1 Release 3.19.0 (08 Apr 2022)
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 * Fix Rust v0 name demangling.
14 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
15 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
16 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with HardenedBSD
18 * ================== PLATFORM CHANGES =================
21 - ignore the "v8.x" architecture levels, only look at actual CPU features
22 present. Fixes mismatch detected between RDMA and atomics features
23 preventing startup on some QEMU configurations.
24 - Implement LD{,A}XP and ST{,L}XP
25 - Fix incorrect code emitted for doubleword CAS.
28 - Fix sys_ipc semtimedop syscall
29 - Fix VFLRX and WFLRX instructions
30 - Fix EXRL instruction with negative offset
33 - Reimplement the vbpermq instruction support to generate less Iops and
34 avoid overflowing internal buffers.
35 - Fix checking for scv support to avoid "Facility 'SCV' unavailable (12),
36 exception" messages in dmsg.
37 - Fix setting condition code for Vector Compare quad word instructions.
38 - Fix fix lxsibzx, lxsihzx and lxsihzx instructions so they only load
39 their respective sized data.
40 - Fix the prefixed stq instruction in PC relative mode.
42 * ==================== TOOL CHANGES ===================
45 - Speed up --track-origins=yes for large (in the range of hundreds to
46 thousands of megabytes) mmap/munmaps.
48 - Several fixes for new versions of libstd++ using new posix try_lock
51 * ==================== FIXED BUGS ====================
53 The following bugs have been fixed or resolved. Note that "n-i-bz"
54 stands for "not in bugzilla" -- that is, a bug that was reported to us
55 but never got a bugzilla entry. We encourage you to file bugs in
56 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
57 than mailing the developers (or mailing lists) directly -- bugs that
58 are not entered into bugzilla tend to get forgotten about or ignored.
60 403802 leak_cpp_interior fails with some reachable blocks different than expected
61 435732 memcheck/tests/leak_cpp_interior fails with gcc11
62 444242 s390x: Valgrind crashes on EXRL with negative offset
63 444399 arm64: unhandled instruction 0xC87F2D89 (LD{,A}XP and ST{,L}XP).
65 444481 gdb_server test failures on s390x
66 444495 dhat/tests/copy fails on s390x
67 444552 memcheck/tests/sem fails on s390x with glibc 2.34
68 444571 PPC, fix the lxsibzx and lxsihzx so they only load their respective
70 444836 PPC, pstq instruction for R=1 is not storing to the correct address.
71 444925 fexecve syscall wrapper not properly implemented
72 445032 valgrind/memcheck crash with SIGSEGV when SIGVTALRM timer used and
74 445300 [PATCH] Fix building tests with Musl
75 445354 arm64 backend: incorrect code emitted for doubleword CAS
76 445415 arm64 front end: alignment checks missing for atomic instructions
77 445504 Using C++ condition_variable results in bogus "mutex is locked simultaneously by two threads" warning
78 445607 Unhandled amd64-freebsd syscall: 247
79 445668 Inline stack frame generation is broken for Rust binaries
80 445916 Demangle Rust v0 symbols with .llvm suffix
81 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false positives
82 446138 DRD/Helgrind with std::timed_mutex::try_lock_until false positives
83 446281 Add a DRD suppression for fwrite
84 446103 Memcheck: `--track-origins=yes` causes extreme slowdowns for large mmap/munmap
85 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false
86 446251 TARGET_SIGNAL_THR added to enum target_signal
87 446823 FreeBSD - missing syscalls when using libzm4
88 447991 s390x: Valgrind indicates illegal instruction on wflrx
89 447995 Valgrind segfault on power10 due to hwcap checking code
90 449483 Powerpc: vcmpgtsq., vcmpgtuq,, vcmpequq. instructions not setting the
91 condition code correctly.
92 449672 ppc64 --track-origins=yes failures because of bad cmov addHRegUse
93 449838 sigsegv liburing the 'impossible' happened for io_uring_setup
94 450025 Powerc: ACC file not implemented as a logical overlay of the VSR
96 450437 Warn for execve syscall with argv or argv[0] being NULL
97 450536 Powerpc: valgrind throws 'facility scv unavailable exception'
98 451626 Syscall param bpf(attr->raw_tracepoint.name) points to unaddressable byte(s)
99 451827 [ppc64le] VEX temporary storage exhausted with several vbpermq instructions
100 451843 valgrind fails to start on a FreeBSD system which enforces W^X
102 To see details of a given bug, visit
103 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
104 where XXXXXX is the bug number as listed above.
106 (3.19.0.RC1: 02 Apr 2022)
109 Release 3.18.0 (15 Oct 2021)
110 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
112 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
113 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
114 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
115 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
116 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
117 AMD64/macOS 10.13 and nanoMIPS/Linux.
119 * ==================== CORE CHANGES ===================
121 * The libiberty demangler has been updated, which brings support for
122 Rust v0 name demangling. [Update: alas, due to a bug, this support
123 isn't working in 3.18.0.]
125 * __libc_freeres isn't called anymore after the program recieves a
126 fatal signal. Causing some internal glibc resources to hang around,
127 but preventing any crashes after the program has ended.
129 * The DWARF reader is now very much faster at startup when just
130 --read-inline-info=yes (the default in most cases) is given.
132 * glibc 2.34, which moved various functions from libpthread.so into
133 libc.so, is now supported.
135 * ================== PLATFORM CHANGES =================
139 - v8.2 scalar and vector FABD, FACGE, FACGT and FADD.
140 - v8.2 FP compare & conditional compare instructions.
141 - Zero variants of v8.2 FP compare instructions.
145 - Support the miscellaneous-instruction-extensions facility 3 and
146 the vector-enhancements facility 2. This enables programs
147 compiled with "-march=arch13" or "-march=z15" to be executed
152 - ISA 3.1 support is now complete
153 - ISA 3.0 support for the darn instruction added.
154 - ISA 3.0 support for the vector system call instruction scv added.
155 - ISA 3.0 support for the copy, paste and cpabort instructions added.
157 * Support for X86/FreeBSD and AMD64/FreeBSD has been added.
159 * ==================== OTHER CHANGES ====================
161 * Memcheck on amd64: minor fixes to remove some false positive
164 * ==================== FIXED BUGS ====================
166 The following bugs have been fixed or resolved. Note that "n-i-bz"
167 stands for "not in bugzilla" -- that is, a bug that was reported to us
168 but never got a bugzilla entry. We encourage you to file bugs in
169 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
170 than mailing the developers (or mailing lists) directly -- bugs that
171 are not entered into bugzilla tend to get forgotten about or ignored.
173 208531 [PATCH]: FreeBSD support for valgrind
174 368960 WARNING: unhandled amd64-linux syscall: 163 (acct)
175 407589 [Linux] Add support for C11 aligned_alloc() and GNU reallocarray()
176 423963 Error in child thread when CLONE_PIDFD is used
177 426148 crash with "impossible happened" when running BPF CO-RE programs
178 429375 PPC ISA 3.1 support is missing, part 9
179 431157 PPC_FEATURE2_SCV needs to be masked in AT_HWCAP2
180 431306 Update demangler to support Rust v0 name mangling
181 432387 s390x: z15 instructions support
182 433437 FreeBSD support, part 1
183 433438 FreeBSD support, part 2
184 433439 FreeBSD support, part 3
185 433469 FreeBSD support, part 4
186 433473 FreeBSD support, part 5
187 433477 FreeBSD support, part 6
188 433479 FreeBSD support, part 7
189 433504 FreeBSD support, part 8
190 433506 FreeBSD support, part 9
191 433507 FreeBSD support, part 10
192 433508 FreeBSD support, part 11
193 433510 FreeBSD support, part 12
194 433801 PPC ISA 3.1 support is missing, part 10 (ISA 3.1 support complete)
195 433863 s390x: memcheck/tests/s390x/{cds,cs,csg} failures
196 434296 s390x: False-positive memcheck diagnostics from vector string
198 434840 PPC64 darn instruction not supported
199 435665 PPC ISA 3.0 copy, paste, cpabort instructions are not supported
200 435908 valgrind tries to fetch from deubginfod for files which already
201 have debug information
202 438871 unhandled instruction bytes: 0xF3 0x49 0xF 0x6F 0x9C 0x24 0x60 0x2
203 439046 valgrind is unusably large when linked with lld
204 439090 Implement close_range(2)
205 439326 Valgrind 3.17.0 won't compile with Intel 2021 oneAPI compilers
206 439590 glibc-2.34 breaks suppressions against obj:*/lib*/libc-2.*so*
207 440670 unhandled ppc64le-linux syscall: 252 statfs64 and 253 fstatfs64
208 440906 Fix impossible constraint issue in P10 testcase.
209 441512 Remove a unneeded / unnecessary prefix check.
210 441534 Update the expected output for test_isa_3_1_VRT.
211 442061 very slow execution under Fedora 34 (readdwarf3)
212 443031 Gcc -many change requires explicit .machine directives
213 443033 Add support for the ISA 3.0 mcrxrx instruction
214 443034 Sraw, srawi, srad, sradi, mfs
215 443178 Powerpc, test jm-mfspr expected output needs to be updated.
216 443179 Need new test for the lxvx and stxvx instructions on ISA 2.07 and
218 443180 The subnormal test and the ISA 3.0 test generate compiler warnings
219 443314 In the latest GIT version, Valgrind with "--trace-flags" crashes
221 443605 Don't call final_tidyup (__libc_freeres) on FatalSignal
223 To see details of a given bug, visit
224 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
225 where XXXXXX is the bug number as listed below.
227 (3.18.0.RC1: 12 Oct 2021)
228 (3.18.0: 15 Oct 2021)
232 Release 3.17.0 (19 Mar 2021)
233 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
235 3.17.0 fixes a number of bugs and adds some functional changes: support for
236 GCC 11, Clang 11, DWARF5 debuginfo, the 'debuginfod' debuginfo server, and
237 some new instructions for Arm64, S390 and POWER. There are also some tool
240 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
241 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
242 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
243 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
244 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
246 * ==================== CORE CHANGES ===================
248 * DWARF version 5 support. Valgrind can now read DWARF version 5 debuginfo as
251 * Valgrind now supports debuginfod, an HTTP server for distributing ELF/DWARF
252 debugging information. When a debuginfo file cannot be found locally,
253 Valgrind is able to query debuginfod servers for the file using its
254 build-id. See the user manual for more information about debuginfod support.
256 * ================== PLATFORM CHANGES =================
260 - Inaccuracies resulting from double-rounding in the simulation of
261 floating-point multiply-add/subtract instructions have been fixed. These
262 should now behave exactly as the hardware does.
264 - Partial support for the ARM v8.2 instruction set. v8.2 support work is
265 ongoing. Support for the half-word variants of at least the following
266 instructions has been added:
268 FABS <Vd>.<T>, <Vn>.<T>
270 FNEG <Vd>.<T>, <Vn>.<T>
272 FSQRT <Vd>.<T>, <Vn>.<T>
277 - Implement the new instructions/features that were added to z/Architecture
278 with the vector-enhancements facility 1. Also cover the instructions from
279 the vector-packed-decimal facility that are defined outside the chapter
280 "Vector Decimal Instructions", but not the ones from that chapter itself.
282 For a detailed list of newly supported instructions see the updates to
283 `docs/internals/s390-opcodes.csv'.
285 Since the miscellaneous instruction extensions facility 2 was already
286 added in Valgrind 3.16.0, this completes the support necessary to run
287 general programs built with `--march=z14' under Valgrind. The
288 vector-packed-decimal facility is currently not exploited by the standard
289 toolchain and libraries.
293 - Various bug fixes. Fix for the sync field to limit setting just two of
294 the two bits in the L-field. Fix the write size for the stxsibx and
295 stxsihx instructions. Fix the modsw and modsd instructions.
297 - Partial support for ISA 3.1 has been added. Support for the VSX PCV mask
298 instructions, bfloat16 GER instructions, and bfloat16 to/from float 32-bit
299 conversion instructions are still missing.
301 * ==================== TOOL CHANGES ====================
303 * General tool changes
305 - All the tools and their vgpreload libraries are now installed under
306 libexec because they cannot be executed directly and should be run through
307 the valgrind executable. This should be an internal, not user visible,
308 change, but might impact valgrind packagers.
310 - The --track-fds option now respects -q, --quiet and won't output anything
311 if no file descriptors are leaked. It also won't report the standard stdin
312 (0), stdout (1) or stderr (2) descriptors as being leaked with
313 --trace-fds=yes anymore. To track whether the standard file descriptors
314 are still open at the end of the program run use --trace-fds=all.
318 - DHAT has been extended, with two new modes of operation. The new
319 --mode=copy flag triggers copy profiling, which records calls to memcpy,
320 strcpy, and similar functions. The new --mode=ad-hoc flag triggers ad hoc
321 profiling, which records calls to the DHAT_AD_HOC_EVENT client request in
322 the new dhat/dhat.h file. This is useful for learning more about hot code
323 paths. See the user manual for more information about the new modes.
325 - Because of these changes, DHAT's file format has changed. DHAT output
326 files produced with earlier versions of DHAT will not work with this
327 version of DHAT's viewer, and DHAT output files produced with this version
328 of DHAT will not work with earlier versions of DHAT's viewer.
330 * ==================== FIXED BUGS ====================
332 The following bugs have been fixed or resolved. Note that "n-i-bz"
333 stands for "not in bugzilla" -- that is, a bug that was reported to us
334 but never got a bugzilla entry. We encourage you to file bugs in
335 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
336 than mailing the developers (or mailing lists) directly -- bugs that
337 are not entered into bugzilla tend to get forgotten about or ignored.
339 To see details of a given bug, visit
340 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
341 where XXXXXX is the bug number as listed below.
343 140178 open("/proc/self/exe", ...); doesn't quite work
344 140939 --track-fds reports leakage of stdout/in/err and doesn't respect -q
345 217695 malloc/calloc/realloc/memalign failure doesn't set errno to ENOMEM
346 338633 gdbserver_tests/nlcontrolc.vgtest hangs on arm64
347 345077 linux syscall execveat support (linux 3.19)
348 361770 Missing F_ADD_SEALS
349 369029 handle linux syscalls sched_getattr and sched_setattr
350 384729 __libc_freeres inhibits cross-platform valgrind
351 388787 Support for C++17 new/delete
352 391853 Makefile.all.am:L247 and @SOLARIS_UNDEF_LARGESOURCE@ being empty
353 396656 Warnings while reading debug info
354 397605 ioctl FICLONE mishandled
355 401416 Compile failure with openmpi 4.0
356 408663 Suppression file for musl libc
357 404076 s390x: z14 vector instructions not implemented
358 410743 shmat() calls for 32-bit programs fail when running in 64-bit valgrind
359 (actually affected all x86 and nanomips regardless of host bitness)
360 413547 regression test does not check for Arm 64 features.
361 414268 Enable AArch64 feature detection and decoding for v8.x instructions
362 415293 Incorrect call-graph tracking due to new _dl_runtime_resolve_xsave*
363 422174 unhandled instruction bytes: 0x48 0xE9 (REX prefixed JMP instruction)
364 422261 platform selection fails for unqualified client name
365 422623 epoll_ctl warns for uninitialized padding on non-amd64 64bit arches
366 423021 PPC: Add missing ISA 3.0 documentation link and HWCAPS test.
367 423195 PPC ISA 3.1 support is missing, part 1
368 423361 Adds io_uring support on arm64/aarch64 (and all other arches)
369 424012 crash with readv/writev having invalid but not NULL arg2 iovec
370 424298 amd64: Implement RDSEED
371 425232 PPC ISA 3.1 support is missing, part 2
372 425820 Failure to recognize vpcmpeqq as a dependency breaking idiom.
373 426014 arm64: implement fmadd and fmsub as Iop_MAdd/Sub
374 426123 PPC ISA 3.1 support is missing, part 3
375 426144 Fix "condition variable has not been initialized" on Fedora 33.
376 427400 PPC ISA 3.1 support is missing, part 4
377 427401 PPC ISA 3.1 support is missing, part 5
378 427404 PPC ISA 3.1 support is missing, part 6
379 427870 lmw, lswi and related PowerPC insns aren't allowed on ppc64le
380 427787 Support new faccessat2 linux syscall (439)
381 427969 debuginfo section duplicates a section in the main ELF file
382 428035 drd: Unbreak the musl build
383 428648 s390_emit_load_mem panics due to 20-bit offset for vector load
384 428716 cppcheck detects potential leak in VEX/useful/smchash.c
385 428909 helgrind: need to intercept duplicate libc definitions for Fedora 33
386 429352 PPC ISA 3.1 support is missing, part 7
387 429354 PPC ISA 3.1 support is missing, part 8
388 429692 unhandled ppc64le-linux syscall: 147 (getsid)
389 429864 s390x: C++ atomic test_and_set yields false-positive memcheck
391 429952 Errors when building regtest with clang
392 430354 ppc stxsibx and stxsihx instructions write too much data
393 430429 valgrind.h doesn't compile on s390x with clang
394 430485 expr_is_guardable doesn't handle Iex_Qop
395 431556 Complete arm64 FADDP v8.2 instruction support
396 432102 Add support for DWARF5 as produced by GCC11
397 432161 Addition of arm64 v8.2 FADDP, FNEG and FSQRT
398 432381 drd: Process STACK_REGISTER client requests
399 432552 [AArch64] invalid error emitted for pre-decremented byte/hword addresses
400 432672 vg_regtest: test-specific environment variables not reset between tests
401 432809 VEX should support REX.W + POPF
402 432861 PPC modsw and modsd give incorrect results for 1 mod 12
403 432870 gdbserver_tests:nlcontrolc hangs with newest glibc2.33 x86-64
404 432215 Add debuginfod functionality
405 433323 Use pkglibexecdir as vglibdir
406 433500 DRD regtest faulures when libstdc++ and libgcc debuginfo are installed
407 433629 valgrind/README has type "abd" instead of "and"
408 433641 Rust std::sys::unix::fs::try_statx Syscall param fstatat(file_name)
409 433898 arm64: Handle sp, lr, fp as DwReg in CfiExpr
410 434193 GCC 9+ inlined strcmp causes "Conditional jump or move [..] value" report
411 n-i-bz helgrind: If hg_cli__realloc fails, return NULL.
412 n-i-bz arm64 front end: avoid Memcheck false positives relating to CPUID
414 (3.17.0.RC1: 13 Mar 2021)
415 (3.17.0.RC2: 17 Mar 2021)
416 (3.17.0: 19 Mar 2021)
420 Release 3.16.1 (22 June 2020)
421 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
423 3.16.1 fixes two critical bugs discovered after 3.16.0 was frozen. It also
424 fixes character encoding problems in the documentation HTML.
426 422677 PPC sync instruction L field should only be 2 bits in ISA 3.0
427 422715 32-bit x86: vex: the `impossible' happened: expr_is_guardable: unhandled expr
429 (3.16.1, 22 June 2020, 36d6727e1d768333a536f274491e5879cab2c2f7)
433 Release 3.16.0 (27 May 2020)
434 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
436 3.16.0 is a feature release with many improvements and the usual collection of
439 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
440 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
441 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
442 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
443 support for X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
445 * ==================== CORE CHANGES ===================
447 * It is now possible to dynamically change the value of many command line
448 options while your program (or its children) are running under Valgrind.
450 To see the list of dynamically changeable options, run
451 "valgrind --help-dyn-options".
453 You can change the options from the shell by using vgdb to launch
454 the monitor command "v.clo <clo option>...".
455 The same monitor command can be used from a gdb connected
456 to the valgrind gdbserver.
457 Your program can also change the dynamically changeable options using
458 the client request VALGRIND_CLO_CHANGE(option).
460 * ================== PLATFORM CHANGES =================
462 * MIPS: preliminary support for nanoMIPS instruction set has been added.
464 * ==================== TOOL CHANGES ====================
468 - The implicit memcpy done by each call to realloc now counts towards the
469 read and write counts of resized heap blocks, making those counts higher
474 - cg_annotate's --auto and --show-percs options now default to 'yes', because
475 they are usually wanted.
479 - callgrind_annotate's --auto and --show-percs options now default to 'yes',
480 because they are usually wanted.
482 - The command option --collect-systime has been enhanced to specify
483 the unit used to record the elapsed time spent during system calls.
484 The command option now accepts the values no|yes|msec|usec|nsec,
485 where yes is a synonym of msec. When giving the value nsec, the
486 system cpu time of system calls is also recorded.
490 - Several memcheck options are now dynamically changeable.
491 Use valgrind --help-dyn-options to list them.
493 - The release 3.15 introduced a backward incompatible change for
494 some suppression entries related to preadv and pwritev syscalls.
495 When reading a suppression entry using the unsupported 3.14 format,
496 valgrind will now produce a warning to say the suppression entry will not
497 work, and suggest the needed change.
499 - Significantly fewer false positive errors on optimised code generated by
500 Clang and GCC. In particular, Memcheck now deals better with the
501 situation where the compiler will transform C-level "A && B" into "B && A"
502 under certain circumstances (in which the transformation is valid).
503 Handling of integer equality/non-equality checks on partially defined
504 values is also improved on some architectures.
508 - The exprimental Stack and Global Array Checking tool has been removed.
509 It only ever worked on x86 and amd64, and even on those it had a
510 high false positive rate and was slow. An alternative for detecting
511 stack and global array overruns is using the AddressSanitizer (ASAN)
512 facility of the GCC and Clang compilers, which require you to rebuild
513 your code with -fsanitize=address.
515 * ==================== OTHER CHANGES ====================
517 * New and modified GDB server monitor features:
519 - Option -T tells vgdb to output a timestamp in the vgdb information messages.
521 - The gdbserver monitor commands that require an address and an optional
522 length argument now accepts the alternate 'C like' syntax "address[length]".
523 For example, the memcheck command "monitor who_points_at 0x12345678 120"
524 can now also be given as "monitor who_points_at 0x12345678[120]".
526 * ==================== FIXED BUGS ====================
528 The following bugs have been fixed or resolved. Note that "n-i-bz"
529 stands for "not in bugzilla" -- that is, a bug that was reported to us
530 but never got a bugzilla entry. We encourage you to file bugs in
531 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
532 than mailing the developers (or mailing lists) directly -- bugs that
533 are not entered into bugzilla tend to get forgotten about or ignored.
535 To see details of a given bug, visit
536 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
537 where XXXXXX is the bug number as listed below.
539 343099 Linux setns syscall wrapper missing, unhandled syscall: 308
540 == 368923 WARNING: unhandled arm64-linux syscall: 268 (setns)
541 == 369031 WARNING: unhandled amd64-linux syscall: 308 (setns)
542 385386 Assertion failed "szB >= CACHE_ENTRY_SIZE" at m_debuginfo/image.c:517
543 400162 Patch: Guard against __GLIBC_PREREQ for musl libc
544 400593 In Coregrind, use statx for some internal syscalls if [f]stat[64] fail
545 400872 Add nanoMIPS support to Valgrind
546 403212 drd/tests/trylock hangs on FreeBSD
547 404406 s390x: z14 miscellaneous instructions not implemented
548 405201 Incorrect size of struct vki_siginfo on 64-bit Linux architectures
549 406561 mcinfcallWSRU gdbserver_test fails on ppc64
550 406824 Unsupported baseline
551 407218 Add support for the copy_file_range syscall
552 407307 Intercept stpcpy also in ld.so for arm64
553 407376 Update Xen support to 4.12 (4.13, actually) and add more coverage
555 407764 drd cond_post_wait gets wrong (?) condition on s390x z13 system
556 408009 Expose rdrand and f16c even on avx if host cpu supports them
557 408091 Missing pkey syscalls
558 408414 Add support for missing for preadv2 and pwritev2 syscalls
559 409141 Valgrind hangs when SIGKILLed
560 409206 Support for Linux PPS and PTP ioctls
561 409367 exit_group() after signal to thread waiting in futex() causes hangs
562 409429 amd64: recognize 'cmpeq' variants as a dependency breaking idiom
563 409780 References to non-existent configure.in
564 410556 Add support for BLKIO{MIN,OPT} and BLKALIGNOFF ioctls
565 410599 Non-deterministic behaviour of pth_self_kill_15_other test
566 410757 discrepancy for preadv2/pwritev2 syscalls across different versions
567 411134 Allow the user to change a set of command line options during execution
568 411451 amd64->IR of bt/btc/bts/btr with immediate clears zero flag
569 412344 Problem setting mips flags with specific paths
570 412408 unhandled arm-linux syscall: 124 - adjtime - on arm-linux
571 413119 Ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
572 413330 avx-1 test fails on AMD EPYC 7401P 24-Core Processor
573 413603 callgrind_annotate/cg_annotate truncate function names at '#'
574 414565 Specific use case bug found in SysRes VG_(do_sys_sigprocmask)
575 415136 ARMv8.1 Compare-and-Swap instructions are not supported
576 415757 vex x86->IR: 0x66 0xF 0xCE 0x4F (bswapw)
577 416239 valgrind crashes when handling clock_adjtime
578 416285 Use prlimit64 in VG_(getrlimit) and VG_(setrlimit)
579 416286 DRD reports "conflicting load" error on std::mutex::lock()
580 416301 s390x: "compare and signal" not supported
581 416387 finit_module and bpf syscalls are unhandled on arm64
582 416464 Fix false reports for uninitialized memory for PR_CAPBSET_READ/DROP
583 416667 gcc10 ppc64le impossible constraint in 'asm' in test_isa.
584 416753 new 32bit time syscalls for 2038+
585 417075 pwritev(vector[...]) suppression ignored
586 417075 is not fixed, but incompatible supp entries are detected
587 and a warning is produced for these.
588 417187 [MIPS] Conditional branch problem since 'grail' changes
589 417238 Test memcheck/tests/vbit-test fails on mips64 BE
590 417266 Make memcheck/tests/linux/sigqueue usable with musl
591 417281 s390x: /bin/true segfaults with "grail" enabled
592 417427 commit to fix vki_siginfo_t definition created numerous regression
594 417452 s390_insn_store_emit: dst->tag for HRcVec128
595 417578 Add suppressions for glibc DTV leaks
596 417906 clone with CLONE_VFORK and no CLONE_VM fails
597 418004 Grail code additions break ppc64.
598 418435 s390x: spurious "Conditional jump or move depends on uninitialised [..]"
599 418997 s390x: Support Iex_ITE for float and vector types
600 419503 s390x: Avoid modifying registers returned from isel functions
601 421321 gcc10 arm64 build needs __getauxval for linking with libgcc
602 421570 std_mutex fails on Arm v8.1 h/w
603 434035 vgdb might crash if valgrind is killed
604 n-i-bz Fix minor one time leaks in dhat.
605 n-i-bz Add --run-cxx-freeres=no in outer args to avoid inner crashes.
606 n-i-bz Add support for the Linux io_uring system calls
607 n-i-bz sys_statx: don't complain if both |filename| and |buf| are NULL.
608 n-i-bz Fix non-glibc build of test suite with s390x_features
609 n-i-bz MinGW, include/valgrind.h: Fix detection of 64-bit mode
610 423195 PPC ISA 3.1 support is missing, part 1
612 (3.16.0.RC1: 18 May 2020, git 6052ee66a0cf5234e8e2a2b49a8760226bc13b92)
613 (3.16.0.RC2: 19 May 2020, git 940ec1ca69a09f7fdae3e800b7359f85c13c4b37)
614 (3.16.0: 27 May 2020, git bf5e647edb9e96cbd5c57cc944984402eeee296d)
618 Release 3.15.0 (12 April 2019)
619 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
621 3.15.0 is a feature release with many improvements and the usual collection of
624 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
625 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
626 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
627 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
628 support for X86/macOS 10.13 and AMD64/macOS 10.13.
630 * ==================== CORE CHANGES ===================
632 * The XTree Massif output format now makes use of the information obtained
633 when specifying --read-inline-info=yes.
635 * amd64 (x86_64): the RDRAND and F16C insn set extensions are now supported.
637 * ==================== TOOL CHANGES ====================
641 - DHAT been thoroughly overhauled, improved, and given a GUI. As a result,
642 it has been promoted from an experimental tool to a regular tool. Run it
643 with --tool=dhat instead of --tool=exp-dhat.
645 - DHAT now prints only minimal data when the program ends, instead writing
646 the bulk of the profiling data to a file. As a result, the --show-top-n
647 and --sort-by options have been removed.
649 - Profile results can be viewed with the new viewer, dh_view.html. When
650 a run ends, a short message is printed, explaining how to view the result.
652 - See the documentation for more details.
656 - cg_annotate has a new option, --show-percs, which prints percentages next
661 - callgrind_annotate has a new option, --show-percs, which prints percentages
662 next to all event counts.
664 - callgrind_annotate now inserts commas in call counts, and
665 sort the caller/callee lists in the call tree.
669 - The default value for --read-inline-info is now "yes" on
670 Linux/Android/Solaris. It is still "no" on other OS.
674 - The option --xtree-leak=yes (to output leak result in xtree format)
675 automatically activates the option --show-leak-kinds=all, as xtree
676 visualisation tools such as kcachegrind can in any case select what kind
677 of leak to visualise.
679 - There has been further work to avoid false positives. In particular,
680 integer equality on partially defined inputs (C == and !=) is now handled
683 * ==================== OTHER CHANGES ====================
685 * The new option --show-error-list=no|yes displays, at the end of the run, the
686 list of detected errors and the used suppressions. Prior to this change,
687 showing this information could only be done by specifying "-v -v", but that
688 also produced a lot of other possibly-non-useful messages. The option -s is
689 equivalent to --show-error-list=yes.
691 * ==================== FIXED BUGS ====================
693 The following bugs have been fixed or resolved. Note that "n-i-bz"
694 stands for "not in bugzilla" -- that is, a bug that was reported to us
695 but never got a bugzilla entry. We encourage you to file bugs in
696 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
697 than mailing the developers (or mailing lists) directly -- bugs that
698 are not entered into bugzilla tend to get forgotten about or ignored.
700 To see details of a given bug, visit
701 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
702 where XXXXXX is the bug number as listed below.
704 385411 s390x: z13 vector floating-point instructions not implemented
705 397187 z13 vector register support for vgdb gdbserver
706 398183 Vex errors with _mm256_shuffle_epi8/vpshufb
707 398870 Please add support for instruction vcvtps2ph
708 399287 amd64 front end: Illegal Instruction vcmptrueps
709 399301 Use inlined frames in Massif XTree output.
710 399322 Improve callgrind_annotate output
711 399444 VEX/priv/guest_s390_toIR.c:17407: (style) Mismatching assignment [..]
712 400164 helgrind test encounters mips x-compiler warnings and assembler error
713 400490 s390x: VRs allocated as if separate from FPRs
714 400491 s390x: Operand of LOCH treated as unsigned integer
715 400975 Compile error: error: '-mips64r2' conflicts with the other architecture
716 options, which specify a mips64 processor
717 401112 LLVM 5.0 generates comparison against partially initialized data
718 401277 More bugs in z13 support
719 401454 Add a --show-percs option to cg_annotate and callgrind_annotate.
720 401578 drd: crashes sometimes on fork()
721 401627 memcheck errors with glibc avx2 optimized wcsncmp
722 401822 none/tests/ppc64/jm-vmx fails and produces assembler warnings
723 401827 none/tests/ppc64/test_isa_2_06_part3 failure on ppc64le (xvrsqrtesp)
724 401828 none/tests/ppc64/test_isa_2_06_part1 failure on ppc64le (fcfids and
726 402006 mark helper regs defined in final_tidyup before freeres_wrapper call
727 402048 WARNING: unhandled ppc64[be|le]-linux syscall: 26 (ptrace)
728 402123 invalid assembler opcodes for mips32r2
729 402134 assertion fail in mc_translate.c (noteTmpUsesIn) Iex_VECRET on arm64
730 402327 Warning: DWARF2 CFI reader: unhandled DW_OP_ opcode 0x13 (DW_OP_drop)
731 402341 drd/tests/tsan_thread_wrappers_pthread.h:369: suspicious code ?
732 402351 mips64 libvexmultiarch_test fails on s390x
734 402395 coregrind/vgdb-invoker-solaris.c: 2 * poor error checking
735 402480 Do not use %rsp in clobber list
736 402481 vbit-test fails on x86 for Iop_CmpEQ64 iselInt64Expr Sar64
737 402515 Implement new option --show-error-list=no|yes / -s
738 402519 POWER 3.0 addex instruction incorrectly implemented
739 402781 Redo the cache used to process indirect branch targets
740 403123 vex amd64->IR:0xF3 0x48 0xF 0xAE 0xD3 (wrfsbase)
741 403552 s390x: wrong facility bit checked for vector facility
742 404054 memcheck powerpc subfe x, x, x initializes x to 0 or -1 based on CA
743 404638 Add VG_(replaceIndexXA)
744 404843 s390x: backtrace sometimes ends prematurely
745 404888 autotools cleanup series
746 405079 unhandled ppc64le-linux syscall: 131 (quotactl)
747 405182 Valgrind fails to build with Clang
748 405205 filter_libc: remove the line holding the futex syscall error entirely
749 405356 PPC64, xvcvsxdsp, xvcvuxdsp are supposed to write the 32-bit result to
750 the upper and lower 32-bits of the 64-bit result
751 405362 PPC64, vmsummbm instruction doesn't handle overflow case correctly
752 405363 PPC64, xvcvdpsxws, xvcvdpuxws, do not handle NaN arguments correctly.
753 405365 PPC64, function _get_maxmin_fp_NaN() doesn't handle QNaN, SNaN case
755 405403 s390x disassembler cannot be used on x86
756 405430 Use gcc -Wimplicit-fallthrough=2 by default if available
757 405458 MIPS mkFormVEC arguments swapped?
758 405716 drd: Fix an integer overflow in the stack margin calculation
759 405722 Support arm64 core dump
760 405733 PPC64, xvcvdpsp should write 32-bit result to upper and lower 32-bits
761 of the 64-bit destination field.
762 405734 PPC64, vrlwnm, vrlwmi, vrldrm, vrldmi do not work properly when me < mb
763 405782 "VEX temporary storage exhausted" when attempting to debug slic3r-pe
764 406198 none/tests/ppc64/test_isa_3_0_other test sporadically including CA
766 406256 PPC64, vector floating point instructions don't handle subnormal
767 according to VSCR[NJ] bit setting.
768 406352 cachegrind/callgrind fails ann tests because of missing a.c
769 406354 dhat is broken on x86 (32bit)
770 406355 mcsignopass, mcsigpass, mcbreak fail due to difference in gdb output
771 406357 gdbserver_tests fails because of gdb output change
772 406360 memcheck/tests/libstdc++.supp needs more supression variants
773 406422 none/tests/amd64-linux/map_32bits.vgtest fails too easily
774 406465 arm64 insn selector fails on "t0 = <expr>" where <expr> has type Ity_F16
775 407340 PPC64, does not support the vlogefp, vexptefp instructions.
776 n-i-bz add syswrap for PTRACE_GET|SET_THREAD_AREA on amd64.
777 n-i-bz Fix callgrind_annotate non deterministic order for equal total
778 n-i-bz callgrind_annotate --threshold=100 does not print all functions.
779 n-i-bz callgrind_annotate Use of uninitialized value in numeric gt (>)
780 n-i-bz amd64 (x86_64): RDRAND and F16C insn set extensions are supported
782 (3.15.0.RC1: 8 April 2019, git ce94d674de5b99df173aad4c3ee48fc2a92e5d9c)
783 (3.15.0.RC2: 11 April 2019, git 0c8be9bbede189ec580ec270521811766429595f)
784 (3.15.0: 14 April 2019, git 270037da8b508954f0f7d703a0bebf5364eec548)
788 Release 3.14.0 (9 October 2018)
789 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
791 3.14.0 is a feature release with many improvements and the usual collection of
794 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
795 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
796 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
797 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12. There is also preliminary
798 support for X86/macOS 10.13, AMD64/macOS 10.13.
800 * ==================== CORE CHANGES ===================
802 * The new option --keep-debuginfo=no|yes (default no) can be used to retain
803 debug info for unloaded code. This allows saved stack traces (e.g. for
804 memory leaks) to include file/line info for code that has been dlclose'd (or
805 similar). See the user manual for more information and known limitations.
807 * Ability to specify suppressions based on source file name and line number.
809 * Majorly overhauled register allocator. No end-user changes, but the JIT
810 generates code a bit more quickly now.
812 * ================== PLATFORM CHANGES =================
814 * Preliminary support for macOS 10.13 has been added.
816 * mips: support for MIPS32/MIPS64 Revision 6 has been added.
818 * mips: support for MIPS SIMD architecture (MSA) has been added.
820 * mips: support for MIPS N32 ABI has been added.
822 * s390: partial support for vector instructions (integer and string) has been
825 * ==================== TOOL CHANGES ====================
827 * Helgrind: Addition of a flag
828 --delta-stacktrace=no|yes [yes on linux amd64/x86]
829 which specifies how full history stack traces should be computed.
830 Setting this to =yes can speed up Helgrind by 25% when using
831 --history-level=full.
833 * Memcheck: reduced false positive rate for optimised code created by Clang 6
834 / LLVM 6 on x86, amd64 and arm64. In particular, Memcheck analyses code
835 blocks more carefully to determine where it can avoid expensive definedness
836 checks without loss of precision. This is controlled by the flag
837 --expensive-definedness-checks=no|auto|yes [auto].
839 * ==================== OTHER CHANGES ====================
841 * Valgrind is now buildable with link-time optimisation (LTO). A new
842 configure option --enable-lto=yes allows building Valgrind with LTO. If the
843 toolchain supports it, this produces a smaller/faster Valgrind (up to 10%).
844 Note that if you are doing Valgrind development, --enable-lto=yes massively
845 slows down the build process.
847 * ==================== FIXED BUGS ====================
849 The following bugs have been fixed or resolved. Note that "n-i-bz"
850 stands for "not in bugzilla" -- that is, a bug that was reported to us
851 but never got a bugzilla entry. We encourage you to file bugs in
852 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
853 than mailing the developers (or mailing lists) directly -- bugs that
854 are not entered into bugzilla tend to get forgotten about or ignored.
856 To see details of a given bug, visit
857 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
858 where XXXXXX is the bug number as listed below.
860 79362 Debug info is lost for .so files when they are dlclose'd
861 208052 strlcpy error when n = 0
862 255603 exp-sgcheck Assertion '!already_present' failed
863 338252 building valgrind with -flto (link time optimisation) fails
864 345763 MIPS N32 ABI support
865 368913 WARNING: unhandled arm64-linux syscall: 117 (ptrace)
866 == 388664 unhandled arm64-linux syscall: 117 (ptrace)
867 372347 Replacement problem of the additional c++14/c++17 new/delete operators
868 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
869 376257 helgrind history full speed up using a cached stack
870 379373 Fix syscall param msg->desc.port.name points to uninitialised byte(s)
872 379748 Fix missing pselect syscall (OS X 10.11)
873 379754 Fix missing syscall ulock_wait (OS X 10.12)
874 380397 s390x: __GI_strcspn() replacemenet needed
875 381162 possible array overrun in VEX register allocator
876 381272 ppc64 doesn't compile test_isa_2_06_partx.c without VSX support
877 381274 powerpc too chatty even with --sigill-diagnostics=no
878 381289 epoll_pwait can have a NULL sigmask
879 381553 VEX register allocator v3
880 381556 arm64: Handle feature registers access on 4.11 Linux kernel or later
881 381769 Use ucontext_t instead of struct ucontext
882 381805 arm32 needs ld.so index hardwire for new glibc security fixes
883 382256 gz compiler flag test doesn't work for gold
884 382407 vg_perf needs "--terse" command line option
885 382515 "Assertion 'di->have_dinfo' failed." on wine's dlls/mscoree/tests/[..]
886 382563 MIPS MSA ASE support
887 382998 xml-socket doesn't work
888 383275 massif: m_xarray.c:162 (ensureSpaceXA): Assertion '!xa->arr' failed
889 383723 Fix missing kevent_qos syscall (macOS 10.11)
890 == 385604 illegal hardware instruction (OpenCV cv::namedWindow)
891 384096 Mention AddrCheck at Memcheck's command line option [..]
892 384230 vex x86->IR: 0x67 0xE8 0xAB 0x68
893 == 384156 vex x86->IR: 0x67 0xE8 0x6B 0x6A
894 == 386115 vex x86->IR: 0x67 0xE8 0xD3 0x8B any program
895 == 388407 vex x86->IR: 0x67 0xE8 0xAB 0x29
896 == 394903 vex x86->IR: 0x67 0xE8 0x1B 0xDA
897 384337 performance improvements to VEX register allocator v2 and v3
898 384526 reduce number of spill insns generated by VEX register allocator v3
899 384584 Callee saved regs listed first for AMD64, X86, and PPC architectures
900 384631 Sanitise client args as printed with -v
901 384633 Add a simple progress-reporting facility
902 384987 VEX regalloc: allocate caller-save registers for short lived vregs
903 385055 PPC VEX temporary storage exhausted
904 385182 PPC64 is missing support for the DSCR
905 385183 PPC64, Add support for xscmpeqdp, xscmpgtdp, xscmpgedp, xsmincdp
906 385207 PPC64, generate_store_FPRF() generates too many Iops
907 385208 PPC64, xxperm instruction exhausts temporary memory
908 385210 PPC64, vpermr instruction could exhaust temporary memory
909 385279 unhandled syscall: mach:43 (mach_generate_activity_id)
910 == 395136 valgrind: m_syswrap/syswrap-main.c:438 (Bool eq_Syscall[..]
911 == 387045 Valgrind crashing on High Sierra when testing any newly [..]
912 385334 PPC64, fix vpermr, xxperm, xxpermr mask value.
913 385408 s390x: z13 vector "support" instructions not implemented
914 385409 s390x: z13 vector integer instructions not implemented
915 385410 s390x: z13 vector string instructions not implemented
916 385412 s390x: new non-vector z13 instructions not implemented
917 385868 glibc ld.so _dl_runtime_resolve_avx_slow conditional jump warning.
918 385912 none/tests/rlimit_nofile fails on newer glibc/kernel.
919 385939 Optionally exit on the first error
920 386318 valgrind.org/info/tools.html is missing SGCheck
921 386425 running valgrind + wine on armv7l gives illegal opcode
922 386397 PPC64, valgrind truncates powerpc timebase to 32-bits.
923 387410 MIPSr6 support
924 387664 Memcheck: make expensive-definedness-checks be the default
925 387712 s390x cgijnl reports Conditional jump depends on uninitialised value
926 387766 asm shifts cause false positive "Conditional jump or move depends
927 on uninitialised value"
928 387773 .gnu_debugaltlink paths resolve relative to .debug file, not symlink
929 388174 valgrind with Wine quits with "Assertion 'cfsi_fits' failed"
930 388786 Support bpf syscall in amd64 Linux
931 388862 Add replacements for wmemchr and wcsnlen on Linux
932 389065 valgrind meets gcc flag -Wlogical-op
933 389373 exp-sgcheck the 'impossible' happened as Ist_LoadG is not instrumented
934 390471 suppression by specification of source-file line number
935 390723 make xtree dump files world wide readable, similar to log files
936 391164 constraint bug in tests/ppc64/test_isa_2_07_part1.c for mtfprwa
937 391861 Massif Assertion 'n_ips >= 1 && n_ips <= VG_(clo_backtrace_size)'
938 392118 unhandled amd64-linux syscall: 332 (statx)
939 392449 callgrind not clearing the number of calls properly
940 393017 Add missing support for xsmaxcdp instruction, bug fixes for xsmincdp,
941 lxssp, stxssp and stxvl instructions.
942 393023 callgrind_control risks using the wrong vgdb
943 393062 build-id ELF phdrs read causes "debuginfo reader: ensure_valid failed"
944 393099 posix_memalign() invalid write if alignment == 0
945 393146 failing assert "is_DebugInfo_active(di)"
946 395709 PPC64 is missing support for the xvnegsp instruction
947 395682 Accept read-only PT_LOAD segments and .rodata by ld -z separate-code
949 396475 valgrind OS-X build: config.h not found (out-of-tree macOS builds)
950 395991 arm-linux: wine's unit tests enter a signal delivery loop [..]
951 396839 s390x: Trap instructions not implemented
952 396887 arch_prctl should return EINVAL on unknown option
953 == 397286 crash before launching binary (Unsupported arch_prctl option)
954 == 397393 valgrind: the 'impossible' happened: (Archlinux)
955 == 397521 valgrind: the 'impossible' happened: Unsupported [..]
956 396906 compile tests failure on mips32-linux: broken inline asm in tests on
958 397012 glibc ld.so uses arch_prctl on i386
959 397089 amd64: Incorrect decoding of three-register vmovss/vmovsd opcode 11h
960 397354 utimensat should ignore timespec tv_sec if tv_nsec is UTIME_NOW/OMIT
961 397424 glibc 2.27 and gdb_server tests
962 398028 Assertion `cfsi_fits` failing in simple C program
963 398066 s390x: cgijl dep1, 0 reports false unitialised values warning
965 n-i-bz Fix missing workq_ops operations (macOS)
966 n-i-bz fix bug in strspn replacement
967 n-i-bz Add support for the Linux BLKFLSBUF ioctl
968 n-i-bz Add support for the Linux BLKREPORTZONE and BLKRESETZONE ioctls
969 n-i-bz Fix possible stack trashing by semctl syscall wrapping
970 n-i-bz Add support for the Linux membarrier() system call
971 n-i-bz x86 front end: recognise and handle UD2 correctly
972 n-i-bz Signal delivery for x86-linux: ensure that the stack pointer is
973 correctly aligned before entering the handler.
975 (3.14.0.RC1: 30 September 2018, git c2aeea2d28acb0639bcc8cc1e4ab115067db1eae)
976 (3.14.0.RC2: 3 October 2018, git 3e214c4858a6fdd5697e767543a0c19e30505582)
977 (3.14.0: 9 October 2018, git 353a3587bb0e2757411f9138f5e936728ed6cc4f)
981 Release 3.13.0 (15 June 2017)
982 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
984 3.13.0 is a feature release with many improvements and the usual collection of
987 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
988 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
989 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
990 X86/Solaris, AMD64/Solaris and AMD64/MacOSX 10.12.
992 * ==================== CORE CHANGES ===================
994 * The translation cache size has been increased to keep up with the demands of
995 large applications. The maximum number of sectors has increased from 24 to
996 48. The default number of sectors has increased from 16 to 32 on all
997 targets except Android, where the increase is from 6 to 12.
999 * The amount of memory that Valgrind can use has been increased from 64GB to
1000 128GB. In particular this means your application can allocate up to about
1001 60GB when running on Memcheck.
1003 * Valgrind's default load address has been changed from 0x3800'0000 to
1004 0x5800'0000, so as to make it possible to load larger executables. This
1005 should make it possible to load executables of size at least 1200MB.
1007 * A massive spaceleak caused by reading compressed debuginfo files has been
1008 fixed. Valgrind should now be entirely usable with gcc-7.0 "-gz" created
1011 * The C++ demangler has been updated.
1013 * Support for demangling Rust symbols has been added.
1015 * A new representation of stack traces, the "XTree", has been added. An XTree
1016 is a tree of stacktraces with data associated with the stacktraces. This is
1017 used by various tools (Memcheck, Helgrind, Massif) to report on the heap
1018 consumption of your program. Reporting is controlled by the new options
1019 --xtree-memory=none|allocs|full and --xtree-memory-file=<file>.
1021 A report can also be produced on demand using the gdbserver monitor command
1022 'xtmemory [<filename>]>'. The XTree can be output in 2 formats: 'callgrind
1023 format' and 'massif format. The existing visualisers for these formats (e.g.
1024 callgrind_annotate, KCachegrind, ms_print) can be used to visualise and
1025 analyse these reports.
1027 Memcheck can also produce XTree leak reports using the Callgrind file
1028 format. For more details, see the user manual.
1030 * ================== PLATFORM CHANGES =================
1032 * ppc64: support for ISA 3.0B and various fixes for existing 3.0 support
1034 * amd64: fixes for JIT failure problems on long AVX2 code blocks
1036 * amd64 and x86: support for CET prefixes has been added
1038 * arm32: a few missing ARMv8 instructions have been implemented
1040 * arm64, mips64, mips32: an alternative implementation of Load-Linked and
1041 Store-Conditional instructions has been added. This is to deal with
1042 processor implementations that implement the LL/SC specifications strictly
1043 and as a result cause Valgrind to hang in certain situations. The
1044 alternative implementation is automatically enabled at startup, as required.
1045 You can use the option --sim-hints=fallback-llsc to force-enable it if you
1048 * Support for OSX 10.12 has been improved.
1050 * On Linux, clone handling has been improved to honour CLONE_VFORK that
1051 involves a child stack. Note however that CLONE_VFORK | CLONE_VM is handled
1052 like CLONE_VFORK (by removing CLONE_VM), so applications that depend on
1053 CLONE_VM exact semantics will (still) not work.
1055 * The TileGX/Linux port has been removed because it appears to be both unused
1058 * ==================== TOOL CHANGES ====================
1062 - Memcheck should give fewer false positives when running optimised
1063 Clang/LLVM generated code.
1065 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1067 - New command line options --xtree-leak=no|yes and --xtree-leak-file=<file>
1068 to produce the end of execution leak report in a xtree callgrind format
1071 - New option 'xtleak' in the memcheck leak_check monitor command, to produce
1072 the leak report in an xtree file.
1076 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1078 - For some workloads (typically, for big applications), Massif memory
1079 consumption and CPU consumption has decreased significantly.
1083 - Support for --xtree-memory and 'xtmemory [<filename>]>'.
1085 - addition of client request VALGRIND_HG_GNAT_DEPENDENT_MASTER_JOIN, useful
1086 for Ada gnat compiled applications.
1088 * ==================== OTHER CHANGES ====================
1090 * For Valgrind developers: in an outer/inner setup, the outer Valgrind will
1091 append the inner guest stacktrace to the inner host stacktrace. This helps
1092 to investigate the errors reported by the outer, when they are caused by the
1093 inner guest program (such as an inner regtest). See README_DEVELOPERS for
1096 * To allow fast detection of callgrind files by desktop environments and file
1097 managers, the format was extended to have an optional first line that
1098 uniquely identifies the format ("# callgrind format"). Callgrind creates
1099 this line now, as does the new xtree functionality.
1101 * File name template arguments (such as --log-file, --xtree-memory-file, ...)
1102 have a new %n format letter that is replaced by a sequence number.
1104 * "--version -v" now shows the SVN revision numbers from which Valgrind was
1107 * ==================== FIXED BUGS ====================
1109 The following bugs have been fixed or resolved. Note that "n-i-bz"
1110 stands for "not in bugzilla" -- that is, a bug that was reported to us
1111 but never got a bugzilla entry. We encourage you to file bugs in
1112 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1113 than mailing the developers (or mailing lists) directly -- bugs that
1114 are not entered into bugzilla tend to get forgotten about or ignored.
1116 To see details of a given bug, visit
1117 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1118 where XXXXXX is the bug number as listed below.
1120 162848 --log-file output isn't split when a program forks
1121 340777 Illegal instruction on mips (ar71xx)
1122 341481 MIPS64: Iop_CmpNE32 triggers false warning on MIPS64 platforms
1123 342040 Valgrind mishandles clone with CLONE_VFORK | CLONE_VM that clones
1124 to a different stack.
1125 344139 x86 stack-seg overrides, needed by the Wine people
1126 344524 store conditional of guest applications always fail - observed on
1128 348616 Wine/valgrind: noted but unhandled ioctl 0x5390 [..] (DVD_READ_STRUCT)
1129 352395 Please provide SVN revision info in --version -v
1130 352767 Wine/valgrind: noted but unhandled ioctl 0x5307 [..] (CDROMSTOP)
1131 356374 Assertion 'DRD_(g_threadinfo)[tid].pt_threadid !=
1132 INVALID_POSIX_THREADID' failed
1133 358213 helgrind/drd bar_bad testcase hangs or crashes with new glibc pthread
1134 barrier implementation
1135 358697 valgrind.h: Some code remains even when defining NVALGRIND
1136 359202 Add musl libc configure/compile
1137 360415 amd64 instructions ADCX and ADOX are not implemented in VEX
1138 == 372828 (vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10)
1139 360429 unhandled ioctl 0x530d with no size/direction hints (CDROMREADMODE1)
1140 362223 assertion failed when .valgrindrc is a directory instead of a file
1141 367543 bt/btc/btr/bts x86/x86_64 instructions are poorly-handled wrt flags
1142 367942 Segfault vgPlain_do_sys_sigaction (m_signals.c:1138)
1143 368507 can't malloc chunks larger than about 34GB
1144 368529 Android arm target link error, missing atexit and pthread_atfork
1145 368863 WARNING: unhandled arm64-linux syscall: 100 (get_robust_list)
1146 368865 WARNING: unhandled arm64-linux syscall: 272 (kcmp)
1147 368868 disInstr(arm64): unhandled instruction 0xD53BE000 = cntfrq_el0 (ARMv8)
1148 368917 WARNING: unhandled arm64-linux syscall: 218 (request_key)
1149 368918 WARNING: unhandled arm64-linux syscall: 127 (sched_rr_get_interval)
1150 368922 WARNING: unhandled arm64-linux syscall: 161 (sethostname)
1151 368924 WARNING: unhandled arm64-linux syscall: 84 (sync_file_range)
1152 368925 WARNING: unhandled arm64-linux syscall: 130 (tkill)
1153 368926 WARNING: unhandled arm64-linux syscall: 97 (unshare)
1154 369459 valgrind on arm64 violates the ARMv8 spec (ldxr/stxr)
1155 370028 Reduce the number of compiler warnings on MIPS platforms
1156 370635 arm64 missing syscall getcpu
1157 371225 Fix order of timer_{gettime,getoverrun,settime} syscalls on arm64
1158 371227 Clean AArch64 syscall table
1159 371412 Rename wrap_sys_shmat to sys_shmat like other wrappers
1160 371471 Valgrind complains about non legit memory leaks on placement new (C++)
1161 371491 handleAddrOverrides() is [incorrect] when ASO prefix is used
1162 371503 disInstr(arm64): unhandled instruction 0xF89F0000
1163 371869 support '%' in symbol Z-encoding
1164 371916 execution tree xtree concept
1165 372120 c++ demangler demangles symbols which are not c++
1166 372185 Support of valgrind on ARMv8 with 32 bit executable
1167 372188 vex amd64->IR: 0x66 0xF 0x3A 0x62 0x4A 0x10 0x10 0x48 (PCMPxSTRx $0x10)
1168 372195 Power PC, xxsel instruction is not always recognized.
1169 372504 Hanging on exit_group
1170 372600 process loops forever when fatal signals are arriving quickly
1171 372794 LibVEX (arm32 front end): 'Assertion szBlg2 <= 3' failed
1172 373046 Stacks registered by core are never deregistered
1173 373069 memcheck/tests/leak_cpp_interior fails with GCC 5.1+
1174 373086 Implement additional Xen hypercalls
1175 373192 Calling posix_spawn in glibc 2.24 completely broken
1176 373488 Support for fanotify API on ARM64 architecture
1177 == 368864 WARNING: unhandled arm64-linux syscall: 262 (fanotify_init)
1178 373555 Rename BBPTR to GSPTR as it denotes guest state pointer only
1179 373938 const IRExpr arguments for matchIRExpr()
1180 374719 some spelling fixes
1181 374963 increase valgrind's load address to prevent mmap failure
1182 375514 valgrind_get_tls_addr() does not work in case of static TLS
1183 375772 +1 error in get_elf_symbol_info() when computing value of 'hi' address
1184 for ML_(find_rx_mapping)()
1185 375806 Test helgrind/tests/tc22_exit_w_lock fails with glibc 2.24
1186 375839 Temporary storage exhausted, with long sequence of vfmadd231ps insns
1187 == 377159 "vex: the `impossible' happened" still present
1188 == 375150 Assertion 'tres.status == VexTransOK' failed
1189 == 378068 valgrind crashes on AVX2 function in FFmpeg
1190 376142 Segfaults on MIPS Cavium Octeon boards
1191 376279 disInstr(arm64): unhandled instruction 0xD50320FF
1192 376455 Solaris: unhandled syscall lgrpsys(180)
1193 376518 Solaris: unhandled fast trap getlgrp(6)
1194 376611 ppc64 and arm64 don't know about prlimit64 syscall
1195 376729 PPC64, remove R2 from the clobber list
1197 376956 syswrap of SNDDRV and DRM_IOCTL_VERSION causing some addresses
1198 to be wrongly marked as addressable
1199 377066 Some Valgrind unit tests fail to compile on Ubuntu 16.10 with
1200 PIE enabled by default
1201 377376 memcheck/tests/linux/getregset fails with glibc2.24
1202 377427 PPC64, lxv instruction failing on odd destination register
1203 377478 PPC64: ISA 3.0 setup fixes
1204 377698 Missing memory check for futex() uaddr arg for FUTEX_WAKE
1205 and FUTEX_WAKE_BITSET, check only 4 args for FUTEX_WAKE_BITSET,
1206 and 2 args for FUTEX_TRYLOCK_PI
1207 377717 Fix massive space leak when reading compressed debuginfo sections
1208 377891 Update Xen 4.6 domctl wrappers
1209 377930 fcntl syscall wrapper is missing flock structure check
1210 378524 libvexmultiarch_test regression on s390x and ppc64
1211 378535 Valgrind reports INTERNAL ERROR in execve syscall wrapper
1212 378673 Update libiberty demangler
1213 378931 Add ISA 3.0B additional isnstructions, add OV32, CA32 setting support
1214 379039 syscall wrapper for prctl(PR_SET_NAME) must not check more than 16 bytes
1215 379094 Valgrind reports INTERNAL ERROR in rt_sigsuspend syscall wrapper
1216 379371 UNKNOWN task message [id 3444, to mach_task_self(), reply 0x603]
1217 (task_register_dyld_image_infos)
1218 379372 UNKNOWN task message [id 3447, to mach_task_self(), reply 0x603]
1219 (task_register_dyld_shared_cache_image_info)
1220 379390 unhandled syscall: mach:70 (host_create_mach_voucher_trap)
1221 379473 MIPS: add support for rdhwr cycle counter register
1222 379504 remove TileGX/Linux port
1223 379525 Support more x86 nop opcodes
1224 379838 disAMode(x86): not an addr!
1225 379703 PC ISA 3.0 fixes: stxvx, stxv, xscmpexpdp instructions
1226 379890 arm: unhandled instruction: 0xEBAD 0x1B05 (sub.w fp, sp, r5, lsl #4)
1227 379895 clock_gettime does not execute POST syscall wrapper
1228 379925 PPC64, mtffs does not set the FPCC and C bits in the FPSCR correctly
1229 379966 WARNING: unhandled amd64-linux syscall: 313 (finit_module)
1230 380200 xtree generated callgrind files refer to files without directory name
1231 380202 Assertion failure for cache line size (cls == 64) on aarch64.
1232 380397 s390x: __GI_strcspn() replacement needed
1233 n-i-bz Fix pub_tool_basics.h build issue with g++ 4.4.7.
1235 (3.13.0.RC1: 2 June 2017, vex r3386, valgrind r16434)
1236 (3.13.0.RC2: 9 June 2017, vex r3389, valgrind r16443)
1237 (3.13.0: 14 June 2017, vex r3396, valgrind r16446)
1241 Release 3.12.0 (20 October 2016)
1242 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1244 3.12.0 is a feature release with many improvements and the usual
1245 collection of bug fixes.
1247 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1248 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1249 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1250 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1251 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1252 X86/MacOSX 10.11/12, AMD64/MacOSX 10.11/12 and TILEGX/Linux.
1254 * ================== PLATFORM CHANGES =================
1256 * POWER: Support for ISA 3.0 has been added
1258 * mips: support for O32 FPXX ABI has been added.
1259 * mips: improved recognition of different processors
1260 * mips: determination of page size now done at run time
1262 * amd64: Partial support for AMD FMA4 instructions.
1264 * arm, arm64: Support for v8 crypto and CRC instructions.
1266 * Improvements and robustification of the Solaris port.
1268 * Preliminary support for MacOS 10.12 (Sierra) has been added.
1270 Whilst 3.12.0 continues to support the 32-bit x86 instruction set, we
1271 would prefer users to migrate to 64-bit x86 (a.k.a amd64 or x86_64)
1272 where possible. Valgrind's support for 32-bit x86 has stagnated in
1273 recent years and has fallen far behind that for 64-bit x86
1274 instructions. By contrast 64-bit x86 is well supported, up to and
1277 * ==================== TOOL CHANGES ====================
1281 - Added meta mempool support for describing a custom allocator which:
1282 - Auto-frees all chunks assuming that destroying a pool destroys all
1284 - Uses itself to allocate other memory blocks
1286 - New flag --ignore-range-below-sp to ignore memory accesses below
1287 the stack pointer, if you really have to. The related flag
1288 --workaround-gcc296-bugs=yes is now deprecated. Use
1289 --ignore-range-below-sp=1024-1 as a replacement.
1293 - Improved thread startup time significantly on non-Linux platforms.
1297 - Added collection of the metric "tot-blocks-allocd"
1299 * ==================== OTHER CHANGES ====================
1301 * Replacement/wrapping of malloc/new related functions is now done not just
1302 for system libraries by default, but for any globally defined malloc/new
1303 related function (both in shared libraries and statically linked alternative
1304 malloc implementations). The dynamic (runtime) linker is excluded, though.
1305 To only intercept malloc/new related functions in
1306 system libraries use --soname-synonyms=somalloc=nouserintercepts (where
1307 "nouserintercepts" can be any non-existing library name).
1308 This new functionality is not implemented for MacOS X.
1310 * The maximum number of callers in a suppression entry is now equal to
1311 the maximum size for --num-callers (500).
1312 Note that --gen-suppressions=yes|all similarly generates suppressions
1313 containing up to --num-callers frames.
1315 * New and modified GDB server monitor features:
1317 - Valgrind's gdbserver now accepts the command 'catch syscall'.
1318 Note that you must have GDB >= 7.11 to use 'catch syscall' with
1321 * New option --run-cxx-freeres=<yes|no> can be used to change whether
1322 __gnu_cxx::__freeres() cleanup function is called or not. Default is
1325 * Valgrind is able to read compressed debuginfo sections in two formats:
1326 - zlib ELF gABI format with SHF_COMPRESSED flag (gcc option -gz=zlib)
1327 - zlib GNU format with .zdebug sections (gcc option -gz=zlib-gnu)
1329 * Modest JIT-cost improvements: the cost of instrumenting code blocks
1330 for the most common use case (x86_64-linux, Memcheck) has been
1333 * Improved performance for programs that do a lot of discarding of
1334 instruction address ranges of 8KB or less.
1336 * The C++ symbol demangler has been updated.
1338 * More robustness against invalid syscall parameters on Linux.
1340 * ==================== FIXED BUGS ====================
1342 The following bugs have been fixed or resolved. Note that "n-i-bz"
1343 stands for "not in bugzilla" -- that is, a bug that was reported to us
1344 but never got a bugzilla entry. We encourage you to file bugs in
1345 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1346 than mailing the developers (or mailing lists) directly -- bugs that
1347 are not entered into bugzilla tend to get forgotten about or ignored.
1349 To see details of a given bug, visit
1350 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1351 where XXXXXX is the bug number as listed below.
1353 191069 Exiting due to signal not reported in XML output
1354 199468 Suppressions: stack size limited to 25
1355 while --num-callers allows more frames
1356 212352 vex amd64 unhandled opc_aux = 0x 2, first_opcode == 0xDC (FCOM)
1357 278744 cvtps2pd with redundant RexW
1358 303877 valgrind doesn't support compressed debuginfo sections.
1359 345307 Warning about "still reachable" memory when using libstdc++ from gcc 5
1360 348345 Assertion fails for negative lineno
1361 348924 MIPS: Load doubles through memory so the code compiles with the FPXX ABI
1362 351282 V 3.10.1 MIPS softfloat build broken with GCC 4.9.3 / binutils 2.25.1
1363 351692 Dumps created by valgrind are not readable by gdb (mips32 specific)
1364 351804 Crash on generating suppressions for "printf" call on OS X 10.10
1365 352197 mips: mmap2() not wrapped correctly for page size > 4096
1366 353083 arm64 doesn't implement various xattr system calls
1367 353084 arm64 doesn't support sigpending system call
1368 353137 www: update info for Supported Platforms
1369 353138 www: update "The Valgrind Developers" page
1370 353370 don't advertise RDRAND in cpuid for Core-i7-4910-like avx2 machine
1373 353384 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x62 (pcmpXstrX $0x62)
1374 353398 WARNING: unhandled amd64-solaris syscall: 207
1375 353660 XML in auxwhat tag not escaping reserved symbols properly
1376 353680 s390x: Crash with certain glibc versions due to non-implemented TBEGIN
1377 353727 amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x72 (pcmpXstrX $0x72)
1378 353802 ELF debug info reader confused with multiple .rodata sections
1379 353891 Assert 'bad_scanned_addr < VG_ROUNDDN(start+len, sizeof(Addr))' failed
1380 353917 unhandled amd64-solaris syscall fchdir(120)
1381 353920 unhandled amd64-solaris syscall: 170
1382 354274 arm: unhandled instruction: 0xEBAD 0x0AC1 (sub.w sl, sp, r1, lsl #3)
1383 354392 unhandled amd64-solaris syscall: 171
1384 354797 Vbit test does not include Iops for Power 8 instruction support
1385 354883 tst->os_state.pthread - magic_delta assertion failure on OSX 10.11
1389 354933 Fix documentation of --kernel-variant=android-no-hw-tls option
1390 355188 valgrind should intercept all malloc related global functions
1391 355454 do not intercept malloc related symbols from the runtime linker
1392 355455 stderr.exp of test cases wrapmalloc and wrapmallocstatic overconstrained
1393 356044 Dwarf line info reader misinterprets is_stmt register
1394 356112 mips: replace addi with addiu
1395 356393 valgrind (vex) crashes because isZeroU happened
1398 356676 arm64-linux: unhandled syscalls 125, 126 (sched_get_priority_max/min)
1399 356678 arm64-linux: unhandled syscall 232 (mincore)
1400 356817 valgrind.h triggers compiler errors on MSVC when defining NVALGRIND
1401 356823 Unsupported ARM instruction: stlex
1402 357059 x86/amd64: SSE cvtpi2ps with memory source does transition to MMX state
1403 357338 Unhandled instruction for SHA instructions libcrypto Boring SSL
1404 357673 crash if I try to run valgrind with a binary link with libcurl
1405 357833 Setting RLIMIT_DATA to zero breaks with linux 4.5+
1406 357871 pthread_spin_destroy not properly wrapped
1407 357887 Calls to VG_(fclose) do not close the file descriptor
1408 357932 amd64->IR: accept redundant REX prefixes for {minsd,maxsd} m128, xmm.
1409 358030 support direct socket calls on x86 32bit (new in linux 4.3)
1410 358478 drd/tests/std_thread.cpp doesn't build with GCC6
1411 359133 Assertion 'eltSzB <= ddpa->poolSzB' failed
1412 359181 Buffer Overflow during Demangling
1413 359201 futex syscall "skips" argument 5 if op is FUTEX_WAIT_BITSET
1414 359289 s390x: popcnt (B9E1) not implemented
1415 359472 The Power PC vsubuqm instruction doesn't always give the correct result
1416 359503 Add missing syscalls for aarch64 (arm64)
1417 359645 "You need libc6-dbg" help message could be more helpful
1418 359703 s390: wire up separate socketcalls system calls
1419 359724 getsockname might crash - deref_UInt should call safe_to_deref
1420 359733 amd64 implement ld.so strchr/index override like x86
1421 359767 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 1/5
1422 359829 Power PC test suite none/tests/ppc64/test_isa_2_07.c uses
1424 359838 arm64: Unhandled instruction 0xD5033F5F (clrex)
1425 359871 Incorrect mask handling in ppoll
1426 359952 Unrecognised PCMPESTRM variants (0x70, 0x19)
1427 360008 Contents of Power vr registers contents is not printed correctly when
1428 the --vgdb-shadow-registers=yes option is used
1429 360035 POWER PC instruction bcdadd and bcdsubtract generate result with
1430 non-zero shadow bits
1431 360378 arm64: Unhandled instruction 0x5E280844 (sha1h s4, s2)
1432 360425 arm64 unsupported instruction ldpsw
1434 360519 none/tests/arm64/memory.vgtest might fail with newer gcc
1435 360571 Error about the Android Runtime reading below the stack pointer on ARM
1436 360574 Wrong parameter type for an ashmem ioctl() call on Android and ARM64
1437 360749 kludge for multiple .rodata sections on Solaris no longer needed
1438 360752 raise the number of reserved fds in m_main.c from 10 to 12
1439 361207 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 2/5
1440 361226 s390x: risbgn (EC59) not implemented
1441 361253 [s390x] ex_clone.c:42: undefined reference to `pthread_create'
1442 361354 ppc64[le]: wire up separate socketcalls system calls
1443 361615 Inconsistent termination for multithreaded process terminated by signal
1444 361926 Unhandled Solaris syscall: sysfs(84)
1445 362009 V dumps core on unimplemented functionality before threads are created
1446 362329 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 3/5
1447 362894 missing (broken) support for wbit field on mtfsfi instruction (ppc64)
1448 362935 [AsusWRT] Assertion 'sizeof(TTEntryC) <= 88' failed
1449 362953 Request for an update to the Valgrind Developers page
1450 363680 add renameat2() support
1451 363705 arm64 missing syscall name_to_handle_at and open_by_handle_at
1452 363714 ppc64 missing syscalls sync, waitid and name_to/open_by_handle_at
1453 363858 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 4/5
1454 364058 clarify in manual limitations of array overruns detections
1455 364413 pselect sycallwrapper mishandles NULL sigmask
1456 364728 Power PC, missing support for several HW registers in
1457 get_otrack_shadow_offset_wrk()
1458 364948 Valgrind does not support the IBM POWER ISA 3.0 instructions, part 5/5
1459 365273 Invalid write to stack location reported after signal handler runs
1460 365912 ppc64BE segfault during jm-insns test (RELRO)
1461 366079 FPXX Support for MIPS32 Valgrind
1462 366138 Fix configure errors out when using Xcode 8 (clang 8.0.0)
1463 366344 Multiple unhandled instruction for Aarch64
1464 (0x0EE0E020, 0x1AC15800, 0x4E284801, 0x5E040023, 0x5E056060)
1465 367995 Integration of memcheck with custom memory allocator
1466 368120 x86_linux asm _start functions do not keep 16-byte aligned stack pointer
1467 368412 False positive result for altivec capability check
1468 368416 Add tc06_two_races_xml.exp output for ppc64
1469 368419 Perf Events ioctls not implemented
1470 368461 mmapunmap test fails on ppc64
1471 368823 run_a_thread_NORETURN assembly code typo for VGP_arm64_linux target
1472 369000 AMD64 fma4 instructions unsupported.
1473 369169 ppc64 fails jm_int_isa_2_07 test
1474 369175 jm_vec_isa_2_07 test crashes on ppc64
1475 369209 valgrind loops and eats up all memory if cwd doesn't exist.
1476 369356 pre_mem_read_sockaddr syscall wrapper can crash with bad sockaddr
1477 369359 msghdr_foreachfield can crash when handling bad iovec
1478 369360 Bad sigprocmask old or new sets can crash valgrind
1479 369361 vmsplice syscall wrapper crashes on bad iovec
1480 369362 Bad sigaction arguments crash valgrind
1481 369383 x86 sys_modify_ldt wrapper crashes on bad ptr
1482 369402 Bad set/get_thread_area pointer crashes valgrind
1483 369441 bad lvec argument crashes process_vm_readv/writev syscall wrappers
1484 369446 valgrind crashes on unknown fcntl command
1485 369439 S390x: Unhandled insns RISBLG/RISBHG and LDE/LDER
1486 369468 Remove quadratic metapool algorithm using VG_(HT_remove_at_Iter)
1487 370265 ISA 3.0 HW cap stuff needs updating
1488 371128 BCD add and subtract instructions on Power BE in 32-bit mode do not work
1489 372195 Power PC, xxsel instruction is not always recognized
1491 n-i-bz Fix incorrect (or infinite loop) unwind on RHEL7 x86 and amd64
1492 n-i-bz massif --pages-as-heap=yes does not report peak caused by mmap+munmap
1493 n-i-bz false positive leaks due to aspacemgr merging heap & non heap segments
1494 n-i-bz Fix ppoll_alarm exclusion on OS X
1495 n-i-bz Document brk segment limitation, reference manual in limit reached msg.
1496 n-i-bz Fix clobber list in none/tests/amd64/xacq_xrel.c [valgrind r15737]
1497 n-i-bz Bump allowed shift value for "add.w reg, sp, reg, lsl #N" [vex r3206]
1498 n-i-bz amd64: memcheck false positive with shr %edx
1499 n-i-bz arm3: Allow early writeback of SP base register in "strd rD, [sp, #-16]"
1500 n-i-bz ppc: Fix two cases of PPCAvFpOp vs PPCFpOp enum confusion
1501 n-i-bz arm: Fix incorrect register-number constraint check for LDAEX{,B,H,D}
1502 n-i-bz DHAT: added collection of the metric "tot-blocks-allocd"
1504 (3.12.0.RC1: 20 October 2016, vex r3282, valgrind r16094)
1505 (3.12.0.RC2: 20 October 2016, vex r3282, valgrind r16096)
1506 (3.12.0: 21 October 2016, vex r3282, valgrind r16098)
1510 Release 3.11.0 (22 September 2015)
1511 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1513 3.11.0 is a feature release with many improvements and the usual
1514 collection of bug fixes.
1516 This release supports X86/Linux, AMD64/Linux, ARM32/Linux,
1517 ARM64/Linux, PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux,
1518 MIPS32/Linux, MIPS64/Linux, ARM/Android, ARM64/Android,
1519 MIPS32/Android, X86/Android, X86/Solaris, AMD64/Solaris, X86/MacOSX
1520 10.10 and AMD64/MacOSX 10.10. There is also preliminary support for
1521 X86/MacOSX 10.11, AMD64/MacOSX 10.11 and TILEGX/Linux.
1523 * ================== PLATFORM CHANGES =================
1525 * Support for Solaris/x86 and Solaris/amd64 has been added.
1527 * Preliminary support for Mac OS X 10.11 (El Capitan) has been added.
1529 * Preliminary support for the Tilera TileGX architecture has been added.
1531 * s390x: It is now required for the host to have the "long displacement"
1532 facility. The oldest supported machine model is z990.
1534 * x86: on an SSE2 only host, Valgrind in 32 bit mode now claims to be a
1535 Pentium 4. 3.10.1 wrongly claimed to be a Core 2, which is SSSE3.
1537 * The JIT's register allocator is significantly faster, making the JIT
1538 as a whole somewhat faster, so JIT-intensive activities, for example
1539 program startup, are modestly faster, around 5%.
1541 * There have been changes to the default settings of several command
1542 line flags, as detailed below.
1544 * Intel AVX2 support is more complete (64 bit targets only). On AVX2
1545 capable hosts, the simulated CPUID will now indicate AVX2 support.
1547 * ==================== TOOL CHANGES ====================
1551 - The default value for --leak-check-heuristics has been changed from
1552 "none" to "all". This helps to reduce the number of possibly
1553 lost blocks, in particular for C++ applications.
1555 - The default value for --keep-stacktraces has been changed from
1556 "malloc-then-free" to "malloc-and-free". This has a small cost in
1557 memory (one word per malloc-ed block) but allows Memcheck to show the
1558 3 stacktraces of a dangling reference: where the block was allocated,
1559 where it was freed, and where it is acccessed after being freed.
1561 - The default value for --partial-loads-ok has been changed from "no" to
1562 "yes", so as to avoid false positive errors resulting from some kinds
1563 of vectorised loops.
1565 - A new monitor command 'xb <addr> <len>' shows the validity bits of
1566 <len> bytes at <addr>. The monitor command 'xb' is easier to use
1567 than get_vbits when you need to associate byte data value with
1568 their corresponding validity bits.
1570 - The 'block_list' monitor command has been enhanced:
1571 o it can print a range of loss records
1572 o it now accepts an optional argument 'limited <max_blocks>'
1573 to control the number of blocks printed.
1574 o if a block has been found using a heuristic, then
1575 'block_list' now shows the heuristic after the block size.
1576 o the loss records/blocks to print can be limited to the blocks
1577 found via specified heuristics.
1579 - The C helper functions used to instrument loads on
1580 x86-{linux,solaris} and arm-linux (both 32-bit only) have been
1581 replaced by handwritten assembly sequences. This gives speedups
1582 in the region of 0% to 7% for those targets only.
1584 - A new command line option, --expensive-definedness-checks=yes|no,
1585 has been added. This is useful for avoiding occasional invalid
1586 uninitialised-value errors in optimised code. Watch out for
1587 runtime degradation, as this can be up to 25%. As always, though,
1588 the slowdown is highly application specific. The default setting
1593 - A new monitor command 'all_snapshots <filename>' dumps all
1594 snapshots taken so far.
1598 - Significant memory reduction and moderate speedups for
1599 --history-level=full for applications accessing a lot of memory
1600 with many different stacktraces.
1602 - The default value for --conflict-cache-size=N has been doubled to
1603 2000000. Users that were not using the default value should
1604 preferably also double the value they give.
1606 The default was changed due to the changes in the "full history"
1607 implementation. Doubling the value gives on average a slightly more
1608 complete history and uses similar memory (or significantly less memory
1609 in the worst case) than the previous implementation.
1611 - The Helgrind monitor command 'info locks' now accepts an optional
1612 argument 'lock_addr', which shows information about the lock at the
1615 - When using --history-level=full, the new Helgrind monitor command
1616 'accesshistory <addr> [<len>]' will show the recorded accesses for
1617 <len> (or 1) bytes at <addr>.
1619 * ==================== OTHER CHANGES ====================
1621 * The default value for the --smc-check option has been changed from
1622 "stack" to "all-non-file" on targets that provide automatic D-I
1623 cache coherence (x86, amd64 and s390x). The result is to provide,
1624 by default, transparent support for JIT generated and self-modifying
1625 code on all targets.
1627 * Mac OS X only: the default value for the --dsymutil option has been
1628 changed from "no" to "yes", since any serious usage on Mac OS X
1629 always required it to be "yes".
1631 * The command line options --db-attach and --db-command have been removed.
1632 They were deprecated in 3.10.0.
1634 * When a process dies due to a signal, Valgrind now shows the signal
1635 and the stacktrace at default verbosity (i.e. verbosity 1).
1637 * The address description logic used by Memcheck and Helgrind now
1638 describes addresses in anonymous segments, file mmap-ed segments,
1639 shared memory segments and the brk data segment.
1641 * The new option --error-markers=<begin>,<end> can be used to mark the
1642 begin/end of errors in textual output mode, to facilitate
1643 searching/extracting errors in output files that mix valgrind errors
1644 with program output.
1646 * The new option --max-threads=<number> can be used to change the number
1647 of threads valgrind can handle. The default is 500 threads which
1648 should be more than enough for most applications.
1650 * The new option --valgrind-stacksize=<number> can be used to change the
1651 size of the private thread stacks used by Valgrind. This is useful
1652 for reducing memory use or increasing the stack size if Valgrind
1653 segfaults due to stack overflow.
1655 * The new option --avg-transtab-entry-size=<number> can be used to specify
1656 the expected instrumented block size, either to reduce memory use or
1657 to avoid excessive retranslation.
1659 * Valgrind can be built with Intel's ICC compiler, version 14.0 or later.
1661 * New and modified GDB server monitor features:
1663 - When a signal is reported in GDB, you can now use the GDB convenience
1664 variable $_siginfo to examine detailed signal information.
1666 - Valgrind's gdbserver now allows the user to change the signal
1667 to deliver to the process. So, use 'signal SIGNAL' to continue execution
1668 with SIGNAL instead of the signal reported to GDB. Use 'signal 0' to
1669 continue without passing the signal to the process.
1671 - With GDB >= 7.10, the command 'target remote'
1672 will automatically load the executable file of the process running
1673 under Valgrind. This means you do not need to specify the executable
1674 file yourself, GDB will discover it itself. See GDB documentation about
1675 'qXfer:exec-file:read' packet for more info.
1677 * ==================== FIXED BUGS ====================
1679 The following bugs have been fixed or resolved. Note that "n-i-bz"
1680 stands for "not in bugzilla" -- that is, a bug that was reported to us
1681 but never got a bugzilla entry. We encourage you to file bugs in
1682 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1683 than mailing the developers (or mailing lists) directly -- bugs that
1684 are not entered into bugzilla tend to get forgotten about or ignored.
1686 To see details of a given bug, visit
1687 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1688 where XXXXXX is the bug number as listed below.
1690 116002 VG_(printf): Problems with justification of strings and integers
1691 155125 avoid cutting away file:lineno after long function name
1692 197259 Unsupported arch_prtctl PR_SET_GS option
1693 201152 ppc64: Assertion in ppc32g_dirtyhelper_MFSPR_268_269
1694 201216 Fix Valgrind does not support pthread_sigmask() on OS X
1695 201435 Fix Darwin: -v does not show kernel version
1696 208217 "Warning: noted but unhandled ioctl 0x2000747b" on Mac OS X
1697 211256 Fixed an outdated comment regarding the default platform.
1698 211529 Incomplete call stacks for code compiled by newer versions of MSVC
1699 211926 Avoid compilation warnings in valgrind.h with -pedantic
1700 212291 Fix unhandled syscall: unix:132 (mkfifo) on OS X
1702 226609 Crediting upstream authors in man page
1703 231257 Valgrind omits path when executing script from shebang line
1704 254164 OS X task_info: UNKNOWN task message [id 3405, to mach_task_self() [..]
1705 294065 Improve the pdb file reader by avoiding hardwired absolute pathnames
1706 269360 s390x: Fix addressing mode selection for compare-and-swap
1707 302630 Memcheck: Assertion failed: 'sizeof(UWord) == sizeof(UInt)'
1709 312989 ioctl handling needs to do POST handling on generic ioctls and [..]
1710 319274 Fix unhandled syscall: unix:410 (sigsuspend_nocancel) on OS X
1711 324181 mmap does not handle MAP_32BIT (handle it now, rather than fail it)
1712 327745 Fix valgrind 3.9.0 build fails on Mac OS X 10.6.8
1713 330147 libmpiwrap PMPI_Get_count returns undefined value
1714 333051 mmap of huge pages fails due to incorrect alignment
1716 334802 valgrind does not always explain why a given option is bad
1717 335618 mov.w rN, pc/sp (ARM32)
1718 335785 amd64->IR 0xC4 0xE2 0x75 0x2F (vmaskmovpd)
1723 335907 segfault when running wine's ddrawex/tests/surface.c under valgrind
1724 338602 AVX2 bit in CPUID missing
1725 338606 Strange message for scripts with invalid interpreter
1726 338731 ppc: Fix testuite build for toolchains not supporting -maltivec
1727 338995 shmat with hugepages (SHM_HUGETLB) fails with EINVAL
1728 339045 Getting valgrind to compile and run on OS X Yosemite (10.10)
1730 339156 gdbsrv not called for fatal signal
1731 339215 Valgrind 3.10.0 contain 2013 in copyrights notice
1732 339288 support Cavium Octeon MIPS specific BBIT*32 instructions
1733 339636 Use fxsave64 and fxrstor64 mnemonics instead of old-school rex64 prefix
1734 339442 Fix testsuite build failure on OS X 10.9
1735 339542 Enable compilation with Intel's ICC compiler
1736 339563 The DVB demux DMX_STOP ioctl doesn't have a wrapper
1737 339688 Mac-specific ASM does not support .version directive (cpuid,
1738 tronical and pushfpopf tests)
1739 339745 Valgrind crash when check Marmalade app (partial fix)
1740 339755 Fix known deliberate memory leak in setenv() on Mac OS X 10.9
1741 339778 Linux/TileGx platform support for Valgrind
1742 339780 Fix known uninitialised read in pthread_rwlock_init() on Mac OS X 10.9
1743 339789 Fix none/tests/execve test on Mac OS X 10.9
1744 339808 Fix none/tests/rlimit64_nofile test on Mac OS X 10.9
1745 339820 vex amd64->IR: 0x66 0xF 0x3A 0x63 0xA 0x42 0x74 0x9 (pcmpistri $0x42)
1746 340115 Fix none/tests/cmdline[1|2] tests on systems which define TMPDIR
1747 340392 Allow user to select more accurate definedness checking in memcheck
1748 to avoid invalid complaints on optimised code
1749 340430 Fix some grammatical weirdness in the manual.
1750 341238 Recognize GCC5/DWARFv5 DW_LANG constants (Go, C11, C++11, C++14)
1751 341419 Signal handler ucontext_t not filled out correctly on OS X
1752 341539 VG_(describe_addr) should not describe address as belonging to client
1753 segment if it is past the heap end
1754 341613 Enable building of manythreads and thread-exits tests on Mac OS X
1755 341615 Fix none/tests/darwin/access_extended test on Mac OS X
1756 341698 Valgrind's AESKEYGENASSIST gives wrong result in words 0 and 2 [..]
1757 341789 aarch64: shmat fails with valgrind on ARMv8
1758 341997 MIPS64: Cavium OCTEON insns - immediate operand handled incorrectly
1759 342008 valgrind.h needs type cast [..] for clang/llvm in 64-bit mode
1760 342038 Unhandled syscalls on aarch64 (mbind/get/set_mempolicy)
1761 342063 wrong format specifier for test mcblocklistsearch in gdbserver_tests
1762 342117 Hang when loading PDB file for MSVC compiled Firefox under Wine
1763 342221 socket connect false positive uninit memory for unknown af family
1764 342353 Allow dumping full massif output while valgrind is still running
1765 342571 Valgrind chokes on AVX compare intrinsic with _CMP_GE_QS
1769 342603 Add I2C_SMBUS ioctl support
1770 342635 OS X 10.10 (Yosemite) - missing system calls and fcntl code
1771 342683 Mark memory past the initial brk limit as unaddressable
1772 342783 arm: unhandled instruction 0xEEFE1ACA = "vcvt.s32.f32 s3, s3, #12"
1773 342795 Internal glibc __GI_mempcpy call should be intercepted
1774 342841 s390x: Support instructions fiebr(a) and fidbr(a)
1775 343012 Unhandled syscall 319 (memfd_create)
1776 343069 Patch updating v4l2 API support
1777 343173 helgrind crash during stack unwind
1778 343219 fix GET_STARTREGS for arm
1779 343303 Fix known deliberate memory leak in setenv() on Mac OS X 10.10
1780 343306 OS X 10.10: UNKNOWN mach_msg unhandled MACH_SEND_TRAILER option
1781 343332 Unhandled instruction 0x9E310021 (fcvtmu) on aarch64
1782 343335 unhandled instruction 0x1E638400 (fccmp) aarch64
1783 343523 OS X mach_ports_register: UNKNOWN task message [id 3403, to [..]
1784 343525 OS X host_get_special_port: UNKNOWN host message [id 412, to [..]
1785 343597 ppc64le: incorrect use of offseof macro
1786 343649 OS X host_create_mach_voucher: UNKNOWN host message [id 222, to [..]
1787 343663 OS X 10.10 Memchecj always reports a leak regardless of [..]
1788 343732 Unhandled syscall 144 (setgid) on aarch64
1789 343733 Unhandled syscall 187 (msgctl and related) on aarch64
1790 343802 s390x: False positive "conditional jump or move depends on [..]
1791 343902 --vgdb=yes doesn't break when --xml=yes is used
1792 343967 Don't warn about setuid/setgid/setcap executable for directories
1793 343978 Recognize DWARF5/GCC5 DW_LANG_Fortran 2003 and 2008 constants
1794 344007 accept4 syscall unhandled on arm64 (242) and ppc64 (344)
1795 344033 Helgrind on ARM32 loses track of mutex state in pthread_cond_wait
1796 344054 www - update info for Solaris/illumos
1797 344416 'make regtest' does not work cleanly on OS X
1798 344235 Remove duplicate include of pub_core_aspacemgr.h
1799 344279 syscall sendmmsg on arm64 (269) and ppc32/64 (349) unhandled
1800 344295 syscall recvmmsg on arm64 (243) and ppc32/64 (343) unhandled
1801 344307 2 unhandled syscalls on aarch64/arm64: umount2(39), mount (40)
1802 344314 callgrind_annotate ... warnings about commands containing newlines
1803 344318 socketcall should wrap recvmmsg and sendmmsg
1804 344337 Fix unhandled syscall: mach:41 (_kernelrpc_mach_port_guard_trap)
1805 344416 Fix 'make regtest' does not work cleanly on OS X
1806 344499 Fix compilation for Linux kernel >= 4.0.0
1807 344512 OS X: unhandled syscall: unix:348 (__pthread_chdir),
1808 unix:349 (__pthread_fchdir)
1809 344559 Garbage collection of unused segment names in address space manager
1810 344560 Fix stack traces missing penultimate frame on OS X
1811 344621 Fix memcheck/tests/err_disable4 test on OS X
1812 344686 Fix suppression for pthread_rwlock_init on OS X 10.10
1813 344702 Fix missing libobjc suppressions on OS X 10.10
1815 344936 Fix unhandled syscall: unix:473 (readlinkat) on OS X 10.10
1816 344939 Fix memcheck/tests/xml1 on OS X 10.10
1817 345016 helgrind/tests/locked_vs_unlocked2 is failing sometimes
1818 345079 Fix build problems in VEX/useful/test_main.c
1819 345126 Incorrect handling of VIDIOC_G_AUDIO and G_AUDOUT
1820 345177 arm64: prfm (reg) not implemented
1821 345215 Performance improvements for the register allocator
1822 345248 add support for Solaris OS in valgrind
1823 345338 TIOCGSERIAL and TIOCSSERIAL ioctl support on Linux
1824 345394 Fix memcheck/tests/strchr on OS X
1825 345637 Fix memcheck/tests/sendmsg on OS X
1826 345695 Add POWERPC support for AT_DCACHESIZE and HWCAP2
1827 345824 Fix aspacem segment mismatch: seen with none/tests/bigcode
1828 345887 Fix an assertion in the address space manager
1829 345928 amd64: callstack only contains current function for small stacks
1830 345984 disInstr(arm): unhandled instruction: 0xEE193F1E
1831 345987 MIPS64: Implement cavium LHX instruction
1832 346031 MIPS: Implement support for the CvmCount register (rhwr %0, 31)
1833 346185 Fix typo saving altivec register v24
1834 346267 Compiler warnings for PPC64 code on call to LibVEX_GuestPPC64_get_XER()
1835 and LibVEX_GuestPPC64_get_CR()
1836 346270 Regression tests none/tests/jm_vec/isa_2_07 and
1837 none/tests/test_isa_2_07_part2 have failures on PPC64 little endian
1838 346307 fuse filesystem syscall deadlocks
1839 346324 PPC64 missing support for lbarx, lharx, stbcx and sthcx instructions
1840 346411 MIPS: SysRes::_valEx handling is incorrect
1841 346416 Add support for LL_IOC_PATH2FID and LL_IOC_GETPARENT Lustre ioctls
1842 346474 PPC64 Power 8, spr TEXASRU register not supported
1843 346487 Compiler generates "note" about a future ABI change for PPC64
1844 346562 MIPS64: lwl/lwr instructions are performing 64bit loads
1845 and causing spurious "invalid read of size 8" warnings
1846 346801 Fix link error on OS X: _vgModuleLocal_sf_maybe_extend_stack
1847 347151 Fix suppression for pthread_rwlock_init on OS X 10.8
1848 347233 Fix memcheck/tests/strchr on OS X 10.10 (Haswell)
1849 347322 Power PC regression test cleanup
1850 347379 valgrind --leak-check=full leak errors from system libs on OS X 10.8
1852 347389 unhandled syscall: 373 (Linux ARM syncfs)
1853 347686 Patch set to cleanup PPC64 regtests
1854 347978 Remove bash dependencies where not needed
1855 347982 OS X: undefined symbols for architecture x86_64: "_global" [..]
1856 347988 Memcheck: the 'impossible' happened: unexpected size for Addr (OSX/wine)
1858 348102 Patch updating v4l2 API support
1859 348247 amd64 front end: jno jumps wrongly when overflow is not set
1860 348269 Improve mmap MAP_HUGETLB support.
1861 348334 (ppc) valgrind does not simulate dcbfl - then my program terminates
1862 348345 Assertion fails for negative lineno
1863 348377 Unsupported ARM instruction: yield
1864 348565 Fix detection of command line option availability for clang
1865 348574 vex amd64->IR pcmpistri SSE4.2 unsupported (pcmpistri $0x18)
1866 348728 Fix broken check for VIDIOC_G_ENC_INDEX
1867 348748 Fix redundant condition
1868 348890 Fix clang warning about unsupported --param inline-unit-growth=900
1869 348949 Bogus "ERROR: --ignore-ranges: suspiciously large range"
1870 349034 Add Lustre ioctls LL_IOC_GROUP_LOCK and LL_IOC_GROUP_UNLOCK
1871 349086 Fix UNKNOWN task message [id 3406, to mach_task_self(), [..]
1872 349087 Fix UNKNOWN task message [id 3410, to mach_task_self(), [..]
1873 349626 Implemented additional Xen hypercalls
1874 349769 Clang/osx: ld: warning: -read_only_relocs cannot be used with x86_64
1875 349790 Clean up of the hardware capability checking utilities.
1876 349828 memcpy intercepts memmove causing src/dst overlap error (ppc64 ld.so)
1877 349874 Fix typos in source code
1878 349879 memcheck: add handwritten assembly for helperc_LOADV*
1879 349941 di_notify_mmap might create wrong start/size DebugInfoMapping
1880 350062 vex x86->IR: 0x66 0xF 0x3A 0xB (ROUNDSD) on OS X
1881 350202 Add limited param to 'monitor block_list'
1882 350290 s390x: Support instructions fixbr(a)
1883 350359 memcheck/tests/x86/fxsave hangs indefinetely on OS X
1884 350809 Fix none/tests/async-sigs for Solaris
1885 350811 Remove reference to --db-attach which has been removed.
1886 350813 Memcheck/x86: enable handwritten assembly helpers for x86/Solaris too
1887 350854 hard-to-understand code in VG_(load_ELF)()
1888 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
1889 351386 Solaris: Cannot run ld.so.1 under Valgrind
1890 351474 Fix VG_(iseqsigset) as obvious
1891 351531 Typo in /include/vki/vki-xen-physdev.h header guard
1892 351756 Intercept platform_memchr$VARIANT$Haswell on OS X
1893 351858 ldsoexec support on Solaris
1894 351873 Newer gcc doesn't allow __builtin_tabortdc[i] in ppc32 mode
1895 352130 helgrind reports false races for printfs using mempcpy on FILE* state
1896 352284 s390: Conditional jump depends on uninitialised value(s) in vfprintf
1897 352320 arm64 crash on none/tests/nestedfs
1898 352765 Vbit test fails on Power 6
1899 352768 The mbar instruction is missing from the Power PC support
1900 352769 Power PC program priority register (PPR) is not supported
1901 n-i-bz Provide implementations of certain compiler builtins to support
1902 compilers that may not provide those
1903 n-i-bz Old STABS code is still being compiled, but never used. Remove it.
1904 n-i-bz Fix compilation on distros with glibc < 2.5
1905 n-i-bz (vex 3098) Avoid generation of Neon insns on non-Neon hosts
1906 n-i-bz Enable rt_sigpending syscall on ppc64 linux.
1907 n-i-bz mremap did not work properly on shared memory
1908 n-i-bz Fix incorrect sizeof expression in syswrap-xen.c reported by Coverity
1909 n-i-bz In VALGRIND_PRINTF write out thread name, if any, to xml
1911 (3.11.0.TEST1: 8 September 2015, vex r3187, valgrind r15646)
1912 (3.11.0.TEST2: 21 September 2015, vex r3193, valgrind r15667)
1913 (3.11.0: 22 September 2015, vex r3195, valgrind r15674)
1917 Release 3.10.1 (25 November 2014)
1918 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1919 3.10.1 is a bug fix release. It fixes various bugs reported in 3.10.0
1920 and backports fixes for all reported missing AArch64 ARMv8 instructions
1921 and syscalls from the trunk. If you package or deliver 3.10.0 for others
1922 to use, you might want to consider upgrading to 3.10.1 instead.
1924 The following bugs have been fixed or resolved. Note that "n-i-bz"
1925 stands for "not in bugzilla" -- that is, a bug that was reported to us
1926 but never got a bugzilla entry. We encourage you to file bugs in
1927 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
1928 than mailing the developers (or mailing lists) directly -- bugs that
1929 are not entered into bugzilla tend to get forgotten about or ignored.
1931 To see details of a given bug, visit
1932 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
1933 where XXXXXX is the bug number as listed below.
1935 335440 arm64: ld1 (single structure) is not implemented
1936 335713 arm64: unhanded instruction: prfm (immediate)
1937 339020 ppc64: memcheck/tests/ppc64/power_ISA2_05 failing in nightly build
1938 339182 ppc64: AvSplat ought to load destination vector register with [..]
1939 339336 PPC64 store quad instruction (stq) is not supposed to change [..]
1940 339433 ppc64 lxvw4x instruction uses four 32-byte loads
1941 339645 Use correct tag names in sys_getdents/64 wrappers
1942 339706 Fix false positive for ioctl(TIOCSIG) on linux
1943 339721 assertion 'check_sibling == sibling' failed in readdwarf3.c ...
1944 339853 arm64 times syscall unknown
1945 339855 arm64 unhandled getsid/setsid syscalls
1946 339858 arm64 dmb sy not implemented
1947 339926 Unhandled instruction 0x1E674001 (frintx) on aarm64
1948 339927 Unhandled instruction 0x9E7100C6 (fcvtmu) on aarch64
1949 339938 disInstr(arm64): unhandled instruction 0x4F8010A4 (fmla)
1951 339940 arm64: unhandled syscall: 83 (sys_fdatasync) + patch
1952 340033 arm64: unhandled insn dmb ishld and some other isb-dmb-dsb variants
1953 340028 unhandled syscalls for arm64 (msync, pread64, setreuid and setregid)
1954 340036 arm64: Unhandled instruction ld4 (multiple structures, no offset)
1955 340236 arm64: unhandled syscalls: mknodat, fchdir, chroot, fchownat
1956 340509 arm64: unhandled instruction fcvtas
1957 340630 arm64: fchmod (52) and fchown (55) syscalls not recognized
1958 340632 arm64: unhandled instruction fcvtas
1959 340722 Resolve "UNKNOWN attrlist flags 0:0x10000000"
1960 340725 AVX2: Incorrect decoding of vpbroadcast{b,w} reg,reg forms
1961 340788 warning: unhandled syscall: 318 (getrandom)
1962 340807 disInstr(arm): unhandled instruction: 0xEE989B20
1963 340856 disInstr(arm64): unhandled instruction 0x1E634C45 (fcsel)
1964 340922 arm64: unhandled getgroups/setgroups syscalls
1965 350251 Fix typo in VEX utility program (test_main.c).
1966 350407 arm64: unhandled instruction ucvtf (vector, integer)
1967 350809 none/tests/async-sigs breaks when run under cron on Solaris
1968 350811 update README.solaris after r15445
1969 350813 Use handwritten memcheck assembly helpers on x86/Solaris [..]
1970 350854 strange code in VG_(load_ELF)()
1971 351140 arm64 syscalls setuid (146) and setresgid (149) not implemented
1972 n-i-bz DRD and Helgrind: Handle Imbe_CancelReservation (clrex on ARM)
1973 n-i-bz Add missing ]] to terminate CDATA.
1974 n-i-bz Glibc versions prior to 2.5 do not define PTRACE_GETSIGINFO
1975 n-i-bz Enable sys_fadvise64_64 on arm32.
1976 n-i-bz Add test cases for all remaining AArch64 SIMD, FP and memory insns.
1977 n-i-bz Add test cases for all known arm64 load/store instructions.
1978 n-i-bz PRE(sys_openat): when checking whether ARG1 == VKI_AT_FDCWD [..]
1979 n-i-bz Add detection of old ppc32 magic instructions from bug 278808.
1980 n-i-bz exp-dhat: Implement missing function "dh_malloc_usable_size".
1981 n-i-bz arm64: Implement "fcvtpu w, s".
1982 n-i-bz arm64: implement ADDP and various others
1983 n-i-bz arm64: Implement {S,U}CVTF (scalar, fixedpt).
1984 n-i-bz arm64: enable FCVT{A,N}S X,S.
1986 (3.10.1: 25 November 2014, vex r3026, valgrind r14785)
1990 Release 3.10.0 (10 September 2014)
1991 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1993 3.10.0 is a feature release with many improvements and the usual
1994 collection of bug fixes.
1996 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
1997 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
1998 MIPS64/Linux, ARM/Android, MIPS32/Android, X86/Android, X86/MacOSX 10.9
1999 and AMD64/MacOSX 10.9. Support for MacOSX 10.8 and 10.9 is
2000 significantly improved relative to the 3.9.0 release.
2002 * ================== PLATFORM CHANGES =================
2004 * Support for the 64-bit ARM Architecture (AArch64 ARMv8). This port
2005 is mostly complete, and is usable, but some SIMD instructions are as
2008 * Support for little-endian variant of the 64-bit POWER architecture.
2010 * Support for Android on MIPS32.
2012 * Support for 64bit FPU on MIPS32 platforms.
2014 * Both 32- and 64-bit executables are supported on MacOSX 10.8 and 10.9.
2016 * Configuration for and running on Android targets has changed.
2017 See README.android in the source tree for details.
2019 * ================== DEPRECATED FEATURES =================
2021 * --db-attach is now deprecated and will be removed in the next
2022 valgrind feature release. The built-in GDB server capabilities are
2023 superior and should be used instead. Learn more here:
2024 http://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.gdbserver
2026 * ==================== TOOL CHANGES ====================
2030 - Client code can now selectively disable and re-enable reporting of
2031 invalid address errors in specific ranges using the new client
2032 requests VALGRIND_DISABLE_ADDR_ERROR_REPORTING_IN_RANGE and
2033 VALGRIND_ENABLE_ADDR_ERROR_REPORTING_IN_RANGE.
2035 - Leak checker: there is a new leak check heuristic called
2036 "length64". This is used to detect interior pointers pointing 8
2037 bytes inside a block, on the assumption that the first 8 bytes
2038 holds the value "block size - 8". This is used by
2039 sqlite3MemMalloc, for example.
2041 - Checking of system call parameters: if a syscall parameter
2042 (e.g. bind struct sockaddr, sendmsg struct msghdr, ...) has
2043 several fields not initialised, an error is now reported for each
2044 field. Previously, an error was reported only for the first
2045 uninitialised field.
2047 - Mismatched alloc/free checking: a new flag
2048 --show-mismatched-frees=no|yes [yes] makes it possible to turn off
2049 such checks if necessary.
2053 - Improvements to error messages:
2055 o Race condition error message involving heap allocated blocks also
2056 show the thread number that allocated the raced-on block.
2058 o All locks referenced by an error message are now announced.
2059 Previously, some error messages only showed the lock addresses.
2061 o The message indicating where a lock was first observed now also
2062 describes the address/location of the lock.
2064 - Helgrind now understands the Ada task termination rules and
2065 creates a happens-before relationship between a terminated task
2066 and its master. This avoids some false positives and avoids a big
2067 memory leak when a lot of Ada tasks are created and terminated.
2068 The interceptions are only activated with forthcoming releases of
2069 gnatpro >= 7.3.0w-20140611 and gcc >= 5.0.
2071 - A new GDB server monitor command "info locks" giving the list of
2072 locks, their location, and their status.
2076 - callgrind_control now supports the --vgdb-prefix argument,
2077 which is needed if valgrind was started with this same argument.
2079 * ==================== OTHER CHANGES ====================
2081 * Unwinding through inlined function calls. Stack unwinding can now
2082 make use of Dwarf3 inlined-unwind information if it is available.
2083 The practical effect is that inlined calls become visible in stack
2084 traces. The suppression matching machinery has been adjusted
2085 accordingly. This is controlled by the new option
2086 --read-inline-info=yes|no. Currently this is enabled by default
2087 only on Linux and Android targets and only for the tools Memcheck,
2090 * Valgrind can now read EXIDX unwind information on 32-bit ARM
2091 targets. If an object contains both CFI and EXIDX unwind
2092 information, Valgrind will prefer the CFI over the EXIDX. This
2093 facilitates unwinding through system libraries on arm-android
2096 * Address description logic has been improved and is now common
2097 between Memcheck and Helgrind, resulting in better address
2098 descriptions for some kinds of error messages.
2100 * Error messages about dubious arguments (eg, to malloc or calloc) are
2101 output like other errors. This means that they can be suppressed
2102 and they have a stack trace.
2104 * The C++ demangler has been updated for better C++11 support.
2106 * New and modified GDB server monitor features:
2108 - Thread local variables/storage (__thread) can now be displayed.
2110 - The GDB server monitor command "v.info location <address>"
2111 displays information about an address. The information produced
2112 depends on the tool and on the options given to valgrind.
2113 Possibly, the following are described: global variables, local
2114 (stack) variables, allocated or freed blocks, ...
2116 - The option "--vgdb-stop-at=event1,event2,..." allows the user to
2117 ask the GDB server to stop at the start of program execution, at
2118 the end of the program execution and on Valgrind internal errors.
2120 - A new monitor command "v.info stats" shows various Valgrind core
2121 and tool statistics.
2123 - A new monitor command "v.set hostvisibility" allows the GDB server
2124 to provide access to Valgrind internal host status/memory.
2126 * A new option "--aspace-minaddr=<address>" can in some situations
2127 allow the use of more memory by decreasing the address above which
2128 Valgrind maps memory. It can also be used to solve address
2129 conflicts with system libraries by increasing the default value.
2130 See user manual for details.
2132 * The amount of memory used by Valgrind to store debug info (unwind
2133 info, line number information and symbol data) has been
2134 significantly reduced, even though Valgrind now reads more
2135 information in order to support unwinding of inlined function calls.
2137 * Dwarf3 handling with --read-var-info=yes has been improved:
2139 - Ada and C struct containing VLAs no longer cause a "bad DIE" error
2141 - Code compiled with
2142 -ffunction-sections -fdata-sections -Wl,--gc-sections
2143 no longer causes assertion failures.
2145 * Improved checking for the --sim-hints= and --kernel-variant=
2146 options. Unknown strings are now detected and reported to the user
2149 * The semantics of stack start/end boundaries in the valgrind.h
2150 VALGRIND_STACK_REGISTER client request has been clarified and
2151 documented. The convention is that start and end are respectively
2152 the lowest and highest addressable bytes of the stack.
2154 * ==================== FIXED BUGS ====================
2156 The following bugs have been fixed or resolved. Note that "n-i-bz"
2157 stands for "not in bugzilla" -- that is, a bug that was reported to us
2158 but never got a bugzilla entry. We encourage you to file bugs in
2159 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2160 than mailing the developers (or mailing lists) directly -- bugs that
2161 are not entered into bugzilla tend to get forgotten about or ignored.
2163 To see details of a given bug, visit
2164 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2165 where XXXXXX is the bug number as listed below.
2167 175819 Support for ipv6 socket reporting with --track-fds
2168 232510 make distcheck fails
2169 249435 Analyzing wine programs with callgrind triggers a crash
2170 278972 support for inlined function calls in stacktraces and suppression
2172 291310 FXSAVE instruction marks memory as undefined on amd64
2173 303536 ioctl for SIOCETHTOOL (ethtool(8)) isn't wrapped
2174 308729 vex x86->IR: unhandled instruction bytes 0xf 0x5 (syscall)
2175 315199 vgcore file for threaded app does not show which thread crashed
2176 315952 tun/tap ioctls are not supported
2177 323178 Unhandled instruction: PLDW register (ARM)
2178 323179 Unhandled instruction: PLDW immediate (ARM)
2179 324050 Helgrind: SEGV because of unaligned stack when using movdqa
2180 325110 Add test-cases for Power ISA 2.06 insns: divdo/divdo. and divduo/divduo.
2181 325124 [MIPSEL] Compilation error
2182 325477 Phase 4 support for IBM Power ISA 2.07
2183 325538 cavium octeon mips64, valgrind reported "dumping core" [...]
2184 325628 Phase 5 support for IBM Power ISA 2.07
2185 325714 Empty vgcore but RLIMIT_CORE is big enough (too big)
2186 325751 Missing the two privileged Power PC Transactional Memory Instructions
2187 325816 Phase 6 support for IBM Power ISA 2.07
2188 325856 Make SGCheck fail gracefully on unsupported platforms
2189 326026 Iop names for count leading zeros/sign bits incorrectly imply [..]
2190 326436 DRD: False positive in libstdc++ std::list::push_back
2191 326444 Cavium MIPS Octeon Specific Load Indexed Instructions
2192 326462 Refactor vgdb to isolate invoker stuff into separate module
2193 326469 amd64->IR: 0x66 0xF 0x3A 0x63 0xC1 0xE (pcmpistri 0x0E)
2194 326623 DRD: false positive conflict report in a field assignment
2195 326724 Valgrind does not compile on OSX 1.9 Mavericks
2196 326816 Intercept for __strncpy_sse2_unaligned missing?
2197 326921 coregrind fails to compile m_trampoline.S with MIPS/Linux port of V
2198 326983 Clear direction flag after tests on amd64.
2199 327212 Do not prepend the current directory to absolute path names.
2200 327223 Support for Cavium MIPS Octeon Atomic and Count Instructions
2201 327238 Callgrind Assertion 'passed <= last_bb->cjmp_count' failed
2202 327284 s390x: Fix translation of the risbg instruction
2203 327639 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x34
2204 327837 dwz compressed alternate .debug_info and .debug_str not read correctly
2205 327916 DW_TAG_typedef may have no name
2206 327943 s390x: add a redirection for the 'index' function
2207 328100 XABORT not implemented
2208 328205 Implement additional Xen hypercalls
2209 328454 add support Backtraces with ARM unwind tables (EXIDX)
2210 328455 s390x: SIGILL after emitting wrong register pair for ldxbr
2211 328711 valgrind.1 manpage "memcheck options" section is badly generated
2212 328878 vex amd64->IR pcmpestri SSE4.2 instruction is unsupported 0x14
2213 329612 Incorrect handling of AT_BASE for image execution
2214 329694 clang warns about using uninitialized variable
2215 329956 valgrind crashes when lmw/stmw instructions are used on ppc64
2216 330228 mmap must align to VKI_SHMLBA on mips32
2217 330257 LLVM does not support `-mno-dynamic-no-pic` option
2218 330319 amd64->IR: unhandled instruction bytes: 0xF 0x1 0xD5 (xend)
2219 330459 --track-fds=yes doesn't track eventfds
2220 330469 Add clock_adjtime syscall support
2221 330594 Missing sysalls on PowerPC / uClibc
2222 330622 Add test to regression suite for POWER instruction: dcbzl
2223 330939 Support for AMD's syscall instruction on x86
2225 330941 Typo in PRE(poll) syscall wrapper
2226 331057 unhandled instruction: 0xEEE01B20 (vfma.f64) (has patch)
2227 331254 Fix expected output for memcheck/tests/dw4
2228 331255 Fix race condition in test none/tests/coolo_sigaction
2229 331257 Fix type of jump buffer in test none/tests/faultstatus
2230 331305 configure uses bash specific syntax
2231 331337 s390x WARNING: unhandled syscall: 326 (dup3)
2232 331380 Syscall param timer_create(evp) points to uninitialised byte(s)
2233 331476 Patch to handle ioctl 0x5422 on Linux (x86 and amd64)
2234 331829 Unexpected ioctl opcode sign extension
2235 331830 ppc64: WARNING: unhandled syscall: 96/97
2236 331839 drd/tests/sem_open specifies invalid semaphore name
2237 331847 outcome of drd/tests/thread_name is nondeterministic
2238 332037 Valgrind cannot handle Thumb "add pc, reg"
2239 332055 drd asserts on platforms with VG_STACK_REDZONE_SZB == 0 and
2240 consistency checks enabled
2241 332263 intercepts for pthread_rwlock_timedrdlock and
2242 pthread_rwlock_timedwrlock are incorrect
2243 332265 drd could do with post-rwlock_init and pre-rwlock_destroy
2245 332276 Implement additional Xen hypercalls
2246 332658 ldrd.w r1, r2, [PC, #imm] does not adjust for 32bit alignment
2247 332765 Fix ms_print to create temporary files in a proper directory
2248 333072 drd: Add semaphore annotations
2249 333145 Tests for missaligned PC+#imm access for arm
2250 333228 AAarch64 Missing instruction encoding: mrs %[reg], ctr_el0
2251 333230 AAarch64 missing instruction encodings: dc, ic, dsb.
2252 333248 WARNING: unhandled syscall: unix:443
2253 333428 ldr.w pc [rD, #imm] instruction leads to assertion
2254 333501 cachegrind: assertion: Cache set count is not a power of two.
2257 333666 Recognize MPX instructions and bnd prefix.
2258 333788 Valgrind does not support the CDROM_DISC_STATUS ioctl (has patch)
2259 333817 Valgrind reports the memory areas written to by the SG_IO
2261 334049 lzcnt fails silently (x86_32)
2262 334384 Valgrind does not have support Little Endian support for
2264 334585 recvmmsg unhandled (+patch) (arm)
2265 334705 sendmsg and recvmsg should guard against bogus msghdr fields.
2266 334727 Build fails with -Werror=format-security
2267 334788 clarify doc about --log-file initial program directory
2268 334834 PPC64 Little Endian support, patch 2
2269 334836 PPC64 Little Endian support, patch 3 testcase fixes
2270 334936 patch to fix false positives on alsa SNDRV_CTL_* ioctls
2271 335034 Unhandled ioctl: HCIGETDEVLIST
2272 335155 vgdb, fix error print statement.
2273 335262 arm64: movi 8bit version is not supported
2274 335263 arm64: dmb instruction is not implemented
2275 335441 unhandled ioctl 0x8905 (SIOCATMARK) when running wine under valgrind
2276 335496 arm64: sbc/abc instructions are not implemented
2277 335554 arm64: unhandled instruction: abs
2278 335564 arm64: unhandled instruction: fcvtpu Xn, Sn
2279 335735 arm64: unhandled instruction: cnt
2280 335736 arm64: unhandled instruction: uaddlv
2281 335848 arm64: unhandled instruction: {s,u}cvtf
2282 335902 arm64: unhandled instruction: sli
2283 335903 arm64: unhandled instruction: umull (vector)
2284 336055 arm64: unhandled instruction: mov (element)
2285 336062 arm64: unhandled instruction: shrn{,2}
2286 336139 mip64: [...] valgrind hangs and spins on a single core [...]
2287 336189 arm64: unhandled Instruction: mvn
2288 336435 Valgrind hangs in pthread_spin_lock consuming 100% CPU
2289 336619 valgrind --read-var-info=yes doesn't handle DW_TAG_restrict_type
2290 336772 Make moans about unknown ioctls more informative
2291 336957 Add a section about the Solaris/illumos port on the webpage
2292 337094 ifunc wrapper is broken on ppc64
2293 337285 fcntl commands F_OFD_SETLK, F_OFD_SETLKW, and F_OFD_GETLK not supported
2294 337528 leak check heuristic for block prefixed by length as 64bit number
2295 337740 Implement additional Xen hypercalls
2296 337762 guest_arm64_toIR.c:4166 (dis_ARM64_load_store): Assertion `0' failed.
2297 337766 arm64-linux: unhandled syscalls mlock (228) and mlockall (230)
2298 337871 deprecate --db-attach
2299 338023 Add support for all V4L2/media ioctls
2300 338024 inlined functions are not shown if DW_AT_ranges is used
2301 338106 Add support for 'kcmp' syscall
2302 338115 DRD: computed conflict set differs from actual after fork
2303 338160 implement display of thread local storage in gdbsrv
2304 338205 configure.ac and check for -Wno-tautological-compare
2305 338300 coredumps are missing one byte of every segment
2306 338445 amd64 vbit-test fails with unknown opcodes used by arm64 VEX
2307 338499 --sim-hints parsing broken due to wrong order in tokens
2308 338615 suppress glibc 2.20 optimized strcmp implementation for ARMv7
2309 338681 Unable to unwind through clone thread created on i386-linux
2310 338698 race condition between gdbsrv and vgdb on startup
2311 338703 helgrind on arm-linux gets false positives in dynamic loader
2312 338791 alt dwz files can be relative of debug/main file
2313 338878 on MacOS: assertion 'VG_IS_PAGE_ALIGNED(clstack_end+1)' failed
2314 338932 build V-trunk with gcc-trunk
2315 338974 glibc 2.20 changed size of struct sigaction sa_flags field on s390
2316 345079 Fix build problems in VEX/useful/test_main.c
2317 n-i-bz Fix KVM_CREATE_IRQCHIP ioctl handling
2318 n-i-bz s390x: Fix memory corruption for multithreaded applications
2319 n-i-bz vex arm->IR: allow PC as basereg in some LDRD cases
2320 n-i-bz internal error in Valgrind if vgdb transmit signals when ptrace invoked
2321 n-i-bz Fix mingw64 support in valgrind.h (dev@, 9 May 2014)
2322 n-i-bz drd manual: Document how to C++11 programs that use class "std::thread"
2323 n-i-bz Add command-line option --default-suppressions
2324 n-i-bz Add support for BLKDISCARDZEROES ioctl
2325 n-i-bz ppc32/64: fix a regression with the mtfsb0/mtfsb1 instructions
2326 n-i-bz Add support for sys_pivot_root and sys_unshare
2328 (3.10.0.BETA1: 2 September 2014, vex r2940, valgrind r14428)
2329 (3.10.0.BETA2: 8 September 2014, vex r2950, valgrind r14503)
2330 (3.10.0: 10 September 2014, vex r2950, valgrind r14514)
2334 Release 3.9.0 (31 October 2013)
2335 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2336 3.9.0 is a feature release with many improvements and the usual
2337 collection of bug fixes.
2339 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2340 PPC64/Linux, S390X/Linux, MIPS32/Linux, MIPS64/Linux, ARM/Android,
2341 X86/Android, X86/MacOSX 10.7 and AMD64/MacOSX 10.7. Support for
2342 MacOSX 10.8 is significantly improved relative to the 3.8.0 release.
2344 * ================== PLATFORM CHANGES =================
2346 * Support for MIPS64 LE and BE running Linux. Valgrind has been
2347 tested on MIPS64 Debian Squeeze and Debian Wheezy distributions.
2349 * Support for MIPS DSP ASE on MIPS32 platforms.
2351 * Support for s390x Decimal Floating Point instructions on hosts that
2352 have the DFP facility installed.
2354 * Support for POWER8 (Power ISA 2.07) instructions
2356 * Support for Intel AVX2 instructions. This is available only on 64
2359 * Initial support for Intel Transactional Synchronization Extensions,
2362 * Initial support for Hardware Transactional Memory on POWER.
2364 * Improved support for MacOSX 10.8 (64-bit only). Memcheck can now
2365 run large GUI apps tolerably well.
2367 * ==================== TOOL CHANGES ====================
2371 - Improvements in handling of vectorised code, leading to
2372 significantly fewer false error reports. You need to use the flag
2373 --partial-loads-ok=yes to get the benefits of these changes.
2375 - Better control over the leak checker. It is now possible to
2376 specify which leak kinds (definite/indirect/possible/reachable)
2377 should be displayed, which should be regarded as errors, and which
2378 should be suppressed by a given leak suppression. This is done
2379 using the options --show-leak-kinds=kind1,kind2,..,
2380 --errors-for-leak-kinds=kind1,kind2,.. and an optional
2381 "match-leak-kinds:" line in suppression entries, respectively.
2383 Note that generated leak suppressions contain this new line and
2384 are therefore more specific than in previous releases. To get the
2385 same behaviour as previous releases, remove the "match-leak-kinds:"
2386 line from generated suppressions before using them.
2388 - Reduced "possible leak" reports from the leak checker by the use
2389 of better heuristics. The available heuristics provide detection
2390 of valid interior pointers to std::stdstring, to new[] allocated
2391 arrays with elements having destructors and to interior pointers
2392 pointing to an inner part of a C++ object using multiple
2393 inheritance. They can be selected individually using the
2394 option --leak-check-heuristics=heur1,heur2,...
2396 - Better control of stacktrace acquisition for heap-allocated
2397 blocks. Using the --keep-stacktraces option, it is possible to
2398 control independently whether a stack trace is acquired for each
2399 allocation and deallocation. This can be used to create better
2400 "use after free" errors or to decrease Valgrind's resource
2401 consumption by recording less information.
2403 - Better reporting of leak suppression usage. The list of used
2404 suppressions (shown when the -v option is given) now shows, for
2405 each leak suppressions, how many blocks and bytes it suppressed
2406 during the last leak search.
2410 - False errors resulting from the use of statically initialised
2411 mutexes and condition variables (PTHREAD_MUTEX_INITIALISER, etc)
2414 - False errors resulting from the use of pthread_cond_waits that
2415 timeout, have been removed.
2417 * ==================== OTHER CHANGES ====================
2419 * Some attempt to tune Valgrind's space requirements to the expected
2420 capabilities of the target:
2422 - The default size of the translation cache has been reduced from 8
2423 sectors to 6 on Android platforms, since each sector occupies
2424 about 40MB when using Memcheck.
2426 - The default size of the translation cache has been increased to 16
2427 sectors on all other platforms, reflecting the fact that large
2428 applications require instrumentation and storage of huge amounts
2429 of code. For similar reasons, the number of memory mapped
2430 segments that can be tracked has been increased by a factor of 6.
2432 - In all cases, the maximum number of sectors in the translation
2433 cache can be controlled by the new flag --num-transtab-sectors.
2435 * Changes in how debug info (line numbers, etc) is read:
2437 - Valgrind no longer temporarily mmaps the entire object to read
2438 from it. Instead, reading is done through a small fixed sized
2439 buffer. This avoids virtual memory usage spikes when Valgrind
2440 reads debuginfo from large shared objects.
2442 - A new experimental remote debug info server. Valgrind can read
2443 debug info from a different machine (typically, a build host)
2444 where debuginfo objects are stored. This can save a lot of time
2445 and hassle when running Valgrind on resource-constrained targets
2446 (phones, tablets) when the full debuginfo objects are stored
2447 somewhere else. This is enabled by the --debuginfo-server=
2450 - Consistency checking between main and debug objects can be
2451 disabled using the --allow-mismatched-debuginfo option.
2453 * Stack unwinding by stack scanning, on ARM. Unwinding by stack
2454 scanning can recover stack traces in some cases when the normal
2455 unwind mechanisms fail. Stack scanning is best described as "a
2456 nasty, dangerous and misleading hack" and so is disabled by default.
2457 Use --unw-stack-scan-thresh and --unw-stack-scan-frames to enable
2460 * Detection and merging of recursive stack frame cycles. When your
2461 program has recursive algorithms, this limits the memory used by
2462 Valgrind for recorded stack traces and avoids recording
2463 uninteresting repeated calls. This is controlled by the command
2464 line option --merge-recursive-frame and by the monitor command
2465 "v.set merge-recursive-frames".
2467 * File name and line numbers for used suppressions. The list of used
2468 suppressions (shown when the -v option is given) now shows, for each
2469 used suppression, the file name and line number where the suppression
2472 * New and modified GDB server monitor features:
2474 - valgrind.h has a new client request, VALGRIND_MONITOR_COMMAND,
2475 that can be used to execute gdbserver monitor commands from the
2478 - A new monitor command, "v.info open_fds", that gives the list of
2479 open file descriptors and additional details.
2481 - An optional message in the "v.info n_errs_found" monitor command,
2482 for example "v.info n_errs_found test 1234 finished", allowing a
2483 comment string to be added to the process output, perhaps for the
2484 purpose of separating errors of different tests or test phases.
2486 - A new monitor command "v.info execontext" that shows information
2487 about the stack traces recorded by Valgrind.
2489 - A new monitor command "v.do expensive_sanity_check_general" to run
2490 some internal consistency checks.
2492 * New flag --sigill-diagnostics to control whether a diagnostic
2493 message is printed when the JIT encounters an instruction it can't
2494 translate. The actual behavior -- delivery of SIGILL to the
2495 application -- is unchanged.
2497 * The maximum amount of memory that Valgrind can use on 64 bit targets
2498 has been increased from 32GB to 64GB. This should make it possible
2499 to run applications on Memcheck that natively require up to about 35GB.
2501 * ==================== FIXED BUGS ====================
2503 The following bugs have been fixed or resolved. Note that "n-i-bz"
2504 stands for "not in bugzilla" -- that is, a bug that was reported to us
2505 but never got a bugzilla entry. We encourage you to file bugs in
2506 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2507 than mailing the developers (or mailing lists) directly -- bugs that
2508 are not entered into bugzilla tend to get forgotten about or ignored.
2510 To see details of a given bug, visit
2511 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2512 where XXXXXX is the bug number as listed below.
2514 123837 system call: 4th argument is optional, depending on cmd
2515 135425 memcheck should tell you where Freed blocks were Mallocd
2516 164485 VG_N_SEGNAMES and VG_N_SEGMENTS are (still) too small
2517 207815 Adds some of the drm ioctls to syswrap-linux.c
2518 251569 vex amd64->IR: 0xF 0x1 0xF9 0xBF 0x90 0xD0 0x3 0x0 (RDTSCP)
2519 252955 Impossible to compile with ccache
2520 253519 Memcheck reports auxv pointer accesses as invalid reads.
2521 263034 Crash when loading some PPC64 binaries
2522 269599 Increase deepest backtrace
2523 274695 s390x: Support "compare to/from logical" instructions (z196)
2524 275800 s390x: Autodetect cache info (part 2)
2525 280271 Valgrind reports possible memory leaks on still-reachable std::string
2526 284540 Memcheck shouldn't count suppressions matching still-reachable [..]
2527 289578 Backtraces with ARM unwind tables (stack scan flags)
2528 296311 Wrong stack traces due to -fomit-frame-pointer (x86)
2529 304832 ppc32: build failure
2530 305431 Use find_buildid shdr fallback for separate .debug files
2531 305728 Add support for AVX2 instructions
2532 305948 ppc64: code generation for ShlD64 / ShrD64 asserts
2533 306035 s390x: Fix IR generation for LAAG and friends
2534 306054 s390x: Condition code computation for convert-to-int/logical
2535 306098 s390x: alternate opcode form for convert to/from fixed
2536 306587 Fix cache line detection from auxiliary vector for PPC.
2537 306783 Mips unhandled syscall : 4025 / 4079 / 4182
2538 307038 DWARF2 CFI reader: unhandled DW_OP_ opcode 0x8 (DW_OP_const1u et al)
2539 307082 HG false positive: pthread_cond_destroy: destruction of unknown CV
2540 307101 sys_capget second argument can be NULL
2541 307103 sys_openat: If pathname is absolute, then dirfd is ignored.
2542 307106 amd64->IR: f0 0f c0 02 (lock xadd byte)
2543 307113 s390x: DFP support
2544 307141 valgrind does't work in mips-linux system
2545 307155 filter_gdb should filter out syscall-template.S T_PSEUDO
2546 307285 x86_amd64 feature test for avx in test suite is wrong
2547 307290 memcheck overlap testcase needs memcpy version filter
2548 307463 Please add "&limit=0" to the "all open bugs" link
2549 307465 --show-possibly-lost=no should reduce the error count / exit code
2550 307557 Leaks on Mac OS X 10.7.5 libraries at ImageLoader::recursiveInit[..]
2551 307729 pkgconfig support broken valgrind.pc
2552 307828 Memcheck false errors SSE optimized wcscpy, wcscmp, wcsrchr, wcschr
2553 307955 Building valgrind 3.7.0-r4 fails in Gentoo AMD64 when using clang
2554 308089 Unhandled syscall on ppc64: prctl
2555 308135 PPC32 MPC8xx has 16 bytes cache size
2556 308321 testsuite memcheck filter interferes with gdb_filter
2558 308341 vgdb should report process exit (or fatal signal)
2559 308427 s390 memcheck reports tsearch cjump/cmove depends on uninit
2560 308495 Remove build dependency on installed Xen headers
2561 308573 Internal error on 64-bit instruction executed in 32-bit mode
2563 308627 pmovmskb validity bit propagation is imprecise
2564 308644 vgdb command for having the info for the track-fds option
2565 308711 give more info about aspacemgr and arenas in out_of_memory
2566 308717 ARM: implement fixed-point VCVT.F64.[SU]32
2567 308718 ARM implement SMLALBB family of instructions
2568 308886 Missing support for PTRACE_SET/GETREGSET
2569 308930 syscall name_to_handle_at (303 on amd64) not handled
2570 309229 V-bit tester does not report number of tests generated
2571 309323 print unrecognized instuction on MIPS
2572 309425 Provide a --sigill-diagnostics flag to suppress illegal [..]
2573 309427 SSE optimized stpncpy trigger uninitialised value [..] errors
2574 309430 Self hosting ppc64 encounters a vassert error on operand type
2575 309600 valgrind is a bit confused about 0-sized sections
2576 309823 Generate errors for still reachable blocks
2577 309921 PCMPISTRI validity bit propagation is imprecise
2578 309922 none/tests/ppc64/test_dfp5 sometimes fails
2579 310169 The Iop_CmpORD class of Iops is not supported by the vbit checker.
2580 310424 --read-var-info does not properly describe static variables
2581 310792 search additional path for debug symbols
2582 310931 s390x: Message-security assist (MSA) instruction extension [..]
2583 311100 PPC DFP implementation of the integer operands is inconsistent [..]
2584 311318 ARM: "128-bit constant is not implemented" error message
2585 311407 ssse3 bcopy (actually converted memcpy) causes invalid read [..]
2586 311690 V crashes because it redirects branches inside of a redirected function
2587 311880 x86_64: make regtest hangs at shell_valid1
2588 311922 WARNING: unhandled syscall: 170
2590 312171 ppc: insn selection for DFP
2591 312571 Rounding mode call wrong for the DFP Iops [..]
2592 312620 Change to Iop_D32toD64 [..] for s390 DFP support broke ppc [..]
2593 312913 Dangling pointers error should also report the alloc stack trace
2594 312980 Building on Mountain Lion generates some compiler warnings
2595 313267 Adding MIPS64/Linux port to Valgrind
2598 313811 Buffer overflow in assert_fail
2599 314099 coverity pointed out error in VEX guest_ppc_toIR.c insn_suffix
2600 314269 ppc: dead code in insn selection
2601 314718 ARM: implement integer divide instruction (sdiv and udiv)
2602 315345 cl-format.xml and callgrind/dump.c don't agree on using cfl= or cfi=
2603 315441 sendmsg syscall should ignore unset msghdr msg_flags
2604 315534 msgrcv inside a thread causes valgrind to hang (block)
2605 315545 Assertion '(UChar*)sec->tt[tteNo].tcptr <= (UChar*)hcode' failed
2606 315689 disInstr(thumb): unhandled instruction: 0xF852 0x0E10 (LDRT)
2607 315738 disInstr(arm): unhandled instruction: 0xEEBE0BEE (vcvt.s32.f64)
2608 315959 valgrind man page has bogus SGCHECK (and no BBV) OPTIONS section
2609 316144 valgrind.1 manpage contains unknown ??? strings [..]
2610 316145 callgrind command line options in manpage reference (unknown) [..]
2611 316145 callgrind command line options in manpage reference [..]
2612 316181 drd: Fixed a 4x slowdown for certain applications
2613 316503 Valgrind does not support SSE4 "movntdqa" instruction
2614 316535 Use of |signed int| instead of |size_t| in valgrind messages
2615 316696 fluidanimate program of parsec 2.1 stuck
2616 316761 syscall open_by_handle_at (304 on amd64, 342 on x86) not handled
2617 317091 Use -Wl,-Ttext-segment when static linking if possible [..]
2618 317186 "Impossible happens" when occurs VCVT instruction on ARM
2619 317318 Support for Threading Building Blocks "scalable_malloc"
2620 317444 amd64->IR: 0xC4 0x41 0x2C 0xC2 0xD2 0x8 (vcmpeq_uqps)
2621 317461 Fix BMI assembler configure check and avx2/bmi/fma vgtest prereqs
2622 317463 bmi testcase IR SANITY CHECK FAILURE
2623 317506 memcheck/tests/vbit-test fails with unknown opcode after [..]
2624 318050 libmpiwrap fails to compile with out-of-source build
2625 318203 setsockopt handling needs to handle SOL_SOCKET/SO_ATTACH_FILTER
2626 318643 annotate_trace_memory tests infinite loop on arm and ppc [..]
2627 318773 amd64->IR: 0xF3 0x48 0x0F 0xBC 0xC2 0xC3 0x66 0x0F
2628 318929 Crash with: disInstr(thumb): 0xF321 0x0001 (ssat16)
2629 318932 Add missing PPC64 and PPC32 system call support
2630 319235 --db-attach=yes is broken with Yama (ptrace scoping) enabled
2631 319395 Crash with unhandled instruction on STRT (Thumb) instructions
2632 319494 VEX Makefile-gcc standalone build update after r2702
2633 319505 [MIPSEL] Crash: unhandled UNRAY operator.
2634 319858 disInstr(thumb): unhandled instruction on instruction STRBT
2635 319932 disInstr(thumb): unhandled instruction on instruction STRHT
2636 320057 Problems when we try to mmap more than 12 memory pages on MIPS32
2637 320063 Memory from PTRACE_GET_THREAD_AREA is reported uninitialised
2638 320083 disInstr(thumb): unhandled instruction on instruction LDRBT
2639 320116 bind on AF_BLUETOOTH produces warnings because of sockaddr_rc padding
2640 320131 WARNING: unhandled syscall: 369 on ARM (prlimit64)
2641 320211 Stack buffer overflow in ./coregrind/m_main.c with huge TMPDIR
2642 320661 vgModuleLocal_read_elf_debug_info(): "Assertion '!di->soname'
2643 320895 add fanotify support (patch included)
2644 320998 vex amd64->IR pcmpestri and pcmpestrm SSE4.2 instruction
2645 321065 Valgrind updates for Xen 4.3
2646 321148 Unhandled instruction: PLI (Thumb 1, 2, 3)
2647 321363 Unhandled instruction: SSAX (ARM + Thumb)
2648 321364 Unhandled instruction: SXTAB16 (ARM + Thumb)
2649 321466 Unhandled instruction: SHASX (ARM + Thumb)
2650 321467 Unhandled instruction: SHSAX (ARM + Thumb)
2651 321468 Unhandled instruction: SHSUB16 (ARM + Thumb)
2652 321619 Unhandled instruction: SHSUB8 (ARM + Thumb)
2653 321620 Unhandled instruction: UASX (ARM + Thumb)
2654 321621 Unhandled instruction: USAX (ARM + Thumb)
2655 321692 Unhandled instruction: UQADD16 (ARM + Thumb)
2656 321693 Unhandled instruction: LDRSBT (Thumb)
2657 321694 Unhandled instruction: UQASX (ARM + Thumb)
2658 321696 Unhandled instruction: UQSAX (Thumb + ARM)
2659 321697 Unhandled instruction: UHASX (ARM + Thumb)
2660 321703 Unhandled instruction: UHSAX (ARM + Thumb)
2661 321704 Unhandled instruction: REVSH (ARM + Thumb)
2662 321730 Add cg_diff and cg_merge man pages
2663 321738 Add vgdb and valgrind-listener man pages
2665 321891 Unhandled instruction: LDRHT (Thumb)
2666 321960 pthread_create() then alloca() causing invalid stack write errors
2667 321969 ppc32 and ppc64 don't support [lf]setxattr
2668 322254 Show threadname together with tid if set by application
2669 322294 Add initial support for IBM Power ISA 2.07
2670 322368 Assertion failure in wqthread_hijack under OS X 10.8
2671 322563 vex mips->IR: 0x70 0x83 0xF0 0x3A
2672 322807 VALGRIND_PRINTF_BACKTRACE writes callstack to xml and text to stderr
2673 322851 0bXXX binary literal syntax is not standard
2674 323035 Unhandled instruction: LDRSHT(Thumb)
2675 323036 Unhandled instruction: SMMLS (ARM and Thumb)
2676 323116 The memcheck/tests/ppc64/power_ISA2_05.c fails to build [..]
2677 323175 Unhandled instruction: SMLALD (ARM + Thumb)
2678 323177 Unhandled instruction: SMLSLD (ARM + Thumb)
2679 323432 Calling pthread_cond_destroy() or pthread_mutex_destroy() [..]
2680 323437 Phase 2 support for IBM Power ISA 2.07
2681 323713 Support mmxext (integer sse) subset on i386 (athlon)
2682 323803 Transactional memory instructions are not supported for Power
2683 323893 SSE3 not available on amd cpus in valgrind
2684 323905 Probable false positive from Valgrind/drd on close()
2685 323912 valgrind.h header isn't compatible for mingw64
2686 324047 Valgrind doesn't support [LDR,ST]{S}[B,H]T ARM instructions
2687 324149 helgrind: When pthread_cond_timedwait returns ETIMEDOUT [..]
2688 324181 mmap does not handle MAP_32BIT
2689 324227 memcheck false positive leak when a thread calls exit+block [..]
2690 324421 Support for fanotify API on ARM architecture
2691 324514 gdbserver monitor cmd output behaviour consistency [..]
2692 324518 ppc64: Emulation of dcbt instructions does not handle [..]
2693 324546 none/tests/ppc32 test_isa_2_07_part2 requests -m64
2694 324582 When access is made to freed memory, report both allocation [..]
2695 324594 Fix overflow computation for Power ISA 2.06 insns: mulldo/mulldo.
2696 324765 ppc64: illegal instruction when executing none/tests/ppc64/jm-misc
2697 324816 Incorrect VEX implementation for xscvspdp/xvcvspdp for SNaN inputs
2698 324834 Unhandled instructions in Microsoft C run-time for x86_64
2699 324894 Phase 3 support for IBM Power ISA 2.07
2700 326091 drd: Avoid false race reports from optimized strlen() impls
2701 326113 valgrind libvex hwcaps error on AMD64
2702 n-i-bz Some wrong command line options could be ignored
2703 n-i-bz patch to allow fair-sched on android
2704 n-i-bz report error for vgdb snapshot requested before execution
2705 n-i-bz same as 303624 (fixed in 3.8.0), but for x86 android
2707 (3.9.0: 31 October 2013, vex r2796, valgrind r13708)
2711 Release 3.8.1 (19 September 2012)
2712 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2713 3.8.1 is a bug fix release. It fixes some assertion failures in 3.8.0
2714 that occur moderately frequently in real use cases, adds support for
2715 some missing instructions on ARM, and fixes a deadlock condition on
2716 MacOSX. If you package or deliver 3.8.0 for others to use, you might
2717 want to consider upgrading to 3.8.1 instead.
2719 The following bugs have been fixed or resolved. Note that "n-i-bz"
2720 stands for "not in bugzilla" -- that is, a bug that was reported to us
2721 but never got a bugzilla entry. We encourage you to file bugs in
2722 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2723 than mailing the developers (or mailing lists) directly -- bugs that
2724 are not entered into bugzilla tend to get forgotten about or ignored.
2726 To see details of a given bug, visit
2727 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2728 where XXXXXX is the bug number as listed below.
2731 289584 Unhandled instruction: 0xF 0x29 0xE5 (MOVAPS)
2732 295808 amd64->IR: 0xF3 0xF 0xBC 0xC0 (TZCNT)
2733 298281 wcslen causes false(?) uninitialised value warnings
2734 301281 valgrind hangs on OS X when the process calls system()
2735 304035 disInstr(arm): unhandled instruction 0xE1023053
2736 304867 implement MOVBE instruction in x86 mode
2737 304980 Assertion 'lo <= hi' failed in vgModuleLocal_find_rx_mapping
2738 305042 amd64: implement 0F 7F encoding of movq between two registers
2739 305199 ARM: implement QDADD and QDSUB
2740 305321 amd64->IR: 0xF 0xD 0xC (prefetchw)
2741 305513 killed by fatal signal: SIGSEGV
2742 305690 DRD reporting invalid semaphore when sem_trywait fails
2743 305926 Invalid alignment checks for some AVX instructions
2744 306297 disInstr(thumb): unhandled instruction 0xE883 0x000C
2745 306310 3.8.0 release tarball missing some files
2746 306612 RHEL 6 glibc-2.X default suppressions need /lib*/libc-*patterns
2747 306664 vex amd64->IR: 0x66 0xF 0x3A 0x62 0xD1 0x46 0x66 0xF
2748 n-i-bz shmat of a segment > 4Gb does not work
2749 n-i-bz simulate_control_c script wrong USR1 signal number on mips
2750 n-i-bz vgdb ptrace calls wrong on mips [...]
2751 n-i-bz Fixes for more MPI false positives
2752 n-i-bz exp-sgcheck's memcpy causes programs to segfault
2753 n-i-bz OSX build w/ clang: asserts at startup
2754 n-i-bz Incorrect undef'dness prop for Iop_DPBtoBCD and Iop_BCDtoDPB
2755 n-i-bz fix a couple of union tag-vs-field mixups
2756 n-i-bz OSX: use __NR_poll_nocancel rather than __NR_poll
2758 The following bugs were fixed in 3.8.0 but not listed in this NEWS
2761 254088 Valgrind should know about UD2 instruction
2764 304754 NEWS blows TeX's little mind
2766 (3.8.1: 19 September 2012, vex r2537, valgrind r12996)
2770 Release 3.8.0 (10 August 2012)
2771 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2772 3.8.0 is a feature release with many improvements and the usual
2773 collection of bug fixes.
2775 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
2776 PPC64/Linux, S390X/Linux, MIPS/Linux, ARM/Android, X86/Android,
2777 X86/MacOSX 10.6/10.7 and AMD64/MacOSX 10.6/10.7. Support for recent
2778 distros and toolchain components (glibc 2.16, gcc 4.7) has been added.
2779 There is initial support for MacOSX 10.8, but it is not usable for
2780 serious work at present.
2782 * ================== PLATFORM CHANGES =================
2784 * Support for MIPS32 platforms running Linux. Valgrind has been
2785 tested on MIPS32 and MIPS32r2 platforms running different Debian
2786 Squeeze and MeeGo distributions. Both little-endian and big-endian
2787 cores are supported. The tools Memcheck, Massif and Lackey have
2788 been tested and are known to work. See README.mips for more details.
2790 * Preliminary support for Android running on x86.
2792 * Preliminary (as-yet largely unusable) support for MacOSX 10.8.
2794 * Support for Intel AVX instructions and for AES instructions. This
2795 support is available only for 64 bit code.
2797 * Support for POWER Decimal Floating Point instructions.
2799 * ==================== TOOL CHANGES ====================
2801 * Non-libc malloc implementations are now supported. This is useful
2802 for tools that replace malloc (Memcheck, Massif, DRD, Helgrind).
2803 Using the new option --soname-synonyms, such tools can be informed
2804 that the malloc implementation is either linked statically into the
2805 executable, or is present in some other shared library different
2806 from libc.so. This makes it possible to process statically linked
2807 programs, and programs using other malloc libraries, for example
2808 TCMalloc or JEMalloc.
2810 * For tools that provide their own replacement for malloc et al, the
2811 option --redzone-size=<number> allows users to specify the size of
2812 the padding blocks (redzones) added before and after each client
2813 allocated block. Smaller redzones decrease the memory needed by
2814 Valgrind. Bigger redzones increase the chance to detect blocks
2815 overrun or underrun. Prior to this change, the redzone size was
2816 hardwired to 16 bytes in Memcheck.
2820 - The leak_check GDB server monitor command now can
2821 control the maximum nr of loss records to output.
2823 - Reduction of memory use for applications allocating
2824 many blocks and/or having many partially defined bytes.
2826 - Addition of GDB server monitor command 'block_list' that lists
2827 the addresses/sizes of the blocks of a leak search loss record.
2829 - Addition of GDB server monitor command 'who_points_at' that lists
2830 the locations pointing at a block.
2832 - If a redzone size > 0 is given, VALGRIND_MALLOCLIKE_BLOCK now will
2833 detect an invalid access of these redzones, by marking them
2834 noaccess. Similarly, if a redzone size is given for a memory
2835 pool, VALGRIND_MEMPOOL_ALLOC will mark the redzones no access.
2836 This still allows to find some bugs if the user has forgotten to
2837 mark the pool superblock noaccess.
2839 - Performance of memory leak check has been improved, especially in
2840 cases where there are many leaked blocks and/or many suppression
2841 rules used to suppress leak reports.
2843 - Reduced noise (false positive) level on MacOSX 10.6/10.7, due to
2844 more precise analysis, which is important for LLVM/Clang
2845 generated code. This is at the cost of somewhat reduced
2846 performance. Note there is no change to analysis precision or
2847 costs on Linux targets.
2851 - Added even more facilities that can help finding the cause of a data
2852 race, namely the command-line option --ptrace-addr and the macro
2853 DRD_STOP_TRACING_VAR(x). More information can be found in the manual.
2855 - Fixed a subtle bug that could cause false positive data race reports.
2857 * ==================== OTHER CHANGES ====================
2859 * The C++ demangler has been updated so as to work well with C++
2860 compiled by up to at least g++ 4.6.
2862 * Tool developers can make replacement/wrapping more flexible thanks
2863 to the new option --soname-synonyms. This was reported above, but
2864 in fact is very general and applies to all function
2865 replacement/wrapping, not just to malloc-family functions.
2867 * Round-robin scheduling of threads can be selected, using the new
2868 option --fair-sched= yes. Prior to this change, the pipe-based
2869 thread serialisation mechanism (which is still the default) could
2870 give very unfair scheduling. --fair-sched=yes improves
2871 responsiveness of interactive multithreaded applications, and
2872 improves repeatability of results from the thread checkers Helgrind
2875 * For tool developers: support to run Valgrind on Valgrind has been
2876 improved. We can now routinely Valgrind on Helgrind or Memcheck.
2878 * gdbserver now shows the float shadow registers as integer
2879 rather than float values, as the shadow values are mostly
2880 used as bit patterns.
2882 * Increased limit for the --num-callers command line flag to 500.
2884 * Performance improvements for error matching when there are many
2885 suppression records in use.
2887 * Improved support for DWARF4 debugging information (bug 284184).
2889 * Initial support for DWZ compressed Dwarf debug info.
2891 * Improved control over the IR optimiser's handling of the tradeoff
2892 between performance and precision of exceptions. Specifically,
2893 --vex-iropt-precise-memory-exns has been removed and replaced by
2894 --vex-iropt-register-updates, with extended functionality. This
2895 allows the Valgrind gdbserver to always show up to date register
2898 * Modest performance gains through the use of translation chaining for
2901 * ==================== FIXED BUGS ====================
2903 The following bugs have been fixed or resolved. Note that "n-i-bz"
2904 stands for "not in bugzilla" -- that is, a bug that was reported to us
2905 but never got a bugzilla entry. We encourage you to file bugs in
2906 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
2907 than mailing the developers (or mailing lists) directly -- bugs that
2908 are not entered into bugzilla tend to get forgotten about or ignored.
2910 To see details of a given bug, visit
2911 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
2912 where XXXXXX is the bug number as listed below.
2914 197914 Building valgrind from svn now requires automake-1.10
2915 203877 increase to 16Mb maximum allowed alignment for memalign et al
2916 219156 Handle statically linked malloc or other malloc lib (e.g. tcmalloc)
2917 247386 make perf does not run all performance tests
2918 270006 Valgrind scheduler unfair
2919 270777 Adding MIPS/Linux port to Valgrind
2920 270796 s390x: Removed broken support for the TS insn
2921 271438 Fix configure for proper SSE4.2 detection
2922 273114 s390x: Support TR, TRE, TROO, TROT, TRTO, and TRTT instructions
2923 273475 Add support for AVX instructions
2924 274078 improved configure logic for mpicc
2925 276993 fix mremap 'no thrash checks'
2926 278313 Fedora 15/x64: err read debug info with --read-var-info=yes flag
2927 281482 memcheck incorrect byte allocation count in realloc() for silly argument
2928 282230 group allocator for small fixed size, use it for MC_Chunk/SEc vbit
2929 283413 Fix wrong sanity check
2930 283671 Robustize alignment computation in LibVEX_Alloc
2931 283961 Adding support for some HCI IOCTLs
2932 284124 parse_type_DIE: confused by: DWARF 4
2933 284864 == 273475 (Add support for AVX instructions)
2934 285219 Too-restrictive constraints for Thumb2 "SP plus/minus register"
2935 285662 (MacOSX): Memcheck needs to replace memcpy/memmove
2936 285725 == 273475 (Add support for AVX instructions)
2937 286261 add wrapper for linux I2C_RDWR ioctl
2938 286270 vgpreload is not friendly to 64->32 bit execs, gives ld.so warnings
2939 286374 Running cachegrind with --branch-sim=yes on 64-bit PowerPC program fails
2940 286384 configure fails "checking for a supported version of gcc"
2941 286497 == 273475 (Add support for AVX instructions)
2942 286596 == 273475 (Add support for AVX instructions)
2943 286917 disInstr(arm): unhandled instruction: QADD (also QSUB)
2944 287175 ARM: scalar VFP fixed-point VCVT instructions not handled
2945 287260 Incorrect conditional jump or move depends on uninitialised value(s)
2946 287301 vex amd64->IR: 0x66 0xF 0x38 0x41 0xC0 0xB8 0x0 0x0 (PHMINPOSUW)
2947 287307 == 273475 (Add support for AVX instructions)
2948 287858 VG_(strerror): unknown error
2949 288298 (MacOSX) unhandled syscall shm_unlink
2950 288995 == 273475 (Add support for AVX instructions)
2951 289470 Loading of large Mach-O thin binaries fails.
2952 289656 == 273475 (Add support for AVX instructions)
2953 289699 vgdb connection in relay mode erroneously closed due to buffer overrun
2954 289823 == 293754 (PCMPxSTRx not implemented for 16-bit characters)
2955 289839 s390x: Provide support for unicode conversion instructions
2956 289939 monitor cmd 'leak_check' with details about leaked or reachable blocks
2957 290006 memcheck doesn't mark %xmm as initialized after "pcmpeqw %xmm %xmm"
2958 290655 Add support for AESKEYGENASSIST instruction
2959 290719 valgrind-3.7.0 fails with automake-1.11.2 due to"pkglibdir" usage
2960 290974 vgdb must align pages to VKI_SHMLBA (16KB) on ARM
2961 291253 ES register not initialised in valgrind simulation
2962 291568 Fix 3DNOW-related crashes with baseline x86_64 CPU (w patch)
2963 291865 s390x: Support the "Compare Double and Swap" family of instructions
2964 292300 == 273475 (Add support for AVX instructions)
2965 292430 unrecognized instruction in __intel_get_new_mem_ops_cpuid
2966 292493 == 273475 (Add support for AVX instructions)
2967 292626 Missing fcntl F_SETOWN_EX and F_GETOWN_EX support
2968 292627 Missing support for some SCSI ioctls
2969 292628 none/tests/x86/bug125959-x86.c triggers undefined behavior
2970 292841 == 273475 (Add support for AVX instructions)
2971 292993 implement the getcpu syscall on amd64-linux
2972 292995 Implement the “cross memory attach” syscalls introduced in Linux 3.2
2973 293088 Add some VEX sanity checks for ppc64 unhandled instructions
2974 293751 == 290655 (Add support for AESKEYGENASSIST instruction)
2975 293754 PCMPxSTRx not implemented for 16-bit characters
2976 293755 == 293754 (No tests for PCMPxSTRx on 16-bit characters)
2977 293808 CLFLUSH not supported by latest VEX for amd64
2978 294047 valgrind does not correctly emulate prlimit64(..., RLIMIT_NOFILE, ...)
2979 294048 MPSADBW instruction not implemented
2980 294055 regtest none/tests/shell fails when locale is not set to C
2981 294185 INT 0x44 (and others) not supported on x86 guest, but used by Jikes RVM
2982 294190 --vgdb-error=xxx can be out of sync with errors shown to the user
2983 294191 amd64: fnsave/frstor and 0x66 size prefixes on FP instructions
2984 294260 disInstr_AMD64: disInstr miscalculated next %rip
2985 294523 --partial-loads-ok=yes causes false negatives
2986 294617 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A
2987 294736 vex amd64->IR: 0x48 0xF 0xD7 0xD6 0x48 0x83
2988 294812 patch allowing to run (on x86 at least) helgrind/drd on tool.
2989 295089 can not annotate source for both helgrind and drd
2990 295221 POWER Processor decimal floating point instruction support missing
2991 295427 building for i386 with clang on darwin11 requires "-new_linker linker"
2992 295428 coregrind/m_main.c has incorrect x86 assembly for darwin
2993 295590 Helgrind: Assertion 'cvi->nWaiters > 0' failed
2994 295617 ARM - Add some missing syscalls
2995 295799 Missing \n with get_vbits in gdbserver when line is % 80 [...]
2996 296229 Linux user input device ioctls missing wrappers
2997 296318 ELF Debug info improvements (more than one rx/rw mapping)
2998 296422 Add translation chaining support
2999 296457 vex amd64->IR: 0x66 0xF 0x3A 0xDF 0xD1 0x1 0xE8 0x6A (dup of AES)
3000 296792 valgrind 3.7.0: add SIOCSHWTSTAMP (0x89B0) ioctl wrapper
3001 296983 Fix build issues on x86_64/ppc64 without 32-bit toolchains
3002 297078 gdbserver signal handling problems [..]
3003 297147 drd false positives on newly allocated memory
3004 297329 disallow decoding of IBM Power DFP insns on some machines
3005 297497 POWER Processor decimal floating point instruction support missing
3006 297701 Another alias for strncasecmp_l in libc-2.13.so
3007 297911 'invalid write' not reported when using APIs for custom mem allocators.
3008 297976 s390x: revisit EX implementation
3009 297991 Valgrind interferes with mmap()+ftell()
3010 297992 Support systems missing WIFCONTINUED (e.g. pre-2.6.10 Linux)
3011 297993 Fix compilation of valgrind with gcc -g3.
3012 298080 POWER Processor DFP support missing, part 3
3013 298227 == 273475 (Add support for AVX instructions)
3014 298335 == 273475 (Add support for AVX instructions)
3015 298354 Unhandled ARM Thumb instruction 0xEB0D 0x0585 (streq)
3016 298394 s390x: Don't bail out on an unknown machine model. [..]
3017 298421 accept4() syscall (366) support is missing for ARM
3018 298718 vex amd64->IR: 0xF 0xB1 0xCB 0x9C 0x8F 0x45
3019 298732 valgrind installation problem in ubuntu with kernel version 3.x
3020 298862 POWER Processor DFP instruction support missing, part 4
3021 298864 DWARF reader mis-parses DW_FORM_ref_addr
3022 298943 massif asserts with --pages-as-heap=yes when brk is changing [..]
3023 299053 Support DWARF4 DW_AT_high_pc constant form
3024 299104 == 273475 (Add support for AVX instructions)
3025 299316 Helgrind: hg_main.c:628 (map_threads_lookup): Assertion 'thr' failed.
3026 299629 dup3() syscall (358) support is missing for ARM
3027 299694 POWER Processor DFP instruction support missing, part 5
3028 299756 Ignore --free-fill for MEMPOOL_FREE and FREELIKE client requests
3029 299803 == 273475 (Add support for AVX instructions)
3030 299804 == 273475 (Add support for AVX instructions)
3031 299805 == 273475 (Add support for AVX instructions)
3032 300140 ARM - Missing (T1) SMMUL
3033 300195 == 296318 (ELF Debug info improvements (more than one rx/rw mapping))
3034 300389 Assertion `are_valid_hwcaps(VexArchAMD64, [..])' failed.
3035 300414 FCOM and FCOMP unimplemented for amd64 guest
3036 301204 infinite loop in canonicaliseSymtab with ifunc symbol
3037 301229 == 203877 (increase to 16Mb maximum allowed alignment for memalign etc)
3038 301265 add x86 support to Android build
3039 301984 configure script doesn't detect certain versions of clang
3040 302205 Fix compiler warnings for POWER VEX code and POWER test cases
3041 302287 Unhandled movbe instruction on Atom processors
3042 302370 PPC: fnmadd, fnmsub, fnmadds, fnmsubs insns always negate the result
3043 302536 Fix for the POWER Valgrind regression test: memcheck-ISA2.0.
3044 302578 Unrecognized isntruction 0xc5 0x32 0xc2 0xca 0x09 vcmpngess
3045 302656 == 273475 (Add support for AVX instructions)
3046 302709 valgrind for ARM needs extra tls support for android emulator [..]
3047 302827 add wrapper for CDROM_GET_CAPABILITY
3048 302901 Valgrind crashes with dwz optimized debuginfo
3049 302918 Enable testing of the vmaddfp and vnsubfp instructions in the testsuite
3050 303116 Add support for the POWER instruction popcntb
3051 303127 Power test suite fixes for frsqrte, vrefp, and vrsqrtefp instructions.
3052 303250 Assertion `instrs_in->arr_used <= 10000' failed w/ OpenSSL code
3053 303466 == 273475 (Add support for AVX instructions)
3054 303624 segmentation fault on Android 4.1 (e.g. on Galaxy Nexus OMAP)
3055 303963 strstr() function produces wrong results under valgrind callgrind
3056 304054 CALL_FN_xx macros need to enforce stack alignment
3057 304561 tee system call not supported
3058 715750 (MacOSX): Incorrect invalid-address errors near 0xFFFFxxxx (mozbug#)
3059 n-i-bz Add missing gdbserver xml files for shadow registers for ppc32
3060 n-i-bz Bypass gcc4.4/4.5 code gen bugs causing out of memory or asserts
3061 n-i-bz Fix assert in gdbserver for watchpoints watching the same address
3062 n-i-bz Fix false positive in sys_clone on amd64 when optional args [..]
3063 n-i-bz s390x: Shadow registers can now be examined using vgdb
3065 (3.8.0-TEST3: 9 August 2012, vex r2465, valgrind r12865)
3066 (3.8.0: 10 August 2012, vex r2465, valgrind r12866)
3070 Release 3.7.0 (5 November 2011)
3071 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3072 3.7.0 is a feature release with many significant improvements and the
3073 usual collection of bug fixes.
3075 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3076 PPC64/Linux, S390X/Linux, ARM/Android, X86/Darwin and AMD64/Darwin.
3077 Support for recent distros and toolchain components (glibc 2.14, gcc
3078 4.6, MacOSX 10.7) has been added.
3080 * ================== PLATFORM CHANGES =================
3082 * Support for IBM z/Architecture (s390x) running Linux. Valgrind can
3083 analyse 64-bit programs running on z/Architecture. Most user space
3084 instructions up to and including z10 are supported. Valgrind has
3085 been tested extensively on z9, z10, and z196 machines running SLES
3086 10/11, RedHat 5/6m, and Fedora. The Memcheck and Massif tools are
3087 known to work well. Callgrind, Helgrind, and DRD work reasonably
3088 well on z9 and later models. See README.s390 for more details.
3090 * Preliminary support for MacOSX 10.7 and XCode 4. Both 32- and
3091 64-bit processes are supported. Some complex threaded applications
3092 (Firefox) are observed to hang when run as 32 bit applications,
3093 whereas 64-bit versions run OK. The cause is unknown. Memcheck
3094 will likely report some false errors. In general, expect some rough
3095 spots. This release also supports MacOSX 10.6, but drops support
3098 * Preliminary support for Android (on ARM). Valgrind can now run
3099 large applications (eg, Firefox) on (eg) a Samsung Nexus S. See
3100 README.android for more details, plus instructions on how to get
3103 * Support for the IBM Power ISA 2.06 (Power7 instructions)
3105 * General correctness and performance improvements for ARM/Linux, and,
3106 by extension, ARM/Android.
3108 * Further solidification of support for SSE 4.2 in 64-bit mode. AVX
3109 instruction set support is under development but is not available in
3112 * Support for AIX5 has been removed.
3114 * ==================== TOOL CHANGES ====================
3116 * Memcheck: some incremental changes:
3118 - reduction of memory use in some circumstances
3120 - improved handling of freed memory, which in some circumstances
3121 can cause detection of use-after-free that would previously have
3124 - fix of a longstanding bug that could cause false negatives (missed
3125 errors) in programs doing vector saturated narrowing instructions.
3127 * Helgrind: performance improvements and major memory use reductions,
3128 particularly for large, long running applications which perform many
3129 synchronisation (lock, unlock, etc) events. Plus many smaller
3132 - display of locksets for both threads involved in a race
3134 - general improvements in formatting/clarity of error messages
3136 - addition of facilities and documentation regarding annotation
3137 of thread safe reference counted C++ classes
3139 - new flag --check-stack-refs=no|yes [yes], to disable race checking
3140 on thread stacks (a performance hack)
3142 - new flag --free-is-write=no|yes [no], to enable detection of races
3143 where one thread accesses heap memory but another one frees it,
3144 without any coordinating synchronisation event
3146 * DRD: enabled XML output; added support for delayed thread deletion
3147 in order to detect races that occur close to the end of a thread
3148 (--join-list-vol); fixed a memory leak triggered by repeated client
3149 memory allocatation and deallocation; improved Darwin support.
3151 * exp-ptrcheck: this tool has been renamed to exp-sgcheck
3153 * exp-sgcheck: this tool has been reduced in scope so as to improve
3154 performance and remove checking that Memcheck does better.
3155 Specifically, the ability to check for overruns for stack and global
3156 arrays is unchanged, but the ability to check for overruns of heap
3157 blocks has been removed. The tool has accordingly been renamed to
3158 exp-sgcheck ("Stack and Global Array Checking").
3160 * ==================== OTHER CHANGES ====================
3162 * GDB server: Valgrind now has an embedded GDB server. That means it
3163 is possible to control a Valgrind run from GDB, doing all the usual
3164 things that GDB can do (single stepping, breakpoints, examining
3165 data, etc). Tool-specific functionality is also available. For
3166 example, it is possible to query the definedness state of variables
3167 or memory from within GDB when running Memcheck; arbitrarily large
3168 memory watchpoints are supported, etc. To use the GDB server, start
3169 Valgrind with the flag --vgdb-error=0 and follow the on-screen
3172 * Improved support for unfriendly self-modifying code: a new option
3173 --smc-check=all-non-file is available. This adds the relevant
3174 consistency checks only to code that originates in non-file-backed
3175 mappings. In effect this confines the consistency checking only to
3176 code that is or might be JIT generated, and avoids checks on code
3177 that must have been compiled ahead of time. This significantly
3178 improves performance on applications that generate code at run time.
3180 * It is now possible to build a working Valgrind using Clang-2.9 on
3183 * new client requests VALGRIND_{DISABLE,ENABLE}_ERROR_REPORTING.
3184 These enable and disable error reporting on a per-thread, and
3185 nestable, basis. This is useful for hiding errors in particularly
3186 troublesome pieces of code. The MPI wrapper library (libmpiwrap.c)
3187 now uses this facility.
3189 * Added the --mod-funcname option to cg_diff.
3191 * ==================== FIXED BUGS ====================
3193 The following bugs have been fixed or resolved. Note that "n-i-bz"
3194 stands for "not in bugzilla" -- that is, a bug that was reported to us
3195 but never got a bugzilla entry. We encourage you to file bugs in
3196 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3197 mailing the developers (or mailing lists) directly -- bugs that are
3198 not entered into bugzilla tend to get forgotten about or ignored.
3200 To see details of a given bug, visit
3201 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3202 where XXXXXX is the bug number as listed below.
3204 79311 malloc silly arg warning does not give stack trace
3205 210935 port valgrind.h (not valgrind) to win32 to support client requests
3206 214223 valgrind SIGSEGV on startup gcc 4.4.1 ppc32 (G4) Ubuntu 9.10
3207 243404 Port to zSeries
3208 243935 Helgrind: incorrect handling of ANNOTATE_HAPPENS_BEFORE()/AFTER()
3209 247223 non-x86: Suppress warning: 'regparm' attribute directive ignored
3210 250101 huge "free" memory usage due to m_mallocfree.c fragmentation
3211 253206 Some fixes for the faultstatus testcase
3212 255223 capget testcase fails when running as root
3213 256703 xlc_dbl_u32.c testcase broken
3214 256726 Helgrind tests have broken inline asm
3215 259977 == 214223 (Valgrind segfaults doing __builtin_longjmp)
3216 264800 testcase compile failure on zseries
3217 265762 make public VEX headers compilable by G++ 3.x
3218 265771 assertion in jumps.c (r11523) fails with glibc-2.3
3219 266753 configure script does not give the user the option to not use QtCore
3220 266931 gen_insn_test.pl is broken
3221 266961 ld-linux.so.2 i?86-linux strlen issues
3222 266990 setns instruction causes false positive
3223 267020 Make directory for temporary files configurable at run-time.
3224 267342 == 267997 (segmentation fault on Mac OS 10.6)
3225 267383 Assertion 'vgPlain_strlen(dir) + vgPlain_strlen(file) + 1 < 256' failed
3226 267413 Assertion 'DRD_(g_threadinfo)[tid].synchr_nesting >= 1' failed.
3227 267488 regtest: darwin support for 64-bit build
3228 267552 SIGSEGV (misaligned_stack_error) with DRD, but not with other tools
3229 267630 Add support for IBM Power ISA 2.06 -- stage 1
3230 267769 == 267997 (Darwin: memcheck triggers segmentation fault)
3231 267819 Add client request for informing the core about reallocation
3232 267925 laog data structure quadratic for a single sequence of lock
3233 267968 drd: (vgDrd_thread_set_joinable): Assertion '0 <= (int)tid ..' failed
3234 267997 MacOSX: 64-bit V segfaults on launch when built with Xcode 4.0.1
3235 268513 missed optimizations in fold_Expr
3236 268619 s390x: fpr - gpr transfer facility
3237 268620 s390x: reconsider "long displacement" requirement
3238 268621 s390x: improve IR generation for XC
3239 268715 s390x: FLOGR is not universally available
3240 268792 == 267997 (valgrind seg faults on startup when compiled with Xcode 4)
3241 268930 s390x: MHY is not universally available
3242 269078 arm->IR: unhandled instruction SUB (SP minus immediate/register)
3243 269079 Support ptrace system call on ARM
3244 269144 missing "Bad option" error message
3245 269209 conditional load and store facility (z196)
3246 269354 Shift by zero on x86 can incorrectly clobber CC_NDEP
3247 269641 == 267997 (valgrind segfaults immediately (segmentation fault))
3248 269736 s390x: minor code generation tweaks
3249 269778 == 272986 (valgrind.h: swap roles of VALGRIND_DO_CLIENT_REQUEST() ..)
3250 269863 s390x: remove unused function parameters
3251 269864 s390x: tweak s390_emit_load_cc
3252 269884 == 250101 (overhead for huge blocks exhausts space too soon)
3253 270082 s390x: Make sure to point the PSW address to the next address on SIGILL
3254 270115 s390x: rewrite some testcases
3255 270309 == 267997 (valgrind crash on startup)
3256 270320 add support for Linux FIOQSIZE ioctl() call
3257 270326 segfault while trying to sanitize the environment passed to execle
3258 270794 IBM POWER7 support patch causes regression in none/tests
3259 270851 IBM POWER7 fcfidus instruction causes memcheck to fail
3260 270856 IBM POWER7 xsnmaddadp instruction causes memcheck to fail on 32bit app
3261 270925 hyper-optimized strspn() in /lib64/libc-2.13.so needs fix
3262 270959 s390x: invalid use of R0 as base register
3263 271042 VSX configure check fails when it should not
3264 271043 Valgrind build fails with assembler error on ppc64 with binutils 2.21
3265 271259 s390x: fix code confusion
3266 271337 == 267997 (Valgrind segfaults on MacOS X)
3267 271385 s390x: Implement Ist_MBE
3268 271501 s390x: misc cleanups
3269 271504 s390x: promote likely and unlikely
3270 271579 ppc: using wrong enum type
3271 271615 unhandled instruction "popcnt" (arch=amd10h)
3272 271730 Fix bug when checking ioctls: duplicate check
3273 271776 s390x: provide STFLE instruction support
3274 271779 s390x: provide clock instructions like STCK
3275 271799 Darwin: ioctls without an arg report a memory error
3276 271820 arm: fix type confusion
3277 271917 pthread_cond_timedwait failure leads to not-locked false positive
3278 272067 s390x: fix DISP20 macro
3279 272615 A typo in debug output in mc_leakcheck.c
3280 272661 callgrind_annotate chokes when run from paths containing regex chars
3281 272893 amd64->IR: 0x66 0xF 0x38 0x2B 0xC1 0x66 0xF 0x7F == (closed as dup)
3282 272955 Unhandled syscall error for pwrite64 on ppc64 arch
3283 272967 make documentation build-system more robust
3284 272986 Fix gcc-4.6 warnings with valgrind.h
3285 273318 amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38 (missing PCMPxSTRx case)
3286 273318 unhandled PCMPxSTRx case: vex amd64->IR: 0x66 0xF 0x3A 0x61 0xC1 0x38
3287 273431 valgrind segfaults in evalCfiExpr (debuginfo.c:2039)
3288 273465 Callgrind: jumps.c:164 (new_jcc): Assertion '(0 <= jmp) && ...'
3289 273536 Build error: multiple definition of `vgDrd_pthread_cond_initializer'
3290 273640 ppc64-linux: unhandled syscalls setresuid(164) and setresgid(169)
3291 273729 == 283000 (Illegal opcode for SSE2 "roundsd" instruction)
3292 273778 exp-ptrcheck: unhandled sysno == 259
3293 274089 exp-ptrcheck: unhandled sysno == 208
3294 274378 s390x: Various dispatcher tweaks
3295 274447 WARNING: unhandled syscall: 340
3296 274776 amd64->IR: 0x66 0xF 0x38 0x2B 0xC5 0x66
3297 274784 == 267997 (valgrind ls -l results in Segmentation Fault)
3298 274926 valgrind does not build against linux-3
3299 275148 configure FAIL with glibc-2.14
3300 275151 Fedora 15 / glibc-2.14 'make regtest' FAIL
3301 275168 Make Valgrind work for MacOSX 10.7 Lion
3302 275212 == 275284 (lots of false positives from __memcpy_ssse3_back et al)
3303 275278 valgrind does not build on Linux kernel 3.0.* due to silly
3304 275284 Valgrind memcpy/memmove redirection stopped working in glibc 2.14/x86_64
3305 275308 Fix implementation for ppc64 fres instruc
3306 275339 s390x: fix testcase compile warnings
3307 275517 s390x: Provide support for CKSM instruction
3308 275710 s390x: get rid of redundant address mode calculation
3309 275815 == 247894 (Valgrind doesn't know about Linux readahead(2) syscall)
3310 275852 == 250101 (valgrind uses all swap space and is killed)
3311 276784 Add support for IBM Power ISA 2.06 -- stage 3
3312 276987 gdbsrv: fix tests following recent commits
3313 277045 Valgrind crashes with unhandled DW_OP_ opcode 0x2a
3314 277199 The test_isa_2_06_part1.c in none/tests/ppc64 should be a symlink
3315 277471 Unhandled syscall: 340
3316 277610 valgrind crashes in VG_(lseek)(core_fd, phdrs[idx].p_offset, ...)
3317 277653 ARM: support Thumb2 PLD instruction
3318 277663 ARM: NEON float VMUL by scalar incorrect
3319 277689 ARM: tests for VSTn with register post-index are broken
3320 277694 ARM: BLX LR instruction broken in ARM mode
3321 277780 ARM: VMOV.F32 (immediate) instruction is broken
3322 278057 fuse filesystem syscall deadlocks
3323 278078 Unimplemented syscall 280 on ppc32
3324 278349 F_GETPIPE_SZ and F_SETPIPE_SZ Linux fcntl commands
3325 278454 VALGRIND_STACK_DEREGISTER has wrong output type
3326 278502 == 275284 (Valgrind confuses memcpy() and memmove())
3327 278892 gdbsrv: factorize gdb version handling, fix doc and typos
3328 279027 Support for MVCL and CLCL instruction
3329 279027 s390x: Provide support for CLCL and MVCL instructions
3330 279062 Remove a redundant check in the insn selector for ppc.
3331 279071 JDK creates PTEST with redundant REX.W prefix
3332 279212 gdbsrv: add monitor cmd v.info scheduler.
3333 279378 exp-ptrcheck: the 'impossible' happened on mkfifo call
3334 279698 memcheck discards valid-bits for packuswb
3335 279795 memcheck reports uninitialised values for mincore on amd64
3336 279994 Add support for IBM Power ISA 2.06 -- stage 3
3337 280083 mempolicy syscall check errors
3338 280290 vex amd64->IR: 0x66 0xF 0x38 0x28 0xC1 0x66 0xF 0x6F
3339 280710 s390x: config files for nightly builds
3340 280757 /tmp dir still used by valgrind even if TMPDIR is specified
3341 280965 Valgrind breaks fcntl locks when program does mmap
3342 281138 WARNING: unhandled syscall: 340
3343 281241 == 275168 (valgrind useless on Macos 10.7.1 Lion)
3344 281304 == 275168 (Darwin: dyld "cannot load inserted library")
3345 281305 == 275168 (unhandled syscall: unix:357 on Darwin 11.1)
3346 281468 s390x: handle do_clone and gcc clones in call traces
3347 281488 ARM: VFP register corruption
3348 281828 == 275284 (false memmove warning: "Source and destination overlap")
3349 281883 s390x: Fix system call wrapper for "clone".
3350 282105 generalise 'reclaimSuperBlock' to also reclaim splittable superblock
3351 282112 Unhandled instruction bytes: 0xDE 0xD9 0x9B 0xDF (fcompp)
3352 282238 SLES10: make check fails
3353 282979 strcasestr needs replacement with recent(>=2.12) glibc
3354 283000 vex amd64->IR: 0x66 0xF 0x3A 0xA 0xC0 0x9 0xF3 0xF
3355 283243 Regression in ppc64 memcheck tests
3356 283325 == 267997 (Darwin: V segfaults on startup when built with Xcode 4.0)
3357 283427 re-connect epoll_pwait syscall on ARM linux
3358 283600 gdbsrv: android: port vgdb.c
3359 283709 none/tests/faultstatus needs to account for page size
3360 284305 filter_gdb needs enhancement to work on ppc64
3361 284384 clang 3.1 -Wunused-value warnings in valgrind.h, memcheck.h
3362 284472 Thumb2 ROR.W encoding T2 not implemented
3363 284621 XML-escape process command line in XML output
3364 n-i-bz cachegrind/callgrind: handle CPUID information for Core iX Intel CPUs
3365 that have non-power-of-2 sizes (also AMDs)
3366 n-i-bz don't be spooked by libraries mashed by elfhack
3367 n-i-bz don't be spooked by libxul.so linked with gold
3368 n-i-bz improved checking for VALGRIND_CHECK_MEM_IS_DEFINED
3370 (3.7.0-TEST1: 27 October 2011, vex r2228, valgrind r12245)
3371 (3.7.0.RC1: 1 November 2011, vex r2231, valgrind r12257)
3372 (3.7.0: 5 November 2011, vex r2231, valgrind r12258)
3376 Release 3.6.1 (16 February 2011)
3377 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3378 3.6.1 is a bug fix release. It adds support for some SSE4
3379 instructions that were omitted in 3.6.0 due to lack of time. Initial
3380 support for glibc-2.13 has been added. A number of bugs causing
3381 crashing or assertion failures have been fixed.
3383 The following bugs have been fixed or resolved. Note that "n-i-bz"
3384 stands for "not in bugzilla" -- that is, a bug that was reported to us
3385 but never got a bugzilla entry. We encourage you to file bugs in
3386 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3387 mailing the developers (or mailing lists) directly -- bugs that are
3388 not entered into bugzilla tend to get forgotten about or ignored.
3390 To see details of a given bug, visit
3391 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3392 where XXXXXX is the bug number as listed below.
3394 188572 Valgrind on Mac should suppress setenv() mem leak
3395 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 (proper FX{SAVE,RSTOR} support)
3396 210481 vex amd64->IR: Assertion `sz == 2 || sz == 4' failed (REX.W POPQ)
3397 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3398 250038 ppc64: Altivec LVSR and LVSL instructions fail their regtest
3399 254420 memory pool tracking broken
3400 254957 Test code failing to compile due to changes in memcheck.h
3401 255009 helgrind/drd: crash on chmod with invalid parameter
3402 255130 readdwarf3.c parse_type_DIE confused by GNAT Ada types
3403 255355 helgrind/drd: crash on threaded programs doing fork
3405 255418 (SSE4.x) rint call compiled with ICC
3406 255822 --gen-suppressions can create invalid files: "too many callers [...]"
3407 255888 closing valgrindoutput tag outputted to log-stream on error
3408 255963 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x9 0xDB 0x0 (ROUNDPD)
3409 255966 Slowness when using mempool annotations
3410 256387 vex x86->IR: 0xD4 0xA 0x2 0x7 (AAD and AAM)
3411 256600 super-optimized strcasecmp() false positive
3412 256669 vex amd64->IR: Unhandled LOOPNEL insn on amd64
3413 256968 (SSE4.x) vex amd64->IR: 0x66 0xF 0x38 0x10 0xD3 0x66 (BLENDVPx)
3414 257011 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0xE 0xFD 0xA0 (PBLENDW)
3415 257063 (SSE4.x) vex amd64->IR: 0x66 0xF 0x3A 0x8 0xC0 0x0 (ROUNDPS)
3416 257276 Missing case in memcheck --track-origins=yes
3417 258870 (SSE4.x) Add support for EXTRACTPS SSE 4.1 instruction
3418 261966 (SSE4.x) support for CRC32B and CRC32Q is lacking (also CRC32{W,L})
3419 262985 VEX regression in valgrind 3.6.0 in handling PowerPC VMX
3420 262995 (SSE4.x) crash when trying to valgrind gcc-snapshot (PCMPxSTRx $0)
3421 263099 callgrind_annotate counts Ir improperly [...]
3422 263877 undefined coprocessor instruction on ARMv7
3423 265964 configure FAIL with glibc-2.13
3424 n-i-bz Fix compile error w/ icc-12.x in guest_arm_toIR.c
3425 n-i-bz Docs: fix bogus descriptions for VALGRIND_CREATE_BLOCK et al
3426 n-i-bz Massif: don't assert on shmat() with --pages-as-heap=yes
3427 n-i-bz Bug fixes and major speedups for the exp-DHAT space profiler
3428 n-i-bz DRD: disable --free-is-write due to implementation difficulties
3430 (3.6.1: 16 February 2011, vex r2103, valgrind r11561).
3434 Release 3.6.0 (21 October 2010)
3435 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3436 3.6.0 is a feature release with many significant improvements and the
3437 usual collection of bug fixes.
3439 This release supports X86/Linux, AMD64/Linux, ARM/Linux, PPC32/Linux,
3440 PPC64/Linux, X86/Darwin and AMD64/Darwin. Support for recent distros
3441 and toolchain components (glibc 2.12, gcc 4.5, OSX 10.6) has been added.
3443 -------------------------
3445 Here are some highlights. Details are shown further down:
3447 * Support for ARM/Linux.
3449 * Support for recent Linux distros: Ubuntu 10.10 and Fedora 14.
3451 * Support for Mac OS X 10.6, both 32- and 64-bit executables.
3453 * Support for the SSE4.2 instruction set.
3455 * Enhancements to the Callgrind profiler, including the ability to
3456 handle CPUs with three levels of cache.
3458 * A new experimental heap profiler, DHAT.
3460 * A huge number of bug fixes and small enhancements.
3462 -------------------------
3464 Here are details of the above changes, together with descriptions of
3465 many other changes, and a list of fixed bugs.
3467 * ================== PLATFORM CHANGES =================
3469 * Support for ARM/Linux. Valgrind now runs on ARMv7 capable CPUs
3470 running Linux. It is known to work on Ubuntu 10.04, Ubuntu 10.10,
3471 and Maemo 5, so you can run Valgrind on your Nokia N900 if you want.
3473 This requires a CPU capable of running the ARMv7-A instruction set
3474 (Cortex A5, A8 and A9). Valgrind provides fairly complete coverage
3475 of the user space instruction set, including ARM and Thumb integer
3476 code, VFPv3, NEON and V6 media instructions. The Memcheck,
3477 Cachegrind and Massif tools work properly; other tools work to
3480 * Support for recent Linux distros (Ubuntu 10.10 and Fedora 14), along
3481 with support for recent releases of the underlying toolchain
3482 components, notably gcc-4.5 and glibc-2.12.
3484 * Support for Mac OS X 10.6, both 32- and 64-bit executables. 64-bit
3485 support also works much better on OS X 10.5, and is as solid as
3488 * Support for the SSE4.2 instruction set. SSE4.2 is supported in
3489 64-bit mode. In 32-bit mode, support is only available up to and
3490 including SSSE3. Some exceptions: SSE4.2 AES instructions are not
3491 supported in 64-bit mode, and 32-bit mode does in fact support the
3492 bare minimum SSE4 instructions to needed to run programs on Mac OS X
3493 10.6 on 32-bit targets.
3495 * Support for IBM POWER6 cpus has been improved. The Power ISA up to
3496 and including version 2.05 is supported.
3498 * ==================== TOOL CHANGES ====================
3500 * Cachegrind has a new processing script, cg_diff, which finds the
3501 difference between two profiles. It's very useful for evaluating
3502 the performance effects of a change in a program.
3504 Related to this change, the meaning of cg_annotate's (rarely-used)
3505 --threshold option has changed; this is unlikely to affect many
3506 people, if you do use it please see the user manual for details.
3508 * Callgrind now can do branch prediction simulation, similar to
3509 Cachegrind. In addition, it optionally can count the number of
3510 executed global bus events. Both can be used for a better
3511 approximation of a "Cycle Estimation" as derived event (you need to
3512 update the event formula in KCachegrind yourself).
3514 * Cachegrind and Callgrind now refer to the LL (last-level) cache
3515 rather than the L2 cache. This is to accommodate machines with
3516 three levels of caches -- if Cachegrind/Callgrind auto-detects the
3517 cache configuration of such a machine it will run the simulation as
3518 if the L2 cache isn't present. This means the results are less
3519 likely to match the true result for the machine, but
3520 Cachegrind/Callgrind's results are already only approximate, and
3521 should not be considered authoritative. The results are still
3522 useful for giving a general idea about a program's locality.
3524 * Massif has a new option, --pages-as-heap, which is disabled by
3525 default. When enabled, instead of tracking allocations at the level
3526 of heap blocks (as allocated with malloc/new/new[]), it instead
3527 tracks memory allocations at the level of memory pages (as mapped by
3528 mmap, brk, etc). Each mapped page is treated as its own block.
3529 Interpreting the page-level output is harder than the heap-level
3530 output, but this option is useful if you want to account for every
3531 byte of memory used by a program.
3533 * DRD has two new command-line options: --free-is-write and
3534 --trace-alloc. The former allows to detect reading from already freed
3535 memory, and the latter allows tracing of all memory allocations and
3538 * DRD has several new annotations. Custom barrier implementations can
3539 now be annotated, as well as benign races on static variables.
3541 * DRD's happens before / happens after annotations have been made more
3542 powerful, so that they can now also be used to annotate e.g. a smart
3543 pointer implementation.
3545 * Helgrind's annotation set has also been drastically improved, so as
3546 to provide to users a general set of annotations to describe locks,
3547 semaphores, barriers and condition variables. Annotations to
3548 describe thread-safe reference counted heap objects have also been
3551 * Memcheck has a new command-line option, --show-possibly-lost, which
3552 is enabled by default. When disabled, the leak detector will not
3553 show possibly-lost blocks.
3555 * A new experimental heap profiler, DHAT (Dynamic Heap Analysis Tool),
3556 has been added. DHAT keeps track of allocated heap blocks, and also
3557 inspects every memory reference to see which block (if any) is being
3558 accessed. This gives a lot of insight into block lifetimes,
3559 utilisation, turnover, liveness, and the location of hot and cold
3560 fields. You can use DHAT to do hot-field profiling.
3562 * ==================== OTHER CHANGES ====================
3564 * Improved support for unfriendly self-modifying code: the extra
3565 overhead incurred by --smc-check=all has been reduced by
3566 approximately a factor of 5 as compared with 3.5.0.
3568 * Ability to show directory names for source files in error messages.
3569 This is combined with a flexible mechanism for specifying which
3570 parts of the paths should be shown. This is enabled by the new flag
3573 * A new flag, --require-text-symbol, which will stop the run if a
3574 specified symbol is not found it a given shared object when it is
3575 loaded into the process. This makes advanced working with function
3576 intercepting and wrapping safer and more reliable.
3578 * Improved support for the Valkyrie GUI, version 2.0.0. GUI output
3579 and control of Valgrind is now available for the tools Memcheck and
3580 Helgrind. XML output from Valgrind is available for Memcheck,
3581 Helgrind and exp-Ptrcheck.
3583 * More reliable stack unwinding on amd64-linux, particularly in the
3584 presence of function wrappers, and with gcc-4.5 compiled code.
3586 * Modest scalability (performance improvements) for massive
3587 long-running applications, particularly for those with huge amounts
3590 * Support for analyzing programs running under Wine with has been
3591 improved. The header files <valgrind/valgrind.h>,
3592 <valgrind/memcheck.h> and <valgrind/drd.h> can now be used in
3593 Windows-programs compiled with MinGW or one of the Microsoft Visual
3596 * A rare but serious error in the 64-bit x86 CPU simulation was fixed.
3597 The 32-bit simulator was not affected. This did not occur often,
3598 but when it did would usually crash the program under test.
3601 * A large number of bugs were fixed. These are shown below.
3603 * A number of bugs were investigated, and were candidates for fixing,
3604 but are not fixed in 3.6.0, due to lack of developer time. They may
3605 get fixed in later releases. They are:
3607 194402 vex amd64->IR: 0x48 0xF 0xAE 0x4 0x24 0x49 (FXSAVE64)
3608 212419 false positive "lock order violated" (A+B vs A)
3609 213685 Undefined value propagates past dependency breaking instruction
3610 216837 Incorrect instrumentation of NSOperationQueue on Darwin
3611 237920 valgrind segfault on fork failure
3612 242137 support for code compiled by LLVM-2.8
3613 242423 Another unknown Intel cache config value
3614 243232 Inconsistent Lock Orderings report with trylock
3615 243483 ppc: callgrind triggers VEX assertion failure
3616 243935 Helgrind: implementation of ANNOTATE_HAPPENS_BEFORE() is wrong
3617 244677 Helgrind crash hg_main.c:616 (map_threads_lookup): Assertion
3619 246152 callgrind internal error after pthread_cancel on 32 Bit Linux
3620 249435 Analyzing wine programs with callgrind triggers a crash
3621 250038 ppc64: Altivec lvsr and lvsl instructions fail their regtest
3622 250065 Handling large allocations
3623 250101 huge "free" memory usage due to m_mallocfree.c
3624 "superblocks fragmentation"
3625 251569 vex amd64->IR: 0xF 0x1 0xF9 0x8B 0x4C 0x24 (RDTSCP)
3626 252091 Callgrind on ARM does not detect function returns correctly
3627 252600 [PATCH] Allow lhs to be a pointer for shl/shr
3628 254420 memory pool tracking broken
3629 n-i-bz support for adding symbols for JIT generated code
3632 The following bugs have been fixed or resolved. Note that "n-i-bz"
3633 stands for "not in bugzilla" -- that is, a bug that was reported to us
3634 but never got a bugzilla entry. We encourage you to file bugs in
3635 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
3636 mailing the developers (or mailing lists) directly -- bugs that are
3637 not entered into bugzilla tend to get forgotten about or ignored.
3639 To see details of a given bug, visit
3640 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
3641 where XXXXXX is the bug number as listed below.
3643 135264 dcbzl instruction missing
3645 153699 Valgrind should report unaligned reads with movdqa
3647 190429 Valgrind reports lost of errors in ld.so
3648 with x86_64 2.9.90 glibc
3649 197266 valgrind appears to choke on the xmms instruction
3651 197988 Crash when demangling very large symbol names
3652 202315 unhandled syscall: 332 (inotify_init1)
3653 203256 Add page-level profiling to Massif
3654 205093 dsymutil=yes needs quotes, locking (partial fix)
3655 205241 Snow Leopard 10.6 support (partial fix)
3656 206600 Leak checker fails to upgrade indirect blocks when their
3657 parent becomes reachable
3658 210935 port valgrind.h (not valgrind) to win32 so apps run under
3659 wine can make client requests
3660 211410 vex amd64->IR: 0x15 0xFF 0xFF 0x0 0x0 0x89
3661 within Linux ip-stack checksum functions
3662 212335 unhandled instruction bytes: 0xF3 0xF 0xBD 0xC0
3664 213685 Undefined value propagates past dependency breaking instruction
3666 215914 Valgrind inserts bogus empty environment variable
3668 219538 adjtimex syscall wrapper wrong in readonly adjtime mode
3669 222545 shmat fails under valgind on some arm targets
3670 222560 ARM NEON support
3673 232509 Docs build fails with formatting inside <title></title> elements
3675 235642 [PATCH] syswrap-linux.c: support evdev EVIOCG* ioctls
3676 236546 vex x86->IR: 0x66 0xF 0x3A 0xA
3677 237202 vex amd64->IR: 0xF3 0xF 0xB8 0xC0 0x49 0x3B
3678 237371 better support for VALGRIND_MALLOCLIKE_BLOCK
3679 237485 symlink (syscall 57) is not supported on Mac OS
3680 237723 sysno == 101 exp-ptrcheck: the 'impossible' happened:
3682 238208 is_just_below_ESP doesn't take into account red-zone
3683 238345 valgrind passes wrong $0 when executing a shell script
3684 238679 mq_timedreceive syscall doesn't flag the reception buffer
3686 238696 fcntl command F_DUPFD_CLOEXEC not supported
3687 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3688 238713 unhandled instruction bytes: 0x66 0xF 0x29 0xC6
3689 238745 3.5.0 Make fails on PPC Altivec opcodes, though configure
3691 239992 vex amd64->IR: 0x48 0xF 0xC4 0xC1 0x0 0x48
3697 242606 unhandled syscall: setegid (in Ptrcheck)
3698 242814 Helgrind "Impossible has happened" during
3699 QApplication::initInstance();
3700 243064 Valgrind attempting to read debug information from iso
3701 243270 Make stack unwinding in Valgrind wrappers more reliable
3702 243884 exp-ptrcheck: the 'impossible happened: unhandled syscall
3703 sysno = 277 (mq_open)
3704 244009 exp-ptrcheck unknown syscalls in analyzing lighttpd
3705 244493 ARM VFP d16-d31 registers support
3706 244670 add support for audit_session_self syscall on Mac OS 10.6
3707 244921 The xml report of helgrind tool is not well format
3708 244923 In the xml report file, the <preamble> not escape the
3709 xml char, eg '<','&','>'
3710 245535 print full path names in plain text reports
3711 245925 x86-64 red zone handling problem
3712 246258 Valgrind not catching integer underruns + new [] s
3713 246311 reg/reg cmpxchg doesn't work on amd64
3714 246549 unhandled syscall unix:277 while testing 32-bit Darwin app
3715 246888 Improve Makefile.vex.am
3716 247510 [OS X 10.6] Memcheck reports unaddressable bytes passed
3717 to [f]chmod_extended
3718 247526 IBM POWER6 (ISA 2.05) support is incomplete
3719 247561 Some leak testcases fails due to reachable addresses in
3721 247875 sizeofIRType to handle Ity_I128
3722 247894 [PATCH] unhandled syscall sys_readahead
3723 247980 Doesn't honor CFLAGS passed to configure
3724 248373 darwin10.supp is empty in the trunk
3725 248822 Linux FIBMAP ioctl has int parameter instead of long
3726 248893 [PATCH] make readdwarf.c big endianess safe to enable
3727 unwinding on big endian systems
3728 249224 Syscall 336 not supported (SYS_proc_info)
3730 249775 Incorrect scheme for detecting NEON capabilities of host CPU
3731 249943 jni JVM init fails when using valgrind
3732 249991 Valgrind incorrectly declares AESKEYGENASSIST support
3734 249996 linux/arm: unhandled syscall: 181 (__NR_pwrite64)
3735 250799 frexp$fenv_access_off function generates SIGILL
3736 250998 vex x86->IR: unhandled instruction bytes: 0x66 0x66 0x66 0x2E
3737 251251 support pclmulqdq insn
3738 251362 valgrind: ARM: attach to debugger either fails or provokes
3740 251674 Unhandled syscall 294
3743 254257 Add support for debugfiles found by build-id
3744 254550 [PATCH] Implement DW_ATE_UTF (DWARF4)
3745 254646 Wrapped functions cause stack misalignment on OS X
3746 (and possibly Linux)
3747 254556 ARM: valgrinding anything fails with SIGSEGV for 0xFFFF0FA0
3749 (3.6.0: 21 October 2010, vex r2068, valgrind r11471).
3753 Release 3.5.0 (19 August 2009)
3754 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3755 3.5.0 is a feature release with many significant improvements and the
3756 usual collection of bug fixes. The main improvement is that Valgrind
3757 now works on Mac OS X.
3759 This release supports X86/Linux, AMD64/Linux, PPC32/Linux, PPC64/Linux
3760 and X86/Darwin. Support for recent distros and toolchain components
3761 (glibc 2.10, gcc 4.5) has been added.
3763 -------------------------
3765 Here is a short summary of the changes. Details are shown further
3768 * Support for Mac OS X (10.5.x).
3770 * Improvements and simplifications to Memcheck's leak checker.
3772 * Clarification and simplifications in various aspects of Valgrind's
3775 * XML output for Helgrind and Ptrcheck.
3777 * Performance and stability improvements for Helgrind and DRD.
3779 * Genuinely atomic support for x86/amd64/ppc atomic instructions.
3781 * A new experimental tool, BBV, useful for computer architecture
3784 * Improved Wine support, including ability to read Windows PDB
3787 -------------------------
3789 Here are details of the above changes, followed by descriptions of
3790 many other minor changes, and a list of fixed bugs.
3793 * Valgrind now runs on Mac OS X. (Note that Mac OS X is sometimes
3794 called "Darwin" because that is the name of the OS core, which is the
3795 level that Valgrind works at.)
3799 - It requires OS 10.5.x (Leopard). Porting to 10.4.x is not planned
3800 because it would require work and 10.4 is only becoming less common.
3802 - 32-bit programs on x86 and AMD64 (a.k.a x86-64) machines are supported
3803 fairly well. For 10.5.x, 32-bit programs are the default even on
3804 64-bit machines, so it handles most current programs.
3806 - 64-bit programs on x86 and AMD64 (a.k.a x86-64) machines are not
3807 officially supported, but simple programs at least will probably work.
3808 However, start-up is slow.
3810 - PowerPC machines are not supported.
3812 Things that don't work:
3814 - The Ptrcheck tool.
3816 - Objective-C garbage collection.
3820 - If you have Rogue Amoeba's "Instant Hijack" program installed,
3821 Valgrind will fail with a SIGTRAP at start-up. See
3822 https://bugs.kde.org/show_bug.cgi?id=193917 for details and a
3827 - You will likely find --dsymutil=yes a useful option, as error
3828 messages may be imprecise without it.
3830 - Mac OS X support is new and therefore will be less robust than the
3831 Linux support. Please report any bugs you find.
3833 - Threaded programs may run more slowly than on Linux.
3835 Many thanks to Greg Parker for developing this port over several years.
3838 * Memcheck's leak checker has been improved.
3840 - The results for --leak-check=summary now match the summary results
3841 for --leak-check=full. Previously they could differ because
3842 --leak-check=summary counted "indirectly lost" blocks and
3843 "suppressed" blocks as "definitely lost".
3845 - Blocks that are only reachable via at least one interior-pointer,
3846 but are directly pointed to by a start-pointer, were previously
3847 marked as "still reachable". They are now correctly marked as
3850 - The default value for the --leak-resolution option has been
3851 changed from "low" to "high". In general, this means that more
3852 leak reports will be produced, but each leak report will describe
3853 fewer leaked blocks.
3855 - With --leak-check=full, "definitely lost" and "possibly lost"
3856 leaks are now considered as proper errors, ie. they are counted
3857 for the "ERROR SUMMARY" and affect the behaviour of
3858 --error-exitcode. These leaks are not counted as errors if
3859 --leak-check=summary is specified, however.
3861 - Documentation for the leak checker has been improved.
3864 * Various aspects of Valgrind's text output have changed.
3866 - Valgrind's start-up message has changed. It is shorter but also
3867 includes the command being run, which makes it easier to use
3868 --trace-children=yes. An example:
3870 - Valgrind's shut-down messages have also changed. This is most
3871 noticeable with Memcheck, where the leak summary now occurs before
3872 the error summary. This change was necessary to allow leaks to be
3873 counted as proper errors (see the description of the leak checker
3874 changes above for more details). This was also necessary to fix a
3875 longstanding bug in which uses of suppressions against leaks were
3876 not "counted", leading to difficulties in maintaining suppression
3877 files (see https://bugs.kde.org/show_bug.cgi?id=186790).
3879 - Behavior of -v has changed. In previous versions, -v printed out
3880 a mixture of marginally-user-useful information, and tool/core
3881 statistics. The statistics printing has now been moved to its own
3882 flag, --stats=yes. This means -v is less verbose and more likely
3883 to convey useful end-user information.
3885 - The format of some (non-XML) stack trace entries has changed a
3886 little. Previously there were six possible forms:
3888 0x80483BF: really (a.c:20)
3889 0x80483BF: really (in /foo/a.out)
3891 0x80483BF: (within /foo/a.out)
3892 0x80483BF: ??? (a.c:20)
3895 The third and fourth of these forms have been made more consistent
3896 with the others. The six possible forms are now:
3898 0x80483BF: really (a.c:20)
3899 0x80483BF: really (in /foo/a.out)
3900 0x80483BF: really (in ???)
3901 0x80483BF: ??? (in /foo/a.out)
3902 0x80483BF: ??? (a.c:20)
3905 Stack traces produced when --xml=yes is specified are different
3909 * Helgrind and Ptrcheck now support XML output, so they can be used
3910 from GUI tools. Also, the XML output mechanism has been
3913 - The XML format has been overhauled and generalised, so it is more
3914 suitable for error reporting tools in general. The Memcheck
3915 specific aspects of it have been removed. The new format, which
3916 is an evolution of the old format, is described in
3917 docs/internals/xml-output-protocol4.txt.
3919 - Memcheck has been updated to use the new format.
3921 - Helgrind and Ptrcheck are now able to emit output in this format.
3923 - The XML output mechanism has been overhauled. XML is now output
3924 to its own file descriptor, which means that:
3926 * Valgrind can output text and XML independently.
3928 * The longstanding problem of XML output being corrupted by
3929 unexpected un-tagged text messages is solved.
3931 As before, the destination for text output is specified using
3932 --log-file=, --log-fd= or --log-socket=.
3934 As before, XML output for a tool is enabled using --xml=yes.
3936 Because there's a new XML output channel, the XML output
3937 destination is now specified by --xml-file=, --xml-fd= or
3940 Initial feedback has shown this causes some confusion. To
3941 clarify, the two envisaged usage scenarios are:
3943 (1) Normal text output. In this case, do not specify --xml=yes
3944 nor any of --xml-file=, --xml-fd= or --xml-socket=.
3946 (2) XML output. In this case, specify --xml=yes, and one of
3947 --xml-file=, --xml-fd= or --xml-socket= to select the XML
3948 destination, one of --log-file=, --log-fd= or --log-socket=
3949 to select the destination for any remaining text messages,
3950 and, importantly, -q.
3952 -q makes Valgrind completely silent on the text channel,
3953 except in the case of critical failures, such as Valgrind
3954 itself segfaulting, or failing to read debugging information.
3955 Hence, in this scenario, it suffices to check whether or not
3956 any output appeared on the text channel. If yes, then it is
3957 likely to be a critical error which should be brought to the
3958 attention of the user. If no (the text channel produced no
3959 output) then it can be assumed that the run was successful.
3961 This allows GUIs to make the critical distinction they need to
3962 make (did the run fail or not?) without having to search or
3963 filter the text output channel in any way.
3965 It is also recommended to use --child-silent-after-fork=yes in
3969 * Improvements and changes in Helgrind:
3971 - XML output, as described above
3973 - Checks for consistent association between pthread condition
3974 variables and their associated mutexes are now performed.
3976 - pthread_spinlock functions are supported.
3978 - Modest performance improvements.
3980 - Initial (skeletal) support for describing the behaviour of
3981 non-POSIX synchronisation objects through ThreadSanitizer
3982 compatible ANNOTATE_* macros.
3984 - More controllable tradeoffs between performance and the level of
3985 detail of "previous" accesses in a race. There are now three
3988 * --history-level=full. This is the default, and was also the
3989 default in 3.4.x. It shows both stacks involved in a race, but
3990 requires a lot of memory and can be very slow in programs that
3991 do many inter-thread synchronisation events.
3993 * --history-level=none. This only shows the later stack involved
3994 in a race. This can be much faster than --history-level=full,
3995 but makes it much more difficult to find the other access
3996 involved in the race.
3998 The new intermediate setting is
4000 * --history-level=approx
4002 For the earlier (other) access, two stacks are presented. The
4003 earlier access is guaranteed to be somewhere in between the two
4004 program points denoted by those stacks. This is not as useful
4005 as showing the exact stack for the previous access (as per
4006 --history-level=full), but it is better than nothing, and it's
4007 almost as fast as --history-level=none.
4010 * New features and improvements in DRD:
4012 - The error messages printed by DRD are now easier to interpret.
4013 Instead of using two different numbers to identify each thread
4014 (Valgrind thread ID and DRD thread ID), DRD does now identify
4015 threads via a single number (the DRD thread ID). Furthermore
4016 "first observed at" information is now printed for all error
4017 messages related to synchronization objects.
4019 - Added support for named semaphores (sem_open() and sem_close()).
4021 - Race conditions between pthread_barrier_wait() and
4022 pthread_barrier_destroy() calls are now reported.
4024 - Added support for custom allocators through the macros
4025 VALGRIND_MALLOCLIKE_BLOCK() VALGRIND_FREELIKE_BLOCK() (defined in
4026 in <valgrind/valgrind.h>). An alternative for these two macros is
4027 the new client request VG_USERREQ__DRD_CLEAN_MEMORY (defined in
4030 - Added support for annotating non-POSIX synchronization objects
4031 through several new ANNOTATE_*() macros.
4033 - OpenMP: added support for the OpenMP runtime (libgomp) included
4034 with gcc versions 4.3.0 and 4.4.0.
4038 - Added two new command-line options (--first-race-only and
4039 --segment-merging-interval).
4042 * Genuinely atomic support for x86/amd64/ppc atomic instructions
4044 Valgrind will now preserve (memory-access) atomicity of LOCK-
4045 prefixed x86/amd64 instructions, and any others implying a global
4046 bus lock. Ditto for PowerPC l{w,d}arx/st{w,d}cx. instructions.
4048 This means that Valgrinded processes will "play nicely" in
4049 situations where communication with other processes, or the kernel,
4050 is done through shared memory and coordinated with such atomic
4051 instructions. Prior to this change, such arrangements usually
4052 resulted in hangs, races or other synchronisation failures, because
4053 Valgrind did not honour atomicity of such instructions.
4056 * A new experimental tool, BBV, has been added. BBV generates basic
4057 block vectors for use with the SimPoint analysis tool, which allows
4058 a program's overall behaviour to be approximated by running only a
4059 fraction of it. This is useful for computer architecture
4060 researchers. You can run BBV by specifying --tool=exp-bbv (the
4061 "exp-" prefix is short for "experimental"). BBV was written by
4065 * Modestly improved support for running Windows applications under
4066 Wine. In particular, initial support for reading Windows .PDB debug
4067 information has been added.
4070 * A new Memcheck client request VALGRIND_COUNT_LEAK_BLOCKS has been
4071 added. It is similar to VALGRIND_COUNT_LEAKS but counts blocks
4075 * The Valgrind client requests VALGRIND_PRINTF and
4076 VALGRIND_PRINTF_BACKTRACE have been changed slightly. Previously,
4077 the string was always printed immediately on its own line. Now, the
4078 string will be added to a buffer but not printed until a newline is
4079 encountered, or other Valgrind output is printed (note that for
4080 VALGRIND_PRINTF_BACKTRACE, the back-trace itself is considered
4081 "other Valgrind output"). This allows you to use multiple
4082 VALGRIND_PRINTF calls to build up a single output line, and also to
4083 print multiple output lines with a single request (by embedding
4084 multiple newlines in the string).
4087 * The graphs drawn by Massif's ms_print program have changed slightly:
4089 - The half-height chars '.' and ',' are no longer drawn, because
4090 they are confusing. The --y option can be used if the default
4091 y-resolution is not high enough.
4093 - Horizontal lines are now drawn after the top of a snapshot if
4094 there is a gap until the next snapshot. This makes it clear that
4095 the memory usage has not dropped to zero between snapshots.
4098 * Something that happened in 3.4.0, but wasn't clearly announced: the
4099 option --read-var-info=yes can be used by some tools (Memcheck,
4100 Helgrind and DRD). When enabled, it causes Valgrind to read DWARF3
4101 variable type and location information. This makes those tools
4102 start up more slowly and increases memory consumption, but
4103 descriptions of data addresses in error messages become more
4107 * exp-Omega, an experimental instantaneous leak-detecting tool, was
4108 disabled in 3.4.0 due to a lack of interest and maintenance,
4109 although the source code was still in the distribution. The source
4110 code has now been removed from the distribution. For anyone
4111 interested, the removal occurred in SVN revision r10247.
4114 * Some changes have been made to the build system.
4116 - VEX/ is now integrated properly into the build system. This means
4117 that dependency tracking within VEX/ now works properly, "make
4118 install" will work without requiring "make" before it, and
4119 parallel builds (ie. 'make -j') now work (previously a
4120 .NOTPARALLEL directive was used to serialize builds, ie. 'make -j'
4121 was effectively ignored).
4123 - The --with-vex configure option has been removed. It was of
4124 little use and removing it simplified the build system.
4126 - The location of some install files has changed. This should not
4127 affect most users. Those who might be affected:
4129 * For people who use Valgrind with MPI programs, the installed
4130 libmpiwrap.so library has moved from
4131 $(INSTALL)/<platform>/libmpiwrap.so to
4132 $(INSTALL)/libmpiwrap-<platform>.so.
4134 * For people who distribute standalone Valgrind tools, the
4135 installed libraries such as $(INSTALL)/<platform>/libcoregrind.a
4136 have moved to $(INSTALL)/libcoregrind-<platform>.a.
4138 These changes simplify the build system.
4140 - Previously, all the distributed suppression (*.supp) files were
4141 installed. Now, only default.supp is installed. This should not
4142 affect users as the other installed suppression files were not
4143 read; the fact that they were installed was a mistake.
4146 * KNOWN LIMITATIONS:
4148 - Memcheck is unusable with the Intel compiler suite version 11.1,
4149 when it generates code for SSE2-and-above capable targets. This
4150 is because of icc's use of highly optimised inlined strlen
4151 implementations. It causes Memcheck to report huge numbers of
4152 false errors even in simple programs. Helgrind and DRD may also
4155 Versions 11.0 and earlier may be OK, but this has not been
4159 The following bugs have been fixed or resolved. Note that "n-i-bz"
4160 stands for "not in bugzilla" -- that is, a bug that was reported to us
4161 but never got a bugzilla entry. We encourage you to file bugs in
4162 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4163 mailing the developers (or mailing lists) directly -- bugs that are
4164 not entered into bugzilla tend to get forgotten about or ignored.
4166 To see details of a given bug, visit
4167 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
4168 where XXXXXX is the bug number as listed below.
4170 84303 How about a LockCheck tool?
4171 91633 dereference of null ptr in vgPlain_st_basetype
4172 97452 Valgrind doesn't report any pthreads problems
4173 100628 leak-check gets assertion failure when using
4174 VALGRIND_MALLOCLIKE_BLOCK on malloc()ed memory
4175 108528 NPTL pthread cleanup handlers not called
4176 110126 Valgrind 2.4.1 configure.in tramples CFLAGS
4177 110128 mallinfo is not implemented...
4178 110770 VEX: Generated files not always updated when making valgrind
4179 111102 Memcheck: problems with large (memory footprint) applications
4180 115673 Vex's decoder should never assert
4181 117564 False positive: Syscall param clone(child_tidptr) contains
4182 uninitialised byte(s)
4183 119404 executing ssh from inside valgrind fails
4184 133679 Callgrind does not write path names to sources with dwarf debug
4186 135847 configure.in problem with non gnu compilers (and possible fix)
4187 136154 threads.c:273 (vgCallgrind_post_signal): Assertion
4188 '*(vgCallgrind_current_fn_stack.top) == 0' failed.
4189 136230 memcheck reports "possibly lost", should be "still reachable"
4190 137073 NULL arg to MALLOCLIKE_BLOCK causes crash
4191 137904 Valgrind reports a memory leak when using POSIX threads,
4193 139076 valgrind VT_GETSTATE error
4194 142228 complaint of elf_dynamic_do_rela in trivial usage
4195 145347 spurious warning with USBDEVFS_REAPURB
4196 148441 (wine) can't find memory leak in Wine, win32 binary
4198 148742 Leak-check fails assert on exit
4199 149878 add (proper) check for calloc integer overflow
4200 150606 Call graph is broken when using callgrind control
4201 152393 leak errors produce an exit code of 0. I need some way to
4202 cause leak errors to result in a nonzero exit code.
4203 157154 documentation (leak-resolution doc speaks about num-callers
4204 def=4) + what is a loss record
4205 159501 incorrect handling of ALSA ioctls
4206 162020 Valgrinding an empty/zero-byte file crashes valgrind
4207 162482 ppc: Valgrind crashes while reading stabs information
4208 162718 x86: avoid segment selector 0 in sys_set_thread_area()
4209 163253 (wine) canonicaliseSymtab forgot some fields in DiSym
4210 163560 VEX/test_main.c is missing from valgrind-3.3.1
4211 164353 malloc_usable_size() doesn't return a usable size
4212 165468 Inconsistent formatting in memcheck manual -- please fix
4213 169505 main.c:286 (endOfInstr):
4214 Assertion 'ii->cost_offset == *cost_offset' failed
4215 177206 Generate default.supp during compile instead of configure
4216 177209 Configure valt_load_address based on arch+os
4217 177305 eventfd / syscall 323 patch lost
4218 179731 Tests fail to build because of inlining of non-local asm labels
4219 181394 helgrind: libhb_core.c:3762 (msm_write): Assertion
4220 'ordxx == POrd_EQ || ordxx == POrd_LT' failed.
4221 181594 Bogus warning for empty text segment
4222 181707 dwarf doesn't require enumerations to have name
4223 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4224 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4225 Assertion '!already_present' failed.
4226 185359 exp-ptrcheck: unhandled syscall getresuid()
4227 185794 "WARNING: unhandled syscall: 285" (fallocate) on x86_64
4228 185816 Valgrind is unable to handle debug info for files with split
4229 debug info that are prelinked afterwards
4230 185980 [darwin] unhandled syscall: sem_open
4231 186238 bbToIR_AMD64: disInstr miscalculated next %rip
4232 186507 exp-ptrcheck unhandled syscalls prctl, etc.
4233 186790 Suppression pattern used for leaks are not reported
4234 186796 Symbols with length>200 in suppression files are ignored
4235 187048 drd: mutex PTHREAD_PROCESS_SHARED attribute missinterpretation
4236 187416 exp-ptrcheck: support for __NR_{setregid,setreuid,setresuid}
4237 188038 helgrind: hg_main.c:926: mk_SHVAL_fail: the 'impossible' happened
4238 188046 bashisms in the configure script
4239 188127 amd64->IR: unhandled instruction bytes: 0xF0 0xF 0xB0 0xA
4240 188161 memcheck: --track-origins=yes asserts "mc_machine.c:672
4241 (get_otrack_shadow_offset_wrk): the 'impossible' happened."
4242 188248 helgrind: pthread_cleanup_push, pthread_rwlock_unlock,
4243 assertion fail "!lock->heldBy"
4244 188427 Add support for epoll_create1 (with patch)
4245 188530 Support for SIOCGSTAMPNS
4246 188560 Include valgrind.spec in the tarball
4247 188572 Valgrind on Mac should suppress setenv() mem leak
4248 189054 Valgrind fails to build because of duplicate non-local asm labels
4249 189737 vex amd64->IR: unhandled instruction bytes: 0xAC
4250 189762 epoll_create syscall not handled (--tool=exp-ptrcheck)
4251 189763 drd assertion failure: s_threadinfo[tid].is_recording
4252 190219 unhandled syscall: 328 (x86-linux)
4253 190391 dup of 181394; see above
4254 190429 Valgrind reports lots of errors in ld.so with x86_64 2.9.90 glibc
4255 190820 No debug information on powerpc-linux
4256 191095 PATCH: Improve usbdevfs ioctl handling
4257 191182 memcheck: VALGRIND_LEAK_CHECK quadratic when big nr of chunks
4259 191189 --xml=yes should obey --gen-suppressions=all
4260 191192 syslog() needs a suppression on macosx
4261 191271 DARWIN: WARNING: unhandled syscall: 33554697 a.k.a.: 265
4262 191761 getrlimit on MacOSX
4263 191992 multiple --fn-skip only works sometimes; dependent on order
4264 192634 V. reports "aspacem sync_check_mapping_callback:
4265 segment mismatch" on Darwin
4266 192954 __extension__ missing on 2 client requests
4267 194429 Crash at start-up with glibc-2.10.1 and linux-2.6.29
4268 194474 "INSTALL" file has different build instructions than "README"
4269 194671 Unhandled syscall (sem_wait?) from mac valgrind
4270 195069 memcheck: reports leak (memory still reachable) for
4272 195169 drd: (vgDrd_barrier_post_wait):
4273 Assertion 'r->sg[p->post_iteration]' failed.
4274 195268 valgrind --log-file doesn't accept ~/...
4275 195838 VEX abort: LibVEX_N_SPILL_BYTES too small for CPUID boilerplate
4276 195860 WARNING: unhandled syscall: unix:223
4277 196528 need a error suppression for pthread_rwlock_init under os x?
4278 197227 Support aio_* syscalls on Darwin
4279 197456 valgrind should reject --suppressions=(directory)
4280 197512 DWARF2 CFI reader: unhandled CFI instruction 0:10
4281 197591 unhandled syscall 27 (mincore)
4282 197793 Merge DCAS branch to the trunk == 85756, 142103
4283 197794 Avoid duplicate filenames in Vex
4284 197898 make check fails on current SVN
4285 197901 make check fails also under exp-ptrcheck in current SVN
4286 197929 Make --leak-resolution=high the default
4287 197930 Reduce spacing between leak reports
4288 197933 Print command line of client at start-up, and shorten preamble
4289 197966 unhandled syscall 205 (x86-linux, --tool=exp-ptrcheck)
4290 198395 add BBV to the distribution as an experimental tool
4291 198624 Missing syscalls on Darwin: 82, 167, 281, 347
4292 198649 callgrind_annotate doesn't cumulate counters
4293 199338 callgrind_annotate sorting/thresholds are broken for all but Ir
4294 199977 Valgrind complains about an unrecognized instruction in the
4295 atomic_incs test program
4296 200029 valgrind isn't able to read Fedora 12 debuginfo
4297 200760 darwin unhandled syscall: unix:284
4298 200827 DRD doesn't work on Mac OS X
4299 200990 VG_(read_millisecond_timer)() does not work correctly
4300 201016 Valgrind does not support pthread_kill() on Mac OS
4301 201169 Document --read-var-info
4302 201323 Pre-3.5.0 performance sanity checking
4303 201384 Review user manual for the 3.5.0 release
4304 201585 mfpvr not implemented on ppc
4305 201708 tests failing because x86 direction flag is left set
4306 201757 Valgrind doesn't handle any recent sys_futex additions
4307 204377 64-bit valgrind can not start a shell script
4308 (with #!/path/to/shell) if the shell is a 32-bit executable
4309 n-i-bz drd: fixed assertion failure triggered by mutex reinitialization.
4310 n-i-bz drd: fixed a bug that caused incorrect messages to be printed
4311 about memory allocation events with memory access tracing enabled
4312 n-i-bz drd: fixed a memory leak triggered by vector clock deallocation
4314 (3.5.0: 19 Aug 2009, vex r1913, valgrind r10846).
4318 Release 3.4.1 (28 February 2009)
4319 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4320 3.4.1 is a bug-fix release that fixes some regressions and assertion
4321 failures in debug info reading in 3.4.0, most notably incorrect stack
4322 traces on amd64-linux on older (glibc-2.3 based) systems. Various
4323 other debug info problems are also fixed. A number of bugs in the
4324 exp-ptrcheck tool introduced in 3.4.0 have been fixed.
4326 In view of the fact that 3.4.0 contains user-visible regressions
4327 relative to 3.3.x, upgrading to 3.4.1 is recommended. Packagers are
4328 encouraged to ship 3.4.1 in preference to 3.4.0.
4330 The fixed bugs are as follows. Note that "n-i-bz" stands for "not in
4331 bugzilla" -- that is, a bug that was reported to us but never got a
4332 bugzilla entry. We encourage you to file bugs in bugzilla
4333 (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than mailing the
4334 developers (or mailing lists) directly -- bugs that are not entered
4335 into bugzilla tend to get forgotten about or ignored.
4337 n-i-bz Fix various bugs reading icc-11 generated debug info
4338 n-i-bz Fix various bugs reading gcc-4.4 generated debug info
4339 n-i-bz Preliminary support for glibc-2.10 / Fedora 11
4340 n-i-bz Cachegrind and Callgrind: handle non-power-of-two cache sizes,
4341 so as to support (eg) 24k Atom D1 and Core2 with 3/6/12MB L2.
4342 179618 exp-ptrcheck crashed / exit prematurely
4343 179624 helgrind: false positive races with pthread_create and
4344 recv/open/close/read
4345 134207 pkg-config output contains @VG_PLATFORM@
4346 176926 floating point exception at valgrind startup with PPC 440EPX
4347 181594 Bogus warning for empty text segment
4348 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant rex prefixes)
4349 181707 Dwarf3 doesn't require enumerations to have name
4350 185038 exp-ptrcheck: "unhandled syscall: 285" (fallocate) on x86_64
4351 185050 exp-ptrcheck: sg_main.c:727 (add_block_to_GlobalTree):
4352 Assertion '!already_present' failed.
4353 185359 exp-ptrcheck unhandled syscall getresuid()
4355 (3.4.1.RC1: 24 Feb 2008, vex r1884, valgrind r9253).
4356 (3.4.1: 28 Feb 2008, vex r1884, valgrind r9293).
4360 Release 3.4.0 (2 January 2009)
4361 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4362 3.4.0 is a feature release with many significant improvements and the
4363 usual collection of bug fixes. This release supports X86/Linux,
4364 AMD64/Linux, PPC32/Linux and PPC64/Linux. Support for recent distros
4365 (using gcc 4.4, glibc 2.8 and 2.9) has been added.
4367 3.4.0 brings some significant tool improvements. Memcheck can now
4368 report the origin of uninitialised values, the thread checkers
4369 Helgrind and DRD are much improved, and we have a new experimental
4370 tool, exp-Ptrcheck, which is able to detect overruns of stack and
4371 global arrays. In detail:
4373 * Memcheck is now able to track the origin of uninitialised values.
4374 When it reports an uninitialised value error, it will try to show
4375 the origin of the value, as either a heap or stack allocation.
4376 Origin tracking is expensive and so is not enabled by default. To
4377 use it, specify --track-origins=yes. Memcheck's speed will be
4378 essentially halved, and memory usage will be significantly
4379 increased. Nevertheless it can drastically reduce the effort
4380 required to identify the root cause of uninitialised value errors,
4381 and so is often a programmer productivity win, despite running more
4384 * A version (1.4.0) of the Valkyrie GUI, that works with Memcheck in
4385 3.4.0, will be released shortly.
4387 * Helgrind's race detection algorithm has been completely redesigned
4388 and reimplemented, to address usability and scalability concerns:
4390 - The new algorithm has a lower false-error rate: it is much less
4391 likely to report races that do not really exist.
4393 - Helgrind will display full call stacks for both accesses involved
4394 in a race. This makes it easier to identify the root causes of
4397 - Limitations on the size of program that can run have been removed.
4399 - Performance has been modestly improved, although that is very
4402 - Direct support for Qt4 threading has been added.
4404 - pthread_barriers are now directly supported.
4406 - Helgrind works well on all supported Linux targets.
4408 * The DRD thread debugging tool has seen major improvements:
4410 - Greatly improved performance and significantly reduced memory
4413 - Support for several major threading libraries (Boost.Thread, Qt4,
4414 glib, OpenMP) has been added.
4416 - Support for atomic instructions, POSIX semaphores, barriers and
4417 reader-writer locks has been added.
4419 - Works now on PowerPC CPUs too.
4421 - Added support for printing thread stack usage at thread exit time.
4423 - Added support for debugging lock contention.
4425 - Added a manual for Drd.
4427 * A new experimental tool, exp-Ptrcheck, has been added. Ptrcheck
4428 checks for misuses of pointers. In that sense it is a bit like
4429 Memcheck. However, Ptrcheck can do things Memcheck can't: it can
4430 detect overruns of stack and global arrays, it can detect
4431 arbitrarily far out-of-bounds accesses to heap blocks, and it can
4432 detect accesses heap blocks that have been freed a very long time
4433 ago (millions of blocks in the past).
4435 Ptrcheck currently works only on x86-linux and amd64-linux. To use
4436 it, use --tool=exp-ptrcheck. A simple manual is provided, as part
4437 of the main Valgrind documentation. As this is an experimental
4438 tool, we would be particularly interested in hearing about your
4439 experiences with it.
4441 * exp-Omega, an experimental instantaneous leak-detecting tool, is no
4442 longer built by default, although the code remains in the repository
4443 and the tarball. This is due to three factors: a perceived lack of
4444 users, a lack of maintenance, and concerns that it may not be
4445 possible to achieve reliable operation using the existing design.
4447 * As usual, support for the latest Linux distros and toolchain
4448 components has been added. It should work well on Fedora Core 10,
4449 OpenSUSE 11.1 and Ubuntu 8.10. gcc-4.4 (in its current pre-release
4450 state) is supported, as is glibc-2.9. The C++ demangler has been
4451 updated so as to work well with C++ compiled by even the most recent
4454 * You can now use frame-level wildcards in suppressions. This was a
4455 frequently-requested enhancement. A line "..." in a suppression now
4456 matches zero or more frames. This makes it easier to write
4457 suppressions which are precise yet insensitive to changes in
4460 * 3.4.0 adds support on x86/amd64 for the SSSE3 instruction set.
4462 * Very basic support for IBM Power6 has been added (64-bit processes only).
4464 * Valgrind is now cross-compilable. For example, it is possible to
4465 cross compile Valgrind on an x86/amd64-linux host, so that it runs
4466 on a ppc32/64-linux target.
4468 * You can set the main thread's stack size at startup using the
4469 new --main-stacksize= flag (subject of course to ulimit settings).
4470 This is useful for running apps that need a lot of stack space.
4472 * The limitation that you can't use --trace-children=yes together
4473 with --db-attach=yes has been removed.
4475 * The following bugs have been fixed. Note that "n-i-bz" stands for
4476 "not in bugzilla" -- that is, a bug that was reported to us but
4477 never got a bugzilla entry. We encourage you to file bugs in
4478 bugzilla (http://bugs.kde.org/enter_valgrind_bug.cgi) rather than
4479 mailing the developers (or mailing lists) directly.
4481 n-i-bz Make return types for some client requests 64-bit clean
4482 n-i-bz glibc 2.9 support
4483 n-i-bz ignore unsafe .valgrindrc's (CVE-2008-4865)
4484 n-i-bz MPI_Init(0,0) is valid but libmpiwrap.c segfaults
4485 n-i-bz Building in an env without gdb gives bogus gdb attach
4486 92456 Tracing the origin of uninitialised memory
4487 106497 Valgrind does not demangle some C++ template symbols
4489 151612 Suppression with "..." (frame-level wildcards in .supp files)
4490 156404 Unable to start oocalc under memcheck on openSUSE 10.3 (64-bit)
4491 159285 unhandled syscall:25 (stime, on x86-linux)
4492 159452 unhandled ioctl 0x8B01 on "valgrind iwconfig"
4493 160954 ppc build of valgrind crashes with illegal instruction (isel)
4494 160956 mallinfo implementation, w/ patch
4495 162092 Valgrind fails to start gnome-system-monitor
4496 162819 malloc_free_fill test doesn't pass on glibc2.8 x86
4497 163794 assertion failure with "--track-origins=yes"
4498 163933 sigcontext.err and .trapno must be set together
4499 163955 remove constraint !(--db-attach=yes && --trace-children=yes)
4500 164476 Missing kernel module loading system calls
4501 164669 SVN regression: mmap() drops posix file locks
4502 166581 Callgrind output corruption when program forks
4503 167288 Patch file for missing system calls on Cell BE
4504 168943 unsupported scas instruction pentium
4505 171645 Unrecognised instruction (MOVSD, non-binutils encoding)
4506 172417 x86->IR: 0x82 ...
4507 172563 amd64->IR: 0xD9 0xF5 - fprem1
4508 173099 .lds linker script generation error
4509 173177 [x86_64] syscalls: 125/126/179 (capget/capset/quotactl)
4510 173751 amd64->IR: 0x48 0xF 0x6F 0x45 (even more redundant prefixes)
4512 174908 --log-file value not expanded correctly for core file
4513 175044 Add lookup_dcookie for amd64
4514 175150 x86->IR: 0xF2 0xF 0x11 0xC1 (movss non-binutils encoding)
4516 Developer-visible changes:
4518 * Valgrind's debug-info reading machinery has been majorly overhauled.
4519 It can now correctly establish the addresses for ELF data symbols,
4520 which is something that has never worked properly before now.
4522 Also, Valgrind can now read DWARF3 type and location information for
4523 stack and global variables. This makes it possible to use the
4524 framework to build tools that rely on knowing the type and locations
4525 of stack and global variables, for example exp-Ptrcheck.
4527 Reading of such information is disabled by default, because most
4528 tools don't need it, and because it is expensive in space and time.
4529 However, you can force Valgrind to read it, using the
4530 --read-var-info=yes flag. Memcheck, Helgrind and DRD are able to
4531 make use of such information, if present, to provide source-level
4532 descriptions of data addresses in the error messages they create.
4534 (3.4.0.RC1: 24 Dec 2008, vex r1878, valgrind r8882).
4535 (3.4.0: 3 Jan 2009, vex r1878, valgrind r8899).