 Build error: multiple definition of `vgDrd_pthread_cond_initializer'
2401 273640 ppc64-linux: unhandled syscalls setresuid(164) and setresgid(169)
2402 273729 == 283000 (Illegal opcode for SSE2 "roundsd" instruction)
2403 273778 exp-ptrcheck: unhandled sysno == 259
2404 274089 exp-ptrcheck: unhandled sysno == 208
2405 274378 s390x: Various dispatcher tweaks
2406 274447 WARNING: unhandled syscall: 340
2407 274776 amd64->IR: 0x66 0xF 0x38 0x2B 0xC5 0x66
2408 274784 == 267997 (valgrind ls -l results in Segmentation Fault)
2409 274926 valgrind does not build against linux-3
2410 275148 configure FAIL with glibc-2.14
2411 275151 Fedora 15 / glibc-2.14 'make regtest' FAIL
2412 275168 Make Valgrind work for MacOSX 10.7 Lion
2413 275212 == 275284 (lots of false positives from __memcpy_ssse3_back et al)
2414 275278 valgrind does not build on Linux kernel 3.0.* due to silly
2415 275284 Valgrind memcpy/memmove redirection stopped working in glibc 2.14/x86_64
2416 275308 Fix implementation for ppc64 fres instruc
2417 275339 s390x: fix testcase compile warnings
2418 275517 s390x: Provide support for CKSM instruction
2419 275710 s390x: get rid of redundant address mode calculation
2420 275815 == 247894 (Valgrind doesn't know about Linux readahead(2) syscall)
2421 275852 == 250101 (valgrind uses all swap space and is killed)
2422 276784 Add support for IBM Power ISA 2.06 -- stage 3
2423 276987 gdbsrv: fix tests following recent commits
2424 277045 Valgrind crashes with unhandled DW_OP_ opcode 0x2a
2425 277199 The test_isa_2_06_part1.c in none/tests/ppc64 should be a symlink
2426 277471 Unhandled syscall: 340
2427 277610 valgrind crashes in VG_(lseek)(core_fd, phdrs[idx].p_offset, ...)
2428 277653 ARM: support Thumb2 PLD instruction
2429 277663 ARM: NEON float VMUL by scalar incorrect
2430 277689 ARM: tests for VSTn with register post-index are broken
2431 277694 ARM: BLX LR instruction broken in ARM mode
2432 277780 ARM: VMOV.F32 (immediate) instruction is broken
2433 278057 fuse filesystem syscall deadlocks
2434 278078 Unimplemented syscall 280 on ppc32
2435 278349 F_GETPIPE_SZ and F_SETPIPE_SZ Linux fcntl commands
2436 278454 VALGRIND_STACK_DEREGISTER has wrong output type
2437 278502 == 275284 (Valgrind confuses memcpy() and memmove())
2438 278892 gdbsrv: factorize gdb version handling, fix doc and typos
2439 279027 Support for MVCL and CLCL instruction
2440 279027 s390x: Provide support for CLCL and MVCL instructions
2441 279062 Remove a redundant check in the insn selector for ppc.
2442 279071 JDK creates PTEST with redundant REX.W prefix
2443 279212 gdbsrv: add monitor cmd v.info scheduler.
2444 279378 exp-ptrcheck: the 'impossible' happened on mkfifo call
2445 279698 memcheck discards valid-bits for packuswb
2446 279795 memcheck reports uninitialised values for mincore on amd64
2447 279994 Add support for IBM Power ISA 2.06 -- stage 3
2448 280083 mempolicy syscall check errors
2449 280290 vex amd64->IR: 0x66 0xF 0x38 0x28 0xC1 0x66 0xF 0x6F
2450 280710 s390x: config files for nightly builds
2451 280757 /tmp dir still used by valgrind even if TMPDIR is specified
2452 280965 Valgrind breaks fcntl locks when program does mmap
2453 281138 WARNING: unhandled syscall: 340
2454 281241 == 275168 (valgrind useless on Macos 10.7.1 Lion)
2455 281304 == 275168 (Darwin: dyld "cannot load inserted library")
2456 281305 == 275168 (unhandled syscall: unix:357 on Darwin 11.1)
2457 281468 s390x: handle do_clone and gcc clones in call traces
2458 281488 ARM: VFP register corruption
2459 281828 == 275284 (false memmove warning: "Source and destination overlap")
2460 281883 s390x: Fix system call wrapper for "clone".
2461 282105 generalise 'reclaimSuperBlock' to also reclaim splittable superblock
2462 282112 Unhandled instruction bytes: 0xDE 0xD9 0x9B 0xDF (fcompp)
2463 282238 SLES10: make check fails
2464 282979 strcasestr needs replacement with recent(>=2.12) glibc
2465 283000 vex amd64->IR: 0x66 0xF 0x3A 0xA 0xC0 0x9 0xF3 0xF
2466 283243 Regression in ppc64 memcheck tests
2467 283325 == 267997 (Darwin: V segfaults on startup when built with Xcode 4.0)
2468 283427 re-connect epoll_pwait syscall on ARM linux
2469 283600 gdbsrv: android: port vgdb.c
2470 283709 none/tests/faultstatus needs to account for page size
2471 284305 filter_gdb needs enhancement to work on ppc64
2472 284384 clang 3.1 -Wunused-value warnings in valgrind.h, memcheck.h
2473 284472 Thumb2 ROR.W encoding T2 not implemented
2474 284621 XML-escape process command line in XML output
2475 n-i-bz cachegrind/callgrind: handle CPUID information for Core iX Intel CPUs
2476 that have non-power-of-2 sizes (also AMDs)
2477 n-i-bz don't be spooked by libraries mashed by elfhack
2478 n-i-bz don't be spooked by libxul.so linked with gold
2479 n-i-bz improved checking for VALGRIND_CHECK_MEM_IS_DEFINED
2481 (3.7.0-TEST1: 27 October 2011, vex r2228, valgrind r12245)
2482 (3.7.0.RC1: 1 November 2011, vex r2231, valgrind r12257)
2483 (3.7.0: 5 November 2011, vex r2231, valgrind r12258)
2487 Release 3.6.1 (16 February 2011)
2488 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2489 3.6.1 is a bug fix release. It adds support for some SSE4
2490 instructions that were omitted in 3.6.0 due to lack of time. Initial
2491 support for glibc-2.13 has been added. A number of bugs causing
2492 crashing or assertion failures have been fixed.
2494 The following bugs have been fixed or resolved. Note that "n-i-bz"
2495 stands for "not in bugzilla" -- that is, a bug that was reported to us
2496 but never got a bugzilla entry. We encourage you to file bugs in
2497 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
2498 mailing the developers (or mailing lists) directly -- bugs that are
2499 not entered into bugzilla tend to get forgotten about or ignored.
2501 To see details of a given bug, visit
2502 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2503 where XXXXXX is the bug number as listed below.
2505 188572 Valgrind on Mac should suppress setenv() mem leak
2506 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 (proper FX{SAVE,RSTOR} support)
2507 210481 vex amd64->IR: Assertion `sz == 2 || sz == 4' failed (REX.W POPQ)
2508 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
2509 250038 ppc64: Altivec LVSR and LVSL instructions fail their regtest
2510 254420 memory pool tracking broken
2511 254957 Test code failing to compile due to changes in memcheck.h
2512 255009 helgrind/drd: crash on chmod with invalid parameter
2513 255130 readdwarf3.c parse_type_DIE confused by GNAT Ada types
2514 255355 helgrind/drd: crash on threaded programs doing fork
2516 255418 (SSE4.x) rint call compiled with ICC
2517 255822 --gen-suppressions can create invalid files: "too many callers [...]"
2518 255888 closing valgrindoutput tag outputted to log-stream on error
2519 255963 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x9 0xDB 0x0 (ROUNDPD)
2520 255966 Slowness when using mempool annotations
2521 256387 vex x86->IR: 0xD4 0xA 0x2 0x7 (AAD and AAM)
2522 256600 super-optimized strcasecmp() false positive
2523 256669 vex amd64->IR: Unhandled LOOPNEL insn on amd64
2524 256968 (SSE4.x) vex amd64->IR: 0x66 0xF 0x38 0x10 0xD3 0x66 (BLENDVPx)
2525 257011 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0xE 0xFD 0xA0 (PBLENDW)
2526 257063 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x8 0xC0 0x0 (ROUNDPS)
2527 257276 Missing case in memcheck --track-origins=yes
2528 258870 (SSE4.x) Add support for EXTRACTPS SSE 4.1 instruction
2529 261966 (SSE4.x) support for CRC32B and CRC32Q is lacking (also CRC32{W,L})
2530 262985 VEX regression in valgrind 3.6.0 in handling PowerPC VMX
2531 262995 (SSE4.x) crash when trying to valgrind gcc-snapshot (PCMPxSTRx $0)
2532 263099 callgrind_annotate counts Ir improperly [...]
2533 263877 undefined coprocessor instruction on ARMv7
2534 265964 configure FAIL with glibc-2.13
2535 n-i-bz Fix compile error w/ icc-12.x in guest_arm_toIR.c
2536 n-i-bz Docs: fix bogus descriptions for VALGRIND_CREATE_BLOCK et al
2537 n-i-bz Massif: don't assert on shmat() with --pages-as-heap=yes
2538 n-i-bz Bug fixes and major speedups for the exp-DHAT space profiler
2539 n-i-bz DRD: disable --free-is-write due to implementation difficulties
2541 (3.6.1: 16 February 2011, vex r2103, valgrind r11561).
2545 Release 3.6.0 (21 October 2010)
2546 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2547 3.6.0 is a feature release with many significant improvements and the
2548 usual collection of bug fixes.
2550 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2551 PPC64/Linux, X86/Darwin and AMD64/Darwin. Support for recent distros
2552 and toolchain components (glibc 2.12, gcc 4.5, OSX 10.6) has been added.
2554 -------------------------
2556 Here are some highlights. Details are shown further down:
2558 * Support for ARM/Linux.
2560 * Support for recent Linux distros: Ubuntu 10.10 and Fedora 14.
2562 * Support for Mac OS X 10.6, both 32- and 64-bit executables.
2564 * Support for the SSE4.2 instruction set.
2566 * Enhancements to the Callgrind profiler, including the ability to
2567 handle CPUs with three levels of cache.
2569 * A new experimental heap profiler, DHAT.
2571 * A huge number of bug fixes and small enhancements.
2573 -------------------------
2575 Here are details of the above changes, together with descriptions of
2576 many other changes, and a list of fixed bugs.
2578 * ================== PLATFORM CHANGES =================
2580 * Support for ARM/Linux. Valgrind now runs on ARMv7 capable CPUs
2581 running Linux. It is known to work on Ubuntu 10.04, Ubuntu 10.10,
2582 and Maemo 5, so you can run Valgrind on your Nokia N900 if you want.
2584 This requires a CPU capable of running the ARMv7-A instruction set
2585 (Cortex A5, A8 and A9). Valgrind provides fairly complete coverage
2586 of the user space instruction set, including ARM and Thumb integer
2587 code, VFPv3, NEON and V6 media instructions. The Memcheck,
2588 Cachegrind and Massif tools work properly; other tools work to
2591 * Support for recent Linux distros (Ubuntu 10.10 and Fedora 14), along
2592 with support for recent releases of the underlying toolchain
2593 components, notably gcc-4.5 and glibc-2.12.
2595 * Support for Mac OS X 10.6, both 32- and 64-bit executables. 64-bit
2596 support also works much better on OS X 10.5, and is as solid as
2599 * Support for the SSE4.2 instruction set. SSE4.2 is supported in
2600 64-bit mode. In 32-bit mode, support is only available up to and
2601 including SSSE3. Some exceptions: SSE4.2 AES instructions are not
2602 supported in 64-bit mode, and 32-bit mode does in fact support the
2603 bare minimum SSE4 instructions to needed to run programs on Mac OS X
2604 10.6 on 32-bit targets.
2606 * Support for IBM POWER6 cpus has been improved. The Power ISA up to
2607 and including version 2.05 is supported.
2609 * ==================== TOOL CHANGES ====================
2611 * Cachegrind has a new processing script, cg_diff, which finds the
2612 difference between two profiles. It's very useful for evaluating
2613 the performance effects of a change in a program.
2615 Related to this change, the meaning of cg_annotate's (rarely-used)
2616 --threshold option has changed; this is unlikely to affect many
2617 people, if you do use it please see the user manual for details.
2619 * Callgrind now can do branch prediction simulation, similar to
2620 Cachegrind. In addition, it optionally can count the number of
2621 executed global bus events. Both can be used for a better
2622 approximation of a "Cycle Estimation" as derived event (you need to
2623 update the event formula in KCachegrind yourself).
2625 * Cachegrind and Callgrind now refer to the LL (last-level) cache
2626 rather than the L2 cache. This is to accommodate machines with
2627 three levels of caches -- if Cachegrind/Callgrind auto-detects the
2628 cache configuration of such a machine it will run the simulation as
2629 if the L2 cache isn't present. This means the results are less
2630 likely to match the true result for the machine, but
2631 Cachegrind/Callgrind's results are already only approximate, and
2632 should not be considered authoritative. The results are still
2633 useful for giving a general idea about a program's locality.
2635 * Massif has a new option, --pages-as-heap, which is disabled by
2636 default. When enabled, instead of tracking allocations at the level
2637 of heap blocks (as allocated with malloc/new/new[]), it instead
2638 tracks memory allocations at the level of memory pages (as mapped by
2639 mmap, brk, etc). Each mapped page is treated as its own block.
2640 Interpreting the page-level output is harder than the heap-level
2641 output, but this option is useful if you want to account for every
2642 byte of memory used by a program.
2644 * DRD has two new command-line options: --free-is-write and
2645 --trace-alloc. The former allows to detect reading from already freed
2646 memory, and the latter allows tracing of all memory allocations and
2649 * DRD has several new annotations. Custom barrier implementations can
2650 now be annotated, as well as benign races on static variables.
2652 * DRD's happens before / happens after annotations have been made more
2653 powerful, so that they can now also be used to annotate e.g. a smart
2654 pointer implementation.
2656 * Helgrind's annotation set has also been drastically improved, so as
2657 to provide to users a general set of annotations to describe locks,
2658 semaphores, barriers and condition variables. Annotations to
2659 describe thread-safe reference counted heap objects have also been
2662 * Memcheck has a new command-line option, --show-possibly-lost, which
2663 is enabled by default. When disabled, the leak detector will not
2664 show possibly-lost blocks.
2666 * A new experimental heap profiler, DHAT (Dynamic Heap Analysis Tool),
2667 has been added. DHAT keeps track of allocated heap blocks, and also
2668 inspects every memory reference to see which block (if any) is being
2669 accessed. This gives a lot of insight into block lifetimes,
2670 utilisation, turnover, liveness, and the location of hot and cold
2671 fields. You can use DHAT to do hot-field profiling.
2673 * ==================== OTHER CHANGES ====================
2675 * Improved support for unfriendly self-modifying code: the extra
2676 overhead incurred by --smc-check=all has been reduced by
2677 approximately a factor of 5 as compared with 3.5.0.
2679 * Ability to show directory names for source files in error messages.
2680 This is combined with a flexible mechanism for specifying which
2681 parts of the paths should be shown. This is enabled by the new flag
2684 * A new flag, --require-text-symbol, which will stop the run if a
2685 specified symbol is not found it a given shared object when it is
2686 loaded into the process. This makes advanced working with function
2687 intercepting and wrapping safer and more reliable.
2689 * Improved support for the Valkyrie GUI, version 2.0.0. GUI output
2690 and control of Valgrind is now available for the tools Memcheck and
2691 Helgrind. XML output from Valgrind is available for Memcheck,
2692 Helgrind and exp-Ptrcheck.
2694 * More reliable stack unwinding on amd64-linux, particularly in the
2695 presence of function wrappers, and with gcc-4.5 compiled code.
2697 * Modest scalability (performance improvements) for massive
2698 long-running applications, particularly for those with huge amounts
2701 * Support for analyzing programs running under Wine with has been
2702 improved. The header files <valgrind/valgrind.h>,
2703 <valgrind/memcheck.h> and <valgrind/drd.h> can now be used in
2704 Windows-programs compiled with MinGW or one of the Microsoft Visual
2707 * A rare but serious error in the 64-bit x86 CPU simulation was fixed.
2708 The 32-bit simulator was not affected. This did not occur often,
2709 but when it did would usually crash the program under test.
2712 * A large number of bugs were fixed. These are shown below.
2714 * A number of bugs were investigated, and were candidates for fixing,
2715 but are not fixed in 3.6.0, due to lack of developer time. They may
2716 get fixed in later releases. They are:
2718 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 0x24 0x49 (FXSAVE64)
2719 212419 false positive "lock order violated" (A+B vs A)
2720 213685 Undefined value propagates past dependency breaking instruction
2721 216837 Incorrect instrumentation of NSOperationQueue on Darwin
2722 237920 valgrind segfault on fork failure
2723 242137 support for code compiled by LLVM-2.8
2724 242423 Another unknown Intel cache config value
2725 243232 Inconsistent Lock Orderings report with trylock
2726 243483 ppc: callgrind triggers VEX assertion failure
2727 243935 Helgrind: implementation of ANNOTATE_HAPPENS_BEFORE() is wrong
2728 244677 Helgrind crash hg_main.c:616 (map_threads_lookup): Assertion
2730 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
2731 249435 Analyzing wine programs with callgrind triggers a crash
2732 250038 ppc64: Altivec lvsr and lvsl instructions fail their regtest
2733 250065 Handling large allocations
2734 250101 huge "free" memory usage due to m_mallocfree.c
2735 "superblocks fragmentation"
2736 251569 vex amd64->IR: 0xF 0x1 0xF9 0x8B 0x4C 0x24 (RDTSCP)
2737 252091 Callgrind on ARM does not detect function returns correctly
2738 252600 [PATCH] Allow lhs to be a pointer for shl/shr
2739 254420 memory pool tracking broken
2740 n-i-bz support for adding symbols for JIT generated code
2743 The following bugs have been fixed or resolved. Note that "n-i-bz"
2744 stands for "not in bugzilla" -- that is, a bug that was reported to us
2745 but never got a bugzilla entry. We encourage you to file bugs in
2746 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
2747 mailing the developers (or mailing lists) directly -- bugs that are
2748 not entered into bugzilla tend to get forgotten about or ignored.
2750 To see details of a given bug, visit
2751 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2752 where XXXXXX is the bug number as listed below.
2754 135264 dcbzl instruction missing
2756 153699 Valgrind should report unaligned reads with movdqa
2758 190429 Valgrind reports lost of errors in ld.so
2759 with x86_64 2.9.90 glibc
2760 197266 valgrind appears to choke on the xmms instruction
2762 197988 Crash when demangling very large symbol names
2763 202315 unhandled syscall: 332 (inotify_init1)
2764 203256 Add page-level profiling to Massif
2765 205093 dsymutil=yes needs quotes, locking (partial fix)
2766 205241 Snow Leopard 10.6 support (partial fix)
2767 206600 Leak checker fails to upgrade indirect blocks when their
2768 parent becomes reachable
2769 210935 port valgrind.h (not valgrind) to win32 so apps run under
2770 wine can make client requests
2771 211410 vex amd64->IR: 0x15 0xFF 0xFF 0x0 0x0 0x89
2772 within Linux ip-stack checksum functions
2773 212335 unhandled instruction bytes: 0xF3 0xF 0xBD 0xC0
2775 213685 Undefined value propagates past dependency breaking instruction
2777 215914 Valgrind inserts bogus empty environment variable
2779 219538 adjtimex syscall wrapper wrong in readonly adjtime mode
2780 222545 shmat fails under valgind on some arm targets
2781 222560 ARM NEON support
2784 232509 Docs build fails with formatting inside <title></title> elements
2786 235642 [PATCH] syswrap-linux.c: support evdev EVIOCG* ioctls
2787 236546 vex x86->IR: 0x66 0xF 0x3A 0xA
2788 237202 vex amd64->IR: 0xF3 0xF 0xB8 0xC0 0x49 0x3B
2789 237371 better support for VALGRIND_MALLOCLIKE_BLOCK
2790 237485 symlink (syscall 57) is not supported on Mac OS
2791 237723 sysno == 101 exp-ptrcheck: the 'impossible' happened:
2793 238208 is_just_below_ESP doesn't take into account red-zone
2794 238345 valgrind passes wrong $0 when executing a shell script
2795 238679 mq_timedreceive syscall doesn't flag the reception buffer
2797 238696 fcntl command F_DUPFD_CLOEXEC not supported
2798 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
2799 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
2800 238745 3.5.0 Make fails on PPC Altivec opcodes, though configure
2802 239992 vex amd64->IR: 0x48 0xF 0xC4 0xC1 0x0 0x48
2808 242606 unhandled syscall: setegid (in Ptrcheck)
2809 242814 Helgrind "Impossible has happened" during
2810 QApplication::initInstance();
2811 243064 Valgrind attempting to read debug information from iso
2812 243270 Make stack unwinding in Valgrind wrappers more reliable
2813 243884 exp-ptrcheck: the 'impossible happened: unhandled syscall
2814 sysno = 277 (mq_open)
2815 244009 exp-ptrcheck unknown syscalls in analyzing lighttpd
2816 244493 ARM VFP d16-d31 registers support
2817 244670 add support for audit_session_self syscall on Mac OS 10.6
2818 244921 The xml report of helgrind tool is not well format
2819 244923 In the xml report file, the <preamble> not escape the
2820 xml char, eg '<','&','>'
2821 245535 print full path names in plain text reports
2822 245925 x86-64 red zone handling problem
2823 246258 Valgrind not catching integer underruns + new [] s
2824 246311 reg/reg cmpxchg doesn't work on amd64
2825 246549 unhandled syscall unix:277 while testing 32-bit Darwin app
2826 246888 Improve Makefile.vex.am
2827 247510 [OS X 10.6] Memcheck reports unaddressable bytes passed
2828 to [f]chmod_extended
2829 247526 IBM POWER6 (ISA 2.05) support is incomplete
2830 247561 Some leak testcases fails due to reachable addresses in
2832 247875 sizeofIRType to handle Ity_I128
2833 247894 [PATCH] unhandled syscall sys_readahead
2834 247980 Doesn't honor CFLAGS passed to configure
2835 248373 darwin10.supp is empty in the trunk
2836 248822 Linux FIBMAP ioctl has int parameter instead of long
2837 248893 [PATCH] make readdwarf.c big endianess safe to enable
2838 unwinding on big endian systems
2839 249224 Syscall 336 not supported (SYS_proc_info)
2841 249775 Incorrect scheme for detecting NEON capabilities of host CPU
2842 249943 jni JVM init fails when using valgrind
2843 249991 Valgrind incorrectly declares AESKEYGENASSIST support
2845 249996 linux/arm: unhandled syscall: 181 (__NR_pwrite64)
2846 250799 frexp$fenv_access_off function generates SIGILL
2847 250998 vex x86->IR: unhandled instruction bytes: 0x66 0x66 0x66 0x2E
2848 251251 support pclmulqdq insn
2849 251362 valgrind: ARM: attach to debugger either fails or provokes
2851 251674 Unhandled syscall 294
2854 254257 Add support for debugfiles found by build-id
2855 254550 [PATCH] Implement DW_ATE_UTF (DWARF4)
2856 254646 Wrapped functions cause stack misalignment on OS X
2857 (and possibly Linux)
2858 254556 ARM: valgrinding anything fails with SIGSEGV for 0xFFFF0FA0
2860 (3.6.0: 21 October 2010, vex r2068, valgrind r11471).
2864 Release 3.5.0 (19 August 2009)
2865 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2866 3.5.0 is a feature release with many significant improvements and the
2867 usual collection of bug fixes. The main improvement is that Valgrind
2868 now works on Mac OS X.
2870 This release supports X86/Linux, AMD64/Linux, PPC32/Linux, PPC64/Linux
2871 and X86/Darwin. Support for recent distros and toolchain components
2872 (glibc 2.10, gcc 4.5) has been added.
2874 -------------------------
2876 Here is a short summary of the changes. Details are shown further
2879 * Support for Mac OS X (10.5.x).
2881 * Improvements and simplifications to Memcheck's leak checker.
2883 * Clarification and simplifications in various aspects of Valgrind's
2886 * XML output for Helgrind and Ptrcheck.
2888 * Performance and stability improvements for Helgrind and DRD.
2890 * Genuinely atomic support for x86/amd64/ppc atomic instructions.
2892 * A new experimental tool, BBV, useful for computer architecture
2895 * Improved Wine support, including ability to read Windows PDB
2898 -------------------------
2900 Here are details of the above changes, followed by descriptions of
2901 many other minor changes, and a list of fixed bugs.
2904 * Valgrind now runs on Mac OS X. (Note that Mac OS X is sometimes
2905 called "Darwin" because that is the name of the OS core, which is the
2906 level that Valgrind works at.)
2910 - It requires OS 10.5.x (Leopard). Porting to 10.4.x is not planned
2911 because it would require work and 10.4 is only becoming less common.
2913 - 32-bit programs on x86 and AMD64 (a.k.a x86-64) machines are supported
2914 fairly well. For 10.5.x, 32-bit programs are the default even on
2915 64-bit machines, so it handles most current programs.
2917 - 64-bit programs on x86 and AMD64 (a.k.a x86-64) machines are not
2918 officially supported, but simple programs at least will probably work.
2919 However, start-up is slow.
2921 - PowerPC machines are not supported.
2923 Things that don't work:
2925 - The Ptrcheck tool.
2927 - Objective-C garbage collection.
2931 - If you have Rogue Amoeba's "Instant Hijack" program installed,
2932 Valgrind will fail with a SIGTRAP at start-up. See
2933 https://bugs.kde.org/show_bug.cgi?id=193917 for details and a
2938 - You will likely find --dsymutil=yes a useful option, as error
2939 messages may be imprecise without it.
2941 - Mac OS X support is new and therefore will be less robust than the
2942 Linux support. Please report any bugs you find.
2944 - Threaded programs may run more slowly than on Linux.
2946 Many thanks to Greg Parker for developing this port over several years.
2949 * Memcheck's leak checker has been improved.
2951 - The results for --leak-check=summary now match the summary results
2952 for --leak-check=full. Previously they could differ because
2953 --leak-check=summary counted "indirectly lost" blocks and
2954 "suppressed" blocks as "definitely lost".
2956 - Blocks that are only reachable via at least one interior-pointer,
2957 but are directly pointed to by a start-pointer, were previously
2958 marked as "still reachable". They are now correctly marked as
2961 - The default value for the --leak-resolution option has been
2962 changed from "low" to "high". In general, this means that more
2963 leak reports will be produced, but each leak report will describe
2964 fewer leaked blocks.
2966 - With --leak-check=full, "definitely lost" and "possibly lost"
2967 leaks are now considered as proper errors, ie. they are counted
2968 for the "ERROR SUMMARY" and affect the behaviour of
2969 --error-exitcode. These leaks are not counted as errors if
2970 --leak-check=summary is specified, however.
2972 - Documentation for the leak checker has been improved.
2975 * Various aspects of Valgrind's text output have changed.
2977 - Valgrind's start-up message has changed. It is shorter but also
2978 includes the command being run, which makes it easier to use
2979 --trace-children=yes. An example:
2981 - Valgrind's shut-down messages have also changed. This is most
2982 noticeable with Memcheck, where the leak summary now occurs before
2983 the error summary. This change was necessary to allow leaks to be
2984 counted as proper errors (see the description of the leak checker
2985 changes above for more details). This was also necessary to fix a
2986 longstanding bug in which uses of suppressions against leaks were
2987 not "counted", leading to difficulties in maintaining suppression
2988 files (see https://bugs.kde.org/show_bug.cgi?id=186790).
2990 - Behavior of -v has changed. In previous versions, -v printed out
2991 a mixture of marginally-user-useful information, and tool/core
2992 statistics. The statistics printing has now been moved to its own
2993 flag, --stats=yes. This means -v is less verbose and more likely
2994 to convey useful end-user information.
2996 - The format of some (non-XML) stack trace entries has changed a
2997 little. Previously there were six possible forms:
2999 0x80483BF: really (a.c:20)
3000 0x80483BF: really (in /foo/a.out)
3002 0x80483BF: (within /foo/a.out)
3003 0x80483BF: ??? (a.c:20)
3006 The third and fourth of these forms have been made more consistent
3007 with the others. The six possible forms are now:
3009 0x80483BF: really (a.c:20)
3010 0x80483BF: really (in /foo/a.out)
3011 0x80483BF: really (in ???)
3012 0x80483BF: ??? (in /foo/a.out)
3013 0x80483BF: ??? (a.c:20)
3016 Stack traces produced when --xml=yes is specified are different
3020 * Helgrind and Ptrcheck now support XML output, so they can be used
3021 from GUI tools. Also, the XML output mechanism has been
3024 - The XML format has been overhauled and generalised, so it is more
3025 suitable for error reporting tools in general. The Memcheck
3026 specific aspects of it have been removed. The new format, which
3027 is an evolution of the old format, is described in
3028 docs/internals/xml-output-protocol4.txt.
3030 - Memcheck has been updated to use the new format.
3032 - Helgrind and Ptrcheck are now able to emit output in this format.
3034 - The XML output mechanism has been overhauled. XML is now output
3035 to its own file descriptor, which means that:
3037 * Valgrind can output text and XML independently.
3039 * The longstanding problem of XML output being corrupted by
3040 unexpected un-tagged text messages is solved.
3042 As before, the destination for text output is specified using
3043 --log-file=, --log-fd= or --log-socket=.
3045 As before, XML output for a tool is enabled using --xml=yes.
3047 Because there's a new XML output channel, the XML output
3048 destination is now specified by --xml-file=, --xml-fd= or
3051 Initial feedback has shown this causes some confusion. To
3052 clarify, the two envisaged usage scenarios are:
3054 (1) Normal text output. In this case, do not specify --xml=yes
3055 nor any of --xml-file=, --xml-fd= or --xml-socket=.
3057 (2) XML output. In this case, specify --xml=yes, and one of
3058 --xml-file=, --xml-fd= or --xml-socket= to select the XML
3059 destination, one of --log-file=, --log-fd= or --log-socket=
3060 to select the destination for any remaining text messages,
3061 and, importantly, -q.
3063 -q makes Valgrind completely silent on the text channel,
3064 except in the case of critical failures, such as Valgrind
3065 itself segfaulting, or failing to read debugging information.
3066 Hence, in this scenario, it suffices to check whether or not
3067 any output appeared on the text channel. If yes, then it is
3068 likely to be a critical error which should be brought to the
3069 attention of the user. If no (the text channel produced no
3070 output) then it can be assumed that the run was successful.
3072 This allows GUIs to make the critical distinction they need to
3073 make (did the run fail or not?) without having to search or
3074 filter the text output channel in any way.
3076 It is also recommended to use --child-silent-after-fork=yes in
3080 * Improvements and changes in Helgrind:
3082 - XML output, as described above
3084 - Checks for consistent association between pthread condition
3085 variables and their associated mutexes are now performed.
3087 - pthread_spinlock functions are supported.
3089 - Modest performance improvements.
3091 - Initial (skeletal) support for describing the behaviour of
3092 non-POSIX synchronisation objects through ThreadSanitizer
3093 compatible ANNOTATE_* macros.
3095 - More controllable tradeoffs between performance and the level of
3096 detail of "previous" accesses in a race. There are now three
3099 * --history-level=full. This is the default, and was also the
3100 default in 3.4.x. It shows both stacks involved in a race, but
3101 requires a lot of memory and can be very slow in programs that
3102 do many inter-thread synchronisation events.
3104 * --history-level=none. This only shows the later stack involved
3105 in a race. This can be much faster than --history-level=full,
3106 but makes it much more difficult to find the other access
3107 involved in the race.
3109 The new intermediate setting is
3111 * --history-level=approx
3113 For the earlier (other) access, two stacks are presented. The
3114 earlier access is guaranteed to be somewhere in between the two
3115 program points denoted by those stacks. This is not as useful
3116 as showing the exact stack for the previous access (as per
3117 --history-level=full), but it is better than nothing, and it's
3118 almost as fast as --history-level=none.
3121 * New features and improvements in DRD:
3123 - The error messages printed by DRD are now easier to interpret.
3124 Instead of using two different numbers to identify each thread
3125 (Valgrind thread ID and DRD thread ID), DRD does now identify
3126 threads via a single number (the DRD thread ID). Furthermore
3127 "first observed at" information is now printed for all error
3128 messages related to synchronization objects.
3130 - Added support for named semaphores (sem_open() and sem_close()).
3132 - Race conditions between pthread_barrier_wait() and
3133 pthread_barrier_destroy() calls are now reported.
3135 - Added support for custom allocators through the macros
3136 VALGRIND_MALLOCLIKE_BLOCK() VALGRIND_FREELIKE_BLOCK() (defined in
3137 in <valgrind/valgrind.h>). An alternative for these two macros is
3138 the new client request VG_USERREQ__DRD_CLEAN_MEMORY (defined in
3141 - Added support for annotating non-POSIX synchronization objects
3142 through several new ANNOTATE_*() macros.
3144 - OpenMP: added support for the OpenMP runtime (libgomp) included
3145 with gcc versions 4.3.0 and 4.4.0.
3149 - Added two new command-line options (--first-race-only and
3150 --segment-merging-interval).
3153 * Genuinely atomic support for x86/amd64/ppc atomic instructions
3155 Valgrind will now preserve (memory-access) atomicity of LOCK-
3156 prefixed x86/amd64 instructions, and any others implying a global
3157 bus lock. Ditto for PowerPC l{w,d}arx/st{w,d}cx. instructions.
3159 This means that Valgrinded processes will "play nicely" in
3160 situations where communication with other processes, or the kernel,
3161 is done through shared memory and coordinated with such atomic
3162 instructions. Prior to this change, such arrangements usually
3163 resulted in hangs, races or other synchronisation failures, because
3164 Valgrind did not honour atomicity of such instructions.
3167 * A new experimental tool, BBV, has been added. BBV generates basic
3168 block vectors for use with the SimPoint analysis tool, which allows
3169 a program's overall behaviour to be approximated by running only a
3170 fraction of it. This is useful for computer architecture
3171 researchers. You can run BBV by specifying --tool=exp-bbv (the
3172 "exp-" prefix is short for "experimental"). BBV was written by
3176 * Modestly improved support for running Windows applications under
3177 Wine. In particular, initial support for reading Windows .PDB debug
3178 information has been added.
3181 * A new Memcheck client request VALGRIND_COUNT_LEAK_BLOCKS has been
3182 added. It is similar to VALGRIND_COUNT_LEAKS but counts blocks
3186 * The Valgrind client requests VALGRIND_PRINTF and
3187 VALGRIND_PRINTF_BACKTRACE have been changed slightly. Previously,
3188 the string was always printed immediately on its own line. Now, the
3189 string will be added to a buffer but not printed until a newline is
3190 encountered, or other Valgrind output is printed (note that for
3191 VALGRIND_PRINTF_BACKTRACE, the back-trace itself is considered
3192 "other Valgrind output"). This allows you to use multiple
3193 VALGRIND_PRINTF calls to build up a single output line, and also to
3194 print multiple output lines with a single request (by embedding
3195 multiple newlines in the string).
3198 * The graphs drawn by Massif's ms_print program have changed slightly:
3200 - The half-height chars '.' and ',' are no longer drawn, because
3201 they are confusing. The --y option can be used if the default
3202 y-resolution is not high enough.
3204 - Horizontal lines are now drawn after the top of a snapshot if
3205 there is a gap until the next snapshot. This makes it clear that
3206 the memory usage has not dropped to zero between snapshots.
3209 * Something that happened in 3.4.0, but wasn't clearly announced: the
3210 option --read-var-info=yes can be used by some tools (Memcheck,
3211 Helgrind and DRD). When enabled, it causes Valgrind to read DWARF3
3212 variable type and location information. This makes those tools
3213 start up more slowly and increases memory consumption, but
3214 descriptions of data addresses in error messages become more
3218 * exp-Omega, an experimental instantaneous leak-detecting tool, was
3219 disabled in 3.4.0 due to a lack of interest and maintenance,
3220 although the source code was still in the distribution. The source
3221 code has now been removed from the distribution. For anyone
3222 interested, the removal occurred in SVN revision r10247.
3225 * Some changes have been made to the build system.
3227 - VEX/ is now integrated properly into the build system. This means
3228 that dependency tracking within VEX/ now works properly, "make
3229 install" will work without requiring "make" before it, and
3230 parallel builds (ie. 'make -j') now work (previously a
3231 .NOTPARALLEL directive was used to serialize builds, ie. 'make -j'
3232 was effectively ignored).
3234 - The --with-vex configure option has been removed. It was of
3235 little use and removing it simplified the build system.
3237 - The location of some install files has changed. This should not
3238 affect most users. Those who might be affected:
3240 * For people who use Valgrind with MPI programs, the installed
3241 libmpiwrap.so library has moved from
3242 $(INSTALL)/<platform>/libmpiwrap.so to
3243 $(INSTALL)/libmpiwrap-<platform>.so.
3245 * For people who distribute standalone Valgrind tools, the
3246 installed libraries such as $(INSTALL)/<platform>/libcoregrind.a
3247 have moved to $(INSTALL)/libcoregrind-<platform>.a.
3249 These changes simplify the build system.
3251 - Previously, all the distributed suppression (*.supp) files were
3252 installed. Now, only default.supp is installed. This should not
3253 affect users as the other installed suppression files were not
3254 read; the fact that they were installed was a mistake.
3257 * KNOWN LIMITATIONS:
3259 - Memcheck is unusable with the Intel compiler suite version 11.1,
3260 when it generates code for SSE2-and-above capable targets. This
3261 is because of icc's use of highly optimised inlined strlen
3262 implementations. It causes Memcheck to report huge numbers of
3263 false errors even in simple programs. Helgrind and DRD may also
3266 Versions 11.0 and earlier may be OK, but this has not been
3270 The following bugs have been fixed or resolved. Note that "n-i-bz"
3271 stands for "not in bugzilla" -- that is, a bug that was reported to us
3272 but never got a bugzilla entry. We encourage you to file bugs in
3273 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3274 mailing the developers (or mailing lists) directly -- bugs that are
3275 not entered into bugzilla tend to get forgotten about or ignored.
3277 To see details of a given bug, visit
3278 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3279 where XXXXXX is the bug number as listed below.
3281 84303 How about a LockCheck tool?
3282 91633 dereference of null ptr in vgPlain_st_basetype
3283 97452 Valgrind doesn't report any pthreads problems
3284 100628 leak-check gets assertion failure when using
3285 VALGRIND_MALLOCLIKE_BLOCK on malloc()ed memory
3286 108528 NPTL pthread cleanup handlers not called
3287 110126 Valgrind 2.4.1 configure.in tramples CFLAGS
3288 110128 mallinfo is not implemented...
3289 110770 VEX: Generated files not always updated when making valgrind
3290 111102 Memcheck: problems with large (memory footprint) applications
3291 115673 Vex's decoder should never assert
3292 117564 False positive: Syscall param clone(child_tidptr) contains
3293 uninitialised byte(s)
3294 119404 executing ssh from inside valgrind fails
3295 133679 Callgrind does not write path names to sources with dwarf debug
3297 135847 configure.in problem with non gnu compilers (and possible fix)
3298 136154 threads.c:273 (vgCallgrind_post_signal): Assertion
3299 '*(vgCallgrind_current_fn_stack.top) == 0' failed.
3300 136230 memcheck reports "possibly lost", should be "still reachable"
3301 137073 NULL arg to MALLOCLIKE_BLOCK causes crash
3302 137904 Valgrind reports a memory leak when using POSIX threads,
3304 139076 valgrind VT_GETSTATE error
3305 142228 complaint of elf_dynamic_do_rela in trivial usage
3306 145347 spurious warning with USBDEVFS_REAPURB
3307 148441 (wine) can't find memory leak in Wine, win32 binary
3309 148742 Leak-check fails assert on exit
3310 149878 add (proper) check for calloc integer overflow
3311 150606 Call graph is broken when using callgrind control
3312 152393 leak errors produce an exit code of 0. I need some way to
3313 cause leak errors to result in a nonzero exit code.
3314 157154 documentation (leak-resolution doc speaks about num-callers
3315 def=4) + what is a loss record
3316 159501 incorrect handling of ALSA ioctls
3317 162020 Valgrinding an empty/zero-byte file crashes valgrind
3318 162482 ppc: Valgrind crashes while reading stabs information
3319 162718 x86: avoid segment selector 0 in sys_set_thread_area()
3320 163253 (wine) canonicaliseSymtab forgot some fields in DiSym
3321 163560 VEX/test_main.c is missing from valgrind-3.3.1
3322 164353 malloc_usable_size() doesn't return a usable size
3323 165468 Inconsistent formatting in memcheck manual -- please fix
3324 169505 main.c:286 (endOfInstr):
3325 Assertion 'ii->cost_offset == *cost_offset' failed
3326 177206 Generate default.supp during compile instead of configure
3327 177209 Configure valt_load_address based on arch+os
3328 177305 eventfd / syscall 323 patch lost
3329 179731 Tests fail to build because of inlining of non-local asm labels
3330 181394 helgrind: libhb_core.c:3762 (msm_write): Assertion
3331 'ordxx == POrd_EQ || ordxx == POrd_LT' failed.
3332 181594 Bogus warning for empty text segment
3333 181707 dwarf doesn't require enumerations to have name
3334 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
3335 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
3336 Assertion '!already_present' failed.
3337 185359 exp-ptrcheck: unhandled syscall getresuid()
3338 185794 "WARNING: unhandled syscall: 285" (fallocate) on x86_64
3339 185816 Valgrind is unable to handle debug info for files with split
3340 debug info that are prelinked afterwards
3341 185980 [darwin] unhandled syscall: sem_open
3342 186238 bbToIR_AMD64: disInstr miscalculated next %rip
3343 186507 exp-ptrcheck unhandled syscalls prctl, etc.
3344 186790 Suppression pattern used for leaks are not reported
3345 186796 Symbols with length>200 in suppression files are ignored
3346 187048 drd: mutex PTHREAD_PROCESS_SHARED attribute missinterpretation
3347 187416 exp-ptrcheck: support for __NR_{setregid,setreuid,setresuid}
3348 188038 helgrind: hg_main.c:926: mk_SHVAL_fail: the 'impossible' happened
3349 188046 bashisms in the configure script
3350 188127 amd64->IR: unhandled instruction bytes: 0xF0 0xF 0xB0 0xA
3351 188161 memcheck: --track-origins=yes asserts "mc_machine.c:672
3352 (get_otrack_shadow_offset_wrk): the 'impossible' happened."
3353 188248 helgrind: pthread_cleanup_push, pthread_rwlock_unlock,
3354 assertion fail "!lock->heldBy"
3355 188427 Add support for epoll_create1 (with patch)
3356 188530 Support for SIOCGSTAMPNS
3357 188560 Include valgrind.spec in the tarball
3358 188572 Valgrind on Mac should suppress setenv() mem leak
3359 189054 Valgrind fails to build because of duplicate non-local asm labels
3360 189737 vex amd64->IR: unhandled instruction bytes: 0xAC
3361 189762 epoll_create syscall not handled (--tool=exp-ptrcheck)
3362 189763 drd assertion failure: s_threadinfo[tid].is_recording
3363 190219 unhandled syscall: 328 (x86-linux)
3364 190391 dup of 181394; see above
3365 190429 Valgrind reports lots of errors in ld.so with x86_64 2.9.90 glibc
3366 190820 No debug information on powerpc-linux
3367 191095 PATCH: Improve usbdevfs ioctl handling
3368 191182 memcheck: VALGRIND_LEAK_CHECK quadratic when big nr of chunks
3370 191189 --xml=yes should obey --gen-suppressions=all
3371 191192 syslog() needs a suppression on macosx
3372 191271 DARWIN: WARNING: unhandled syscall: 33554697 a.k.a.: 265
3373 191761 getrlimit on MacOSX
3374 191992 multiple --fn-skip only works sometimes; dependent on order
3375 192634 V. reports "aspacem sync_check_mapping_callback:
3376 segment mismatch" on Darwin
3377 192954 __extension__ missing on 2 client requests
3378 194429 Crash at start-up with glibc-2.10.1 and linux-2.6.29
3379 194474 "INSTALL" file has different build instructions than "README"
3380 194671 Unhandled syscall (sem_wait?) from mac valgrind
3381 195069 memcheck: reports leak (memory still reachable) for
3383 195169 drd: (vgDrd_barrier_post_wait):
3384 Assertion 'r->sg[p->post_iteration]' failed.
3385 195268 valgrind --log-file doesn't accept ~/...
3386 195838 VEX abort: LibVEX_N_SPILL_BYTES too small for CPUID boilerplate
3387 195860 WARNING: unhandled syscall: unix:223
3388 196528 need a error suppression for pthread_rwlock_init under os x?
3389 197227 Support aio_* syscalls on Darwin
3390 197456 valgrind should reject --suppressions=(directory)
3391 197512 DWARF2 CFI reader: unhandled CFI instruction 0:10
3392 197591 unhandled syscall 27 (mincore)
3393 197793 Merge DCAS branch to the trunk == 85756, 142103
3394 197794 Avoid duplicate filenames in Vex
3395 197898 make check fails on current SVN
3396 197901 make check fails also under exp-ptrcheck in current SVN
3397 197929 Make --leak-resolution=high the default
3398 197930 Reduce spacing between leak reports
3399 197933 Print command line of client at start-up, and shorten preamble
3400 197966 unhandled syscall 205 (x86-linux, --tool=exp-ptrcheck)
3401 198395 add BBV to the distribution as an experimental tool
3402 198624 Missing syscalls on Darwin: 82, 167, 281, 347
3403 198649 callgrind_annotate doesn't cumulate counters
3404 199338 callgrind_annotate sorting/thresholds are broken for all but Ir
3405 199977 Valgrind complains about an unrecognized instruction in the
3406 atomic_incs test program
3407 200029 valgrind isn't able to read Fedora 12 debuginfo
3408 200760 darwin unhandled syscall: unix:284
3409 200827 DRD doesn't work on Mac OS X
3410 200990 VG_(read_millisecond_timer)() does not work correctly
3411 201016 Valgrind does not support pthread_kill() on Mac OS
3412 201169 Document --read-var-info
3413 201323 Pre-3.5.0 performance sanity checking
3414 201384 Review user manual for the 3.5.0 release
3415 201585 mfpvr not implemented on ppc
3416 201708 tests failing because x86 direction flag is left set
3417 201757 Valgrind doesn't handle any recent sys_futex additions
3418 204377 64-bit valgrind can not start a shell script
3419 (with #!/path/to/shell) if the shell is a 32-bit executable
3420 n-i-bz drd: fixed assertion failure triggered by mutex reinitialization.
3421 n-i-bz drd: fixed a bug that caused incorrect messages to be printed
3422 about memory allocation events with memory access tracing enabled
3423 n-i-bz drd: fixed a memory leak triggered by vector clock deallocation
3425 (3.5.0: 19 Aug 2009, vex r1913, valgrind r10846).
3429 Release 3.4.1 (28 February 2009)
3430 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3431 3.4.1 is a bug-fix release that fixes some regressions and assertion
3432 failures in debug info reading in 3.4.0, most notably incorrect stack
3433 traces on amd64-linux on older (glibc-2.3 based) systems. Various
3434 other debug info problems are also fixed. A number of bugs in the
3435 exp-ptrcheck tool introduced in 3.4.0 have been fixed.
3437 In view of the fact that 3.4.0 contains user-visible regressions
3438 relative to 3.3.x, upgrading to 3.4.1 is recommended. Packagers are
3439 encouraged to ship 3.4.1 in preference to 3.4.0.
3441 The fixed bugs are as follows. Note that "n-i-bz" stands for "not in
3442 bugzilla" -- that is, a bug that was reported to us but never got a
3443 bugzilla entry. We encourage you to file bugs in bugzilla
3444 (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than mailing the
3445 developers (or mailing lists) directly -- bugs that are not entered
3446 into bugzilla tend to get forgotten about or ignored.
3448 n-i-bz Fix various bugs reading icc-11 generated debug info
3449 n-i-bz Fix various bugs reading gcc-4.4 generated debug info
3450 n-i-bz Preliminary support for glibc-2.10 / Fedora 11
3451 n-i-bz Cachegrind and Callgrind: handle non-power-of-two cache sizes,
3452 so as to support (eg) 24k Atom D1 and Core2 with 3/6/12MB L2.
3453 179618 exp-ptrcheck crashed / exit prematurely
3454 179624 helgrind: false positive races with pthread_create and
3455 recv/open/close/read
3456 134207 pkg-config output contains @VG_PLATFORM@
3457 176926 floating point exception at valgrind startup with PPC 440EPX
3458 181594 Bogus warning for empty text segment
3459 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant rex prefixes)
3460 181707 Dwarf3 doesn't require enumerations to have name
3461 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
3462 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
3463 Assertion '!already_present' failed.
3464 185359 exp-ptrcheck unhandled syscall getresuid()
3466 (3.4.1.RC1: 24 Feb 2008, vex r1884, valgrind r9253).
3467 (3.4.1: 28 Feb 2008, vex r1884, valgrind r9293).
3471 Release 3.4.0 (2 January 2009)
3472 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3473 3.4.0 is a feature release with many significant improvements and the
3474 usual collection of bug fixes. This release supports X86/Linux,
3475 AMD64/Linux, PPC32/Linux and PPC64/Linux. Support for recent distros
3476 (using gcc 4.4, glibc 2.8 and 2.9) has been added.
3478 3.4.0 brings some significant tool improvements. Memcheck can now
3479 report the origin of uninitialised values, the thread checkers
3480 Helgrind and DRD are much improved, and we have a new experimental
3481 tool, exp-Ptrcheck, which is able to detect overruns of stack and
3482 global arrays. In detail:
3484 * Memcheck is now able to track the origin of uninitialised values.
3485 When it reports an uninitialised value error, it will try to show
3486 the origin of the value, as either a heap or stack allocation.
3487 Origin tracking is expensive and so is not enabled by default. To
3488 use it, specify --track-origins=yes. Memcheck's speed will be
3489 essentially halved, and memory usage will be significantly
3490 increased. Nevertheless it can drastically reduce the effort
3491 required to identify the root cause of uninitialised value errors,
3492 and so is often a programmer productivity win, despite running more
3495 * A version (1.4.0) of the Valkyrie GUI, that works with Memcheck in
3496 3.4.0, will be released shortly.
3498 * Helgrind's race detection algorithm has been completely redesigned
3499 and reimplemented, to address usability and scalability concerns:
3501 - The new algorithm has a lower false-error rate: it is much less
3502 likely to report races that do not really exist.
3504 - Helgrind will display full call stacks for both accesses involved
3505 in a race. This makes it easier to identify the root causes of
3508 - Limitations on the size of program that can run have been removed.
3510 - Performance has been modestly improved, although that is very
3513 - Direct support for Qt4 threading has been added.
3515 - pthread_barriers are now directly supported.
3517 - Helgrind works well on all supported Linux targets.
3519 * The DRD thread debugging tool has seen major improvements:
3521 - Greatly improved performance and significantly reduced memory
3524 - Support for several major threading libraries (Boost.Thread, Qt4,
3525 glib, OpenMP) has been added.
3527 - Support for atomic instructions, POSIX semaphores, barriers and
3528 reader-writer locks has been added.
3530 - Works now on PowerPC CPUs too.
3532 - Added support for printing thread stack usage at thread exit time.
3534 - Added support for debugging lock contention.
3536 - Added a manual for Drd.
3538 * A new experimental tool, exp-Ptrcheck, has been added. Ptrcheck
3539 checks for misuses of pointers. In that sense it is a bit like
3540 Memcheck. However, Ptrcheck can do things Memcheck can't: it can
3541 detect overruns of stack and global arrays, it can detect
3542 arbitrarily far out-of-bounds accesses to heap blocks, and it can
3543 detect accesses heap blocks that have been freed a very long time
3544 ago (millions of blocks in the past).
3546 Ptrcheck currently works only on x86-linux and amd64-linux. To use
3547 it, use --tool=exp-ptrcheck. A simple manual is provided, as part
3548 of the main Valgrind documentation. As this is an experimental
3549 tool, we would be particularly interested in hearing about your
3550 experiences with it.
3552 * exp-Omega, an experimental instantaneous leak-detecting tool, is no
3553 longer built by default, although the code remains in the repository
3554 and the tarball. This is due to three factors: a perceived lack of
3555 users, a lack of maintenance, and concerns that it may not be
3556 possible to achieve reliable operation using the existing design.
3558 * As usual, support for the latest Linux distros and toolchain
3559 components has been added. It should work well on Fedora Core 10,
3560 OpenSUSE 11.1 and Ubuntu 8.10. gcc-4.4 (in its current pre-release
3561 state) is supported, as is glibc-2.9. The C++ demangler has been
3562 updated so as to work well with C++ compiled by even the most recent
3565 * You can now use frame-level wildcards in suppressions. This was a
3566 frequently-requested enhancement. A line "..." in a suppression now
3567 matches zero or more frames. This makes it easier to write
3568 suppressions which are precise yet insensitive to changes in
3571 * 3.4.0 adds support on x86/amd64 for the SSSE3 instruction set.
3573 * Very basic support for IBM Power6 has been added (64-bit processes only).
3575 * Valgrind is now cross-compilable. For example, it is possible to
3576 cross compile Valgrind on an x86/amd64-linux host, so that it runs
3577 on a ppc32/64-linux target.
3579 * You can set the main thread's stack size at startup using the
3580 new --main-stacksize= flag (subject of course to ulimit settings).
3581 This is useful for running apps that need a lot of stack space.
3583 * The limitation that you can't use --trace-children=yes together
3584 with --db-attach=yes has been removed.
3586 * The following bugs have been fixed. Note that "n-i-bz" stands for
3587 "not in bugzilla" -- that is, a bug that was reported to us but
3588 never got a bugzilla entry. We encourage you to file bugs in
3589 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3590 mailing the developers (or mailing lists) directly.
3592 n-i-bz Make return types for some client requests 64-bit clean
3593 n-i-bz glibc 2.9 support
3594 n-i-bz ignore unsafe .valgrindrc's (CVE-2008-4865)
3595 n-i-bz MPI_Init(0,0) is valid but libmpiwrap.c segfaults
3596 n-i-bz Building in an env without gdb gives bogus gdb attach
3597 92456 Tracing the origin of uninitialised memory
3598 106497 Valgrind does not demangle some C++ template symbols
3600 151612 Suppression with "..." (frame-level wildcards in .supp files)
3601 156404 Unable to start oocalc under memcheck on openSUSE 10.3 (64-bit)
3602 159285 unhandled syscall:25 (stime, on x86-linux)
3603 159452 unhandled ioctl 0x8B01 on "valgrind iwconfig"
3604 160954 ppc build of valgrind crashes with illegal instruction (isel)
3605 160956 mallinfo implementation, w/ patch
3606 162092 Valgrind fails to start gnome-system-monitor
3607 162819 malloc_free_fill test doesn't pass on glibc2.8 x86
3608 163794 assertion failure with "--track-origins=yes"
3609 163933 sigcontext.err and .trapno must be set together
3610 163955 remove constraint !(--db-attach=yes && --trace-children=yes)
3611 164476 Missing kernel module loading system calls
3612 164669 SVN regression: mmap() drops posix file locks
3613 166581 Callgrind output corruption when program forks
3614 167288 Patch file for missing system calls on Cell BE
3615 168943 unsupported scas instruction pentium
3616 171645 Unrecognised instruction (MOVSD, non-binutils encoding)
3617 172417 x86->IR: 0x82 ...
3618 172563 amd64->IR: 0xD9 0xF5 - fprem1
3619 173099 .lds linker script generation error
3620 173177 [x86_64] syscalls: 125/126/179 (capget/capset/quotactl)
3621 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant prefixes)
3623 174908 --log-file value not expanded correctly for core file
3624 175044 Add lookup_dcookie for amd64
3625 175150 x86->IR: 0xF2 0xF 0x11 0xC1 (movss non-binutils encoding)
3627 Developer-visible changes:
3629 * Valgrind's debug-info reading machinery has been majorly overhauled.
3630 It can now correctly establish the addresses for ELF data symbols,
3631 which is something that has never worked properly before now.
3633 Also, Valgrind can now read DWARF3 type and location information for
3634 stack and global variables. This makes it possible to use the
3635 framework to build tools that rely on knowing the type and locations
3636 of stack and global variables, for example exp-Ptrcheck.
3638 Reading of such information is disabled by default, because most
3639 tools don't need it, and because it is expensive in space and time.
3640 However, you can force Valgrind to read it, using the
3641 --read-var-info=yes flag. Memcheck, Helgrind and DRD are able to
3642 make use of such information, if present, to provide source-level
3643 descriptions of data addresses in the error messages they create.
3645 (3.4.0.RC1: 24 Dec 2008, vex r1878, valgrind r8882).
3646 (3.4.0: 3 Jan 2009, vex r1878, valgrind r8899).