1 Release 3.21.0 (?? Apr 2023)
2 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
5 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
6 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
7 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
8 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
9 AMD64/macOS 10.13 and nanoMIPS/Linux.
11 * ==================== CORE CHANGES ===================
13 * When GDB is used to debug a program running under valgrind using
14 the valgrind gdbserver, GDB will automatically load some
15 python code provided in valgrind defining GDB front end commands
16 corresponding to the valgrind monitor commands.
17 These GDB front end commands accept the same format as
18 the monitor commands directly sent to the Valgrind gdbserver.
19 These GDB front end commands provide a better integration
20 in the GDB command line interface, so as to use for example
21 GDB auto-completion, command specific help, searching for
22 a command or command help matching a regexp, ...
23 For relevant monitor commands, GDB will evaluate arguments
24 to make the use of monitor commands easier.
25 For example, instead of having to print the address of a variable
26 to pass it to a subsequent monitor command, the GDB front end
27 command will evaluate the address argument. It is for example
29 (gdb) memcheck who_point_at &some_struct sizeof(some_struct)
32 $2 = (some_struct_type *) 0x1130a0 <some_struct>
33 (gdb) p sizeof(some_struct)
35 (gdb) monitor who_point_at 0x1130a0 40
37 * The vgdb utility now supports extended-remote protocol when
38 invoked with --multi. In this mode the GDB run command is
39 supported. Which means you don't need to run gdb and valgrind
40 from different terminals. So for example to start you program
41 in gdb and run it under valgrind you can do:
43 (gdb) set remote exec-file prog
45 (gdb) target extended-remote | vgdb --multi
48 * The behaviour of realloc with a size of zero can now
49 be changed for tools that intercept malloc. Those
50 tools are memcheck, helgrind, drd, massif and dhat.
51 Realloc implementations generally do one of two things
52 - free the memory like free() and return NULL
53 (GNU libc and ptmalloc).
54 - either free the memory and then allocate a
55 minumum siized block or just return the
56 original pointer. Return NULL if the
57 allocation of the minimum sized block fails
58 (jemalloc, musl, snmalloc, Solaris, macOS).
59 When Valgrind is configured and built it will
60 try to match the OS and libc behaviour. However
61 if you are using a non-default library to replace
62 malloc and family (e.g., musl on a glibc Linux or
63 tcmalloc on FreeBSD) then you can use a command line
64 option to change the behaviour of Valgrind:
65 --realloc-zero-bytes-frees=yes|no [yes on Linux glibc, no otherwise]
67 * ================== PLATFORM CHANGES =================
69 * Make the address space limit on FreeBSD amd64 128Gbytes
70 (the same as Linux and Solaris, it was 32Gbytes)
72 * ==================== TOOL CHANGES ===================
75 - When doing a delta leak_search, it is now possible to only
76 output the new loss records compared to the previous leak search.
77 This is available in the memcheck monitor command 'leak_search'
78 by specifying the "new" keyword or in your program by using
79 the client request VALGRIND_DO_NEW_LEAK_CHECK.
80 Whenever a "delta" leak search is done (i.e. when specifying
81 "new" or "increased" or "changed" in the monitor command),
82 the new loss records have a "new" marker.
83 - Valgrind now contains python code that defines GDB memcheck
84 front end monitor commands. See CORE CHANGES.
85 - Performs checks for the use of realloc with a size of zero.
86 This is non-portable and a source of errors. If memcheck
87 detects such a usage it will generate an error
89 followed by the usual callstacks.
90 A switch has been added to allow this to be turned off:
91 --show-realloc-size-zero=yes|no [yes]
94 - The option ---history-backtrace-size=<number> allows to configure
95 the number of entries to record in the stack traces of "old"
96 accesses. Previously, this number was hardcoded to 8.
97 - Valgrind now contains python code that defines GDB helgrind
98 front end monitor commands. See CORE CHANGES.
101 - `cg_annotate` has been rewritten from Perl into Python. The new
102 version is twice as fast, has more flexible argument parsing, and
103 will make future improvements easier.
106 - Valgrind now contains python code that defines GDB callgrind
107 front end monitor commands. See CORE CHANGES.
110 - Valgrind now contains python code that defines GDB massif
111 front end monitor commands. See CORE CHANGES.
113 * ==================== FIXED BUGS ====================
115 The following bugs have been fixed or resolved. Note that "n-i-bz"
116 stands for "not in bugzilla" -- that is, a bug that was reported to us
117 but never got a bugzilla entry. We encourage you to file bugs in
118 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
119 than mailing the developers (or mailing lists) directly -- bugs that
120 are not entered into bugzilla tend to get forgotten about or ignored.
122 170510 Don't warn about ioctl of size 0 without direction hint
123 241072 List tools in --help output
124 327548 false positive while destroying mutex
125 382034 Testcases build fixes for musl
126 351857 confusing error message about valid command line option
127 392331 Spurious lock not held error from inside pthread_cond_timedwait
128 400793 pthread_rwlock_timedwrlock false positive
129 433873 openat2 syscall unimplemented on Linux
130 434057 Add stdio mode to valgrind's gdbserver
131 435441 valgrind fails to interpose malloc on musl 1.2.2 due to weak symbol name and no libc soname
132 444110 priv/guest_ppc_toIR.c:36198:31: warning: duplicated 'if' condition.
133 444488 Use glibc.pthread.stack_cache_size tunable
134 444568 drd/tests/pth_barrier_thr_cr fails on Fedora 38
135 445743 "The impossible happened: mutex is locked simultaneously by two threads"
136 while using mutexes with priority inheritance and signals
137 449309 Missing loopback device ioctl(s)
138 459476 vgdb: allow address reuse to avoid "address already in use" errorsuse" errors
139 460356 s390: Sqrt32Fx4 -- cannot reduce tree
140 462830 WARNING: unhandled amd64-freebsd syscall: 474
141 463027 broken check for MPX instruction support in assembler
142 464476 Firefox fails to start under Valgrind
143 464609 Valgrind memcheck should support Linux pidfd_open
144 464680 Show issues caused by memory policies like selinux deny_execmem
145 464859 Build failures with GCC-13 (drd tsan_unittest)
146 464969 D language demangling
147 465435 m_libcfile.c:66 (vgPlain_safe_fd): Assertion 'newfd >= VG_(fd_hard_limit)' failed.
148 466104 aligned_alloc problems, part 1
149 467482 Build failure on aarch64 Alpine
150 467839 Gdbserver: Improve compatibility of library directory name
151 n-i-bz FreeBSD rfork syscall fail with EINVAL or ENOSYS rather than VG_(unimplemented)
153 To see details of a given bug, visit
154 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
155 where XXXXXX is the bug number as listed above.
157 (3.21.0.RC1: ?? Apr 2023)
159 Release 3.20.0 (24 Oct 2022)
160 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
162 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
163 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
164 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
165 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
166 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
167 AMD64/macOS 10.13 and nanoMIPS/Linux.
169 * ==================== CORE CHANGES ===================
171 * The option "--vgdb-stop-at=event1,event2,..." accepts the new value abexit.
172 This indicates to invoke gdbserver when your program exits abnormally
173 (i.e. with a non zero exit code).
174 * Fix Rust v0 name demangling.
175 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
176 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
177 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with
179 * The option --enable-debuginfod=<no|yes> [default: yes] has been added on
181 * More DWARF5 support as generated by clang14.
183 * ==================== FIXED BUGS ====================
185 The following bugs have been fixed or resolved. Note that "n-i-bz"
186 stands for "not in bugzilla" -- that is, a bug that was reported to us
187 but never got a bugzilla entry. We encourage you to file bugs in
188 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
189 than mailing the developers (or mailing lists) directly -- bugs that
190 are not entered into bugzilla tend to get forgotten about or ignored.
192 131186 writev reports error in (vector[...])
193 434764 iconv_open causes ld.so v2.28+ to use optimised strncmp
194 446754 Improve error codes from alloc functions under memcheck
195 452274 memcheck crashes with Assertion 'sci->status.what == SsIdle' failed
196 452779 Valgrind fails to build on FreeBSD 13.0 with llvm-devel (15.0.0)
197 453055 shared_timed_mutex drd test fails with "Lock shared failed" message
198 453602 Missing command line option to enable/disable debuginfod
199 452802 Handle lld 9+ split RW PT_LOAD segments correctly
200 454040 s390x: False-positive memcheck:cond in memmem on arch13 systems
201 456171 [PATCH] FreeBSD: Don't record address errors when accessing the 'kern.ps_strings' sysctl struct
202 n-i-bz Implement vgdb invoker on FreeBSD
203 458845 PowerPC: The L field for the dcbf and sync instruction should be
205 458915 Remove register cache to fix 458915 gdbserver causes wrong syscall return
206 459031 Documentation on --error-exitcode incomplete
207 459477 XERROR messages lacks ending '\n' in vgdb
208 462007 Implicit int in none/tests/faultstatus.c
210 To see details of a given bug, visit
211 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
212 where XXXXXX is the bug number as listed above.
214 (3.20.0.RC1: 20 Oct 2022)
217 Release 3.19.0 (11 Apr 2022)
218 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
220 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
221 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
222 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
223 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
224 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
225 AMD64/macOS 10.13 and nanoMIPS/Linux.
227 * ==================== CORE CHANGES ===================
229 * Fix Rust v0 name demangling.
230 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
231 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
232 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with HardenedBSD
234 * ================== PLATFORM CHANGES =================
237 - ignore the "v8.x" architecture levels, only look at actual CPU features
238 present. Fixes mismatch detected between RDMA and atomics features
239 preventing startup on some QEMU configurations.
240 - Implement LD{,A}XP and ST{,L}XP
241 - Fix incorrect code emitted for doubleword CAS.
244 - Fix sys_ipc semtimedop syscall
245 - Fix VFLRX and WFLRX instructions
246 - Fix EXRL instruction with negative offset
249 - Reimplement the vbpermq instruction support to generate less Iops and
250 avoid overflowing internal buffers.
251 - Fix checking for scv support to avoid "Facility 'SCV' unavailable (12),
252 exception" messages in dmsg.
253 - Fix setting condition code for Vector Compare quad word instructions.
254 - Fix fix lxsibzx, lxsihzx and lxsihzx instructions so they only load
255 their respective sized data.
256 - Fix the prefixed stq instruction in PC relative mode.
258 * ==================== TOOL CHANGES ===================
261 - Speed up --track-origins=yes for large (in the range of hundreds to
262 thousands of megabytes) mmap/munmaps.
264 - Several fixes for new versions of libstd++ using new posix try_lock
267 * ==================== FIXED BUGS ====================
269 The following bugs have been fixed or resolved. Note that "n-i-bz"
270 stands for "not in bugzilla" -- that is, a bug that was reported to us
271 but never got a bugzilla entry. We encourage you to file bugs in
272 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
273 than mailing the developers (or mailing lists) directly -- bugs that
274 are not entered into bugzilla tend to get forgotten about or ignored.
276 403802 leak_cpp_interior fails with some reachable blocks different than expected
277 435732 memcheck/tests/leak_cpp_interior fails with gcc11
278 444242 s390x: Valgrind crashes on EXRL with negative offset
279 444399 arm64: unhandled instruction 0xC87F2D89 (LD{,A}XP and ST{,L}XP).
281 444481 gdb_server test failures on s390x
282 444495 dhat/tests/copy fails on s390x
283 444552 memcheck/tests/sem fails on s390x with glibc 2.34
284 444571 PPC, fix the lxsibzx and lxsihzx so they only load their respective
286 444836 PPC, pstq instruction for R=1 is not storing to the correct address.
287 444925 fexecve syscall wrapper not properly implemented
288 445032 valgrind/memcheck crash with SIGSEGV when SIGVTALRM timer used and
290 445211 Fix out of tree builds
291 445300 [PATCH] Fix building tests with Musl
292 445011 SIGCHLD is sent when valgrind uses debuginfod-find
293 445354 arm64 backend: incorrect code emitted for doubleword CAS
294 445415 arm64 front end: alignment checks missing for atomic instructions
295 445504 Using C++ condition_variable results in bogus "mutex is locked simultaneously by two threads" warning
296 445607 Unhandled amd64-freebsd syscall: 247
297 445668 Inline stack frame generation is broken for Rust binaries
298 445916 Demangle Rust v0 symbols with .llvm suffix
299 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false positives
300 446138 DRD/Helgrind with std::timed_mutex::try_lock_until false positives
301 446281 Add a DRD suppression for fwrite
302 446103 Memcheck: `--track-origins=yes` causes extreme slowdowns for large mmap/munmap
303 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false
304 446251 TARGET_SIGNAL_THR added to enum target_signal
305 446823 FreeBSD - missing syscalls when using libzm4
306 447991 s390x: Valgrind indicates illegal instruction on wflrx
307 447995 Valgrind segfault on power10 due to hwcap checking code
308 449483 Powerpc: vcmpgtsq., vcmpgtuq,, vcmpequq. instructions not setting the
309 condition code correctly.
310 449672 ppc64 --track-origins=yes failures because of bad cmov addHRegUse
311 449838 sigsegv liburing the 'impossible' happened for io_uring_setup
312 450025 Powerc: ACC file not implemented as a logical overlay of the VSR
314 450437 Warn for execve syscall with argv or argv[0] being NULL
315 450536 Powerpc: valgrind throws 'facility scv unavailable exception'
316 451626 Syscall param bpf(attr->raw_tracepoint.name) points to unaddressable byte(s)
317 451827 [ppc64le] VEX temporary storage exhausted with several vbpermq instructions
318 451843 valgrind fails to start on a FreeBSD system which enforces W^X
320 To see details of a given bug, visit
321 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
322 where XXXXXX is the bug number as listed above.
324 (3.19.0.RC1: 02 Apr 2022)
325 (3.19.0.RC2: 08 Apr 2022)
328 Release 3.18.0 (15 Oct 2021)
329 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
331 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
332 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
333 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
334 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
335 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
336 AMD64/macOS 10.13 and nanoMIPS/Linux.
338 * ==================== CORE CHANGES ===================
340 * The libiberty demangler has been updated, which brings support for
341 Rust v0 name demangling. [Update: alas, due to a bug, this support
342 isn't working in 3.18.0.]
344 * __libc_freeres isn't called anymore after the program recieves a
345 fatal signal. Causing some internal glibc resources to hang around,
346 but preventing any crashes after the program has ended.
348 * The DWARF reader is now very much faster at startup when just
349 --read-inline-info=yes (the default in most cases) is given.
351 * glibc 2.34, which moved various functions from libpthread.so into
352 libc.so, is now supported.
354 * ================== PLATFORM CHANGES =================
358 - v8.2 scalar and vector FABD, FACGE, FACGT and FADD.
359 - v8.2 FP compare & conditional compare instructions.
360 - Zero variants of v8.2 FP compare instructions.
364 - Support the miscellaneous-instruction-extensions facility 3 and
365 the vector-enhancements facility 2. This enables programs
366 compiled with "-march=arch13" or "-march=z15" to be executed
371 - ISA 3.1 support is now complete
372 - ISA 3.0 support for the darn instruction added.
373 - ISA 3.0 support for the vector system call instruction scv added.
374 - ISA 3.0 support for the copy, paste and cpabort instructions added.
376 * Support for X86/FreeBSD and AMD64/FreeBSD has been added.
378 * ==================== OTHER CHANGES ====================
380 * Memcheck on amd64: minor fixes to remove some false positive
383 * ==================== FIXED BUGS ====================
385 The following bugs have been fixed or resolved. Note that "n-i-bz"
386 stands for "not in bugzilla" -- that is, a bug that was reported to us
387 but never got a bugzilla entry. We encourage you to file bugs in
388 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
389 than mailing the developers (or mailing lists) directly -- bugs that
390 are not entered into bugzilla tend to get forgotten about or ignored.
392 208531 [PATCH]: FreeBSD support for valgrind
393 368960 WARNING: unhandled amd64-linux syscall: 163 (acct)
394 407589 [Linux] Add support for C11 aligned_alloc() and GNU reallocarray()
395 423963 Error in child thread when CLONE_PIDFD is used
396 426148 crash with "impossible happened" when running BPF CO-RE programs
397 429375 PPC ISA 3.1 support is missing, part 9
398 431157 PPC_FEATURE2_SCV needs to be masked in AT_HWCAP2
399 431306 Update demangler to support Rust v0 name mangling
400 432387 s390x: z15 instructions support
401 433437 FreeBSD support, part 1
402 433438 FreeBSD support, part 2
403 433439 FreeBSD support, part 3
404 433469 FreeBSD support, part 4
405 433473 FreeBSD support, part 5
406 433477 FreeBSD support, part 6
407 433479 FreeBSD support, part 7
408 433504 FreeBSD support, part 8
409 433506 FreeBSD support, part 9
410 433507 FreeBSD support, part 10
411 433508 FreeBSD support, part 11
412 433510 FreeBSD support, part 12
413 433801 PPC ISA 3.1 support is missing, part 10 (ISA 3.1 support complete)
414 433863 s390x: memcheck/tests/s390x/{cds,cs,csg} failures
415 434296 s390x: False-positive memcheck diagnostics from vector string
417 434840 PPC64 darn instruction not supported
418 435665 PPC ISA 3.0 copy, paste, cpabort instructions are not supported
419 435908 valgrind tries to fetch from deubginfod for files which already
420 have debug information
421 438871 unhandled instruction bytes: 0xF3 0x49 0xF 0x6F 0x9C 0x24 0x60 0x2
422 439046 valgrind is unusably large when linked with lld
423 439090 Implement close_range(2)
424 439326 Valgrind 3.17.0 won't compile with Intel 2021 oneAPI compilers
425 439590 glibc-2.34 breaks suppressions against obj:*/lib*/libc-2.*so*
426 440670 unhandled ppc64le-linux syscall: 252 statfs64 and 253 fstatfs64
427 440906 Fix impossible constraint issue in P10 testcase.
428 441512 Remove a unneeded / unnecessary prefix check.
429 441534 Update the expected output for test_isa_3_1_VRT.
430 442061 very slow execution under Fedora 34 (readdwarf3)
431 443031 Gcc -many change requires explicit .machine directives
432 443033 Add support for the ISA 3.0 mcrxrx instruction
433 443034 Sraw, srawi, srad, sradi, mfs
434 443178 Powerpc, test jm-mfspr expected output needs to be updated.
435 443179 Need new test for the lxvx and stxvx instructions on ISA 2.07 and
437 443180 The subnormal test and the ISA 3.0 test generate compiler warnings
438 443314 In the latest GIT version, Valgrind with "--trace-flags" crashes
440 443605 Don't call final_tidyup (__libc_freeres) on FatalSignal
442 To see details of a given bug, visit
443 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
444 where XXXXXX is the bug number as listed below.
446 (3.18.0.RC1: 12 Oct 2021)
447 (3.18.0: 15 Oct 2021)
451 Release 3.17.0 (19 Mar 2021)
452 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
454 3.17.0 fixes a number of bugs and adds some functional changes: support for
455 GCC 11, Clang 11, DWARF5 debuginfo, the 'debuginfod' debuginfo server, and
456 some new instructions for Arm64, S390 and POWER. There are also some tool
459 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
460 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
461 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
462 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
463 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
465 * ==================== CORE CHANGES ===================
467 * DWARF version 5 support. Valgrind can now read DWARF version 5 debuginfo as
470 * Valgrind now supports debuginfod, an HTTP server for distributing ELF/DWARF
471 debugging information. When a debuginfo file cannot be found locally,
472 Valgrind is able to query debuginfod servers for the file using its
473 build-id. See the user manual for more information about debuginfod support.
475 * ================== PLATFORM CHANGES =================
479 - Inaccuracies resulting from double-rounding in the simulation of
480 floating-point multiply-add/subtract instructions have been fixed. These
481 should now behave exactly as the hardware does.
483 - Partial support for the ARM v8.2 instruction set. v8.2 support work is
484 ongoing. Support for the half-word variants of at least the following
485 instructions has been added:
487 FABS <Vd>.<T>, <Vn>.<T>
489 FNEG <Vd>.<T>, <Vn>.<T>
491 FSQRT <Vd>.<T>, <Vn>.<T>
496 - Implement the new instructions/features that were added to z/Architecture
497 with the vector-enhancements facility 1. Also cover the instructions from
498 the vector-packed-decimal facility that are defined outside the chapter
499 "Vector Decimal Instructions", but not the ones from that chapter itself.
501 For a detailed list of newly supported instructions see the updates to
502 `docs/internals/s390-opcodes.csv'.
504 Since the miscellaneous instruction extensions facility 2 was already
505 added in Valgrind 3.16.0, this completes the support necessary to run
506 general programs built with `--march=z14' under Valgrind. The
507 vector-packed-decimal facility is currently not exploited by the standard
508 toolchain and libraries.
512 - Various bug fixes. Fix for the sync field to limit setting just two of
513 the two bits in the L-field. Fix the write size for the stxsibx and
514 stxsihx instructions. Fix the modsw and modsd instructions.
516 - Partial support for ISA 3.1 has been added. Support for the VSX PCV mask
517 instructions, bfloat16 GER instructions, and bfloat16 to/from float 32-bit
518 conversion instructions are still missing.
520 * ==================== TOOL CHANGES ====================
522 * General tool changes
524 - All the tools and their vgpreload libraries are now installed under
525 libexec because they cannot be executed directly and should be run through
526 the valgrind executable. This should be an internal, not user visible,
527 change, but might impact valgrind packagers.
529 - The --track-fds option now respects -q, --quiet and won't output anything
530 if no file descriptors are leaked. It also won't report the standard stdin
531 (0), stdout (1) or stderr (2) descriptors as being leaked with
532 --trace-fds=yes anymore. To track whether the standard file descriptors
533 are still open at the end of the program run use --trace-fds=all.
537 - DHAT has been extended, with two new modes of operation. The new
538 --mode=copy flag triggers copy profiling, which records calls to memcpy,
539 strcpy, and similar functions. The new --mode=ad-hoc flag triggers ad hoc
540 profiling, which records calls to the DHAT_AD_HOC_EVENT client request in
541 the new dhat/dhat.h file. This is useful for learning more about hot code
542 paths. See the user manual for more information about the new modes.
544 - Because of these changes, DHAT's file format has changed. DHAT output
545 files produced with earlier versions of DHAT will not work with this
546 version of DHAT's viewer, and DHAT output files produced with this version
547 of DHAT will not work with earlier versions of DHAT's viewer.
549 * ==================== FIXED BUGS ====================
551 The following bugs have been fixed or resolved. Note that "n-i-bz"
552 stands for "not in bugzilla" -- that is, a bug that was reported to us
553 but never got a bugzilla entry. We encourage you to file bugs in
554 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
555 than mailing the developers (or mailing lists) directly -- bugs that
556 are not entered into bugzilla tend to get forgotten about or ignored.
558 To see details of a given bug, visit
559 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
560 where XXXXXX is the bug number as listed below.
562 140178 open("/proc/self/exe", ...); doesn't quite work
563 140939 --track-fds reports leakage of stdout/in/err and doesn't respect -q
564 217695 malloc/calloc/realloc/memalign failure doesn't set errno to ENOMEM
565 338633 gdbserver_tests/nlcontrolc.vgtest hangs on arm64
566 345077 linux syscall execveat support (linux 3.19)
567 361770 Missing F_ADD_SEALS
568 369029 handle linux syscalls sched_getattr and sched_setattr
569 384729 __libc_freeres inhibits cross-platform valgrind
570 388787 Support for C++17 new/delete
571 391853 Makefile.all.am:L247 and @SOLARIS_UNDEF_LARGESOURCE@ being empty
572 396656 Warnings while reading debug info
573 397605 ioctl FICLONE mishandled
574 401416 Compile failure with openmpi 4.0
575 408663 Suppression file for musl libc
576 404076 s390x: z14 vector instructions not implemented
577 410743 shmat() calls for 32-bit programs fail when running in 64-bit valgrind
578 (actually affected all x86 and nanomips regardless of host bitness)
579 413547 regression test does not check for Arm 64 features.
580 414268 Enable AArch64 feature detection and decoding for v8.x instructions
581 415293 Incorrect call-graph tracking due to new _dl_runtime_resolve_xsave*
582 422174 unhandled instruction bytes: 0x48 0xE9 (REX prefixed JMP instruction)
583 422261 platform selection fails for unqualified client name
584 422623 epoll_ctl warns for uninitialized padding on non-amd64 64bit arches
585 423021 PPC: Add missing ISA 3.0 documentation link and HWCAPS test.
586 423195 PPC ISA 3.1 support is missing, part 1
587 423361 Adds io_uring support on arm64/aarch64 (and all other arches)
588 424012 crash with readv/writev having invalid but not NULL arg2 iovec
589 424298 amd64: Implement RDSEED
590 425232 PPC ISA 3.1 support is missing, part 2
591 425820 Failure to recognize vpcmpeqq as a dependency breaking idiom.
592 426014 arm64: implement fmadd and fmsub as Iop_MAdd/Sub
593 426123 PPC ISA 3.1 support is missing, part 3
594 426144 Fix "condition variable has not been initialized" on Fedora 33.
595 427400 PPC ISA 3.1 support is missing, part 4
596 427401 PPC ISA 3.1 support is missing, part 5
597 427404 PPC ISA 3.1 support is missing, part 6
598 427870 lmw, lswi and related PowerPC insns aren't allowed on ppc64le
599 427787 Support new faccessat2 linux syscall (439)
600 427969 debuginfo section duplicates a section in the main ELF file
601 428035 drd: Unbreak the musl build
602 428648 s390_emit_load_mem panics due to 20-bit offset for vector load
603 428716 cppcheck detects potential leak in VEX/useful/smchash.c
604 428909 helgrind: need to intercept duplicate libc definitions for Fedora 33
605 429352 PPC ISA 3.1 support is missing, part 7
606 429354 PPC ISA 3.1 support is missing, part 8
607 429692 unhandled ppc64le-linux syscall: 147 (getsid)
608 429864 s390x: C++ atomic test_and_set yields false-positive memcheck
610 429952 Errors when building regtest with clang
611 430354 ppc stxsibx and stxsihx instructions write too much data
612 430429 valgrind.h doesn't compile on s390x with clang
613 430485 expr_is_guardable doesn't handle Iex_Qop
614 431556 Complete arm64 FADDP v8.2 instruction support
615 432102 Add support for DWARF5 as produced by GCC11
616 432161 Addition of arm64 v8.2 FADDP, FNEG and FSQRT
617 432381 drd: Process STACK_REGISTER client requests
618 432552 [AArch64] invalid error emitted for pre-decremented byte/hword addresses
619 432672 vg_regtest: test-specific environment variables not reset between tests
620 432809 VEX should support REX.W + POPF
621 432861 PPC modsw and modsd give incorrect results for 1 mod 12
622 432870 gdbserver_tests:nlcontrolc hangs with newest glibc2.33 x86-64
623 432215 Add debuginfod functionality
624 433323 Use pkglibexecdir as vglibdir
625 433500 DRD regtest faulures when libstdc++ and libgcc debuginfo are installed
626 433629 valgrind/README has type "abd" instead of "and"
627 433641 Rust std::sys::unix::fs::try_statx Syscall param fstatat(file_name)
628 433898 arm64: Handle sp, lr, fp as DwReg in CfiExpr
629 434193 GCC 9+ inlined strcmp causes "Conditional jump or move [..] value" report
630 n-i-bz helgrind: If hg_cli__realloc fails, return NULL.
631 n-i-bz arm64 front end: avoid Memcheck false positives relating to CPUID
633 (3.17.0.RC1: 13 Mar 2021)
634 (3.17.0.RC2: 17 Mar 2021)
635 (3.17.0: 19 Mar 2021)
639 Release 3.16.1 (22 June 2020)
640 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
642 3.16.1 fixes two critical bugs discovered after 3.16.0 was frozen. It also
643 fixes character encoding problems in the documentation HTML.
645 422677 PPC sync instruction L field should only be 2 bits in ISA 3.0
646 422715 32-bit x86: vex: the `impossible' happened: expr_is_guardable: unhandled expr
648 (3.16.1, 22 June 2020, 36d6727e1d768333a536f274491e5879cab2c2f7)
652 Release 3.16.0 (27 May 2020)
653 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
655 3.16.0 is a feature release with many improvements and the usual collection of
658 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
659 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
660 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
661 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
662 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
664 * ==================== CORE CHANGES ===================
666 * It is now possible to dynamically change the value of many command line
667 options while your program (or its children) are running under Valgrind.
669 To see the list of dynamically changeable options, run
670 "valgrind --help-dyn-options".
672 You can change the options from the shell by using vgdb to launch
673 the monitor command "v.clo <clo option>...".
674 The same monitor command can be used from a gdb connected
675 to the valgrind gdbserver.
676 Your program can also change the dynamically changeable options using
677 the client request VALGRIND_CLO_CHANGE(option).
679 * ================== PLATFORM CHANGES =================
681 * MIPS: preliminary support for nanoMIPS instruction set has been added.
683 * ==================== TOOL CHANGES ====================
687 - The implicit memcpy done by each call to realloc now counts towards the
688 read and write counts of resized heap blocks, making those counts higher
693 - cg_annotate's --auto and --show-percs options now default to 'yes', because
694 they are usually wanted.
698 - callgrind_annotate's --auto and --show-percs options now default to 'yes',
699 because they are usually wanted.
701 - The command option --collect-systime has been enhanced to specify
702 the unit used to record the elapsed time spent during system calls.
703 The command option now accepts the values no|yes|msec|usec|nsec,
704 where yes is a synonym of msec. When giving the value nsec, the
705 system cpu time of system calls is also recorded.
709 - Several memcheck options are now dynamically changeable.
710 Use valgrind --help-dyn-options to list them.
712 - The release 3.15 introduced a backward incompatible change for
713 some suppression entries related to preadv and pwritev syscalls.
714 When reading a suppression entry using the unsupported 3.14 format,
715 valgrind will now produce a warning to say the suppression entry will not
716 work, and suggest the needed change.
718 - Significantly fewer false positive errors on optimised code generated by
719 Clang and GCC. In particular, Memcheck now deals better with the
720 situation where the compiler will transform C-level "A && B" into "B && A"
721 under certain circumstances (in which the transformation is valid).
722 Handling of integer equality/non-equality checks on partially defined
723 values is also improved on some architectures.
727 - The exprimental Stack and Global Array Checking tool has been removed.
728 It only ever worked on x86 and amd64, and even on those it had a
729 high false positive rate and was slow. An alternative for detecting
730 stack and global array overruns is using the AddressSanitizer (ASAN)
731 facility of the GCC and Clang compilers, which require you to rebuild
732 your code with -fsanitize=address.
734 * ==================== OTHER CHANGES ====================
736 * New and modified GDB server monitor features:
738 - Option -T tells vgdb to output a timestamp in the vgdb information messages.
740 - The gdbserver monitor commands that require an address and an optional
741 length argument now accepts the alternate 'C like' syntax "address[length]".
742 For example, the memcheck command "monitor who_points_at 0x12345678 120"
743 can now also be given as "monitor who_points_at 0x12345678[120]".
745 * ==================== FIXED BUGS ====================
747 The following bugs have been fixed or resolved. Note that "n-i-bz"
748 stands for "not in bugzilla" -- that is, a bug that was reported to us
749 but never got a bugzilla entry. We encourage you to file bugs in
750 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
751 than mailing the developers (or mailing lists) directly -- bugs that
752 are not entered into bugzilla tend to get forgotten about or ignored.
754 To see details of a given bug, visit
755 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
756 where XXXXXX is the bug number as listed below.
758 343099 Linux setns syscall wrapper missing, unhandled syscall: 308
759 == 368923 WARNING: unhandled arm64-linux syscall: 268 (setns)
760 == 369031 WARNING: unhandled amd64-linux syscall: 308 (setns)
761 385386 Assertion failed "szB >= CACHE_ENTRY_SIZE" at m_debuginfo/image.c:517
762 400162 Patch: Guard against __GLIBC_PREREQ for musl libc
763 400593 In Coregrind, use statx for some internal syscalls if [f]stat[64] fail
764 400872 Add nanoMIPS support to Valgrind
765 403212 drd/tests/trylock hangs on FreeBSD
766 404406 s390x: z14 miscellaneous instructions not implemented
767 405201 Incorrect size of struct vki_siginfo on 64-bit Linux architectures
768 406561 mcinfcallWSRU gdbserver_test fails on ppc64
769 406824 Unsupported baseline
770 407218 Add support for the copy_file_range syscall
771 407307 Intercept stpcpy also in ld.so for arm64
772 407376 Update Xen support to 4.12 (4.13, actually) and add more coverage
774 407764 drd cond_post_wait gets wrong (?) condition on s390x z13 system
775 408009 Expose rdrand and f16c even on avx if host cpu supports them
776 408091 Missing pkey syscalls
777 408414 Add support for missing for preadv2 and pwritev2 syscalls
778 409141 Valgrind hangs when SIGKILLed
779 409206 Support for Linux PPS and PTP ioctls
780 409367 exit_group() after signal to thread waiting in futex() causes hangs
781 409429 amd64: recognize 'cmpeq' variants as a dependency breaking idiom
782 409780 References to non-existent configure.in
783 410556 Add support for BLKIO{MIN,OPT} and BLKALIGNOFF ioctls
784 410599 Non-deterministic behaviour of pth_self_kill_15_other test
785 410757 discrepancy for preadv2/pwritev2 syscalls across different versions
786 411134 Allow the user to change a set of command line options during execution
787 411451 amd64->IR of bt/btc/bts/btr with immediate clears zero flag
788 412344 Problem setting mips flags with specific paths
789 412408 unhandled arm-linux syscall: 124 - adjtime - on arm-linux
790 413119 Ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
791 413330 avx-1 test fails on AMD EPYC 7401P 24-Core Processor
792 413603 callgrind_annotate/cg_annotate truncate function names at '#'
793 414565 Specific use case bug found in SysRes VG_(do_sys_sigprocmask)
794 415136 ARMv8.1 Compare-and-Swap instructions are not supported
795 415757 vex x86->IR: 0x66 0xF 0xCE 0x4F (bswapw)
796 416239 valgrind crashes when handling clock_adjtime
797 416285 Use prlimit64 in VG_(getrlimit) and VG_(setrlimit)
798 416286 DRD reports "conflicting load" error on std::mutex::lock()
799 416301 s390x: "compare and signal" not supported
800 416387 finit_module and bpf syscalls are unhandled on arm64
801 416464 Fix false reports for uninitialized memory for PR_CAPBSET_READ/DROP
802 416667 gcc10 ppc64le impossible constraint in 'asm' in test_isa.
803 416753 new 32bit time syscalls for 2038+
804 417075 pwritev(vector[...]) suppression ignored
805 417075 is not fixed, but incompatible supp entries are detected
806 and a warning is produced for these.
807 417187 [MIPS] Conditional branch problem since 'grail' changes
808 417238 Test memcheck/tests/vbit-test fails on mips64 BE
809 417266 Make memcheck/tests/linux/sigqueue usable with musl
810 417281 s390x: /bin/true segfaults with "grail" enabled
811 417427 commit to fix vki_siginfo_t definition created numerous regression
813 417452 s390_insn_store_emit: dst->tag for HRcVec128
814 417578 Add suppressions for glibc DTV leaks
815 417906 clone with CLONE_VFORK and no CLONE_VM fails
816 418004 Grail code additions break ppc64.
817 418435 s390x: spurious "Conditional jump or move depends on uninitialised [..]"
818 418997 s390x: Support Iex_ITE for float and vector types
819 419503 s390x: Avoid modifying registers returned from isel functions
820 421321 gcc10 arm64 build needs __getauxval for linking with libgcc
821 421570 std_mutex fails on Arm v8.1 h/w
822 434035 vgdb might crash if valgrind is killed
823 n-i-bz Fix minor one time leaks in dhat.
824 n-i-bz Add --run-cxx-freeres=no in outer args to avoid inner crashes.
825 n-i-bz Add support for the Linux io_uring system calls
826 n-i-bz sys_statx: don't complain if both |filename| and |buf| are NULL.
827 n-i-bz Fix non-glibc build of test suite with s390x_features
828 n-i-bz MinGW, include/valgrind.h: Fix detection of 64-bit mode
829 423195 PPC ISA 3.1 support is missing, part 1
831 (3.16.0.RC1: 18 May 2020, git 6052ee66a0cf5234e8e2a2b49a8760226bc13b92)
832 (3.16.0.RC2: 19 May 2020, git 940ec1ca69a09f7fdae3e800b7359f85c13c4b37)
833 (3.16.0: 27 May 2020, git bf5e647edb9e96cbd5c57cc944984402eeee296d)
837 Release 3.15.0 (12 April 2019)
838 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
840 3.15.0 is a feature release with many improvements and the usual collection of
843 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
844 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
845 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
846 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
847 support for X86/macOS 10.13 and AMD64/macOS 10.13.
849 * ==================== CORE CHANGES ===================
851 * The XTree Massif output format now makes use of the information obtained
852 when specifying --read-inline-info=yes.
854 * amd64 (x86_64): the RDRAND and F16C insn set extensions are now supported.
856 * ==================== TOOL CHANGES ====================
860 - DHAT been thoroughly overhauled, improved, and given a GUI. As a result,
861 it has been promoted from an experimental tool to a regular tool. Run it
862 with --tool=dhat instead of --tool=exp-dhat.
864 - DHAT now prints only minimal data when the program ends, instead writing
865 the bulk of the profiling data to a file. As a result, the --show-top-n
866 and --sort-by options have been removed.
868 - Profile results can be viewed with the new viewer, dh_view.html. When
869 a run ends, a short message is printed, explaining how to view the result.
871 - See the documentation for more details.
875 - cg_annotate has a new option, --show-percs, which prints percentages next
880 - callgrind_annotate has a new option, --show-percs, which prints percentages
881 next to all event counts.
883 - callgrind_annotate now inserts commas in call counts, and
884 sort the caller/callee lists in the call tree.
888 - The default value for --read-inline-info is now "yes" on
889 Linux/Android/Solaris. It is still "no" on other OS.
893 - The option --xtree-leak=yes (to output leak result in xtree format)
894 automatically activates the option --show-leak-kinds=all, as xtree
895 visualisation tools such as kcachegrind can in any case select what kind
896 of leak to visualise.
898 - There has been further work to avoid false positives. In particular,
899 integer equality on partially defined inputs (C == and !=) is now handled
902 * ==================== OTHER CHANGES ====================
904 * The new option --show-error-list=no|yes displays, at the end of the run, the
905 list of detected errors and the used suppressions. Prior to this change,
906 showing this information could only be done by specifying "-v -v", but that
907 also produced a lot of other possibly-non-useful messages. The option -s is
908 equivalent to --show-error-list=yes.
910 * ==================== FIXED BUGS ====================
912 The following bugs have been fixed or resolved. Note that "n-i-bz"
913 stands for "not in bugzilla" -- that is, a bug that was reported to us
914 but never got a bugzilla entry. We encourage you to file bugs in
915 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
916 than mailing the developers (or mailing lists) directly -- bugs that
917 are not entered into bugzilla tend to get forgotten about or ignored.
919 To see details of a given bug, visit
920 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
921 where XXXXXX is the bug number as listed below.
923 385411 s390x: z13 vector floating-point instructions not implemented
924 397187 z13 vector register support for vgdb gdbserver
925 398183 Vex errors with _mm256_shuffle_epi8/vpshufb
926 398870 Please add support for instruction vcvtps2ph
927 399287 amd64 front end: Illegal Instruction vcmptrueps
928 399301 Use inlined frames in Massif XTree output.
929 399322 Improve callgrind_annotate output
930 399444 VEX/priv/guest_s390_toIR.c:17407: (style) Mismatching assignment [..]
931 400164 helgrind test encounters mips x-compiler warnings and assembler error
932 400490 s390x: VRs allocated as if separate from FPRs
933 400491 s390x: Operand of LOCH treated as unsigned integer
934 400975 Compile error: error: '-mips64r2' conflicts with the other architecture
935 options, which specify a mips64 processor
936 401112 LLVM 5.0 generates comparison against partially initialized data
937 401277 More bugs in z13 support
938 401454 Add a --show-percs option to cg_annotate and callgrind_annotate.
939 401578 drd: crashes sometimes on fork()
940 401627 memcheck errors with glibc avx2 optimized wcsncmp
941 401822 none/tests/ppc64/jm-vmx fails and produces assembler warnings
942 401827 none/tests/ppc64/test_isa_2_06_part3 failure on ppc64le (xvrsqrtesp)
943 401828 none/tests/ppc64/test_isa_2_06_part1 failure on ppc64le (fcfids and
945 402006 mark helper regs defined in final_tidyup before freeres_wrapper call
946 402048 WARNING: unhandled ppc64[be|le]-linux syscall: 26 (ptrace)
947 402123 invalid assembler opcodes for mips32r2
948 402134 assertion fail in mc_translate.c (noteTmpUsesIn) Iex_VECRET on arm64
949 402327 Warning: DWARF2 CFI reader: unhandled DW_OP_ opcode 0x13 (DW_OP_drop)
950 402341 drd/tests/tsan_thread_wrappers_pthread.h:369: suspicious code ?
951 402351 mips64 libvexmultiarch_test fails on s390x
953 402395 coregrind/vgdb-invoker-solaris.c: 2 * poor error checking
954 402480 Do not use %rsp in clobber list
955 402481 vbit-test fails on x86 for Iop_CmpEQ64 iselInt64Expr Sar64
956 402515 Implement new option --show-error-list=no|yes / -s
957 402519 POWER 3.0 addex instruction incorrectly implemented
958 402781 Redo the cache used to process indirect branch targets
959 403123 vex amd64->IR:0xF3 0x48 0xF 0xAE 0xD3 (wrfsbase)
960 403552 s390x: wrong facility bit checked for vector facility
961 404054 memcheck powerpc subfe x, x, x initializes x to 0 or -1 based on CA
962 404638 Add VG_(replaceIndexXA)
963 404843 s390x: backtrace sometimes ends prematurely
964 404888 autotools cleanup series
965 405079 unhandled ppc64le-linux syscall: 131 (quotactl)
966 405182 Valgrind fails to build with Clang
967 405205 filter_libc: remove the line holding the futex syscall error entirely
968 405356 PPC64, xvcvsxdsp, xvcvuxdsp are supposed to write the 32-bit result to
969 the upper and lower 32-bits of the 64-bit result
970 405362 PPC64, vmsummbm instruction doesn't handle overflow case correctly
971 405363 PPC64, xvcvdpsxws, xvcvdpuxws, do not handle NaN arguments correctly.
972 405365 PPC64, function _get_maxmin_fp_NaN() doesn't handle QNaN, SNaN case
974 405403 s390x disassembler cannot be used on x86
975 405430 Use gcc -Wimplicit-fallthrough=2 by default if available
976 405458 MIPS mkFormVEC arguments swapped?
977 405716 drd: Fix an integer overflow in the stack margin calculation
978 405722 Support arm64 core dump
979 405733 PPC64, xvcvdpsp should write 32-bit result to upper and lower 32-bits
980 of the 64-bit destination field.
981 405734 PPC64, vrlwnm, vrlwmi, vrldrm, vrldmi do not work properly when me < mb
982 405782 "VEX temporary storage exhausted" when attempting to debug slic3r-pe
983 406198 none/tests/ppc64/test_isa_3_0_other test sporadically including CA
985 406256 PPC64, vector floating point instructions don't handle subnormal
986 according to VSCR[NJ] bit setting.
987 406352 cachegrind/callgrind fails ann tests because of missing a.c
988 406354 dhat is broken on x86 (32bit)
989 406355 mcsignopass, mcsigpass, mcbreak fail due to difference in gdb output
990 406357 gdbserver_tests fails because of gdb output change
991 406360 memcheck/tests/libstdc++.supp needs more supression variants
992 406422 none/tests/amd64-linux/map_32bits.vgtest fails too easily
993 406465 arm64 insn selector fails on "t0 = <expr>" where <expr> has type Ity_F16
994 407340 PPC64, does not support the vlogefp, vexptefp instructions.
995 n-i-bz add syswrap for PTRACE_GET|SET_THREAD_AREA on amd64.
996 n-i-bz Fix callgrind_annotate non deterministic order for equal total
997 n-i-bz callgrind_annotate --threshold=100 does not print all functions.
998 n-i-bz callgrind_annotate Use of uninitialized value in numeric gt (>)
999 n-i-bz amd64 (x86_64): RDRAND and F16C insn set extensions are supported
1001 (3.15.0.RC1: 8 April 2019, git ce94d674de5b99df173aad4c3ee48fc2a92e5d9c)
1002 (3.15.0.RC2: 11 April 2019, git 0c8be9bbede189ec580ec270521811766429595f)
1003 (3.15.0: 14 April 2019, git 270037da8b508954f0f7d703a0bebf5364eec548)
1007 Release 3.14.0 (9 October 2018)
1008 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1010 3.14.0 is a feature release with many improvements and the usual collection of
1013 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1014 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1015 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
1016 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
1017 support for X86/macOS 10.13, AMD64/macOS 10.13.
1019 * ==================== CORE CHANGES ===================
1021 * The new option --keep-debuginfo=no|yes (default no) can be used to retain
1022 debug info for unloaded code. This allows saved stack traces (e.g. for
1023 memory leaks) to include file/line info for code that has been dlclose'd (or
1024 similar). See the user manual for more information and known limitations.
1026 * Ability to specify suppressions based on source file name and line number.
1028 * Majorly overhauled register allocator. No end-user changes, but the JIT
1029 generates code a bit more quickly now.
1031 * ================== PLATFORM CHANGES =================
1033 * Preliminary support for macOS 10.13 has been added.
1035 * mips: support for MIPS32/MIPS64 Revision 6 has been added.
1037 * mips: support for MIPS SIMD architecture (MSA) has been added.
1039 * mips: support for MIPS N32 ABI has been added.
1041 * s390: partial support for vector instructions (integer and string) has been
1044 * ==================== TOOL CHANGES ====================
1046 * Helgrind: Addition of a flag
1047 --delta-stacktrace=no|yes [yes on linux amd64/x86]
1048 which specifies how full history stack traces should be computed.
1049 Setting this to =yes can speed up Helgrind by 25% when using
1050 --history-level=full.
1052 * Memcheck: reduced false positive rate for optimised code created by Clang 6
1053 / LLVM 6 on x86, amd64 and arm64. In particular, Memcheck analyses code
1054 blocks more carefully to determine where it can avoid expensive definedness
1055 checks without loss of precision. This is controlled by the flag
1056 --expensive-definedness-checks=no|auto|yes [auto].
1058 * ==================== OTHER CHANGES ====================
1060 * Valgrind is now buildable with link-time optimisation (LTO). A new
1061 configure option --enable-lto=yes allows building Valgrind with LTO. If the
1062 toolchain supports it, this produces a smaller/faster Valgrind (up to 10%).
1063 Note that if you are doing Valgrind development, --enable-lto=yes massively
1064 slows down the build process.
1066 * ==================== FIXED BUGS ====================
1068 The following bugs have been fixed or resolved. Note that "n-i-bz"
1069 stands for "not in bugzilla" -- that is, a bug that was reported to us
1070 but never got a bugzilla entry. We encourage you to file bugs in
1071 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1072 than mailing the developers (or mailing lists) directly -- bugs that
1073 are not entered into bugzilla tend to get forgotten about or ignored.
1075 To see details of a given bug, visit
1076 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1077 where XXXXXX is the bug number as listed below.
1079 79362 Debug info is lost for .so files when they are dlclose'd
1080 208052 strlcpy error when n = 0
1081 255603 exp-sgcheck Assertion '!already_present' failed
1082 338252 building valgrind with -flto (link time optimisation) fails
1083 345763 MIPS N32 ABI support
1084 368913 WARNING: unhandled arm64-linux syscall: 117 (ptrace)
1085 == 388664 unhandled arm64-linux syscall: 117 (ptrace)
1086 372347 Replacement problem of the additional c++14/c++17 new/delete operators
1087 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
1088 376257 helgrind history full speed up using a cached stack
1089 379373 Fix syscall param msg->desc.port.name points to uninitialised byte(s)
1091 379748 Fix missing pselect syscall (OS X 10.11)
1092 379754 Fix missing syscall ulock_wait (OS X 10.12)
1093 380397 s390x: __GI_strcspn() replacemenet needed
1094 381162 possible array overrun in VEX register allocator
1095 381272 ppc64 doesn't compile test_isa_2_06_partx.c without VSX support
1096 381274 powerpc too chatty even with --sigill-diagnostics=no
1097 381289 epoll_pwait can have a NULL sigmask
1098 381553 VEX register allocator v3
1099 381556 arm64: Handle feature registers access on 4.11 Linux kernel or later
1100 381769 Use ucontext_t instead of struct ucontext
1101 381805 arm32 needs ld.so index hardwire for new glibc security fixes
1102 382256 gz compiler flag test doesn't work for gold
1103 382407 vg_perf needs "--terse" command line option
1104 382515 "Assertion 'di->have_dinfo' failed." on wine's dlls/mscoree/tests/[..]
1105 382563 MIPS MSA ASE support
1106 382998 xml-socket doesn't work
1107 383275 massif: m_xarray.c:162 (ensureSpaceXA): Assertion '!xa->arr' failed
1108 383723 Fix missing kevent_qos syscall (macOS 10.11)
1109 == 385604 illegal hardware instruction (OpenCV cv::namedWindow)
1110 384096 Mention AddrCheck at Memcheck's command line option [..]
1111 384230 vex x86->IR: 0x67 0xE8 0xAB 0x68
1112 == 384156 vex x86->IR: 0x67 0xE8 0x6B 0x6A
1113 == 386115 vex x86->IR: 0x67 0xE8 0xD3 0x8B any program
1114 == 388407 vex x86->IR: 0x67 0xE8 0xAB 0x29
1115 == 394903 vex x86->IR: 0x67 0xE8 0x1B 0xDA
1116 384337 performance improvements to VEX register allocator v2 and v3
1117 384526 reduce number of spill insns generated by VEX register allocator v3
1118 384584 Callee saved regs listed first for AMD64, X86, and PPC architectures
1119 384631 Sanitise client args as printed with -v
1120 384633 Add a simple progress-reporting facility
1121 384987 VEX regalloc: allocate caller-save registers for short lived vregs
1122 385055 PPC VEX temporary storage exhausted
1123 385182 PPC64 is missing support for the DSCR
1124 385183 PPC64, Add support for xscmpeqdp, xscmpgtdp, xscmpgedp, xsmincdp
1125 385207 PPC64, generate_store_FPRF() generates too many Iops
1126 385208 PPC64, xxperm instruction exhausts temporary memory
1127 385210 PPC64, vpermr instruction could exhaust temporary memory
1128 385279 unhandled syscall: mach:43 (mach_generate_activity_id)
1129 == 395136 valgrind: m_syswrap/syswrap-main.c:438 (Bool eq_Syscall[..]
1130 == 387045 Valgrind crashing on High Sierra when testing any newly [..]
1131 385334 PPC64, fix vpermr, xxperm, xxpermr mask value.
1132 385408 s390x: z13 vector "support" instructions not implemented
1133 385409 s390x: z13 vector integer instructions not implemented
1134 385410 s390x: z13 vector string instructions not implemented
1135 385412 s390x: new non-vector z13 instructions not implemented
1136 385868 glibc ld.so _dl_runtime_resolve_avx_slow conditional jump warning.
1137 385912 none/tests/rlimit_nofile fails on newer glibc/kernel.
1138 385939 Optionally exit on the first error
1139 386318 valgrind.org/info/tools.html is missing SGCheck
1140 386425 running valgrind + wine on armv7l gives illegal opcode
1141 386397 PPC64, valgrind truncates powerpc timebase to 32-bits.
1142 387410 MIPSr6 support
1143 387664 Memcheck: make expensive-definedness-checks be the default
1144 387712 s390x cgijnl reports Conditional jump depends on uninitialised value
1145 387766 asm shifts cause false positive "Conditional jump or move depends
1146 on uninitialised value"
1147 387773 .gnu_debugaltlink paths resolve relative to .debug file, not symlink
1148 388174 valgrind with Wine quits with "Assertion 'cfsi_fits' failed"
1149 388786 Support bpf syscall in amd64 Linux
1150 388862 Add replacements for wmemchr and wcsnlen on Linux
1151 389065 valgrind meets gcc flag -Wlogical-op
1152 389373 exp-sgcheck the 'impossible' happened as Ist_LoadG is not instrumented
1153 390471 suppression by specification of source-file line number
1154 390723 make xtree dump files world wide readable, similar to log files
1155 391164 constraint bug in tests/ppc64/test_isa_2_07_part1.c for mtfprwa
1156 391861 Massif Assertion 'n_ips >= 1 && n_ips <= VG_(clo_backtrace_size)'
1157 392118 unhandled amd64-linux syscall: 332 (statx)
1158 392449 callgrind not clearing the number of calls properly
1159 393017 Add missing support for xsmaxcdp instruction, bug fixes for xsmincdp,
1160 lxssp, stxssp and stxvl instructions.
1161 393023 callgrind_control risks using the wrong vgdb
1162 393062 build-id ELF phdrs read causes "debuginfo reader: ensure_valid failed"
1163 393099 posix_memalign() invalid write if alignment == 0
1164 393146 failing assert "is_DebugInfo_active(di)"
1165 395709 PPC64 is missing support for the xvnegsp instruction
1166 395682 Accept read-only PT_LOAD segments and .rodata by ld -z separate-code
1168 396475 valgrind OS-X build: config.h not found (out-of-tree macOS builds)
1169 395991 arm-linux: wine's unit tests enter a signal delivery loop [..]
1170 396839 s390x: Trap instructions not implemented
1171 396887 arch_prctl should return EINVAL on unknown option
1172 == 397286 crash before launching binary (Unsupported arch_prctl option)
1173 == 397393 valgrind: the 'impossible' happened: (Archlinux)
1174 == 397521 valgrind: the 'impossible' happened: Unsupported [..]
1175 396906 compile tests failure on mips32-linux: broken inline asm in tests on
1177 397012 glibc ld.so uses arch_prctl on i386
1178 397089 amd64: Incorrect decoding of three-register vmovss/vmovsd opcode 11h
1179 397354 utimensat should ignore timespec tv_sec if tv_nsec is UTIME_NOW/OMIT
1180 397424 glibc 2.27 and gdb_server tests
1181 398028 Assertion `cfsi_fits` failing in simple C program
1182 398066 s390x: cgijl dep1, 0 reports false unitialised values warning
1184 n-i-bz Fix missing workq_ops operations (macOS)
1185 n-i-bz fix bug in strspn replacement
1186 n-i-bz Add support for the Linux BLKFLSBUF ioctl
1187 n-i-bz Add support for the Linux BLKREPORTZONE and BLKRESETZONE ioctls
1188 n-i-bz Fix possible stack trashing by semctl syscall wrapping
1189 n-i-bz Add support for the Linux membarrier() system call
1190 n-i-bz x86 front end: recognise and handle UD2 correctly
1191 n-i-bz Signal delivery for x86-linux: ensure that the stack pointer is
1192 correctly aligned before entering the handler.
1194 (3.14.0.RC1: 30 September 2018, git c2aeea2d28acb0639bcc8cc1e4ab115067db1eae)
1195 (3.14.0.RC2: 3 October 2018, git 3e214c4858a6fdd5697e767543a0c19e30505582)
1196 (3.14.0: 9 October 2018, git 353a3587bb0e2757411f9138f5e936728ed6cc4f)
1200 Release 3.13.0 (15 June 2017)
1201 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1203 3.13.0 is a feature release with many improvements and the usual collection of
1206 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1207 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1208 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
1209 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12.
1211 * ==================== CORE CHANGES ===================
1213 * The translation cache size has been increased to keep up with the demands of
1214 large applications. The maximum number of sectors has increased from 24 to
1215 48. The default number of sectors has increased from 16 to 32 on all
1216 targets except Android, where the increase is from 6 to 12.
1218 * The amount of memory that Valgrind can use has been increased from 64GB to
1219 128GB. In particular this means your application can allocate up to about
1220 60GB when running on Memcheck.
1222 * Valgrind's default load address has been changed from 0x3800'0000 to
1223 0x5800'0000, so as to make it possible to load larger executables. This
1224 should make it possible to load executables of size at least 1200MB.
1226 * A massive spaceleak caused by reading compressed debuginfo files has been
1227 fixed. Valgrind should now be entirely usable with gcc-7.0 "-gz" created
1230 * The C++ demangler has been updated.
1232 * Support for demangling Rust symbols has been added.
1234 * A new representation of stack traces, the "XTree", has been added. An XTree
1235 is a tree of stacktraces with data associated with the stacktraces. This is
1236 used by various tools (Memcheck, Helgrind, Massif) to report on the heap
1237 consumption of your program. Reporting is controlled by the new options
1238 --xtree-memory=none|allocs|full and --xtree-memory-file=<file>.
1240 A report can also be produced on demand using the gdbserver monitor command
1241 'xtmemory [<filename>]>'. The XTree can be output in 2 formats: 'callgrind
1242 format' and 'massif format. The existing visualisers for these formats (e.g.
1243 callgrind_annotate, KCachegrind, ms_print) can be used to visualise and
1244 analyse these reports.
1246 Memcheck can also produce XTree leak reports using the Callgrind file
1247 format. For more details, see the user manual.
1249 * ================== PLATFORM CHANGES =================
1251 * ppc64: support for ISA 3.0B and various fixes for existing 3.0 support
1253 * amd64: fixes for JIT failure problems on long AVX2 code blocks
1255 * amd64 and x86: support for CET prefixes has been added
1257 * arm32: a few missing ARMv8 instructions have been implemented
1259 * arm64, mips64, mips32: an alternative implementation of Load-Linked and
1260 Store-Conditional instructions has been added. This is to deal with
1261 processor implementations that implement the LL/SC specifications strictly
1262 and as a result cause Valgrind to hang in certain situations. The
1263 alternative implementation is automatically enabled at startup, as required.
1264 You can use the option --sim-hints=fallback-llsc to force-enable it if you
1267 * Support for OSX 10.12 has been improved.
1269 * On Linux, clone handling has been improved to honour CLONE_VFORK that
1270 involves a child stack. Note however that CLONE_VFORK | CLONE_VM is handled
1271 like CLONE_VFORK (by removing CLONE_VM), so applications that depend on
1272 CLONE_VM exact semantics will (still) not work.
1274 * The TileGX/Linux port has been removed because it appears to be both unused
1277 * ==================== TOOL CHANGES ====================
1281 - Memcheck should give fewer false positives when running optimised
1282 Clang/LLVM generated code.
1284 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1286 - New command line options --xtree-leak=no|yes and --xtree-leak-file=<file>
1287 to produce the end of execution leak report in a xtree callgrind format
1290 - New option 'xtleak' in the memcheck leak_check monitor command, to produce
1291 the leak report in an xtree file.
1295 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1297 - For some workloads (typically, for big applications), Massif memory
1298 consumption and CPU consumption has decreased significantly.
1302 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1304 - addition of client request VALGRIND_HG_GNAT_DEPENDENT_MASTER_JOIN, useful
1305 for Ada gnat compiled applications.
1307 * ==================== OTHER CHANGES ====================
1309 * For Valgrind developers: in an outer/inner setup, the outer Valgrind will
1310 append the inner guest stacktrace to the inner host stacktrace. This helps
1311 to investigate the errors reported by the outer, when they are caused by the
1312 inner guest program (such as an inner regtest). See README_DEVELOPERS for
1315 * To allow fast detection of callgrind files by desktop environments and file
1316 managers, the format was extended to have an optional first line that
1317 uniquely identifies the format ("# callgrind format"). Callgrind creates
1318 this line now, as does the new xtree functionality.
1320 * File name template arguments (such as --log-file, --xtree-memory-file, ...)
1321 have a new %n format letter that is replaced by a sequence number.
1323 * "--version -v" now shows the SVN revision numbers from which Valgrind was
1326 * ==================== FIXED BUGS ====================
1328 The following bugs have been fixed or resolved. Note that "n-i-bz"
1329 stands for "not in bugzilla" -- that is, a bug that was reported to us
1330 but never got a bugzilla entry. We encourage you to file bugs in
1331 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1332 than mailing the developers (or mailing lists) directly -- bugs that
1333 are not entered into bugzilla tend to get forgotten about or ignored.
1335 To see details of a given bug, visit
1336 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1337 where XXXXXX is the bug number as listed below.
1339 162848 --log-file output isn't split when a program forks
1340 340777 Illegal instruction on mips (ar71xx)
1341 341481 MIPS64: Iop_CmpNE32 triggers false warning on MIPS64 platforms
1342 342040 Valgrind mishandles clone with CLONE_VFORK | CLONE_VM that clones
1343 to a different stack.
1344 344139 x86 stack-seg overrides, needed by the Wine people
1345 344524 store conditional of guest applications always fail - observed on
1347 348616 Wine/valgrind: noted but unhandled ioctl 0x5390 [..] (DVD_READ_STRUCT)
1348 352395 Please provide SVN revision info in --version -v
1349 352767 Wine/valgrind: noted but unhandled ioctl 0x5307 [..] (CDROMSTOP)
1350 356374 Assertion 'DRD_(g_threadinfo)[tid].pt_threadid !=
1351 INVALID_POSIX_THREADID' failed
1352 358213 helgrind/drd bar_bad testcase hangs or crashes with new glibc pthread
1353 barrier implementation
1354 358697 valgrind.h: Some code remains even when defining NVALGRIND
1355 359202 Add musl libc configure/compile
1356 360415 amd64 instructions ADCX and ADOX are not implemented in VEX
1357 == 372828 (vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10)
1358 360429 unhandled ioctl 0x530d with no size/direction hints (CDROMREADMODE1)
1359 362223 assertion failed when .valgrindrc is a directory instead of a file
1360 367543 bt/btc/btr/bts x86/x86_64 instructions are poorly-handled wrt flags
1361 367942 Segfault vgPlain_do_sys_sigaction (m_signals.c:1138)
1362 368507 can't malloc chunks larger than about 34GB
1363 368529 Android arm target link error, missing atexit and pthread_atfork
1364 368863 WARNING: unhandled arm64-linux syscall: 100 (get_robust_list)
1365 368865 WARNING: unhandled arm64-linux syscall: 272 (kcmp)
1366 368868 disInstr(arm64): unhandled instruction 0xD53BE000 = cntfrq_el0 (ARMv8)
1367 368917 WARNING: unhandled arm64-linux syscall: 218 (request_key)
1368 368918 WARNING: unhandled arm64-linux syscall: 127 (sched_rr_get_interval)
1369 368922 WARNING: unhandled arm64-linux syscall: 161 (sethostname)
1370 368924 WARNING: unhandled arm64-linux syscall: 84 (sync_file_range)
1371 368925 WARNING: unhandled arm64-linux syscall: 130 (tkill)
1372 368926 WARNING: unhandled arm64-linux syscall: 97 (unshare)
1373 369459 valgrind on arm64 violates the ARMv8 spec (ldxr/stxr)
1374 370028 Reduce the number of compiler warnings on MIPS platforms
1375 370635 arm64 missing syscall getcpu
1376 371225 Fix order of timer_{gettime,getoverrun,settime} syscalls on arm64
1377 371227 Clean AArch64 syscall table
1378 371412 Rename wrap_sys_shmat to sys_shmat like other wrappers
1379 371471 Valgrind complains about non legit memory leaks on placement new (C++)
1380 371491 handleAddrOverrides() is [incorrect] when ASO prefix is used
1381 371503 disInstr(arm64): unhandled instruction 0xF89F0000
1382 371869 support '%' in symbol Z-encoding
1383 371916 execution tree xtree concept
1384 372120 c++ demangler demangles symbols which are not c++
1385 372185 Support of valgrind on ARMv8 with 32 bit executable
1386 372188 vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10 0x10 0x48 (PCMPxSTRx $0x10)
1387 372195 Power PC, xxsel instruction is not always recognized.
1388 372504 Hanging on exit_group
1389 372600 process loops forever when fatal signals are arriving quickly
1390 372794 LibVEX (arm32 front end): 'Assertion szBlg2 <= 3' failed
1391 373046 Stacks registered by core are never deregistered
1392 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
1393 373086 Implement additional Xen hypercalls
1394 373192 Calling posix_spawn in glibc 2.24 completely broken
1395 373488 Support for fanotify API on ARM64 architecture
1396 == 368864 WARNING: unhandled arm64-linux syscall: 262 (fanotify_init)
1397 373555 Rename BBPTR to GSPTR as it denotes guest state pointer only
1398 373938 const IRExpr arguments for matchIRExpr()
1399 374719 some spelling fixes
1400 374963 increase valgrind's load address to prevent mmap failure
1401 375514 valgrind_get_tls_addr() does not work in case of static TLS
1402 375772 +1 error in get_elf_symbol_info() when computing value of 'hi' address
1403 for ML_(find_rx_mapping)()
1404 375806 Test helgrind/tests/tc22_exit_w_lock fails with glibc 2.24
1405 375839 Temporary storage exhausted, with long sequence of vfmadd231ps insns
1406 == 377159 "vex: the `impossible' happened" still present
1407 == 375150 Assertion 'tres.status == VexTransOK' failed
1408 == 378068 valgrind crashes on AVX2 function in FFmpeg
1409 376142 Segfaults on MIPS Cavium Octeon boards
1410 376279 disInstr(arm64): unhandled instruction 0xD50320FF
1411 376455 Solaris: unhandled syscall lgrpsys(180)
1412 376518 Solaris: unhandled fast trap getlgrp(6)
1413 376611 ppc64 and arm64 don't know about prlimit64 syscall
1414 376729 PPC64, remove R2 from the clobber list
1416 376956 syswrap of SNDDRV and DRM_IOCTL_VERSION causing some addresses
1417 to be wrongly marked as addressable
1418 377066 Some Valgrind unit tests fail to compile on Ubuntu 16.10 with
1419 PIE enabled by default
1420 377376 memcheck/tests/linux/getregset fails with glibc2.24
1421 377427 PPC64, lxv instruction failing on odd destination register
1422 377478 PPC64: ISA 3.0 setup fixes
1423 377698 Missing memory check for futex() uaddr arg for FUTEX_WAKE
1424 and FUTEX_WAKE_BITSET, check only 4 args for FUTEX_WAKE_BITSET,
1425 and 2 args for FUTEX_TRYLOCK_PI
1426 377717 Fix massive space leak when reading compressed debuginfo sections
1427 377891 Update Xen 4.6 domctl wrappers
1428 377930 fcntl syscall wrapper is missing flock structure check
1429 378524 libvexmultiarch_test regression on s390x and ppc64
1430 378535 Valgrind reports INTERNAL ERROR in execve syscall wrapper
1431 378673 Update libiberty demangler
1432 378931 Add ISA 3.0B additional isnstructions, add OV32, CA32 setting support
1433 379039 syscall wrapper for prctl(PR_SET_NAME) must not check more than 16 bytes
1434 379094 Valgrind reports INTERNAL ERROR in rt_sigsuspend syscall wrapper
1435 379371 UNKNOWN task message [id 3444, to mach_task_self(), reply 0x603]
1436 (task_register_dyld_image_infos)
1437 379372 UNKNOWN task message [id 3447, to mach_task_self(), reply 0x603]
1438 (task_register_dyld_shared_cache_image_info)
1439 379390 unhandled syscall: mach:70 (host_create_mach_voucher_trap)
1440 379473 MIPS: add support for rdhwr cycle counter register
1441 379504 remove TileGX/Linux port
1442 379525 Support more x86 nop opcodes
1443 379838 disAMode(x86): not an addr!
1444 379703 PC ISA 3.0 fixes: stxvx, stxv, xscmpexpdp instructions
1445 379890 arm: unhandled instruction: 0xEBAD 0x1B05 (sub.w fp, sp, r5, lsl #4)
1446 379895 clock_gettime does not execute POST syscall wrapper
1447 379925 PPC64, mtffs does not set the FPCC and C bits in the FPSCR correctly
1448 379966 WARNING: unhandled amd64-linux syscall: 313 (finit_module)
1449 380200 xtree generated callgrind files refer to files without directory name
1450 380202 Assertion failure for cache line size (cls == 64) on aarch64.
1451 380397 s390x: __GI_strcspn() replacement needed
1452 n-i-bz Fix pub_tool_basics.h build issue with g++ 4.4.7.
1454 (3.13.0.RC1: 2 June 2017, vex r3386, valgrind r16434)
1455 (3.13.0.RC2: 9 June 2017, vex r3389, valgrind r16443)
1456 (3.13.0: 14 June 2017, vex r3396, valgrind r16446)
1460 Release 3.12.0 (20 October 2016)
1461 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1463 3.12.0 is a feature release with many improvements and the usual
1464 collection of bug fixes.
1466 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1467 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1468 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1469 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1470 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1471 X86/MacOSX 10.11/12, AMD64/MacOSX 10.11/12 and TILEGX/Linux.
1473 * ================== PLATFORM CHANGES =================
1475 * POWER: Support for ISA 3.0 has been added
1477 * mips: support for O32 FPXX ABI has been added.
1478 * mips: improved recognition of different processors
1479 * mips: determination of page size now done at run time
1481 * amd64: Partial support for AMD FMA4 instructions.
1483 * arm, arm64: Support for v8 crypto and CRC instructions.
1485 * Improvements and robustification of the Solaris port.
1487 * Preliminary support for MacOS 10.12 (Sierra) has been added.
1489 Whilst 3.12.0 continues to support the 32-bit x86 instruction set, we
1490 would prefer users to migrate to 64-bit x86 (a.k.a amd64 or x86_64)
1491 where possible. Valgrind's support for 32-bit x86 has stagnated in
1492 recent years and has fallen far behind that for 64-bit x86
1493 instructions. By contrast 64-bit x86 is well supported, up to and
1496 * ==================== TOOL CHANGES ====================
1500 - Added meta mempool support for describing a custom allocator which:
1501 - Auto-frees all chunks assuming that destroying a pool destroys all
1503 - Uses itself to allocate other memory blocks
1505 - New flag --ignore-range-below-sp to ignore memory accesses below
1506 the stack pointer, if you really have to. The related flag
1507 --workaround-gcc296-bugs=yes is now deprecated. Use
1508 --ignore-range-below-sp=1024-1 as a replacement.
1512 - Improved thread startup time significantly on non-Linux platforms.
1516 - Added collection of the metric "tot-blocks-allocd"
1518 * ==================== OTHER CHANGES ====================
1520 * Replacement/wrapping of malloc/new related functions is now done not just
1521 for system libraries by default, but for any globally defined malloc/new
1522 related function (both in shared libraries and statically linked alternative
1523 malloc implementations). The dynamic (runtime) linker is excluded, though.
1524 To only intercept malloc/new related functions in
1525 system libraries use --soname-synonyms=somalloc=nouserintercepts (where
1526 "nouserintercepts" can be any non-existing library name).
1527 This new functionality is not implemented for MacOS X.
1529 * The maximum number of callers in a suppression entry is now equal to
1530 the maximum size for --num-callers (500).
1531 Note that --gen-suppressions=yes|all similarly generates suppressions
1532 containing up to --num-callers frames.
1534 * New and modified GDB server monitor features:
1536 - Valgrind's gdbserver now accepts the command 'catch syscall'.
1537 Note that you must have GDB >= 7.11 to use 'catch syscall' with
1540 * New option --run-cxx-freeres=<yes|no> can be used to change whether
1541 __gnu_cxx::__freeres() cleanup function is called or not. Default is
1544 * Valgrind is able to read compressed debuginfo sections in two formats:
1545 - zlib ELF gABI format with SHF_COMPRESSED flag (gcc option -gz=zlib)
1546 - zlib GNU format with .zdebug sections (gcc option -gz=zlib-gnu)
1548 * Modest JIT-cost improvements: the cost of instrumenting code blocks
1549 for the most common use case (x86_64-linux, Memcheck) has been
1552 * Improved performance for programs that do a lot of discarding of
1553 instruction address ranges of 8KB or less.
1555 * The C++ symbol demangler has been updated.
1557 * More robustness against invalid syscall parameters on Linux.
1559 * ==================== FIXED BUGS ====================
1561 The following bugs have been fixed or resolved. Note that "n-i-bz"
1562 stands for "not in bugzilla" -- that is, a bug that was reported to us
1563 but never got a bugzilla entry. We encourage you to file bugs in
1564 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1565 than mailing the developers (or mailing lists) directly -- bugs that
1566 are not entered into bugzilla tend to get forgotten about or ignored.
1568 To see details of a given bug, visit
1569 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1570 where XXXXXX is the bug number as listed below.
1572 191069 Exiting due to signal not reported in XML output
1573 199468 Suppressions: stack size limited to 25
1574 while --num-callers allows more frames
1575 212352 vex amd64 unhandled opc_aux = 0x 2, first_opcode == 0xDC (FCOM)
1576 278744 cvtps2pd with redundant RexW
1577 303877 valgrind doesn't support compressed debuginfo sections.
1578 345307 Warning about "still reachable" memory when using libstdc++ from gcc 5
1579 348345 Assertion fails for negative lineno
1580 348924 MIPS: Load doubles through memory so the code compiles with the FPXX ABI
1581 351282 V 3.10.1 MIPS softfloat build broken with GCC 4.9.3 / binutils 2.25.1
1582 351692 Dumps created by valgrind are not readable by gdb (mips32 specific)
1583 351804 Crash on generating suppressions for "printf" call on OS X 10.10
1584 352197 mips: mmap2() not wrapped correctly for page size > 4096
1585 353083 arm64 doesn't implement various xattr system calls
1586 353084 arm64 doesn't support sigpending system call
1587 353137 www: update info for Supported Platforms
1588 353138 www: update "The Valgrind Developers" page
1589 353370 don't advertise RDRAND in cpuid for Core-i7-4910-like avx2 machine
1592 353384 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x62 (pcmpXstrX $0x62)
1593 353398 WARNING: unhandled amd64-solaris syscall: 207
1594 353660 XML in auxwhat tag not escaping reserved symbols properly
1595 353680 s390x: Crash with certain glibc versions due to non-implemented TBEGIN
1596 353727 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x72 (pcmpXstrX $0x72)
1597 353802 ELF debug info reader confused with multiple .rodata sections
1598 353891 Assert 'bad_scanned_addr < VG_ROUNDDN(start+len, sizeof(Addr))' failed
1599 353917 unhandled amd64-solaris syscall fchdir(120)
1600 353920 unhandled amd64-solaris syscall: 170
1601 354274 arm: unhandled instruction: 0xEBAD 0x0AC1 (sub.w sl, sp, r1, lsl #3)
1602 354392 unhandled amd64-solaris syscall: 171
1603 354797 Vbit test does not include Iops for Power 8 instruction support
1604 354883 tst->os_state.pthread - magic_delta assertion failure on OSX 10.11
1608 354933 Fix documentation of --kernel-variant=android-no-hw-tls option
1609 355188 valgrind should intercept all malloc related global functions
1610 355454 do not intercept malloc related symbols from the runtime linker
1611 355455 stderr.exp of test cases wrapmalloc and wrapmallocstatic overconstrained
1612 356044 Dwarf line info reader misinterprets is_stmt register
1613 356112 mips: replace addi with addiu
1614 356393 valgrind (vex) crashes because isZeroU happened
1617 356676 arm64-linux: unhandled syscalls 125, 126 (sched_get_priority_max/min)
1618 356678 arm64-linux: unhandled syscall 232 (mincore)
1619 356817 valgrind.h triggers compiler errors on MSVC when defining NVALGRIND
1620 356823 Unsupported ARM instruction: stlex
1621 357059 x86/amd64: SSE cvtpi2ps with memory source does transition to MMX state
1622 357338 Unhandled instruction for SHA instructions libcrypto Boring SSL
1623 357673 crash if I try to run valgrind with a binary link with libcurl
1624 357833 Setting RLIMIT_DATA to zero breaks with linux 4.5+
1625 357871 pthread_spin_destroy not properly wrapped
1626 357887 Calls to VG_(fclose) do not close the file descriptor
1627 357932 amd64->IR: accept redundant REX prefixes for {minsd,maxsd} m128, xmm.
1628 358030 support direct socket calls on x86 32bit (new in linux 4.3)
1629 358478 drd/tests/std_thread.cpp doesn't build with GCC6
1630 359133 Assertion 'eltSzB <= ddpa->poolSzB' failed
1631 359181 Buffer Overflow during Demangling
1632 359201 futex syscall "skips" argument 5 if op is FUTEX_WAIT_BITSET
1633 359289 s390x: popcnt (B9E1) not implemented
1634 359472 The Power PC vsubuqm instruction doesn't always give the correct result
1635 359503 Add missing syscalls for aarch64 (arm64)
1636 359645 "You need libc6-dbg" help message could be more helpful
1637 359703 s390: wire up separate socketcalls system calls
1638 359724 getsockname might crash - deref_UInt should call safe_to_deref
1639 359733 amd64 implement ld.so strchr/index override like x86
1640 359767 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 1/5
1641 359829 Power PC test suite none/tests/ppc64/test_isa_2_07.c uses
1643 359838 arm64: Unhandled instruction 0xD5033F5F (clrex)
1644 359871 Incorrect mask handling in ppoll
1645 359952 Unrecognised PCMPESTRM variants (0x70, 0x19)
1646 360008 Contents of Power vr registers contents is not printed correctly when
1647 the --vgdb-shadow-registers=yes option is used
1648 360035 POWER PC instruction bcdadd and bcdsubtract generate result with
1649 non-zero shadow bits
1650 360378 arm64: Unhandled instruction 0x5E280844 (sha1h s4, s2)
1651 360425 arm64 unsupported instruction ldpsw
1653 360519 none/tests/arm64/memory.vgtest might fail with newer gcc
1654 360571 Error about the Android Runtime reading below the stack pointer on ARM
1655 360574 Wrong parameter type for an ashmem ioctl() call on Android and ARM64
1656 360749 kludge for multiple .rodata sections on Solaris no longer needed
1657 360752 raise the number of reserved fds in m_main.c from 10 to 12
1658 361207 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 2/5
1659 361226 s390x: risbgn (EC59) not implemented
1660 361253 [s390x] ex_clone.c:42: undefined reference to `pthread_create'
1661 361354 ppc64[le]: wire up separate socketcalls system calls
1662 361615 Inconsistent termination for multithreaded process terminated by signal
1663 361926 Unhandled Solaris syscall: sysfs(84)
1664 362009 V dumps core on unimplemented functionality before threads are created
1665 362329 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 3/5
1666 362894 missing (broken) support for wbit field on mtfsfi instruction (ppc64)
1667 362935 [AsusWRT] Assertion 'sizeof(TTEntryC) <= 88' failed
1668 362953 Request for an update to the Valgrind Developers page
1669 363680 add renameat2() support
1670 363705 arm64 missing syscall name_to_handle_at and open_by_handle_at
1671 363714 ppc64 missing syscalls sync, waitid and name_to/open_by_handle_at
1672 363858 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 4/5
1673 364058 clarify in manual limitations of array overruns detections
1674 364413 pselect sycallwrapper mishandles NULL sigmask
1675 364728 Power PC, missing support for several HW registers in
1676 get_otrack_shadow_offset_wrk()
1677 364948 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 5/5
1678 365273 Invalid write to stack location reported after signal handler runs
1679 365912 ppc64BE segfault during jm-insns test (RELRO)
1680 366079 FPXX Support for MIPS32 Valgrind
1681 366138 Fix configure errors out when using Xcode 8 (clang 8.0.0)
1682 366344 Multiple unhandled instruction for Aarch64
1683 (0x0EE0E020, 0x1AC15800, 0x4E284801, 0x5E040023, 0x5E056060)
1684 367995 Integration of memcheck with custom memory allocator
1685 368120 x86_linux asm _start functions do not keep 16-byte aligned stack pointer
1686 368412 False positive result for altivec capability check
1687 368416 Add tc06_two_races_xml.exp output for ppc64
1688 368419 Perf Events ioctls not implemented
1689 368461 mmapunmap test fails on ppc64
1690 368823 run_a_thread_NORETURN assembly code typo for VGP_arm64_linux target
1691 369000 AMD64 fma4 instructions unsupported.
1692 369169 ppc64 fails jm_int_isa_2_07 test
1693 369175 jm_vec_isa_2_07 test crashes on ppc64
1694 369209 valgrind loops and eats up all memory if cwd doesn't exist.
1695 369356 pre_mem_read_sockaddr syscall wrapper can crash with bad sockaddr
1696 369359 msghdr_foreachfield can crash when handling bad iovec
1697 369360 Bad sigprocmask old or new sets can crash valgrind
1698 369361 vmsplice syscall wrapper crashes on bad iovec
1699 369362 Bad sigaction arguments crash valgrind
1700 369383 x86 sys_modify_ldt wrapper crashes on bad ptr
1701 369402 Bad set/get_thread_area pointer crashes valgrind
1702 369441 bad lvec argument crashes process_vm_readv/writev syscall wrappers
1703 369446 valgrind crashes on unknown fcntl command
1704 369439 S390x: Unhandled insns RISBLG/RISBHG and LDE/LDER
1705 369468 Remove quadratic metapool algorithm using VG_(HT_remove_at_Iter)
1706 370265 ISA 3.0 HW cap stuff needs updating
1707 371128 BCD add and subtract instructions on Power BE in 32-bit mode do not work
1708 372195 Power PC, xxsel instruction is not always recognized
1710 n-i-bz Fix incorrect (or infinite loop) unwind on RHEL7 x86 and amd64
1711 n-i-bz massif --pages-as-heap=yes does not report peak caused by mmap+munmap
1712 n-i-bz false positive leaks due to aspacemgr merging heap & non heap segments
1713 n-i-bz Fix ppoll_alarm exclusion on OS X
1714 n-i-bz Document brk segment limitation, reference manual in limit reached msg.
1715 n-i-bz Fix clobber list in none/tests/amd64/xacq_xrel.c [valgrind r15737]
1716 n-i-bz Bump allowed shift value for "add.w reg, sp, reg, lsl #N" [vex r3206]
1717 n-i-bz amd64: memcheck false positive with shr %edx
1718 n-i-bz arm3: Allow early writeback of SP base register in "strd rD, [sp, #-16]"
1719 n-i-bz ppc: Fix two cases of PPCAvFpOp vs PPCFpOp enum confusion
1720 n-i-bz arm: Fix incorrect register-number constraint check for LDAEX{,B,H,D}
1721 n-i-bz DHAT: added collection of the metric "tot-blocks-allocd"
1723 (3.12.0.RC1: 20 October 2016, vex r3282, valgrind r16094)
1724 (3.12.0.RC2: 20 October 2016, vex r3282, valgrind r16096)
1725 (3.12.0: 21 October 2016, vex r3282, valgrind r16098)
1729 Release 3.11.0 (22 September 2015)
1730 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1732 3.11.0 is a feature release with many improvements and the usual
1733 collection of bug fixes.
1735 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1736 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1737 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1738 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1739 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1740 X86/MacOSX 10.11, AMD64/MacOSX 10.11 and TILEGX/Linux.
1742 * ================== PLATFORM CHANGES =================
1744 * Support for Solaris/x86 and Solaris/amd64 has been added.
1746 * Preliminary support for Mac OS X 10.11 (El Capitan) has been added.
1748 * Preliminary support for the Tilera TileGX architecture has been added.
1750 * s390x: It is now required for the host to have the "long displacement"
1751 facility. The oldest supported machine model is z990.
1753 * x86: on an SSE2 only host, Valgrind in 32 bit mode now claims to be a
1754 Pentium 4. 3.10.1 wrongly claimed to be a Core 2, which is SSSE3.
1756 * The JIT's register allocator is significantly faster, making the JIT
1757 as a whole somewhat faster, so JIT-intensive activities, for example
1758 program startup, are modestly faster, around 5%.
1760 * There have been changes to the default settings of several command
1761 line flags, as detailed below.
1763 * Intel AVX2 support is more complete (64 bit targets only). On AVX2
1764 capable hosts, the simulated CPUID will now indicate AVX2 support.
1766 * ==================== TOOL CHANGES ====================
1770 - The default value for --leak-check-heuristics has been changed from
1771 "none" to "all". This helps to reduce the number of possibly
1772 lost blocks, in particular for C++ applications.
1774 - The default value for --keep-stacktraces has been changed from
1775 "malloc-then-free" to "malloc-and-free". This has a small cost in
1776 memory (one word per malloc-ed block) but allows Memcheck to show the
1777 3 stacktraces of a dangling reference: where the block was allocated,
1778 where it was freed, and where it is acccessed after being freed.
1780 - The default value for --partial-loads-ok has been changed from "no" to
1781 "yes", so as to avoid false positive errors resulting from some kinds
1782 of vectorised loops.
1784 - A new monitor command 'xb <addr> <len>' shows the validity bits of
1785 <len> bytes at <addr>. The monitor command 'xb' is easier to use
1786 than get_vbits when you need to associate byte data value with
1787 their corresponding validity bits.
1789 - The 'block_list' monitor command has been enhanced:
1790 o it can print a range of loss records
1791 o it now accepts an optional argument 'limited <max_blocks>'
1792 to control the number of blocks printed.
1793 o if a block has been found using a heuristic, then
1794 'block_list' now shows the heuristic after the block size.
1795 o the loss records/blocks to print can be limited to the blocks
1796 found via specified heuristics.
1798 - The C helper functions used to instrument loads on
1799 x86-{linux,solaris} and arm-linux (both 32-bit only) have been
1800 replaced by handwritten assembly sequences. This gives speedups
1801 in the region of 0% to 7% for those targets only.
1803 - A new command line option, --expensive-definedness-checks=yes|no,
1804 has been added. This is useful for avoiding occasional invalid
1805 uninitialised-value errors in optimised code. Watch out for
1806 runtime degradation, as this can be up to 25%. As always, though,
1807 the slowdown is highly application specific. The default setting
1812 - A new monitor command 'all_snapshots <filename>' dumps all
1813 snapshots taken so far.
1817 - Significant memory reduction and moderate speedups for
1818 --history-level=full for applications accessing a lot of memory
1819 with many different stacktraces.
1821 - The default value for --conflict-cache-size=N has been doubled to
1822 2000000. Users that were not using the default value should
1823 preferably also double the value they give.
1825 The default was changed due to the changes in the "full history"
1826 implementation. Doubling the value gives on average a slightly more
1827 complete history and uses similar memory (or significantly less memory
1828 in the worst case) than the previous implementation.
1830 - The Helgrind monitor command 'info locks' now accepts an optional
1831 argument 'lock_addr', which shows information about the lock at the
1834 - When using --history-level=full, the new Helgrind monitor command
1835 'accesshistory <addr> [<len>]' will show the recorded accesses for
1836 <len> (or 1) bytes at <addr>.
1838 * ==================== OTHER CHANGES ====================
1840 * The default value for the --smc-check option has been changed from
1841 "stack" to "all-non-file" on targets that provide automatic D-I
1842 cache coherence (x86, amd64 and s390x). The result is to provide,
1843 by default, transparent support for JIT generated and self-modifying
1844 code on all targets.
1846 * Mac OS X only: the default value for the --dsymutil option has been
1847 changed from "no" to "yes", since any serious usage on Mac OS X
1848 always required it to be "yes".
1850 * The command line options --db-attach and --db-command have been removed.
1851 They were deprecated in 3.10.0.
1853 * When a process dies due to a signal, Valgrind now shows the signal
1854 and the stacktrace at default verbosity (i.e. verbosity 1).
1856 * The address description logic used by Memcheck and Helgrind now
1857 describes addresses in anonymous segments, file mmap-ed segments,
1858 shared memory segments and the brk data segment.
1860 * The new option --error-markers=<begin>,<end> can be used to mark the
1861 begin/end of errors in textual output mode, to facilitate
1862 searching/extracting errors in output files that mix valgrind errors
1863 with program output.
1865 * The new option --max-threads=<number> can be used to change the number
1866 of threads valgrind can handle. The default is 500 threads which
1867 should be more than enough for most applications.
1869 * The new option --valgrind-stacksize=<number> can be used to change the
1870 size of the private thread stacks used by Valgrind. This is useful
1871 for reducing memory use or increasing the stack size if Valgrind
1872 segfaults due to stack overflow.
1874 * The new option --avg-transtab-entry-size=<number> can be used to specify
1875 the expected instrumented block size, either to reduce memory use or
1876 to avoid excessive retranslation.
1878 * Valgrind can be built with Intel's ICC compiler, version 14.0 or later.
1880 * New and modified GDB server monitor features:
1882 - When a signal is reported in GDB, you can now use the GDB convenience
1883 variable $_siginfo to examine detailed signal information.
1885 - Valgrind's gdbserver now allows the user to change the signal
1886 to deliver to the process. So, use 'signal SIGNAL' to continue execution
1887 with SIGNAL instead of the signal reported to GDB. Use 'signal 0' to
1888 continue without passing the signal to the process.
1890 - With GDB >= 7.10, the command 'target remote'
1891 will automatically load the executable file of the process running
1892 under Valgrind. This means you do not need to specify the executable
1893 file yourself, GDB will discover it itself. See GDB documentation about
1894 'qXfer:exec-file:read' packet for more info.
1896 * ==================== FIXED BUGS ====================
1898 The following bugs have been fixed or resolved. Note that "n-i-bz"
1899 stands for "not in bugzilla" -- that is, a bug that was reported to us
1900 but never got a bugzilla entry. We encourage you to file bugs in
1901 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1902 than mailing the developers (or mailing lists) directly -- bugs that
1903 are not entered into bugzilla tend to get forgotten about or ignored.
1905 To see details of a given bug, visit
1906 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1907 where XXXXXX is the bug number as listed below.
1909 116002 VG_(printf): Problems with justification of strings and integers
1910 155125 avoid cutting away file:lineno after long function name
1911 197259 Unsupported arch_prtctl PR_SET_GS option
1912 201152 ppc64: Assertion in ppc32g_dirtyhelper_MFSPR_268_269
1913 201216 Fix Valgrind does not support pthread_sigmask() on OS X
1914 201435 Fix Darwin: -v does not show kernel version
1915 208217 "Warning: noted but unhandled ioctl 0x2000747b" on Mac OS X
1916 211256 Fixed an outdated comment regarding the default platform.
1917 211529 Incomplete call stacks for code compiled by newer versions of MSVC
1918 211926 Avoid compilation warnings in valgrind.h with -pedantic
1919 212291 Fix unhandled syscall: unix:132 (mkfifo) on OS X
1921 226609 Crediting upstream authors in man page
1922 231257 Valgrind omits path when executing script from shebang line
1923 254164 OS X task_info: UNKNOWN task message [id 3405, to mach_task_self() [..]
1924 294065 Improve the pdb file reader by avoiding hardwired absolute pathnames
1925 269360 s390x: Fix addressing mode selection for compare-and-swap
1926 302630 Memcheck: Assertion failed: 'sizeof(UWord) == sizeof(UInt)'
1928 312989 ioctl handling needs to do POST handling on generic ioctls and [..]
1929 319274 Fix unhandled syscall: unix:410 (sigsuspend_nocancel) on OS X
1930 324181 mmap does not handle MAP_32BIT (handle it now, rather than fail it)
1931 327745 Fix valgrind 3.9.0 build fails on Mac OS X 10.6.8
1932 330147 libmpiwrap PMPI_Get_count returns undefined value
1933 333051 mmap of huge pages fails due to incorrect alignment
1935 334802 valgrind does not always explain why a given option is bad
1936 335618 mov.w rN, pc/sp (ARM32)
1937 335785 amd64->IR 0xC4 0xE2 0x75 0x2F (vmaskmovpd)
1942 335907 segfault when running wine's ddrawex/tests/surface.c under valgrind
1943 338602 AVX2 bit in CPUID missing
1944 338606 Strange message for scripts with invalid interpreter
1945 338731 ppc: Fix testuite build for toolchains not supporting -maltivec
1946 338995 shmat with hugepages (SHM_HUGETLB) fails with EINVAL
1947 339045 Getting valgrind to compile and run on OS X Yosemite (10.10)
1949 339156 gdbsrv not called for fatal signal
1950 339215 Valgrind 3.10.0 contain 2013 in copyrights notice
1951 339288 support Cavium Octeon MIPS specific BBIT*32 instructions
1952 339636 Use fxsave64 and fxrstor64 mnemonics instead of old-school rex64 prefix
1953 339442 Fix testsuite build failure on OS X 10.9
1954 339542 Enable compilation with Intel's ICC compiler
1955 339563 The DVB demux DMX_STOP ioctl doesn't have a wrapper
1956 339688 Mac-specific ASM does not support .version directive (cpuid,
1957 tronical and pushfpopf tests)
1958 339745 Valgrind crash when check Marmalade app (partial fix)
1959 339755 Fix known deliberate memory leak in setenv() on Mac OS X 10.9
1960 339778 Linux/TileGx platform support for Valgrind
1961 339780 Fix known uninitialised read in pthread_rwlock_init() on Mac OS X 10.9
1962 339789 Fix none/tests/execve test on Mac OS X 10.9
1963 339808 Fix none/tests/rlimit64_nofile test on Mac OS X 10.9
1964 339820 vex amd64->IR: 0x66 0xF 0x3A 0x63 0xA 0x42 0x74 0x9 (pcmpistri $0x42)
1965 340115 Fix none/tests/cmdline[1|2] tests on systems which define TMPDIR
1966 340392 Allow user to select more accurate definedness checking in memcheck
1967 to avoid invalid complaints on optimised code
1968 340430 Fix some grammatical weirdness in the manual.
1969 341238 Recognize GCC5/DWARFv5 DW_LANG constants (Go, C11, C++11, C++14)
1970 341419 Signal handler ucontext_t not filled out correctly on OS X
1971 341539 VG_(describe_addr) should not describe address as belonging to client
1972 segment if it is past the heap end
1973 341613 Enable building of manythreads and thread-exits tests on Mac OS X
1974 341615 Fix none/tests/darwin/access_extended test on Mac OS X
1975 341698 Valgrind's AESKEYGENASSIST gives wrong result in words 0 and 2 [..]
1976 341789 aarch64: shmat fails with valgrind on ARMv8
1977 341997 MIPS64: Cavium OCTEON insns - immediate operand handled incorrectly
1978 342008 valgrind.h needs type cast [..] for clang/llvm in 64-bit mode
1979 342038 Unhandled syscalls on aarch64 (mbind/get/set_mempolicy)
1980 342063 wrong format specifier for test mcblocklistsearch in gdbserver_tests
1981 342117 Hang when loading PDB file for MSVC compiled Firefox under Wine
1982 342221 socket connect false positive uninit memory for unknown af family
1983 342353 Allow dumping full massif output while valgrind is still running
1984 342571 Valgrind chokes on AVX compare intrinsic with _CMP_GE_QS
1988 342603 Add I2C_SMBUS ioctl support
1989 342635 OS X 10.10 (Yosemite) - missing system calls and fcntl code
1990 342683 Mark memory past the initial brk limit as unaddressable
1991 342783 arm: unhandled instruction 0xEEFE1ACA = "vcvt.s32.f32 s3, s3, #12"
1992 342795 Internal glibc __GI_mempcpy call should be intercepted
1993 342841 s390x: Support instructions fiebr(a) and fidbr(a)
1994 343012 Unhandled syscall 319 (memfd_create)
1995 343069 Patch updating v4l2 API support
1996 343173 helgrind crash during stack unwind
1997 343219 fix GET_STARTREGS for arm
1998 343303 Fix known deliberate memory leak in setenv() on Mac OS X 10.10
1999 343306 OS X 10.10: UNKNOWN mach_msg unhandled MACH_SEND_TRAILER option
2000 343332 Unhandled instruction 0x9E310021 (fcvtmu) on aarch64
2001 343335 unhandled instruction 0x1E638400 (fccmp) aarch64
2002 343523 OS X mach_ports_register: UNKNOWN task message [id 3403, to [..]
2003 343525 OS X host_get_special_port: UNKNOWN host message [id 412, to [..]
2004 343597 ppc64le: incorrect use of offseof macro
2005 343649 OS X host_create_mach_voucher: UNKNOWN host message [id 222, to [..]
2006 343663 OS X 10.10 Memchecj always reports a leak regardless of [..]
2007 343732 Unhandled syscall 144 (setgid) on aarch64
2008 343733 Unhandled syscall 187 (msgctl and related) on aarch64
2009 343802 s390x: False positive "conditional jump or move depends on [..]
2010 343902 --vgdb=yes doesn't break when --xml=yes is used
2011 343967 Don't warn about setuid/setgid/setcap executable for directories
2012 343978 Recognize DWARF5/GCC5 DW_LANG_Fortran 2003 and 2008 constants
2013 344007 accept4 syscall unhandled on arm64 (242) and ppc64 (344)
2014 344033 Helgrind on ARM32 loses track of mutex state in pthread_cond_wait
2015 344054 www - update info for Solaris/illumos
2016 344416 'make regtest' does not work cleanly on OS X
2017 344235 Remove duplicate include of pub_core_aspacemgr.h
2018 344279 syscall sendmmsg on arm64 (269) and ppc32/64 (349) unhandled
2019 344295 syscall recvmmsg on arm64 (243) and ppc32/64 (343) unhandled
2020 344307 2 unhandled syscalls on aarch64/arm64: umount2(39), mount (40)
2021 344314 callgrind_annotate ... warnings about commands containing newlines
2022 344318 socketcall should wrap recvmmsg and sendmmsg
2023 344337 Fix unhandled syscall: mach:41 (_kernelrpc_mach_port_guard_trap)
2024 344416 Fix 'make regtest' does not work cleanly on OS X
2025 344499 Fix compilation for Linux kernel >= 4.0.0
2026 344512 OS X: unhandled syscall: unix:348 (__pthread_chdir),
2027 unix:349 (__pthread_fchdir)
2028 344559 Garbage collection of unused segment names in address space manager
2029 344560 Fix stack traces missing penultimate frame on OS X
2030 344621 Fix memcheck/tests/err_disable4 test on OS X
2031 344686 Fix suppression for pthread_rwlock_init on OS X 10.10
2032 344702 Fix missing libobjc suppressions on OS X 10.10
2034 344936 Fix unhandled syscall: unix:473 (readlinkat) on OS X 10.10
2035 344939 Fix memcheck/tests/xml1 on OS X 10.10
2036 345016 helgrind/tests/locked_vs_unlocked2 is failing sometimes
2037 345079 Fix build problems in VEX/useful/test_main.c
2038 345126 Incorrect handling of VIDIOC_G_AUDIO and G_AUDOUT
2039 345177 arm64: prfm (reg) not implemented
2040 345215 Performance improvements for the register allocator
2041 345248 add support for Solaris OS in valgrind
2042 345338 TIOCGSERIAL and TIOCSSERIAL ioctl support on Linux
2043 345394 Fix memcheck/tests/strchr on OS X
2044 345637 Fix memcheck/tests/sendmsg on OS X
2045 345695 Add POWERPC support for AT_DCACHESIZE and HWCAP2
2046 345824 Fix aspacem segment mismatch: seen with none/tests/bigcode
2047 345887 Fix an assertion in the address space manager
2048 345928 amd64: callstack only contains current function for small stacks
2049 345984 disInstr(arm): unhandled instruction: 0xEE193F1E
2050 345987 MIPS64: Implement cavium LHX instruction
2051 346031 MIPS: Implement support for the CvmCount register (rhwr %0, 31)
2052 346185 Fix typo saving altivec register v24
2053 346267 Compiler warnings for PPC64 code on call to LibVEX_GuestPPC64_get_XER()
2054 and LibVEX_GuestPPC64_get_CR()
2055 346270 Regression tests none/tests/jm_vec/isa_2_07 and
2056 none/tests/test_isa_2_07_part2 have failures on PPC64 little endian
2057 346307 fuse filesystem syscall deadlocks
2058 346324 PPC64 missing support for lbarx, lharx, stbcx and sthcx instructions
2059 346411 MIPS: SysRes::_valEx handling is incorrect
2060 346416 Add support for LL_IOC_PATH2FID and LL_IOC_GETPARENT Lustre ioctls
2061 346474 PPC64 Power 8, spr TEXASRU register not supported
2062 346487 Compiler generates "note" about a future ABI change for PPC64
2063 346562 MIPS64: lwl/lwr instructions are performing 64bit loads
2064 and causing spurious "invalid read of size 8" warnings
2065 346801 Fix link error on OS X: _vgModuleLocal_sf_maybe_extend_stack
2066 347151 Fix suppression for pthread_rwlock_init on OS X 10.8
2067 347233 Fix memcheck/tests/strchr on OS X 10.10 (Haswell)
2068 347322 Power PC regression test cleanup
2069 347379 valgrind --leak-check=full leak errors from system libs on OS X 10.8
2071 347389 unhandled syscall: 373 (Linux ARM syncfs)
2072 347686 Patch set to cleanup PPC64 regtests
2073 347978 Remove bash dependencies where not needed
2074 347982 OS X: undefined symbols for architecture x86_64: "_global" [..]
2075 347988 Memcheck: the 'impossible' happened: unexpected size for Addr (OSX/wine)
2077 348102 Patch updating v4l2 API support
2078 348247 amd64 front end: jno jumps wrongly when overflow is not set
2079 348269 Improve mmap MAP_HUGETLB support.
2080 348334 (ppc) valgrind does not simulate dcbfl - then my program terminates
2081 348345 Assertion fails for negative lineno
2082 348377 Unsupported ARM instruction: yield
2083 348565 Fix detection of command line option availability for clang
2084 348574 vex amd64->IR pcmpistri SSE4.2 unsupported (pcmpistri $0x18)
2085 348728 Fix broken check for VIDIOC_G_ENC_INDEX
2086 348748 Fix redundant condition
2087 348890 Fix clang warning about unsupported --param inline-unit-growth=900
2088 348949 Bogus "ERROR: --ignore-ranges: suspiciously large range"
2089 349034 Add Lustre ioctls LL_IOC_GROUP_LOCK and LL_IOC_GROUP_UNLOCK
2090 349086 Fix UNKNOWN task message [id 3406, to mach_task_self(), [..]
2091 349087 Fix UNKNOWN task message [id 3410, to mach_task_self(), [..]
2092 349626 Implemented additional Xen hypercalls
2093 349769 Clang/osx: ld: warning: -read_only_relocs cannot be used with x86_64
2094 349790 Clean up of the hardware capability checking utilities.
2095 349828 memcpy intercepts memmove causing src/dst overlap error (ppc64 ld.so)
2096 349874 Fix typos in source code
2097 349879 memcheck: add handwritten assembly for helperc_LOADV*
2098 349941 di_notify_mmap might create wrong start/size DebugInfoMapping
2099 350062 vex x86->IR: 0x66 0xF 0x3A 0xB (ROUNDSD) on OS X
2100 350202 Add limited param to 'monitor block_list'
2101 350290 s390x: Support instructions fixbr(a)
2102 350359 memcheck/tests/x86/fxsave hangs indefinetely on OS X
2103 350809 Fix none/tests/async-sigs for Solaris
2104 350811 Remove reference to --db-attach which has been removed.
2105 350813 Memcheck/x86: enable handwritten assembly helpers for x86/Solaris too
2106 350854 hard-to-understand code in VG_(load_ELF)()
2107 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
2108 351386 Solaris: Cannot run ld.so.1 under Valgrind
2109 351474 Fix VG_(iseqsigset) as obvious
2110 351531 Typo in /include/vki/vki-xen-physdev.h header guard
2111 351756 Intercept platform_memchr$VARIANT$Haswell on OS X
2112 351858 ldsoexec support on Solaris
2113 351873 Newer gcc doesn't allow __builtin_tabortdc[i] in ppc32 mode
2114 352130 helgrind reports false races for printfs using mempcpy on FILE* state
2115 352284 s390: Conditional jump depends on uninitialised value(s) in vfprintf
2116 352320 arm64 crash on none/tests/nestedfs
2117 352765 Vbit test fails on Power 6
2118 352768 The mbar instruction is missing from the Power PC support
2119 352769 Power PC program priority register (PPR) is not supported
2120 n-i-bz Provide implementations of certain compiler builtins to support
2121 compilers that may not provide those
2122 n-i-bz Old STABS code is still being compiled, but never used. Remove it.
2123 n-i-bz Fix compilation on distros with glibc < 2.5
2124 n-i-bz (vex 3098) Avoid generation of Neon insns on non-Neon hosts
2125 n-i-bz Enable rt_sigpending syscall on ppc64 linux.
2126 n-i-bz mremap did not work properly on shared memory
2127 n-i-bz Fix incorrect sizeof expression in syswrap-xen.c reported by Coverity
2128 n-i-bz In VALGRIND_PRINTF write out thread name, if any, to xml
2130 (3.11.0.TEST1: 8 September 2015, vex r3187, valgrind r15646)
2131 (3.11.0.TEST2: 21 September 2015, vex r3193, valgrind r15667)
2132 (3.11.0: 22 September 2015, vex r3195, valgrind r15674)
2136 Release 3.10.1 (25 November 2014)
2137 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2138 3.10.1 is a bug fix release. It fixes various bugs reported in 3.10.0
2139 and backports fixes for all reported missing AArch64 ARMv8 instructions
2140 and syscalls from the trunk. If you package or deliver 3.10.0 for others
2141 to use, you might want to consider upgrading to 3.10.1 instead.
2143 The following bugs have been fixed or resolved. Note that "n-i-bz"
2144 stands for "not in bugzilla" -- that is, a bug that was reported to us
2145 but never got a bugzilla entry. We encourage you to file bugs in
2146 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2147 than mailing the developers (or mailing lists) directly -- bugs that
2148 are not entered into bugzilla tend to get forgotten about or ignored.
2150 To see details of a given bug, visit
2151 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2152 where XXXXXX is the bug number as listed below.
2154 335440 arm64: ld1 (single structure) is not implemented
2155 335713 arm64: unhanded instruction: prfm (immediate)
2156 339020 ppc64: memcheck/tests/ppc64/power_ISA2_05 failing in nightly build
2157 339182 ppc64: AvSplat ought to load destination vector register with [..]
2158 339336 PPC64 store quad instruction (stq) is not supposed to change [..]
2159 339433 ppc64 lxvw4x instruction uses four 32-byte loads
2160 339645 Use correct tag names in sys_getdents/64 wrappers
2161 339706 Fix false positive for ioctl(TIOCSIG) on linux
2162 339721 assertion 'check_sibling == sibling' failed in readdwarf3.c ...
2163 339853 arm64 times syscall unknown
2164 339855 arm64 unhandled getsid/setsid syscalls
2165 339858 arm64 dmb sy not implemented
2166 339926 Unhandled instruction 0x1E674001 (frintx) on aarm64
2167 339927 Unhandled instruction 0x9E7100C6 (fcvtmu) on aarch64
2168 339938 disInstr(arm64): unhandled instruction 0x4F8010A4 (fmla)
2170 339940 arm64: unhandled syscall: 83 (sys_fdatasync) + patch
2171 340033 arm64: unhandled insn dmb ishld and some other isb-dmb-dsb variants
2172 340028 unhandled syscalls for arm64 (msync, pread64, setreuid and setregid)
2173 340036 arm64: Unhandled instruction ld4 (multiple structures, no offset)
2174 340236 arm64: unhandled syscalls: mknodat, fchdir, chroot, fchownat
2175 340509 arm64: unhandled instruction fcvtas
2176 340630 arm64: fchmod (52) and fchown (55) syscalls not recognized
2177 340632 arm64: unhandled instruction fcvtas
2178 340722 Resolve "UNKNOWN attrlist flags 0:0x10000000"
2179 340725 AVX2: Incorrect decoding of vpbroadcast{b,w} reg,reg forms
2180 340788 warning: unhandled syscall: 318 (getrandom)
2181 340807 disInstr(arm): unhandled instruction: 0xEE989B20
2182 340856 disInstr(arm64): unhandled instruction 0x1E634C45 (fcsel)
2183 340922 arm64: unhandled getgroups/setgroups syscalls
2184 350251 Fix typo in VEX utility program (test_main.c).
2185 350407 arm64: unhandled instruction ucvtf (vector, integer)
2186 350809 none/tests/async-sigs breaks when run under cron on Solaris
2187 350811 update README.solaris after r15445
2188 350813 Use handwritten memcheck assembly helpers on x86/Solaris [..]
2189 350854 strange code in VG_(load_ELF)()
2190 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
2191 n-i-bz DRD and Helgrind: Handle Imbe_CancelReservation (clrex on ARM)
2192 n-i-bz Add missing ]] to terminate CDATA.
2193 n-i-bz Glibc versions prior to 2.5 do not define PTRACE_GETSIGINFO
2194 n-i-bz Enable sys_fadvise64_64 on arm32.
2195 n-i-bz Add test cases for all remaining AArch64 SIMD, FP and memory insns.
2196 n-i-bz Add test cases for all known arm64 load/store instructions.
2197 n-i-bz PRE(sys_openat): when checking whether ARG1 == VKI_AT_FDCWD [..]
2198 n-i-bz Add detection of old ppc32 magic instructions from bug 278808.
2199 n-i-bz exp-dhat: Implement missing function "dh_malloc_usable_size".
2200 n-i-bz arm64: Implement "fcvtpu w, s".
2201 n-i-bz arm64: implement ADDP and various others
2202 n-i-bz arm64: Implement {S,U}CVTF (scalar, fixedpt).
2203 n-i-bz arm64: enable FCVT{A,N}S X,S.
2205 (3.10.1: 25 November 2014, vex r3026, valgrind r14785)
2209 Release 3.10.0 (10 September 2014)
2210 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2212 3.10.0 is a feature release with many improvements and the usual
2213 collection of bug fixes.
2215 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
2216 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
2217 MIPS64/Linux, ARM/Android, MIPS32/Android, X86/Android, X86/MacOSX 10.9
2218 and AMD64/MacOSX 10.9. Support for MacOSX 10.8 and 10.9 is
2219 significantly improved relative to the 3.9.0 release.
2221 * ================== PLATFORM CHANGES =================
2223 * Support for the 64-bit ARM Architecture (AArch64 ARMv8). This port
2224 is mostly complete, and is usable, but some SIMD instructions are as
2227 * Support for little-endian variant of the 64-bit POWER architecture.
2229 * Support for Android on MIPS32.
2231 * Support for 64bit FPU on MIPS32 platforms.
2233 * Both 32- and 64-bit executables are supported on MacOSX 10.8 and 10.9.
2235 * Configuration for and running on Android targets has changed.
2236 See README.android in the source tree for details.
2238 * ================== DEPRECATED FEATURES =================
2240 * --db-attach is now deprecated and will be removed in the next
2241 valgrind feature release. The built-in GDB server capabilities are
2242 superior and should be used instead. Learn more here:
2243 http://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.gdbserver
2245 * ==================== TOOL CHANGES ====================
2249 - Client code can now selectively disable and re-enable reporting of
2250 invalid address errors in specific ranges using the new client
2251 requests VALGRIND_DISABLE_ADDR_ERROR_REPORTING_IN_RANGE and
2252 VALGRIND_ENABLE_ADDR_ERROR_REPORTING_IN_RANGE.
2254 - Leak checker: there is a new leak check heuristic called
2255 "length64". This is used to detect interior pointers pointing 8
2256 bytes inside a block, on the assumption that the first 8 bytes
2257 holds the value "block size - 8". This is used by
2258 sqlite3MemMalloc, for example.
2260 - Checking of system call parameters: if a syscall parameter
2261 (e.g. bind struct sockaddr, sendmsg struct msghdr, ...) has
2262 several fields not initialised, an error is now reported for each
2263 field. Previously, an error was reported only for the first
2264 uninitialised field.
2266 - Mismatched alloc/free checking: a new flag
2267 --show-mismatched-frees=no|yes [yes] makes it possible to turn off
2268 such checks if necessary.
2272 - Improvements to error messages:
2274 o Race condition error message involving heap allocated blocks also
2275 show the thread number that allocated the raced-on block.
2277 o All locks referenced by an error message are now announced.
2278 Previously, some error messages only showed the lock addresses.
2280 o The message indicating where a lock was first observed now also
2281 describes the address/location of the lock.
2283 - Helgrind now understands the Ada task termination rules and
2284 creates a happens-before relationship between a terminated task
2285 and its master. This avoids some false positives and avoids a big
2286 memory leak when a lot of Ada tasks are created and terminated.
2287 The interceptions are only activated with forthcoming releases of
2288 gnatpro >= 7.3.0w-20140611 and gcc >= 5.0.
2290 - A new GDB server monitor command "info locks" giving the list of
2291 locks, their location, and their status.
2295 - callgrind_control now supports the --vgdb-prefix argument,
2296 which is needed if valgrind was started with this same argument.
2298 * ==================== OTHER CHANGES ====================
2300 * Unwinding through inlined function calls. Stack unwinding can now
2301 make use of Dwarf3 inlined-unwind information if it is available.
2302 The practical effect is that inlined calls become visible in stack
2303 traces. The suppression matching machinery has been adjusted
2304 accordingly. This is controlled by the new option
2305 --read-inline-info=yes|no. Currently this is enabled by default
2306 only on Linux and Android targets and only for the tools Memcheck,
2309 * Valgrind can now read EXIDX unwind information on 32-bit ARM
2310 targets. If an object contains both CFI and EXIDX unwind
2311 information, Valgrind will prefer the CFI over the EXIDX. This
2312 facilitates unwinding through system libraries on arm-android
2315 * Address description logic has been improved and is now common
2316 between Memcheck and Helgrind, resulting in better address
2317 descriptions for some kinds of error messages.
2319 * Error messages about dubious arguments (eg, to malloc or calloc) are
2320 output like other errors. This means that they can be suppressed
2321 and they have a stack trace.
2323 * The C++ demangler has been updated for better C++11 support.
2325 * New and modified GDB server monitor features:
2327 - Thread local variables/storage (__thread) can now be displayed.
2329 - The GDB server monitor command "v.info location <address>"
2330 displays information about an address. The information produced
2331 depends on the tool and on the options given to valgrind.
2332 Possibly, the following are described: global variables, local
2333 (stack) variables, allocated or freed blocks, ...
2335 - The option "--vgdb-stop-at=event1,event2,..." allows the user to
2336 ask the GDB server to stop at the start of program execution, at
2337 the end of the program execution and on Valgrind internal errors.
2339 - A new monitor command "v.info stats" shows various Valgrind core
2340 and tool statistics.
2342 - A new monitor command "v.set hostvisibility" allows the GDB server
2343 to provide access to Valgrind internal host status/memory.
2345 * A new option "--aspace-minaddr=<address>" can in some situations
2346 allow the use of more memory by decreasing the address above which
2347 Valgrind maps memory. It can also be used to solve address
2348 conflicts with system libraries by increasing the default value.
2349 See user manual for details.
2351 * The amount of memory used by Valgrind to store debug info (unwind
2352 info, line number information and symbol data) has been
2353 significantly reduced, even though Valgrind now reads more
2354 information in order to support unwinding of inlined function calls.
2356 * Dwarf3 handling with --read-var-info=yes has been improved:
2358 - Ada and C struct containing VLAs no longer cause a "bad DIE" error
2360 - Code compiled with
2361 -ffunction-sections -fdata-sections -Wl,--gc-sections
2362 no longer causes assertion failures.
2364 * Improved checking for the --sim-hints= and --kernel-variant=
2365 options. Unknown strings are now detected and reported to the user
2368 * The semantics of stack start/end boundaries in the valgrind.h
2369 VALGRIND_STACK_REGISTER client request has been clarified and
2370 documented. The convention is that start and end are respectively
2371 the lowest and highest addressable bytes of the stack.
2373 * ==================== FIXED BUGS ====================
2375 The following bugs have been fixed or resolved. Note that "n-i-bz"
2376 stands for "not in bugzilla" -- that is, a bug that was reported to us
2377 but never got a bugzilla entry. We encourage you to file bugs in
2378 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2379 than mailing the developers (or mailing lists) directly -- bugs that
2380 are not entered into bugzilla tend to get forgotten about or ignored.
2382 To see details of a given bug, visit
2383 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2384 where XXXXXX is the bug number as listed below.
2386 175819 Support for ipv6 socket reporting with --track-fds
2387 232510 make distcheck fails
2388 249435 Analyzing wine programs with callgrind triggers a crash
2389 278972 support for inlined function calls in stacktraces and suppression
2391 291310 FXSAVE instruction marks memory as undefined on amd64
2392 303536 ioctl for SIOCETHTOOL (ethtool(8)) isn't wrapped
2393 308729 vex x86->IR: unhandled instruction bytes 0xf 0x5 (syscall)
2394 315199 vgcore file for threaded app does not show which thread crashed
2395 315952 tun/tap ioctls are not supported
2396 323178 Unhandled instruction: PLDW register (ARM)
2397 323179 Unhandled instruction: PLDW immediate (ARM)
2398 324050 Helgrind: SEGV because of unaligned stack when using movdqa
2399 325110 Add test-cases for Power ISA 2.06 insns: divdo/divdo. and divduo/divduo.
2400 325124 [MIPSEL] Compilation error
2401 325477 Phase 4 support for IBM Power ISA 2.07
2402 325538 cavium octeon mips64, valgrind reported "dumping core" [...]
2403 325628 Phase 5 support for IBM Power ISA 2.07
2404 325714 Empty vgcore but RLIMIT_CORE is big enough (too big)
2405 325751 Missing the two privileged Power PC Transactional Memory Instructions
2406 325816 Phase 6 support for IBM Power ISA 2.07
2407 325856 Make SGCheck fail gracefully on unsupported platforms
2408 326026 Iop names for count leading zeros/sign bits incorrectly imply [..]
2409 326436 DRD: False positive in libstdc++ std::list::push_back
2410 326444 Cavium MIPS Octeon Specific Load Indexed Instructions
2411 326462 Refactor vgdb to isolate invoker stuff into separate module
2412 326469 amd64->IR: 0x66 0xF 0x3A 0x63 0xC1 0xE (pcmpistri 0x0E)
2413 326623 DRD: false positive conflict report in a field assignment
2414 326724 Valgrind does not compile on OSX 1.9 Mavericks
2415 326816 Intercept for __strncpy_sse2_unaligned missing?
2416 326921 coregrind fails to compile m_trampoline.S with MIPS/Linux port of V
2417 326983 Clear direction flag after tests on amd64.
2418 327212 Do not prepend the current directory to absolute path names.
2419 327223 Support for Cavium MIPS Octeon Atomic and Count Instructions
2420 327238 Callgrind Assertion 'passed <= last_bb->cjmp_count' failed
2421 327284 s390x: Fix translation of the risbg instruction
2422 327639 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x34
2423 327837 dwz compressed alternate .debug_info and .debug_str not read correctly
2424 327916 DW_TAG_typedef may have no name
2425 327943 s390x: add a redirection for the 'index' function
2426 328100 XABORT not implemented
2427 328205 Implement additional Xen hypercalls
2428 328454 add support Backtraces with ARM unwind tables (EXIDX)
2429 328455 s390x: SIGILL after emitting wrong register pair for ldxbr
2430 328711 valgrind.1 manpage "memcheck options" section is badly generated
2431 328878 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x14
2432 329612 Incorrect handling of AT_BASE for image execution
2433 329694 clang warns about using uninitialized variable
2434 329956 valgrind crashes when lmw/stmw instructions are used on ppc64
2435 330228 mmap must align to VKI_SHMLBA on mips32
2436 330257 LLVM does not support `-mno-dynamic-no-pic` option
2437 330319 amd64->IR: unhandled instruction bytes: 0xF 0x1 0xD5 (xend)
2438 330459 --track-fds=yes doesn't track eventfds
2439 330469 Add clock_adjtime syscall support
2440 330594 Missing sysalls on PowerPC / uClibc
2441 330622 Add test to regression suite for POWER instruction: dcbzl
2442 330939 Support for AMD's syscall instruction on x86
2444 330941 Typo in PRE(poll) syscall wrapper
2445 331057 unhandled instruction: 0xEEE01B20 (vfma.f64) (has patch)
2446 331254 Fix expected output for memcheck/tests/dw4
2447 331255 Fix race condition in test none/tests/coolo_sigaction
2448 331257 Fix type of jump buffer in test none/tests/faultstatus
2449 331305 configure uses bash specific syntax
2450 331337 s390x WARNING: unhandled syscall: 326 (dup3)
2451 331380 Syscall param timer_create(evp) points to uninitialised byte(s)
2452 331476 Patch to handle ioctl 0x5422 on Linux (x86 and amd64)
2453 331829 Unexpected ioctl opcode sign extension
2454 331830 ppc64: WARNING: unhandled syscall: 96/97
2455 331839 drd/tests/sem_open specifies invalid semaphore name
2456 331847 outcome of drd/tests/thread_name is nondeterministic
2457 332037 Valgrind cannot handle Thumb "add pc, reg"
2458 332055 drd asserts on platforms with VG_STACK_REDZONE_SZB == 0 and
2459 consistency checks enabled
2460 332263 intercepts for pthread_rwlock_timedrdlock and
2461 pthread_rwlock_timedwrlock are incorrect
2462 332265 drd could do with post-rwlock_init and pre-rwlock_destroy
2464 332276 Implement additional Xen hypercalls
2465 332658 ldrd.w r1, r2, [PC, #imm] does not adjust for 32bit alignment
2466 332765 Fix ms_print to create temporary files in a proper directory
2467 333072 drd: Add semaphore annotations
2468 333145 Tests for missaligned PC+#imm access for arm
2469 333228 AAarch64 Missing instruction encoding: mrs %[reg], ctr_el0
2470 333230 AAarch64 missing instruction encodings: dc, ic, dsb.
2471 333248 WARNING: unhandled syscall: unix:443
2472 333428 ldr.w pc [rD, #imm] instruction leads to assertion
2473 333501 cachegrind: assertion: Cache set count is not a power of two.
2476 333666 Recognize MPX instructions and bnd prefix.
2477 333788 Valgrind does not support the CDROM_DISC_STATUS ioctl (has patch)
2478 333817 Valgrind reports the memory areas written to by the SG_IO
2480 334049 lzcnt fails silently (x86_32)
2481 334384 Valgrind does not have support Little Endian support for
2483 334585 recvmmsg unhandled (+patch) (arm)
2484 334705 sendmsg and recvmsg should guard against bogus msghdr fields.
2485 334727 Build fails with -Werror=format-security
2486 334788 clarify doc about --log-file initial program directory
2487 334834 PPC64 Little Endian support, patch 2
2488 334836 PPC64 Little Endian support, patch 3 testcase fixes
2489 334936 patch to fix false positives on alsa SNDRV_CTL_* ioctls
2490 335034 Unhandled ioctl: HCIGETDEVLIST
2491 335155 vgdb, fix error print statement.
2492 335262 arm64: movi 8bit version is not supported
2493 335263 arm64: dmb instruction is not implemented
2494 335441 unhandled ioctl 0x8905 (SIOCATMARK) when running wine under valgrind
2495 335496 arm64: sbc/abc instructions are not implemented
2496 335554 arm64: unhandled instruction: abs
2497 335564 arm64: unhandled instruction: fcvtpu Xn, Sn
2498 335735 arm64: unhandled instruction: cnt
2499 335736 arm64: unhandled instruction: uaddlv
2500 335848 arm64: unhandled instruction: {s,u}cvtf
2501 335902 arm64: unhandled instruction: sli
2502 335903 arm64: unhandled instruction: umull (vector)
2503 336055 arm64: unhandled instruction: mov (element)
2504 336062 arm64: unhandled instruction: shrn{,2}
2505 336139 mip64: [...] valgrind hangs and spins on a single core [...]
2506 336189 arm64: unhandled Instruction: mvn
2507 336435 Valgrind hangs in pthread_spin_lock consuming 100% CPU
2508 336619 valgrind --read-var-info=yes doesn't handle DW_TAG_restrict_type
2509 336772 Make moans about unknown ioctls more informative
2510 336957 Add a section about the Solaris/illumos port on the webpage
2511 337094 ifunc wrapper is broken on ppc64
2512 337285 fcntl commands F_OFD_SETLK, F_OFD_SETLKW, and F_OFD_GETLK not supported
2513 337528 leak check heuristic for block prefixed by length as 64bit number
2514 337740 Implement additional Xen hypercalls
2515 337762 guest_arm64_toIR.c:4166 (dis_ARM64_load_store): Assertion `0' failed.
2516 337766 arm64-linux: unhandled syscalls mlock (228) and mlockall (230)
2517 337871 deprecate --db-attach
2518 338023 Add support for all V4L2/media ioctls
2519 338024 inlined functions are not shown if DW_AT_ranges is used
2520 338106 Add support for 'kcmp' syscall
2521 338115 DRD: computed conflict set differs from actual after fork
2522 338160 implement display of thread local storage in gdbsrv
2523 338205 configure.ac and check for -Wno-tautological-compare
2524 338300 coredumps are missing one byte of every segment
2525 338445 amd64 vbit-test fails with unknown opcodes used by arm64 VEX
2526 338499 --sim-hints parsing broken due to wrong order in tokens
2527 338615 suppress glibc 2.20 optimized strcmp implementation for ARMv7
2528 338681 Unable to unwind through clone thread created on i386-linux
2529 338698 race condition between gdbsrv and vgdb on startup
2530 338703 helgrind on arm-linux gets false positives in dynamic loader
2531 338791 alt dwz files can be relative of debug/main file
2532 338878 on MacOS: assertion 'VG_IS_PAGE_ALIGNED(clstack_end+1)' failed
2533 338932 build V-trunk with gcc-trunk
2534 338974 glibc 2.20 changed size of struct sigaction sa_flags field on s390
2535 345079 Fix build problems in VEX/useful/test_main.c
2536 n-i-bz Fix KVM_CREATE_IRQCHIP ioctl handling
2537 n-i-bz s390x: Fix memory corruption for multithreaded applications
2538 n-i-bz vex arm->IR: allow PC as basereg in some LDRD cases
2539 n-i-bz internal error in Valgrind if vgdb transmit signals when ptrace invoked
2540 n-i-bz Fix mingw64 support in valgrind.h (dev@, 9 May 2014)
2541 n-i-bz drd manual: Document how to C++11 programs that use class "std::thread"
2542 n-i-bz Add command-line option --default-suppressions
2543 n-i-bz Add support for BLKDISCARDZEROES ioctl
2544 n-i-bz ppc32/64: fix a regression with the mtfsb0/mtfsb1 instructions
2545 n-i-bz Add support for sys_pivot_root and sys_unshare
2547 (3.10.0.BETA1: 2 September 2014, vex r2940, valgrind r14428)
2548 (3.10.0.BETA2: 8 September 2014, vex r2950, valgrind r14503)
2549 (3.10.0: 10 September 2014, vex r2950, valgrind r14514)
2553 Release 3.9.0 (31 October 2013)
2554 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2555 3.9.0 is a feature release with many improvements and the usual
2556 collection of bug fixes.
2558 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2559 PPC64/Linux, S390X/Linux, MIPS32/Linux, MIPS64/Linux, ARM/Android,
2560 X86/Android, X86/MacOSX 10.7 and AMD64/MacOSX 10.7. Support for
2561 MacOSX 10.8 is significantly improved relative to the 3.8.0 release.
2563 * ================== PLATFORM CHANGES =================
2565 * Support for MIPS64 LE and BE running Linux. Valgrind has been
2566 tested on MIPS64 Debian Squeeze and Debian Wheezy distributions.
2568 * Support for MIPS DSP ASE on MIPS32 platforms.
2570 * Support for s390x Decimal Floating Point instructions on hosts that
2571 have the DFP facility installed.
2573 * Support for POWER8 (Power ISA 2.07) instructions
2575 * Support for Intel AVX2 instructions. This is available only on 64
2578 * Initial support for Intel Transactional Synchronization Extensions,
2581 * Initial support for Hardware Transactional Memory on POWER.
2583 * Improved support for MacOSX 10.8 (64-bit only). Memcheck can now
2584 run large GUI apps tolerably well.
2586 * ==================== TOOL CHANGES ====================
2590 - Improvements in handling of vectorised code, leading to
2591 significantly fewer false error reports. You need to use the flag
2592 --partial-loads-ok=yes to get the benefits of these changes.
2594 - Better control over the leak checker. It is now possible to
2595 specify which leak kinds (definite/indirect/possible/reachable)
2596 should be displayed, which should be regarded as errors, and which
2597 should be suppressed by a given leak suppression. This is done
2598 using the options --show-leak-kinds=kind1,kind2,..,
2599 --errors-for-leak-kinds=kind1,kind2,.. and an optional
2600 "match-leak-kinds:" line in suppression entries, respectively.
2602 Note that generated leak suppressions contain this new line and
2603 are therefore more specific than in previous releases. To get the
2604 same behaviour as previous releases, remove the "match-leak-kinds:"
2605 line from generated suppressions before using them.
2607 - Reduced "possible leak" reports from the leak checker by the use
2608 of better heuristics. The available heuristics provide detection
2609 of valid interior pointers to std::stdstring, to new[] allocated
2610 arrays with elements having destructors and to interior pointers
2611 pointing to an inner part of a C++ object using multiple
2612 inheritance. They can be selected individually using the
2613 option --leak-check-heuristics=heur1,heur2,...
2615 - Better control of stacktrace acquisition for heap-allocated
2616 blocks. Using the --keep-stacktraces option, it is possible to
2617 control independently whether a stack trace is acquired for each
2618 allocation and deallocation. This can be used to create better
2619 "use after free" errors or to decrease Valgrind's resource
2620 consumption by recording less information.
2622 - Better reporting of leak suppression usage. The list of used
2623 suppressions (shown when the -v option is given) now shows, for
2624 each leak suppressions, how many blocks and bytes it suppressed
2625 during the last leak search.
2629 - False errors resulting from the use of statically initialised
2630 mutexes and condition variables (PTHREAD_MUTEX_INITIALISER, etc)
2633 - False errors resulting from the use of pthread_cond_waits that
2634 timeout, have been removed.
2636 * ==================== OTHER CHANGES ====================
2638 * Some attempt to tune Valgrind's space requirements to the expected
2639 capabilities of the target:
2641 - The default size of the translation cache has been reduced from 8
2642 sectors to 6 on Android platforms, since each sector occupies
2643 about 40MB when using Memcheck.
2645 - The default size of the translation cache has been increased to 16
2646 sectors on all other platforms, reflecting the fact that large
2647 applications require instrumentation and storage of huge amounts
2648 of code. For similar reasons, the number of memory mapped
2649 segments that can be tracked has been increased by a factor of 6.
2651 - In all cases, the maximum number of sectors in the translation
2652 cache can be controlled by the new flag --num-transtab-sectors.
2654 * Changes in how debug info (line numbers, etc) is read:
2656 - Valgrind no longer temporarily mmaps the entire object to read
2657 from it. Instead, reading is done through a small fixed sized
2658 buffer. This avoids virtual memory usage spikes when Valgrind
2659 reads debuginfo from large shared objects.
2661 - A new experimental remote debug info server. Valgrind can read
2662 debug info from a different machine (typically, a build host)
2663 where debuginfo objects are stored. This can save a lot of time
2664 and hassle when running Valgrind on resource-constrained targets
2665 (phones, tablets) when the full debuginfo objects are stored
2666 somewhere else. This is enabled by the --debuginfo-server=
2669 - Consistency checking between main and debug objects can be
2670 disabled using the --allow-mismatched-debuginfo option.
2672 * Stack unwinding by stack scanning, on ARM. Unwinding by stack
2673 scanning can recover stack traces in some cases when the normal
2674 unwind mechanisms fail. Stack scanning is best described as "a
2675 nasty, dangerous and misleading hack" and so is disabled by default.
2676 Use --unw-stack-scan-thresh and --unw-stack-scan-frames to enable
2679 * Detection and merging of recursive stack frame cycles. When your
2680 program has recursive algorithms, this limits the memory used by
2681 Valgrind for recorded stack traces and avoids recording
2682 uninteresting repeated calls. This is controlled by the command
2683 line option --merge-recursive-frame and by the monitor command
2684 "v.set merge-recursive-frames".
2686 * File name and line numbers for used suppressions. The list of used
2687 suppressions (shown when the -v option is given) now shows, for each
2688 used suppression, the file name and line number where the suppression
2691 * New and modified GDB server monitor features:
2693 - valgrind.h has a new client request, VALGRIND_MONITOR_COMMAND,
2694 that can be used to execute gdbserver monitor commands from the
2697 - A new monitor command, "v.info open_fds", that gives the list of
2698 open file descriptors and additional details.
2700 - An optional message in the "v.info n_errs_found" monitor command,
2701 for example "v.info n_errs_found test 1234 finished", allowing a
2702 comment string to be added to the process output, perhaps for the
2703 purpose of separating errors of different tests or test phases.
2705 - A new monitor command "v.info execontext" that shows information
2706 about the stack traces recorded by Valgrind.
2708 - A new monitor command "v.do expensive_sanity_check_general" to run
2709 some internal consistency checks.
2711 * New flag --sigill-diagnostics to control whether a diagnostic
2712 message is printed when the JIT encounters an instruction it can't
2713 translate. The actual behavior -- delivery of SIGILL to the
2714 application -- is unchanged.
2716 * The maximum amount of memory that Valgrind can use on 64 bit targets
2717 has been increased from 32GB to 64GB. This should make it possible
2718 to run applications on Memcheck that natively require up to about 35GB.
2720 * ==================== FIXED BUGS ====================
2722 The following bugs have been fixed or resolved. Note that "n-i-bz"
2723 stands for "not in bugzilla" -- that is, a bug that was reported to us
2724 but never got a bugzilla entry. We encourage you to file bugs in
2725 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2726 than mailing the developers (or mailing lists) directly -- bugs that
2727 are not entered into bugzilla tend to get forgotten about or ignored.
2729 To see details of a given bug, visit
2730 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2731 where XXXXXX is the bug number as listed below.
2733 123837 system call: 4th argument is optional, depending on cmd
2734 135425 memcheck should tell you where Freed blocks were Mallocd
2735 164485 VG_N_SEGNAMES and VG_N_SEGMENTS are (still) too small
2736 207815 Adds some of the drm ioctls to syswrap-linux.c
2737 251569 vex amd64->IR: 0xF 0x1 0xF9 0xBF 0x90 0xD0 0x3 0x0 (RDTSCP)
2738 252955 Impossible to compile with ccache
2739 253519 Memcheck reports auxv pointer accesses as invalid reads.
2740 263034 Crash when loading some PPC64 binaries
2741 269599 Increase deepest backtrace
2742 274695 s390x: Support "compare to/from logical" instructions (z196)
2743 275800 s390x: Autodetect cache info (part 2)
2744 280271 Valgrind reports possible memory leaks on still-reachable std::string
2745 284540 Memcheck shouldn't count suppressions matching still-reachable [..]
2746 289578 Backtraces with ARM unwind tables (stack scan flags)
2747 296311 Wrong stack traces due to -fomit-frame-pointer (x86)
2748 304832 ppc32: build failure
2749 305431 Use find_buildid shdr fallback for separate .debug files
2750 305728 Add support for AVX2 instructions
2751 305948 ppc64: code generation for ShlD64 / ShrD64 asserts
2752 306035 s390x: Fix IR generation for LAAG and friends
2753 306054 s390x: Condition code computation for convert-to-int/logical
2754 306098 s390x: alternate opcode form for convert to/from fixed
2755 306587 Fix cache line detection from auxiliary vector for PPC.
2756 306783 Mips unhandled syscall : 4025 / 4079 / 4182
2757 307038 DWARF2 CFI reader: unhandled DW_OP_ opcode 0x8 (DW_OP_const1u et al)
2758 307082 HG false positive: pthread_cond_destroy: destruction of unknown CV
2759 307101 sys_capget second argument can be NULL
2760 307103 sys_openat: If pathname is absolute, then dirfd is ignored.
2761 307106 amd64->IR: f0 0f c0 02 (lock xadd byte)
2762 307113 s390x: DFP support
2763 307141 valgrind does't work in mips-linux system
2764 307155 filter_gdb should filter out syscall-template.S T_PSEUDO
2765 307285 x86_amd64 feature test for avx in test suite is wrong
2766 307290 memcheck overlap testcase needs memcpy version filter
2767 307463 Please add "&limit=0" to the "all open bugs" link
2768 307465 --show-possibly-lost=no should reduce the error count / exit code
2769 307557 Leaks on Mac OS X 10.7.5 libraries at ImageLoader::recursiveInit[..]
2770 307729 pkgconfig support broken valgrind.pc
2771 307828 Memcheck false errors SSE optimized wcscpy, wcscmp, wcsrchr, wcschr
2772 307955 Building valgrind 3.7.0-r4 fails in Gentoo AMD64 when using clang
2773 308089 Unhandled syscall on ppc64: prctl
2774 308135 PPC32 MPC8xx has 16 bytes cache size
2775 308321 testsuite memcheck filter interferes with gdb_filter
2777 308341 vgdb should report process exit (or fatal signal)
2778 308427 s390 memcheck reports tsearch cjump/cmove depends on uninit
2779 308495 Remove build dependency on installed Xen headers
2780 308573 Internal error on 64-bit instruction executed in 32-bit mode
2782 308627 pmovmskb validity bit propagation is imprecise
2783 308644 vgdb command for having the info for the track-fds option
2784 308711 give more info about aspacemgr and arenas in out_of_memory
2785 308717 ARM: implement fixed-point VCVT.F64.[SU]32
2786 308718 ARM implement SMLALBB family of instructions
2787 308886 Missing support for PTRACE_SET/GETREGSET
2788 308930 syscall name_to_handle_at (303 on amd64) not handled
2789 309229 V-bit tester does not report number of tests generated
2790 309323 print unrecognized instuction on MIPS
2791 309425 Provide a --sigill-diagnostics flag to suppress illegal [..]
2792 309427 SSE optimized stpncpy trigger uninitialised value [..] errors
2793 309430 Self hosting ppc64 encounters a vassert error on operand type
2794 309600 valgrind is a bit confused about 0-sized sections
2795 309823 Generate errors for still reachable blocks
2796 309921 PCMPISTRI validity bit propagation is imprecise
2797 309922 none/tests/ppc64/test_dfp5 sometimes fails
2798 310169 The Iop_CmpORD class of Iops is not supported by the vbit checker.
2799 310424 --read-var-info does not properly describe static variables
2800 310792 search additional path for debug symbols
2801 310931 s390x: Message-security assist (MSA) instruction extension [..]
2802 311100 PPC DFP implementation of the integer operands is inconsistent [..]
2803 311318 ARM: "128-bit constant is not implemented" error message
2804 311407 ssse3 bcopy (actually converted memcpy) causes invalid read [..]
2805 311690 V crashes because it redirects branches inside of a redirected function
2806 311880 x86_64: make regtest hangs at shell_valid1
2807 311922 WARNING: unhandled syscall: 170
2809 312171 ppc: insn selection for DFP
2810 312571 Rounding mode call wrong for the DFP Iops [..]
2811 312620 Change to Iop_D32toD64 [..] for s390 DFP support broke ppc [..]
2812 312913 Dangling pointers error should also report the alloc stack trace
2813 312980 Building on Mountain Lion generates some compiler warnings
2814 313267 Adding MIPS64/Linux port to Valgrind
2817 313811 Buffer overflow in assert_fail
2818 314099 coverity pointed out error in VEX guest_ppc_toIR.c insn_suffix
2819 314269 ppc: dead code in insn selection
2820 314718 ARM: implement integer divide instruction (sdiv and udiv)
2821 315345 cl-format.xml and callgrind/dump.c don't agree on using cfl= or cfi=
2822 315441 sendmsg syscall should ignore unset msghdr msg_flags
2823 315534 msgrcv inside a thread causes valgrind to hang (block)
2824 315545 Assertion '(UChar*)sec->tt[tteNo].tcptr <= (UChar*)hcode' failed
2825 315689 disInstr(thumb): unhandled instruction: 0xF852 0x0E10 (LDRT)
2826 315738 disInstr(arm): unhandled instruction: 0xEEBE0BEE (vcvt.s32.f64)
2827 315959 valgrind man page has bogus SGCHECK (and no BBV) OPTIONS section
2828 316144 valgrind.1 manpage contains unknown ??? strings [..]
2829 316145 callgrind command line options in manpage reference (unknown) [..]
2830 316145 callgrind command line options in manpage reference [..]
2831 316181 drd: Fixed a 4x slowdown for certain applications
2832 316503 Valgrind does not support SSE4 "movntdqa" instruction
2833 316535 Use of |signed int| instead of |size_t| in valgrind messages
2834 316696 fluidanimate program of parsec 2.1 stuck
2835 316761 syscall open_by_handle_at (304 on amd64, 342 on x86) not handled
2836 317091 Use -Wl,-Ttext-segment when static linking if possible [..]
2837 317186 "Impossible happens" when occurs VCVT instruction on ARM
2838 317318 Support for Threading Building Blocks "scalable_malloc"
2839 317444 amd64->IR: 0xC4 0x41 0x2C 0xC2 0xD2 0x8 (vcmpeq_uqps)
2840 317461 Fix BMI assembler configure check and avx2/bmi/fma vgtest prereqs
2841 317463 bmi testcase IR SANITY CHECK FAILURE
2842 317506 memcheck/tests/vbit-test fails with unknown opcode after [..]
2843 318050 libmpiwrap fails to compile with out-of-source build
2844 318203 setsockopt handling needs to handle SOL_SOCKET/SO_ATTACH_FILTER
2845 318643 annotate_trace_memory tests infinite loop on arm and ppc [..]
2846 318773 amd64->IR: 0xF3 0x48 0x0F 0xBC 0xC2 0xC3 0x66 0x0F
2847 318929 Crash with: disInstr(thumb): 0xF321 0x0001 (ssat16)
2848 318932 Add missing PPC64 and PPC32 system call support
2849 319235 --db-attach=yes is broken with Yama (ptrace scoping) enabled
2850 319395 Crash with unhandled instruction on STRT (Thumb) instructions
2851 319494 VEX Makefile-gcc standalone build update after r2702
2852 319505 [MIPSEL] Crash: unhandled UNRAY operator.
2853 319858 disInstr(thumb): unhandled instruction on instruction STRBT
2854 319932 disInstr(thumb): unhandled instruction on instruction STRHT
2855 320057 Problems when we try to mmap more than 12 memory pages on MIPS32
2856 320063 Memory from PTRACE_GET_THREAD_AREA is reported uninitialised
2857 320083 disInstr(thumb): unhandled instruction on instruction LDRBT
2858 320116 bind on AF_BLUETOOTH produces warnings because of sockaddr_rc padding
2859 320131 WARNING: unhandled syscall: 369 on ARM (prlimit64)
2860 320211 Stack buffer overflow in ./coregrind/m_main.c with huge TMPDIR
2861 320661 vgModuleLocal_read_elf_debug_info(): "Assertion '!di->soname'
2862 320895 add fanotify support (patch included)
2863 320998 vex amd64->IR pcmpestri and pcmpestrm SSE4.2 instruction
2864 321065 Valgrind updates for Xen 4.3
2865 321148 Unhandled instruction: PLI (Thumb 1, 2, 3)
2866 321363 Unhandled instruction: SSAX (ARM + Thumb)
2867 321364 Unhandled instruction: SXTAB16 (ARM + Thumb)
2868 321466 Unhandled instruction: SHASX (ARM + Thumb)
2869 321467 Unhandled instruction: SHSAX (ARM + Thumb)
2870 321468 Unhandled instruction: SHSUB16 (ARM + Thumb)
2871 321619 Unhandled instruction: SHSUB8 (ARM + Thumb)
2872 321620 Unhandled instruction: UASX (ARM + Thumb)
2873 321621 Unhandled instruction: USAX (ARM + Thumb)
2874 321692 Unhandled instruction: UQADD16 (ARM + Thumb)
2875 321693 Unhandled instruction: LDRSBT (Thumb)
2876 321694 Unhandled instruction: UQASX (ARM + Thumb)
2877 321696 Unhandled instruction: UQSAX (Thumb + ARM)
2878 321697 Unhandled instruction: UHASX (ARM + Thumb)
2879 321703 Unhandled instruction: UHSAX (ARM + Thumb)
2880 321704 Unhandled instruction: REVSH (ARM + Thumb)
2881 321730 Add cg_diff and cg_merge man pages
2882 321738 Add vgdb and valgrind-listener man pages
2884 321891 Unhandled instruction: LDRHT (Thumb)
2885 321960 pthread_create() then alloca() causing invalid stack write errors
2886 321969 ppc32 and ppc64 don't support [lf]setxattr
2887 322254 Show threadname together with tid if set by application
2888 322294 Add initial support for IBM Power ISA 2.07
2889 322368 Assertion failure in wqthread_hijack under OS X 10.8
2890 322563 vex mips->IR: 0x70 0x83 0xF0 0x3A
2891 322807 VALGRIND_PRINTF_BACKTRACE writes callstack to xml and text to stderr
2892 322851 0bXXX binary literal syntax is not standard
2893 323035 Unhandled instruction: LDRSHT(Thumb)
2894 323036 Unhandled instruction: SMMLS (ARM and Thumb)
2895 323116 The memcheck/tests/ppc64/power_ISA2_05.c fails to build [..]
2896 323175 Unhandled instruction: SMLALD (ARM + Thumb)
2897 323177 Unhandled instruction: SMLSLD (ARM + Thumb)
2898 323432 Calling pthread_cond_destroy() or pthread_mutex_destroy() [..]
2899 323437 Phase 2 support for IBM Power ISA 2.07
2900 323713 Support mmxext (integer sse) subset on i386 (athlon)
2901 323803 Transactional memory instructions are not supported for Power
2902 323893 SSE3 not available on amd cpus in valgrind
2903 323905 Probable false positive from Valgrind/drd on close()
2904 323912 valgrind.h header isn't compatible for mingw64
2905 324047 Valgrind doesn't support [LDR,ST]{S}[B,H]T ARM instructions
2906 324149 helgrind: When pthread_cond_timedwait returns ETIMEDOUT [..]
2907 324181 mmap does not handle MAP_32BIT
2908 324227 memcheck false positive leak when a thread calls exit+block [..]
2909 324421 Support for fanotify API on ARM architecture
2910 324514 gdbserver monitor cmd output behaviour consistency [..]
2911 324518 ppc64: Emulation of dcbt instructions does not handle [..]
2912 324546 none/tests/ppc32 test_isa_2_07_part2 requests -m64
2913 324582 When access is made to freed memory, report both allocation [..]
2914 324594 Fix overflow computation for Power ISA 2.06 insns: mulldo/mulldo.
2915 324765 ppc64: illegal instruction when executing none/tests/ppc64/jm-misc
2916 324816 Incorrect VEX implementation for xscvspdp/xvcvspdp for SNaN inputs
2917 324834 Unhandled instructions in Microsoft C run-time for x86_64
2918 324894 Phase 3 support for IBM Power ISA 2.07
2919 326091 drd: Avoid false race reports from optimized strlen() impls
2920 326113 valgrind libvex hwcaps error on AMD64
2921 n-i-bz Some wrong command line options could be ignored
2922 n-i-bz patch to allow fair-sched on android
2923 n-i-bz report error for vgdb snapshot requested before execution
2924 n-i-bz same as 303624 (fixed in 3.8.0), but for x86 android
2926 (3.9.0: 31 October 2013, vex r2796, valgrind r13708)
2930 Release 3.8.1 (19 September 2012)
2931 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2932 3.8.1 is a bug fix release. It fixes some assertion failures in 3.8.0
2933 that occur moderately frequently in real use cases, adds support for
2934 some missing instructions on ARM, and fixes a deadlock condition on
2935 MacOSX. If you package or deliver 3.8.0 for others to use, you might
2936 want to consider upgrading to 3.8.1 instead.
2938 The following bugs have been fixed or resolved. Note that "n-i-bz"
2939 stands for "not in bugzilla" -- that is, a bug that was reported to us
2940 but never got a bugzilla entry. We encourage you to file bugs in
2941 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2942 than mailing the developers (or mailing lists) directly -- bugs that
2943 are not entered into bugzilla tend to get forgotten about or ignored.
2945 To see details of a given bug, visit
2946 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2947 where XXXXXX is the bug number as listed below.
2950 289584 Unhandled instruction: 0xF 0x29 0xE5 (MOVAPS)
2951 295808 amd64->IR: 0xF3 0xF 0xBC 0xC0 (TZCNT)
2952 298281 wcslen causes false(?) uninitialised value warnings
2953 301281 valgrind hangs on OS X when the process calls system()
2954 304035 disInstr(arm): unhandled instruction 0xE1023053
2955 304867 implement MOVBE instruction in x86 mode
2956 304980 Assertion 'lo <= hi' failed in vgModuleLocal_find_rx_mapping
2957 305042 amd64: implement 0F 7F encoding of movq between two registers
2958 305199 ARM: implement QDADD and QDSUB
2959 305321 amd64->IR: 0xF 0xD 0xC (prefetchw)
2960 305513 killed by fatal signal: SIGSEGV
2961 305690 DRD reporting invalid semaphore when sem_trywait fails
2962 305926 Invalid alignment checks for some AVX instructions
2963 306297 disInstr(thumb): unhandled instruction 0xE883 0x000C
2964 306310 3.8.0 release tarball missing some files
2965 306612 RHEL 6 glibc-2.X default suppressions need /lib*/libc-*patterns
2966 306664 vex amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x46 0x66 0xF
2967 n-i-bz shmat of a segment > 4Gb does not work
2968 n-i-bz simulate_control_c script wrong USR1 signal number on mips
2969 n-i-bz vgdb ptrace calls wrong on mips [...]
2970 n-i-bz Fixes for more MPI false positives
2971 n-i-bz exp-sgcheck's memcpy causes programs to segfault
2972 n-i-bz OSX build w/ clang: asserts at startup
2973 n-i-bz Incorrect undef'dness prop for Iop_DPBtoBCD and Iop_BCDtoDPB
2974 n-i-bz fix a couple of union tag-vs-field mixups
2975 n-i-bz OSX: use __NR_poll_nocancel rather than __NR_poll
2977 The following bugs were fixed in 3.8.0 but not listed in this NEWS
2980 254088 Valgrind should know about UD2 instruction
2983 304754 NEWS blows TeX's little mind
2985 (3.8.1: 19 September 2012, vex r2537, valgrind r12996)
2989 Release 3.8.0 (10 August 2012)
2990 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2991 3.8.0 is a feature release with many improvements and the usual
2992 collection of bug fixes.
2994 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2995 PPC64/Linux, S390X/Linux, MIPS/Linux, ARM/Android, X86/Android,
2996 X86/MacOSX 10.6/10.7 and AMD64/MacOSX 10.6/10.7. Support for recent
2997 distros and toolchain components (glibc 2.16, gcc 4.7) has been added.
2998 There is initial support for MacOSX 10.8, but it is not usable for
2999 serious work at present.
3001 * ================== PLATFORM CHANGES =================
3003 * Support for MIPS32 platforms running Linux. Valgrind has been
3004 tested on MIPS32 and MIPS32r2 platforms running different Debian
3005 Squeeze and MeeGo distributions. Both little-endian and big-endian
3006 cores are supported. The tools Memcheck, Massif and Lackey have
3007 been tested and are known to work. See README.mips for more details.
3009 * Preliminary support for Android running on x86.
3011 * Preliminary (as-yet largely unusable) support for MacOSX 10.8.
3013 * Support for Intel AVX instructions and for AES instructions. This
3014 support is available only for 64 bit code.
3016 * Support for POWER Decimal Floating Point instructions.
3018 * ==================== TOOL CHANGES ====================
3020 * Non-libc malloc implementations are now supported. This is useful
3021 for tools that replace malloc (Memcheck, Massif, DRD, Helgrind).
3022 Using the new option --soname-synonyms, such tools can be informed
3023 that the malloc implementation is either linked statically into the
3024 executable, or is present in some other shared library different
3025 from libc.so. This makes it possible to process statically linked
3026 programs, and programs using other malloc libraries, for example
3027 TCMalloc or JEMalloc.
3029 * For tools that provide their own replacement for malloc et al, the
3030 option --redzone-size=<number> allows users to specify the size of
3031 the padding blocks (redzones) added before and after each client
3032 allocated block. Smaller redzones decrease the memory needed by
3033 Valgrind. Bigger redzones increase the chance to detect blocks
3034 overrun or underrun. Prior to this change, the redzone size was
3035 hardwired to 16 bytes in Memcheck.
3039 - The leak_check GDB server monitor command now can
3040 control the maximum nr of loss records to output.
3042 - Reduction of memory use for applications allocating
3043 many blocks and/or having many partially defined bytes.
3045 - Addition of GDB server monitor command 'block_list' that lists
3046 the addresses/sizes of the blocks of a leak search loss record.
3048 - Addition of GDB server monitor command 'who_points_at' that lists
3049 the locations pointing at a block.
3051 - If a redzone size > 0 is given, VALGRIND_MALLOCLIKE_BLOCK now will
3052 detect an invalid access of these redzones, by marking them
3053 noaccess. Similarly, if a redzone size is given for a memory
3054 pool, VALGRIND_MEMPOOL_ALLOC will mark the redzones no access.
3055 This still allows to find some bugs if the user has forgotten to
3056 mark the pool superblock noaccess.
3058 - Performance of memory leak check has been improved, especially in
3059 cases where there are many leaked blocks and/or many suppression
3060 rules used to suppress leak reports.
3062 - Reduced noise (false positive) level on MacOSX 10.6/10.7, due to
3063 more precise analysis, which is important for LLVM/Clang
3064 generated code. This is at the cost of somewhat reduced
3065 performance. Note there is no change to analysis precision or
3066 costs on Linux targets.
3070 - Added even more facilities that can help finding the cause of a data
3071 race, namely the command-line option --ptrace-addr and the macro
3072 DRD_STOP_TRACING_VAR(x). More information can be found in the manual.
3074 - Fixed a subtle bug that could cause false positive data race reports.
3076 * ==================== OTHER CHANGES ====================
3078 * The C++ demangler has been updated so as to work well with C++
3079 compiled by up to at least g++ 4.6.
3081 * Tool developers can make replacement/wrapping more flexible thanks
3082 to the new option --soname-synonyms. This was reported above, but
3083 in fact is very general and applies to all function
3084 replacement/wrapping, not just to malloc-family functions.
3086 * Round-robin scheduling of threads can be selected, using the new
3087 option --fair-sched= yes. Prior to this change, the pipe-based
3088 thread serialisation mechanism (which is still the default) could
3089 give very unfair scheduling. --fair-sched=yes improves
3090 responsiveness of interactive multithreaded applications, and
3091 improves repeatability of results from the thread checkers Helgrind
3094 * For tool developers: support to run Valgrind on Valgrind has been
3095 improved. We can now routinely Valgrind on Helgrind or Memcheck.
3097 * gdbserver now shows the float shadow registers as integer
3098 rather than float values, as the shadow values are mostly
3099 used as bit patterns.
3101 * Increased limit for the --num-callers command line flag to 500.
3103 * Performance improvements for error matching when there are many
3104 suppression records in use.
3106 * Improved support for DWARF4 debugging information (bug 284184).
3108 * Initial support for DWZ compressed Dwarf debug info.
3110 * Improved control over the IR optimiser's handling of the tradeoff
3111 between performance and precision of exceptions. Specifically,
3112 --vex-iropt-precise-memory-exns has been removed and replaced by
3113 --vex-iropt-register-updates, with extended functionality. This
3114 allows the Valgrind gdbserver to always show up to date register
3117 * Modest performance gains through the use of translation chaining for
3120 * ==================== FIXED BUGS ====================
3122 The following bugs have been fixed or resolved. Note that "n-i-bz"
3123 stands for "not in bugzilla" -- that is, a bug that was reported to us
3124 but never got a bugzilla entry. We encourage you to file bugs in
3125 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
3126 than mailing the developers (or mailing lists) directly -- bugs that
3127 are not entered into bugzilla tend to get forgotten about or ignored.
3129 To see details of a given bug, visit
3130 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3131 where XXXXXX is the bug number as listed below.
3133 197914 Building valgrind from svn now requires automake-1.10
3134 203877 increase to 16Mb maximum allowed alignment for memalign et al
3135 219156 Handle statically linked malloc or other malloc lib (e.g. tcmalloc)
3136 247386 make perf does not run all performance tests
3137 270006 Valgrind scheduler unfair
3138 270777 Adding MIPS/Linux port to Valgrind
3139 270796 s390x: Removed broken support for the TS insn
3140 271438 Fix configure for proper SSE4.2 detection
3141 273114 s390x: Support TR, TRE, TROO, TROT, TRTO, and TRTT instructions
3142 273475 Add support for AVX instructions
3143 274078 improved configure logic for mpicc
3144 276993 fix mremap 'no thrash checks'
3145 278313 Fedora 15/x64: err read debug info with --read-var-info=yes flag
3146 281482 memcheck incorrect byte allocation count in realloc() for silly argument
3147 282230 group allocator for small fixed size, use it for MC_Chunk/SEc vbit
3148 283413 Fix wrong sanity check
3149 283671 Robustize alignment computation in LibVEX_Alloc
3150 283961 Adding support for some HCI IOCTLs
3151 284124 parse_type_DIE: confused by: DWARF 4
3152 284864 == 273475 (Add support for AVX instructions)
3153 285219 Too-restrictive constraints for Thumb2 "SP plus/minus register"
3154 285662 (MacOSX): Memcheck needs to replace memcpy/memmove
3155 285725 == 273475 (Add support for AVX instructions)
3156 286261 add wrapper for linux I2C_RDWR ioctl
3157 286270 vgpreload is not friendly to 64->32 bit execs, gives ld.so warnings
3158 286374 Running cachegrind with --branch-sim=yes on 64-bit PowerPC program fails
3159 286384 configure fails "checking for a supported version of gcc"
3160 286497 == 273475 (Add support for AVX instructions)
3161 286596 == 273475 (Add support for AVX instructions)
3162 286917 disInstr(arm): unhandled instruction: QADD (also QSUB)
3163 287175 ARM: scalar VFP fixed-point VCVT instructions not handled
3164 287260 Incorrect conditional jump or move depends on uninitialised value(s)
3165 287301 vex amd64->IR: 0x66 0xF 0x38 0x41 0xC0 0xB8 0x0 0x0 (PHMINPOSUW)
3166 287307 == 273475 (Add support for AVX instructions)
3167 287858 VG_(strerror): unknown error
3168 288298 (MacOSX) unhandled syscall shm_unlink
3169 288995 == 273475 (Add support for AVX instructions)
3170 289470 Loading of large Mach-O thin binaries fails.
3171 289656 == 273475 (Add support for AVX instructions)
3172 289699 vgdb connection in relay mode erroneously closed due to buffer overrun
3173 289823 == 293754 (PCMPxSTRx not implemented for 16-bit characters)
3174 289839 s390x: Provide support for unicode conversion instructions
3175 289939 monitor cmd 'leak_check' with details about leaked or reachable blocks
3176 290006 memcheck doesn't mark %xmm as initialized after "pcmpeqw %xmm %xmm"
3177 290655 Add support for AESKEYGENASSIST instruction
3178 290719 valgrind-3.7.0 fails with automake-1.11.2 due to"pkglibdir" usage
3179 290974 vgdb must align pages to VKI_SHMLBA (16KB) on ARM
3180 291253 ES register not initialised in valgrind simulation
3181 291568 Fix 3DNOW-related crashes with baseline x86_64 CPU (w patch)
3182 291865 s390x: Support the "Compare Double and Swap" family of instructions
3183 292300 == 273475 (Add support for AVX instructions)
3184 292430 unrecognized instruction in __intel_get_new_mem_ops_cpuid
3185 292493 == 273475 (Add support for AVX instructions)
3186 292626 Missing fcntl F_SETOWN_EX and F_GETOWN_EX support
3187 292627 Missing support for some SCSI ioctls
3188 292628 none/tests/x86/bug125959-x86.c triggers undefined behavior
3189 292841 == 273475 (Add support for AVX instructions)
3190 292993 implement the getcpu syscall on amd64-linux
3191 292995 Implement the “cross memory attach” syscalls introduced in Linux 3.2
3192 293088 Add some VEX sanity checks for ppc64 unhandled instructions
3193 293751 == 290655 (Add support for AESKEYGENASSIST instruction)
3194 293754 PCMPxSTRx not implemented for 16-bit characters
3195 293755 == 293754 (No tests for PCMPxSTRx on 16-bit characters)
3196 293808 CLFLUSH not supported by latest VEX for amd64
3197 294047 valgrind does not correctly emulate prlimit64(..., RLIMIT_NOFILE, ...)
3198 294048 MPSADBW instruction not implemented
3199 294055 regtest none/tests/shell fails when locale is not set to C
3200 294185 INT 0x44 (and others) not supported on x86 guest, but used by Jikes RVM
3201 294190 --vgdb-error=xxx can be out of sync with errors shown to the user
3202 294191 amd64: fnsave/frstor and 0x66 size prefixes on FP instructions
3203 294260 disInstr_AMD64: disInstr miscalculated next %rip
3204 294523 --partial-loads-ok=yes causes false negatives
3205 294617 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A
3206 294736 vex amd64->IR: 0x48 0xF 0xD7 0xD6 0x48 0x83
3207 294812 patch allowing to run (on x86 at least) helgrind/drd on tool.
3208 295089 can not annotate source for both helgrind and drd
3209 295221 POWER Processor decimal floating point instruction support missing
3210 295427 building for i386 with clang on darwin11 requires "-new_linker linker"
3211 295428 coregrind/m_main.c has incorrect x86 assembly for darwin
3212 295590 Helgrind: Assertion 'cvi->nWaiters > 0' failed
3213 295617 ARM - Add some missing syscalls
3214 295799 Missing \n with get_vbits in gdbserver when line is % 80 [...]
3215 296229 Linux user input device ioctls missing wrappers
3216 296318 ELF Debug info improvements (more than one rx/rw mapping)
3217 296422 Add translation chaining support
3218 296457 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A (dup of AES)
3219 296792 valgrind 3.7.0: add SIOCSHWTSTAMP (0x89B0) ioctl wrapper
3220 296983 Fix build issues on x86_64/ppc64 without 32-bit toolchains
3221 297078 gdbserver signal handling problems [..]
3222 297147 drd false positives on newly allocated memory
3223 297329 disallow decoding of IBM Power DFP insns on some machines
3224 297497 POWER Processor decimal floating point instruction support missing
3225 297701 Another alias for strncasecmp_l in libc-2.13.so
3226 297911 'invalid write' not reported when using APIs for custom mem allocators.
3227 297976 s390x: revisit EX implementation
3228 297991 Valgrind interferes with mmap()+ftell()
3229 297992 Support systems missing WIFCONTINUED (e.g. pre-2.6.10 Linux)
3230 297993 Fix compilation of valgrind with gcc -g3.
3231 298080 POWER Processor DFP support missing, part 3
3232 298227 == 273475 (Add support for AVX instructions)
3233 298335 == 273475 (Add support for AVX instructions)
3234 298354 Unhandled ARM Thumb instruction 0xEB0D 0x0585 (streq)
3235 298394 s390x: Don't bail out on an unknown machine model. [..]
3236 298421 accept4() syscall (366) support is missing for ARM
3237 298718 vex amd64->IR: 0xF 0xB1 0xCB 0x9C 0x8F 0x45
3238 298732 valgrind installation problem in ubuntu with kernel version 3.x
3239 298862 POWER Processor DFP instruction support missing, part 4
3240 298864 DWARF reader mis-parses DW_FORM_ref_addr
3241 298943 massif asserts with --pages-as-heap=yes when brk is changing [..]
3242 299053 Support DWARF4 DW_AT_high_pc constant form
3243 299104 == 273475 (Add support for AVX instructions)
3244 299316 Helgrind: hg_main.c:628 (map_threads_lookup): Assertion 'thr' failed.
3245 299629 dup3() syscall (358) support is missing for ARM
3246 299694 POWER Processor DFP instruction support missing, part 5
3247 299756 Ignore --free-fill for MEMPOOL_FREE and FREELIKE client requests
3248 299803 == 273475 (Add support for AVX instructions)
3249 299804 == 273475 (Add support for AVX instructions)
3250 299805 == 273475 (Add support for AVX instructions)
3251 300140 ARM - Missing (T1) SMMUL
3252 300195 == 296318 (ELF Debug info improvements (more than one rx/rw mapping))
3253 300389 Assertion `are_valid_hwcaps(VexArchAMD64, [..])' failed.
3254 300414 FCOM and FCOMP unimplemented for amd64 guest
3255 301204 infinite loop in canonicaliseSymtab with ifunc symbol
3256 301229 == 203877 (increase to 16Mb maximum allowed alignment for memalign etc)
3257 301265 add x86 support to Android build
3258 301984 configure script doesn't detect certain versions of clang
3259 302205 Fix compiler warnings for POWER VEX code and POWER test cases
3260 302287 Unhandled movbe instruction on Atom processors
3261 302370 PPC: fnmadd, fnmsub, fnmadds, fnmsubs insns always negate the result
3262 302536 Fix for the POWER Valgrind regression test: memcheck-ISA2.0.
3263 302578 Unrecognized isntruction 0xc5 0x32 0xc2 0xca 0x09 vcmpngess
3264 302656 == 273475 (Add support for AVX instructions)
3265 302709 valgrind for ARM needs extra tls support for android emulator [..]
3266 302827 add wrapper for CDROM_GET_CAPABILITY
3267 302901 Valgrind crashes with dwz optimized debuginfo
3268 302918 Enable testing of the vmaddfp and vnsubfp instructions in the testsuite
3269 303116 Add support for the POWER instruction popcntb
3270 303127 Power test suite fixes for frsqrte, vrefp, and vrsqrtefp instructions.
3271 303250 Assertion `instrs_in->arr_used <= 10000' failed w/ OpenSSL code
3272 303466 == 273475 (Add support for AVX instructions)
3273 303624 segmentation fault on Android 4.1 (e.g. on Galaxy Nexus OMAP)
3274 303963 strstr() function produces wrong results under valgrind callgrind
3275 304054 CALL_FN_xx macros need to enforce stack alignment
3276 304561 tee system call not supported
3277 715750 (MacOSX): Incorrect invalid-address errors near 0xFFFFxxxx (mozbug#)
3278 n-i-bz Add missing gdbserver xml files for shadow registers for ppc32
3279 n-i-bz Bypass gcc4.4/4.5 code gen bugs causing out of memory or asserts
3280 n-i-bz Fix assert in gdbserver for watchpoints watching the same address
3281 n-i-bz Fix false positive in sys_clone on amd64 when optional args [..]
3282 n-i-bz s390x: Shadow registers can now be examined using vgdb
3284 (3.8.0-TEST3: 9 August 2012, vex r2465, valgrind r12865)
3285 (3.8.0: 10 August 2012, vex r2465, valgrind r12866)
3289 Release 3.7.0 (5 November 2011)
3290 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3291 3.7.0 is a feature release with many significant improvements and the
3292 usual collection of bug fixes.
3294 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3295 PPC64/Linux, S390X/Linux, ARM/Android, X86/Darwin and AMD64/Darwin.
3296 Support for recent distros and toolchain components (glibc 2.14, gcc
3297 4.6, MacOSX 10.7) has been added.
3299 * ================== PLATFORM CHANGES =================
3301 * Support for IBM z/Architecture (s390x) running Linux. Valgrind can
3302 analyse 64-bit programs running on z/Architecture. Most user space
3303 instructions up to and including z10 are supported. Valgrind has
3304 been tested extensively on z9, z10, and z196 machines running SLES
3305 10/11, RedHat 5/6m, and Fedora. The Memcheck and Massif tools are
3306 known to work well. Callgrind, Helgrind, and DRD work reasonably
3307 well on z9 and later models. See README.s390 for more details.
3309 * Preliminary support for MacOSX 10.7 and XCode 4. Both 32- and
3310 64-bit processes are supported. Some complex threaded applications
3311 (Firefox) are observed to hang when run as 32 bit applications,
3312 whereas 64-bit versions run OK. The cause is unknown. Memcheck
3313 will likely report some false errors. In general, expect some rough
3314 spots. This release also supports MacOSX 10.6, but drops support
3317 * Preliminary support for Android (on ARM). Valgrind can now run
3318 large applications (eg, Firefox) on (eg) a Samsung Nexus S. See
3319 README.android for more details, plus instructions on how to get
3322 * Support for the IBM Power ISA 2.06 (Power7 instructions)
3324 * General correctness and performance improvements for ARM/Linux, and,
3325 by extension, ARM/Android.
3327 * Further solidification of support for SSE 4.2 in 64-bit mode. AVX
3328 instruction set support is under development but is not available in
3331 * Support for AIX5 has been removed.
3333 * ==================== TOOL CHANGES ====================
3335 * Memcheck: some incremental changes:
3337 - reduction of memory use in some circumstances
3339 - improved handling of freed memory, which in some circumstances
3340 can cause detection of use-after-free that would previously have
3343 - fix of a longstanding bug that could cause false negatives (missed
3344 errors) in programs doing vector saturated narrowing instructions.
3346 * Helgrind: performance improvements and major memory use reductions,
3347 particularly for large, long running applications which perform many
3348 synchronisation (lock, unlock, etc) events. Plus many smaller
3351 - display of locksets for both threads involved in a race
3353 - general improvements in formatting/clarity of error messages
3355 - addition of facilities and documentation regarding annotation
3356 of thread safe reference counted C++ classes
3358 - new flag --check-stack-refs=no|yes [yes], to disable race checking
3359 on thread stacks (a performance hack)
3361 - new flag --free-is-write=no|yes [no], to enable detection of races
3362 where one thread accesses heap memory but another one frees it,
3363 without any coordinating synchronisation event
3365 * DRD: enabled XML output; added support for delayed thread deletion
3366 in order to detect races that occur close to the end of a thread
3367 (--join-list-vol); fixed a memory leak triggered by repeated client
3368 memory allocatation and deallocation; improved Darwin support.
3370 * exp-ptrcheck: this tool has been renamed to exp-sgcheck
3372 * exp-sgcheck: this tool has been reduced in scope so as to improve
3373 performance and remove checking that Memcheck does better.
3374 Specifically, the ability to check for overruns for stack and global
3375 arrays is unchanged, but the ability to check for overruns of heap
3376 blocks has been removed. The tool has accordingly been renamed to
3377 exp-sgcheck ("Stack and Global Array Checking").
3379 * ==================== OTHER CHANGES ====================
3381 * GDB server: Valgrind now has an embedded GDB server. That means it
3382 is possible to control a Valgrind run from GDB, doing all the usual
3383 things that GDB can do (single stepping, breakpoints, examining
3384 data, etc). Tool-specific functionality is also available. For
3385 example, it is possible to query the definedness state of variables
3386 or memory from within GDB when running Memcheck; arbitrarily large
3387 memory watchpoints are supported, etc. To use the GDB server, start
3388 Valgrind with the flag --vgdb-error=0 and follow the on-screen
3391 * Improved support for unfriendly self-modifying code: a new option
3392 --smc-check=all-non-file is available. This adds the relevant
3393 consistency checks only to code that originates in non-file-backed
3394 mappings. In effect this confines the consistency checking only to
3395 code that is or might be JIT generated, and avoids checks on code
3396 that must have been compiled ahead of time. This significantly
3397 improves performance on applications that generate code at run time.
3399 * It is now possible to build a working Valgrind using Clang-2.9 on
3402 * new client requests VALGRIND_{DISABLE,ENABLE}_ERROR_REPORTING.
3403 These enable and disable error reporting on a per-thread, and
3404 nestable, basis. This is useful for hiding errors in particularly
3405 troublesome pieces of code. The MPI wrapper library (libmpiwrap.c)
3406 now uses this facility.
3408 * Added the --mod-funcname option to cg_diff.
3410 * ==================== FIXED BUGS ====================
3412 The following bugs have been fixed or resolved. Note that "n-i-bz"
3413 stands for "not in bugzilla" -- that is, a bug that was reported to us
3414 but never got a bugzilla entry. We encourage you to file bugs in
3415 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3416 mailing the developers (or mailing lists) directly -- bugs that are
3417 not entered into bugzilla tend to get forgotten about or ignored.
3419 To see details of a given bug, visit
3420 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3421 where XXXXXX is the bug number as listed below.
3423 79311 malloc silly arg warning does not give stack trace
3424 210935 port valgrind.h (not valgrind) to win32 to support client requests
3425 214223 valgrind SIGSEGV on startup gcc 4.4.1 ppc32 (G4) Ubuntu 9.10
3426 243404 Port to zSeries
3427 243935 Helgrind: incorrect handling of ANNOTATE_HAPPENS_BEFORE()/AFTER()
3428 247223 non-x86: Suppress warning: 'regparm' attribute directive ignored
3429 250101 huge "free" memory usage due to m_mallocfree.c fragmentation
3430 253206 Some fixes for the faultstatus testcase
3431 255223 capget testcase fails when running as root
3432 256703 xlc_dbl_u32.c testcase broken
3433 256726 Helgrind tests have broken inline asm
3434 259977 == 214223 (Valgrind segfaults doing __builtin_longjmp)
3435 264800 testcase compile failure on zseries
3436 265762 make public VEX headers compilable by G++ 3.x
3437 265771 assertion in jumps.c (r11523) fails with glibc-2.3
3438 266753 configure script does not give the user the option to not use QtCore
3439 266931 gen_insn_test.pl is broken
3440 266961 ld-linux.so.2 i?86-linux strlen issues
3441 266990 setns instruction causes false positive
3442 267020 Make directory for temporary files configurable at run-time.
3443 267342 == 267997 (segmentation fault on Mac OS 10.6)
3444 267383 Assertion 'vgPlain_strlen(dir) + vgPlain_strlen(file) + 1 < 256' failed
3445 267413 Assertion 'DRD_(g_threadinfo)[tid].synchr_nesting >= 1' failed.
3446 267488 regtest: darwin support for 64-bit build
3447 267552 SIGSEGV (misaligned_stack_error) with DRD, but not with other tools
3448 267630 Add support for IBM Power ISA 2.06 -- stage 1
3449 267769 == 267997 (Darwin: memcheck triggers segmentation fault)
3450 267819 Add client request for informing the core about reallocation
3451 267925 laog data structure quadratic for a single sequence of lock
3452 267968 drd: (vgDrd_thread_set_joinable): Assertion '0 <= (int)tid ..' failed
3453 267997 MacOSX: 64-bit V segfaults on launch when built with Xcode 4.0.1
3454 268513 missed optimizations in fold_Expr
3455 268619 s390x: fpr - gpr transfer facility
3456 268620 s390x: reconsider "long displacement" requirement
3457 268621 s390x: improve IR generation for XC
3458 268715 s390x: FLOGR is not universally available
3459 268792 == 267997 (valgrind seg faults on startup when compiled with Xcode 4)
3460 268930 s390x: MHY is not universally available
3461 269078 arm->IR: unhandled instruction SUB (SP minus immediate/register)
3462 269079 Support ptrace system call on ARM
3463 269144 missing "Bad option" error message
3464 269209 conditional load and store facility (z196)
3465 269354 Shift by zero on x86 can incorrectly clobber CC_NDEP
3466 269641 == 267997 (valgrind segfaults immediately (segmentation fault))
3467 269736 s390x: minor code generation tweaks
3468 269778 == 272986 (valgrind.h: swap roles of VALGRIND_DO_CLIENT_REQUEST() ..)
3469 269863 s390x: remove unused function parameters
3470 269864 s390x: tweak s390_emit_load_cc
3471 269884 == 250101 (overhead for huge blocks exhausts space too soon)
3472 270082 s390x: Make sure to point the PSW address to the next address on SIGILL
3473 270115 s390x: rewrite some testcases
3474 270309 == 267997 (valgrind crash on startup)
3475 270320 add support for Linux FIOQSIZE ioctl() call
3476 270326 segfault while trying to sanitize the environment passed to execle
3477 270794 IBM POWER7 support patch causes regression in none/tests
3478 270851 IBM POWER7 fcfidus instruction causes memcheck to fail
3479 270856 IBM POWER7 xsnmaddadp instruction causes memcheck to fail on 32bit app
3480 270925 hyper-optimized strspn() in /lib64/libc-2.13.so needs fix
3481 270959 s390x: invalid use of R0 as base register
3482 271042 VSX configure check fails when it should not
3483 271043 Valgrind build fails with assembler error on ppc64 with binutils 2.21
3484 271259 s390x: fix code confusion
3485 271337 == 267997 (Valgrind segfaults on MacOS X)
3486 271385 s390x: Implement Ist_MBE
3487 271501 s390x: misc cleanups
3488 271504 s390x: promote likely and unlikely
3489 271579 ppc: using wrong enum type
3490 271615 unhandled instruction "popcnt" (arch=amd10h)
3491 271730 Fix bug when checking ioctls: duplicate check
3492 271776 s390x: provide STFLE instruction support
3493 271779 s390x: provide clock instructions like STCK
3494 271799 Darwin: ioctls without an arg report a memory error
3495 271820 arm: fix type confusion
3496 271917 pthread_cond_timedwait failure leads to not-locked false positive
3497 272067 s390x: fix DISP20 macro
3498 272615 A typo in debug output in mc_leakcheck.c
3499 272661 callgrind_annotate chokes when run from paths containing regex chars
3500 272893 amd64->IR: 0x66 0xF 0x38 0x2B 0xC1 0x66 0xF 0x7F == (closed as dup)
3501 272955 Unhandled syscall error for pwrite64 on ppc64 arch
3502 272967 make documentation build-system more robust
3503 272986 Fix gcc-4.6 warnings with valgrind.h
3504 273318 amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38 (missing PCMPxSTRx case)
3505 273318 unhandled PCMPxSTRx case: vex amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38
3506 273431 valgrind segfaults in evalCfiExpr (debuginfo.c:2039)
3507 273465 Callgrind: jumps.c:164 (new_jcc): Assertion '(0 <= jmp) && ...'
3508 273536 Build error: multiple definition of `vgDrd_pthread_cond_initializer'
3509 273640 ppc64-linux: unhandled syscalls setresuid(164) and setresgid(169)
3510 273729 == 283000 (Illegal opcode for SSE2 "roundsd" instruction)
3511 273778 exp-ptrcheck: unhandled sysno == 259
3512 274089 exp-ptrcheck: unhandled sysno == 208
3513 274378 s390x: Various dispatcher tweaks
3514 274447 WARNING: unhandled syscall: 340
3515 274776 amd64->IR: 0x66 0xF 0x38 0x2B 0xC5 0x66
3516 274784 == 267997 (valgrind ls -l results in Segmentation Fault)
3517 274926 valgrind does not build against linux-3
3518 275148 configure FAIL with glibc-2.14
3519 275151 Fedora 15 / glibc-2.14 'make regtest' FAIL
3520 275168 Make Valgrind work for MacOSX 10.7 Lion
3521 275212 == 275284 (lots of false positives from __memcpy_ssse3_back et al)
3522 275278 valgrind does not build on Linux kernel 3.0.* due to silly
3523 275284 Valgrind memcpy/memmove redirection stopped working in glibc 2.14/x86_64
3524 275308 Fix implementation for ppc64 fres instruc
3525 275339 s390x: fix testcase compile warnings
3526 275517 s390x: Provide support for CKSM instruction
3527 275710 s390x: get rid of redundant address mode calculation
3528 275815 == 247894 (Valgrind doesn't know about Linux readahead(2) syscall)
3529 275852 == 250101 (valgrind uses all swap space and is killed)
3530 276784 Add support for IBM Power ISA 2.06 -- stage 3
3531 276987 gdbsrv: fix tests following recent commits
3532 277045 Valgrind crashes with unhandled DW_OP_ opcode 0x2a
3533 277199 The test_isa_2_06_part1.c in none/tests/ppc64 should be a symlink
3534 277471 Unhandled syscall: 340
3535 277610 valgrind crashes in VG_(lseek)(core_fd, phdrs[idx].p_offset, ...)
3536 277653 ARM: support Thumb2 PLD instruction
3537 277663 ARM: NEON float VMUL by scalar incorrect
3538 277689 ARM: tests for VSTn with register post-index are broken
3539 277694 ARM: BLX LR instruction broken in ARM mode
3540 277780 ARM: VMOV.F32 (immediate) instruction is broken
3541 278057 fuse filesystem syscall deadlocks
3542 278078 Unimplemented syscall 280 on ppc32
3543 278349 F_GETPIPE_SZ and F_SETPIPE_SZ Linux fcntl commands
3544 278454 VALGRIND_STACK_DEREGISTER has wrong output type
3545 278502 == 275284 (Valgrind confuses memcpy() and memmove())
3546 278892 gdbsrv: factorize gdb version handling, fix doc and typos
3547 279027 Support for MVCL and CLCL instruction
3548 279027 s390x: Provide support for CLCL and MVCL instructions
3549 279062 Remove a redundant check in the insn selector for ppc.
3550 279071 JDK creates PTEST with redundant REX.W prefix
3551 279212 gdbsrv: add monitor cmd v.info scheduler.
3552 279378 exp-ptrcheck: the 'impossible' happened on mkfifo call
3553 279698 memcheck discards valid-bits for packuswb
3554 279795 memcheck reports uninitialised values for mincore on amd64
3555 279994 Add support for IBM Power ISA 2.06 -- stage 3
3556 280083 mempolicy syscall check errors
3557 280290 vex amd64->IR: 0x66 0xF 0x38 0x28 0xC1 0x66 0xF 0x6F
3558 280710 s390x: config files for nightly builds
3559 280757 /tmp dir still used by valgrind even if TMPDIR is specified
3560 280965 Valgrind breaks fcntl locks when program does mmap
3561 281138 WARNING: unhandled syscall: 340
3562 281241 == 275168 (valgrind useless on Macos 10.7.1 Lion)
3563 281304 == 275168 (Darwin: dyld "cannot load inserted library")
3564 281305 == 275168 (unhandled syscall: unix:357 on Darwin 11.1)
3565 281468 s390x: handle do_clone and gcc clones in call traces
3566 281488 ARM: VFP register corruption
3567 281828 == 275284 (false memmove warning: "Source and destination overlap")
3568 281883 s390x: Fix system call wrapper for "clone".
3569 282105 generalise 'reclaimSuperBlock' to also reclaim splittable superblock
3570 282112 Unhandled instruction bytes: 0xDE 0xD9 0x9B 0xDF (fcompp)
3571 282238 SLES10: make check fails
3572 282979 strcasestr needs replacement with recent(>=2.12) glibc
3573 283000 vex amd64->IR: 0x66 0xF 0x3A 0xA 0xC0 0x9 0xF3 0xF
3574 283243 Regression in ppc64 memcheck tests
3575 283325 == 267997 (Darwin: V segfaults on startup when built with Xcode 4.0)
3576 283427 re-connect epoll_pwait syscall on ARM linux
3577 283600 gdbsrv: android: port vgdb.c
3578 283709 none/tests/faultstatus needs to account for page size
3579 284305 filter_gdb needs enhancement to work on ppc64
3580 284384 clang 3.1 -Wunused-value warnings in valgrind.h, memcheck.h
3581 284472 Thumb2 ROR.W encoding T2 not implemented
3582 284621 XML-escape process command line in XML output
3583 n-i-bz cachegrind/callgrind: handle CPUID information for Core iX Intel CPUs
3584 that have non-power-of-2 sizes (also AMDs)
3585 n-i-bz don't be spooked by libraries mashed by elfhack
3586 n-i-bz don't be spooked by libxul.so linked with gold
3587 n-i-bz improved checking for VALGRIND_CHECK_MEM_IS_DEFINED
3589 (3.7.0-TEST1: 27 October 2011, vex r2228, valgrind r12245)
3590 (3.7.0.RC1: 1 November 2011, vex r2231, valgrind r12257)
3591 (3.7.0: 5 November 2011, vex r2231, valgrind r12258)
3595 Release 3.6.1 (16 February 2011)
3596 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3597 3.6.1 is a bug fix release. It adds support for some SSE4
3598 instructions that were omitted in 3.6.0 due to lack of time. Initial
3599 support for glibc-2.13 has been added. A number of bugs causing
3600 crashing or assertion failures have been fixed.
3602 The following bugs have been fixed or resolved. Note that "n-i-bz"
3603 stands for "not in bugzilla" -- that is, a bug that was reported to us
3604 but never got a bugzilla entry. We encourage you to file bugs in
3605 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3606 mailing the developers (or mailing lists) directly -- bugs that are
3607 not entered into bugzilla tend to get forgotten about or ignored.
3609 To see details of a given bug, visit
3610 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3611 where XXXXXX is the bug number as listed below.
3613 188572 Valgrind on Mac should suppress setenv() mem leak
3614 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 (proper FX{SAVE,RSTOR} support)
3615 210481 vex amd64->IR: Assertion `sz == 2 || sz == 4' failed (REX.W POPQ)
3616 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3617 250038 ppc64: Altivec LVSR and LVSL instructions fail their regtest
3618 254420 memory pool tracking broken
3619 254957 Test code failing to compile due to changes in memcheck.h
3620 255009 helgrind/drd: crash on chmod with invalid parameter
3621 255130 readdwarf3.c parse_type_DIE confused by GNAT Ada types
3622 255355 helgrind/drd: crash on threaded programs doing fork
3624 255418 (SSE4.x) rint call compiled with ICC
3625 255822 --gen-suppressions can create invalid files: "too many callers [...]"
3626 255888 closing valgrindoutput tag outputted to log-stream on error
3627 255963 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x9 0xDB 0x0 (ROUNDPD)
3628 255966 Slowness when using mempool annotations
3629 256387 vex x86->IR: 0xD4 0xA 0x2 0x7 (AAD and AAM)
3630 256600 super-optimized strcasecmp() false positive
3631 256669 vex amd64->IR: Unhandled LOOPNEL insn on amd64
3632 256968 (SSE4.x) vex amd64->IR: 0x66 0xF 0x38 0x10 0xD3 0x66 (BLENDVPx)
3633 257011 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0xE 0xFD 0xA0 (PBLENDW)
3634 257063 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x8 0xC0 0x0 (ROUNDPS)
3635 257276 Missing case in memcheck --track-origins=yes
3636 258870 (SSE4.x) Add support for EXTRACTPS SSE 4.1 instruction
3637 261966 (SSE4.x) support for CRC32B and CRC32Q is lacking (also CRC32{W,L})
3638 262985 VEX regression in valgrind 3.6.0 in handling PowerPC VMX
3639 262995 (SSE4.x) crash when trying to valgrind gcc-snapshot (PCMPxSTRx $0)
3640 263099 callgrind_annotate counts Ir improperly [...]
3641 263877 undefined coprocessor instruction on ARMv7
3642 265964 configure FAIL with glibc-2.13
3643 n-i-bz Fix compile error w/ icc-12.x in guest_arm_toIR.c
3644 n-i-bz Docs: fix bogus descriptions for VALGRIND_CREATE_BLOCK et al
3645 n-i-bz Massif: don't assert on shmat() with --pages-as-heap=yes
3646 n-i-bz Bug fixes and major speedups for the exp-DHAT space profiler
3647 n-i-bz DRD: disable --free-is-write due to implementation difficulties
3649 (3.6.1: 16 February 2011, vex r2103, valgrind r11561).
3653 Release 3.6.0 (21 October 2010)
3654 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3655 3.6.0 is a feature release with many significant improvements and the
3656 usual collection of bug fixes.
3658 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3659 PPC64/Linux, X86/Darwin and AMD64/Darwin. Support for recent distros
3660 and toolchain components (glibc 2.12, gcc 4.5, OSX 10.6) has been added.
3662 -------------------------
3664 Here are some highlights. Details are shown further down:
3666 * Support for ARM/Linux.
3668 * Support for recent Linux distros: Ubuntu 10.10 and Fedora 14.
3670 * Support for Mac OS X 10.6, both 32- and 64-bit executables.
3672 * Support for the SSE4.2 instruction set.
3674 * Enhancements to the Callgrind profiler, including the ability to
3675 handle CPUs with three levels of cache.
3677 * A new experimental heap profiler, DHAT.
3679 * A huge number of bug fixes and small enhancements.
3681 -------------------------
3683 Here are details of the above changes, together with descriptions of
3684 many other changes, and a list of fixed bugs.
3686 * ================== PLATFORM CHANGES =================
3688 * Support for ARM/Linux. Valgrind now runs on ARMv7 capable CPUs
3689 running Linux. It is known to work on Ubuntu 10.04, Ubuntu 10.10,
3690 and Maemo 5, so you can run Valgrind on your Nokia N900 if you want.
3692 This requires a CPU capable of running the ARMv7-A instruction set
3693 (Cortex A5, A8 and A9). Valgrind provides fairly complete coverage
3694 of the user space instruction set, including ARM and Thumb integer
3695 code, VFPv3, NEON and V6 media instructions. The Memcheck,
3696 Cachegrind and Massif tools work properly; other tools work to
3699 * Support for recent Linux distros (Ubuntu 10.10 and Fedora 14), along
3700 with support for recent releases of the underlying toolchain
3701 components, notably gcc-4.5 and glibc-2.12.
3703 * Support for Mac OS X 10.6, both 32- and 64-bit executables. 64-bit
3704 support also works much better on OS X 10.5, and is as solid as
3707 * Support for the SSE4.2 instruction set. SSE4.2 is supported in
3708 64-bit mode. In 32-bit mode, support is only available up to and
3709 including SSSE3. Some exceptions: SSE4.2 AES instructions are not
3710 supported in 64-bit mode, and 32-bit mode does in fact support the
3711 bare minimum SSE4 instructions to needed to run programs on Mac OS X
3712 10.6 on 32-bit targets.
3714 * Support for IBM POWER6 cpus has been improved. The Power ISA up to
3715 and including version 2.05 is supported.
3717 * ==================== TOOL CHANGES ====================
3719 * Cachegrind has a new processing script, cg_diff, which finds the
3720 difference between two profiles. It's very useful for evaluating
3721 the performance effects of a change in a program.
3723 Related to this change, the meaning of cg_annotate's (rarely-used)
3724 --threshold option has changed; this is unlikely to affect many
3725 people, if you do use it please see the user manual for details.
3727 * Callgrind now can do branch prediction simulation, similar to
3728 Cachegrind. In addition, it optionally can count the number of
3729 executed global bus events. Both can be used for a better
3730 approximation of a "Cycle Estimation" as derived event (you need to
3731 update the event formula in KCachegrind yourself).
3733 * Cachegrind and Callgrind now refer to the LL (last-level) cache
3734 rather than the L2 cache. This is to accommodate machines with
3735 three levels of caches -- if Cachegrind/Callgrind auto-detects the
3736 cache configuration of such a machine it will run the simulation as
3737 if the L2 cache isn't present. This means the results are less
3738 likely to match the true result for the machine, but
3739 Cachegrind/Callgrind's results are already only approximate, and
3740 should not be considered authoritative. The results are still
3741 useful for giving a general idea about a program's locality.
3743 * Massif has a new option, --pages-as-heap, which is disabled by
3744 default. When enabled, instead of tracking allocations at the level
3745 of heap blocks (as allocated with malloc/new/new[]), it instead
3746 tracks memory allocations at the level of memory pages (as mapped by
3747 mmap, brk, etc). Each mapped page is treated as its own block.
3748 Interpreting the page-level output is harder than the heap-level
3749 output, but this option is useful if you want to account for every
3750 byte of memory used by a program.
3752 * DRD has two new command-line options: --free-is-write and
3753 --trace-alloc. The former allows to detect reading from already freed
3754 memory, and the latter allows tracing of all memory allocations and
3757 * DRD has several new annotations. Custom barrier implementations can
3758 now be annotated, as well as benign races on static variables.
3760 * DRD's happens before / happens after annotations have been made more
3761 powerful, so that they can now also be used to annotate e.g. a smart
3762 pointer implementation.
3764 * Helgrind's annotation set has also been drastically improved, so as
3765 to provide to users a general set of annotations to describe locks,
3766 semaphores, barriers and condition variables. Annotations to
3767 describe thread-safe reference counted heap objects have also been
3770 * Memcheck has a new command-line option, --show-possibly-lost, which
3771 is enabled by default. When disabled, the leak detector will not
3772 show possibly-lost blocks.
3774 * A new experimental heap profiler, DHAT (Dynamic Heap Analysis Tool),
3775 has been added. DHAT keeps track of allocated heap blocks, and also
3776 inspects every memory reference to see which block (if any) is being
3777 accessed. This gives a lot of insight into block lifetimes,
3778 utilisation, turnover, liveness, and the location of hot and cold
3779 fields. You can use DHAT to do hot-field profiling.
3781 * ==================== OTHER CHANGES ====================
3783 * Improved support for unfriendly self-modifying code: the extra
3784 overhead incurred by --smc-check=all has been reduced by
3785 approximately a factor of 5 as compared with 3.5.0.
3787 * Ability to show directory names for source files in error messages.
3788 This is combined with a flexible mechanism for specifying which
3789 parts of the paths should be shown. This is enabled by the new flag
3792 * A new flag, --require-text-symbol, which will stop the run if a
3793 specified symbol is not found it a given shared object when it is
3794 loaded into the process. This makes advanced working with function
3795 intercepting and wrapping safer and more reliable.
3797 * Improved support for the Valkyrie GUI, version 2.0.0. GUI output
3798 and control of Valgrind is now available for the tools Memcheck and
3799 Helgrind. XML output from Valgrind is available for Memcheck,
3800 Helgrind and exp-Ptrcheck.
3802 * More reliable stack unwinding on amd64-linux, particularly in the
3803 presence of function wrappers, and with gcc-4.5 compiled code.
3805 * Modest scalability (performance improvements) for massive
3806 long-running applications, particularly for those with huge amounts
3809 * Support for analyzing programs running under Wine with has been
3810 improved. The header files <valgrind/valgrind.h>,
3811 <valgrind/memcheck.h> and <valgrind/drd.h> can now be used in
3812 Windows-programs compiled with MinGW or one of the Microsoft Visual
3815 * A rare but serious error in the 64-bit x86 CPU simulation was fixed.
3816 The 32-bit simulator was not affected. This did not occur often,
3817 but when it did would usually crash the program under test.
3820 * A large number of bugs were fixed. These are shown below.
3822 * A number of bugs were investigated, and were candidates for fixing,
3823 but are not fixed in 3.6.0, due to lack of developer time. They may
3824 get fixed in later releases. They are:
3826 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 0x24 0x49 (FXSAVE64)
3827 212419 false positive "lock order violated" (A+B vs A)
3828 213685 Undefined value propagates past dependency breaking instruction
3829 216837 Incorrect instrumentation of NSOperationQueue on Darwin
3830 237920 valgrind segfault on fork failure
3831 242137 support for code compiled by LLVM-2.8
3832 242423 Another unknown Intel cache config value
3833 243232 Inconsistent Lock Orderings report with trylock
3834 243483 ppc: callgrind triggers VEX assertion failure
3835 243935 Helgrind: implementation of ANNOTATE_HAPPENS_BEFORE() is wrong
3836 244677 Helgrind crash hg_main.c:616 (map_threads_lookup): Assertion
3838 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3839 249435 Analyzing wine programs with callgrind triggers a crash
3840 250038 ppc64: Altivec lvsr and lvsl instructions fail their regtest
3841 250065 Handling large allocations
3842 250101 huge "free" memory usage due to m_mallocfree.c
3843 "superblocks fragmentation"
3844 251569 vex amd64->IR: 0xF 0x1 0xF9 0x8B 0x4C 0x24 (RDTSCP)
3845 252091 Callgrind on ARM does not detect function returns correctly
3846 252600 [PATCH] Allow lhs to be a pointer for shl/shr
3847 254420 memory pool tracking broken
3848 n-i-bz support for adding symbols for JIT generated code
3851 The following bugs have been fixed or resolved. Note that "n-i-bz"
3852 stands for "not in bugzilla" -- that is, a bug that was reported to us
3853 but never got a bugzilla entry. We encourage you to file bugs in
3854 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3855 mailing the developers (or mailing lists) directly -- bugs that are
3856 not entered into bugzilla tend to get forgotten about or ignored.
3858 To see details of a given bug, visit
3859 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3860 where XXXXXX is the bug number as listed below.
3862 135264 dcbzl instruction missing
3864 153699 Valgrind should report unaligned reads with movdqa
3866 190429 Valgrind reports lost of errors in ld.so
3867 with x86_64 2.9.90 glibc
3868 197266 valgrind appears to choke on the xmms instruction
3870 197988 Crash when demangling very large symbol names
3871 202315 unhandled syscall: 332 (inotify_init1)
3872 203256 Add page-level profiling to Massif
3873 205093 dsymutil=yes needs quotes, locking (partial fix)
3874 205241 Snow Leopard 10.6 support (partial fix)
3875 206600 Leak checker fails to upgrade indirect blocks when their
3876 parent becomes reachable
3877 210935 port valgrind.h (not valgrind) to win32 so apps run under
3878 wine can make client requests
3879 211410 vex amd64->IR: 0x15 0xFF 0xFF 0x0 0x0 0x89
3880 within Linux ip-stack checksum functions
3881 212335 unhandled instruction bytes: 0xF3 0xF 0xBD 0xC0
3883 213685 Undefined value propagates past dependency breaking instruction
3885 215914 Valgrind inserts bogus empty environment variable
3887 219538 adjtimex syscall wrapper wrong in readonly adjtime mode
3888 222545 shmat fails under valgind on some arm targets
3889 222560 ARM NEON support
3892 232509 Docs build fails with formatting inside <title></title> elements
3894 235642 [PATCH] syswrap-linux.c: support evdev EVIOCG* ioctls
3895 236546 vex x86->IR: 0x66 0xF 0x3A 0xA
3896 237202 vex amd64->IR: 0xF3 0xF 0xB8 0xC0 0x49 0x3B
3897 237371 better support for VALGRIND_MALLOCLIKE_BLOCK
3898 237485 symlink (syscall 57) is not supported on Mac OS
3899 237723 sysno == 101 exp-ptrcheck: the 'impossible' happened:
3901 238208 is_just_below_ESP doesn't take into account red-zone
3902 238345 valgrind passes wrong $0 when executing a shell script
3903 238679 mq_timedreceive syscall doesn't flag the reception buffer
3905 238696 fcntl command F_DUPFD_CLOEXEC not supported
3906 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3907 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3908 238745 3.5.0 Make fails on PPC Altivec opcodes, though configure
3910 239992 vex amd64->IR: 0x48 0xF 0xC4 0xC1 0x0 0x48
3916 242606 unhandled syscall: setegid (in Ptrcheck)
3917 242814 Helgrind "Impossible has happened" during
3918 QApplication::initInstance();
3919 243064 Valgrind attempting to read debug information from iso
3920 243270 Make stack unwinding in Valgrind wrappers more reliable
3921 243884 exp-ptrcheck: the 'impossible happened: unhandled syscall
3922 sysno = 277 (mq_open)
3923 244009 exp-ptrcheck unknown syscalls in analyzing lighttpd
3924 244493 ARM VFP d16-d31 registers support
3925 244670 add support for audit_session_self syscall on Mac OS 10.6
3926 244921 The xml report of helgrind tool is not well format
3927 244923 In the xml report file, the <preamble> not escape the
3928 xml char, eg '<','&','>'
3929 245535 print full path names in plain text reports
3930 245925 x86-64 red zone handling problem
3931 246258 Valgrind not catching integer underruns + new [] s
3932 246311 reg/reg cmpxchg doesn't work on amd64
3933 246549 unhandled syscall unix:277 while testing 32-bit Darwin app
3934 246888 Improve Makefile.vex.am
3935 247510 [OS X 10.6] Memcheck reports unaddressable bytes passed
3936 to [f]chmod_extended
3937 247526 IBM POWER6 (ISA 2.05) support is incomplete
3938 247561 Some leak testcases fails due to reachable addresses in
3940 247875 sizeofIRType to handle Ity_I128
3941 247894 [PATCH] unhandled syscall sys_readahead
3942 247980 Doesn't honor CFLAGS passed to configure
3943 248373 darwin10.supp is empty in the trunk
3944 248822 Linux FIBMAP ioctl has int parameter instead of long
3945 248893 [PATCH] make readdwarf.c big endianess safe to enable
3946 unwinding on big endian systems
3947 249224 Syscall 336 not supported (SYS_proc_info)
3949 249775 Incorrect scheme for detecting NEON capabilities of host CPU
3950 249943 jni JVM init fails when using valgrind
3951 249991 Valgrind incorrectly declares AESKEYGENASSIST support
3953 249996 linux/arm: unhandled syscall: 181 (__NR_pwrite64)
3954 250799 frexp$fenv_access_off function generates SIGILL
3955 250998 vex x86->IR: unhandled instruction bytes: 0x66 0x66 0x66 0x2E
3956 251251 support pclmulqdq insn
3957 251362 valgrind: ARM: attach to debugger either fails or provokes
3959 251674 Unhandled syscall 294
3962 254257 Add support for debugfiles found by build-id
3963 254550 [PATCH] Implement DW_ATE_UTF (DWARF4)
3964 254646 Wrapped functions cause stack misalignment on OS X
3965 (and possibly Linux)
3966 254556 ARM: valgrinding anything fails with SIGSEGV for 0xFFFF0FA0
3968 (3.6.0: 21 October 2010, vex r2068, valgrind r11471).
3972 Release 3.5.0 (19 August 2009)
3973 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3974 3.5.0 is a feature release with many significant improvements and the
3975 usual collection of bug fixes. The main improvement is that Valgrind
3976 now works on Mac OS X.
3978 This release supports X86/Linux, AMD64/Linux, PPC32/Linux, PPC64/Linux
3979 and X86/Darwin. Support for recent distros and toolchain components
3980 (glibc 2.10, gcc 4.5) has been added.
3982 -------------------------
3984 Here is a short summary of the changes. Details are shown further
3987 * Support for Mac OS X (10.5.x).
3989 * Improvements and simplifications to Memcheck's leak checker.
3991 * Clarification and simplifications in various aspects of Valgrind's
3994 * XML output for Helgrind and Ptrcheck.
3996 * Performance and stability improvements for Helgrind and DRD.
3998 * Genuinely atomic support for x86/amd64/ppc atomic instructions.
4000 * A new experimental tool, BBV, useful for computer architecture
4003 * Improved Wine support, including ability to read Windows PDB
4006 -------------------------
4008 Here are details of the above changes, followed by descriptions of
4009 many other minor changes, and a list of fixed bugs.
4012 * Valgrind now runs on Mac OS X. (Note that Mac OS X is sometimes
4013 called "Darwin" because that is the name of the OS core, which is the
4014 level that Valgrind works at.)
4018 - It requires OS 10.5.x (Leopard). Porting to 10.4.x is not planned
4019 because it would require work and 10.4 is only becoming less common.
4021 - 32-bit programs on x86 and AMD64 (a.k.a x86-64) machines are supported
4022 fairly well. For 10.5.x, 32-bit programs are the default even on
4023 64-bit machines, so it handles most current programs.
4025 - 64-bit programs on x86 and AMD64 (a.k.a x86-64) machines are not
4026 officially supported, but simple programs at least will probably work.
4027 However, start-up is slow.
4029 - PowerPC machines are not supported.
4031 Things that don't work:
4033 - The Ptrcheck tool.
4035 - Objective-C garbage collection.
4039 - If you have Rogue Amoeba's "Instant Hijack" program installed,
4040 Valgrind will fail with a SIGTRAP at start-up. See
4041 https://bugs.kde.org/show_bug.cgi?id=193917 for details and a
4046 - You will likely find --dsymutil=yes a useful option, as error
4047 messages may be imprecise without it.
4049 - Mac OS X support is new and therefore will be less robust than the
4050 Linux support. Please report any bugs you find.
4052 - Threaded programs may run more slowly than on Linux.
4054 Many thanks to Greg Parker for developing this port over several years.
4057 * Memcheck's leak checker has been improved.
4059 - The results for --leak-check=summary now match the summary results
4060 for --leak-check=full. Previously they could differ because
4061 --leak-check=summary counted "indirectly lost" blocks and
4062 "suppressed" blocks as "definitely lost".
4064 - Blocks that are only reachable via at least one interior-pointer,
4065 but are directly pointed to by a start-pointer, were previously
4066 marked as "still reachable". They are now correctly marked as
4069 - The default value for the --leak-resolution option has been
4070 changed from "low" to "high". In general, this means that more
4071 leak reports will be produced, but each leak report will describe
4072 fewer leaked blocks.
4074 - With --leak-check=full, "definitely lost" and "possibly lost"
4075 leaks are now considered as proper errors, ie. they are counted
4076 for the "ERROR SUMMARY" and affect the behaviour of
4077 --error-exitcode. These leaks are not counted as errors if
4078 --leak-check=summary is specified, however.
4080 - Documentation for the leak checker has been improved.
4083 * Various aspects of Valgrind's text output have changed.
4085 - Valgrind's start-up message has changed. It is shorter but also
4086 includes the command being run, which makes it easier to use
4087 --trace-children=yes. An example:
4089 - Valgrind's shut-down messages have also changed. This is most
4090 noticeable with Memcheck, where the leak summary now occurs before
4091 the error summary. This change was necessary to allow leaks to be
4092 counted as proper errors (see the description of the leak checker
4093 changes above for more details). This was also necessary to fix a
4094 longstanding bug in which uses of suppressions against leaks were
4095 not "counted", leading to difficulties in maintaining suppression
4096 files (see https://bugs.kde.org/show_bug.cgi?id=186790).
4098 - Behavior of -v has changed. In previous versions, -v printed out
4099 a mixture of marginally-user-useful information, and tool/core
4100 statistics. The statistics printing has now been moved to its own
4101 flag, --stats=yes. This means -v is less verbose and more likely
4102 to convey useful end-user information.
4104 - The format of some (non-XML) stack trace entries has changed a
4105 little. Previously there were six possible forms:
4107 0x80483BF: really (a.c:20)
4108 0x80483BF: really (in /foo/a.out)
4110 0x80483BF: (within /foo/a.out)
4111 0x80483BF: ??? (a.c:20)
4114 The third and fourth of these forms have been made more consistent
4115 with the others. The six possible forms are now:
4117 0x80483BF: really (a.c:20)
4118 0x80483BF: really (in /foo/a.out)
4119 0x80483BF: really (in ???)
4120 0x80483BF: ??? (in /foo/a.out)
4121 0x80483BF: ??? (a.c:20)
4124 Stack traces produced when --xml=yes is specified are different
4128 * Helgrind and Ptrcheck now support XML output, so they can be used
4129 from GUI tools. Also, the XML output mechanism has been
4132 - The XML format has been overhauled and generalised, so it is more
4133 suitable for error reporting tools in general. The Memcheck
4134 specific aspects of it have been removed. The new format, which
4135 is an evolution of the old format, is described in
4136 docs/internals/xml-output-protocol4.txt.
4138 - Memcheck has been updated to use the new format.
4140 - Helgrind and Ptrcheck are now able to emit output in this format.
4142 - The XML output mechanism has been overhauled. XML is now output
4143 to its own file descriptor, which means that:
4145 * Valgrind can output text and XML independently.
4147 * The longstanding problem of XML output being corrupted by
4148 unexpected un-tagged text messages is solved.
4150 As before, the destination for text output is specified using
4151 --log-file=, --log-fd= or --log-socket=.
4153 As before, XML output for a tool is enabled using --xml=yes.
4155 Because there's a new XML output channel, the XML output
4156 destination is now specified by --xml-file=, --xml-fd= or
4159 Initial feedback has shown this causes some confusion. To
4160 clarify, the two envisaged usage scenarios are:
4162 (1) Normal text output. In this case, do not specify --xml=yes
4163 nor any of --xml-file=, --xml-fd= or --xml-socket=.
4165 (2) XML output. In this case, specify --xml=yes, and one of
4166 --xml-file=, --xml-fd= or --xml-socket= to select the XML
4167 destination, one of --log-file=, --log-fd= or --log-socket=
4168 to select the destination for any remaining text messages,
4169 and, importantly, -q.
4171 -q makes Valgrind completely silent on the text channel,
4172 except in the case of critical failures, such as Valgrind
4173 itself segfaulting, or failing to read debugging information.
4174 Hence, in this scenario, it suffices to check whether or not
4175 any output appeared on the text channel. If yes, then it is
4176 likely to be a critical error which should be brought to the
4177 attention of the user. If no (the text channel produced no
4178 output) then it can be assumed that the run was successful.
4180 This allows GUIs to make the critical distinction they need to
4181 make (did the run fail or not?) without having to search or
4182 filter the text output channel in any way.
4184 It is also recommended to use --child-silent-after-fork=yes in
4188 * Improvements and changes in Helgrind:
4190 - XML output, as described above
4192 - Checks for consistent association between pthread condition
4193 variables and their associated mutexes are now performed.
4195 - pthread_spinlock functions are supported.
4197 - Modest performance improvements.
4199 - Initial (skeletal) support for describing the behaviour of
4200 non-POSIX synchronisation objects through ThreadSanitizer
4201 compatible ANNOTATE_* macros.
4203 - More controllable tradeoffs between performance and the level of
4204 detail of "previous" accesses in a race. There are now three
4207 * --history-level=full. This is the default, and was also the
4208 default in 3.4.x. It shows both stacks involved in a race, but
4209 requires a lot of memory and can be very slow in programs that
4210 do many inter-thread synchronisation events.
4212 * --history-level=none. This only shows the later stack involved
4213 in a race. This can be much faster than --history-level=full,
4214 but makes it much more difficult to find the other access
4215 involved in the race.
4217 The new intermediate setting is
4219 * --history-level=approx
4221 For the earlier (other) access, two stacks are presented. The
4222 earlier access is guaranteed to be somewhere in between the two
4223 program points denoted by those stacks. This is not as useful
4224 as showing the exact stack for the previous access (as per
4225 --history-level=full), but it is better than nothing, and it's
4226 almost as fast as --history-level=none.
4229 * New features and improvements in DRD:
4231 - The error messages printed by DRD are now easier to interpret.
4232 Instead of using two different numbers to identify each thread
4233 (Valgrind thread ID and DRD thread ID), DRD does now identify
4234 threads via a single number (the DRD thread ID). Furthermore
4235 "first observed at" information is now printed for all error
4236 messages related to synchronization objects.
4238 - Added support for named semaphores (sem_open() and sem_close()).
4240 - Race conditions between pthread_barrier_wait() and
4241 pthread_barrier_destroy() calls are now reported.
4243 - Added support for custom allocators through the macros
4244 VALGRIND_MALLOCLIKE_BLOCK() VALGRIND_FREELIKE_BLOCK() (defined in
4245 in <valgrind/valgrind.h>). An alternative for these two macros is
4246 the new client request VG_USERREQ__DRD_CLEAN_MEMORY (defined in
4249 - Added support for annotating non-POSIX synchronization objects
4250 through several new ANNOTATE_*() macros.
4252 - OpenMP: added support for the OpenMP runtime (libgomp) included
4253 with gcc versions 4.3.0 and 4.4.0.
4257 - Added two new command-line options (--first-race-only and
4258 --segment-merging-interval).
4261 * Genuinely atomic support for x86/amd64/ppc atomic instructions
4263 Valgrind will now preserve (memory-access) atomicity of LOCK-
4264 prefixed x86/amd64 instructions, and any others implying a global
4265 bus lock. Ditto for PowerPC l{w,d}arx/st{w,d}cx. instructions.
4267 This means that Valgrinded processes will "play nicely" in
4268 situations where communication with other processes, or the kernel,
4269 is done through shared memory and coordinated with such atomic
4270 instructions. Prior to this change, such arrangements usually
4271 resulted in hangs, races or other synchronisation failures, because
4272 Valgrind did not honour atomicity of such instructions.
4275 * A new experimental tool, BBV, has been added. BBV generates basic
4276 block vectors for use with the SimPoint analysis tool, which allows
4277 a program's overall behaviour to be approximated by running only a
4278 fraction of it. This is useful for computer architecture
4279 researchers. You can run BBV by specifying --tool=exp-bbv (the
4280 "exp-" prefix is short for "experimental"). BBV was written by
4284 * Modestly improved support for running Windows applications under
4285 Wine. In particular, initial support for reading Windows .PDB debug
4286 information has been added.
4289 * A new Memcheck client request VALGRIND_COUNT_LEAK_BLOCKS has been
4290 added. It is similar to VALGRIND_COUNT_LEAKS but counts blocks
4294 * The Valgrind client requests VALGRIND_PRINTF and
4295 VALGRIND_PRINTF_BACKTRACE have been changed slightly. Previously,
4296 the string was always printed immediately on its own line. Now, the
4297 string will be added to a buffer but not printed until a newline is
4298 encountered, or other Valgrind output is printed (note that for
4299 VALGRIND_PRINTF_BACKTRACE, the back-trace itself is considered
4300 "other Valgrind output"). This allows you to use multiple
4301 VALGRIND_PRINTF calls to build up a single output line, and also to
4302 print multiple output lines with a single request (by embedding
4303 multiple newlines in the string).
4306 * The graphs drawn by Massif's ms_print program have changed slightly:
4308 - The half-height chars '.' and ',' are no longer drawn, because
4309 they are confusing. The --y option can be used if the default
4310 y-resolution is not high enough.
4312 - Horizontal lines are now drawn after the top of a snapshot if
4313 there is a gap until the next snapshot. This makes it clear that
4314 the memory usage has not dropped to zero between snapshots.
4317 * Something that happened in 3.4.0, but wasn't clearly announced: the
4318 option --read-var-info=yes can be used by some tools (Memcheck,
4319 Helgrind and DRD). When enabled, it causes Valgrind to read DWARF3
4320 variable type and location information. This makes those tools
4321 start up more slowly and increases memory consumption, but
4322 descriptions of data addresses in error messages become more
4326 * exp-Omega, an experimental instantaneous leak-detecting tool, was
4327 disabled in 3.4.0 due to a lack of interest and maintenance,
4328 although the source code was still in the distribution. The source
4329 code has now been removed from the distribution. For anyone
4330 interested, the removal occurred in SVN revision r10247.
4333 * Some changes have been made to the build system.
4335 - VEX/ is now integrated properly into the build system. This means
4336 that dependency tracking within VEX/ now works properly, "make
4337 install" will work without requiring "make" before it, and
4338 parallel builds (ie. 'make -j') now work (previously a
4339 .NOTPARALLEL directive was used to serialize builds, ie. 'make -j'
4340 was effectively ignored).
4342 - The --with-vex configure option has been removed. It was of
4343 little use and removing it simplified the build system.
4345 - The location of some install files has changed. This should not
4346 affect most users. Those who might be affected:
4348 * For people who use Valgrind with MPI programs, the installed
4349 libmpiwrap.so library has moved from
4350 $(INSTALL)/<platform>/libmpiwrap.so to
4351 $(INSTALL)/libmpiwrap-<platform>.so.
4353 * For people who distribute standalone Valgrind tools, the
4354 installed libraries such as $(INSTALL)/<platform>/libcoregrind.a
4355 have moved to $(INSTALL)/libcoregrind-<platform>.a.
4357 These changes simplify the build system.
4359 - Previously, all the distributed suppression (*.supp) files were
4360 installed. Now, only default.supp is installed. This should not
4361 affect users as the other installed suppression files were not
4362 read; the fact that they were installed was a mistake.
4365 * KNOWN LIMITATIONS:
4367 - Memcheck is unusable with the Intel compiler suite version 11.1,
4368 when it generates code for SSE2-and-above capable targets. This
4369 is because of icc's use of highly optimised inlined strlen
4370 implementations. It causes Memcheck to report huge numbers of
4371 false errors even in simple programs. Helgrind and DRD may also
4374 Versions 11.0 and earlier may be OK, but this has not been
4378 The following bugs have been fixed or resolved. Note that "n-i-bz"
4379 stands for "not in bugzilla" -- that is, a bug that was reported to us
4380 but never got a bugzilla entry. We encourage you to file bugs in
4381 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4382 mailing the developers (or mailing lists) directly -- bugs that are
4383 not entered into bugzilla tend to get forgotten about or ignored.
4385 To see details of a given bug, visit
4386 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
4387 where XXXXXX is the bug number as listed below.
4389 84303 How about a LockCheck tool?
4390 91633 dereference of null ptr in vgPlain_st_basetype
4391 97452 Valgrind doesn't report any pthreads problems
4392 100628 leak-check gets assertion failure when using
4393 VALGRIND_MALLOCLIKE_BLOCK on malloc()ed memory
4394 108528 NPTL pthread cleanup handlers not called
4395 110126 Valgrind 2.4.1 configure.in tramples CFLAGS
4396 110128 mallinfo is not implemented...
4397 110770 VEX: Generated files not always updated when making valgrind
4398 111102 Memcheck: problems with large (memory footprint) applications
4399 115673 Vex's decoder should never assert
4400 117564 False positive: Syscall param clone(child_tidptr) contains
4401 uninitialised byte(s)
4402 119404 executing ssh from inside valgrind fails
4403 133679 Callgrind does not write path names to sources with dwarf debug
4405 135847 configure.in problem with non gnu compilers (and possible fix)
4406 136154 threads.c:273 (vgCallgrind_post_signal): Assertion
4407 '*(vgCallgrind_current_fn_stack.top) == 0' failed.
4408 136230 memcheck reports "possibly lost", should be "still reachable"
4409 137073 NULL arg to MALLOCLIKE_BLOCK causes crash
4410 137904 Valgrind reports a memory leak when using POSIX threads,
4412 139076 valgrind VT_GETSTATE error
4413 142228 complaint of elf_dynamic_do_rela in trivial usage
4414 145347 spurious warning with USBDEVFS_REAPURB
4415 148441 (wine) can't find memory leak in Wine, win32 binary
4417 148742 Leak-check fails assert on exit
4418 149878 add (proper) check for calloc integer overflow
4419 150606 Call graph is broken when using callgrind control
4420 152393 leak errors produce an exit code of 0. I need some way to
4421 cause leak errors to result in a nonzero exit code.
4422 157154 documentation (leak-resolution doc speaks about num-callers
4423 def=4) + what is a loss record
4424 159501 incorrect handling of ALSA ioctls
4425 162020 Valgrinding an empty/zero-byte file crashes valgrind
4426 162482 ppc: Valgrind crashes while reading stabs information
4427 162718 x86: avoid segment selector 0 in sys_set_thread_area()
4428 163253 (wine) canonicaliseSymtab forgot some fields in DiSym
4429 163560 VEX/test_main.c is missing from valgrind-3.3.1
4430 164353 malloc_usable_size() doesn't return a usable size
4431 165468 Inconsistent formatting in memcheck manual -- please fix
4432 169505 main.c:286 (endOfInstr):
4433 Assertion 'ii->cost_offset == *cost_offset' failed
4434 177206 Generate default.supp during compile instead of configure
4435 177209 Configure valt_load_address based on arch+os
4436 177305 eventfd / syscall 323 patch lost
4437 179731 Tests fail to build because of inlining of non-local asm labels
4438 181394 helgrind: libhb_core.c:3762 (msm_write): Assertion
4439 'ordxx == POrd_EQ || ordxx == POrd_LT' failed.
4440 181594 Bogus warning for empty text segment
4441 181707 dwarf doesn't require enumerations to have name
4442 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4443 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4444 Assertion '!already_present' failed.
4445 185359 exp-ptrcheck: unhandled syscall getresuid()
4446 185794 "WARNING: unhandled syscall: 285" (fallocate) on x86_64
4447 185816 Valgrind is unable to handle debug info for files with split
4448 debug info that are prelinked afterwards
4449 185980 [darwin] unhandled syscall: sem_open
4450 186238 bbToIR_AMD64: disInstr miscalculated next %rip
4451 186507 exp-ptrcheck unhandled syscalls prctl, etc.
4452 186790 Suppression pattern used for leaks are not reported
4453 186796 Symbols with length>200 in suppression files are ignored
4454 187048 drd: mutex PTHREAD_PROCESS_SHARED attribute missinterpretation
4455 187416 exp-ptrcheck: support for __NR_{setregid,setreuid,setresuid}
4456 188038 helgrind: hg_main.c:926: mk_SHVAL_fail: the 'impossible' happened
4457 188046 bashisms in the configure script
4458 188127 amd64->IR: unhandled instruction bytes: 0xF0 0xF 0xB0 0xA
4459 188161 memcheck: --track-origins=yes asserts "mc_machine.c:672
4460 (get_otrack_shadow_offset_wrk): the 'impossible' happened."
4461 188248 helgrind: pthread_cleanup_push, pthread_rwlock_unlock,
4462 assertion fail "!lock->heldBy"
4463 188427 Add support for epoll_create1 (with patch)
4464 188530 Support for SIOCGSTAMPNS
4465 188560 Include valgrind.spec in the tarball
4466 188572 Valgrind on Mac should suppress setenv() mem leak
4467 189054 Valgrind fails to build because of duplicate non-local asm labels
4468 189737 vex amd64->IR: unhandled instruction bytes: 0xAC
4469 189762 epoll_create syscall not handled (--tool=exp-ptrcheck)
4470 189763 drd assertion failure: s_threadinfo[tid].is_recording
4471 190219 unhandled syscall: 328 (x86-linux)
4472 190391 dup of 181394; see above
4473 190429 Valgrind reports lots of errors in ld.so with x86_64 2.9.90 glibc
4474 190820 No debug information on powerpc-linux
4475 191095 PATCH: Improve usbdevfs ioctl handling
4476 191182 memcheck: VALGRIND_LEAK_CHECK quadratic when big nr of chunks
4478 191189 --xml=yes should obey --gen-suppressions=all
4479 191192 syslog() needs a suppression on macosx
4480 191271 DARWIN: WARNING: unhandled syscall: 33554697 a.k.a.: 265
4481 191761 getrlimit on MacOSX
4482 191992 multiple --fn-skip only works sometimes; dependent on order
4483 192634 V. reports "aspacem sync_check_mapping_callback:
4484 segment mismatch" on Darwin
4485 192954 __extension__ missing on 2 client requests
4486 194429 Crash at start-up with glibc-2.10.1 and linux-2.6.29
4487 194474 "INSTALL" file has different build instructions than "README"
4488 194671 Unhandled syscall (sem_wait?) from mac valgrind
4489 195069 memcheck: reports leak (memory still reachable) for
4491 195169 drd: (vgDrd_barrier_post_wait):
4492 Assertion 'r->sg[p->post_iteration]' failed.
4493 195268 valgrind --log-file doesn't accept ~/...
4494 195838 VEX abort: LibVEX_N_SPILL_BYTES too small for CPUID boilerplate
4495 195860 WARNING: unhandled syscall: unix:223
4496 196528 need a error suppression for pthread_rwlock_init under os x?
4497 197227 Support aio_* syscalls on Darwin
4498 197456 valgrind should reject --suppressions=(directory)
4499 197512 DWARF2 CFI reader: unhandled CFI instruction 0:10
4500 197591 unhandled syscall 27 (mincore)
4501 197793 Merge DCAS branch to the trunk == 85756, 142103
4502 197794 Avoid duplicate filenames in Vex
4503 197898 make check fails on current SVN
4504 197901 make check fails also under exp-ptrcheck in current SVN
4505 197929 Make --leak-resolution=high the default
4506 197930 Reduce spacing between leak reports
4507 197933 Print command line of client at start-up, and shorten preamble
4508 197966 unhandled syscall 205 (x86-linux, --tool=exp-ptrcheck)
4509 198395 add BBV to the distribution as an experimental tool
4510 198624 Missing syscalls on Darwin: 82, 167, 281, 347
4511 198649 callgrind_annotate doesn't cumulate counters
4512 199338 callgrind_annotate sorting/thresholds are broken for all but Ir
4513 199977 Valgrind complains about an unrecognized instruction in the
4514 atomic_incs test program
4515 200029 valgrind isn't able to read Fedora 12 debuginfo
4516 200760 darwin unhandled syscall: unix:284
4517 200827 DRD doesn't work on Mac OS X
4518 200990 VG_(read_millisecond_timer)() does not work correctly
4519 201016 Valgrind does not support pthread_kill() on Mac OS
4520 201169 Document --read-var-info
4521 201323 Pre-3.5.0 performance sanity checking
4522 201384 Review user manual for the 3.5.0 release
4523 201585 mfpvr not implemented on ppc
4524 201708 tests failing because x86 direction flag is left set
4525 201757 Valgrind doesn't handle any recent sys_futex additions
4526 204377 64-bit valgrind can not start a shell script
4527 (with #!/path/to/shell) if the shell is a 32-bit executable
4528 n-i-bz drd: fixed assertion failure triggered by mutex reinitialization.
4529 n-i-bz drd: fixed a bug that caused incorrect messages to be printed
4530 about memory allocation events with memory access tracing enabled
4531 n-i-bz drd: fixed a memory leak triggered by vector clock deallocation
4533 (3.5.0: 19 Aug 2009, vex r1913, valgrind r10846).
4537 Release 3.4.1 (28 February 2009)
4538 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4539 3.4.1 is a bug-fix release that fixes some regressions and assertion
4540 failures in debug info reading in 3.4.0, most notably incorrect stack
4541 traces on amd64-linux on older (glibc-2.3 based) systems. Various
4542 other debug info problems are also fixed. A number of bugs in the
4543 exp-ptrcheck tool introduced in 3.4.0 have been fixed.
4545 In view of the fact that 3.4.0 contains user-visible regressions
4546 relative to 3.3.x, upgrading to 3.4.1 is recommended. Packagers are
4547 encouraged to ship 3.4.1 in preference to 3.4.0.
4549 The fixed bugs are as follows. Note that "n-i-bz" stands for "not in
4550 bugzilla" -- that is, a bug that was reported to us but never got a
4551 bugzilla entry. We encourage you to file bugs in bugzilla
4552 (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than mailing the
4553 developers (or mailing lists) directly -- bugs that are not entered
4554 into bugzilla tend to get forgotten about or ignored.
4556 n-i-bz Fix various bugs reading icc-11 generated debug info
4557 n-i-bz Fix various bugs reading gcc-4.4 generated debug info
4558 n-i-bz Preliminary support for glibc-2.10 / Fedora 11
4559 n-i-bz Cachegrind and Callgrind: handle non-power-of-two cache sizes,
4560 so as to support (eg) 24k Atom D1 and Core2 with 3/6/12MB L2.
4561 179618 exp-ptrcheck crashed / exit prematurely
4562 179624 helgrind: false positive races with pthread_create and
4563 recv/open/close/read
4564 134207 pkg-config output contains @VG_PLATFORM@
4565 176926 floating point exception at valgrind startup with PPC 440EPX
4566 181594 Bogus warning for empty text segment
4567 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant rex prefixes)
4568 181707 Dwarf3 doesn't require enumerations to have name
4569 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4570 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4571 Assertion '!already_present' failed.
4572 185359 exp-ptrcheck unhandled syscall getresuid()
4574 (3.4.1.RC1: 24 Feb 2008, vex r1884, valgrind r9253).
4575 (3.4.1: 28 Feb 2008, vex r1884, valgrind r9293).
4579 Release 3.4.0 (2 January 2009)
4580 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4581 3.4.0 is a feature release with many significant improvements and the
4582 usual collection of bug fixes. This release supports X86/Linux,
4583 AMD64/Linux, PPC32/Linux and PPC64/Linux. Support for recent distros
4584 (using gcc 4.4, glibc 2.8 and 2.9) has been added.
4586 3.4.0 brings some significant tool improvements. Memcheck can now
4587 report the origin of uninitialised values, the thread checkers
4588 Helgrind and DRD are much improved, and we have a new experimental
4589 tool, exp-Ptrcheck, which is able to detect overruns of stack and
4590 global arrays. In detail:
4592 * Memcheck is now able to track the origin of uninitialised values.
4593 When it reports an uninitialised value error, it will try to show
4594 the origin of the value, as either a heap or stack allocation.
4595 Origin tracking is expensive and so is not enabled by default. To
4596 use it, specify --track-origins=yes. Memcheck's speed will be
4597 essentially halved, and memory usage will be significantly
4598 increased. Nevertheless it can drastically reduce the effort
4599 required to identify the root cause of uninitialised value errors,
4600 and so is often a programmer productivity win, despite running more
4603 * A version (1.4.0) of the Valkyrie GUI, that works with Memcheck in
4604 3.4.0, will be released shortly.
4606 * Helgrind's race detection algorithm has been completely redesigned
4607 and reimplemented, to address usability and scalability concerns:
4609 - The new algorithm has a lower false-error rate: it is much less
4610 likely to report races that do not really exist.
4612 - Helgrind will display full call stacks for both accesses involved
4613 in a race. This makes it easier to identify the root causes of
4616 - Limitations on the size of program that can run have been removed.
4618 - Performance has been modestly improved, although that is very
4621 - Direct support for Qt4 threading has been added.
4623 - pthread_barriers are now directly supported.
4625 - Helgrind works well on all supported Linux targets.
4627 * The DRD thread debugging tool has seen major improvements:
4629 - Greatly improved performance and significantly reduced memory
4632 - Support for several major threading libraries (Boost.Thread, Qt4,
4633 glib, OpenMP) has been added.
4635 - Support for atomic instructions, POSIX semaphores, barriers and
4636 reader-writer locks has been added.
4638 - Works now on PowerPC CPUs too.
4640 - Added support for printing thread stack usage at thread exit time.
4642 - Added support for debugging lock contention.
4644 - Added a manual for Drd.
4646 * A new experimental tool, exp-Ptrcheck, has been added. Ptrcheck
4647 checks for misuses of pointers. In that sense it is a bit like
4648 Memcheck. However, Ptrcheck can do things Memcheck can't: it can
4649 detect overruns of stack and global arrays, it can detect
4650 arbitrarily far out-of-bounds accesses to heap blocks, and it can
4651 detect accesses heap blocks that have been freed a very long time
4652 ago (millions of blocks in the past).
4654 Ptrcheck currently works only on x86-linux and amd64-linux. To use
4655 it, use --tool=exp-ptrcheck. A simple manual is provided, as part
4656 of the main Valgrind documentation. As this is an experimental
4657 tool, we would be particularly interested in hearing about your
4658 experiences with it.
4660 * exp-Omega, an experimental instantaneous leak-detecting tool, is no
4661 longer built by default, although the code remains in the repository
4662 and the tarball. This is due to three factors: a perceived lack of
4663 users, a lack of maintenance, and concerns that it may not be
4664 possible to achieve reliable operation using the existing design.
4666 * As usual, support for the latest Linux distros and toolchain
4667 components has been added. It should work well on Fedora Core 10,
4668 OpenSUSE 11.1 and Ubuntu 8.10. gcc-4.4 (in its current pre-release
4669 state) is supported, as is glibc-2.9. The C++ demangler has been
4670 updated so as to work well with C++ compiled by even the most recent
4673 * You can now use frame-level wildcards in suppressions. This was a
4674 frequently-requested enhancement. A line "..." in a suppression now
4675 matches zero or more frames. This makes it easier to write
4676 suppressions which are precise yet insensitive to changes in
4679 * 3.4.0 adds support on x86/amd64 for the SSSE3 instruction set.
4681 * Very basic support for IBM Power6 has been added (64-bit processes only).
4683 * Valgrind is now cross-compilable. For example, it is possible to
4684 cross compile Valgrind on an x86/amd64-linux host, so that it runs
4685 on a ppc32/64-linux target.
4687 * You can set the main thread's stack size at startup using the
4688 new --main-stacksize= flag (subject of course to ulimit settings).
4689 This is useful for running apps that need a lot of stack space.
4691 * The limitation that you can't use --trace-children=yes together
4692 with --db-attach=yes has been removed.
4694 * The following bugs have been fixed. Note that "n-i-bz" stands for
4695 "not in bugzilla" -- that is, a bug that was reported to us but
4696 never got a bugzilla entry. We encourage you to file bugs in
4697 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4698 mailing the developers (or mailing lists) directly.
4700 n-i-bz Make return types for some client requests 64-bit clean
4701 n-i-bz glibc 2.9 support
4702 n-i-bz ignore unsafe .valgrindrc's (CVE-2008-4865)
4703 n-i-bz MPI_Init(0,0) is valid but libmpiwrap.c segfaults
4704 n-i-bz Building in an env without gdb gives bogus gdb attach
4705 92456 Tracing the origin of uninitialised memory
4706 106497 Valgrind does not demangle some C++ template symbols
4708 151612 Suppression with "..." (frame-level wildcards in .supp files)
4709 156404 Unable to start oocalc under memcheck on openSUSE 10.3 (64-bit)
4710 159285 unhandled syscall:25 (stime, on x86-linux)
4711 159452 unhandled ioctl 0x8B01 on "valgrind iwconfig"
4712 160954 ppc build of valgrind crashes with illegal instruction (isel)
4713 160956 mallinfo implementation, w/ patch
4714 162092 Valgrind fails to start gnome-system-monitor
4715 162819 malloc_free_fill test doesn't pass on glibc2.8 x86
4716 163794 assertion failure with "--track-origins=yes"
4717 163933 sigcontext.err and .trapno must be set together
4718 163955 remove constraint !(--db-attach=yes && --trace-children=yes)
4719 164476 Missing kernel module loading system calls
4720 164669 SVN regression: mmap() drops posix file locks
4721 166581 Callgrind output corruption when program forks
4722 167288 Patch file for missing system calls on Cell BE
4723 168943 unsupported scas instruction pentium
4724 171645 Unrecognised instruction (MOVSD, non-binutils encoding)
4725 172417 x86->IR: 0x82 ...
4726 172563 amd64->IR: 0xD9 0xF5 - fprem1
4727 173099 .lds linker script generation error
4728 173177 [x86_64] syscalls: 125/126/179 (capget/capset/quotactl)
4729 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant prefixes)
4731 174908 --log-file value not expanded correctly for core file
4732 175044 Add lookup_dcookie for amd64
4733 175150 x86->IR: 0xF2 0xF 0x11 0xC1 (movss non-binutils encoding)
4735 Developer-visible changes:
4737 * Valgrind's debug-info reading machinery has been majorly overhauled.
4738 It can now correctly establish the addresses for ELF data symbols,
4739 which is something that has never worked properly before now.
4741 Also, Valgrind can now read DWARF3 type and location information for
4742 stack and global variables. This makes it possible to use the
4743 framework to build tools that rely on knowing the type and locations
4744 of stack and global variables, for example exp-Ptrcheck.
4746 Reading of such information is disabled by default, because most
4747 tools don't need it, and because it is expensive in space and time.
4748 However, you can force Valgrind to read it, using the
4749 --read-var-info=yes flag. Memcheck, Helgrind and DRD are able to
4750 make use of such information, if present, to provide source-level
4751 descriptions of data addresses in the error messages they create.
4753 (3.4.0.RC1: 24 Dec 2008, vex r1878, valgrind r8882).
4754 (3.4.0: 3 Jan 2009, vex r1878, valgrind r8899).