1 Release 3.23.0 (?? Apr 2024)
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 * --track-fds=yes will now also warn about double closing of file
14 descriptors. Printing the context where the file descriptor was
15 originally opened and where it was previously closed.
17 * ================== PLATFORM CHANGES =================
19 * ==================== TOOL CHANGES ===================
21 * ==================== FIXED BUGS ====================
23 The following bugs have been fixed or resolved. Note that "n-i-bz"
24 stands for "not in bugzilla" -- that is, a bug that was reported to us
25 but never got a bugzilla entry. We encourage you to file bugs in
26 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
27 than mailing the developers (or mailing lists) directly -- bugs that
28 are not entered into bugzilla tend to get forgotten about or ignored.
31 283429 ARM leak checking needs CLEAR_CALLER_SAVED_REGS
32 281059 Cannot connect to Oracle using valgrind
33 369723 __builtin_longjmp not supported in clang/llvm on Android arm64 target
34 390269 unhandled amd64-darwin syscall: unix:464 (openat_nocancel)
35 401284 False positive "Source and destination overlap in strncat"
36 428364 Signals inside io_uring_enter not handled
37 437790 valgrind reports "Conditional jump or move depends on uninitialised
38 value" in memchr of macOS 10.12-10.15
39 460616 disInstr(arm64): unhandled instruction 0x4E819402 (dotprod/ASIMDDP)
40 466762 Add redirs for C23 free_sized() and free_aligned_sized()
41 466884 Missing writev uninit padding suppression for _XSend
42 471036 disInstr_AMD64: disInstr miscalculated next %rip on RORX imm8, m32/64, r32/6
43 471222 support tracking of file descriptors being double closed
44 475498 Add reallocarray wrapper
45 476320 Build failure with GCC
46 476331 clean up generated/distributed filter scripts
47 476535 Difference in allocation size for massif/tests/overloaded-new between
48 clang++/libc++ and g++/libstdc++
49 476548 valgrind 3.22.0 fails on assertion when loading debuginfo file
51 476708 valgrind-monitor.py regular expressions should use raw strings
52 476780 Extend strlcat and strlcpy wrappers to GNU libc
53 476787 Build of Valgrind 3.21.0 fails when SOLARIS_PT_SUNDWTRACE_THRP is
55 476887 WARNING: unhandled amd64-freebsd syscall: 578
56 477198 Add fchmodat2 syscall on linux
57 477628 Add mremap support for Solaris
58 477630 Include ucontext.h rather than sys/ucontext.h in Solaris sources
59 477719 vgdb incorrectly replies to qRcmd packet
60 478211 Redundant code for vgdb.c and Valgrind core tools
61 478624 Valgrind incompatibility with binutils-2.42 on x86 with new nop patterns
62 (unhandled instruction bytes: 0x2E 0x8D 0xB4 0x26
63 478837 valgrind fails to read debug info for rust binaries
64 479041 Executables without RW sections do not trigger debuginfo reading
65 480052 WARNING: unhandled amd64-freebsd syscall: 580
66 480126 Build failure on Raspberry Pi 5 / OS 6.1.0-rpi7-rpi-v8
67 480405 valgrind 3.22.0 "m_debuginfo/image.c:586 (set_CEnt):
68 Assertion '!sr_isError(sr)' failed."
69 480488 Add support for FreeBSD 13.3
70 480706 Unhandled syscall 325 (mlock2)
71 481131 [PATCH] x86 regtest: fix clobber lists in generated asm statements
72 n-i-bz Add redirect for memccpy
74 To see details of a given bug, visit
75 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
76 where XXXXXX is the bug number as listed above.
78 (3.23.0.RC1: ?? Apr 2024)
80 Release 3.22.0 (31 Oct 2023)
81 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
83 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
84 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
85 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
86 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
87 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
88 AMD64/macOS 10.13 and nanoMIPS/Linux.
90 * ==================== CORE CHANGES ===================
92 * A new configure option --with-gdbscripts-dir lets you install
93 the gdb valgrind python monitor scripts in a specific location.
94 For example a distro could use it to install the scripts in a
95 safe load location --with-gdbscripts-dir=%{_datadir}/gdb/auto-load
96 It is also possible to configure --without-gdb-scripts-dir so no
97 .debug_gdb_scripts section is added to the vgpreload library and
98 no valgrind-monitor python scripts are installed at all.
100 * ================== PLATFORM CHANGES =================
102 * Support has been added for FreeBSD 14 and FreeBSD 15.
103 * Add support for the folllowing FreeBSD system calls:
104 close_range, kqueuex, membarrier, timerfd_create,
105 timerfd_settime and timerfd_gettime (all added in FreeBSD 15).
107 * ==================== TOOL CHANGES ===================
109 * Memcheck now tests and warns about the values used for
110 alignment and size. These apply to various functions: memalign,
111 posix_memalign and aligned_alloc in C and various overloads
112 of operators new and delete in C++. The kinds of error that can
114 - invalid alignment, for instance the alignment is usually required
116 - mismatched alignment between aligned allocation and aligned
118 - mismatched size when sized delete is used
119 - bad size for functions that have implementation defined behaviour
120 when the requested size is zero
123 - You can now profile part of a program's execution using the new
124 `CACHEGRIND_START_INSTRUMENTATION` and `CACHEGRIND_STOP_INSTRUMENTATION`
125 client requests, along with the new `--instr-at-start` option. The
126 behaviour is the same as Callgrind's equivalent functionality.
128 * ==================== FIXED BUGS ====================
130 The following bugs have been fixed or resolved. Note that "n-i-bz"
131 stands for "not in bugzilla" -- that is, a bug that was reported to us
132 but never got a bugzilla entry. We encourage you to file bugs in
133 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
134 than mailing the developers (or mailing lists) directly -- bugs that
135 are not entered into bugzilla tend to get forgotten about or ignored.
137 390871 ELF debug info reader confused with multiple .rodata* sections
138 417993 vbit-test fail on s390x with Iop_Add32: spurious dependency on uninit
139 426751 Valgrind reports "still reachable" memory using musl
140 (alpine running inside docker)
141 432801 Valgrind 3.16.1 reports a jump based on uninitialized memory somehow
142 related to clang and signals
143 433857 Add validation to C++17 aligned new/delete alignment size
144 433859 Add mismatched detection to C++ 17 aligned new/delete
145 460192 Add epoll_pwait2
146 461074 DWARF2 CFI reader: unhandled DW_OP_ 0x11 (consts) DW_OP_ 0x92 (bregx)
147 465782 s390x: Valgrind doesn't compile with Clang on s390x
148 466105 aligned_alloc problems, part 2
149 467441 Add mismatched detection to C++ 14 sized delete
150 469049 link failure on ppc64 (big endian) valgrind 3.20
151 469146 massif --ignore-fn does not ignore inlined functions
152 469768 Make it possible to install gdb scripts in a different location
153 470121 Can't run callgrind_control with valgrind 3.21.0 because of perl errors
154 470132 s390x: Assertion failure on VGM instruction
155 470520 Multiple realloc zero errors crash in MC_(eq_Error)
156 470713 Failure on the Yosys project: valgrind: m_libcfile.c:1802
157 (Bool vgPlain_realpath(const HChar *, HChar *)):
158 Assertion 'resolved' failed
159 470830 Don't print actions vgdb me ... continue for vgdb --multi mode
160 470978 s390x: Valgrind cannot start qemu-kvm when "sysctl vm.allocate_pgste=0"
161 471311 gdb --multi mode stdout redirecting to stderr
162 471807 Add support for lazy reading and downloading of DWARF debuginfo
163 472219 Syscall param ppoll(ufds.events) points to uninitialised byte(s)
164 472875 none/tests/s390x/dfp-1 failure
165 472963 Broken regular expression in configure.ac
166 473604 Fix bug472219.c compile failure with Clang 16
167 473677 make check compile failure with Clang 16 based on GCC 13.x
168 473745 must-be-redirected function - strlen
169 473870 FreeBSD 14 applications fail early at startup
170 473944 Handle mold linker split RW PT_LOAD segments correctly
171 474332 aligned_alloc under Valgrind returns nullptr when alignment is not a multiple of sizeof(void *)
172 475650 DRD does not work with C11 threads
173 475652 Missing suppression for __wcsncpy_avx2 (strncpy-avx2.S:308)?
174 476108 vg_replace_malloc DELETE checks size
175 n-i-bz Allow arguments with spaces in .valgrindrc files
176 n-i-bz FreeBSD fixed reading of Valgrind tools own debuginfo
178 To see details of a given bug, visit
179 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
180 where XXXXXX is the bug number as listed above.
182 (3.22.0.RC1: 17 Oct 2023)
183 (3.22.0.RC2: 26 Oct 2023)
185 Release 3.21.0 (28 Apr 2023)
186 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
188 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
189 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
190 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
191 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
192 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
193 AMD64/macOS 10.13 and nanoMIPS/Linux.
195 * ==================== CORE CHANGES ===================
197 * When GDB is used to debug a program running under valgrind using
198 the valgrind gdbserver, GDB will automatically load some
199 python code provided in valgrind defining GDB front end commands
200 corresponding to the valgrind monitor commands.
201 These GDB front end commands accept the same format as
202 the monitor commands directly sent to the Valgrind gdbserver.
203 These GDB front end commands provide a better integration
204 in the GDB command line interface, so as to use for example
205 GDB auto-completion, command specific help, searching for
206 a command or command help matching a regexp, ...
207 For relevant monitor commands, GDB will evaluate arguments
208 to make the use of monitor commands easier.
209 For example, instead of having to print the address of a variable
210 to pass it to a subsequent monitor command, the GDB front end
211 command will evaluate the address argument. It is for example
213 (gdb) memcheck who_points_at &some_struct sizeof(some_struct)
216 $2 = (some_struct_type *) 0x1130a0 <some_struct>
217 (gdb) p sizeof(some_struct)
219 (gdb) monitor who_point_at 0x1130a0 40
221 * The vgdb utility now supports extended-remote protocol when
222 invoked with --multi. In this mode the GDB run command is
223 supported. Which means you don't need to run gdb and valgrind
224 from different terminals. So for example to start your program
225 in gdb and run it under valgrind you can do:
227 (gdb) set remote exec-file prog
229 (gdb) target extended-remote | vgdb --multi
232 * The behaviour of realloc with a size of zero can now
233 be changed for tools that intercept malloc. Those
234 tools are memcheck, helgrind, drd, massif and dhat.
235 Realloc implementations generally do one of two things
236 - free the memory like free() and return NULL
237 (GNU libc and ptmalloc).
238 - either free the memory and then allocate a
239 minimum sized block or just return the
240 original pointer. Return NULL if the
241 allocation of the minimum sized block fails
242 (jemalloc, musl, snmalloc, Solaris, macOS).
243 When Valgrind is configured and built it will
244 try to match the OS and libc behaviour. However
245 if you are using a non-default library to replace
246 malloc and family (e.g., musl on a glibc Linux or
247 tcmalloc on FreeBSD) then you can use a command line
248 option to change the behaviour of Valgrind:
249 --realloc-zero-bytes-frees=yes|no [yes on Linux glibc, no otherwise]
251 * ================== PLATFORM CHANGES =================
253 * Make the address space limit on FreeBSD amd64 128Gbytes
254 (the same as Linux and Solaris, it was 32Gbytes)
256 * ==================== TOOL CHANGES ===================
259 - When doing a delta leak_search, it is now possible to only
260 output the new loss records compared to the previous leak search.
261 This is available in the memcheck monitor command 'leak_search'
262 by specifying the "new" keyword or in your program by using
263 the client request VALGRIND_DO_NEW_LEAK_CHECK.
264 Whenever a "delta" leak search is done (i.e. when specifying
265 "new" or "increased" or "changed" in the monitor command),
266 the new loss records have a "new" marker.
267 - Valgrind now contains python code that defines GDB memcheck
268 front end monitor commands. See CORE CHANGES.
269 - Performs checks for the use of realloc with a size of zero.
270 This is non-portable and a source of errors. If memcheck
271 detects such a usage it will generate an error
272 realloc() with size 0
273 followed by the usual callstacks.
274 A switch has been added to allow this to be turned off:
275 --show-realloc-size-zero=yes|no [yes]
278 - The option ---history-backtrace-size=<number> allows to configure
279 the number of entries to record in the stack traces of "old"
280 accesses. Previously, this number was hardcoded to 8.
281 - Valgrind now contains python code that defines GDB helgrind
282 front end monitor commands. See CORE CHANGES.
285 - `--cache-sim=no` is now the default. The cache simulation is old and
286 unlikely to match any real modern machine. This means only the `Ir`
287 event are gathered by default, but that is by far the most useful
289 - `cg_annotate`, `cg_diff`, and `cg_merge` have been rewritten in
290 Python. As a result, they all have more flexible command line
291 argument handling, e.g. supporting `--show-percs` and
292 `--no-show-percs` forms as well as the existing `--show-percs=yes`
293 and `--show-percs=no`.
294 - `cg_annotate` has some functional changes.
295 - It's much faster, e.g. 3-4x on common cases.
296 - It now supports diffing (with `--diff`, `--mod-filename`, and
297 `--mod-funcname`) and merging (by passing multiple data files).
298 - It now provides more information at the file and function level.
299 There are now "File:function" and "Function:file" sections. These
300 are very useful for programs that use inlining a lot.
301 - Support for user-annotated files and the `-I`/`--include` option
302 has been removed, because it was of little use and blocked other
304 - The `--auto` option is renamed `--annotate`, though the old
305 `--auto=yes`/`--auto=no` forms are still supported.
306 - `cg_diff` and `cg_merge` are now deprecated, because `cg_annotate`
307 now does a better job of diffing and merging.
308 - The Cachegrind output file format has changed very slightly, but in
309 ways nobody is likely to notice.
312 - Valgrind now contains python code that defines GDB callgrind
313 front end monitor commands. See CORE CHANGES.
316 - Valgrind now contains python code that defines GDB massif
317 front end monitor commands. See CORE CHANGES.
320 - A new kind of user request has been added which allows you to
321 override the 1024 byte limit on access count histograms for blocks
322 of memory. The client request is DHAT_HISTOGRAM_MEMORY.
324 * ==================== FIXED BUGS ====================
326 The following bugs have been fixed or resolved. Note that "n-i-bz"
327 stands for "not in bugzilla" -- that is, a bug that was reported to us
328 but never got a bugzilla entry. We encourage you to file bugs in
329 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
330 than mailing the developers (or mailing lists) directly -- bugs that
331 are not entered into bugzilla tend to get forgotten about or ignored.
333 170510 Don't warn about ioctl of size 0 without direction hint
334 241072 List tools in --help output
335 327548 false positive while destroying mutex
336 382034 Testcases build fixes for musl
337 351857 confusing error message about valid command line option
338 374596 inconsistent RDTSCP support on x86_64
339 392331 Spurious lock not held error from inside pthread_cond_timedwait
340 397083 Likely false positive "uninitialised value(s)" for __wmemchr_avx2 and __wmemcmp_avx2_movbe
341 400793 pthread_rwlock_timedwrlock false positive
342 419054 Unhandled syscall getcpu on arm32
343 433873 openat2 syscall unimplemented on Linux
344 434057 Add stdio mode to valgrind's gdbserver
345 435441 valgrind fails to interpose malloc on musl 1.2.2 due to weak symbol name and no libc soname
346 436413 Warn about realloc of size zero
347 439685 compiler warning in callgrind/main.c
348 444110 priv/guest_ppc_toIR.c:36198:31: warning: duplicated 'if' condition.
349 444487 hginfo test detects an extra lock inside data symbol "_rtld_local"
350 444488 Use glibc.pthread.stack_cache_size tunable
351 444568 drd/tests/pth_barrier_thr_cr fails on Fedora 38
352 445743 "The impossible happened: mutex is locked simultaneously by two threads"
353 while using mutexes with priority inheritance and signals
354 449309 Missing loopback device ioctl(s)
355 459476 vgdb: allow address reuse to avoid "address already in use" errorsuse" errors
356 460356 s390: Sqrt32Fx4 -- cannot reduce tree
357 462830 WARNING: unhandled amd64-freebsd syscall: 474
358 463027 broken check for MPX instruction support in assembler
359 464103 Enhancement: add a client request to DHAT to mark memory to be histogrammed
360 464476 Firefox fails to start under Valgrind
361 464609 Valgrind memcheck should support Linux pidfd_open
362 464680 Show issues caused by memory policies like selinux deny_execmem
363 464859 Build failures with GCC-13 (drd tsan_unittest)
364 464969 D language demangling
365 465435 m_libcfile.c:66 (vgPlain_safe_fd): Assertion 'newfd >= VG_(fd_hard_limit)' failed.
366 466104 aligned_alloc problems, part 1
367 467036 Add time cost statistics for Regtest
368 467482 Build failure on aarch64 Alpine
369 467714 fdleak_* and rlimit tests fail when parent process has more than
370 64 descriptors opened
371 467839 Gdbserver: Improve compatibility of library directory name
372 468401 [PATCH] Add a style file for clang-format
373 468556 Build failure for vgdb
374 468606 build: remove "Valgrind relies on GCC" check/output
375 469097 ppc64(be) doesn't support SCV syscall instruction
376 n-i-bz FreeBSD rfork syscall fail with EINVAL or ENOSYS rather than VG_(unimplemented)
378 To see details of a given bug, visit
379 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
380 where XXXXXX is the bug number as listed above.
382 * ==================== KNOWN ISSUES ===================
384 * configure --enable-lto=yes is know to not work in all setups.
385 See bug 469049. Workaround: Build without LTO.
387 (3.21.0.RC1: 14 Apr 2023)
388 (3.21.0.RC2: 21 Apr 2023)
390 Release 3.20.0 (24 Oct 2022)
391 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
393 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
394 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
395 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
396 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
397 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
398 AMD64/macOS 10.13 and nanoMIPS/Linux.
400 * ==================== CORE CHANGES ===================
402 * The option "--vgdb-stop-at=event1,event2,..." accepts the new value abexit.
403 This indicates to invoke gdbserver when your program exits abnormally
404 (i.e. with a non zero exit code).
405 * Fix Rust v0 name demangling.
406 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
407 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
408 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with
410 * The option --enable-debuginfod=<no|yes> [default: yes] has been added on
412 * More DWARF5 support as generated by clang14.
414 * ==================== FIXED BUGS ====================
416 The following bugs have been fixed or resolved. Note that "n-i-bz"
417 stands for "not in bugzilla" -- that is, a bug that was reported to us
418 but never got a bugzilla entry. We encourage you to file bugs in
419 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
420 than mailing the developers (or mailing lists) directly -- bugs that
421 are not entered into bugzilla tend to get forgotten about or ignored.
423 131186 writev reports error in (vector[...])
424 434764 iconv_open causes ld.so v2.28+ to use optimised strncmp
425 446754 Improve error codes from alloc functions under memcheck
426 452274 memcheck crashes with Assertion 'sci->status.what == SsIdle' failed
427 452779 Valgrind fails to build on FreeBSD 13.0 with llvm-devel (15.0.0)
428 453055 shared_timed_mutex drd test fails with "Lock shared failed" message
429 453602 Missing command line option to enable/disable debuginfod
430 452802 Handle lld 9+ split RW PT_LOAD segments correctly
431 454040 s390x: False-positive memcheck:cond in memmem on arch13 systems
432 456171 [PATCH] FreeBSD: Don't record address errors when accessing the 'kern.ps_strings' sysctl struct
433 n-i-bz Implement vgdb invoker on FreeBSD
434 458845 PowerPC: The L field for the dcbf and sync instruction should be
436 458915 Remove register cache to fix 458915 gdbserver causes wrong syscall return
437 459031 Documentation on --error-exitcode incomplete
438 459477 XERROR messages lacks ending '\n' in vgdb
439 462007 Implicit int in none/tests/faultstatus.c
441 To see details of a given bug, visit
442 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
443 where XXXXXX is the bug number as listed above.
445 (3.20.0.RC1: 20 Oct 2022)
448 Release 3.19.0 (11 Apr 2022)
449 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
451 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
452 PPC32/Linux, PPC64BE/Linux, PPC64LE/Linux, S390X/Linux, MIPS32/Linux,
453 MIPS64/Linux, ARM/Android, ARM64/Android, MIPS32/Android, X86/Android,
454 X86/Solaris, AMD64/Solaris, AMD64/MacOSX 10.12, X86/FreeBSD and
455 AMD64/FreeBSD. There is also preliminary support for X86/macOS 10.13,
456 AMD64/macOS 10.13 and nanoMIPS/Linux.
458 * ==================== CORE CHANGES ===================
460 * Fix Rust v0 name demangling.
461 * The Linux rseq syscall is now implemented as (silently) returning ENOSYS.
462 * Add FreeBSD syscall wrappers for __specialfd and __realpathat.
463 * Remove FreeBSD dependencies on COMPAT10, which fixes compatibility with HardenedBSD
465 * ================== PLATFORM CHANGES =================
468 - ignore the "v8.x" architecture levels, only look at actual CPU features
469 present. Fixes mismatch detected between RDMA and atomics features
470 preventing startup on some QEMU configurations.
471 - Implement LD{,A}XP and ST{,L}XP
472 - Fix incorrect code emitted for doubleword CAS.
475 - Fix sys_ipc semtimedop syscall
476 - Fix VFLRX and WFLRX instructions
477 - Fix EXRL instruction with negative offset
480 - Reimplement the vbpermq instruction support to generate less Iops and
481 avoid overflowing internal buffers.
482 - Fix checking for scv support to avoid "Facility 'SCV' unavailable (12),
483 exception" messages in dmsg.
484 - Fix setting condition code for Vector Compare quad word instructions.
485 - Fix fix lxsibzx, lxsihzx and lxsihzx instructions so they only load
486 their respective sized data.
487 - Fix the prefixed stq instruction in PC relative mode.
489 * ==================== TOOL CHANGES ===================
492 - Speed up --track-origins=yes for large (in the range of hundreds to
493 thousands of megabytes) mmap/munmaps.
495 - Several fixes for new versions of libstd++ using new posix try_lock
498 * ==================== FIXED BUGS ====================
500 The following bugs have been fixed or resolved. Note that "n-i-bz"
501 stands for "not in bugzilla" -- that is, a bug that was reported to us
502 but never got a bugzilla entry. We encourage you to file bugs in
503 bugzilla (https://bugs.kde.org/enter_bug.cgi?product=valgrind) rather
504 than mailing the developers (or mailing lists) directly -- bugs that
505 are not entered into bugzilla tend to get forgotten about or ignored.
507 403802 leak_cpp_interior fails with some reachable blocks different than expected
508 435732 memcheck/tests/leak_cpp_interior fails with gcc11
509 444242 s390x: Valgrind crashes on EXRL with negative offset
510 444399 arm64: unhandled instruction 0xC87F2D89 (LD{,A}XP and ST{,L}XP).
512 444481 gdb_server test failures on s390x
513 444495 dhat/tests/copy fails on s390x
514 444552 memcheck/tests/sem fails on s390x with glibc 2.34
515 444571 PPC, fix the lxsibzx and lxsihzx so they only load their respective
517 444836 PPC, pstq instruction for R=1 is not storing to the correct address.
518 444925 fexecve syscall wrapper not properly implemented
519 445032 valgrind/memcheck crash with SIGSEGV when SIGVTALRM timer used and
521 445211 Fix out of tree builds
522 445300 [PATCH] Fix building tests with Musl
523 445011 SIGCHLD is sent when valgrind uses debuginfod-find
524 445354 arm64 backend: incorrect code emitted for doubleword CAS
525 445415 arm64 front end: alignment checks missing for atomic instructions
526 445504 Using C++ condition_variable results in bogus "mutex is locked simultaneously by two threads" warning
527 445607 Unhandled amd64-freebsd syscall: 247
528 445668 Inline stack frame generation is broken for Rust binaries
529 445916 Demangle Rust v0 symbols with .llvm suffix
530 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false positives
531 446138 DRD/Helgrind with std::timed_mutex::try_lock_until false positives
532 446281 Add a DRD suppression for fwrite
533 446103 Memcheck: `--track-origins=yes` causes extreme slowdowns for large mmap/munmap
534 446139 DRD/Helgrind with std::shared_timed_mutex::try_lock_until and try_lock_shared_until false
535 446251 TARGET_SIGNAL_THR added to enum target_signal
536 446823 FreeBSD - missing syscalls when using libzm4
537 447991 s390x: Valgrind indicates illegal instruction on wflrx
538 447995 Valgrind segfault on power10 due to hwcap checking code
539 449483 Powerpc: vcmpgtsq., vcmpgtuq,, vcmpequq. instructions not setting the
540 condition code correctly.
541 449672 ppc64 --track-origins=yes failures because of bad cmov addHRegUse
542 449838 sigsegv liburing the 'impossible' happened for io_uring_setup
543 450025 Powerc: ACC file not implemented as a logical overlay of the VSR
545 450437 Warn for execve syscall with argv or argv[0] being NULL
546 450536 Powerpc: valgrind throws 'facility scv unavailable exception'
547 451626 Syscall param bpf(attr->raw_tracepoint.name) points to unaddressable byte(s)
548 451827 [ppc64le] VEX temporary storage exhausted with several vbpermq instructions
549 451843 valgrind fails to start on a FreeBSD system which enforces W^X
551 To see details of a given bug, visit
552 https://bugs.kde.org/show_bug.cgi?id=XXXXXX
553 where XXXXXX is the bug number as listed above.
555 (3.19.0.RC1: 02 Apr 2022)
556 (3.19.0.RC2: 08 Apr 2022)