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 - `--cache-sim=no` is now the default. The cache simulation is old and
102 unlikely to match any real modern machine. This means only the `Ir`
103 event are gathered by default, but that is by far the most useful
105 - `cg_annotate`, `cg_diff`, and `cg_merge` have been rewritten in
106 Python. As a result, they all have more flexible command line
107 argument handling, e.g. supporting `--show-percs` and
108 `--no-show-percs` forms as well as the existing `--show-percs=yes`
109 and `--show-percs=no`.
110 - `cg_annotate` has some functional changes.
111 - It's much faster, e.g. 3-4x on common cases.
112 - It now supports diffing (with `--diff`, `--mod-filename`, and
113 `--mod-funcname`) and merging (by passing multiple data files).
114 - It now provides more information at the file and function level.
115 There are now "File:function" and "Function:file" sections. These
116 are very useful for programs that use inlining a lot.
117 - Support for user-annotated files and the `-I`/`--include` option
118 has been removed, because it was of little use and blocked other
120 - The `--auto` option is renamed `--annotate`, though the old
121 `--auto=yes`/`--auto=no` forms are still supported.
122 - `cg_diff` and `cg_merge` are now deprecated, because `cg_annotate`
123 now does a better job of diffing and merging.
124 - The Cachegrind output file format has changed very slightly, but in
125 ways nobody is likely to notice.
128 - Valgrind now contains python code that defines GDB callgrind
129 front end monitor commands. See CORE CHANGES.
132 - Valgrind now contains python code that defines GDB massif
133 front end monitor commands. See CORE CHANGES.
135 * ==================== FIXED BUGS ====================
137 The following bugs have been fixed or resolved. Note that "n-i-bz"
138 stands for "not in bugzilla" -- that is, a bug that was reported to us
139 but never got a bugzilla entry. We encourage you to file bugs in
140 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
141 than mailing the developers (or mailing lists) directly -- bugs that
142 are not entered into bugzilla tend to get forgotten about or ignored.
144 170510 Don't warn about ioctl of size 0 without direction hint
145 241072 List tools in --help output
146 327548 false positive while destroying mutex
147 382034 Testcases build fixes for musl
148 351857 confusing error message about valid command line option
149 374596 inconsistent RDTSCP support on x86_64
150 392331 Spurious lock not held error from inside pthread_cond_timedwait
151 397083 Likely false positive "uninitialised value(s)" for __wmemchr_avx2 and __wmemcmp_avx2_movbe
152 400793 pthread_rwlock_timedwrlock false positive
153 419054 Unhandled syscall getcpu on arm32
154 433873 openat2 syscall unimplemented on Linux
155 434057 Add stdio mode to valgrind's gdbserver
156 435441 valgrind fails to interpose malloc on musl 1.2.2 due to weak symbol name and no libc soname
157 436413 Warn about realloc of size zero
158 439685 compiler warning in callgrind/main.c
159 444110 priv/guest_ppc_toIR.c:36198:31: warning: duplicated 'if' condition.
160 444487 hginfo test detects an extra lock inside data symbol "_rtld_local"
161 444488 Use glibc.pthread.stack_cache_size tunable
162 444568 drd/tests/pth_barrier_thr_cr fails on Fedora 38
163 445743 "The impossible happened: mutex is locked simultaneously by two threads"
164 while using mutexes with priority inheritance and signals
165 449309 Missing loopback device ioctl(s)
166 459476 vgdb: allow address reuse to avoid "address already in use" errorsuse" errors
167 460356 s390: Sqrt32Fx4 -- cannot reduce tree
168 462830 WARNING: unhandled amd64-freebsd syscall: 474
169 463027 broken check for MPX instruction support in assembler
170 464103 Enhancement: add a client request to DHAT to mark memory to be histogrammed
171 464476 Firefox fails to start under Valgrind
172 464609 Valgrind memcheck should support Linux pidfd_open
173 464680 Show issues caused by memory policies like selinux deny_execmem
174 464859 Build failures with GCC-13 (drd tsan_unittest)
175 464969 D language demangling
176 465435 m_libcfile.c:66 (vgPlain_safe_fd): Assertion 'newfd >= VG_(fd_hard_limit)' failed.
177 466104 aligned_alloc problems, part 1
178 467036 Add time cost statistics for Regtest
179 467482 Build failure on aarch64 Alpine
180 467714 fdleak_* and rlimit tests fail when parent process has more than
181 64 descriptors opened
182 467839 Gdbserver: Improve compatibility of library directory name
183 468401 [PATCH] Add a style file for clang-format
184 468556 Build failure for vgdb
185 468606 build: remove "Valgrind relies on GCC" check/output
186 n-i-bz FreeBSD rfork syscall fail with EINVAL or ENOSYS rather than VG_(unimplemented)
188 To see details of a given bug, visit
189 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
190 where XXXXXX is the bug number as listed above.
192 (3.21.0.RC1: 14 Apr 2023)
193 (3.21.0.RC2: 21 Apr 2023)
195 Release 3.20.0 (24 Oct 2022)
196 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
198 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
199 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
200 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
201 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
202 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
203 AMD64/macOS 10.13 and nanoMIPS/Linux.
205 * ==================== CORE CHANGES ===================
207 * The option "--vgdb-stop-at=event1,event2,..." accepts the new value abexit.
208 This indicates to invoke gdbserver when your program exits abnormally
209 (i.e. with a non zero exit code).
210 * Fix Rust v0 name demangling.
211 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
212 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
213 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with
215 * The option --enable-debuginfod=<no|yes> [default: yes] has been added on
217 * More DWARF5 support as generated by clang14.
219 * ==================== FIXED BUGS ====================
221 The following bugs have been fixed or resolved. Note that "n-i-bz"
222 stands for "not in bugzilla" -- that is, a bug that was reported to us
223 but never got a bugzilla entry. We encourage you to file bugs in
224 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
225 than mailing the developers (or mailing lists) directly -- bugs that
226 are not entered into bugzilla tend to get forgotten about or ignored.
228 131186 writev reports error in (vector[...])
229 434764 iconv_open causes ld.so v2.28+ to use optimised strncmp
230 446754 Improve error codes from alloc functions under memcheck
231 452274 memcheck crashes with Assertion 'sci->status.what == SsIdle' failed
232 452779 Valgrind fails to build on FreeBSD 13.0 with llvm-devel (15.0.0)
233 453055 shared_timed_mutex drd test fails with "Lock shared failed" message
234 453602 Missing command line option to enable/disable debuginfod
235 452802 Handle lld 9+ split RW PT_LOAD segments correctly
236 454040 s390x: False-positive memcheck:cond in memmem on arch13 systems
237 456171 [PATCH] FreeBSD: Don't record address errors when accessing the 'kern.ps_strings' sysctl struct
238 n-i-bz Implement vgdb invoker on FreeBSD
239 458845 PowerPC: The L field for the dcbf and sync instruction should be
241 458915 Remove register cache to fix 458915 gdbserver causes wrong syscall return
242 459031 Documentation on --error-exitcode incomplete
243 459477 XERROR messages lacks ending '\n' in vgdb
244 462007 Implicit int in none/tests/faultstatus.c
246 To see details of a given bug, visit
247 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
248 where XXXXXX is the bug number as listed above.
250 (3.20.0.RC1: 20 Oct 2022)
253 Release 3.19.0 (11 Apr 2022)
254 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
256 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
257 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
258 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
259 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
260 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
261 AMD64/macOS 10.13 and nanoMIPS/Linux.
263 * ==================== CORE CHANGES ===================
265 * Fix Rust v0 name demangling.
266 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
267 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
268 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with HardenedBSD
270 * ================== PLATFORM CHANGES =================
273 - ignore the "v8.x" architecture levels, only look at actual CPU features
274 present. Fixes mismatch detected between RDMA and atomics features
275 preventing startup on some QEMU configurations.
276 - Implement LD{,A}XP and ST{,L}XP
277 - Fix incorrect code emitted for doubleword CAS.
280 - Fix sys_ipc semtimedop syscall
281 - Fix VFLRX and WFLRX instructions
282 - Fix EXRL instruction with negative offset
285 - Reimplement the vbpermq instruction support to generate less Iops and
286 avoid overflowing internal buffers.
287 - Fix checking for scv support to avoid "Facility 'SCV' unavailable (12),
288 exception" messages in dmsg.
289 - Fix setting condition code for Vector Compare quad word instructions.
290 - Fix fix lxsibzx, lxsihzx and lxsihzx instructions so they only load
291 their respective sized data.
292 - Fix the prefixed stq instruction in PC relative mode.
294 * ==================== TOOL CHANGES ===================
297 - Speed up --track-origins=yes for large (in the range of hundreds to
298 thousands of megabytes) mmap/munmaps.
300 - Several fixes for new versions of libstd++ using new posix try_lock
303 * ==================== FIXED BUGS ====================
305 The following bugs have been fixed or resolved. Note that "n-i-bz"
306 stands for "not in bugzilla" -- that is, a bug that was reported to us
307 but never got a bugzilla entry. We encourage you to file bugs in
308 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
309 than mailing the developers (or mailing lists) directly -- bugs that
310 are not entered into bugzilla tend to get forgotten about or ignored.
312 403802 leak_cpp_interior fails with some reachable blocks different than expected
313 435732 memcheck/tests/leak_cpp_interior fails with gcc11
314 444242 s390x: Valgrind crashes on EXRL with negative offset
315 444399 arm64: unhandled instruction 0xC87F2D89 (LD{,A}XP and ST{,L}XP).
317 444481 gdb_server test failures on s390x
318 444495 dhat/tests/copy fails on s390x
319 444552 memcheck/tests/sem fails on s390x with glibc 2.34
320 444571 PPC, fix the lxsibzx and lxsihzx so they only load their respective
322 444836 PPC, pstq instruction for R=1 is not storing to the correct address.
323 444925 fexecve syscall wrapper not properly implemented
324 445032 valgrind/memcheck crash with SIGSEGV when SIGVTALRM timer used and
326 445211 Fix out of tree builds
327 445300 [PATCH] Fix building tests with Musl
328 445011 SIGCHLD is sent when valgrind uses debuginfod-find
329 445354 arm64 backend: incorrect code emitted for doubleword CAS
330 445415 arm64 front end: alignment checks missing for atomic instructions
331 445504 Using C++ condition_variable results in bogus "mutex is locked simultaneously by two threads" warning
332 445607 Unhandled amd64-freebsd syscall: 247
333 445668 Inline stack frame generation is broken for Rust binaries
334 445916 Demangle Rust v0 symbols with .llvm suffix
335 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false positives
336 446138 DRD/Helgrind with std::timed_mutex::try_lock_until false positives
337 446281 Add a DRD suppression for fwrite
338 446103 Memcheck: `--track-origins=yes` causes extreme slowdowns for large mmap/munmap
339 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false
340 446251 TARGET_SIGNAL_THR added to enum target_signal
341 446823 FreeBSD - missing syscalls when using libzm4
342 447991 s390x: Valgrind indicates illegal instruction on wflrx
343 447995 Valgrind segfault on power10 due to hwcap checking code
344 449483 Powerpc: vcmpgtsq., vcmpgtuq,, vcmpequq. instructions not setting the
345 condition code correctly.
346 449672 ppc64 --track-origins=yes failures because of bad cmov addHRegUse
347 449838 sigsegv liburing the 'impossible' happened for io_uring_setup
348 450025 Powerc: ACC file not implemented as a logical overlay of the VSR
350 450437 Warn for execve syscall with argv or argv[0] being NULL
351 450536 Powerpc: valgrind throws 'facility scv unavailable exception'
352 451626 Syscall param bpf(attr->raw_tracepoint.name) points to unaddressable byte(s)
353 451827 [ppc64le] VEX temporary storage exhausted with several vbpermq instructions
354 451843 valgrind fails to start on a FreeBSD system which enforces W^X
356 To see details of a given bug, visit
357 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
358 where XXXXXX is the bug number as listed above.
360 (3.19.0.RC1: 02 Apr 2022)
361 (3.19.0.RC2: 08 Apr 2022)
364 Release 3.18.0 (15 Oct 2021)
365 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
367 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
368 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
369 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
370 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
371 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
372 AMD64/macOS 10.13 and nanoMIPS/Linux.
374 * ==================== CORE CHANGES ===================
376 * The libiberty demangler has been updated, which brings support for
377 Rust v0 name demangling. [Update: alas, due to a bug, this support
378 isn't working in 3.18.0.]
380 * __libc_freeres isn't called anymore after the program recieves a
381 fatal signal. Causing some internal glibc resources to hang around,
382 but preventing any crashes after the program has ended.
384 * The DWARF reader is now very much faster at startup when just
385 --read-inline-info=yes (the default in most cases) is given.
387 * glibc 2.34, which moved various functions from libpthread.so into
388 libc.so, is now supported.
390 * ================== PLATFORM CHANGES =================
394 - v8.2 scalar and vector FABD, FACGE, FACGT and FADD.
395 - v8.2 FP compare & conditional compare instructions.
396 - Zero variants of v8.2 FP compare instructions.
400 - Support the miscellaneous-instruction-extensions facility 3 and
401 the vector-enhancements facility 2. This enables programs
402 compiled with "-march=arch13" or "-march=z15" to be executed
407 - ISA 3.1 support is now complete
408 - ISA 3.0 support for the darn instruction added.
409 - ISA 3.0 support for the vector system call instruction scv added.
410 - ISA 3.0 support for the copy, paste and cpabort instructions added.
412 * Support for X86/FreeBSD and AMD64/FreeBSD has been added.
414 * ==================== OTHER CHANGES ====================
416 * Memcheck on amd64: minor fixes to remove some false positive
419 * ==================== FIXED BUGS ====================
421 The following bugs have been fixed or resolved. Note that "n-i-bz"
422 stands for "not in bugzilla" -- that is, a bug that was reported to us
423 but never got a bugzilla entry. We encourage you to file bugs in
424 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
425 than mailing the developers (or mailing lists) directly -- bugs that
426 are not entered into bugzilla tend to get forgotten about or ignored.
428 208531 [PATCH]: FreeBSD support for valgrind
429 368960 WARNING: unhandled amd64-linux syscall: 163 (acct)
430 407589 [Linux] Add support for C11 aligned_alloc() and GNU reallocarray()
431 423963 Error in child thread when CLONE_PIDFD is used
432 426148 crash with "impossible happened" when running BPF CO-RE programs
433 429375 PPC ISA 3.1 support is missing, part 9
434 431157 PPC_FEATURE2_SCV needs to be masked in AT_HWCAP2
435 431306 Update demangler to support Rust v0 name mangling
436 432387 s390x: z15 instructions support
437 433437 FreeBSD support, part 1
438 433438 FreeBSD support, part 2
439 433439 FreeBSD support, part 3
440 433469 FreeBSD support, part 4
441 433473 FreeBSD support, part 5
442 433477 FreeBSD support, part 6
443 433479 FreeBSD support, part 7
444 433504 FreeBSD support, part 8
445 433506 FreeBSD support, part 9
446 433507 FreeBSD support, part 10
447 433508 FreeBSD support, part 11
448 433510 FreeBSD support, part 12
449 433801 PPC ISA 3.1 support is missing, part 10 (ISA 3.1 support complete)
450 433863 s390x: memcheck/tests/s390x/{cds,cs,csg} failures
451 434296 s390x: False-positive memcheck diagnostics from vector string
453 434840 PPC64 darn instruction not supported
454 435665 PPC ISA 3.0 copy, paste, cpabort instructions are not supported
455 435908 valgrind tries to fetch from deubginfod for files which already
456 have debug information
457 438871 unhandled instruction bytes: 0xF3 0x49 0xF 0x6F 0x9C 0x24 0x60 0x2
458 439046 valgrind is unusably large when linked with lld
459 439090 Implement close_range(2)
460 439326 Valgrind 3.17.0 won't compile with Intel 2021 oneAPI compilers
461 439590 glibc-2.34 breaks suppressions against obj:*/lib*/libc-2.*so*
462 440670 unhandled ppc64le-linux syscall: 252 statfs64 and 253 fstatfs64
463 440906 Fix impossible constraint issue in P10 testcase.
464 441512 Remove a unneeded / unnecessary prefix check.
465 441534 Update the expected output for test_isa_3_1_VRT.
466 442061 very slow execution under Fedora 34 (readdwarf3)
467 443031 Gcc -many change requires explicit .machine directives
468 443033 Add support for the ISA 3.0 mcrxrx instruction
469 443034 Sraw, srawi, srad, sradi, mfs
470 443178 Powerpc, test jm-mfspr expected output needs to be updated.
471 443179 Need new test for the lxvx and stxvx instructions on ISA 2.07 and
473 443180 The subnormal test and the ISA 3.0 test generate compiler warnings
474 443314 In the latest GIT version, Valgrind with "--trace-flags" crashes
476 443605 Don't call final_tidyup (__libc_freeres) on FatalSignal
478 To see details of a given bug, visit
479 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
480 where XXXXXX is the bug number as listed below.
482 (3.18.0.RC1: 12 Oct 2021)
483 (3.18.0: 15 Oct 2021)
487 Release 3.17.0 (19 Mar 2021)
488 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
490 3.17.0 fixes a number of bugs and adds some functional changes: support for
491 GCC 11, Clang 11, DWARF5 debuginfo, the 'debuginfod' debuginfo server, and
492 some new instructions for Arm64, S390 and POWER. There are also some tool
495 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
496 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
497 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
498 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
499 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
501 * ==================== CORE CHANGES ===================
503 * DWARF version 5 support. Valgrind can now read DWARF version 5 debuginfo as
506 * Valgrind now supports debuginfod, an HTTP server for distributing ELF/DWARF
507 debugging information. When a debuginfo file cannot be found locally,
508 Valgrind is able to query debuginfod servers for the file using its
509 build-id. See the user manual for more information about debuginfod support.
511 * ================== PLATFORM CHANGES =================
515 - Inaccuracies resulting from double-rounding in the simulation of
516 floating-point multiply-add/subtract instructions have been fixed. These
517 should now behave exactly as the hardware does.
519 - Partial support for the ARM v8.2 instruction set. v8.2 support work is
520 ongoing. Support for the half-word variants of at least the following
521 instructions has been added:
523 FABS <Vd>.<T>, <Vn>.<T>
525 FNEG <Vd>.<T>, <Vn>.<T>
527 FSQRT <Vd>.<T>, <Vn>.<T>
532 - Implement the new instructions/features that were added to z/Architecture
533 with the vector-enhancements facility 1. Also cover the instructions from
534 the vector-packed-decimal facility that are defined outside the chapter
535 "Vector Decimal Instructions", but not the ones from that chapter itself.
537 For a detailed list of newly supported instructions see the updates to
538 `docs/internals/s390-opcodes.csv'.
540 Since the miscellaneous instruction extensions facility 2 was already
541 added in Valgrind 3.16.0, this completes the support necessary to run
542 general programs built with `--march=z14' under Valgrind. The
543 vector-packed-decimal facility is currently not exploited by the standard
544 toolchain and libraries.
548 - Various bug fixes. Fix for the sync field to limit setting just two of
549 the two bits in the L-field. Fix the write size for the stxsibx and
550 stxsihx instructions. Fix the modsw and modsd instructions.
552 - Partial support for ISA 3.1 has been added. Support for the VSX PCV mask
553 instructions, bfloat16 GER instructions, and bfloat16 to/from float 32-bit
554 conversion instructions are still missing.
556 * ==================== TOOL CHANGES ====================
558 * General tool changes
560 - All the tools and their vgpreload libraries are now installed under
561 libexec because they cannot be executed directly and should be run through
562 the valgrind executable. This should be an internal, not user visible,
563 change, but might impact valgrind packagers.
565 - The --track-fds option now respects -q, --quiet and won't output anything
566 if no file descriptors are leaked. It also won't report the standard stdin
567 (0), stdout (1) or stderr (2) descriptors as being leaked with
568 --trace-fds=yes anymore. To track whether the standard file descriptors
569 are still open at the end of the program run use --trace-fds=all.
573 - DHAT has been extended, with two new modes of operation. The new
574 --mode=copy flag triggers copy profiling, which records calls to memcpy,
575 strcpy, and similar functions. The new --mode=ad-hoc flag triggers ad hoc
576 profiling, which records calls to the DHAT_AD_HOC_EVENT client request in
577 the new dhat/dhat.h file. This is useful for learning more about hot code
578 paths. See the user manual for more information about the new modes.
580 - Because of these changes, DHAT's file format has changed. DHAT output
581 files produced with earlier versions of DHAT will not work with this
582 version of DHAT's viewer, and DHAT output files produced with this version
583 of DHAT will not work with earlier versions of DHAT's viewer.
585 * ==================== FIXED BUGS ====================
587 The following bugs have been fixed or resolved. Note that "n-i-bz"
588 stands for "not in bugzilla" -- that is, a bug that was reported to us
589 but never got a bugzilla entry. We encourage you to file bugs in
590 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
591 than mailing the developers (or mailing lists) directly -- bugs that
592 are not entered into bugzilla tend to get forgotten about or ignored.
594 To see details of a given bug, visit
595 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
596 where XXXXXX is the bug number as listed below.
598 140178 open("/proc/self/exe", ...); doesn't quite work
599 140939 --track-fds reports leakage of stdout/in/err and doesn't respect -q
600 217695 malloc/calloc/realloc/memalign failure doesn't set errno to ENOMEM
601 338633 gdbserver_tests/nlcontrolc.vgtest hangs on arm64
602 345077 linux syscall execveat support (linux 3.19)
603 361770 Missing F_ADD_SEALS
604 369029 handle linux syscalls sched_getattr and sched_setattr
605 384729 __libc_freeres inhibits cross-platform valgrind
606 388787 Support for C++17 new/delete
607 391853 Makefile.all.am:L247 and @SOLARIS_UNDEF_LARGESOURCE@ being empty
608 396656 Warnings while reading debug info
609 397605 ioctl FICLONE mishandled
610 401416 Compile failure with openmpi 4.0
611 408663 Suppression file for musl libc
612 404076 s390x: z14 vector instructions not implemented
613 410743 shmat() calls for 32-bit programs fail when running in 64-bit valgrind
614 (actually affected all x86 and nanomips regardless of host bitness)
615 413547 regression test does not check for Arm 64 features.
616 414268 Enable AArch64 feature detection and decoding for v8.x instructions
617 415293 Incorrect call-graph tracking due to new _dl_runtime_resolve_xsave*
618 422174 unhandled instruction bytes: 0x48 0xE9 (REX prefixed JMP instruction)
619 422261 platform selection fails for unqualified client name
620 422623 epoll_ctl warns for uninitialized padding on non-amd64 64bit arches
621 423021 PPC: Add missing ISA 3.0 documentation link and HWCAPS test.
622 423195 PPC ISA 3.1 support is missing, part 1
623 423361 Adds io_uring support on arm64/aarch64 (and all other arches)
624 424012 crash with readv/writev having invalid but not NULL arg2 iovec
625 424298 amd64: Implement RDSEED
626 425232 PPC ISA 3.1 support is missing, part 2
627 425820 Failure to recognize vpcmpeqq as a dependency breaking idiom.
628 426014 arm64: implement fmadd and fmsub as Iop_MAdd/Sub
629 426123 PPC ISA 3.1 support is missing, part 3
630 426144 Fix "condition variable has not been initialized" on Fedora 33.
631 427400 PPC ISA 3.1 support is missing, part 4
632 427401 PPC ISA 3.1 support is missing, part 5
633 427404 PPC ISA 3.1 support is missing, part 6
634 427870 lmw, lswi and related PowerPC insns aren't allowed on ppc64le
635 427787 Support new faccessat2 linux syscall (439)
636 427969 debuginfo section duplicates a section in the main ELF file
637 428035 drd: Unbreak the musl build
638 428648 s390_emit_load_mem panics due to 20-bit offset for vector load
639 428716 cppcheck detects potential leak in VEX/useful/smchash.c
640 428909 helgrind: need to intercept duplicate libc definitions for Fedora 33
641 429352 PPC ISA 3.1 support is missing, part 7
642 429354 PPC ISA 3.1 support is missing, part 8
643 429692 unhandled ppc64le-linux syscall: 147 (getsid)
644 429864 s390x: C++ atomic test_and_set yields false-positive memcheck
646 429952 Errors when building regtest with clang
647 430354 ppc stxsibx and stxsihx instructions write too much data
648 430429 valgrind.h doesn't compile on s390x with clang
649 430485 expr_is_guardable doesn't handle Iex_Qop
650 431556 Complete arm64 FADDP v8.2 instruction support
651 432102 Add support for DWARF5 as produced by GCC11
652 432161 Addition of arm64 v8.2 FADDP, FNEG and FSQRT
653 432381 drd: Process STACK_REGISTER client requests
654 432552 [AArch64] invalid error emitted for pre-decremented byte/hword addresses
655 432672 vg_regtest: test-specific environment variables not reset between tests
656 432809 VEX should support REX.W + POPF
657 432861 PPC modsw and modsd give incorrect results for 1 mod 12
658 432870 gdbserver_tests:nlcontrolc hangs with newest glibc2.33 x86-64
659 432215 Add debuginfod functionality
660 433323 Use pkglibexecdir as vglibdir
661 433500 DRD regtest faulures when libstdc++ and libgcc debuginfo are installed
662 433629 valgrind/README has type "abd" instead of "and"
663 433641 Rust std::sys::unix::fs::try_statx Syscall param fstatat(file_name)
664 433898 arm64: Handle sp, lr, fp as DwReg in CfiExpr
665 434193 GCC 9+ inlined strcmp causes "Conditional jump or move [..] value" report
666 n-i-bz helgrind: If hg_cli__realloc fails, return NULL.
667 n-i-bz arm64 front end: avoid Memcheck false positives relating to CPUID
669 (3.17.0.RC1: 13 Mar 2021)
670 (3.17.0.RC2: 17 Mar 2021)
671 (3.17.0: 19 Mar 2021)
675 Release 3.16.1 (22 June 2020)
676 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
678 3.16.1 fixes two critical bugs discovered after 3.16.0 was frozen. It also
679 fixes character encoding problems in the documentation HTML.
681 422677 PPC sync instruction L field should only be 2 bits in ISA 3.0
682 422715 32-bit x86: vex: the `impossible' happened: expr_is_guardable: unhandled expr
684 (3.16.1, 22 June 2020, 36d6727e1d768333a536f274491e5879cab2c2f7)
688 Release 3.16.0 (27 May 2020)
689 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
691 3.16.0 is a feature release with many improvements and the usual collection of
694 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
695 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
696 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
697 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
698 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
700 * ==================== CORE CHANGES ===================
702 * It is now possible to dynamically change the value of many command line
703 options while your program (or its children) are running under Valgrind.
705 To see the list of dynamically changeable options, run
706 "valgrind --help-dyn-options".
708 You can change the options from the shell by using vgdb to launch
709 the monitor command "v.clo <clo option>...".
710 The same monitor command can be used from a gdb connected
711 to the valgrind gdbserver.
712 Your program can also change the dynamically changeable options using
713 the client request VALGRIND_CLO_CHANGE(option).
715 * ================== PLATFORM CHANGES =================
717 * MIPS: preliminary support for nanoMIPS instruction set has been added.
719 * ==================== TOOL CHANGES ====================
723 - The implicit memcpy done by each call to realloc now counts towards the
724 read and write counts of resized heap blocks, making those counts higher
729 - cg_annotate's --auto and --show-percs options now default to 'yes', because
730 they are usually wanted.
734 - callgrind_annotate's --auto and --show-percs options now default to 'yes',
735 because they are usually wanted.
737 - The command option --collect-systime has been enhanced to specify
738 the unit used to record the elapsed time spent during system calls.
739 The command option now accepts the values no|yes|msec|usec|nsec,
740 where yes is a synonym of msec. When giving the value nsec, the
741 system cpu time of system calls is also recorded.
745 - Several memcheck options are now dynamically changeable.
746 Use valgrind --help-dyn-options to list them.
748 - The release 3.15 introduced a backward incompatible change for
749 some suppression entries related to preadv and pwritev syscalls.
750 When reading a suppression entry using the unsupported 3.14 format,
751 valgrind will now produce a warning to say the suppression entry will not
752 work, and suggest the needed change.
754 - Significantly fewer false positive errors on optimised code generated by
755 Clang and GCC. In particular, Memcheck now deals better with the
756 situation where the compiler will transform C-level "A && B" into "B && A"
757 under certain circumstances (in which the transformation is valid).
758 Handling of integer equality/non-equality checks on partially defined
759 values is also improved on some architectures.
763 - The exprimental Stack and Global Array Checking tool has been removed.
764 It only ever worked on x86 and amd64, and even on those it had a
765 high false positive rate and was slow. An alternative for detecting
766 stack and global array overruns is using the AddressSanitizer (ASAN)
767 facility of the GCC and Clang compilers, which require you to rebuild
768 your code with -fsanitize=address.
770 * ==================== OTHER CHANGES ====================
772 * New and modified GDB server monitor features:
774 - Option -T tells vgdb to output a timestamp in the vgdb information messages.
776 - The gdbserver monitor commands that require an address and an optional
777 length argument now accepts the alternate 'C like' syntax "address[length]".
778 For example, the memcheck command "monitor who_points_at 0x12345678 120"
779 can now also be given as "monitor who_points_at 0x12345678[120]".
781 * ==================== FIXED BUGS ====================
783 The following bugs have been fixed or resolved. Note that "n-i-bz"
784 stands for "not in bugzilla" -- that is, a bug that was reported to us
785 but never got a bugzilla entry. We encourage you to file bugs in
786 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
787 than mailing the developers (or mailing lists) directly -- bugs that
788 are not entered into bugzilla tend to get forgotten about or ignored.
790 To see details of a given bug, visit
791 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
792 where XXXXXX is the bug number as listed below.
794 343099 Linux setns syscall wrapper missing, unhandled syscall: 308
795 == 368923 WARNING: unhandled arm64-linux syscall: 268 (setns)
796 == 369031 WARNING: unhandled amd64-linux syscall: 308 (setns)
797 385386 Assertion failed "szB >= CACHE_ENTRY_SIZE" at m_debuginfo/image.c:517
798 400162 Patch: Guard against __GLIBC_PREREQ for musl libc
799 400593 In Coregrind, use statx for some internal syscalls if [f]stat[64] fail
800 400872 Add nanoMIPS support to Valgrind
801 403212 drd/tests/trylock hangs on FreeBSD
802 404406 s390x: z14 miscellaneous instructions not implemented
803 405201 Incorrect size of struct vki_siginfo on 64-bit Linux architectures
804 406561 mcinfcallWSRU gdbserver_test fails on ppc64
805 406824 Unsupported baseline
806 407218 Add support for the copy_file_range syscall
807 407307 Intercept stpcpy also in ld.so for arm64
808 407376 Update Xen support to 4.12 (4.13, actually) and add more coverage
810 407764 drd cond_post_wait gets wrong (?) condition on s390x z13 system
811 408009 Expose rdrand and f16c even on avx if host cpu supports them
812 408091 Missing pkey syscalls
813 408414 Add support for missing for preadv2 and pwritev2 syscalls
814 409141 Valgrind hangs when SIGKILLed
815 409206 Support for Linux PPS and PTP ioctls
816 409367 exit_group() after signal to thread waiting in futex() causes hangs
817 409429 amd64: recognize 'cmpeq' variants as a dependency breaking idiom
818 409780 References to non-existent configure.in
819 410556 Add support for BLKIO{MIN,OPT} and BLKALIGNOFF ioctls
820 410599 Non-deterministic behaviour of pth_self_kill_15_other test
821 410757 discrepancy for preadv2/pwritev2 syscalls across different versions
822 411134 Allow the user to change a set of command line options during execution
823 411451 amd64->IR of bt/btc/bts/btr with immediate clears zero flag
824 412344 Problem setting mips flags with specific paths
825 412408 unhandled arm-linux syscall: 124 - adjtime - on arm-linux
826 413119 Ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
827 413330 avx-1 test fails on AMD EPYC 7401P 24-Core Processor
828 413603 callgrind_annotate/cg_annotate truncate function names at '#'
829 414565 Specific use case bug found in SysRes VG_(do_sys_sigprocmask)
830 415136 ARMv8.1 Compare-and-Swap instructions are not supported
831 415757 vex x86->IR: 0x66 0xF 0xCE 0x4F (bswapw)
832 416239 valgrind crashes when handling clock_adjtime
833 416285 Use prlimit64 in VG_(getrlimit) and VG_(setrlimit)
834 416286 DRD reports "conflicting load" error on std::mutex::lock()
835 416301 s390x: "compare and signal" not supported
836 416387 finit_module and bpf syscalls are unhandled on arm64
837 416464 Fix false reports for uninitialized memory for PR_CAPBSET_READ/DROP
838 416667 gcc10 ppc64le impossible constraint in 'asm' in test_isa.
839 416753 new 32bit time syscalls for 2038+
840 417075 pwritev(vector[...]) suppression ignored
841 417075 is not fixed, but incompatible supp entries are detected
842 and a warning is produced for these.
843 417187 [MIPS] Conditional branch problem since 'grail' changes
844 417238 Test memcheck/tests/vbit-test fails on mips64 BE
845 417266 Make memcheck/tests/linux/sigqueue usable with musl
846 417281 s390x: /bin/true segfaults with "grail" enabled
847 417427 commit to fix vki_siginfo_t definition created numerous regression
849 417452 s390_insn_store_emit: dst->tag for HRcVec128
850 417578 Add suppressions for glibc DTV leaks
851 417906 clone with CLONE_VFORK and no CLONE_VM fails
852 418004 Grail code additions break ppc64.
853 418435 s390x: spurious "Conditional jump or move depends on uninitialised [..]"
854 418997 s390x: Support Iex_ITE for float and vector types
855 419503 s390x: Avoid modifying registers returned from isel functions
856 421321 gcc10 arm64 build needs __getauxval for linking with libgcc
857 421570 std_mutex fails on Arm v8.1 h/w
858 434035 vgdb might crash if valgrind is killed
859 n-i-bz Fix minor one time leaks in dhat.
860 n-i-bz Add --run-cxx-freeres=no in outer args to avoid inner crashes.
861 n-i-bz Add support for the Linux io_uring system calls
862 n-i-bz sys_statx: don't complain if both |filename| and |buf| are NULL.
863 n-i-bz Fix non-glibc build of test suite with s390x_features
864 n-i-bz MinGW, include/valgrind.h: Fix detection of 64-bit mode
865 423195 PPC ISA 3.1 support is missing, part 1
867 (3.16.0.RC1: 18 May 2020, git 6052ee66a0cf5234e8e2a2b49a8760226bc13b92)
868 (3.16.0.RC2: 19 May 2020, git 940ec1ca69a09f7fdae3e800b7359f85c13c4b37)
869 (3.16.0: 27 May 2020, git bf5e647edb9e96cbd5c57cc944984402eeee296d)
873 Release 3.15.0 (12 April 2019)
874 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
876 3.15.0 is a feature release with many improvements and the usual collection of
879 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
880 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
881 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
882 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
883 support for X86/macOS 10.13 and AMD64/macOS 10.13.
885 * ==================== CORE CHANGES ===================
887 * The XTree Massif output format now makes use of the information obtained
888 when specifying --read-inline-info=yes.
890 * amd64 (x86_64): the RDRAND and F16C insn set extensions are now supported.
892 * ==================== TOOL CHANGES ====================
896 - DHAT been thoroughly overhauled, improved, and given a GUI. As a result,
897 it has been promoted from an experimental tool to a regular tool. Run it
898 with --tool=dhat instead of --tool=exp-dhat.
900 - DHAT now prints only minimal data when the program ends, instead writing
901 the bulk of the profiling data to a file. As a result, the --show-top-n
902 and --sort-by options have been removed.
904 - Profile results can be viewed with the new viewer, dh_view.html. When
905 a run ends, a short message is printed, explaining how to view the result.
907 - See the documentation for more details.
911 - cg_annotate has a new option, --show-percs, which prints percentages next
916 - callgrind_annotate has a new option, --show-percs, which prints percentages
917 next to all event counts.
919 - callgrind_annotate now inserts commas in call counts, and
920 sort the caller/callee lists in the call tree.
924 - The default value for --read-inline-info is now "yes" on
925 Linux/Android/Solaris. It is still "no" on other OS.
929 - The option --xtree-leak=yes (to output leak result in xtree format)
930 automatically activates the option --show-leak-kinds=all, as xtree
931 visualisation tools such as kcachegrind can in any case select what kind
932 of leak to visualise.
934 - There has been further work to avoid false positives. In particular,
935 integer equality on partially defined inputs (C == and !=) is now handled
938 * ==================== OTHER CHANGES ====================
940 * The new option --show-error-list=no|yes displays, at the end of the run, the
941 list of detected errors and the used suppressions. Prior to this change,
942 showing this information could only be done by specifying "-v -v", but that
943 also produced a lot of other possibly-non-useful messages. The option -s is
944 equivalent to --show-error-list=yes.
946 * ==================== FIXED BUGS ====================
948 The following bugs have been fixed or resolved. Note that "n-i-bz"
949 stands for "not in bugzilla" -- that is, a bug that was reported to us
950 but never got a bugzilla entry. We encourage you to file bugs in
951 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
952 than mailing the developers (or mailing lists) directly -- bugs that
953 are not entered into bugzilla tend to get forgotten about or ignored.
955 To see details of a given bug, visit
956 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
957 where XXXXXX is the bug number as listed below.
959 385411 s390x: z13 vector floating-point instructions not implemented
960 397187 z13 vector register support for vgdb gdbserver
961 398183 Vex errors with _mm256_shuffle_epi8/vpshufb
962 398870 Please add support for instruction vcvtps2ph
963 399287 amd64 front end: Illegal Instruction vcmptrueps
964 399301 Use inlined frames in Massif XTree output.
965 399322 Improve callgrind_annotate output
966 399444 VEX/priv/guest_s390_toIR.c:17407: (style) Mismatching assignment [..]
967 400164 helgrind test encounters mips x-compiler warnings and assembler error
968 400490 s390x: VRs allocated as if separate from FPRs
969 400491 s390x: Operand of LOCH treated as unsigned integer
970 400975 Compile error: error: '-mips64r2' conflicts with the other architecture
971 options, which specify a mips64 processor
972 401112 LLVM 5.0 generates comparison against partially initialized data
973 401277 More bugs in z13 support
974 401454 Add a --show-percs option to cg_annotate and callgrind_annotate.
975 401578 drd: crashes sometimes on fork()
976 401627 memcheck errors with glibc avx2 optimized wcsncmp
977 401822 none/tests/ppc64/jm-vmx fails and produces assembler warnings
978 401827 none/tests/ppc64/test_isa_2_06_part3 failure on ppc64le (xvrsqrtesp)
979 401828 none/tests/ppc64/test_isa_2_06_part1 failure on ppc64le (fcfids and
981 402006 mark helper regs defined in final_tidyup before freeres_wrapper call
982 402048 WARNING: unhandled ppc64[be|le]-linux syscall: 26 (ptrace)
983 402123 invalid assembler opcodes for mips32r2
984 402134 assertion fail in mc_translate.c (noteTmpUsesIn) Iex_VECRET on arm64
985 402327 Warning: DWARF2 CFI reader: unhandled DW_OP_ opcode 0x13 (DW_OP_drop)
986 402341 drd/tests/tsan_thread_wrappers_pthread.h:369: suspicious code ?
987 402351 mips64 libvexmultiarch_test fails on s390x
989 402395 coregrind/vgdb-invoker-solaris.c: 2 * poor error checking
990 402480 Do not use %rsp in clobber list
991 402481 vbit-test fails on x86 for Iop_CmpEQ64 iselInt64Expr Sar64
992 402515 Implement new option --show-error-list=no|yes / -s
993 402519 POWER 3.0 addex instruction incorrectly implemented
994 402781 Redo the cache used to process indirect branch targets
995 403123 vex amd64->IR:0xF3 0x48 0xF 0xAE 0xD3 (wrfsbase)
996 403552 s390x: wrong facility bit checked for vector facility
997 404054 memcheck powerpc subfe x, x, x initializes x to 0 or -1 based on CA
998 404638 Add VG_(replaceIndexXA)
999 404843 s390x: backtrace sometimes ends prematurely
1000 404888 autotools cleanup series
1001 405079 unhandled ppc64le-linux syscall: 131 (quotactl)
1002 405182 Valgrind fails to build with Clang
1003 405205 filter_libc: remove the line holding the futex syscall error entirely
1004 405356 PPC64, xvcvsxdsp, xvcvuxdsp are supposed to write the 32-bit result to
1005 the upper and lower 32-bits of the 64-bit result
1006 405362 PPC64, vmsummbm instruction doesn't handle overflow case correctly
1007 405363 PPC64, xvcvdpsxws, xvcvdpuxws, do not handle NaN arguments correctly.
1008 405365 PPC64, function _get_maxmin_fp_NaN() doesn't handle QNaN, SNaN case
1010 405403 s390x disassembler cannot be used on x86
1011 405430 Use gcc -Wimplicit-fallthrough=2 by default if available
1012 405458 MIPS mkFormVEC arguments swapped?
1013 405716 drd: Fix an integer overflow in the stack margin calculation
1014 405722 Support arm64 core dump
1015 405733 PPC64, xvcvdpsp should write 32-bit result to upper and lower 32-bits
1016 of the 64-bit destination field.
1017 405734 PPC64, vrlwnm, vrlwmi, vrldrm, vrldmi do not work properly when me < mb
1018 405782 "VEX temporary storage exhausted" when attempting to debug slic3r-pe
1019 406198 none/tests/ppc64/test_isa_3_0_other test sporadically including CA
1021 406256 PPC64, vector floating point instructions don't handle subnormal
1022 according to VSCR[NJ] bit setting.
1023 406352 cachegrind/callgrind fails ann tests because of missing a.c
1024 406354 dhat is broken on x86 (32bit)
1025 406355 mcsignopass, mcsigpass, mcbreak fail due to difference in gdb output
1026 406357 gdbserver_tests fails because of gdb output change
1027 406360 memcheck/tests/libstdc++.supp needs more supression variants
1028 406422 none/tests/amd64-linux/map_32bits.vgtest fails too easily
1029 406465 arm64 insn selector fails on "t0 = <expr>" where <expr> has type Ity_F16
1030 407340 PPC64, does not support the vlogefp, vexptefp instructions.
1031 n-i-bz add syswrap for PTRACE_GET|SET_THREAD_AREA on amd64.
1032 n-i-bz Fix callgrind_annotate non deterministic order for equal total
1033 n-i-bz callgrind_annotate --threshold=100 does not print all functions.
1034 n-i-bz callgrind_annotate Use of uninitialized value in numeric gt (>)
1035 n-i-bz amd64 (x86_64): RDRAND and F16C insn set extensions are supported
1037 (3.15.0.RC1: 8 April 2019, git ce94d674de5b99df173aad4c3ee48fc2a92e5d9c)
1038 (3.15.0.RC2: 11 April 2019, git 0c8be9bbede189ec580ec270521811766429595f)
1039 (3.15.0: 14 April 2019, git 270037da8b508954f0f7d703a0bebf5364eec548)
1043 Release 3.14.0 (9 October 2018)
1044 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1046 3.14.0 is a feature release with many improvements and the usual collection of
1049 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1050 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1051 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
1052 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
1053 support for X86/macOS 10.13, AMD64/macOS 10.13.
1055 * ==================== CORE CHANGES ===================
1057 * The new option --keep-debuginfo=no|yes (default no) can be used to retain
1058 debug info for unloaded code. This allows saved stack traces (e.g. for
1059 memory leaks) to include file/line info for code that has been dlclose'd (or
1060 similar). See the user manual for more information and known limitations.
1062 * Ability to specify suppressions based on source file name and line number.
1064 * Majorly overhauled register allocator. No end-user changes, but the JIT
1065 generates code a bit more quickly now.
1067 * ================== PLATFORM CHANGES =================
1069 * Preliminary support for macOS 10.13 has been added.
1071 * mips: support for MIPS32/MIPS64 Revision 6 has been added.
1073 * mips: support for MIPS SIMD architecture (MSA) has been added.
1075 * mips: support for MIPS N32 ABI has been added.
1077 * s390: partial support for vector instructions (integer and string) has been
1080 * ==================== TOOL CHANGES ====================
1082 * Helgrind: Addition of a flag
1083 --delta-stacktrace=no|yes [yes on linux amd64/x86]
1084 which specifies how full history stack traces should be computed.
1085 Setting this to =yes can speed up Helgrind by 25% when using
1086 --history-level=full.
1088 * Memcheck: reduced false positive rate for optimised code created by Clang 6
1089 / LLVM 6 on x86, amd64 and arm64. In particular, Memcheck analyses code
1090 blocks more carefully to determine where it can avoid expensive definedness
1091 checks without loss of precision. This is controlled by the flag
1092 --expensive-definedness-checks=no|auto|yes [auto].
1094 * ==================== OTHER CHANGES ====================
1096 * Valgrind is now buildable with link-time optimisation (LTO). A new
1097 configure option --enable-lto=yes allows building Valgrind with LTO. If the
1098 toolchain supports it, this produces a smaller/faster Valgrind (up to 10%).
1099 Note that if you are doing Valgrind development, --enable-lto=yes massively
1100 slows down the build process.
1102 * ==================== FIXED BUGS ====================
1104 The following bugs have been fixed or resolved. Note that "n-i-bz"
1105 stands for "not in bugzilla" -- that is, a bug that was reported to us
1106 but never got a bugzilla entry. We encourage you to file bugs in
1107 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1108 than mailing the developers (or mailing lists) directly -- bugs that
1109 are not entered into bugzilla tend to get forgotten about or ignored.
1111 To see details of a given bug, visit
1112 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1113 where XXXXXX is the bug number as listed below.
1115 79362 Debug info is lost for .so files when they are dlclose'd
1116 208052 strlcpy error when n = 0
1117 255603 exp-sgcheck Assertion '!already_present' failed
1118 338252 building valgrind with -flto (link time optimisation) fails
1119 345763 MIPS N32 ABI support
1120 368913 WARNING: unhandled arm64-linux syscall: 117 (ptrace)
1121 == 388664 unhandled arm64-linux syscall: 117 (ptrace)
1122 372347 Replacement problem of the additional c++14/c++17 new/delete operators
1123 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
1124 376257 helgrind history full speed up using a cached stack
1125 379373 Fix syscall param msg->desc.port.name points to uninitialised byte(s)
1127 379748 Fix missing pselect syscall (OS X 10.11)
1128 379754 Fix missing syscall ulock_wait (OS X 10.12)
1129 380397 s390x: __GI_strcspn() replacemenet needed
1130 381162 possible array overrun in VEX register allocator
1131 381272 ppc64 doesn't compile test_isa_2_06_partx.c without VSX support
1132 381274 powerpc too chatty even with --sigill-diagnostics=no
1133 381289 epoll_pwait can have a NULL sigmask
1134 381553 VEX register allocator v3
1135 381556 arm64: Handle feature registers access on 4.11 Linux kernel or later
1136 381769 Use ucontext_t instead of struct ucontext
1137 381805 arm32 needs ld.so index hardwire for new glibc security fixes
1138 382256 gz compiler flag test doesn't work for gold
1139 382407 vg_perf needs "--terse" command line option
1140 382515 "Assertion 'di->have_dinfo' failed." on wine's dlls/mscoree/tests/[..]
1141 382563 MIPS MSA ASE support
1142 382998 xml-socket doesn't work
1143 383275 massif: m_xarray.c:162 (ensureSpaceXA): Assertion '!xa->arr' failed
1144 383723 Fix missing kevent_qos syscall (macOS 10.11)
1145 == 385604 illegal hardware instruction (OpenCV cv::namedWindow)
1146 384096 Mention AddrCheck at Memcheck's command line option [..]
1147 384230 vex x86->IR: 0x67 0xE8 0xAB 0x68
1148 == 384156 vex x86->IR: 0x67 0xE8 0x6B 0x6A
1149 == 386115 vex x86->IR: 0x67 0xE8 0xD3 0x8B any program
1150 == 388407 vex x86->IR: 0x67 0xE8 0xAB 0x29
1151 == 394903 vex x86->IR: 0x67 0xE8 0x1B 0xDA
1152 384337 performance improvements to VEX register allocator v2 and v3
1153 384526 reduce number of spill insns generated by VEX register allocator v3
1154 384584 Callee saved regs listed first for AMD64, X86, and PPC architectures
1155 384631 Sanitise client args as printed with -v
1156 384633 Add a simple progress-reporting facility
1157 384987 VEX regalloc: allocate caller-save registers for short lived vregs
1158 385055 PPC VEX temporary storage exhausted
1159 385182 PPC64 is missing support for the DSCR
1160 385183 PPC64, Add support for xscmpeqdp, xscmpgtdp, xscmpgedp, xsmincdp
1161 385207 PPC64, generate_store_FPRF() generates too many Iops
1162 385208 PPC64, xxperm instruction exhausts temporary memory
1163 385210 PPC64, vpermr instruction could exhaust temporary memory
1164 385279 unhandled syscall: mach:43 (mach_generate_activity_id)
1165 == 395136 valgrind: m_syswrap/syswrap-main.c:438 (Bool eq_Syscall[..]
1166 == 387045 Valgrind crashing on High Sierra when testing any newly [..]
1167 385334 PPC64, fix vpermr, xxperm, xxpermr mask value.
1168 385408 s390x: z13 vector "support" instructions not implemented
1169 385409 s390x: z13 vector integer instructions not implemented
1170 385410 s390x: z13 vector string instructions not implemented
1171 385412 s390x: new non-vector z13 instructions not implemented
1172 385868 glibc ld.so _dl_runtime_resolve_avx_slow conditional jump warning.
1173 385912 none/tests/rlimit_nofile fails on newer glibc/kernel.
1174 385939 Optionally exit on the first error
1175 386318 valgrind.org/info/tools.html is missing SGCheck
1176 386425 running valgrind + wine on armv7l gives illegal opcode
1177 386397 PPC64, valgrind truncates powerpc timebase to 32-bits.
1178 387410 MIPSr6 support
1179 387664 Memcheck: make expensive-definedness-checks be the default
1180 387712 s390x cgijnl reports Conditional jump depends on uninitialised value
1181 387766 asm shifts cause false positive "Conditional jump or move depends
1182 on uninitialised value"
1183 387773 .gnu_debugaltlink paths resolve relative to .debug file, not symlink
1184 388174 valgrind with Wine quits with "Assertion 'cfsi_fits' failed"
1185 388786 Support bpf syscall in amd64 Linux
1186 388862 Add replacements for wmemchr and wcsnlen on Linux
1187 389065 valgrind meets gcc flag -Wlogical-op
1188 389373 exp-sgcheck the 'impossible' happened as Ist_LoadG is not instrumented
1189 390471 suppression by specification of source-file line number
1190 390723 make xtree dump files world wide readable, similar to log files
1191 391164 constraint bug in tests/ppc64/test_isa_2_07_part1.c for mtfprwa
1192 391861 Massif Assertion 'n_ips >= 1 && n_ips <= VG_(clo_backtrace_size)'
1193 392118 unhandled amd64-linux syscall: 332 (statx)
1194 392449 callgrind not clearing the number of calls properly
1195 393017 Add missing support for xsmaxcdp instruction, bug fixes for xsmincdp,
1196 lxssp, stxssp and stxvl instructions.
1197 393023 callgrind_control risks using the wrong vgdb
1198 393062 build-id ELF phdrs read causes "debuginfo reader: ensure_valid failed"
1199 393099 posix_memalign() invalid write if alignment == 0
1200 393146 failing assert "is_DebugInfo_active(di)"
1201 395709 PPC64 is missing support for the xvnegsp instruction
1202 395682 Accept read-only PT_LOAD segments and .rodata by ld -z separate-code
1204 396475 valgrind OS-X build: config.h not found (out-of-tree macOS builds)
1205 395991 arm-linux: wine's unit tests enter a signal delivery loop [..]
1206 396839 s390x: Trap instructions not implemented
1207 396887 arch_prctl should return EINVAL on unknown option
1208 == 397286 crash before launching binary (Unsupported arch_prctl option)
1209 == 397393 valgrind: the 'impossible' happened: (Archlinux)
1210 == 397521 valgrind: the 'impossible' happened: Unsupported [..]
1211 396906 compile tests failure on mips32-linux: broken inline asm in tests on
1213 397012 glibc ld.so uses arch_prctl on i386
1214 397089 amd64: Incorrect decoding of three-register vmovss/vmovsd opcode 11h
1215 397354 utimensat should ignore timespec tv_sec if tv_nsec is UTIME_NOW/OMIT
1216 397424 glibc 2.27 and gdb_server tests
1217 398028 Assertion `cfsi_fits` failing in simple C program
1218 398066 s390x: cgijl dep1, 0 reports false unitialised values warning
1220 n-i-bz Fix missing workq_ops operations (macOS)
1221 n-i-bz fix bug in strspn replacement
1222 n-i-bz Add support for the Linux BLKFLSBUF ioctl
1223 n-i-bz Add support for the Linux BLKREPORTZONE and BLKRESETZONE ioctls
1224 n-i-bz Fix possible stack trashing by semctl syscall wrapping
1225 n-i-bz Add support for the Linux membarrier() system call
1226 n-i-bz x86 front end: recognise and handle UD2 correctly
1227 n-i-bz Signal delivery for x86-linux: ensure that the stack pointer is
1228 correctly aligned before entering the handler.
1230 (3.14.0.RC1: 30 September 2018, git c2aeea2d28acb0639bcc8cc1e4ab115067db1eae)
1231 (3.14.0.RC2: 3 October 2018, git 3e214c4858a6fdd5697e767543a0c19e30505582)
1232 (3.14.0: 9 October 2018, git 353a3587bb0e2757411f9138f5e936728ed6cc4f)
1236 Release 3.13.0 (15 June 2017)
1237 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1239 3.13.0 is a feature release with many improvements and the usual collection of
1242 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1243 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1244 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
1245 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12.
1247 * ==================== CORE CHANGES ===================
1249 * The translation cache size has been increased to keep up with the demands of
1250 large applications. The maximum number of sectors has increased from 24 to
1251 48. The default number of sectors has increased from 16 to 32 on all
1252 targets except Android, where the increase is from 6 to 12.
1254 * The amount of memory that Valgrind can use has been increased from 64GB to
1255 128GB. In particular this means your application can allocate up to about
1256 60GB when running on Memcheck.
1258 * Valgrind's default load address has been changed from 0x3800'0000 to
1259 0x5800'0000, so as to make it possible to load larger executables. This
1260 should make it possible to load executables of size at least 1200MB.
1262 * A massive spaceleak caused by reading compressed debuginfo files has been
1263 fixed. Valgrind should now be entirely usable with gcc-7.0 "-gz" created
1266 * The C++ demangler has been updated.
1268 * Support for demangling Rust symbols has been added.
1270 * A new representation of stack traces, the "XTree", has been added. An XTree
1271 is a tree of stacktraces with data associated with the stacktraces. This is
1272 used by various tools (Memcheck, Helgrind, Massif) to report on the heap
1273 consumption of your program. Reporting is controlled by the new options
1274 --xtree-memory=none|allocs|full and --xtree-memory-file=<file>.
1276 A report can also be produced on demand using the gdbserver monitor command
1277 'xtmemory [<filename>]>'. The XTree can be output in 2 formats: 'callgrind
1278 format' and 'massif format. The existing visualisers for these formats (e.g.
1279 callgrind_annotate, KCachegrind, ms_print) can be used to visualise and
1280 analyse these reports.
1282 Memcheck can also produce XTree leak reports using the Callgrind file
1283 format. For more details, see the user manual.
1285 * ================== PLATFORM CHANGES =================
1287 * ppc64: support for ISA 3.0B and various fixes for existing 3.0 support
1289 * amd64: fixes for JIT failure problems on long AVX2 code blocks
1291 * amd64 and x86: support for CET prefixes has been added
1293 * arm32: a few missing ARMv8 instructions have been implemented
1295 * arm64, mips64, mips32: an alternative implementation of Load-Linked and
1296 Store-Conditional instructions has been added. This is to deal with
1297 processor implementations that implement the LL/SC specifications strictly
1298 and as a result cause Valgrind to hang in certain situations. The
1299 alternative implementation is automatically enabled at startup, as required.
1300 You can use the option --sim-hints=fallback-llsc to force-enable it if you
1303 * Support for OSX 10.12 has been improved.
1305 * On Linux, clone handling has been improved to honour CLONE_VFORK that
1306 involves a child stack. Note however that CLONE_VFORK | CLONE_VM is handled
1307 like CLONE_VFORK (by removing CLONE_VM), so applications that depend on
1308 CLONE_VM exact semantics will (still) not work.
1310 * The TileGX/Linux port has been removed because it appears to be both unused
1313 * ==================== TOOL CHANGES ====================
1317 - Memcheck should give fewer false positives when running optimised
1318 Clang/LLVM generated code.
1320 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1322 - New command line options --xtree-leak=no|yes and --xtree-leak-file=<file>
1323 to produce the end of execution leak report in a xtree callgrind format
1326 - New option 'xtleak' in the memcheck leak_check monitor command, to produce
1327 the leak report in an xtree file.
1331 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1333 - For some workloads (typically, for big applications), Massif memory
1334 consumption and CPU consumption has decreased significantly.
1338 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1340 - addition of client request VALGRIND_HG_GNAT_DEPENDENT_MASTER_JOIN, useful
1341 for Ada gnat compiled applications.
1343 * ==================== OTHER CHANGES ====================
1345 * For Valgrind developers: in an outer/inner setup, the outer Valgrind will
1346 append the inner guest stacktrace to the inner host stacktrace. This helps
1347 to investigate the errors reported by the outer, when they are caused by the
1348 inner guest program (such as an inner regtest). See README_DEVELOPERS for
1351 * To allow fast detection of callgrind files by desktop environments and file
1352 managers, the format was extended to have an optional first line that
1353 uniquely identifies the format ("# callgrind format"). Callgrind creates
1354 this line now, as does the new xtree functionality.
1356 * File name template arguments (such as --log-file, --xtree-memory-file, ...)
1357 have a new %n format letter that is replaced by a sequence number.
1359 * "--version -v" now shows the SVN revision numbers from which Valgrind was
1362 * ==================== FIXED BUGS ====================
1364 The following bugs have been fixed or resolved. Note that "n-i-bz"
1365 stands for "not in bugzilla" -- that is, a bug that was reported to us
1366 but never got a bugzilla entry. We encourage you to file bugs in
1367 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1368 than mailing the developers (or mailing lists) directly -- bugs that
1369 are not entered into bugzilla tend to get forgotten about or ignored.
1371 To see details of a given bug, visit
1372 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1373 where XXXXXX is the bug number as listed below.
1375 162848 --log-file output isn't split when a program forks
1376 340777 Illegal instruction on mips (ar71xx)
1377 341481 MIPS64: Iop_CmpNE32 triggers false warning on MIPS64 platforms
1378 342040 Valgrind mishandles clone with CLONE_VFORK | CLONE_VM that clones
1379 to a different stack.
1380 344139 x86 stack-seg overrides, needed by the Wine people
1381 344524 store conditional of guest applications always fail - observed on
1383 348616 Wine/valgrind: noted but unhandled ioctl 0x5390 [..] (DVD_READ_STRUCT)
1384 352395 Please provide SVN revision info in --version -v
1385 352767 Wine/valgrind: noted but unhandled ioctl 0x5307 [..] (CDROMSTOP)
1386 356374 Assertion 'DRD_(g_threadinfo)[tid].pt_threadid !=
1387 INVALID_POSIX_THREADID' failed
1388 358213 helgrind/drd bar_bad testcase hangs or crashes with new glibc pthread
1389 barrier implementation
1390 358697 valgrind.h: Some code remains even when defining NVALGRIND
1391 359202 Add musl libc configure/compile
1392 360415 amd64 instructions ADCX and ADOX are not implemented in VEX
1393 == 372828 (vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10)
1394 360429 unhandled ioctl 0x530d with no size/direction hints (CDROMREADMODE1)
1395 362223 assertion failed when .valgrindrc is a directory instead of a file
1396 367543 bt/btc/btr/bts x86/x86_64 instructions are poorly-handled wrt flags
1397 367942 Segfault vgPlain_do_sys_sigaction (m_signals.c:1138)
1398 368507 can't malloc chunks larger than about 34GB
1399 368529 Android arm target link error, missing atexit and pthread_atfork
1400 368863 WARNING: unhandled arm64-linux syscall: 100 (get_robust_list)
1401 368865 WARNING: unhandled arm64-linux syscall: 272 (kcmp)
1402 368868 disInstr(arm64): unhandled instruction 0xD53BE000 = cntfrq_el0 (ARMv8)
1403 368917 WARNING: unhandled arm64-linux syscall: 218 (request_key)
1404 368918 WARNING: unhandled arm64-linux syscall: 127 (sched_rr_get_interval)
1405 368922 WARNING: unhandled arm64-linux syscall: 161 (sethostname)
1406 368924 WARNING: unhandled arm64-linux syscall: 84 (sync_file_range)
1407 368925 WARNING: unhandled arm64-linux syscall: 130 (tkill)
1408 368926 WARNING: unhandled arm64-linux syscall: 97 (unshare)
1409 369459 valgrind on arm64 violates the ARMv8 spec (ldxr/stxr)
1410 370028 Reduce the number of compiler warnings on MIPS platforms
1411 370635 arm64 missing syscall getcpu
1412 371225 Fix order of timer_{gettime,getoverrun,settime} syscalls on arm64
1413 371227 Clean AArch64 syscall table
1414 371412 Rename wrap_sys_shmat to sys_shmat like other wrappers
1415 371471 Valgrind complains about non legit memory leaks on placement new (C++)
1416 371491 handleAddrOverrides() is [incorrect] when ASO prefix is used
1417 371503 disInstr(arm64): unhandled instruction 0xF89F0000
1418 371869 support '%' in symbol Z-encoding
1419 371916 execution tree xtree concept
1420 372120 c++ demangler demangles symbols which are not c++
1421 372185 Support of valgrind on ARMv8 with 32 bit executable
1422 372188 vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10 0x10 0x48 (PCMPxSTRx $0x10)
1423 372195 Power PC, xxsel instruction is not always recognized.
1424 372504 Hanging on exit_group
1425 372600 process loops forever when fatal signals are arriving quickly
1426 372794 LibVEX (arm32 front end): 'Assertion szBlg2 <= 3' failed
1427 373046 Stacks registered by core are never deregistered
1428 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
1429 373086 Implement additional Xen hypercalls
1430 373192 Calling posix_spawn in glibc 2.24 completely broken
1431 373488 Support for fanotify API on ARM64 architecture
1432 == 368864 WARNING: unhandled arm64-linux syscall: 262 (fanotify_init)
1433 373555 Rename BBPTR to GSPTR as it denotes guest state pointer only
1434 373938 const IRExpr arguments for matchIRExpr()
1435 374719 some spelling fixes
1436 374963 increase valgrind's load address to prevent mmap failure
1437 375514 valgrind_get_tls_addr() does not work in case of static TLS
1438 375772 +1 error in get_elf_symbol_info() when computing value of 'hi' address
1439 for ML_(find_rx_mapping)()
1440 375806 Test helgrind/tests/tc22_exit_w_lock fails with glibc 2.24
1441 375839 Temporary storage exhausted, with long sequence of vfmadd231ps insns
1442 == 377159 "vex: the `impossible' happened" still present
1443 == 375150 Assertion 'tres.status == VexTransOK' failed
1444 == 378068 valgrind crashes on AVX2 function in FFmpeg
1445 376142 Segfaults on MIPS Cavium Octeon boards
1446 376279 disInstr(arm64): unhandled instruction 0xD50320FF
1447 376455 Solaris: unhandled syscall lgrpsys(180)
1448 376518 Solaris: unhandled fast trap getlgrp(6)
1449 376611 ppc64 and arm64 don't know about prlimit64 syscall
1450 376729 PPC64, remove R2 from the clobber list
1452 376956 syswrap of SNDDRV and DRM_IOCTL_VERSION causing some addresses
1453 to be wrongly marked as addressable
1454 377066 Some Valgrind unit tests fail to compile on Ubuntu 16.10 with
1455 PIE enabled by default
1456 377376 memcheck/tests/linux/getregset fails with glibc2.24
1457 377427 PPC64, lxv instruction failing on odd destination register
1458 377478 PPC64: ISA 3.0 setup fixes
1459 377698 Missing memory check for futex() uaddr arg for FUTEX_WAKE
1460 and FUTEX_WAKE_BITSET, check only 4 args for FUTEX_WAKE_BITSET,
1461 and 2 args for FUTEX_TRYLOCK_PI
1462 377717 Fix massive space leak when reading compressed debuginfo sections
1463 377891 Update Xen 4.6 domctl wrappers
1464 377930 fcntl syscall wrapper is missing flock structure check
1465 378524 libvexmultiarch_test regression on s390x and ppc64
1466 378535 Valgrind reports INTERNAL ERROR in execve syscall wrapper
1467 378673 Update libiberty demangler
1468 378931 Add ISA 3.0B additional isnstructions, add OV32, CA32 setting support
1469 379039 syscall wrapper for prctl(PR_SET_NAME) must not check more than 16 bytes
1470 379094 Valgrind reports INTERNAL ERROR in rt_sigsuspend syscall wrapper
1471 379371 UNKNOWN task message [id 3444, to mach_task_self(), reply 0x603]
1472 (task_register_dyld_image_infos)
1473 379372 UNKNOWN task message [id 3447, to mach_task_self(), reply 0x603]
1474 (task_register_dyld_shared_cache_image_info)
1475 379390 unhandled syscall: mach:70 (host_create_mach_voucher_trap)
1476 379473 MIPS: add support for rdhwr cycle counter register
1477 379504 remove TileGX/Linux port
1478 379525 Support more x86 nop opcodes
1479 379838 disAMode(x86): not an addr!
1480 379703 PC ISA 3.0 fixes: stxvx, stxv, xscmpexpdp instructions
1481 379890 arm: unhandled instruction: 0xEBAD 0x1B05 (sub.w fp, sp, r5, lsl #4)
1482 379895 clock_gettime does not execute POST syscall wrapper
1483 379925 PPC64, mtffs does not set the FPCC and C bits in the FPSCR correctly
1484 379966 WARNING: unhandled amd64-linux syscall: 313 (finit_module)
1485 380200 xtree generated callgrind files refer to files without directory name
1486 380202 Assertion failure for cache line size (cls == 64) on aarch64.
1487 380397 s390x: __GI_strcspn() replacement needed
1488 n-i-bz Fix pub_tool_basics.h build issue with g++ 4.4.7.
1490 (3.13.0.RC1: 2 June 2017, vex r3386, valgrind r16434)
1491 (3.13.0.RC2: 9 June 2017, vex r3389, valgrind r16443)
1492 (3.13.0: 14 June 2017, vex r3396, valgrind r16446)
1496 Release 3.12.0 (20 October 2016)
1497 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1499 3.12.0 is a feature release with many improvements and the usual
1500 collection of bug fixes.
1502 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1503 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1504 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1505 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1506 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1507 X86/MacOSX 10.11/12, AMD64/MacOSX 10.11/12 and TILEGX/Linux.
1509 * ================== PLATFORM CHANGES =================
1511 * POWER: Support for ISA 3.0 has been added
1513 * mips: support for O32 FPXX ABI has been added.
1514 * mips: improved recognition of different processors
1515 * mips: determination of page size now done at run time
1517 * amd64: Partial support for AMD FMA4 instructions.
1519 * arm, arm64: Support for v8 crypto and CRC instructions.
1521 * Improvements and robustification of the Solaris port.
1523 * Preliminary support for MacOS 10.12 (Sierra) has been added.
1525 Whilst 3.12.0 continues to support the 32-bit x86 instruction set, we
1526 would prefer users to migrate to 64-bit x86 (a.k.a amd64 or x86_64)
1527 where possible. Valgrind's support for 32-bit x86 has stagnated in
1528 recent years and has fallen far behind that for 64-bit x86
1529 instructions. By contrast 64-bit x86 is well supported, up to and
1532 * ==================== TOOL CHANGES ====================
1536 - Added meta mempool support for describing a custom allocator which:
1537 - Auto-frees all chunks assuming that destroying a pool destroys all
1539 - Uses itself to allocate other memory blocks
1541 - New flag --ignore-range-below-sp to ignore memory accesses below
1542 the stack pointer, if you really have to. The related flag
1543 --workaround-gcc296-bugs=yes is now deprecated. Use
1544 --ignore-range-below-sp=1024-1 as a replacement.
1548 - Improved thread startup time significantly on non-Linux platforms.
1552 - Added collection of the metric "tot-blocks-allocd"
1554 * ==================== OTHER CHANGES ====================
1556 * Replacement/wrapping of malloc/new related functions is now done not just
1557 for system libraries by default, but for any globally defined malloc/new
1558 related function (both in shared libraries and statically linked alternative
1559 malloc implementations). The dynamic (runtime) linker is excluded, though.
1560 To only intercept malloc/new related functions in
1561 system libraries use --soname-synonyms=somalloc=nouserintercepts (where
1562 "nouserintercepts" can be any non-existing library name).
1563 This new functionality is not implemented for MacOS X.
1565 * The maximum number of callers in a suppression entry is now equal to
1566 the maximum size for --num-callers (500).
1567 Note that --gen-suppressions=yes|all similarly generates suppressions
1568 containing up to --num-callers frames.
1570 * New and modified GDB server monitor features:
1572 - Valgrind's gdbserver now accepts the command 'catch syscall'.
1573 Note that you must have GDB >= 7.11 to use 'catch syscall' with
1576 * New option --run-cxx-freeres=<yes|no> can be used to change whether
1577 __gnu_cxx::__freeres() cleanup function is called or not. Default is
1580 * Valgrind is able to read compressed debuginfo sections in two formats:
1581 - zlib ELF gABI format with SHF_COMPRESSED flag (gcc option -gz=zlib)
1582 - zlib GNU format with .zdebug sections (gcc option -gz=zlib-gnu)
1584 * Modest JIT-cost improvements: the cost of instrumenting code blocks
1585 for the most common use case (x86_64-linux, Memcheck) has been
1588 * Improved performance for programs that do a lot of discarding of
1589 instruction address ranges of 8KB or less.
1591 * The C++ symbol demangler has been updated.
1593 * More robustness against invalid syscall parameters on Linux.
1595 * ==================== FIXED BUGS ====================
1597 The following bugs have been fixed or resolved. Note that "n-i-bz"
1598 stands for "not in bugzilla" -- that is, a bug that was reported to us
1599 but never got a bugzilla entry. We encourage you to file bugs in
1600 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1601 than mailing the developers (or mailing lists) directly -- bugs that
1602 are not entered into bugzilla tend to get forgotten about or ignored.
1604 To see details of a given bug, visit
1605 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1606 where XXXXXX is the bug number as listed below.
1608 191069 Exiting due to signal not reported in XML output
1609 199468 Suppressions: stack size limited to 25
1610 while --num-callers allows more frames
1611 212352 vex amd64 unhandled opc_aux = 0x 2, first_opcode == 0xDC (FCOM)
1612 278744 cvtps2pd with redundant RexW
1613 303877 valgrind doesn't support compressed debuginfo sections.
1614 345307 Warning about "still reachable" memory when using libstdc++ from gcc 5
1615 348345 Assertion fails for negative lineno
1616 348924 MIPS: Load doubles through memory so the code compiles with the FPXX ABI
1617 351282 V 3.10.1 MIPS softfloat build broken with GCC 4.9.3 / binutils 2.25.1
1618 351692 Dumps created by valgrind are not readable by gdb (mips32 specific)
1619 351804 Crash on generating suppressions for "printf" call on OS X 10.10
1620 352197 mips: mmap2() not wrapped correctly for page size > 4096
1621 353083 arm64 doesn't implement various xattr system calls
1622 353084 arm64 doesn't support sigpending system call
1623 353137 www: update info for Supported Platforms
1624 353138 www: update "The Valgrind Developers" page
1625 353370 don't advertise RDRAND in cpuid for Core-i7-4910-like avx2 machine
1628 353384 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x62 (pcmpXstrX $0x62)
1629 353398 WARNING: unhandled amd64-solaris syscall: 207
1630 353660 XML in auxwhat tag not escaping reserved symbols properly
1631 353680 s390x: Crash with certain glibc versions due to non-implemented TBEGIN
1632 353727 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x72 (pcmpXstrX $0x72)
1633 353802 ELF debug info reader confused with multiple .rodata sections
1634 353891 Assert 'bad_scanned_addr < VG_ROUNDDN(start+len, sizeof(Addr))' failed
1635 353917 unhandled amd64-solaris syscall fchdir(120)
1636 353920 unhandled amd64-solaris syscall: 170
1637 354274 arm: unhandled instruction: 0xEBAD 0x0AC1 (sub.w sl, sp, r1, lsl #3)
1638 354392 unhandled amd64-solaris syscall: 171
1639 354797 Vbit test does not include Iops for Power 8 instruction support
1640 354883 tst->os_state.pthread - magic_delta assertion failure on OSX 10.11
1644 354933 Fix documentation of --kernel-variant=android-no-hw-tls option
1645 355188 valgrind should intercept all malloc related global functions
1646 355454 do not intercept malloc related symbols from the runtime linker
1647 355455 stderr.exp of test cases wrapmalloc and wrapmallocstatic overconstrained
1648 356044 Dwarf line info reader misinterprets is_stmt register
1649 356112 mips: replace addi with addiu
1650 356393 valgrind (vex) crashes because isZeroU happened
1653 356676 arm64-linux: unhandled syscalls 125, 126 (sched_get_priority_max/min)
1654 356678 arm64-linux: unhandled syscall 232 (mincore)
1655 356817 valgrind.h triggers compiler errors on MSVC when defining NVALGRIND
1656 356823 Unsupported ARM instruction: stlex
1657 357059 x86/amd64: SSE cvtpi2ps with memory source does transition to MMX state
1658 357338 Unhandled instruction for SHA instructions libcrypto Boring SSL
1659 357673 crash if I try to run valgrind with a binary link with libcurl
1660 357833 Setting RLIMIT_DATA to zero breaks with linux 4.5+
1661 357871 pthread_spin_destroy not properly wrapped
1662 357887 Calls to VG_(fclose) do not close the file descriptor
1663 357932 amd64->IR: accept redundant REX prefixes for {minsd,maxsd} m128, xmm.
1664 358030 support direct socket calls on x86 32bit (new in linux 4.3)
1665 358478 drd/tests/std_thread.cpp doesn't build with GCC6
1666 359133 Assertion 'eltSzB <= ddpa->poolSzB' failed
1667 359181 Buffer Overflow during Demangling
1668 359201 futex syscall "skips" argument 5 if op is FUTEX_WAIT_BITSET
1669 359289 s390x: popcnt (B9E1) not implemented
1670 359472 The Power PC vsubuqm instruction doesn't always give the correct result
1671 359503 Add missing syscalls for aarch64 (arm64)
1672 359645 "You need libc6-dbg" help message could be more helpful
1673 359703 s390: wire up separate socketcalls system calls
1674 359724 getsockname might crash - deref_UInt should call safe_to_deref
1675 359733 amd64 implement ld.so strchr/index override like x86
1676 359767 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 1/5
1677 359829 Power PC test suite none/tests/ppc64/test_isa_2_07.c uses
1679 359838 arm64: Unhandled instruction 0xD5033F5F (clrex)
1680 359871 Incorrect mask handling in ppoll
1681 359952 Unrecognised PCMPESTRM variants (0x70, 0x19)
1682 360008 Contents of Power vr registers contents is not printed correctly when
1683 the --vgdb-shadow-registers=yes option is used
1684 360035 POWER PC instruction bcdadd and bcdsubtract generate result with
1685 non-zero shadow bits
1686 360378 arm64: Unhandled instruction 0x5E280844 (sha1h s4, s2)
1687 360425 arm64 unsupported instruction ldpsw
1689 360519 none/tests/arm64/memory.vgtest might fail with newer gcc
1690 360571 Error about the Android Runtime reading below the stack pointer on ARM
1691 360574 Wrong parameter type for an ashmem ioctl() call on Android and ARM64
1692 360749 kludge for multiple .rodata sections on Solaris no longer needed
1693 360752 raise the number of reserved fds in m_main.c from 10 to 12
1694 361207 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 2/5
1695 361226 s390x: risbgn (EC59) not implemented
1696 361253 [s390x] ex_clone.c:42: undefined reference to `pthread_create'
1697 361354 ppc64[le]: wire up separate socketcalls system calls
1698 361615 Inconsistent termination for multithreaded process terminated by signal
1699 361926 Unhandled Solaris syscall: sysfs(84)
1700 362009 V dumps core on unimplemented functionality before threads are created
1701 362329 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 3/5
1702 362894 missing (broken) support for wbit field on mtfsfi instruction (ppc64)
1703 362935 [AsusWRT] Assertion 'sizeof(TTEntryC) <= 88' failed
1704 362953 Request for an update to the Valgrind Developers page
1705 363680 add renameat2() support
1706 363705 arm64 missing syscall name_to_handle_at and open_by_handle_at
1707 363714 ppc64 missing syscalls sync, waitid and name_to/open_by_handle_at
1708 363858 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 4/5
1709 364058 clarify in manual limitations of array overruns detections
1710 364413 pselect sycallwrapper mishandles NULL sigmask
1711 364728 Power PC, missing support for several HW registers in
1712 get_otrack_shadow_offset_wrk()
1713 364948 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 5/5
1714 365273 Invalid write to stack location reported after signal handler runs
1715 365912 ppc64BE segfault during jm-insns test (RELRO)
1716 366079 FPXX Support for MIPS32 Valgrind
1717 366138 Fix configure errors out when using Xcode 8 (clang 8.0.0)
1718 366344 Multiple unhandled instruction for Aarch64
1719 (0x0EE0E020, 0x1AC15800, 0x4E284801, 0x5E040023, 0x5E056060)
1720 367995 Integration of memcheck with custom memory allocator
1721 368120 x86_linux asm _start functions do not keep 16-byte aligned stack pointer
1722 368412 False positive result for altivec capability check
1723 368416 Add tc06_two_races_xml.exp output for ppc64
1724 368419 Perf Events ioctls not implemented
1725 368461 mmapunmap test fails on ppc64
1726 368823 run_a_thread_NORETURN assembly code typo for VGP_arm64_linux target
1727 369000 AMD64 fma4 instructions unsupported.
1728 369169 ppc64 fails jm_int_isa_2_07 test
1729 369175 jm_vec_isa_2_07 test crashes on ppc64
1730 369209 valgrind loops and eats up all memory if cwd doesn't exist.
1731 369356 pre_mem_read_sockaddr syscall wrapper can crash with bad sockaddr
1732 369359 msghdr_foreachfield can crash when handling bad iovec
1733 369360 Bad sigprocmask old or new sets can crash valgrind
1734 369361 vmsplice syscall wrapper crashes on bad iovec
1735 369362 Bad sigaction arguments crash valgrind
1736 369383 x86 sys_modify_ldt wrapper crashes on bad ptr
1737 369402 Bad set/get_thread_area pointer crashes valgrind
1738 369441 bad lvec argument crashes process_vm_readv/writev syscall wrappers
1739 369446 valgrind crashes on unknown fcntl command
1740 369439 S390x: Unhandled insns RISBLG/RISBHG and LDE/LDER
1741 369468 Remove quadratic metapool algorithm using VG_(HT_remove_at_Iter)
1742 370265 ISA 3.0 HW cap stuff needs updating
1743 371128 BCD add and subtract instructions on Power BE in 32-bit mode do not work
1744 372195 Power PC, xxsel instruction is not always recognized
1746 n-i-bz Fix incorrect (or infinite loop) unwind on RHEL7 x86 and amd64
1747 n-i-bz massif --pages-as-heap=yes does not report peak caused by mmap+munmap
1748 n-i-bz false positive leaks due to aspacemgr merging heap & non heap segments
1749 n-i-bz Fix ppoll_alarm exclusion on OS X
1750 n-i-bz Document brk segment limitation, reference manual in limit reached msg.
1751 n-i-bz Fix clobber list in none/tests/amd64/xacq_xrel.c [valgrind r15737]
1752 n-i-bz Bump allowed shift value for "add.w reg, sp, reg, lsl #N" [vex r3206]
1753 n-i-bz amd64: memcheck false positive with shr %edx
1754 n-i-bz arm3: Allow early writeback of SP base register in "strd rD, [sp, #-16]"
1755 n-i-bz ppc: Fix two cases of PPCAvFpOp vs PPCFpOp enum confusion
1756 n-i-bz arm: Fix incorrect register-number constraint check for LDAEX{,B,H,D}
1757 n-i-bz DHAT: added collection of the metric "tot-blocks-allocd"
1759 (3.12.0.RC1: 20 October 2016, vex r3282, valgrind r16094)
1760 (3.12.0.RC2: 20 October 2016, vex r3282, valgrind r16096)
1761 (3.12.0: 21 October 2016, vex r3282, valgrind r16098)
1765 Release 3.11.0 (22 September 2015)
1766 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1768 3.11.0 is a feature release with many improvements and the usual
1769 collection of bug fixes.
1771 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1772 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1773 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1774 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1775 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1776 X86/MacOSX 10.11, AMD64/MacOSX 10.11 and TILEGX/Linux.
1778 * ================== PLATFORM CHANGES =================
1780 * Support for Solaris/x86 and Solaris/amd64 has been added.
1782 * Preliminary support for Mac OS X 10.11 (El Capitan) has been added.
1784 * Preliminary support for the Tilera TileGX architecture has been added.
1786 * s390x: It is now required for the host to have the "long displacement"
1787 facility. The oldest supported machine model is z990.
1789 * x86: on an SSE2 only host, Valgrind in 32 bit mode now claims to be a
1790 Pentium 4. 3.10.1 wrongly claimed to be a Core 2, which is SSSE3.
1792 * The JIT's register allocator is significantly faster, making the JIT
1793 as a whole somewhat faster, so JIT-intensive activities, for example
1794 program startup, are modestly faster, around 5%.
1796 * There have been changes to the default settings of several command
1797 line flags, as detailed below.
1799 * Intel AVX2 support is more complete (64 bit targets only). On AVX2
1800 capable hosts, the simulated CPUID will now indicate AVX2 support.
1802 * ==================== TOOL CHANGES ====================
1806 - The default value for --leak-check-heuristics has been changed from
1807 "none" to "all". This helps to reduce the number of possibly
1808 lost blocks, in particular for C++ applications.
1810 - The default value for --keep-stacktraces has been changed from
1811 "malloc-then-free" to "malloc-and-free". This has a small cost in
1812 memory (one word per malloc-ed block) but allows Memcheck to show the
1813 3 stacktraces of a dangling reference: where the block was allocated,
1814 where it was freed, and where it is acccessed after being freed.
1816 - The default value for --partial-loads-ok has been changed from "no" to
1817 "yes", so as to avoid false positive errors resulting from some kinds
1818 of vectorised loops.
1820 - A new monitor command 'xb <addr> <len>' shows the validity bits of
1821 <len> bytes at <addr>. The monitor command 'xb' is easier to use
1822 than get_vbits when you need to associate byte data value with
1823 their corresponding validity bits.
1825 - The 'block_list' monitor command has been enhanced:
1826 o it can print a range of loss records
1827 o it now accepts an optional argument 'limited <max_blocks>'
1828 to control the number of blocks printed.
1829 o if a block has been found using a heuristic, then
1830 'block_list' now shows the heuristic after the block size.
1831 o the loss records/blocks to print can be limited to the blocks
1832 found via specified heuristics.
1834 - The C helper functions used to instrument loads on
1835 x86-{linux,solaris} and arm-linux (both 32-bit only) have been
1836 replaced by handwritten assembly sequences. This gives speedups
1837 in the region of 0% to 7% for those targets only.
1839 - A new command line option, --expensive-definedness-checks=yes|no,
1840 has been added. This is useful for avoiding occasional invalid
1841 uninitialised-value errors in optimised code. Watch out for
1842 runtime degradation, as this can be up to 25%. As always, though,
1843 the slowdown is highly application specific. The default setting
1848 - A new monitor command 'all_snapshots <filename>' dumps all
1849 snapshots taken so far.
1853 - Significant memory reduction and moderate speedups for
1854 --history-level=full for applications accessing a lot of memory
1855 with many different stacktraces.
1857 - The default value for --conflict-cache-size=N has been doubled to
1858 2000000. Users that were not using the default value should
1859 preferably also double the value they give.
1861 The default was changed due to the changes in the "full history"
1862 implementation. Doubling the value gives on average a slightly more
1863 complete history and uses similar memory (or significantly less memory
1864 in the worst case) than the previous implementation.
1866 - The Helgrind monitor command 'info locks' now accepts an optional
1867 argument 'lock_addr', which shows information about the lock at the
1870 - When using --history-level=full, the new Helgrind monitor command
1871 'accesshistory <addr> [<len>]' will show the recorded accesses for
1872 <len> (or 1) bytes at <addr>.
1874 * ==================== OTHER CHANGES ====================
1876 * The default value for the --smc-check option has been changed from
1877 "stack" to "all-non-file" on targets that provide automatic D-I
1878 cache coherence (x86, amd64 and s390x). The result is to provide,
1879 by default, transparent support for JIT generated and self-modifying
1880 code on all targets.
1882 * Mac OS X only: the default value for the --dsymutil option has been
1883 changed from "no" to "yes", since any serious usage on Mac OS X
1884 always required it to be "yes".
1886 * The command line options --db-attach and --db-command have been removed.
1887 They were deprecated in 3.10.0.
1889 * When a process dies due to a signal, Valgrind now shows the signal
1890 and the stacktrace at default verbosity (i.e. verbosity 1).
1892 * The address description logic used by Memcheck and Helgrind now
1893 describes addresses in anonymous segments, file mmap-ed segments,
1894 shared memory segments and the brk data segment.
1896 * The new option --error-markers=<begin>,<end> can be used to mark the
1897 begin/end of errors in textual output mode, to facilitate
1898 searching/extracting errors in output files that mix valgrind errors
1899 with program output.
1901 * The new option --max-threads=<number> can be used to change the number
1902 of threads valgrind can handle. The default is 500 threads which
1903 should be more than enough for most applications.
1905 * The new option --valgrind-stacksize=<number> can be used to change the
1906 size of the private thread stacks used by Valgrind. This is useful
1907 for reducing memory use or increasing the stack size if Valgrind
1908 segfaults due to stack overflow.
1910 * The new option --avg-transtab-entry-size=<number> can be used to specify
1911 the expected instrumented block size, either to reduce memory use or
1912 to avoid excessive retranslation.
1914 * Valgrind can be built with Intel's ICC compiler, version 14.0 or later.
1916 * New and modified GDB server monitor features:
1918 - When a signal is reported in GDB, you can now use the GDB convenience
1919 variable $_siginfo to examine detailed signal information.
1921 - Valgrind's gdbserver now allows the user to change the signal
1922 to deliver to the process. So, use 'signal SIGNAL' to continue execution
1923 with SIGNAL instead of the signal reported to GDB. Use 'signal 0' to
1924 continue without passing the signal to the process.
1926 - With GDB >= 7.10, the command 'target remote'
1927 will automatically load the executable file of the process running
1928 under Valgrind. This means you do not need to specify the executable
1929 file yourself, GDB will discover it itself. See GDB documentation about
1930 'qXfer:exec-file:read' packet for more info.
1932 * ==================== FIXED BUGS ====================
1934 The following bugs have been fixed or resolved. Note that "n-i-bz"
1935 stands for "not in bugzilla" -- that is, a bug that was reported to us
1936 but never got a bugzilla entry. We encourage you to file bugs in
1937 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1938 than mailing the developers (or mailing lists) directly -- bugs that
1939 are not entered into bugzilla tend to get forgotten about or ignored.
1941 To see details of a given bug, visit
1942 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1943 where XXXXXX is the bug number as listed below.
1945 116002 VG_(printf): Problems with justification of strings and integers
1946 155125 avoid cutting away file:lineno after long function name
1947 197259 Unsupported arch_prtctl PR_SET_GS option
1948 201152 ppc64: Assertion in ppc32g_dirtyhelper_MFSPR_268_269
1949 201216 Fix Valgrind does not support pthread_sigmask() on OS X
1950 201435 Fix Darwin: -v does not show kernel version
1951 208217 "Warning: noted but unhandled ioctl 0x2000747b" on Mac OS X
1952 211256 Fixed an outdated comment regarding the default platform.
1953 211529 Incomplete call stacks for code compiled by newer versions of MSVC
1954 211926 Avoid compilation warnings in valgrind.h with -pedantic
1955 212291 Fix unhandled syscall: unix:132 (mkfifo) on OS X
1957 226609 Crediting upstream authors in man page
1958 231257 Valgrind omits path when executing script from shebang line
1959 254164 OS X task_info: UNKNOWN task message [id 3405, to mach_task_self() [..]
1960 294065 Improve the pdb file reader by avoiding hardwired absolute pathnames
1961 269360 s390x: Fix addressing mode selection for compare-and-swap
1962 302630 Memcheck: Assertion failed: 'sizeof(UWord) == sizeof(UInt)'
1964 312989 ioctl handling needs to do POST handling on generic ioctls and [..]
1965 319274 Fix unhandled syscall: unix:410 (sigsuspend_nocancel) on OS X
1966 324181 mmap does not handle MAP_32BIT (handle it now, rather than fail it)
1967 327745 Fix valgrind 3.9.0 build fails on Mac OS X 10.6.8
1968 330147 libmpiwrap PMPI_Get_count returns undefined value
1969 333051 mmap of huge pages fails due to incorrect alignment
1971 334802 valgrind does not always explain why a given option is bad
1972 335618 mov.w rN, pc/sp (ARM32)
1973 335785 amd64->IR 0xC4 0xE2 0x75 0x2F (vmaskmovpd)
1978 335907 segfault when running wine's ddrawex/tests/surface.c under valgrind
1979 338602 AVX2 bit in CPUID missing
1980 338606 Strange message for scripts with invalid interpreter
1981 338731 ppc: Fix testuite build for toolchains not supporting -maltivec
1982 338995 shmat with hugepages (SHM_HUGETLB) fails with EINVAL
1983 339045 Getting valgrind to compile and run on OS X Yosemite (10.10)
1985 339156 gdbsrv not called for fatal signal
1986 339215 Valgrind 3.10.0 contain 2013 in copyrights notice
1987 339288 support Cavium Octeon MIPS specific BBIT*32 instructions
1988 339636 Use fxsave64 and fxrstor64 mnemonics instead of old-school rex64 prefix
1989 339442 Fix testsuite build failure on OS X 10.9
1990 339542 Enable compilation with Intel's ICC compiler
1991 339563 The DVB demux DMX_STOP ioctl doesn't have a wrapper
1992 339688 Mac-specific ASM does not support .version directive (cpuid,
1993 tronical and pushfpopf tests)
1994 339745 Valgrind crash when check Marmalade app (partial fix)
1995 339755 Fix known deliberate memory leak in setenv() on Mac OS X 10.9
1996 339778 Linux/TileGx platform support for Valgrind
1997 339780 Fix known uninitialised read in pthread_rwlock_init() on Mac OS X 10.9
1998 339789 Fix none/tests/execve test on Mac OS X 10.9
1999 339808 Fix none/tests/rlimit64_nofile test on Mac OS X 10.9
2000 339820 vex amd64->IR: 0x66 0xF 0x3A 0x63 0xA 0x42 0x74 0x9 (pcmpistri $0x42)
2001 340115 Fix none/tests/cmdline[1|2] tests on systems which define TMPDIR
2002 340392 Allow user to select more accurate definedness checking in memcheck
2003 to avoid invalid complaints on optimised code
2004 340430 Fix some grammatical weirdness in the manual.
2005 341238 Recognize GCC5/DWARFv5 DW_LANG constants (Go, C11, C++11, C++14)
2006 341419 Signal handler ucontext_t not filled out correctly on OS X
2007 341539 VG_(describe_addr) should not describe address as belonging to client
2008 segment if it is past the heap end
2009 341613 Enable building of manythreads and thread-exits tests on Mac OS X
2010 341615 Fix none/tests/darwin/access_extended test on Mac OS X
2011 341698 Valgrind's AESKEYGENASSIST gives wrong result in words 0 and 2 [..]
2012 341789 aarch64: shmat fails with valgrind on ARMv8
2013 341997 MIPS64: Cavium OCTEON insns - immediate operand handled incorrectly
2014 342008 valgrind.h needs type cast [..] for clang/llvm in 64-bit mode
2015 342038 Unhandled syscalls on aarch64 (mbind/get/set_mempolicy)
2016 342063 wrong format specifier for test mcblocklistsearch in gdbserver_tests
2017 342117 Hang when loading PDB file for MSVC compiled Firefox under Wine
2018 342221 socket connect false positive uninit memory for unknown af family
2019 342353 Allow dumping full massif output while valgrind is still running
2020 342571 Valgrind chokes on AVX compare intrinsic with _CMP_GE_QS
2024 342603 Add I2C_SMBUS ioctl support
2025 342635 OS X 10.10 (Yosemite) - missing system calls and fcntl code
2026 342683 Mark memory past the initial brk limit as unaddressable
2027 342783 arm: unhandled instruction 0xEEFE1ACA = "vcvt.s32.f32 s3, s3, #12"
2028 342795 Internal glibc __GI_mempcpy call should be intercepted
2029 342841 s390x: Support instructions fiebr(a) and fidbr(a)
2030 343012 Unhandled syscall 319 (memfd_create)
2031 343069 Patch updating v4l2 API support
2032 343173 helgrind crash during stack unwind
2033 343219 fix GET_STARTREGS for arm
2034 343303 Fix known deliberate memory leak in setenv() on Mac OS X 10.10
2035 343306 OS X 10.10: UNKNOWN mach_msg unhandled MACH_SEND_TRAILER option
2036 343332 Unhandled instruction 0x9E310021 (fcvtmu) on aarch64
2037 343335 unhandled instruction 0x1E638400 (fccmp) aarch64
2038 343523 OS X mach_ports_register: UNKNOWN task message [id 3403, to [..]
2039 343525 OS X host_get_special_port: UNKNOWN host message [id 412, to [..]
2040 343597 ppc64le: incorrect use of offseof macro
2041 343649 OS X host_create_mach_voucher: UNKNOWN host message [id 222, to [..]
2042 343663 OS X 10.10 Memchecj always reports a leak regardless of [..]
2043 343732 Unhandled syscall 144 (setgid) on aarch64
2044 343733 Unhandled syscall 187 (msgctl and related) on aarch64
2045 343802 s390x: False positive "conditional jump or move depends on [..]
2046 343902 --vgdb=yes doesn't break when --xml=yes is used
2047 343967 Don't warn about setuid/setgid/setcap executable for directories
2048 343978 Recognize DWARF5/GCC5 DW_LANG_Fortran 2003 and 2008 constants
2049 344007 accept4 syscall unhandled on arm64 (242) and ppc64 (344)
2050 344033 Helgrind on ARM32 loses track of mutex state in pthread_cond_wait
2051 344054 www - update info for Solaris/illumos
2052 344416 'make regtest' does not work cleanly on OS X
2053 344235 Remove duplicate include of pub_core_aspacemgr.h
2054 344279 syscall sendmmsg on arm64 (269) and ppc32/64 (349) unhandled
2055 344295 syscall recvmmsg on arm64 (243) and ppc32/64 (343) unhandled
2056 344307 2 unhandled syscalls on aarch64/arm64: umount2(39), mount (40)
2057 344314 callgrind_annotate ... warnings about commands containing newlines
2058 344318 socketcall should wrap recvmmsg and sendmmsg
2059 344337 Fix unhandled syscall: mach:41 (_kernelrpc_mach_port_guard_trap)
2060 344416 Fix 'make regtest' does not work cleanly on OS X
2061 344499 Fix compilation for Linux kernel >= 4.0.0
2062 344512 OS X: unhandled syscall: unix:348 (__pthread_chdir),
2063 unix:349 (__pthread_fchdir)
2064 344559 Garbage collection of unused segment names in address space manager
2065 344560 Fix stack traces missing penultimate frame on OS X
2066 344621 Fix memcheck/tests/err_disable4 test on OS X
2067 344686 Fix suppression for pthread_rwlock_init on OS X 10.10
2068 344702 Fix missing libobjc suppressions on OS X 10.10
2070 344936 Fix unhandled syscall: unix:473 (readlinkat) on OS X 10.10
2071 344939 Fix memcheck/tests/xml1 on OS X 10.10
2072 345016 helgrind/tests/locked_vs_unlocked2 is failing sometimes
2073 345079 Fix build problems in VEX/useful/test_main.c
2074 345126 Incorrect handling of VIDIOC_G_AUDIO and G_AUDOUT
2075 345177 arm64: prfm (reg) not implemented
2076 345215 Performance improvements for the register allocator
2077 345248 add support for Solaris OS in valgrind
2078 345338 TIOCGSERIAL and TIOCSSERIAL ioctl support on Linux
2079 345394 Fix memcheck/tests/strchr on OS X
2080 345637 Fix memcheck/tests/sendmsg on OS X
2081 345695 Add POWERPC support for AT_DCACHESIZE and HWCAP2
2082 345824 Fix aspacem segment mismatch: seen with none/tests/bigcode
2083 345887 Fix an assertion in the address space manager
2084 345928 amd64: callstack only contains current function for small stacks
2085 345984 disInstr(arm): unhandled instruction: 0xEE193F1E
2086 345987 MIPS64: Implement cavium LHX instruction
2087 346031 MIPS: Implement support for the CvmCount register (rhwr %0, 31)
2088 346185 Fix typo saving altivec register v24
2089 346267 Compiler warnings for PPC64 code on call to LibVEX_GuestPPC64_get_XER()
2090 and LibVEX_GuestPPC64_get_CR()
2091 346270 Regression tests none/tests/jm_vec/isa_2_07 and
2092 none/tests/test_isa_2_07_part2 have failures on PPC64 little endian
2093 346307 fuse filesystem syscall deadlocks
2094 346324 PPC64 missing support for lbarx, lharx, stbcx and sthcx instructions
2095 346411 MIPS: SysRes::_valEx handling is incorrect
2096 346416 Add support for LL_IOC_PATH2FID and LL_IOC_GETPARENT Lustre ioctls
2097 346474 PPC64 Power 8, spr TEXASRU register not supported
2098 346487 Compiler generates "note" about a future ABI change for PPC64
2099 346562 MIPS64: lwl/lwr instructions are performing 64bit loads
2100 and causing spurious "invalid read of size 8" warnings
2101 346801 Fix link error on OS X: _vgModuleLocal_sf_maybe_extend_stack
2102 347151 Fix suppression for pthread_rwlock_init on OS X 10.8
2103 347233 Fix memcheck/tests/strchr on OS X 10.10 (Haswell)
2104 347322 Power PC regression test cleanup
2105 347379 valgrind --leak-check=full leak errors from system libs on OS X 10.8
2107 347389 unhandled syscall: 373 (Linux ARM syncfs)
2108 347686 Patch set to cleanup PPC64 regtests
2109 347978 Remove bash dependencies where not needed
2110 347982 OS X: undefined symbols for architecture x86_64: "_global" [..]
2111 347988 Memcheck: the 'impossible' happened: unexpected size for Addr (OSX/wine)
2113 348102 Patch updating v4l2 API support
2114 348247 amd64 front end: jno jumps wrongly when overflow is not set
2115 348269 Improve mmap MAP_HUGETLB support.
2116 348334 (ppc) valgrind does not simulate dcbfl - then my program terminates
2117 348345 Assertion fails for negative lineno
2118 348377 Unsupported ARM instruction: yield
2119 348565 Fix detection of command line option availability for clang
2120 348574 vex amd64->IR pcmpistri SSE4.2 unsupported (pcmpistri $0x18)
2121 348728 Fix broken check for VIDIOC_G_ENC_INDEX
2122 348748 Fix redundant condition
2123 348890 Fix clang warning about unsupported --param inline-unit-growth=900
2124 348949 Bogus "ERROR: --ignore-ranges: suspiciously large range"
2125 349034 Add Lustre ioctls LL_IOC_GROUP_LOCK and LL_IOC_GROUP_UNLOCK
2126 349086 Fix UNKNOWN task message [id 3406, to mach_task_self(), [..]
2127 349087 Fix UNKNOWN task message [id 3410, to mach_task_self(), [..]
2128 349626 Implemented additional Xen hypercalls
2129 349769 Clang/osx: ld: warning: -read_only_relocs cannot be used with x86_64
2130 349790 Clean up of the hardware capability checking utilities.
2131 349828 memcpy intercepts memmove causing src/dst overlap error (ppc64 ld.so)
2132 349874 Fix typos in source code
2133 349879 memcheck: add handwritten assembly for helperc_LOADV*
2134 349941 di_notify_mmap might create wrong start/size DebugInfoMapping
2135 350062 vex x86->IR: 0x66 0xF 0x3A 0xB (ROUNDSD) on OS X
2136 350202 Add limited param to 'monitor block_list'
2137 350290 s390x: Support instructions fixbr(a)
2138 350359 memcheck/tests/x86/fxsave hangs indefinetely on OS X
2139 350809 Fix none/tests/async-sigs for Solaris
2140 350811 Remove reference to --db-attach which has been removed.
2141 350813 Memcheck/x86: enable handwritten assembly helpers for x86/Solaris too
2142 350854 hard-to-understand code in VG_(load_ELF)()
2143 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
2144 351386 Solaris: Cannot run ld.so.1 under Valgrind
2145 351474 Fix VG_(iseqsigset) as obvious
2146 351531 Typo in /include/vki/vki-xen-physdev.h header guard
2147 351756 Intercept platform_memchr$VARIANT$Haswell on OS X
2148 351858 ldsoexec support on Solaris
2149 351873 Newer gcc doesn't allow __builtin_tabortdc[i] in ppc32 mode
2150 352130 helgrind reports false races for printfs using mempcpy on FILE* state
2151 352284 s390: Conditional jump depends on uninitialised value(s) in vfprintf
2152 352320 arm64 crash on none/tests/nestedfs
2153 352765 Vbit test fails on Power 6
2154 352768 The mbar instruction is missing from the Power PC support
2155 352769 Power PC program priority register (PPR) is not supported
2156 n-i-bz Provide implementations of certain compiler builtins to support
2157 compilers that may not provide those
2158 n-i-bz Old STABS code is still being compiled, but never used. Remove it.
2159 n-i-bz Fix compilation on distros with glibc < 2.5
2160 n-i-bz (vex 3098) Avoid generation of Neon insns on non-Neon hosts
2161 n-i-bz Enable rt_sigpending syscall on ppc64 linux.
2162 n-i-bz mremap did not work properly on shared memory
2163 n-i-bz Fix incorrect sizeof expression in syswrap-xen.c reported by Coverity
2164 n-i-bz In VALGRIND_PRINTF write out thread name, if any, to xml
2166 (3.11.0.TEST1: 8 September 2015, vex r3187, valgrind r15646)
2167 (3.11.0.TEST2: 21 September 2015, vex r3193, valgrind r15667)
2168 (3.11.0: 22 September 2015, vex r3195, valgrind r15674)
2172 Release 3.10.1 (25 November 2014)
2173 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2174 3.10.1 is a bug fix release. It fixes various bugs reported in 3.10.0
2175 and backports fixes for all reported missing AArch64 ARMv8 instructions
2176 and syscalls from the trunk. If you package or deliver 3.10.0 for others
2177 to use, you might want to consider upgrading to 3.10.1 instead.
2179 The following bugs have been fixed or resolved. Note that "n-i-bz"
2180 stands for "not in bugzilla" -- that is, a bug that was reported to us
2181 but never got a bugzilla entry. We encourage you to file bugs in
2182 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2183 than mailing the developers (or mailing lists) directly -- bugs that
2184 are not entered into bugzilla tend to get forgotten about or ignored.
2186 To see details of a given bug, visit
2187 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2188 where XXXXXX is the bug number as listed below.
2190 335440 arm64: ld1 (single structure) is not implemented
2191 335713 arm64: unhanded instruction: prfm (immediate)
2192 339020 ppc64: memcheck/tests/ppc64/power_ISA2_05 failing in nightly build
2193 339182 ppc64: AvSplat ought to load destination vector register with [..]
2194 339336 PPC64 store quad instruction (stq) is not supposed to change [..]
2195 339433 ppc64 lxvw4x instruction uses four 32-byte loads
2196 339645 Use correct tag names in sys_getdents/64 wrappers
2197 339706 Fix false positive for ioctl(TIOCSIG) on linux
2198 339721 assertion 'check_sibling == sibling' failed in readdwarf3.c ...
2199 339853 arm64 times syscall unknown
2200 339855 arm64 unhandled getsid/setsid syscalls
2201 339858 arm64 dmb sy not implemented
2202 339926 Unhandled instruction 0x1E674001 (frintx) on aarm64
2203 339927 Unhandled instruction 0x9E7100C6 (fcvtmu) on aarch64
2204 339938 disInstr(arm64): unhandled instruction 0x4F8010A4 (fmla)
2206 339940 arm64: unhandled syscall: 83 (sys_fdatasync) + patch
2207 340033 arm64: unhandled insn dmb ishld and some other isb-dmb-dsb variants
2208 340028 unhandled syscalls for arm64 (msync, pread64, setreuid and setregid)
2209 340036 arm64: Unhandled instruction ld4 (multiple structures, no offset)
2210 340236 arm64: unhandled syscalls: mknodat, fchdir, chroot, fchownat
2211 340509 arm64: unhandled instruction fcvtas
2212 340630 arm64: fchmod (52) and fchown (55) syscalls not recognized
2213 340632 arm64: unhandled instruction fcvtas
2214 340722 Resolve "UNKNOWN attrlist flags 0:0x10000000"
2215 340725 AVX2: Incorrect decoding of vpbroadcast{b,w} reg,reg forms
2216 340788 warning: unhandled syscall: 318 (getrandom)
2217 340807 disInstr(arm): unhandled instruction: 0xEE989B20
2218 340856 disInstr(arm64): unhandled instruction 0x1E634C45 (fcsel)
2219 340922 arm64: unhandled getgroups/setgroups syscalls
2220 350251 Fix typo in VEX utility program (test_main.c).
2221 350407 arm64: unhandled instruction ucvtf (vector, integer)
2222 350809 none/tests/async-sigs breaks when run under cron on Solaris
2223 350811 update README.solaris after r15445
2224 350813 Use handwritten memcheck assembly helpers on x86/Solaris [..]
2225 350854 strange code in VG_(load_ELF)()
2226 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
2227 n-i-bz DRD and Helgrind: Handle Imbe_CancelReservation (clrex on ARM)
2228 n-i-bz Add missing ]] to terminate CDATA.
2229 n-i-bz Glibc versions prior to 2.5 do not define PTRACE_GETSIGINFO
2230 n-i-bz Enable sys_fadvise64_64 on arm32.
2231 n-i-bz Add test cases for all remaining AArch64 SIMD, FP and memory insns.
2232 n-i-bz Add test cases for all known arm64 load/store instructions.
2233 n-i-bz PRE(sys_openat): when checking whether ARG1 == VKI_AT_FDCWD [..]
2234 n-i-bz Add detection of old ppc32 magic instructions from bug 278808.
2235 n-i-bz exp-dhat: Implement missing function "dh_malloc_usable_size".
2236 n-i-bz arm64: Implement "fcvtpu w, s".
2237 n-i-bz arm64: implement ADDP and various others
2238 n-i-bz arm64: Implement {S,U}CVTF (scalar, fixedpt).
2239 n-i-bz arm64: enable FCVT{A,N}S X,S.
2241 (3.10.1: 25 November 2014, vex r3026, valgrind r14785)
2245 Release 3.10.0 (10 September 2014)
2246 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2248 3.10.0 is a feature release with many improvements and the usual
2249 collection of bug fixes.
2251 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
2252 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
2253 MIPS64/Linux, ARM/Android, MIPS32/Android, X86/Android, X86/MacOSX 10.9
2254 and AMD64/MacOSX 10.9. Support for MacOSX 10.8 and 10.9 is
2255 significantly improved relative to the 3.9.0 release.
2257 * ================== PLATFORM CHANGES =================
2259 * Support for the 64-bit ARM Architecture (AArch64 ARMv8). This port
2260 is mostly complete, and is usable, but some SIMD instructions are as
2263 * Support for little-endian variant of the 64-bit POWER architecture.
2265 * Support for Android on MIPS32.
2267 * Support for 64bit FPU on MIPS32 platforms.
2269 * Both 32- and 64-bit executables are supported on MacOSX 10.8 and 10.9.
2271 * Configuration for and running on Android targets has changed.
2272 See README.android in the source tree for details.
2274 * ================== DEPRECATED FEATURES =================
2276 * --db-attach is now deprecated and will be removed in the next
2277 valgrind feature release. The built-in GDB server capabilities are
2278 superior and should be used instead. Learn more here:
2279 http://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.gdbserver
2281 * ==================== TOOL CHANGES ====================
2285 - Client code can now selectively disable and re-enable reporting of
2286 invalid address errors in specific ranges using the new client
2287 requests VALGRIND_DISABLE_ADDR_ERROR_REPORTING_IN_RANGE and
2288 VALGRIND_ENABLE_ADDR_ERROR_REPORTING_IN_RANGE.
2290 - Leak checker: there is a new leak check heuristic called
2291 "length64". This is used to detect interior pointers pointing 8
2292 bytes inside a block, on the assumption that the first 8 bytes
2293 holds the value "block size - 8". This is used by
2294 sqlite3MemMalloc, for example.
2296 - Checking of system call parameters: if a syscall parameter
2297 (e.g. bind struct sockaddr, sendmsg struct msghdr, ...) has
2298 several fields not initialised, an error is now reported for each
2299 field. Previously, an error was reported only for the first
2300 uninitialised field.
2302 - Mismatched alloc/free checking: a new flag
2303 --show-mismatched-frees=no|yes [yes] makes it possible to turn off
2304 such checks if necessary.
2308 - Improvements to error messages:
2310 o Race condition error message involving heap allocated blocks also
2311 show the thread number that allocated the raced-on block.
2313 o All locks referenced by an error message are now announced.
2314 Previously, some error messages only showed the lock addresses.
2316 o The message indicating where a lock was first observed now also
2317 describes the address/location of the lock.
2319 - Helgrind now understands the Ada task termination rules and
2320 creates a happens-before relationship between a terminated task
2321 and its master. This avoids some false positives and avoids a big
2322 memory leak when a lot of Ada tasks are created and terminated.
2323 The interceptions are only activated with forthcoming releases of
2324 gnatpro >= 7.3.0w-20140611 and gcc >= 5.0.
2326 - A new GDB server monitor command "info locks" giving the list of
2327 locks, their location, and their status.
2331 - callgrind_control now supports the --vgdb-prefix argument,
2332 which is needed if valgrind was started with this same argument.
2334 * ==================== OTHER CHANGES ====================
2336 * Unwinding through inlined function calls. Stack unwinding can now
2337 make use of Dwarf3 inlined-unwind information if it is available.
2338 The practical effect is that inlined calls become visible in stack
2339 traces. The suppression matching machinery has been adjusted
2340 accordingly. This is controlled by the new option
2341 --read-inline-info=yes|no. Currently this is enabled by default
2342 only on Linux and Android targets and only for the tools Memcheck,
2345 * Valgrind can now read EXIDX unwind information on 32-bit ARM
2346 targets. If an object contains both CFI and EXIDX unwind
2347 information, Valgrind will prefer the CFI over the EXIDX. This
2348 facilitates unwinding through system libraries on arm-android
2351 * Address description logic has been improved and is now common
2352 between Memcheck and Helgrind, resulting in better address
2353 descriptions for some kinds of error messages.
2355 * Error messages about dubious arguments (eg, to malloc or calloc) are
2356 output like other errors. This means that they can be suppressed
2357 and they have a stack trace.
2359 * The C++ demangler has been updated for better C++11 support.
2361 * New and modified GDB server monitor features:
2363 - Thread local variables/storage (__thread) can now be displayed.
2365 - The GDB server monitor command "v.info location <address>"
2366 displays information about an address. The information produced
2367 depends on the tool and on the options given to valgrind.
2368 Possibly, the following are described: global variables, local
2369 (stack) variables, allocated or freed blocks, ...
2371 - The option "--vgdb-stop-at=event1,event2,..." allows the user to
2372 ask the GDB server to stop at the start of program execution, at
2373 the end of the program execution and on Valgrind internal errors.
2375 - A new monitor command "v.info stats" shows various Valgrind core
2376 and tool statistics.
2378 - A new monitor command "v.set hostvisibility" allows the GDB server
2379 to provide access to Valgrind internal host status/memory.
2381 * A new option "--aspace-minaddr=<address>" can in some situations
2382 allow the use of more memory by decreasing the address above which
2383 Valgrind maps memory. It can also be used to solve address
2384 conflicts with system libraries by increasing the default value.
2385 See user manual for details.
2387 * The amount of memory used by Valgrind to store debug info (unwind
2388 info, line number information and symbol data) has been
2389 significantly reduced, even though Valgrind now reads more
2390 information in order to support unwinding of inlined function calls.
2392 * Dwarf3 handling with --read-var-info=yes has been improved:
2394 - Ada and C struct containing VLAs no longer cause a "bad DIE" error
2396 - Code compiled with
2397 -ffunction-sections -fdata-sections -Wl,--gc-sections
2398 no longer causes assertion failures.
2400 * Improved checking for the --sim-hints= and --kernel-variant=
2401 options. Unknown strings are now detected and reported to the user
2404 * The semantics of stack start/end boundaries in the valgrind.h
2405 VALGRIND_STACK_REGISTER client request has been clarified and
2406 documented. The convention is that start and end are respectively
2407 the lowest and highest addressable bytes of the stack.
2409 * ==================== FIXED BUGS ====================
2411 The following bugs have been fixed or resolved. Note that "n-i-bz"
2412 stands for "not in bugzilla" -- that is, a bug that was reported to us
2413 but never got a bugzilla entry. We encourage you to file bugs in
2414 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2415 than mailing the developers (or mailing lists) directly -- bugs that
2416 are not entered into bugzilla tend to get forgotten about or ignored.
2418 To see details of a given bug, visit
2419 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2420 where XXXXXX is the bug number as listed below.
2422 175819 Support for ipv6 socket reporting with --track-fds
2423 232510 make distcheck fails
2424 249435 Analyzing wine programs with callgrind triggers a crash
2425 278972 support for inlined function calls in stacktraces and suppression
2427 291310 FXSAVE instruction marks memory as undefined on amd64
2428 303536 ioctl for SIOCETHTOOL (ethtool(8)) isn't wrapped
2429 308729 vex x86->IR: unhandled instruction bytes 0xf 0x5 (syscall)
2430 315199 vgcore file for threaded app does not show which thread crashed
2431 315952 tun/tap ioctls are not supported
2432 323178 Unhandled instruction: PLDW register (ARM)
2433 323179 Unhandled instruction: PLDW immediate (ARM)
2434 324050 Helgrind: SEGV because of unaligned stack when using movdqa
2435 325110 Add test-cases for Power ISA 2.06 insns: divdo/divdo. and divduo/divduo.
2436 325124 [MIPSEL] Compilation error
2437 325477 Phase 4 support for IBM Power ISA 2.07
2438 325538 cavium octeon mips64, valgrind reported "dumping core" [...]
2439 325628 Phase 5 support for IBM Power ISA 2.07
2440 325714 Empty vgcore but RLIMIT_CORE is big enough (too big)
2441 325751 Missing the two privileged Power PC Transactional Memory Instructions
2442 325816 Phase 6 support for IBM Power ISA 2.07
2443 325856 Make SGCheck fail gracefully on unsupported platforms
2444 326026 Iop names for count leading zeros/sign bits incorrectly imply [..]
2445 326436 DRD: False positive in libstdc++ std::list::push_back
2446 326444 Cavium MIPS Octeon Specific Load Indexed Instructions
2447 326462 Refactor vgdb to isolate invoker stuff into separate module
2448 326469 amd64->IR: 0x66 0xF 0x3A 0x63 0xC1 0xE (pcmpistri 0x0E)
2449 326623 DRD: false positive conflict report in a field assignment
2450 326724 Valgrind does not compile on OSX 1.9 Mavericks
2451 326816 Intercept for __strncpy_sse2_unaligned missing?
2452 326921 coregrind fails to compile m_trampoline.S with MIPS/Linux port of V
2453 326983 Clear direction flag after tests on amd64.
2454 327212 Do not prepend the current directory to absolute path names.
2455 327223 Support for Cavium MIPS Octeon Atomic and Count Instructions
2456 327238 Callgrind Assertion 'passed <= last_bb->cjmp_count' failed
2457 327284 s390x: Fix translation of the risbg instruction
2458 327639 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x34
2459 327837 dwz compressed alternate .debug_info and .debug_str not read correctly
2460 327916 DW_TAG_typedef may have no name
2461 327943 s390x: add a redirection for the 'index' function
2462 328100 XABORT not implemented
2463 328205 Implement additional Xen hypercalls
2464 328454 add support Backtraces with ARM unwind tables (EXIDX)
2465 328455 s390x: SIGILL after emitting wrong register pair for ldxbr
2466 328711 valgrind.1 manpage "memcheck options" section is badly generated
2467 328878 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x14
2468 329612 Incorrect handling of AT_BASE for image execution
2469 329694 clang warns about using uninitialized variable
2470 329956 valgrind crashes when lmw/stmw instructions are used on ppc64
2471 330228 mmap must align to VKI_SHMLBA on mips32
2472 330257 LLVM does not support `-mno-dynamic-no-pic` option
2473 330319 amd64->IR: unhandled instruction bytes: 0xF 0x1 0xD5 (xend)
2474 330459 --track-fds=yes doesn't track eventfds
2475 330469 Add clock_adjtime syscall support
2476 330594 Missing sysalls on PowerPC / uClibc
2477 330622 Add test to regression suite for POWER instruction: dcbzl
2478 330939 Support for AMD's syscall instruction on x86
2480 330941 Typo in PRE(poll) syscall wrapper
2481 331057 unhandled instruction: 0xEEE01B20 (vfma.f64) (has patch)
2482 331254 Fix expected output for memcheck/tests/dw4
2483 331255 Fix race condition in test none/tests/coolo_sigaction
2484 331257 Fix type of jump buffer in test none/tests/faultstatus
2485 331305 configure uses bash specific syntax
2486 331337 s390x WARNING: unhandled syscall: 326 (dup3)
2487 331380 Syscall param timer_create(evp) points to uninitialised byte(s)
2488 331476 Patch to handle ioctl 0x5422 on Linux (x86 and amd64)
2489 331829 Unexpected ioctl opcode sign extension
2490 331830 ppc64: WARNING: unhandled syscall: 96/97
2491 331839 drd/tests/sem_open specifies invalid semaphore name
2492 331847 outcome of drd/tests/thread_name is nondeterministic
2493 332037 Valgrind cannot handle Thumb "add pc, reg"
2494 332055 drd asserts on platforms with VG_STACK_REDZONE_SZB == 0 and
2495 consistency checks enabled
2496 332263 intercepts for pthread_rwlock_timedrdlock and
2497 pthread_rwlock_timedwrlock are incorrect
2498 332265 drd could do with post-rwlock_init and pre-rwlock_destroy
2500 332276 Implement additional Xen hypercalls
2501 332658 ldrd.w r1, r2, [PC, #imm] does not adjust for 32bit alignment
2502 332765 Fix ms_print to create temporary files in a proper directory
2503 333072 drd: Add semaphore annotations
2504 333145 Tests for missaligned PC+#imm access for arm
2505 333228 AAarch64 Missing instruction encoding: mrs %[reg], ctr_el0
2506 333230 AAarch64 missing instruction encodings: dc, ic, dsb.
2507 333248 WARNING: unhandled syscall: unix:443
2508 333428 ldr.w pc [rD, #imm] instruction leads to assertion
2509 333501 cachegrind: assertion: Cache set count is not a power of two.
2512 333666 Recognize MPX instructions and bnd prefix.
2513 333788 Valgrind does not support the CDROM_DISC_STATUS ioctl (has patch)
2514 333817 Valgrind reports the memory areas written to by the SG_IO
2516 334049 lzcnt fails silently (x86_32)
2517 334384 Valgrind does not have support Little Endian support for
2519 334585 recvmmsg unhandled (+patch) (arm)
2520 334705 sendmsg and recvmsg should guard against bogus msghdr fields.
2521 334727 Build fails with -Werror=format-security
2522 334788 clarify doc about --log-file initial program directory
2523 334834 PPC64 Little Endian support, patch 2
2524 334836 PPC64 Little Endian support, patch 3 testcase fixes
2525 334936 patch to fix false positives on alsa SNDRV_CTL_* ioctls
2526 335034 Unhandled ioctl: HCIGETDEVLIST
2527 335155 vgdb, fix error print statement.
2528 335262 arm64: movi 8bit version is not supported
2529 335263 arm64: dmb instruction is not implemented
2530 335441 unhandled ioctl 0x8905 (SIOCATMARK) when running wine under valgrind
2531 335496 arm64: sbc/abc instructions are not implemented
2532 335554 arm64: unhandled instruction: abs
2533 335564 arm64: unhandled instruction: fcvtpu Xn, Sn
2534 335735 arm64: unhandled instruction: cnt
2535 335736 arm64: unhandled instruction: uaddlv
2536 335848 arm64: unhandled instruction: {s,u}cvtf
2537 335902 arm64: unhandled instruction: sli
2538 335903 arm64: unhandled instruction: umull (vector)
2539 336055 arm64: unhandled instruction: mov (element)
2540 336062 arm64: unhandled instruction: shrn{,2}
2541 336139 mip64: [...] valgrind hangs and spins on a single core [...]
2542 336189 arm64: unhandled Instruction: mvn
2543 336435 Valgrind hangs in pthread_spin_lock consuming 100% CPU
2544 336619 valgrind --read-var-info=yes doesn't handle DW_TAG_restrict_type
2545 336772 Make moans about unknown ioctls more informative
2546 336957 Add a section about the Solaris/illumos port on the webpage
2547 337094 ifunc wrapper is broken on ppc64
2548 337285 fcntl commands F_OFD_SETLK, F_OFD_SETLKW, and F_OFD_GETLK not supported
2549 337528 leak check heuristic for block prefixed by length as 64bit number
2550 337740 Implement additional Xen hypercalls
2551 337762 guest_arm64_toIR.c:4166 (dis_ARM64_load_store): Assertion `0' failed.
2552 337766 arm64-linux: unhandled syscalls mlock (228) and mlockall (230)
2553 337871 deprecate --db-attach
2554 338023 Add support for all V4L2/media ioctls
2555 338024 inlined functions are not shown if DW_AT_ranges is used
2556 338106 Add support for 'kcmp' syscall
2557 338115 DRD: computed conflict set differs from actual after fork
2558 338160 implement display of thread local storage in gdbsrv
2559 338205 configure.ac and check for -Wno-tautological-compare
2560 338300 coredumps are missing one byte of every segment
2561 338445 amd64 vbit-test fails with unknown opcodes used by arm64 VEX
2562 338499 --sim-hints parsing broken due to wrong order in tokens
2563 338615 suppress glibc 2.20 optimized strcmp implementation for ARMv7
2564 338681 Unable to unwind through clone thread created on i386-linux
2565 338698 race condition between gdbsrv and vgdb on startup
2566 338703 helgrind on arm-linux gets false positives in dynamic loader
2567 338791 alt dwz files can be relative of debug/main file
2568 338878 on MacOS: assertion 'VG_IS_PAGE_ALIGNED(clstack_end+1)' failed
2569 338932 build V-trunk with gcc-trunk
2570 338974 glibc 2.20 changed size of struct sigaction sa_flags field on s390
2571 345079 Fix build problems in VEX/useful/test_main.c
2572 n-i-bz Fix KVM_CREATE_IRQCHIP ioctl handling
2573 n-i-bz s390x: Fix memory corruption for multithreaded applications
2574 n-i-bz vex arm->IR: allow PC as basereg in some LDRD cases
2575 n-i-bz internal error in Valgrind if vgdb transmit signals when ptrace invoked
2576 n-i-bz Fix mingw64 support in valgrind.h (dev@, 9 May 2014)
2577 n-i-bz drd manual: Document how to C++11 programs that use class "std::thread"
2578 n-i-bz Add command-line option --default-suppressions
2579 n-i-bz Add support for BLKDISCARDZEROES ioctl
2580 n-i-bz ppc32/64: fix a regression with the mtfsb0/mtfsb1 instructions
2581 n-i-bz Add support for sys_pivot_root and sys_unshare
2583 (3.10.0.BETA1: 2 September 2014, vex r2940, valgrind r14428)
2584 (3.10.0.BETA2: 8 September 2014, vex r2950, valgrind r14503)
2585 (3.10.0: 10 September 2014, vex r2950, valgrind r14514)
2589 Release 3.9.0 (31 October 2013)
2590 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2591 3.9.0 is a feature release with many improvements and the usual
2592 collection of bug fixes.
2594 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2595 PPC64/Linux, S390X/Linux, MIPS32/Linux, MIPS64/Linux, ARM/Android,
2596 X86/Android, X86/MacOSX 10.7 and AMD64/MacOSX 10.7. Support for
2597 MacOSX 10.8 is significantly improved relative to the 3.8.0 release.
2599 * ================== PLATFORM CHANGES =================
2601 * Support for MIPS64 LE and BE running Linux. Valgrind has been
2602 tested on MIPS64 Debian Squeeze and Debian Wheezy distributions.
2604 * Support for MIPS DSP ASE on MIPS32 platforms.
2606 * Support for s390x Decimal Floating Point instructions on hosts that
2607 have the DFP facility installed.
2609 * Support for POWER8 (Power ISA 2.07) instructions
2611 * Support for Intel AVX2 instructions. This is available only on 64
2614 * Initial support for Intel Transactional Synchronization Extensions,
2617 * Initial support for Hardware Transactional Memory on POWER.
2619 * Improved support for MacOSX 10.8 (64-bit only). Memcheck can now
2620 run large GUI apps tolerably well.
2622 * ==================== TOOL CHANGES ====================
2626 - Improvements in handling of vectorised code, leading to
2627 significantly fewer false error reports. You need to use the flag
2628 --partial-loads-ok=yes to get the benefits of these changes.
2630 - Better control over the leak checker. It is now possible to
2631 specify which leak kinds (definite/indirect/possible/reachable)
2632 should be displayed, which should be regarded as errors, and which
2633 should be suppressed by a given leak suppression. This is done
2634 using the options --show-leak-kinds=kind1,kind2,..,
2635 --errors-for-leak-kinds=kind1,kind2,.. and an optional
2636 "match-leak-kinds:" line in suppression entries, respectively.
2638 Note that generated leak suppressions contain this new line and
2639 are therefore more specific than in previous releases. To get the
2640 same behaviour as previous releases, remove the "match-leak-kinds:"
2641 line from generated suppressions before using them.
2643 - Reduced "possible leak" reports from the leak checker by the use
2644 of better heuristics. The available heuristics provide detection
2645 of valid interior pointers to std::stdstring, to new[] allocated
2646 arrays with elements having destructors and to interior pointers
2647 pointing to an inner part of a C++ object using multiple
2648 inheritance. They can be selected individually using the
2649 option --leak-check-heuristics=heur1,heur2,...
2651 - Better control of stacktrace acquisition for heap-allocated
2652 blocks. Using the --keep-stacktraces option, it is possible to
2653 control independently whether a stack trace is acquired for each
2654 allocation and deallocation. This can be used to create better
2655 "use after free" errors or to decrease Valgrind's resource
2656 consumption by recording less information.
2658 - Better reporting of leak suppression usage. The list of used
2659 suppressions (shown when the -v option is given) now shows, for
2660 each leak suppressions, how many blocks and bytes it suppressed
2661 during the last leak search.
2665 - False errors resulting from the use of statically initialised
2666 mutexes and condition variables (PTHREAD_MUTEX_INITIALISER, etc)
2669 - False errors resulting from the use of pthread_cond_waits that
2670 timeout, have been removed.
2672 * ==================== OTHER CHANGES ====================
2674 * Some attempt to tune Valgrind's space requirements to the expected
2675 capabilities of the target:
2677 - The default size of the translation cache has been reduced from 8
2678 sectors to 6 on Android platforms, since each sector occupies
2679 about 40MB when using Memcheck.
2681 - The default size of the translation cache has been increased to 16
2682 sectors on all other platforms, reflecting the fact that large
2683 applications require instrumentation and storage of huge amounts
2684 of code. For similar reasons, the number of memory mapped
2685 segments that can be tracked has been increased by a factor of 6.
2687 - In all cases, the maximum number of sectors in the translation
2688 cache can be controlled by the new flag --num-transtab-sectors.
2690 * Changes in how debug info (line numbers, etc) is read:
2692 - Valgrind no longer temporarily mmaps the entire object to read
2693 from it. Instead, reading is done through a small fixed sized
2694 buffer. This avoids virtual memory usage spikes when Valgrind
2695 reads debuginfo from large shared objects.
2697 - A new experimental remote debug info server. Valgrind can read
2698 debug info from a different machine (typically, a build host)
2699 where debuginfo objects are stored. This can save a lot of time
2700 and hassle when running Valgrind on resource-constrained targets
2701 (phones, tablets) when the full debuginfo objects are stored
2702 somewhere else. This is enabled by the --debuginfo-server=
2705 - Consistency checking between main and debug objects can be
2706 disabled using the --allow-mismatched-debuginfo option.
2708 * Stack unwinding by stack scanning, on ARM. Unwinding by stack
2709 scanning can recover stack traces in some cases when the normal
2710 unwind mechanisms fail. Stack scanning is best described as "a
2711 nasty, dangerous and misleading hack" and so is disabled by default.
2712 Use --unw-stack-scan-thresh and --unw-stack-scan-frames to enable
2715 * Detection and merging of recursive stack frame cycles. When your
2716 program has recursive algorithms, this limits the memory used by
2717 Valgrind for recorded stack traces and avoids recording
2718 uninteresting repeated calls. This is controlled by the command
2719 line option --merge-recursive-frame and by the monitor command
2720 "v.set merge-recursive-frames".
2722 * File name and line numbers for used suppressions. The list of used
2723 suppressions (shown when the -v option is given) now shows, for each
2724 used suppression, the file name and line number where the suppression
2727 * New and modified GDB server monitor features:
2729 - valgrind.h has a new client request, VALGRIND_MONITOR_COMMAND,
2730 that can be used to execute gdbserver monitor commands from the
2733 - A new monitor command, "v.info open_fds", that gives the list of
2734 open file descriptors and additional details.
2736 - An optional message in the "v.info n_errs_found" monitor command,
2737 for example "v.info n_errs_found test 1234 finished", allowing a
2738 comment string to be added to the process output, perhaps for the
2739 purpose of separating errors of different tests or test phases.
2741 - A new monitor command "v.info execontext" that shows information
2742 about the stack traces recorded by Valgrind.
2744 - A new monitor command "v.do expensive_sanity_check_general" to run
2745 some internal consistency checks.
2747 * New flag --sigill-diagnostics to control whether a diagnostic
2748 message is printed when the JIT encounters an instruction it can't
2749 translate. The actual behavior -- delivery of SIGILL to the
2750 application -- is unchanged.
2752 * The maximum amount of memory that Valgrind can use on 64 bit targets
2753 has been increased from 32GB to 64GB. This should make it possible
2754 to run applications on Memcheck that natively require up to about 35GB.
2756 * ==================== FIXED BUGS ====================
2758 The following bugs have been fixed or resolved. Note that "n-i-bz"
2759 stands for "not in bugzilla" -- that is, a bug that was reported to us
2760 but never got a bugzilla entry. We encourage you to file bugs in
2761 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2762 than mailing the developers (or mailing lists) directly -- bugs that
2763 are not entered into bugzilla tend to get forgotten about or ignored.
2765 To see details of a given bug, visit
2766 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2767 where XXXXXX is the bug number as listed below.
2769 123837 system call: 4th argument is optional, depending on cmd
2770 135425 memcheck should tell you where Freed blocks were Mallocd
2771 164485 VG_N_SEGNAMES and VG_N_SEGMENTS are (still) too small
2772 207815 Adds some of the drm ioctls to syswrap-linux.c
2773 251569 vex amd64->IR: 0xF 0x1 0xF9 0xBF 0x90 0xD0 0x3 0x0 (RDTSCP)
2774 252955 Impossible to compile with ccache
2775 253519 Memcheck reports auxv pointer accesses as invalid reads.
2776 263034 Crash when loading some PPC64 binaries
2777 269599 Increase deepest backtrace
2778 274695 s390x: Support "compare to/from logical" instructions (z196)
2779 275800 s390x: Autodetect cache info (part 2)
2780 280271 Valgrind reports possible memory leaks on still-reachable std::string
2781 284540 Memcheck shouldn't count suppressions matching still-reachable [..]
2782 289578 Backtraces with ARM unwind tables (stack scan flags)
2783 296311 Wrong stack traces due to -fomit-frame-pointer (x86)
2784 304832 ppc32: build failure
2785 305431 Use find_buildid shdr fallback for separate .debug files
2786 305728 Add support for AVX2 instructions
2787 305948 ppc64: code generation for ShlD64 / ShrD64 asserts
2788 306035 s390x: Fix IR generation for LAAG and friends
2789 306054 s390x: Condition code computation for convert-to-int/logical
2790 306098 s390x: alternate opcode form for convert to/from fixed
2791 306587 Fix cache line detection from auxiliary vector for PPC.
2792 306783 Mips unhandled syscall : 4025 / 4079 / 4182
2793 307038 DWARF2 CFI reader: unhandled DW_OP_ opcode 0x8 (DW_OP_const1u et al)
2794 307082 HG false positive: pthread_cond_destroy: destruction of unknown CV
2795 307101 sys_capget second argument can be NULL
2796 307103 sys_openat: If pathname is absolute, then dirfd is ignored.
2797 307106 amd64->IR: f0 0f c0 02 (lock xadd byte)
2798 307113 s390x: DFP support
2799 307141 valgrind does't work in mips-linux system
2800 307155 filter_gdb should filter out syscall-template.S T_PSEUDO
2801 307285 x86_amd64 feature test for avx in test suite is wrong
2802 307290 memcheck overlap testcase needs memcpy version filter
2803 307463 Please add "&limit=0" to the "all open bugs" link
2804 307465 --show-possibly-lost=no should reduce the error count / exit code
2805 307557 Leaks on Mac OS X 10.7.5 libraries at ImageLoader::recursiveInit[..]
2806 307729 pkgconfig support broken valgrind.pc
2807 307828 Memcheck false errors SSE optimized wcscpy, wcscmp, wcsrchr, wcschr
2808 307955 Building valgrind 3.7.0-r4 fails in Gentoo AMD64 when using clang
2809 308089 Unhandled syscall on ppc64: prctl
2810 308135 PPC32 MPC8xx has 16 bytes cache size
2811 308321 testsuite memcheck filter interferes with gdb_filter
2813 308341 vgdb should report process exit (or fatal signal)
2814 308427 s390 memcheck reports tsearch cjump/cmove depends on uninit
2815 308495 Remove build dependency on installed Xen headers
2816 308573 Internal error on 64-bit instruction executed in 32-bit mode
2818 308627 pmovmskb validity bit propagation is imprecise
2819 308644 vgdb command for having the info for the track-fds option
2820 308711 give more info about aspacemgr and arenas in out_of_memory
2821 308717 ARM: implement fixed-point VCVT.F64.[SU]32
2822 308718 ARM implement SMLALBB family of instructions
2823 308886 Missing support for PTRACE_SET/GETREGSET
2824 308930 syscall name_to_handle_at (303 on amd64) not handled
2825 309229 V-bit tester does not report number of tests generated
2826 309323 print unrecognized instuction on MIPS
2827 309425 Provide a --sigill-diagnostics flag to suppress illegal [..]
2828 309427 SSE optimized stpncpy trigger uninitialised value [..] errors
2829 309430 Self hosting ppc64 encounters a vassert error on operand type
2830 309600 valgrind is a bit confused about 0-sized sections
2831 309823 Generate errors for still reachable blocks
2832 309921 PCMPISTRI validity bit propagation is imprecise
2833 309922 none/tests/ppc64/test_dfp5 sometimes fails
2834 310169 The Iop_CmpORD class of Iops is not supported by the vbit checker.
2835 310424 --read-var-info does not properly describe static variables
2836 310792 search additional path for debug symbols
2837 310931 s390x: Message-security assist (MSA) instruction extension [..]
2838 311100 PPC DFP implementation of the integer operands is inconsistent [..]
2839 311318 ARM: "128-bit constant is not implemented" error message
2840 311407 ssse3 bcopy (actually converted memcpy) causes invalid read [..]
2841 311690 V crashes because it redirects branches inside of a redirected function
2842 311880 x86_64: make regtest hangs at shell_valid1
2843 311922 WARNING: unhandled syscall: 170
2845 312171 ppc: insn selection for DFP
2846 312571 Rounding mode call wrong for the DFP Iops [..]
2847 312620 Change to Iop_D32toD64 [..] for s390 DFP support broke ppc [..]
2848 312913 Dangling pointers error should also report the alloc stack trace
2849 312980 Building on Mountain Lion generates some compiler warnings
2850 313267 Adding MIPS64/Linux port to Valgrind
2853 313811 Buffer overflow in assert_fail
2854 314099 coverity pointed out error in VEX guest_ppc_toIR.c insn_suffix
2855 314269 ppc: dead code in insn selection
2856 314718 ARM: implement integer divide instruction (sdiv and udiv)
2857 315345 cl-format.xml and callgrind/dump.c don't agree on using cfl= or cfi=
2858 315441 sendmsg syscall should ignore unset msghdr msg_flags
2859 315534 msgrcv inside a thread causes valgrind to hang (block)
2860 315545 Assertion '(UChar*)sec->tt[tteNo].tcptr <= (UChar*)hcode' failed
2861 315689 disInstr(thumb): unhandled instruction: 0xF852 0x0E10 (LDRT)
2862 315738 disInstr(arm): unhandled instruction: 0xEEBE0BEE (vcvt.s32.f64)
2863 315959 valgrind man page has bogus SGCHECK (and no BBV) OPTIONS section
2864 316144 valgrind.1 manpage contains unknown ??? strings [..]
2865 316145 callgrind command line options in manpage reference (unknown) [..]
2866 316145 callgrind command line options in manpage reference [..]
2867 316181 drd: Fixed a 4x slowdown for certain applications
2868 316503 Valgrind does not support SSE4 "movntdqa" instruction
2869 316535 Use of |signed int| instead of |size_t| in valgrind messages
2870 316696 fluidanimate program of parsec 2.1 stuck
2871 316761 syscall open_by_handle_at (304 on amd64, 342 on x86) not handled
2872 317091 Use -Wl,-Ttext-segment when static linking if possible [..]
2873 317186 "Impossible happens" when occurs VCVT instruction on ARM
2874 317318 Support for Threading Building Blocks "scalable_malloc"
2875 317444 amd64->IR: 0xC4 0x41 0x2C 0xC2 0xD2 0x8 (vcmpeq_uqps)
2876 317461 Fix BMI assembler configure check and avx2/bmi/fma vgtest prereqs
2877 317463 bmi testcase IR SANITY CHECK FAILURE
2878 317506 memcheck/tests/vbit-test fails with unknown opcode after [..]
2879 318050 libmpiwrap fails to compile with out-of-source build
2880 318203 setsockopt handling needs to handle SOL_SOCKET/SO_ATTACH_FILTER
2881 318643 annotate_trace_memory tests infinite loop on arm and ppc [..]
2882 318773 amd64->IR: 0xF3 0x48 0x0F 0xBC 0xC2 0xC3 0x66 0x0F
2883 318929 Crash with: disInstr(thumb): 0xF321 0x0001 (ssat16)
2884 318932 Add missing PPC64 and PPC32 system call support
2885 319235 --db-attach=yes is broken with Yama (ptrace scoping) enabled
2886 319395 Crash with unhandled instruction on STRT (Thumb) instructions
2887 319494 VEX Makefile-gcc standalone build update after r2702
2888 319505 [MIPSEL] Crash: unhandled UNRAY operator.
2889 319858 disInstr(thumb): unhandled instruction on instruction STRBT
2890 319932 disInstr(thumb): unhandled instruction on instruction STRHT
2891 320057 Problems when we try to mmap more than 12 memory pages on MIPS32
2892 320063 Memory from PTRACE_GET_THREAD_AREA is reported uninitialised
2893 320083 disInstr(thumb): unhandled instruction on instruction LDRBT
2894 320116 bind on AF_BLUETOOTH produces warnings because of sockaddr_rc padding
2895 320131 WARNING: unhandled syscall: 369 on ARM (prlimit64)
2896 320211 Stack buffer overflow in ./coregrind/m_main.c with huge TMPDIR
2897 320661 vgModuleLocal_read_elf_debug_info(): "Assertion '!di->soname'
2898 320895 add fanotify support (patch included)
2899 320998 vex amd64->IR pcmpestri and pcmpestrm SSE4.2 instruction
2900 321065 Valgrind updates for Xen 4.3
2901 321148 Unhandled instruction: PLI (Thumb 1, 2, 3)
2902 321363 Unhandled instruction: SSAX (ARM + Thumb)
2903 321364 Unhandled instruction: SXTAB16 (ARM + Thumb)
2904 321466 Unhandled instruction: SHASX (ARM + Thumb)
2905 321467 Unhandled instruction: SHSAX (ARM + Thumb)
2906 321468 Unhandled instruction: SHSUB16 (ARM + Thumb)
2907 321619 Unhandled instruction: SHSUB8 (ARM + Thumb)
2908 321620 Unhandled instruction: UASX (ARM + Thumb)
2909 321621 Unhandled instruction: USAX (ARM + Thumb)
2910 321692 Unhandled instruction: UQADD16 (ARM + Thumb)
2911 321693 Unhandled instruction: LDRSBT (Thumb)
2912 321694 Unhandled instruction: UQASX (ARM + Thumb)
2913 321696 Unhandled instruction: UQSAX (Thumb + ARM)
2914 321697 Unhandled instruction: UHASX (ARM + Thumb)
2915 321703 Unhandled instruction: UHSAX (ARM + Thumb)
2916 321704 Unhandled instruction: REVSH (ARM + Thumb)
2917 321730 Add cg_diff and cg_merge man pages
2918 321738 Add vgdb and valgrind-listener man pages
2920 321891 Unhandled instruction: LDRHT (Thumb)
2921 321960 pthread_create() then alloca() causing invalid stack write errors
2922 321969 ppc32 and ppc64 don't support [lf]setxattr
2923 322254 Show threadname together with tid if set by application
2924 322294 Add initial support for IBM Power ISA 2.07
2925 322368 Assertion failure in wqthread_hijack under OS X 10.8
2926 322563 vex mips->IR: 0x70 0x83 0xF0 0x3A
2927 322807 VALGRIND_PRINTF_BACKTRACE writes callstack to xml and text to stderr
2928 322851 0bXXX binary literal syntax is not standard
2929 323035 Unhandled instruction: LDRSHT(Thumb)
2930 323036 Unhandled instruction: SMMLS (ARM and Thumb)
2931 323116 The memcheck/tests/ppc64/power_ISA2_05.c fails to build [..]
2932 323175 Unhandled instruction: SMLALD (ARM + Thumb)
2933 323177 Unhandled instruction: SMLSLD (ARM + Thumb)
2934 323432 Calling pthread_cond_destroy() or pthread_mutex_destroy() [..]
2935 323437 Phase 2 support for IBM Power ISA 2.07
2936 323713 Support mmxext (integer sse) subset on i386 (athlon)
2937 323803 Transactional memory instructions are not supported for Power
2938 323893 SSE3 not available on amd cpus in valgrind
2939 323905 Probable false positive from Valgrind/drd on close()
2940 323912 valgrind.h header isn't compatible for mingw64
2941 324047 Valgrind doesn't support [LDR,ST]{S}[B,H]T ARM instructions
2942 324149 helgrind: When pthread_cond_timedwait returns ETIMEDOUT [..]
2943 324181 mmap does not handle MAP_32BIT
2944 324227 memcheck false positive leak when a thread calls exit+block [..]
2945 324421 Support for fanotify API on ARM architecture
2946 324514 gdbserver monitor cmd output behaviour consistency [..]
2947 324518 ppc64: Emulation of dcbt instructions does not handle [..]
2948 324546 none/tests/ppc32 test_isa_2_07_part2 requests -m64
2949 324582 When access is made to freed memory, report both allocation [..]
2950 324594 Fix overflow computation for Power ISA 2.06 insns: mulldo/mulldo.
2951 324765 ppc64: illegal instruction when executing none/tests/ppc64/jm-misc
2952 324816 Incorrect VEX implementation for xscvspdp/xvcvspdp for SNaN inputs
2953 324834 Unhandled instructions in Microsoft C run-time for x86_64
2954 324894 Phase 3 support for IBM Power ISA 2.07
2955 326091 drd: Avoid false race reports from optimized strlen() impls
2956 326113 valgrind libvex hwcaps error on AMD64
2957 n-i-bz Some wrong command line options could be ignored
2958 n-i-bz patch to allow fair-sched on android
2959 n-i-bz report error for vgdb snapshot requested before execution
2960 n-i-bz same as 303624 (fixed in 3.8.0), but for x86 android
2962 (3.9.0: 31 October 2013, vex r2796, valgrind r13708)
2966 Release 3.8.1 (19 September 2012)
2967 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2968 3.8.1 is a bug fix release. It fixes some assertion failures in 3.8.0
2969 that occur moderately frequently in real use cases, adds support for
2970 some missing instructions on ARM, and fixes a deadlock condition on
2971 MacOSX. If you package or deliver 3.8.0 for others to use, you might
2972 want to consider upgrading to 3.8.1 instead.
2974 The following bugs have been fixed or resolved. Note that "n-i-bz"
2975 stands for "not in bugzilla" -- that is, a bug that was reported to us
2976 but never got a bugzilla entry. We encourage you to file bugs in
2977 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2978 than mailing the developers (or mailing lists) directly -- bugs that
2979 are not entered into bugzilla tend to get forgotten about or ignored.
2981 To see details of a given bug, visit
2982 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2983 where XXXXXX is the bug number as listed below.
2986 289584 Unhandled instruction: 0xF 0x29 0xE5 (MOVAPS)
2987 295808 amd64->IR: 0xF3 0xF 0xBC 0xC0 (TZCNT)
2988 298281 wcslen causes false(?) uninitialised value warnings
2989 301281 valgrind hangs on OS X when the process calls system()
2990 304035 disInstr(arm): unhandled instruction 0xE1023053
2991 304867 implement MOVBE instruction in x86 mode
2992 304980 Assertion 'lo <= hi' failed in vgModuleLocal_find_rx_mapping
2993 305042 amd64: implement 0F 7F encoding of movq between two registers
2994 305199 ARM: implement QDADD and QDSUB
2995 305321 amd64->IR: 0xF 0xD 0xC (prefetchw)
2996 305513 killed by fatal signal: SIGSEGV
2997 305690 DRD reporting invalid semaphore when sem_trywait fails
2998 305926 Invalid alignment checks for some AVX instructions
2999 306297 disInstr(thumb): unhandled instruction 0xE883 0x000C
3000 306310 3.8.0 release tarball missing some files
3001 306612 RHEL 6 glibc-2.X default suppressions need /lib*/libc-*patterns
3002 306664 vex amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x46 0x66 0xF
3003 n-i-bz shmat of a segment > 4Gb does not work
3004 n-i-bz simulate_control_c script wrong USR1 signal number on mips
3005 n-i-bz vgdb ptrace calls wrong on mips [...]
3006 n-i-bz Fixes for more MPI false positives
3007 n-i-bz exp-sgcheck's memcpy causes programs to segfault
3008 n-i-bz OSX build w/ clang: asserts at startup
3009 n-i-bz Incorrect undef'dness prop for Iop_DPBtoBCD and Iop_BCDtoDPB
3010 n-i-bz fix a couple of union tag-vs-field mixups
3011 n-i-bz OSX: use __NR_poll_nocancel rather than __NR_poll
3013 The following bugs were fixed in 3.8.0 but not listed in this NEWS
3016 254088 Valgrind should know about UD2 instruction
3019 304754 NEWS blows TeX's little mind
3021 (3.8.1: 19 September 2012, vex r2537, valgrind r12996)
3025 Release 3.8.0 (10 August 2012)
3026 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3027 3.8.0 is a feature release with many improvements and the usual
3028 collection of bug fixes.
3030 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3031 PPC64/Linux, S390X/Linux, MIPS/Linux, ARM/Android, X86/Android,
3032 X86/MacOSX 10.6/10.7 and AMD64/MacOSX 10.6/10.7. Support for recent
3033 distros and toolchain components (glibc 2.16, gcc 4.7) has been added.
3034 There is initial support for MacOSX 10.8, but it is not usable for
3035 serious work at present.
3037 * ================== PLATFORM CHANGES =================
3039 * Support for MIPS32 platforms running Linux. Valgrind has been
3040 tested on MIPS32 and MIPS32r2 platforms running different Debian
3041 Squeeze and MeeGo distributions. Both little-endian and big-endian
3042 cores are supported. The tools Memcheck, Massif and Lackey have
3043 been tested and are known to work. See README.mips for more details.
3045 * Preliminary support for Android running on x86.
3047 * Preliminary (as-yet largely unusable) support for MacOSX 10.8.
3049 * Support for Intel AVX instructions and for AES instructions. This
3050 support is available only for 64 bit code.
3052 * Support for POWER Decimal Floating Point instructions.
3054 * ==================== TOOL CHANGES ====================
3056 * Non-libc malloc implementations are now supported. This is useful
3057 for tools that replace malloc (Memcheck, Massif, DRD, Helgrind).
3058 Using the new option --soname-synonyms, such tools can be informed
3059 that the malloc implementation is either linked statically into the
3060 executable, or is present in some other shared library different
3061 from libc.so. This makes it possible to process statically linked
3062 programs, and programs using other malloc libraries, for example
3063 TCMalloc or JEMalloc.
3065 * For tools that provide their own replacement for malloc et al, the
3066 option --redzone-size=<number> allows users to specify the size of
3067 the padding blocks (redzones) added before and after each client
3068 allocated block. Smaller redzones decrease the memory needed by
3069 Valgrind. Bigger redzones increase the chance to detect blocks
3070 overrun or underrun. Prior to this change, the redzone size was
3071 hardwired to 16 bytes in Memcheck.
3075 - The leak_check GDB server monitor command now can
3076 control the maximum nr of loss records to output.
3078 - Reduction of memory use for applications allocating
3079 many blocks and/or having many partially defined bytes.
3081 - Addition of GDB server monitor command 'block_list' that lists
3082 the addresses/sizes of the blocks of a leak search loss record.
3084 - Addition of GDB server monitor command 'who_points_at' that lists
3085 the locations pointing at a block.
3087 - If a redzone size > 0 is given, VALGRIND_MALLOCLIKE_BLOCK now will
3088 detect an invalid access of these redzones, by marking them
3089 noaccess. Similarly, if a redzone size is given for a memory
3090 pool, VALGRIND_MEMPOOL_ALLOC will mark the redzones no access.
3091 This still allows to find some bugs if the user has forgotten to
3092 mark the pool superblock noaccess.
3094 - Performance of memory leak check has been improved, especially in
3095 cases where there are many leaked blocks and/or many suppression
3096 rules used to suppress leak reports.
3098 - Reduced noise (false positive) level on MacOSX 10.6/10.7, due to
3099 more precise analysis, which is important for LLVM/Clang
3100 generated code. This is at the cost of somewhat reduced
3101 performance. Note there is no change to analysis precision or
3102 costs on Linux targets.
3106 - Added even more facilities that can help finding the cause of a data
3107 race, namely the command-line option --ptrace-addr and the macro
3108 DRD_STOP_TRACING_VAR(x). More information can be found in the manual.
3110 - Fixed a subtle bug that could cause false positive data race reports.
3112 * ==================== OTHER CHANGES ====================
3114 * The C++ demangler has been updated so as to work well with C++
3115 compiled by up to at least g++ 4.6.
3117 * Tool developers can make replacement/wrapping more flexible thanks
3118 to the new option --soname-synonyms. This was reported above, but
3119 in fact is very general and applies to all function
3120 replacement/wrapping, not just to malloc-family functions.
3122 * Round-robin scheduling of threads can be selected, using the new
3123 option --fair-sched= yes. Prior to this change, the pipe-based
3124 thread serialisation mechanism (which is still the default) could
3125 give very unfair scheduling. --fair-sched=yes improves
3126 responsiveness of interactive multithreaded applications, and
3127 improves repeatability of results from the thread checkers Helgrind
3130 * For tool developers: support to run Valgrind on Valgrind has been
3131 improved. We can now routinely Valgrind on Helgrind or Memcheck.
3133 * gdbserver now shows the float shadow registers as integer
3134 rather than float values, as the shadow values are mostly
3135 used as bit patterns.
3137 * Increased limit for the --num-callers command line flag to 500.
3139 * Performance improvements for error matching when there are many
3140 suppression records in use.
3142 * Improved support for DWARF4 debugging information (bug 284184).
3144 * Initial support for DWZ compressed Dwarf debug info.
3146 * Improved control over the IR optimiser's handling of the tradeoff
3147 between performance and precision of exceptions. Specifically,
3148 --vex-iropt-precise-memory-exns has been removed and replaced by
3149 --vex-iropt-register-updates, with extended functionality. This
3150 allows the Valgrind gdbserver to always show up to date register
3153 * Modest performance gains through the use of translation chaining for
3156 * ==================== FIXED BUGS ====================
3158 The following bugs have been fixed or resolved. Note that "n-i-bz"
3159 stands for "not in bugzilla" -- that is, a bug that was reported to us
3160 but never got a bugzilla entry. We encourage you to file bugs in
3161 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
3162 than mailing the developers (or mailing lists) directly -- bugs that
3163 are not entered into bugzilla tend to get forgotten about or ignored.
3165 To see details of a given bug, visit
3166 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3167 where XXXXXX is the bug number as listed below.
3169 197914 Building valgrind from svn now requires automake-1.10
3170 203877 increase to 16Mb maximum allowed alignment for memalign et al
3171 219156 Handle statically linked malloc or other malloc lib (e.g. tcmalloc)
3172 247386 make perf does not run all performance tests
3173 270006 Valgrind scheduler unfair
3174 270777 Adding MIPS/Linux port to Valgrind
3175 270796 s390x: Removed broken support for the TS insn
3176 271438 Fix configure for proper SSE4.2 detection
3177 273114 s390x: Support TR, TRE, TROO, TROT, TRTO, and TRTT instructions
3178 273475 Add support for AVX instructions
3179 274078 improved configure logic for mpicc
3180 276993 fix mremap 'no thrash checks'
3181 278313 Fedora 15/x64: err read debug info with --read-var-info=yes flag
3182 281482 memcheck incorrect byte allocation count in realloc() for silly argument
3183 282230 group allocator for small fixed size, use it for MC_Chunk/SEc vbit
3184 283413 Fix wrong sanity check
3185 283671 Robustize alignment computation in LibVEX_Alloc
3186 283961 Adding support for some HCI IOCTLs
3187 284124 parse_type_DIE: confused by: DWARF 4
3188 284864 == 273475 (Add support for AVX instructions)
3189 285219 Too-restrictive constraints for Thumb2 "SP plus/minus register"
3190 285662 (MacOSX): Memcheck needs to replace memcpy/memmove
3191 285725 == 273475 (Add support for AVX instructions)
3192 286261 add wrapper for linux I2C_RDWR ioctl
3193 286270 vgpreload is not friendly to 64->32 bit execs, gives ld.so warnings
3194 286374 Running cachegrind with --branch-sim=yes on 64-bit PowerPC program fails
3195 286384 configure fails "checking for a supported version of gcc"
3196 286497 == 273475 (Add support for AVX instructions)
3197 286596 == 273475 (Add support for AVX instructions)
3198 286917 disInstr(arm): unhandled instruction: QADD (also QSUB)
3199 287175 ARM: scalar VFP fixed-point VCVT instructions not handled
3200 287260 Incorrect conditional jump or move depends on uninitialised value(s)
3201 287301 vex amd64->IR: 0x66 0xF 0x38 0x41 0xC0 0xB8 0x0 0x0 (PHMINPOSUW)
3202 287307 == 273475 (Add support for AVX instructions)
3203 287858 VG_(strerror): unknown error
3204 288298 (MacOSX) unhandled syscall shm_unlink
3205 288995 == 273475 (Add support for AVX instructions)
3206 289470 Loading of large Mach-O thin binaries fails.
3207 289656 == 273475 (Add support for AVX instructions)
3208 289699 vgdb connection in relay mode erroneously closed due to buffer overrun
3209 289823 == 293754 (PCMPxSTRx not implemented for 16-bit characters)
3210 289839 s390x: Provide support for unicode conversion instructions
3211 289939 monitor cmd 'leak_check' with details about leaked or reachable blocks
3212 290006 memcheck doesn't mark %xmm as initialized after "pcmpeqw %xmm %xmm"
3213 290655 Add support for AESKEYGENASSIST instruction
3214 290719 valgrind-3.7.0 fails with automake-1.11.2 due to"pkglibdir" usage
3215 290974 vgdb must align pages to VKI_SHMLBA (16KB) on ARM
3216 291253 ES register not initialised in valgrind simulation
3217 291568 Fix 3DNOW-related crashes with baseline x86_64 CPU (w patch)
3218 291865 s390x: Support the "Compare Double and Swap" family of instructions
3219 292300 == 273475 (Add support for AVX instructions)
3220 292430 unrecognized instruction in __intel_get_new_mem_ops_cpuid
3221 292493 == 273475 (Add support for AVX instructions)
3222 292626 Missing fcntl F_SETOWN_EX and F_GETOWN_EX support
3223 292627 Missing support for some SCSI ioctls
3224 292628 none/tests/x86/bug125959-x86.c triggers undefined behavior
3225 292841 == 273475 (Add support for AVX instructions)
3226 292993 implement the getcpu syscall on amd64-linux
3227 292995 Implement the “cross memory attach” syscalls introduced in Linux 3.2
3228 293088 Add some VEX sanity checks for ppc64 unhandled instructions
3229 293751 == 290655 (Add support for AESKEYGENASSIST instruction)
3230 293754 PCMPxSTRx not implemented for 16-bit characters
3231 293755 == 293754 (No tests for PCMPxSTRx on 16-bit characters)
3232 293808 CLFLUSH not supported by latest VEX for amd64
3233 294047 valgrind does not correctly emulate prlimit64(..., RLIMIT_NOFILE, ...)
3234 294048 MPSADBW instruction not implemented
3235 294055 regtest none/tests/shell fails when locale is not set to C
3236 294185 INT 0x44 (and others) not supported on x86 guest, but used by Jikes RVM
3237 294190 --vgdb-error=xxx can be out of sync with errors shown to the user
3238 294191 amd64: fnsave/frstor and 0x66 size prefixes on FP instructions
3239 294260 disInstr_AMD64: disInstr miscalculated next %rip
3240 294523 --partial-loads-ok=yes causes false negatives
3241 294617 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A
3242 294736 vex amd64->IR: 0x48 0xF 0xD7 0xD6 0x48 0x83
3243 294812 patch allowing to run (on x86 at least) helgrind/drd on tool.
3244 295089 can not annotate source for both helgrind and drd
3245 295221 POWER Processor decimal floating point instruction support missing
3246 295427 building for i386 with clang on darwin11 requires "-new_linker linker"
3247 295428 coregrind/m_main.c has incorrect x86 assembly for darwin
3248 295590 Helgrind: Assertion 'cvi->nWaiters > 0' failed
3249 295617 ARM - Add some missing syscalls
3250 295799 Missing \n with get_vbits in gdbserver when line is % 80 [...]
3251 296229 Linux user input device ioctls missing wrappers
3252 296318 ELF Debug info improvements (more than one rx/rw mapping)
3253 296422 Add translation chaining support
3254 296457 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A (dup of AES)
3255 296792 valgrind 3.7.0: add SIOCSHWTSTAMP (0x89B0) ioctl wrapper
3256 296983 Fix build issues on x86_64/ppc64 without 32-bit toolchains
3257 297078 gdbserver signal handling problems [..]
3258 297147 drd false positives on newly allocated memory
3259 297329 disallow decoding of IBM Power DFP insns on some machines
3260 297497 POWER Processor decimal floating point instruction support missing
3261 297701 Another alias for strncasecmp_l in libc-2.13.so
3262 297911 'invalid write' not reported when using APIs for custom mem allocators.
3263 297976 s390x: revisit EX implementation
3264 297991 Valgrind interferes with mmap()+ftell()
3265 297992 Support systems missing WIFCONTINUED (e.g. pre-2.6.10 Linux)
3266 297993 Fix compilation of valgrind with gcc -g3.
3267 298080 POWER Processor DFP support missing, part 3
3268 298227 == 273475 (Add support for AVX instructions)
3269 298335 == 273475 (Add support for AVX instructions)
3270 298354 Unhandled ARM Thumb instruction 0xEB0D 0x0585 (streq)
3271 298394 s390x: Don't bail out on an unknown machine model. [..]
3272 298421 accept4() syscall (366) support is missing for ARM
3273 298718 vex amd64->IR: 0xF 0xB1 0xCB 0x9C 0x8F 0x45
3274 298732 valgrind installation problem in ubuntu with kernel version 3.x
3275 298862 POWER Processor DFP instruction support missing, part 4
3276 298864 DWARF reader mis-parses DW_FORM_ref_addr
3277 298943 massif asserts with --pages-as-heap=yes when brk is changing [..]
3278 299053 Support DWARF4 DW_AT_high_pc constant form
3279 299104 == 273475 (Add support for AVX instructions)
3280 299316 Helgrind: hg_main.c:628 (map_threads_lookup): Assertion 'thr' failed.
3281 299629 dup3() syscall (358) support is missing for ARM
3282 299694 POWER Processor DFP instruction support missing, part 5
3283 299756 Ignore --free-fill for MEMPOOL_FREE and FREELIKE client requests
3284 299803 == 273475 (Add support for AVX instructions)
3285 299804 == 273475 (Add support for AVX instructions)
3286 299805 == 273475 (Add support for AVX instructions)
3287 300140 ARM - Missing (T1) SMMUL
3288 300195 == 296318 (ELF Debug info improvements (more than one rx/rw mapping))
3289 300389 Assertion `are_valid_hwcaps(VexArchAMD64, [..])' failed.
3290 300414 FCOM and FCOMP unimplemented for amd64 guest
3291 301204 infinite loop in canonicaliseSymtab with ifunc symbol
3292 301229 == 203877 (increase to 16Mb maximum allowed alignment for memalign etc)
3293 301265 add x86 support to Android build
3294 301984 configure script doesn't detect certain versions of clang
3295 302205 Fix compiler warnings for POWER VEX code and POWER test cases
3296 302287 Unhandled movbe instruction on Atom processors
3297 302370 PPC: fnmadd, fnmsub, fnmadds, fnmsubs insns always negate the result
3298 302536 Fix for the POWER Valgrind regression test: memcheck-ISA2.0.
3299 302578 Unrecognized isntruction 0xc5 0x32 0xc2 0xca 0x09 vcmpngess
3300 302656 == 273475 (Add support for AVX instructions)
3301 302709 valgrind for ARM needs extra tls support for android emulator [..]
3302 302827 add wrapper for CDROM_GET_CAPABILITY
3303 302901 Valgrind crashes with dwz optimized debuginfo
3304 302918 Enable testing of the vmaddfp and vnsubfp instructions in the testsuite
3305 303116 Add support for the POWER instruction popcntb
3306 303127 Power test suite fixes for frsqrte, vrefp, and vrsqrtefp instructions.
3307 303250 Assertion `instrs_in->arr_used <= 10000' failed w/ OpenSSL code
3308 303466 == 273475 (Add support for AVX instructions)
3309 303624 segmentation fault on Android 4.1 (e.g. on Galaxy Nexus OMAP)
3310 303963 strstr() function produces wrong results under valgrind callgrind
3311 304054 CALL_FN_xx macros need to enforce stack alignment
3312 304561 tee system call not supported
3313 715750 (MacOSX): Incorrect invalid-address errors near 0xFFFFxxxx (mozbug#)
3314 n-i-bz Add missing gdbserver xml files for shadow registers for ppc32
3315 n-i-bz Bypass gcc4.4/4.5 code gen bugs causing out of memory or asserts
3316 n-i-bz Fix assert in gdbserver for watchpoints watching the same address
3317 n-i-bz Fix false positive in sys_clone on amd64 when optional args [..]
3318 n-i-bz s390x: Shadow registers can now be examined using vgdb
3320 (3.8.0-TEST3: 9 August 2012, vex r2465, valgrind r12865)
3321 (3.8.0: 10 August 2012, vex r2465, valgrind r12866)
3325 Release 3.7.0 (5 November 2011)
3326 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3327 3.7.0 is a feature release with many significant improvements and the
3328 usual collection of bug fixes.
3330 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3331 PPC64/Linux, S390X/Linux, ARM/Android, X86/Darwin and AMD64/Darwin.
3332 Support for recent distros and toolchain components (glibc 2.14, gcc
3333 4.6, MacOSX 10.7) has been added.
3335 * ================== PLATFORM CHANGES =================
3337 * Support for IBM z/Architecture (s390x) running Linux. Valgrind can
3338 analyse 64-bit programs running on z/Architecture. Most user space
3339 instructions up to and including z10 are supported. Valgrind has
3340 been tested extensively on z9, z10, and z196 machines running SLES
3341 10/11, RedHat 5/6m, and Fedora. The Memcheck and Massif tools are
3342 known to work well. Callgrind, Helgrind, and DRD work reasonably
3343 well on z9 and later models. See README.s390 for more details.
3345 * Preliminary support for MacOSX 10.7 and XCode 4. Both 32- and
3346 64-bit processes are supported. Some complex threaded applications
3347 (Firefox) are observed to hang when run as 32 bit applications,
3348 whereas 64-bit versions run OK. The cause is unknown. Memcheck
3349 will likely report some false errors. In general, expect some rough
3350 spots. This release also supports MacOSX 10.6, but drops support
3353 * Preliminary support for Android (on ARM). Valgrind can now run
3354 large applications (eg, Firefox) on (eg) a Samsung Nexus S. See
3355 README.android for more details, plus instructions on how to get
3358 * Support for the IBM Power ISA 2.06 (Power7 instructions)
3360 * General correctness and performance improvements for ARM/Linux, and,
3361 by extension, ARM/Android.
3363 * Further solidification of support for SSE 4.2 in 64-bit mode. AVX
3364 instruction set support is under development but is not available in
3367 * Support for AIX5 has been removed.
3369 * ==================== TOOL CHANGES ====================
3371 * Memcheck: some incremental changes:
3373 - reduction of memory use in some circumstances
3375 - improved handling of freed memory, which in some circumstances
3376 can cause detection of use-after-free that would previously have
3379 - fix of a longstanding bug that could cause false negatives (missed
3380 errors) in programs doing vector saturated narrowing instructions.
3382 * Helgrind: performance improvements and major memory use reductions,
3383 particularly for large, long running applications which perform many
3384 synchronisation (lock, unlock, etc) events. Plus many smaller
3387 - display of locksets for both threads involved in a race
3389 - general improvements in formatting/clarity of error messages
3391 - addition of facilities and documentation regarding annotation
3392 of thread safe reference counted C++ classes
3394 - new flag --check-stack-refs=no|yes [yes], to disable race checking
3395 on thread stacks (a performance hack)
3397 - new flag --free-is-write=no|yes [no], to enable detection of races
3398 where one thread accesses heap memory but another one frees it,
3399 without any coordinating synchronisation event
3401 * DRD: enabled XML output; added support for delayed thread deletion
3402 in order to detect races that occur close to the end of a thread
3403 (--join-list-vol); fixed a memory leak triggered by repeated client
3404 memory allocatation and deallocation; improved Darwin support.
3406 * exp-ptrcheck: this tool has been renamed to exp-sgcheck
3408 * exp-sgcheck: this tool has been reduced in scope so as to improve
3409 performance and remove checking that Memcheck does better.
3410 Specifically, the ability to check for overruns for stack and global
3411 arrays is unchanged, but the ability to check for overruns of heap
3412 blocks has been removed. The tool has accordingly been renamed to
3413 exp-sgcheck ("Stack and Global Array Checking").
3415 * ==================== OTHER CHANGES ====================
3417 * GDB server: Valgrind now has an embedded GDB server. That means it
3418 is possible to control a Valgrind run from GDB, doing all the usual
3419 things that GDB can do (single stepping, breakpoints, examining
3420 data, etc). Tool-specific functionality is also available. For
3421 example, it is possible to query the definedness state of variables
3422 or memory from within GDB when running Memcheck; arbitrarily large
3423 memory watchpoints are supported, etc. To use the GDB server, start
3424 Valgrind with the flag --vgdb-error=0 and follow the on-screen
3427 * Improved support for unfriendly self-modifying code: a new option
3428 --smc-check=all-non-file is available. This adds the relevant
3429 consistency checks only to code that originates in non-file-backed
3430 mappings. In effect this confines the consistency checking only to
3431 code that is or might be JIT generated, and avoids checks on code
3432 that must have been compiled ahead of time. This significantly
3433 improves performance on applications that generate code at run time.
3435 * It is now possible to build a working Valgrind using Clang-2.9 on
3438 * new client requests VALGRIND_{DISABLE,ENABLE}_ERROR_REPORTING.
3439 These enable and disable error reporting on a per-thread, and
3440 nestable, basis. This is useful for hiding errors in particularly
3441 troublesome pieces of code. The MPI wrapper library (libmpiwrap.c)
3442 now uses this facility.
3444 * Added the --mod-funcname option to cg_diff.
3446 * ==================== FIXED BUGS ====================
3448 The following bugs have been fixed or resolved. Note that "n-i-bz"
3449 stands for "not in bugzilla" -- that is, a bug that was reported to us
3450 but never got a bugzilla entry. We encourage you to file bugs in
3451 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3452 mailing the developers (or mailing lists) directly -- bugs that are
3453 not entered into bugzilla tend to get forgotten about or ignored.
3455 To see details of a given bug, visit
3456 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3457 where XXXXXX is the bug number as listed below.
3459 79311 malloc silly arg warning does not give stack trace
3460 210935 port valgrind.h (not valgrind) to win32 to support client requests
3461 214223 valgrind SIGSEGV on startup gcc 4.4.1 ppc32 (G4) Ubuntu 9.10
3462 243404 Port to zSeries
3463 243935 Helgrind: incorrect handling of ANNOTATE_HAPPENS_BEFORE()/AFTER()
3464 247223 non-x86: Suppress warning: 'regparm' attribute directive ignored
3465 250101 huge "free" memory usage due to m_mallocfree.c fragmentation
3466 253206 Some fixes for the faultstatus testcase
3467 255223 capget testcase fails when running as root
3468 256703 xlc_dbl_u32.c testcase broken
3469 256726 Helgrind tests have broken inline asm
3470 259977 == 214223 (Valgrind segfaults doing __builtin_longjmp)
3471 264800 testcase compile failure on zseries
3472 265762 make public VEX headers compilable by G++ 3.x
3473 265771 assertion in jumps.c (r11523) fails with glibc-2.3
3474 266753 configure script does not give the user the option to not use QtCore
3475 266931 gen_insn_test.pl is broken
3476 266961 ld-linux.so.2 i?86-linux strlen issues
3477 266990 setns instruction causes false positive
3478 267020 Make directory for temporary files configurable at run-time.
3479 267342 == 267997 (segmentation fault on Mac OS 10.6)
3480 267383 Assertion 'vgPlain_strlen(dir) + vgPlain_strlen(file) + 1 < 256' failed
3481 267413 Assertion 'DRD_(g_threadinfo)[tid].synchr_nesting >= 1' failed.
3482 267488 regtest: darwin support for 64-bit build
3483 267552 SIGSEGV (misaligned_stack_error) with DRD, but not with other tools
3484 267630 Add support for IBM Power ISA 2.06 -- stage 1
3485 267769 == 267997 (Darwin: memcheck triggers segmentation fault)
3486 267819 Add client request for informing the core about reallocation
3487 267925 laog data structure quadratic for a single sequence of lock
3488 267968 drd: (vgDrd_thread_set_joinable): Assertion '0 <= (int)tid ..' failed
3489 267997 MacOSX: 64-bit V segfaults on launch when built with Xcode 4.0.1
3490 268513 missed optimizations in fold_Expr
3491 268619 s390x: fpr - gpr transfer facility
3492 268620 s390x: reconsider "long displacement" requirement
3493 268621 s390x: improve IR generation for XC
3494 268715 s390x: FLOGR is not universally available
3495 268792 == 267997 (valgrind seg faults on startup when compiled with Xcode 4)
3496 268930 s390x: MHY is not universally available
3497 269078 arm->IR: unhandled instruction SUB (SP minus immediate/register)
3498 269079 Support ptrace system call on ARM
3499 269144 missing "Bad option" error message
3500 269209 conditional load and store facility (z196)
3501 269354 Shift by zero on x86 can incorrectly clobber CC_NDEP
3502 269641 == 267997 (valgrind segfaults immediately (segmentation fault))
3503 269736 s390x: minor code generation tweaks
3504 269778 == 272986 (valgrind.h: swap roles of VALGRIND_DO_CLIENT_REQUEST() ..)
3505 269863 s390x: remove unused function parameters
3506 269864 s390x: tweak s390_emit_load_cc
3507 269884 == 250101 (overhead for huge blocks exhausts space too soon)
3508 270082 s390x: Make sure to point the PSW address to the next address on SIGILL
3509 270115 s390x: rewrite some testcases
3510 270309 == 267997 (valgrind crash on startup)
3511 270320 add support for Linux FIOQSIZE ioctl() call
3512 270326 segfault while trying to sanitize the environment passed to execle
3513 270794 IBM POWER7 support patch causes regression in none/tests
3514 270851 IBM POWER7 fcfidus instruction causes memcheck to fail
3515 270856 IBM POWER7 xsnmaddadp instruction causes memcheck to fail on 32bit app
3516 270925 hyper-optimized strspn() in /lib64/libc-2.13.so needs fix
3517 270959 s390x: invalid use of R0 as base register
3518 271042 VSX configure check fails when it should not
3519 271043 Valgrind build fails with assembler error on ppc64 with binutils 2.21
3520 271259 s390x: fix code confusion
3521 271337 == 267997 (Valgrind segfaults on MacOS X)
3522 271385 s390x: Implement Ist_MBE
3523 271501 s390x: misc cleanups
3524 271504 s390x: promote likely and unlikely
3525 271579 ppc: using wrong enum type
3526 271615 unhandled instruction "popcnt" (arch=amd10h)
3527 271730 Fix bug when checking ioctls: duplicate check
3528 271776 s390x: provide STFLE instruction support
3529 271779 s390x: provide clock instructions like STCK
3530 271799 Darwin: ioctls without an arg report a memory error
3531 271820 arm: fix type confusion
3532 271917 pthread_cond_timedwait failure leads to not-locked false positive
3533 272067 s390x: fix DISP20 macro
3534 272615 A typo in debug output in mc_leakcheck.c
3535 272661 callgrind_annotate chokes when run from paths containing regex chars
3536 272893 amd64->IR: 0x66 0xF 0x38 0x2B 0xC1 0x66 0xF 0x7F == (closed as dup)
3537 272955 Unhandled syscall error for pwrite64 on ppc64 arch
3538 272967 make documentation build-system more robust
3539 272986 Fix gcc-4.6 warnings with valgrind.h
3540 273318 amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38 (missing PCMPxSTRx case)
3541 273318 unhandled PCMPxSTRx case: vex amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38
3542 273431 valgrind segfaults in evalCfiExpr (debuginfo.c:2039)
3543 273465 Callgrind: jumps.c:164 (new_jcc): Assertion '(0 <= jmp) && ...'
3544 273536 Build error: multiple definition of `vgDrd_pthread_cond_initializer'
3545 273640 ppc64-linux: unhandled syscalls setresuid(164) and setresgid(169)
3546 273729 == 283000 (Illegal opcode for SSE2 "roundsd" instruction)
3547 273778 exp-ptrcheck: unhandled sysno == 259
3548 274089 exp-ptrcheck: unhandled sysno == 208
3549 274378 s390x: Various dispatcher tweaks
3550 274447 WARNING: unhandled syscall: 340
3551 274776 amd64->IR: 0x66 0xF 0x38 0x2B 0xC5 0x66
3552 274784 == 267997 (valgrind ls -l results in Segmentation Fault)
3553 274926 valgrind does not build against linux-3
3554 275148 configure FAIL with glibc-2.14
3555 275151 Fedora 15 / glibc-2.14 'make regtest' FAIL
3556 275168 Make Valgrind work for MacOSX 10.7 Lion
3557 275212 == 275284 (lots of false positives from __memcpy_ssse3_back et al)
3558 275278 valgrind does not build on Linux kernel 3.0.* due to silly
3559 275284 Valgrind memcpy/memmove redirection stopped working in glibc 2.14/x86_64
3560 275308 Fix implementation for ppc64 fres instruc
3561 275339 s390x: fix testcase compile warnings
3562 275517 s390x: Provide support for CKSM instruction
3563 275710 s390x: get rid of redundant address mode calculation
3564 275815 == 247894 (Valgrind doesn't know about Linux readahead(2) syscall)
3565 275852 == 250101 (valgrind uses all swap space and is killed)
3566 276784 Add support for IBM Power ISA 2.06 -- stage 3
3567 276987 gdbsrv: fix tests following recent commits
3568 277045 Valgrind crashes with unhandled DW_OP_ opcode 0x2a
3569 277199 The test_isa_2_06_part1.c in none/tests/ppc64 should be a symlink
3570 277471 Unhandled syscall: 340
3571 277610 valgrind crashes in VG_(lseek)(core_fd, phdrs[idx].p_offset, ...)
3572 277653 ARM: support Thumb2 PLD instruction
3573 277663 ARM: NEON float VMUL by scalar incorrect
3574 277689 ARM: tests for VSTn with register post-index are broken
3575 277694 ARM: BLX LR instruction broken in ARM mode
3576 277780 ARM: VMOV.F32 (immediate) instruction is broken
3577 278057 fuse filesystem syscall deadlocks
3578 278078 Unimplemented syscall 280 on ppc32
3579 278349 F_GETPIPE_SZ and F_SETPIPE_SZ Linux fcntl commands
3580 278454 VALGRIND_STACK_DEREGISTER has wrong output type
3581 278502 == 275284 (Valgrind confuses memcpy() and memmove())
3582 278892 gdbsrv: factorize gdb version handling, fix doc and typos
3583 279027 Support for MVCL and CLCL instruction
3584 279027 s390x: Provide support for CLCL and MVCL instructions
3585 279062 Remove a redundant check in the insn selector for ppc.
3586 279071 JDK creates PTEST with redundant REX.W prefix
3587 279212 gdbsrv: add monitor cmd v.info scheduler.
3588 279378 exp-ptrcheck: the 'impossible' happened on mkfifo call
3589 279698 memcheck discards valid-bits for packuswb
3590 279795 memcheck reports uninitialised values for mincore on amd64
3591 279994 Add support for IBM Power ISA 2.06 -- stage 3
3592 280083 mempolicy syscall check errors
3593 280290 vex amd64->IR: 0x66 0xF 0x38 0x28 0xC1 0x66 0xF 0x6F
3594 280710 s390x: config files for nightly builds
3595 280757 /tmp dir still used by valgrind even if TMPDIR is specified
3596 280965 Valgrind breaks fcntl locks when program does mmap
3597 281138 WARNING: unhandled syscall: 340
3598 281241 == 275168 (valgrind useless on Macos 10.7.1 Lion)
3599 281304 == 275168 (Darwin: dyld "cannot load inserted library")
3600 281305 == 275168 (unhandled syscall: unix:357 on Darwin 11.1)
3601 281468 s390x: handle do_clone and gcc clones in call traces
3602 281488 ARM: VFP register corruption
3603 281828 == 275284 (false memmove warning: "Source and destination overlap")
3604 281883 s390x: Fix system call wrapper for "clone".
3605 282105 generalise 'reclaimSuperBlock' to also reclaim splittable superblock
3606 282112 Unhandled instruction bytes: 0xDE 0xD9 0x9B 0xDF (fcompp)
3607 282238 SLES10: make check fails
3608 282979 strcasestr needs replacement with recent(>=2.12) glibc
3609 283000 vex amd64->IR: 0x66 0xF 0x3A 0xA 0xC0 0x9 0xF3 0xF
3610 283243 Regression in ppc64 memcheck tests
3611 283325 == 267997 (Darwin: V segfaults on startup when built with Xcode 4.0)
3612 283427 re-connect epoll_pwait syscall on ARM linux
3613 283600 gdbsrv: android: port vgdb.c
3614 283709 none/tests/faultstatus needs to account for page size
3615 284305 filter_gdb needs enhancement to work on ppc64
3616 284384 clang 3.1 -Wunused-value warnings in valgrind.h, memcheck.h
3617 284472 Thumb2 ROR.W encoding T2 not implemented
3618 284621 XML-escape process command line in XML output
3619 n-i-bz cachegrind/callgrind: handle CPUID information for Core iX Intel CPUs
3620 that have non-power-of-2 sizes (also AMDs)
3621 n-i-bz don't be spooked by libraries mashed by elfhack
3622 n-i-bz don't be spooked by libxul.so linked with gold
3623 n-i-bz improved checking for VALGRIND_CHECK_MEM_IS_DEFINED
3625 (3.7.0-TEST1: 27 October 2011, vex r2228, valgrind r12245)
3626 (3.7.0.RC1: 1 November 2011, vex r2231, valgrind r12257)
3627 (3.7.0: 5 November 2011, vex r2231, valgrind r12258)
3631 Release 3.6.1 (16 February 2011)
3632 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3633 3.6.1 is a bug fix release. It adds support for some SSE4
3634 instructions that were omitted in 3.6.0 due to lack of time. Initial
3635 support for glibc-2.13 has been added. A number of bugs causing
3636 crashing or assertion failures have been fixed.
3638 The following bugs have been fixed or resolved. Note that "n-i-bz"
3639 stands for "not in bugzilla" -- that is, a bug that was reported to us
3640 but never got a bugzilla entry. We encourage you to file bugs in
3641 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3642 mailing the developers (or mailing lists) directly -- bugs that are
3643 not entered into bugzilla tend to get forgotten about or ignored.
3645 To see details of a given bug, visit
3646 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3647 where XXXXXX is the bug number as listed below.
3649 188572 Valgrind on Mac should suppress setenv() mem leak
3650 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 (proper FX{SAVE,RSTOR} support)
3651 210481 vex amd64->IR: Assertion `sz == 2 || sz == 4' failed (REX.W POPQ)
3652 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3653 250038 ppc64: Altivec LVSR and LVSL instructions fail their regtest
3654 254420 memory pool tracking broken
3655 254957 Test code failing to compile due to changes in memcheck.h
3656 255009 helgrind/drd: crash on chmod with invalid parameter
3657 255130 readdwarf3.c parse_type_DIE confused by GNAT Ada types
3658 255355 helgrind/drd: crash on threaded programs doing fork
3660 255418 (SSE4.x) rint call compiled with ICC
3661 255822 --gen-suppressions can create invalid files: "too many callers [...]"
3662 255888 closing valgrindoutput tag outputted to log-stream on error
3663 255963 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x9 0xDB 0x0 (ROUNDPD)
3664 255966 Slowness when using mempool annotations
3665 256387 vex x86->IR: 0xD4 0xA 0x2 0x7 (AAD and AAM)
3666 256600 super-optimized strcasecmp() false positive
3667 256669 vex amd64->IR: Unhandled LOOPNEL insn on amd64
3668 256968 (SSE4.x) vex amd64->IR: 0x66 0xF 0x38 0x10 0xD3 0x66 (BLENDVPx)
3669 257011 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0xE 0xFD 0xA0 (PBLENDW)
3670 257063 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x8 0xC0 0x0 (ROUNDPS)
3671 257276 Missing case in memcheck --track-origins=yes
3672 258870 (SSE4.x) Add support for EXTRACTPS SSE 4.1 instruction
3673 261966 (SSE4.x) support for CRC32B and CRC32Q is lacking (also CRC32{W,L})
3674 262985 VEX regression in valgrind 3.6.0 in handling PowerPC VMX
3675 262995 (SSE4.x) crash when trying to valgrind gcc-snapshot (PCMPxSTRx $0)
3676 263099 callgrind_annotate counts Ir improperly [...]
3677 263877 undefined coprocessor instruction on ARMv7
3678 265964 configure FAIL with glibc-2.13
3679 n-i-bz Fix compile error w/ icc-12.x in guest_arm_toIR.c
3680 n-i-bz Docs: fix bogus descriptions for VALGRIND_CREATE_BLOCK et al
3681 n-i-bz Massif: don't assert on shmat() with --pages-as-heap=yes
3682 n-i-bz Bug fixes and major speedups for the exp-DHAT space profiler
3683 n-i-bz DRD: disable --free-is-write due to implementation difficulties
3685 (3.6.1: 16 February 2011, vex r2103, valgrind r11561).
3689 Release 3.6.0 (21 October 2010)
3690 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3691 3.6.0 is a feature release with many significant improvements and the
3692 usual collection of bug fixes.
3694 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3695 PPC64/Linux, X86/Darwin and AMD64/Darwin. Support for recent distros
3696 and toolchain components (glibc 2.12, gcc 4.5, OSX 10.6) has been added.
3698 -------------------------
3700 Here are some highlights. Details are shown further down:
3702 * Support for ARM/Linux.
3704 * Support for recent Linux distros: Ubuntu 10.10 and Fedora 14.
3706 * Support for Mac OS X 10.6, both 32- and 64-bit executables.
3708 * Support for the SSE4.2 instruction set.
3710 * Enhancements to the Callgrind profiler, including the ability to
3711 handle CPUs with three levels of cache.
3713 * A new experimental heap profiler, DHAT.
3715 * A huge number of bug fixes and small enhancements.
3717 -------------------------
3719 Here are details of the above changes, together with descriptions of
3720 many other changes, and a list of fixed bugs.
3722 * ================== PLATFORM CHANGES =================
3724 * Support for ARM/Linux. Valgrind now runs on ARMv7 capable CPUs
3725 running Linux. It is known to work on Ubuntu 10.04, Ubuntu 10.10,
3726 and Maemo 5, so you can run Valgrind on your Nokia N900 if you want.
3728 This requires a CPU capable of running the ARMv7-A instruction set
3729 (Cortex A5, A8 and A9). Valgrind provides fairly complete coverage
3730 of the user space instruction set, including ARM and Thumb integer
3731 code, VFPv3, NEON and V6 media instructions. The Memcheck,
3732 Cachegrind and Massif tools work properly; other tools work to
3735 * Support for recent Linux distros (Ubuntu 10.10 and Fedora 14), along
3736 with support for recent releases of the underlying toolchain
3737 components, notably gcc-4.5 and glibc-2.12.
3739 * Support for Mac OS X 10.6, both 32- and 64-bit executables. 64-bit
3740 support also works much better on OS X 10.5, and is as solid as
3743 * Support for the SSE4.2 instruction set. SSE4.2 is supported in
3744 64-bit mode. In 32-bit mode, support is only available up to and
3745 including SSSE3. Some exceptions: SSE4.2 AES instructions are not
3746 supported in 64-bit mode, and 32-bit mode does in fact support the
3747 bare minimum SSE4 instructions to needed to run programs on Mac OS X
3748 10.6 on 32-bit targets.
3750 * Support for IBM POWER6 cpus has been improved. The Power ISA up to
3751 and including version 2.05 is supported.
3753 * ==================== TOOL CHANGES ====================
3755 * Cachegrind has a new processing script, cg_diff, which finds the
3756 difference between two profiles. It's very useful for evaluating
3757 the performance effects of a change in a program.
3759 Related to this change, the meaning of cg_annotate's (rarely-used)
3760 --threshold option has changed; this is unlikely to affect many
3761 people, if you do use it please see the user manual for details.
3763 * Callgrind now can do branch prediction simulation, similar to
3764 Cachegrind. In addition, it optionally can count the number of
3765 executed global bus events. Both can be used for a better
3766 approximation of a "Cycle Estimation" as derived event (you need to
3767 update the event formula in KCachegrind yourself).
3769 * Cachegrind and Callgrind now refer to the LL (last-level) cache
3770 rather than the L2 cache. This is to accommodate machines with
3771 three levels of caches -- if Cachegrind/Callgrind auto-detects the
3772 cache configuration of such a machine it will run the simulation as
3773 if the L2 cache isn't present. This means the results are less
3774 likely to match the true result for the machine, but
3775 Cachegrind/Callgrind's results are already only approximate, and
3776 should not be considered authoritative. The results are still
3777 useful for giving a general idea about a program's locality.
3779 * Massif has a new option, --pages-as-heap, which is disabled by
3780 default. When enabled, instead of tracking allocations at the level
3781 of heap blocks (as allocated with malloc/new/new[]), it instead
3782 tracks memory allocations at the level of memory pages (as mapped by
3783 mmap, brk, etc). Each mapped page is treated as its own block.
3784 Interpreting the page-level output is harder than the heap-level
3785 output, but this option is useful if you want to account for every
3786 byte of memory used by a program.
3788 * DRD has two new command-line options: --free-is-write and
3789 --trace-alloc. The former allows to detect reading from already freed
3790 memory, and the latter allows tracing of all memory allocations and
3793 * DRD has several new annotations. Custom barrier implementations can
3794 now be annotated, as well as benign races on static variables.
3796 * DRD's happens before / happens after annotations have been made more
3797 powerful, so that they can now also be used to annotate e.g. a smart
3798 pointer implementation.
3800 * Helgrind's annotation set has also been drastically improved, so as
3801 to provide to users a general set of annotations to describe locks,
3802 semaphores, barriers and condition variables. Annotations to
3803 describe thread-safe reference counted heap objects have also been
3806 * Memcheck has a new command-line option, --show-possibly-lost, which
3807 is enabled by default. When disabled, the leak detector will not
3808 show possibly-lost blocks.
3810 * A new experimental heap profiler, DHAT (Dynamic Heap Analysis Tool),
3811 has been added. DHAT keeps track of allocated heap blocks, and also
3812 inspects every memory reference to see which block (if any) is being
3813 accessed. This gives a lot of insight into block lifetimes,
3814 utilisation, turnover, liveness, and the location of hot and cold
3815 fields. You can use DHAT to do hot-field profiling.
3817 * ==================== OTHER CHANGES ====================
3819 * Improved support for unfriendly self-modifying code: the extra
3820 overhead incurred by --smc-check=all has been reduced by
3821 approximately a factor of 5 as compared with 3.5.0.
3823 * Ability to show directory names for source files in error messages.
3824 This is combined with a flexible mechanism for specifying which
3825 parts of the paths should be shown. This is enabled by the new flag
3828 * A new flag, --require-text-symbol, which will stop the run if a
3829 specified symbol is not found it a given shared object when it is
3830 loaded into the process. This makes advanced working with function
3831 intercepting and wrapping safer and more reliable.
3833 * Improved support for the Valkyrie GUI, version 2.0.0. GUI output
3834 and control of Valgrind is now available for the tools Memcheck and
3835 Helgrind. XML output from Valgrind is available for Memcheck,
3836 Helgrind and exp-Ptrcheck.
3838 * More reliable stack unwinding on amd64-linux, particularly in the
3839 presence of function wrappers, and with gcc-4.5 compiled code.
3841 * Modest scalability (performance improvements) for massive
3842 long-running applications, particularly for those with huge amounts
3845 * Support for analyzing programs running under Wine with has been
3846 improved. The header files <valgrind/valgrind.h>,
3847 <valgrind/memcheck.h> and <valgrind/drd.h> can now be used in
3848 Windows-programs compiled with MinGW or one of the Microsoft Visual
3851 * A rare but serious error in the 64-bit x86 CPU simulation was fixed.
3852 The 32-bit simulator was not affected. This did not occur often,
3853 but when it did would usually crash the program under test.
3856 * A large number of bugs were fixed. These are shown below.
3858 * A number of bugs were investigated, and were candidates for fixing,
3859 but are not fixed in 3.6.0, due to lack of developer time. They may
3860 get fixed in later releases. They are:
3862 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 0x24 0x49 (FXSAVE64)
3863 212419 false positive "lock order violated" (A+B vs A)
3864 213685 Undefined value propagates past dependency breaking instruction
3865 216837 Incorrect instrumentation of NSOperationQueue on Darwin
3866 237920 valgrind segfault on fork failure
3867 242137 support for code compiled by LLVM-2.8
3868 242423 Another unknown Intel cache config value
3869 243232 Inconsistent Lock Orderings report with trylock
3870 243483 ppc: callgrind triggers VEX assertion failure
3871 243935 Helgrind: implementation of ANNOTATE_HAPPENS_BEFORE() is wrong
3872 244677 Helgrind crash hg_main.c:616 (map_threads_lookup): Assertion
3874 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3875 249435 Analyzing wine programs with callgrind triggers a crash
3876 250038 ppc64: Altivec lvsr and lvsl instructions fail their regtest
3877 250065 Handling large allocations
3878 250101 huge "free" memory usage due to m_mallocfree.c
3879 "superblocks fragmentation"
3880 251569 vex amd64->IR: 0xF 0x1 0xF9 0x8B 0x4C 0x24 (RDTSCP)
3881 252091 Callgrind on ARM does not detect function returns correctly
3882 252600 [PATCH] Allow lhs to be a pointer for shl/shr
3883 254420 memory pool tracking broken
3884 n-i-bz support for adding symbols for JIT generated code
3887 The following bugs have been fixed or resolved. Note that "n-i-bz"
3888 stands for "not in bugzilla" -- that is, a bug that was reported to us
3889 but never got a bugzilla entry. We encourage you to file bugs in
3890 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3891 mailing the developers (or mailing lists) directly -- bugs that are
3892 not entered into bugzilla tend to get forgotten about or ignored.
3894 To see details of a given bug, visit
3895 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3896 where XXXXXX is the bug number as listed below.
3898 135264 dcbzl instruction missing
3900 153699 Valgrind should report unaligned reads with movdqa
3902 190429 Valgrind reports lost of errors in ld.so
3903 with x86_64 2.9.90 glibc
3904 197266 valgrind appears to choke on the xmms instruction
3906 197988 Crash when demangling very large symbol names
3907 202315 unhandled syscall: 332 (inotify_init1)
3908 203256 Add page-level profiling to Massif
3909 205093 dsymutil=yes needs quotes, locking (partial fix)
3910 205241 Snow Leopard 10.6 support (partial fix)
3911 206600 Leak checker fails to upgrade indirect blocks when their
3912 parent becomes reachable
3913 210935 port valgrind.h (not valgrind) to win32 so apps run under
3914 wine can make client requests
3915 211410 vex amd64->IR: 0x15 0xFF 0xFF 0x0 0x0 0x89
3916 within Linux ip-stack checksum functions
3917 212335 unhandled instruction bytes: 0xF3 0xF 0xBD 0xC0
3919 213685 Undefined value propagates past dependency breaking instruction
3921 215914 Valgrind inserts bogus empty environment variable
3923 219538 adjtimex syscall wrapper wrong in readonly adjtime mode
3924 222545 shmat fails under valgind on some arm targets
3925 222560 ARM NEON support
3928 232509 Docs build fails with formatting inside <title></title> elements
3930 235642 [PATCH] syswrap-linux.c: support evdev EVIOCG* ioctls
3931 236546 vex x86->IR: 0x66 0xF 0x3A 0xA
3932 237202 vex amd64->IR: 0xF3 0xF 0xB8 0xC0 0x49 0x3B
3933 237371 better support for VALGRIND_MALLOCLIKE_BLOCK
3934 237485 symlink (syscall 57) is not supported on Mac OS
3935 237723 sysno == 101 exp-ptrcheck: the 'impossible' happened:
3937 238208 is_just_below_ESP doesn't take into account red-zone
3938 238345 valgrind passes wrong $0 when executing a shell script
3939 238679 mq_timedreceive syscall doesn't flag the reception buffer
3941 238696 fcntl command F_DUPFD_CLOEXEC not supported
3942 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3943 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3944 238745 3.5.0 Make fails on PPC Altivec opcodes, though configure
3946 239992 vex amd64->IR: 0x48 0xF 0xC4 0xC1 0x0 0x48
3952 242606 unhandled syscall: setegid (in Ptrcheck)
3953 242814 Helgrind "Impossible has happened" during
3954 QApplication::initInstance();
3955 243064 Valgrind attempting to read debug information from iso
3956 243270 Make stack unwinding in Valgrind wrappers more reliable
3957 243884 exp-ptrcheck: the 'impossible happened: unhandled syscall
3958 sysno = 277 (mq_open)
3959 244009 exp-ptrcheck unknown syscalls in analyzing lighttpd
3960 244493 ARM VFP d16-d31 registers support
3961 244670 add support for audit_session_self syscall on Mac OS 10.6
3962 244921 The xml report of helgrind tool is not well format
3963 244923 In the xml report file, the <preamble> not escape the
3964 xml char, eg '<','&','>'
3965 245535 print full path names in plain text reports
3966 245925 x86-64 red zone handling problem
3967 246258 Valgrind not catching integer underruns + new [] s
3968 246311 reg/reg cmpxchg doesn't work on amd64
3969 246549 unhandled syscall unix:277 while testing 32-bit Darwin app
3970 246888 Improve Makefile.vex.am
3971 247510 [OS X 10.6] Memcheck reports unaddressable bytes passed
3972 to [f]chmod_extended
3973 247526 IBM POWER6 (ISA 2.05) support is incomplete
3974 247561 Some leak testcases fails due to reachable addresses in
3976 247875 sizeofIRType to handle Ity_I128
3977 247894 [PATCH] unhandled syscall sys_readahead
3978 247980 Doesn't honor CFLAGS passed to configure
3979 248373 darwin10.supp is empty in the trunk
3980 248822 Linux FIBMAP ioctl has int parameter instead of long
3981 248893 [PATCH] make readdwarf.c big endianess safe to enable
3982 unwinding on big endian systems
3983 249224 Syscall 336 not supported (SYS_proc_info)
3985 249775 Incorrect scheme for detecting NEON capabilities of host CPU
3986 249943 jni JVM init fails when using valgrind
3987 249991 Valgrind incorrectly declares AESKEYGENASSIST support
3989 249996 linux/arm: unhandled syscall: 181 (__NR_pwrite64)
3990 250799 frexp$fenv_access_off function generates SIGILL
3991 250998 vex x86->IR: unhandled instruction bytes: 0x66 0x66 0x66 0x2E
3992 251251 support pclmulqdq insn
3993 251362 valgrind: ARM: attach to debugger either fails or provokes
3995 251674 Unhandled syscall 294
3998 254257 Add support for debugfiles found by build-id
3999 254550 [PATCH] Implement DW_ATE_UTF (DWARF4)
4000 254646 Wrapped functions cause stack misalignment on OS X
4001 (and possibly Linux)
4002 254556 ARM: valgrinding anything fails with SIGSEGV for 0xFFFF0FA0
4004 (3.6.0: 21 October 2010, vex r2068, valgrind r11471).
4008 Release 3.5.0 (19 August 2009)
4009 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4010 3.5.0 is a feature release with many significant improvements and the
4011 usual collection of bug fixes. The main improvement is that Valgrind
4012 now works on Mac OS X.
4014 This release supports X86/Linux, AMD64/Linux, PPC32/Linux, PPC64/Linux
4015 and X86/Darwin. Support for recent distros and toolchain components
4016 (glibc 2.10, gcc 4.5) has been added.
4018 -------------------------
4020 Here is a short summary of the changes. Details are shown further
4023 * Support for Mac OS X (10.5.x).
4025 * Improvements and simplifications to Memcheck's leak checker.
4027 * Clarification and simplifications in various aspects of Valgrind's
4030 * XML output for Helgrind and Ptrcheck.
4032 * Performance and stability improvements for Helgrind and DRD.
4034 * Genuinely atomic support for x86/amd64/ppc atomic instructions.
4036 * A new experimental tool, BBV, useful for computer architecture
4039 * Improved Wine support, including ability to read Windows PDB
4042 -------------------------
4044 Here are details of the above changes, followed by descriptions of
4045 many other minor changes, and a list of fixed bugs.
4048 * Valgrind now runs on Mac OS X. (Note that Mac OS X is sometimes
4049 called "Darwin" because that is the name of the OS core, which is the
4050 level that Valgrind works at.)
4054 - It requires OS 10.5.x (Leopard). Porting to 10.4.x is not planned
4055 because it would require work and 10.4 is only becoming less common.
4057 - 32-bit programs on x86 and AMD64 (a.k.a x86-64) machines are supported
4058 fairly well. For 10.5.x, 32-bit programs are the default even on
4059 64-bit machines, so it handles most current programs.
4061 - 64-bit programs on x86 and AMD64 (a.k.a x86-64) machines are not
4062 officially supported, but simple programs at least will probably work.
4063 However, start-up is slow.
4065 - PowerPC machines are not supported.
4067 Things that don't work:
4069 - The Ptrcheck tool.
4071 - Objective-C garbage collection.
4075 - If you have Rogue Amoeba's "Instant Hijack" program installed,
4076 Valgrind will fail with a SIGTRAP at start-up. See
4077 https://bugs.kde.org/show_bug.cgi?id=193917 for details and a
4082 - You will likely find --dsymutil=yes a useful option, as error
4083 messages may be imprecise without it.
4085 - Mac OS X support is new and therefore will be less robust than the
4086 Linux support. Please report any bugs you find.
4088 - Threaded programs may run more slowly than on Linux.
4090 Many thanks to Greg Parker for developing this port over several years.
4093 * Memcheck's leak checker has been improved.
4095 - The results for --leak-check=summary now match the summary results
4096 for --leak-check=full. Previously they could differ because
4097 --leak-check=summary counted "indirectly lost" blocks and
4098 "suppressed" blocks as "definitely lost".
4100 - Blocks that are only reachable via at least one interior-pointer,
4101 but are directly pointed to by a start-pointer, were previously
4102 marked as "still reachable". They are now correctly marked as
4105 - The default value for the --leak-resolution option has been
4106 changed from "low" to "high". In general, this means that more
4107 leak reports will be produced, but each leak report will describe
4108 fewer leaked blocks.
4110 - With --leak-check=full, "definitely lost" and "possibly lost"
4111 leaks are now considered as proper errors, ie. they are counted
4112 for the "ERROR SUMMARY" and affect the behaviour of
4113 --error-exitcode. These leaks are not counted as errors if
4114 --leak-check=summary is specified, however.
4116 - Documentation for the leak checker has been improved.
4119 * Various aspects of Valgrind's text output have changed.
4121 - Valgrind's start-up message has changed. It is shorter but also
4122 includes the command being run, which makes it easier to use
4123 --trace-children=yes. An example:
4125 - Valgrind's shut-down messages have also changed. This is most
4126 noticeable with Memcheck, where the leak summary now occurs before
4127 the error summary. This change was necessary to allow leaks to be
4128 counted as proper errors (see the description of the leak checker
4129 changes above for more details). This was also necessary to fix a
4130 longstanding bug in which uses of suppressions against leaks were
4131 not "counted", leading to difficulties in maintaining suppression
4132 files (see https://bugs.kde.org/show_bug.cgi?id=186790).
4134 - Behavior of -v has changed. In previous versions, -v printed out
4135 a mixture of marginally-user-useful information, and tool/core
4136 statistics. The statistics printing has now been moved to its own
4137 flag, --stats=yes. This means -v is less verbose and more likely
4138 to convey useful end-user information.
4140 - The format of some (non-XML) stack trace entries has changed a
4141 little. Previously there were six possible forms:
4143 0x80483BF: really (a.c:20)
4144 0x80483BF: really (in /foo/a.out)
4146 0x80483BF: (within /foo/a.out)
4147 0x80483BF: ??? (a.c:20)
4150 The third and fourth of these forms have been made more consistent
4151 with the others. The six possible forms are now:
4153 0x80483BF: really (a.c:20)
4154 0x80483BF: really (in /foo/a.out)
4155 0x80483BF: really (in ???)
4156 0x80483BF: ??? (in /foo/a.out)
4157 0x80483BF: ??? (a.c:20)
4160 Stack traces produced when --xml=yes is specified are different
4164 * Helgrind and Ptrcheck now support XML output, so they can be used
4165 from GUI tools. Also, the XML output mechanism has been
4168 - The XML format has been overhauled and generalised, so it is more
4169 suitable for error reporting tools in general. The Memcheck
4170 specific aspects of it have been removed. The new format, which
4171 is an evolution of the old format, is described in
4172 docs/internals/xml-output-protocol4.txt.
4174 - Memcheck has been updated to use the new format.
4176 - Helgrind and Ptrcheck are now able to emit output in this format.
4178 - The XML output mechanism has been overhauled. XML is now output
4179 to its own file descriptor, which means that:
4181 * Valgrind can output text and XML independently.
4183 * The longstanding problem of XML output being corrupted by
4184 unexpected un-tagged text messages is solved.
4186 As before, the destination for text output is specified using
4187 --log-file=, --log-fd= or --log-socket=.
4189 As before, XML output for a tool is enabled using --xml=yes.
4191 Because there's a new XML output channel, the XML output
4192 destination is now specified by --xml-file=, --xml-fd= or
4195 Initial feedback has shown this causes some confusion. To
4196 clarify, the two envisaged usage scenarios are:
4198 (1) Normal text output. In this case, do not specify --xml=yes
4199 nor any of --xml-file=, --xml-fd= or --xml-socket=.
4201 (2) XML output. In this case, specify --xml=yes, and one of
4202 --xml-file=, --xml-fd= or --xml-socket= to select the XML
4203 destination, one of --log-file=, --log-fd= or --log-socket=
4204 to select the destination for any remaining text messages,
4205 and, importantly, -q.
4207 -q makes Valgrind completely silent on the text channel,
4208 except in the case of critical failures, such as Valgrind
4209 itself segfaulting, or failing to read debugging information.
4210 Hence, in this scenario, it suffices to check whether or not
4211 any output appeared on the text channel. If yes, then it is
4212 likely to be a critical error which should be brought to the
4213 attention of the user. If no (the text channel produced no
4214 output) then it can be assumed that the run was successful.
4216 This allows GUIs to make the critical distinction they need to
4217 make (did the run fail or not?) without having to search or
4218 filter the text output channel in any way.
4220 It is also recommended to use --child-silent-after-fork=yes in
4224 * Improvements and changes in Helgrind:
4226 - XML output, as described above
4228 - Checks for consistent association between pthread condition
4229 variables and their associated mutexes are now performed.
4231 - pthread_spinlock functions are supported.
4233 - Modest performance improvements.
4235 - Initial (skeletal) support for describing the behaviour of
4236 non-POSIX synchronisation objects through ThreadSanitizer
4237 compatible ANNOTATE_* macros.
4239 - More controllable tradeoffs between performance and the level of
4240 detail of "previous" accesses in a race. There are now three
4243 * --history-level=full. This is the default, and was also the
4244 default in 3.4.x. It shows both stacks involved in a race, but
4245 requires a lot of memory and can be very slow in programs that
4246 do many inter-thread synchronisation events.
4248 * --history-level=none. This only shows the later stack involved
4249 in a race. This can be much faster than --history-level=full,
4250 but makes it much more difficult to find the other access
4251 involved in the race.
4253 The new intermediate setting is
4255 * --history-level=approx
4257 For the earlier (other) access, two stacks are presented. The
4258 earlier access is guaranteed to be somewhere in between the two
4259 program points denoted by those stacks. This is not as useful
4260 as showing the exact stack for the previous access (as per
4261 --history-level=full), but it is better than nothing, and it's
4262 almost as fast as --history-level=none.
4265 * New features and improvements in DRD:
4267 - The error messages printed by DRD are now easier to interpret.
4268 Instead of using two different numbers to identify each thread
4269 (Valgrind thread ID and DRD thread ID), DRD does now identify
4270 threads via a single number (the DRD thread ID). Furthermore
4271 "first observed at" information is now printed for all error
4272 messages related to synchronization objects.
4274 - Added support for named semaphores (sem_open() and sem_close()).
4276 - Race conditions between pthread_barrier_wait() and
4277 pthread_barrier_destroy() calls are now reported.
4279 - Added support for custom allocators through the macros
4280 VALGRIND_MALLOCLIKE_BLOCK() VALGRIND_FREELIKE_BLOCK() (defined in
4281 in <valgrind/valgrind.h>). An alternative for these two macros is
4282 the new client request VG_USERREQ__DRD_CLEAN_MEMORY (defined in
4285 - Added support for annotating non-POSIX synchronization objects
4286 through several new ANNOTATE_*() macros.
4288 - OpenMP: added support for the OpenMP runtime (libgomp) included
4289 with gcc versions 4.3.0 and 4.4.0.
4293 - Added two new command-line options (--first-race-only and
4294 --segment-merging-interval).
4297 * Genuinely atomic support for x86/amd64/ppc atomic instructions
4299 Valgrind will now preserve (memory-access) atomicity of LOCK-
4300 prefixed x86/amd64 instructions, and any others implying a global
4301 bus lock. Ditto for PowerPC l{w,d}arx/st{w,d}cx. instructions.
4303 This means that Valgrinded processes will "play nicely" in
4304 situations where communication with other processes, or the kernel,
4305 is done through shared memory and coordinated with such atomic
4306 instructions. Prior to this change, such arrangements usually
4307 resulted in hangs, races or other synchronisation failures, because
4308 Valgrind did not honour atomicity of such instructions.
4311 * A new experimental tool, BBV, has been added. BBV generates basic
4312 block vectors for use with the SimPoint analysis tool, which allows
4313 a program's overall behaviour to be approximated by running only a
4314 fraction of it. This is useful for computer architecture
4315 researchers. You can run BBV by specifying --tool=exp-bbv (the
4316 "exp-" prefix is short for "experimental"). BBV was written by
4320 * Modestly improved support for running Windows applications under
4321 Wine. In particular, initial support for reading Windows .PDB debug
4322 information has been added.
4325 * A new Memcheck client request VALGRIND_COUNT_LEAK_BLOCKS has been
4326 added. It is similar to VALGRIND_COUNT_LEAKS but counts blocks
4330 * The Valgrind client requests VALGRIND_PRINTF and
4331 VALGRIND_PRINTF_BACKTRACE have been changed slightly. Previously,
4332 the string was always printed immediately on its own line. Now, the
4333 string will be added to a buffer but not printed until a newline is
4334 encountered, or other Valgrind output is printed (note that for
4335 VALGRIND_PRINTF_BACKTRACE, the back-trace itself is considered
4336 "other Valgrind output"). This allows you to use multiple
4337 VALGRIND_PRINTF calls to build up a single output line, and also to
4338 print multiple output lines with a single request (by embedding
4339 multiple newlines in the string).
4342 * The graphs drawn by Massif's ms_print program have changed slightly:
4344 - The half-height chars '.' and ',' are no longer drawn, because
4345 they are confusing. The --y option can be used if the default
4346 y-resolution is not high enough.
4348 - Horizontal lines are now drawn after the top of a snapshot if
4349 there is a gap until the next snapshot. This makes it clear that
4350 the memory usage has not dropped to zero between snapshots.
4353 * Something that happened in 3.4.0, but wasn't clearly announced: the
4354 option --read-var-info=yes can be used by some tools (Memcheck,
4355 Helgrind and DRD). When enabled, it causes Valgrind to read DWARF3
4356 variable type and location information. This makes those tools
4357 start up more slowly and increases memory consumption, but
4358 descriptions of data addresses in error messages become more
4362 * exp-Omega, an experimental instantaneous leak-detecting tool, was
4363 disabled in 3.4.0 due to a lack of interest and maintenance,
4364 although the source code was still in the distribution. The source
4365 code has now been removed from the distribution. For anyone
4366 interested, the removal occurred in SVN revision r10247.
4369 * Some changes have been made to the build system.
4371 - VEX/ is now integrated properly into the build system. This means
4372 that dependency tracking within VEX/ now works properly, "make
4373 install" will work without requiring "make" before it, and
4374 parallel builds (ie. 'make -j') now work (previously a
4375 .NOTPARALLEL directive was used to serialize builds, ie. 'make -j'
4376 was effectively ignored).
4378 - The --with-vex configure option has been removed. It was of
4379 little use and removing it simplified the build system.
4381 - The location of some install files has changed. This should not
4382 affect most users. Those who might be affected:
4384 * For people who use Valgrind with MPI programs, the installed
4385 libmpiwrap.so library has moved from
4386 $(INSTALL)/<platform>/libmpiwrap.so to
4387 $(INSTALL)/libmpiwrap-<platform>.so.
4389 * For people who distribute standalone Valgrind tools, the
4390 installed libraries such as $(INSTALL)/<platform>/libcoregrind.a
4391 have moved to $(INSTALL)/libcoregrind-<platform>.a.
4393 These changes simplify the build system.
4395 - Previously, all the distributed suppression (*.supp) files were
4396 installed. Now, only default.supp is installed. This should not
4397 affect users as the other installed suppression files were not
4398 read; the fact that they were installed was a mistake.
4401 * KNOWN LIMITATIONS:
4403 - Memcheck is unusable with the Intel compiler suite version 11.1,
4404 when it generates code for SSE2-and-above capable targets. This
4405 is because of icc's use of highly optimised inlined strlen
4406 implementations. It causes Memcheck to report huge numbers of
4407 false errors even in simple programs. Helgrind and DRD may also
4410 Versions 11.0 and earlier may be OK, but this has not been
4414 The following bugs have been fixed or resolved. Note that "n-i-bz"
4415 stands for "not in bugzilla" -- that is, a bug that was reported to us
4416 but never got a bugzilla entry. We encourage you to file bugs in
4417 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4418 mailing the developers (or mailing lists) directly -- bugs that are
4419 not entered into bugzilla tend to get forgotten about or ignored.
4421 To see details of a given bug, visit
4422 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
4423 where XXXXXX is the bug number as listed below.
4425 84303 How about a LockCheck tool?
4426 91633 dereference of null ptr in vgPlain_st_basetype
4427 97452 Valgrind doesn't report any pthreads problems
4428 100628 leak-check gets assertion failure when using
4429 VALGRIND_MALLOCLIKE_BLOCK on malloc()ed memory
4430 108528 NPTL pthread cleanup handlers not called
4431 110126 Valgrind 2.4.1 configure.in tramples CFLAGS
4432 110128 mallinfo is not implemented...
4433 110770 VEX: Generated files not always updated when making valgrind
4434 111102 Memcheck: problems with large (memory footprint) applications
4435 115673 Vex's decoder should never assert
4436 117564 False positive: Syscall param clone(child_tidptr) contains
4437 uninitialised byte(s)
4438 119404 executing ssh from inside valgrind fails
4439 133679 Callgrind does not write path names to sources with dwarf debug
4441 135847 configure.in problem with non gnu compilers (and possible fix)
4442 136154 threads.c:273 (vgCallgrind_post_signal): Assertion
4443 '*(vgCallgrind_current_fn_stack.top) == 0' failed.
4444 136230 memcheck reports "possibly lost", should be "still reachable"
4445 137073 NULL arg to MALLOCLIKE_BLOCK causes crash
4446 137904 Valgrind reports a memory leak when using POSIX threads,
4448 139076 valgrind VT_GETSTATE error
4449 142228 complaint of elf_dynamic_do_rela in trivial usage
4450 145347 spurious warning with USBDEVFS_REAPURB
4451 148441 (wine) can't find memory leak in Wine, win32 binary
4453 148742 Leak-check fails assert on exit
4454 149878 add (proper) check for calloc integer overflow
4455 150606 Call graph is broken when using callgrind control
4456 152393 leak errors produce an exit code of 0. I need some way to
4457 cause leak errors to result in a nonzero exit code.
4458 157154 documentation (leak-resolution doc speaks about num-callers
4459 def=4) + what is a loss record
4460 159501 incorrect handling of ALSA ioctls
4461 162020 Valgrinding an empty/zero-byte file crashes valgrind
4462 162482 ppc: Valgrind crashes while reading stabs information
4463 162718 x86: avoid segment selector 0 in sys_set_thread_area()
4464 163253 (wine) canonicaliseSymtab forgot some fields in DiSym
4465 163560 VEX/test_main.c is missing from valgrind-3.3.1
4466 164353 malloc_usable_size() doesn't return a usable size
4467 165468 Inconsistent formatting in memcheck manual -- please fix
4468 169505 main.c:286 (endOfInstr):
4469 Assertion 'ii->cost_offset == *cost_offset' failed
4470 177206 Generate default.supp during compile instead of configure
4471 177209 Configure valt_load_address based on arch+os
4472 177305 eventfd / syscall 323 patch lost
4473 179731 Tests fail to build because of inlining of non-local asm labels
4474 181394 helgrind: libhb_core.c:3762 (msm_write): Assertion
4475 'ordxx == POrd_EQ || ordxx == POrd_LT' failed.
4476 181594 Bogus warning for empty text segment
4477 181707 dwarf doesn't require enumerations to have name
4478 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4479 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4480 Assertion '!already_present' failed.
4481 185359 exp-ptrcheck: unhandled syscall getresuid()
4482 185794 "WARNING: unhandled syscall: 285" (fallocate) on x86_64
4483 185816 Valgrind is unable to handle debug info for files with split
4484 debug info that are prelinked afterwards
4485 185980 [darwin] unhandled syscall: sem_open
4486 186238 bbToIR_AMD64: disInstr miscalculated next %rip
4487 186507 exp-ptrcheck unhandled syscalls prctl, etc.
4488 186790 Suppression pattern used for leaks are not reported
4489 186796 Symbols with length>200 in suppression files are ignored
4490 187048 drd: mutex PTHREAD_PROCESS_SHARED attribute missinterpretation
4491 187416 exp-ptrcheck: support for __NR_{setregid,setreuid,setresuid}
4492 188038 helgrind: hg_main.c:926: mk_SHVAL_fail: the 'impossible' happened
4493 188046 bashisms in the configure script
4494 188127 amd64->IR: unhandled instruction bytes: 0xF0 0xF 0xB0 0xA
4495 188161 memcheck: --track-origins=yes asserts "mc_machine.c:672
4496 (get_otrack_shadow_offset_wrk): the 'impossible' happened."
4497 188248 helgrind: pthread_cleanup_push, pthread_rwlock_unlock,
4498 assertion fail "!lock->heldBy"
4499 188427 Add support for epoll_create1 (with patch)
4500 188530 Support for SIOCGSTAMPNS
4501 188560 Include valgrind.spec in the tarball
4502 188572 Valgrind on Mac should suppress setenv() mem leak
4503 189054 Valgrind fails to build because of duplicate non-local asm labels
4504 189737 vex amd64->IR: unhandled instruction bytes: 0xAC
4505 189762 epoll_create syscall not handled (--tool=exp-ptrcheck)
4506 189763 drd assertion failure: s_threadinfo[tid].is_recording
4507 190219 unhandled syscall: 328 (x86-linux)
4508 190391 dup of 181394; see above
4509 190429 Valgrind reports lots of errors in ld.so with x86_64 2.9.90 glibc
4510 190820 No debug information on powerpc-linux
4511 191095 PATCH: Improve usbdevfs ioctl handling
4512 191182 memcheck: VALGRIND_LEAK_CHECK quadratic when big nr of chunks
4514 191189 --xml=yes should obey --gen-suppressions=all
4515 191192 syslog() needs a suppression on macosx
4516 191271 DARWIN: WARNING: unhandled syscall: 33554697 a.k.a.: 265
4517 191761 getrlimit on MacOSX
4518 191992 multiple --fn-skip only works sometimes; dependent on order
4519 192634 V. reports "aspacem sync_check_mapping_callback:
4520 segment mismatch" on Darwin
4521 192954 __extension__ missing on 2 client requests
4522 194429 Crash at start-up with glibc-2.10.1 and linux-2.6.29
4523 194474 "INSTALL" file has different build instructions than "README"
4524 194671 Unhandled syscall (sem_wait?) from mac valgrind
4525 195069 memcheck: reports leak (memory still reachable) for
4527 195169 drd: (vgDrd_barrier_post_wait):
4528 Assertion 'r->sg[p->post_iteration]' failed.
4529 195268 valgrind --log-file doesn't accept ~/...
4530 195838 VEX abort: LibVEX_N_SPILL_BYTES too small for CPUID boilerplate
4531 195860 WARNING: unhandled syscall: unix:223
4532 196528 need a error suppression for pthread_rwlock_init under os x?
4533 197227 Support aio_* syscalls on Darwin
4534 197456 valgrind should reject --suppressions=(directory)
4535 197512 DWARF2 CFI reader: unhandled CFI instruction 0:10
4536 197591 unhandled syscall 27 (mincore)
4537 197793 Merge DCAS branch to the trunk == 85756, 142103
4538 197794 Avoid duplicate filenames in Vex
4539 197898 make check fails on current SVN
4540 197901 make check fails also under exp-ptrcheck in current SVN
4541 197929 Make --leak-resolution=high the default
4542 197930 Reduce spacing between leak reports
4543 197933 Print command line of client at start-up, and shorten preamble
4544 197966 unhandled syscall 205 (x86-linux, --tool=exp-ptrcheck)
4545 198395 add BBV to the distribution as an experimental tool
4546 198624 Missing syscalls on Darwin: 82, 167, 281, 347
4547 198649 callgrind_annotate doesn't cumulate counters
4548 199338 callgrind_annotate sorting/thresholds are broken for all but Ir
4549 199977 Valgrind complains about an unrecognized instruction in the
4550 atomic_incs test program
4551 200029 valgrind isn't able to read Fedora 12 debuginfo
4552 200760 darwin unhandled syscall: unix:284
4553 200827 DRD doesn't work on Mac OS X
4554 200990 VG_(read_millisecond_timer)() does not work correctly
4555 201016 Valgrind does not support pthread_kill() on Mac OS
4556 201169 Document --read-var-info
4557 201323 Pre-3.5.0 performance sanity checking
4558 201384 Review user manual for the 3.5.0 release
4559 201585 mfpvr not implemented on ppc
4560 201708 tests failing because x86 direction flag is left set
4561 201757 Valgrind doesn't handle any recent sys_futex additions
4562 204377 64-bit valgrind can not start a shell script
4563 (with #!/path/to/shell) if the shell is a 32-bit executable
4564 n-i-bz drd: fixed assertion failure triggered by mutex reinitialization.
4565 n-i-bz drd: fixed a bug that caused incorrect messages to be printed
4566 about memory allocation events with memory access tracing enabled
4567 n-i-bz drd: fixed a memory leak triggered by vector clock deallocation
4569 (3.5.0: 19 Aug 2009, vex r1913, valgrind r10846).
4573 Release 3.4.1 (28 February 2009)
4574 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4575 3.4.1 is a bug-fix release that fixes some regressions and assertion
4576 failures in debug info reading in 3.4.0, most notably incorrect stack
4577 traces on amd64-linux on older (glibc-2.3 based) systems. Various
4578 other debug info problems are also fixed. A number of bugs in the
4579 exp-ptrcheck tool introduced in 3.4.0 have been fixed.
4581 In view of the fact that 3.4.0 contains user-visible regressions
4582 relative to 3.3.x, upgrading to 3.4.1 is recommended. Packagers are
4583 encouraged to ship 3.4.1 in preference to 3.4.0.
4585 The fixed bugs are as follows. Note that "n-i-bz" stands for "not in
4586 bugzilla" -- that is, a bug that was reported to us but never got a
4587 bugzilla entry. We encourage you to file bugs in bugzilla
4588 (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than mailing the
4589 developers (or mailing lists) directly -- bugs that are not entered
4590 into bugzilla tend to get forgotten about or ignored.
4592 n-i-bz Fix various bugs reading icc-11 generated debug info
4593 n-i-bz Fix various bugs reading gcc-4.4 generated debug info
4594 n-i-bz Preliminary support for glibc-2.10 / Fedora 11
4595 n-i-bz Cachegrind and Callgrind: handle non-power-of-two cache sizes,
4596 so as to support (eg) 24k Atom D1 and Core2 with 3/6/12MB L2.
4597 179618 exp-ptrcheck crashed / exit prematurely
4598 179624 helgrind: false positive races with pthread_create and
4599 recv/open/close/read
4600 134207 pkg-config output contains @VG_PLATFORM@
4601 176926 floating point exception at valgrind startup with PPC 440EPX
4602 181594 Bogus warning for empty text segment
4603 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant rex prefixes)
4604 181707 Dwarf3 doesn't require enumerations to have name
4605 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4606 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4607 Assertion '!already_present' failed.
4608 185359 exp-ptrcheck unhandled syscall getresuid()
4610 (3.4.1.RC1: 24 Feb 2008, vex r1884, valgrind r9253).
4611 (3.4.1: 28 Feb 2008, vex r1884, valgrind r9293).
4615 Release 3.4.0 (2 January 2009)
4616 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4617 3.4.0 is a feature release with many significant improvements and the
4618 usual collection of bug fixes. This release supports X86/Linux,
4619 AMD64/Linux, PPC32/Linux and PPC64/Linux. Support for recent distros
4620 (using gcc 4.4, glibc 2.8 and 2.9) has been added.
4622 3.4.0 brings some significant tool improvements. Memcheck can now
4623 report the origin of uninitialised values, the thread checkers
4624 Helgrind and DRD are much improved, and we have a new experimental
4625 tool, exp-Ptrcheck, which is able to detect overruns of stack and
4626 global arrays. In detail:
4628 * Memcheck is now able to track the origin of uninitialised values.
4629 When it reports an uninitialised value error, it will try to show
4630 the origin of the value, as either a heap or stack allocation.
4631 Origin tracking is expensive and so is not enabled by default. To
4632 use it, specify --track-origins=yes. Memcheck's speed will be
4633 essentially halved, and memory usage will be significantly
4634 increased. Nevertheless it can drastically reduce the effort
4635 required to identify the root cause of uninitialised value errors,
4636 and so is often a programmer productivity win, despite running more
4639 * A version (1.4.0) of the Valkyrie GUI, that works with Memcheck in
4640 3.4.0, will be released shortly.
4642 * Helgrind's race detection algorithm has been completely redesigned
4643 and reimplemented, to address usability and scalability concerns:
4645 - The new algorithm has a lower false-error rate: it is much less
4646 likely to report races that do not really exist.
4648 - Helgrind will display full call stacks for both accesses involved
4649 in a race. This makes it easier to identify the root causes of
4652 - Limitations on the size of program that can run have been removed.
4654 - Performance has been modestly improved, although that is very
4657 - Direct support for Qt4 threading has been added.
4659 - pthread_barriers are now directly supported.
4661 - Helgrind works well on all supported Linux targets.
4663 * The DRD thread debugging tool has seen major improvements:
4665 - Greatly improved performance and significantly reduced memory
4668 - Support for several major threading libraries (Boost.Thread, Qt4,
4669 glib, OpenMP) has been added.
4671 - Support for atomic instructions, POSIX semaphores, barriers and
4672 reader-writer locks has been added.
4674 - Works now on PowerPC CPUs too.
4676 - Added support for printing thread stack usage at thread exit time.
4678 - Added support for debugging lock contention.
4680 - Added a manual for Drd.
4682 * A new experimental tool, exp-Ptrcheck, has been added. Ptrcheck
4683 checks for misuses of pointers. In that sense it is a bit like
4684 Memcheck. However, Ptrcheck can do things Memcheck can't: it can
4685 detect overruns of stack and global arrays, it can detect
4686 arbitrarily far out-of-bounds accesses to heap blocks, and it can
4687 detect accesses heap blocks that have been freed a very long time
4688 ago (millions of blocks in the past).
4690 Ptrcheck currently works only on x86-linux and amd64-linux. To use
4691 it, use --tool=exp-ptrcheck. A simple manual is provided, as part
4692 of the main Valgrind documentation. As this is an experimental
4693 tool, we would be particularly interested in hearing about your
4694 experiences with it.
4696 * exp-Omega, an experimental instantaneous leak-detecting tool, is no
4697 longer built by default, although the code remains in the repository
4698 and the tarball. This is due to three factors: a perceived lack of
4699 users, a lack of maintenance, and concerns that it may not be
4700 possible to achieve reliable operation using the existing design.
4702 * As usual, support for the latest Linux distros and toolchain
4703 components has been added. It should work well on Fedora Core 10,
4704 OpenSUSE 11.1 and Ubuntu 8.10. gcc-4.4 (in its current pre-release
4705 state) is supported, as is glibc-2.9. The C++ demangler has been
4706 updated so as to work well with C++ compiled by even the most recent
4709 * You can now use frame-level wildcards in suppressions. This was a
4710 frequently-requested enhancement. A line "..." in a suppression now
4711 matches zero or more frames. This makes it easier to write
4712 suppressions which are precise yet insensitive to changes in
4715 * 3.4.0 adds support on x86/amd64 for the SSSE3 instruction set.
4717 * Very basic support for IBM Power6 has been added (64-bit processes only).
4719 * Valgrind is now cross-compilable. For example, it is possible to
4720 cross compile Valgrind on an x86/amd64-linux host, so that it runs
4721 on a ppc32/64-linux target.
4723 * You can set the main thread's stack size at startup using the
4724 new --main-stacksize= flag (subject of course to ulimit settings).
4725 This is useful for running apps that need a lot of stack space.
4727 * The limitation that you can't use --trace-children=yes together
4728 with --db-attach=yes has been removed.
4730 * The following bugs have been fixed. Note that "n-i-bz" stands for
4731 "not in bugzilla" -- that is, a bug that was reported to us but
4732 never got a bugzilla entry. We encourage you to file bugs in
4733 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4734 mailing the developers (or mailing lists) directly.
4736 n-i-bz Make return types for some client requests 64-bit clean
4737 n-i-bz glibc 2.9 support
4738 n-i-bz ignore unsafe .valgrindrc's (CVE-2008-4865)
4739 n-i-bz MPI_Init(0,0) is valid but libmpiwrap.c segfaults
4740 n-i-bz Building in an env without gdb gives bogus gdb attach
4741 92456 Tracing the origin of uninitialised memory
4742 106497 Valgrind does not demangle some C++ template symbols
4744 151612 Suppression with "..." (frame-level wildcards in .supp files)
4745 156404 Unable to start oocalc under memcheck on openSUSE 10.3 (64-bit)
4746 159285 unhandled syscall:25 (stime, on x86-linux)
4747 159452 unhandled ioctl 0x8B01 on "valgrind iwconfig"
4748 160954 ppc build of valgrind crashes with illegal instruction (isel)
4749 160956 mallinfo implementation, w/ patch
4750 162092 Valgrind fails to start gnome-system-monitor
4751 162819 malloc_free_fill test doesn't pass on glibc2.8 x86
4752 163794 assertion failure with "--track-origins=yes"
4753 163933 sigcontext.err and .trapno must be set together
4754 163955 remove constraint !(--db-attach=yes && --trace-children=yes)
4755 164476 Missing kernel module loading system calls
4756 164669 SVN regression: mmap() drops posix file locks
4757 166581 Callgrind output corruption when program forks
4758 167288 Patch file for missing system calls on Cell BE
4759 168943 unsupported scas instruction pentium
4760 171645 Unrecognised instruction (MOVSD, non-binutils encoding)
4761 172417 x86->IR: 0x82 ...
4762 172563 amd64->IR: 0xD9 0xF5 - fprem1
4763 173099 .lds linker script generation error
4764 173177 [x86_64] syscalls: 125/126/179 (capget/capset/quotactl)
4765 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant prefixes)
4767 174908 --log-file value not expanded correctly for core file
4768 175044 Add lookup_dcookie for amd64
4769 175150 x86->IR: 0xF2 0xF 0x11 0xC1 (movss non-binutils encoding)
4771 Developer-visible changes:
4773 * Valgrind's debug-info reading machinery has been majorly overhauled.
4774 It can now correctly establish the addresses for ELF data symbols,
4775 which is something that has never worked properly before now.
4777 Also, Valgrind can now read DWARF3 type and location information for
4778 stack and global variables. This makes it possible to use the
4779 framework to build tools that rely on knowing the type and locations
4780 of stack and global variables, for example exp-Ptrcheck.
4782 Reading of such information is disabled by default, because most
4783 tools don't need it, and because it is expensive in space and time.
4784 However, you can force Valgrind to read it, using the
4785 --read-var-info=yes flag. Memcheck, Helgrind and DRD are able to
4786 make use of such information, if present, to provide source-level
4787 descriptions of data addresses in the error messages they create.
4789 (3.4.0.RC1: 24 Dec 2008, vex r1878, valgrind r8882).
4790 (3.4.0: 3 Jan 2009, vex r1878, valgrind r8899).