jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 1 | |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 2 | Stable release 2.4.0 (March 2005) -- CHANGES RELATIVE TO 2.2.0 |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 3 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 4 | 2.4.0 brings many significant changes and bug fixes. The most |
| 5 | significant user-visible change is that we no longer supply our own |
| 6 | pthread implementation. Instead, Valgrind is finally capable of |
| 7 | running the native thread library, either LinuxThreads or NPTL. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 8 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 9 | This means our libpthread has gone, along with the bugs associated |
| 10 | with it. Valgrind now supports the kernel's threading syscalls, and |
| 11 | lets you use your standard system libpthread. As a result: |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 12 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 13 | * There are many fewer system dependencies and strange library-related |
| 14 | bugs. There is a small performance improvement, and a large |
| 15 | stability improvement. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 16 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 17 | * On the downside, Valgrind can no longer report misuses of the POSIX |
| 18 | PThreads API. It also means that Helgrind currently does not work. |
| 19 | We hope to fix these problems in a future release. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 20 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 21 | Note that running the native thread libraries does not mean Valgrind |
| 22 | is able to provide genuine concurrent execution on SMPs. We still |
| 23 | impose the restriction that only one thread is running at any given |
| 24 | time. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 25 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 26 | There are many other significant changes too: |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 27 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 28 | * Memcheck is (once again) the default tool. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 29 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 30 | * The default stack backtrace is now 12 call frames, rather than 4. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 31 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 32 | * Suppressions can have up to 25 call frame matches, rather than 4. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 33 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 34 | * Memcheck and Addrcheck use less memory. Under some circumstances, |
| 35 | they no longer allocate shadow memory if there are large regions of |
| 36 | memory with the same A/V states - such as an mmaped file. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 37 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 38 | * The memory-leak detector in Memcheck and Addrcheck has been |
| 39 | improved. It now reports more types of memory leak, including |
| 40 | leaked cycles. When reporting leaked memory, it can distinguish |
| 41 | between directly leaked memory (memory with no references), and |
| 42 | indirectly leaked memory (memory only referred to by other leaked |
| 43 | memory). |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 44 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 45 | * Memcheck's confusion over the effect of mprotect() has been fixed: |
| 46 | previously mprotect could erroneously mark undefined data as |
| 47 | defined. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 48 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 49 | * Signal handling is much improved and should be very close to what |
| 50 | you get when running natively. |
| 51 | |
| 52 | One result of this is that Valgrind observes changes to sigcontexts |
| 53 | passed to signal handlers. Such modifications will take effect when |
| 54 | the signal returns. You will need to run with --single-step=yes to |
| 55 | make this useful. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 56 | |
| 57 | * Valgrind is built in Position Independent Executable (PIE) format if |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 58 | your toolchain supports it. This allows it to take advantage of all |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 59 | the available address space on systems with 4Gbyte user address |
| 60 | spaces. |
| 61 | |
| 62 | * Valgrind can now run itself (requires PIE support). |
| 63 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 64 | * Syscall arguments are now checked for validity. Previously all |
| 65 | memory used by syscalls was checked, but now the actual values |
| 66 | passed are also checked. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 67 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 68 | * Syscall wrappers are more robust against bad addresses being passed |
| 69 | to syscalls: they will fail with EFAULT rather than killing Valgrind |
| 70 | with SIGSEGV. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 71 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 72 | * Because clone() is directly supported, some non-pthread uses of it |
| 73 | will work. Partial sharing (where some resources are shared, and |
| 74 | some are not) is not supported. |
| 75 | |
| 76 | * open() and readlink() on /proc/self/exe are supported. |
| 77 | |
| 78 | BUGS FIXED: |
| 79 | |
| 80 | 88520 pipe+fork+dup2 kills the main program |
| 81 | 88604 Valgrind Aborts when using $VALGRIND_OPTS and user progra... |
| 82 | 88614 valgrind: vg_libpthread.c:2323 (read): Assertion `read_pt... |
| 83 | 88703 Stabs parser fails to handle ";" |
| 84 | 88886 ioctl wrappers for TIOCMBIS and TIOCMBIC |
| 85 | 89032 valgrind pthread_cond_timedwait fails |
| 86 | 89106 the 'impossible' happened |
| 87 | 89139 Missing sched_setaffinity & sched_getaffinity |
| 88 | 89198 valgrind lacks support for SIOCSPGRP and SIOCGPGRP |
| 89 | 89263 Missing ioctl translations for scsi-generic and CD playing |
| 90 | 89440 tests/deadlock.c line endings |
| 91 | 89481 `impossible' happened: EXEC FAILED |
| 92 | 89663 valgrind 2.2.0 crash on Redhat 7.2 |
| 93 | 89792 Report pthread_mutex_lock() deadlocks instead of returnin... |
| 94 | 90111 statvfs64 gives invalid error/warning |
| 95 | 90128 crash+memory fault with stabs generated by gnat for a run... |
| 96 | 90778 VALGRIND_CHECK_DEFINED() not as documented in memcheck.h |
| 97 | 90834 cachegrind crashes at end of program without reporting re... |
| 98 | 91028 valgrind: vg_memory.c:229 (vgPlain_unmap_range): Assertio... |
| 99 | 91162 valgrind crash while debugging drivel 1.2.1 |
| 100 | 91199 Unimplemented function |
| 101 | 91325 Signal routing does not propagate the siginfo structure |
| 102 | 91599 Assertion `cv == ((void *)0)' |
| 103 | 91604 rw_lookup clears orig and sends the NULL value to rw_new |
| 104 | 91821 Small problems building valgrind with $top_builddir ne $t... |
| 105 | 91844 signal 11 (SIGSEGV) at get_tcb (libpthread.c:86) in corec... |
| 106 | 92264 UNIMPLEMENTED FUNCTION: pthread_condattr_setpshared |
| 107 | 92331 per-target flags necessitate AM_PROG_CC_C_O |
| 108 | 92420 valgrind doesn't compile with linux 2.6.8.1/9 |
| 109 | 92513 Valgrind 2.2.0 generates some warning messages |
| 110 | 92528 vg_symtab2.c:170 (addLoc): Assertion `loc->size > 0' failed. |
| 111 | 93096 unhandled ioctl 0x4B3A and 0x5601 |
| 112 | 93117 Tool and core interface versions do not match |
| 113 | 93128 Can't run valgrind --tool=memcheck because of unimplement... |
| 114 | 93174 Valgrind can crash if passed bad args to certain syscalls |
| 115 | 93309 Stack frame in new thread is badly aligned |
| 116 | 93328 Wrong types used with sys_sigprocmask() |
| 117 | 93763 /usr/include/asm/msr.h is missing |
| 118 | 93776 valgrind: vg_memory.c:508 (vgPlain_find_map_space): Asser... |
| 119 | 93810 fcntl() argument checking a bit too strict |
| 120 | 94378 Assertion `tst->sigqueue_head != tst->sigqueue_tail' failed. |
| 121 | 94429 valgrind 2.2.0 segfault with mmap64 in glibc 2.3.3 |
| 122 | 94645 Impossible happened: PINSRW mem |
| 123 | 94953 valgrind: the `impossible' happened: SIGSEGV |
| 124 | 95667 Valgrind does not work with any KDE app |
| 125 | 96243 Assertion 'res==0' failed |
| 126 | 96252 stage2 loader of valgrind fails to allocate memory |
| 127 | 96520 All programs crashing at _dl_start (in /lib/ld-2.3.3.so) ... |
| 128 | 96660 ioctl CDROMREADTOCENTRY causes bogus warnings |
| 129 | 96747 After looping in a segfault handler, the impossible happens |
| 130 | 96923 Zero sized arrays crash valgrind trace back with SIGFPE |
| 131 | 96948 valgrind stops with assertion failure regarding mmap2 |
| 132 | 96966 valgrind fails when application opens more than 16 sockets |
| 133 | 97398 valgrind: vg_libpthread.c:2667 Assertion failed |
| 134 | 97407 valgrind: vg_mylibc.c:1226 (vgPlain_safe_fd): Assertion `... |
| 135 | 97427 "Warning: invalid file descriptor -1 in syscall close()" ... |
| 136 | 97785 missing backtrace |
| 137 | 97792 build in obj dir fails - autoconf / makefile cleanup |
| 138 | 97880 pthread_mutex_lock fails from shared library (special ker... |
| 139 | 97975 program aborts without ang VG messages |
| 140 | 98129 Failed when open and close file 230000 times using stdio |
| 141 | 98175 Crashes when using valgrind-2.2.0 with a program using al... |
| 142 | 98288 Massif broken |
| 143 | 98303 UNIMPLEMENTED FUNCTION pthread_condattr_setpshared |
| 144 | 98630 failed--compilation missing warnings.pm, fails to make he... |
| 145 | 98756 Cannot valgrind signal-heavy kdrive X server |
| 146 | 98966 valgrinding the JVM fails with a sanity check assertion |
| 147 | 99035 Valgrind crashes while profiling |
| 148 | 99142 loops with message "Signal 11 being dropped from thread 0... |
| 149 | 99195 threaded apps crash on thread start (using QThread::start... |
| 150 | 99348 Assertion `vgPlain_lseek(core_fd, 0, 1) == phdrs[i].p_off... |
| 151 | 99568 False negative due to mishandling of mprotect |
| 152 | 99738 valgrind memcheck crashes on program that uses sigitimer |
| 153 | 99923 0-sized allocations are reported as leaks |
| 154 | 99949 program seg faults after exit() |
| 155 | 100036 "newSuperblock's request for 1048576 bytes failed" |
| 156 | 100116 valgrind: (pthread_cond_init): Assertion `sizeof(* cond) ... |
| 157 | 100486 memcheck reports "valgrind: the `impossible' happened: V... |
| 158 | 100833 second call to "mremap" fails with EINVAL |
| 159 | 101156 (vgPlain_find_map_space): Assertion `(addr & ((1 << 12)-1... |
| 160 | 101173 Assertion `recDepth >= 0 && recDepth < 500' failed |
| 161 | 101291 creating threads in a forked process fails |
| 162 | 101313 valgrind causes different behavior when resizing a window... |
| 163 | 101423 segfault for c++ array of floats |
| 164 | 101562 valgrind massif dies on SIGINT even with signal handler r... |
| 165 | |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 166 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 167 | Stable release 2.2.0 (31 August 2004) -- CHANGES RELATIVE TO 2.0.0 |
| 168 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
jseward | 7d180d9 | 2004-08-30 18:15:37 +0000 | [diff] [blame] | 169 | 2.2.0 brings nine months worth of improvements and bug fixes. We |
| 170 | believe it to be a worthy successor to 2.0.0. There are literally |
| 171 | hundreds of bug fixes and minor improvements. There are also some |
| 172 | fairly major user-visible changes: |
| 173 | |
| 174 | * A complete overhaul of handling of system calls and signals, and |
| 175 | their interaction with threads. In general, the accuracy of the |
| 176 | system call, thread and signal simulations is much improved: |
| 177 | |
| 178 | - Blocking system calls behave exactly as they do when running |
| 179 | natively (not on valgrind). That is, if a syscall blocks only the |
| 180 | calling thread when running natively, than it behaves the same on |
| 181 | valgrind. No more mysterious hangs because V doesn't know that some |
| 182 | syscall or other, should block only the calling thread. |
| 183 | |
| 184 | - Interrupted syscalls should now give more faithful results. |
| 185 | |
| 186 | - Signal contexts in signal handlers are supported. |
| 187 | |
| 188 | * Improvements to NPTL support to the extent that V now works |
| 189 | properly on NPTL-only setups. |
| 190 | |
| 191 | * Greater isolation between Valgrind and the program being run, so |
| 192 | the program is less likely to inadvertently kill Valgrind by |
| 193 | doing wild writes. |
| 194 | |
| 195 | * Massif: a new space profiling tool. Try it! It's cool, and it'll |
| 196 | tell you in detail where and when your C/C++ code is allocating heap. |
| 197 | Draws pretty .ps pictures of memory use against time. A potentially |
| 198 | powerful tool for making sense of your program's space use. |
| 199 | |
| 200 | * File descriptor leakage checks. When enabled, Valgrind will print out |
| 201 | a list of open file descriptors on exit. |
| 202 | |
| 203 | * Improved SSE2/SSE3 support. |
| 204 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 205 | * Time-stamped output; use --time-stamp=yes |
| 206 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 207 | |
| 208 | |
| 209 | Stable release 2.2.0 (31 August 2004) -- CHANGES RELATIVE TO 2.1.2 |
| 210 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 211 | 2.2.0 is not much different from 2.1.2, released seven weeks ago. |
| 212 | A number of bugs have been fixed, most notably #85658, which gave |
| 213 | problems for quite a few people. There have been many internal |
| 214 | cleanups, but those are not user visible. |
| 215 | |
| 216 | The following bugs have been fixed since 2.1.2: |
| 217 | |
| 218 | 85658 Assert in coregrind/vg_libpthread.c:2326 (open64) != |
| 219 | (void*)0 failed |
| 220 | This bug was reported multiple times, and so the following |
| 221 | duplicates of it are also fixed: 87620, 85796, 85935, 86065, |
| 222 | 86919, 86988, 87917, 88156 |
| 223 | |
| 224 | 80716 Semaphore mapping bug caused by unmap (sem_destroy) |
| 225 | (Was fixed prior to 2.1.2) |
| 226 | |
| 227 | 86987 semctl and shmctl syscalls family is not handled properly |
| 228 | |
| 229 | 86696 valgrind 2.1.2 + RH AS2.1 + librt |
| 230 | |
| 231 | 86730 valgrind locks up at end of run with assertion failure |
| 232 | in __pthread_unwind |
| 233 | |
| 234 | 86641 memcheck doesn't work with Mesa OpenGL/ATI on Suse 9.1 |
| 235 | (also fixes 74298, a duplicate of this) |
| 236 | |
| 237 | 85947 MMX/SSE unhandled instruction 'sfence' |
| 238 | |
| 239 | 84978 Wrong error "Conditional jump or move depends on |
| 240 | uninitialised value" resulting from "sbbl %reg, %reg" |
| 241 | |
| 242 | 86254 ssort() fails when signed int return type from comparison is |
| 243 | too small to handle result of unsigned int subtraction |
| 244 | |
| 245 | 87089 memalign( 4, xxx) makes valgrind assert |
| 246 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 247 | 86407 Add support for low-level parallel port driver ioctls. |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 248 | |
| 249 | 70587 Add timestamps to Valgrind output? (wishlist) |
| 250 | |
| 251 | 84937 vg_libpthread.c:2505 (se_remap): Assertion `res == 0' |
| 252 | (fixed prior to 2.1.2) |
| 253 | |
| 254 | 86317 cannot load libSDL-1.2.so.0 using valgrind |
| 255 | |
| 256 | 86989 memcpy from mac_replace_strmem.c complains about |
| 257 | uninitialized pointers passed when length to copy is zero |
| 258 | |
| 259 | 85811 gnu pascal symbol causes segmentation fault; ok in 2.0.0 |
| 260 | |
| 261 | 79138 writing to sbrk()'d memory causes segfault |
| 262 | |
| 263 | 77369 sched deadlock while signal received during pthread_join |
| 264 | and the joined thread exited |
| 265 | |
| 266 | 88115 In signal handler for SIGFPE, siginfo->si_addr is wrong |
| 267 | under Valgrind |
| 268 | |
| 269 | 78765 Massif crashes on app exit if FP exceptions are enabled |
| 270 | |
| 271 | Additionally there are the following changes, which are not |
| 272 | connected to any bug report numbers, AFAICS: |
| 273 | |
| 274 | * Fix scary bug causing mis-identification of SSE stores vs |
| 275 | loads and so causing memcheck to sometimes give nonsense results |
| 276 | on SSE code. |
| 277 | |
| 278 | * Add support for the POSIX message queue system calls. |
| 279 | |
| 280 | * Fix to allow 32-bit Valgrind to run on AMD64 boxes. Note: this does |
| 281 | NOT allow Valgrind to work with 64-bit executables - only with 32-bit |
| 282 | executables on an AMD64 box. |
| 283 | |
| 284 | * At configure time, only check whether linux/mii.h can be processed |
| 285 | so that we don't generate ugly warnings by trying to compile it. |
| 286 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 287 | * Add support for POSIX clocks and timers. |
| 288 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 289 | |
| 290 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 291 | Developer (cvs head) release 2.1.2 (18 July 2004) |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 292 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 293 | 2.1.2 contains four months worth of bug fixes and refinements. |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 294 | Although officially a developer release, we believe it to be stable |
| 295 | enough for widespread day-to-day use. 2.1.2 is pretty good, so try it |
| 296 | first, although there is a chance it won't work. If so then try 2.0.0 |
| 297 | and tell us what went wrong." 2.1.2 fixes a lot of problems present |
| 298 | in 2.0.0 and is generally a much better product. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 299 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 300 | Relative to 2.1.1, a large number of minor problems with 2.1.1 have |
| 301 | been fixed, and so if you use 2.1.1 you should try 2.1.2. Users of |
| 302 | the last stable release, 2.0.0, might also want to try this release. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 303 | |
| 304 | The following bugs, and probably many more, have been fixed. These |
| 305 | are listed at http://bugs.kde.org. Reporting a bug for valgrind in |
| 306 | the http://bugs.kde.org is much more likely to get you a fix than |
| 307 | mailing developers directly, so please continue to keep sending bugs |
| 308 | there. |
| 309 | |
| 310 | 76869 Crashes when running any tool under Fedora Core 2 test1 |
| 311 | This fixes the problem with returning from a signal handler |
jseward | 0400d4d | 2004-07-17 23:26:46 +0000 | [diff] [blame] | 312 | when VDSOs are turned off in FC2. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 313 | |
jseward | 0400d4d | 2004-07-17 23:26:46 +0000 | [diff] [blame] | 314 | 69508 java 1.4.2 client fails with erroneous "stack size too small". |
| 315 | This fix makes more of the pthread stack attribute related |
| 316 | functions work properly. Java still doesn't work though. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 317 | |
| 318 | 71906 malloc alignment should be 8, not 4 |
| 319 | All memory returned by malloc/new etc is now at least |
| 320 | 8-byte aligned. |
| 321 | |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 322 | 81970 vg_alloc_ThreadState: no free slots available |
| 323 | (closed because the workaround is simple: increase |
| 324 | VG_N_THREADS, rebuild and try again.) |
| 325 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 326 | 78514 Conditional jump or move depends on uninitialized value(s) |
| 327 | (a slight mishanding of FP code in memcheck) |
| 328 | |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 329 | 77952 pThread Support (crash) (due to initialisation-ordering probs) |
| 330 | (also 85118) |
| 331 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 332 | 80942 Addrcheck wasn't doing overlap checking as it should. |
| 333 | 78048 return NULL on malloc/new etc failure, instead of asserting |
| 334 | 73655 operator new() override in user .so files often doesn't get picked up |
| 335 | 83060 Valgrind does not handle native kernel AIO |
| 336 | 69872 Create proper coredumps after fatal signals |
| 337 | 82026 failure with new glibc versions: __libc_* functions are not exported |
| 338 | 70344 UNIMPLEMENTED FUNCTION: tcdrain |
| 339 | 81297 Cancellation of pthread_cond_wait does not require mutex |
| 340 | 82872 Using debug info from additional packages (wishlist) |
| 341 | 83025 Support for ioctls FIGETBSZ and FIBMAP |
| 342 | 83340 Support for ioctl HDIO_GET_IDENTITY |
| 343 | 79714 Support for the semtimedop system call. |
| 344 | 77022 Support for ioctls FBIOGET_VSCREENINFO and FBIOGET_FSCREENINFO |
| 345 | 82098 hp2ps ansification (wishlist) |
| 346 | 83573 Valgrind SIGSEGV on execve |
| 347 | 82999 show which cmdline option was erroneous (wishlist) |
| 348 | 83040 make valgrind VPATH and distcheck-clean (wishlist) |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 349 | 83998 Assertion `newfd > vgPlain_max_fd' failed (see below) |
| 350 | 82722 Unchecked mmap in as_pad leads to mysterious failures later |
| 351 | 78958 memcheck seg faults while running Mozilla |
jseward | a2c0800 | 2004-07-18 09:24:57 +0000 | [diff] [blame] | 352 | 85416 Arguments with colon (e.g. --logsocket) ignored |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 353 | |
| 354 | |
| 355 | Additionally there are the following changes, which are not |
| 356 | connected to any bug report numbers, AFAICS: |
| 357 | |
| 358 | * Rearranged address space layout relative to 2.1.1, so that |
| 359 | Valgrind/tools will run out of memory later than currently in many |
| 360 | circumstances. This is good news esp. for Calltree. It should |
| 361 | be possible for client programs to allocate over 800MB of |
| 362 | memory when using memcheck now. |
| 363 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 364 | * Improved checking when laying out memory. Should hopefully avoid |
| 365 | the random segmentation faults that 2.1.1 sometimes caused. |
| 366 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 367 | * Support for Fedora Core 2 and SuSE 9.1. Improvements to NPTL |
| 368 | support to the extent that V now works properly on NPTL-only setups. |
| 369 | |
| 370 | * Renamed the following options: |
| 371 | --logfile-fd --> --log-fd |
| 372 | --logfile --> --log-file |
| 373 | --logsocket --> --log-socket |
| 374 | to be consistent with each other and other options (esp. --input-fd). |
| 375 | |
| 376 | * Add support for SIOCGMIIPHY, SIOCGMIIREG and SIOCSMIIREG ioctls and |
| 377 | improve the checking of other interface related ioctls. |
| 378 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 379 | * Fix building with gcc-3.4.1. |
| 380 | |
| 381 | * Remove limit on number of semaphores supported. |
| 382 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 383 | * Add support for syscalls: set_tid_address (258), acct (51). |
| 384 | |
| 385 | * Support instruction "repne movs" -- not official but seems to occur. |
| 386 | |
| 387 | * Implement an emulated soft limit for file descriptors in addition to |
| 388 | the current reserved area, which effectively acts as a hard limit. The |
| 389 | setrlimit system call now simply updates the emulated limits as best |
| 390 | as possible - the hard limit is not allowed to move at all and just |
| 391 | returns EPERM if you try and change it. This should stop reductions |
| 392 | in the soft limit causing assertions when valgrind tries to allocate |
| 393 | descriptors from the reserved area. |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 394 | (This actually came from bug #83998). |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 395 | |
nethercote | 76e8fd9 | 2004-07-21 10:37:31 +0000 | [diff] [blame] | 396 | * Major overhaul of Cachegrind implementation. First user-visible change |
nethercote | 5e63bfb | 2004-07-17 12:49:53 +0000 | [diff] [blame] | 397 | is that cachegrind.out files are now typically 90% smaller than they |
nethercote | 76e8fd9 | 2004-07-21 10:37:31 +0000 | [diff] [blame] | 398 | used to be; code annotation times are correspondingly much smaller. |
| 399 | Second user-visible change is that hit/miss counts for code that is |
| 400 | unloaded at run-time is no longer dumped into a single "discard" pile, |
| 401 | but accurately preserved. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 402 | |
| 403 | * Client requests for telling valgrind about memory pools. |
| 404 | |
| 405 | |
| 406 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 407 | Developer (cvs head) release 2.1.1 (12 March 2004) |
| 408 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 409 | 2.1.1 contains some internal structural changes needed for V's |
jseward | e3e7704 | 2004-03-12 21:07:05 +0000 | [diff] [blame] | 410 | long-term future. These don't affect end-users. Most notable |
| 411 | user-visible changes are: |
| 412 | |
| 413 | * Greater isolation between Valgrind and the program being run, so |
| 414 | the program is less likely to inadvertently kill Valgrind by |
| 415 | doing wild writes. |
| 416 | |
| 417 | * Massif: a new space profiling tool. Try it! It's cool, and it'll |
| 418 | tell you in detail where and when your C/C++ code is allocating heap. |
| 419 | Draws pretty .ps pictures of memory use against time. A potentially |
| 420 | powerful tool for making sense of your program's space use. |
| 421 | |
| 422 | * Fixes for many bugs, including support for more SSE2/SSE3 instructions, |
| 423 | various signal/syscall things, and various problems with debug |
| 424 | info readers. |
| 425 | |
| 426 | * Support for glibc-2.3.3 based systems. |
| 427 | |
| 428 | We are now doing automatic overnight build-and-test runs on a variety |
| 429 | of distros. As a result, we believe 2.1.1 builds and runs on: |
| 430 | Red Hat 7.2, 7.3, 8.0, 9, Fedora Core 1, SuSE 8.2, SuSE 9. |
| 431 | |
| 432 | |
| 433 | The following bugs, and probably many more, have been fixed. These |
| 434 | are listed at http://bugs.kde.org. Reporting a bug for valgrind in |
| 435 | the http://bugs.kde.org is much more likely to get you a fix than |
| 436 | mailing developers directly, so please continue to keep sending bugs |
| 437 | there. |
| 438 | |
| 439 | 69616 glibc 2.3.2 w/NPTL is massively different than what valgrind expects |
| 440 | 69856 I don't know how to instrument MMXish stuff (Helgrind) |
| 441 | 73892 valgrind segfaults starting with Objective-C debug info |
| 442 | (fix for S-type stabs) |
| 443 | 73145 Valgrind complains too much about close(<reserved fd>) |
| 444 | 73902 Shadow memory allocation seems to fail on RedHat 8.0 |
| 445 | 68633 VG_N_SEMAPHORES too low (V itself was leaking semaphores) |
| 446 | 75099 impossible to trace multiprocess programs |
| 447 | 76839 the `impossible' happened: disInstr: INT but not 0x80 ! |
| 448 | 76762 vg_to_ucode.c:3748 (dis_push_segreg): Assertion `sz == 4' failed. |
| 449 | 76747 cannot include valgrind.h in c++ program |
| 450 | 76223 parsing B(3,10) gave NULL type => impossible happens |
| 451 | 75604 shmdt handling problem |
| 452 | 76416 Problems with gcc 3.4 snap 20040225 |
| 453 | 75614 using -gstabs when building your programs the `impossible' happened |
| 454 | 75787 Patch for some CDROM ioctls CDORM_GET_MCN, CDROM_SEND_PACKET, |
| 455 | 75294 gcc 3.4 snapshot's libstdc++ have unsupported instructions. |
| 456 | (REP RET) |
| 457 | 73326 vg_symtab2.c:272 (addScopeRange): Assertion `range->size > 0' failed. |
| 458 | 72596 not recognizing __libc_malloc |
| 459 | 69489 Would like to attach ddd to running program |
| 460 | 72781 Cachegrind crashes with kde programs |
| 461 | 73055 Illegal operand at DXTCV11CompressBlockSSE2 (more SSE opcodes) |
| 462 | 73026 Descriptor leak check reports port numbers wrongly |
| 463 | 71705 README_MISSING_SYSCALL_OR_IOCTL out of date |
| 464 | 72643 Improve support for SSE/SSE2 instructions |
| 465 | 72484 valgrind leaves it's own signal mask in place when execing |
| 466 | 72650 Signal Handling always seems to restart system calls |
| 467 | 72006 The mmap system call turns all errors in ENOMEM |
| 468 | 71781 gdb attach is pretty useless |
| 469 | 71180 unhandled instruction bytes: 0xF 0xAE 0x85 0xE8 |
| 470 | 69886 writes to zero page cause valgrind to assert on exit |
| 471 | 71791 crash when valgrinding gimp 1.3 (stabs reader problem) |
| 472 | 69783 unhandled syscall: 218 |
| 473 | 69782 unhandled instruction bytes: 0x66 0xF 0x2B 0x80 |
| 474 | 70385 valgrind fails if the soft file descriptor limit is less |
| 475 | than about 828 |
| 476 | 69529 "rep; nop" should do a yield |
| 477 | 70827 programs with lots of shared libraries report "mmap failed" |
| 478 | for some of them when reading symbols |
| 479 | 71028 glibc's strnlen is optimised enough to confuse valgrind |
| 480 | |
| 481 | |
| 482 | |
| 483 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 484 | Unstable (cvs head) release 2.1.0 (15 December 2003) |
| 485 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 486 | For whatever it's worth, 2.1.0 actually seems pretty darn stable to me |
| 487 | (Julian). It looks eminently usable, and given that it fixes some |
| 488 | significant bugs, may well be worth using on a day-to-day basis. |
| 489 | 2.1.0 is known to build and pass regression tests on: SuSE 9, SuSE |
| 490 | 8.2, RedHat 8. |
| 491 | |
| 492 | 2.1.0 most notably includes Jeremy Fitzhardinge's complete overhaul of |
| 493 | handling of system calls and signals, and their interaction with |
| 494 | threads. In general, the accuracy of the system call, thread and |
| 495 | signal simulations is much improved. Specifically: |
| 496 | |
| 497 | - Blocking system calls behave exactly as they do when running |
| 498 | natively (not on valgrind). That is, if a syscall blocks only the |
| 499 | calling thread when running natively, than it behaves the same on |
| 500 | valgrind. No more mysterious hangs because V doesn't know that some |
| 501 | syscall or other, should block only the calling thread. |
| 502 | |
| 503 | - Interrupted syscalls should now give more faithful results. |
| 504 | |
| 505 | - Finally, signal contexts in signal handlers are supported. As a |
| 506 | result, konqueror on SuSE 9 no longer segfaults when notified of |
| 507 | file changes in directories it is watching. |
| 508 | |
| 509 | Other changes: |
| 510 | |
| 511 | - Robert Walsh's file descriptor leakage checks. When enabled, |
| 512 | Valgrind will print out a list of open file descriptors on |
| 513 | exit. Along with each file descriptor, Valgrind prints out a stack |
| 514 | backtrace of where the file was opened and any details relating to the |
| 515 | file descriptor such as the file name or socket details. |
| 516 | To use, give: --track-fds=yes |
| 517 | |
| 518 | - Implemented a few more SSE/SSE2 instructions. |
| 519 | |
| 520 | - Less crud on the stack when you do 'where' inside a GDB attach. |
| 521 | |
| 522 | - Fixed the following bugs: |
| 523 | 68360: Valgrind does not compile against 2.6.0-testX kernels |
| 524 | 68525: CVS head doesn't compile on C90 compilers |
| 525 | 68566: pkgconfig support (wishlist) |
| 526 | 68588: Assertion `sz == 4' failed in vg_to_ucode.c (disInstr) |
| 527 | 69140: valgrind not able to explicitly specify a path to a binary. |
| 528 | 69432: helgrind asserts encountering a MutexErr when there are |
| 529 | EraserErr suppressions |
| 530 | |
jseward | 0b12daf | 2003-12-14 14:35:28 +0000 | [diff] [blame] | 531 | - Increase the max size of the translation cache from 200k average bbs |
| 532 | to 300k average bbs. Programs on the size of OOo (680m17) are |
| 533 | thrashing the cache at the smaller size, creating large numbers of |
| 534 | retranslations and wasting significant time as a result. |
| 535 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 536 | |
| 537 | |
| 538 | Stable release 2.0.0 (5 Nov 2003) |
| 539 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 540 | |
| 541 | 2.0.0 improves SSE/SSE2 support, fixes some minor bugs, and |
| 542 | improves support for SuSE 9 and the Red Hat "Severn" beta. |
| 543 | |
| 544 | - Further improvements to SSE/SSE2 support. The entire test suite of |
| 545 | the GNU Scientific Library (gsl-1.4) compiled with Intel Icc 7.1 |
| 546 | 20030307Z '-g -O -xW' now works. I think this gives pretty good |
| 547 | coverage of SSE/SSE2 floating point instructions, or at least the |
| 548 | subset emitted by Icc. |
| 549 | |
| 550 | - Also added support for the following instructions: |
| 551 | MOVNTDQ UCOMISD UNPCKLPS UNPCKHPS SQRTSS |
| 552 | PUSH/POP %{FS,GS}, and PUSH %CS (Nb: there is no POP %CS). |
| 553 | |
| 554 | - CFI support for GDB version 6. Needed to enable newer GDBs |
| 555 | to figure out where they are when using --gdb-attach=yes. |
| 556 | |
| 557 | - Fix this: |
| 558 | mc_translate.c:1091 (memcheck_instrument): Assertion |
| 559 | `u_in->size == 4 || u_in->size == 16' failed. |
| 560 | |
| 561 | - Return an error rather than panicing when given a bad socketcall. |
| 562 | |
| 563 | - Fix checking of syscall rt_sigtimedwait(). |
| 564 | |
| 565 | - Implement __NR_clock_gettime (syscall 265). Needed on Red Hat Severn. |
| 566 | |
| 567 | - Fixed bug in overlap check in strncpy() -- it was assuming the src was 'n' |
| 568 | bytes long, when it could be shorter, which could cause false |
| 569 | positives. |
| 570 | |
| 571 | - Support use of select() for very large numbers of file descriptors. |
| 572 | |
| 573 | - Don't fail silently if the executable is statically linked, or is |
| 574 | setuid/setgid. Print an error message instead. |
| 575 | |
| 576 | - Support for old DWARF-1 format line number info. |
| 577 | |
| 578 | |
| 579 | |
| 580 | Snapshot 20031012 (12 October 2003) |
| 581 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 582 | |
| 583 | Three months worth of bug fixes, roughly. Most significant single |
| 584 | change is improved SSE/SSE2 support, mostly thanks to Dirk Mueller. |
| 585 | |
| 586 | 20031012 builds on Red Hat Fedora ("Severn") but doesn't really work |
| 587 | (curiosly, mozilla runs OK, but a modest "ls -l" bombs). I hope to |
| 588 | get a working version out soon. It may or may not work ok on the |
| 589 | forthcoming SuSE 9; I hear positive noises about it but haven't been |
| 590 | able to verify this myself (not until I get hold of a copy of 9). |
| 591 | |
| 592 | A detailed list of changes, in no particular order: |
| 593 | |
| 594 | - Describe --gen-suppressions in the FAQ. |
| 595 | |
| 596 | - Syscall __NR_waitpid supported. |
| 597 | |
| 598 | - Minor MMX bug fix. |
| 599 | |
| 600 | - -v prints program's argv[] at startup. |
| 601 | |
| 602 | - More glibc-2.3 suppressions. |
| 603 | |
| 604 | - Suppressions for stack underrun bug(s) in the c++ support library |
| 605 | distributed with Intel Icc 7.0. |
| 606 | |
| 607 | - Fix problems reading /proc/self/maps. |
| 608 | |
| 609 | - Fix a couple of messages that should have been suppressed by -q, |
| 610 | but weren't. |
| 611 | |
| 612 | - Make Addrcheck understand "Overlap" suppressions. |
| 613 | |
| 614 | - At startup, check if program is statically linked and bail out if so. |
| 615 | |
| 616 | - Cachegrind: Auto-detect Intel Pentium-M, also VIA Nehemiah |
| 617 | |
| 618 | - Memcheck/addrcheck: minor speed optimisations |
| 619 | |
| 620 | - Handle syscall __NR_brk more correctly than before. |
| 621 | |
| 622 | - Fixed incorrect allocate/free mismatch errors when using |
| 623 | operator new(unsigned, std::nothrow_t const&) |
| 624 | operator new[](unsigned, std::nothrow_t const&) |
| 625 | |
| 626 | - Support POSIX pthread spinlocks. |
| 627 | |
| 628 | - Fixups for clean compilation with gcc-3.3.1. |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 629 | |
| 630 | - Implemented more opcodes: |
| 631 | - push %es |
| 632 | - push %ds |
| 633 | - pop %es |
| 634 | - pop %ds |
| 635 | - movntq |
| 636 | - sfence |
| 637 | - pshufw |
| 638 | - pavgb |
| 639 | - ucomiss |
| 640 | - enter |
daywalker | b18d253 | 2003-09-27 20:15:01 +0000 | [diff] [blame] | 641 | - mov imm32, %esp |
| 642 | - all "in" and "out" opcodes |
daywalker | 79aad84 | 2003-09-30 22:58:12 +0000 | [diff] [blame] | 643 | - inc/dec %esp |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 644 | - A whole bunch of SSE/SSE2 instructions |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 645 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 646 | - Memcheck: don't bomb on SSE/SSE2 code. |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 647 | |
sewardj | 9023879 | 2003-05-05 00:23:42 +0000 | [diff] [blame] | 648 | |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 649 | Snapshot 20030725 (25 July 2003) |
| 650 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 651 | |
| 652 | Fixes some minor problems in 20030716. |
| 653 | |
| 654 | - Fix bugs in overlap checking for strcpy/memcpy etc. |
| 655 | |
| 656 | - Do overlap checking with Addrcheck as well as Memcheck. |
| 657 | |
| 658 | - Fix this: |
| 659 | Memcheck: the `impossible' happened: |
| 660 | get_error_name: unexpected type |
| 661 | |
| 662 | - Install headers needed to compile new skins. |
| 663 | |
sewardj | e3dd2e0 | 2003-07-27 08:04:48 +0000 | [diff] [blame] | 664 | - Remove leading spaces and colon in the LD_LIBRARY_PATH / LD_PRELOAD |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 665 | passed to non-traced children. |
| 666 | |
sewardj | b9eda95 | 2003-07-26 21:39:05 +0000 | [diff] [blame] | 667 | - Fix file descriptor leak in valgrind-listener. |
| 668 | |
sewardj | e3dd2e0 | 2003-07-27 08:04:48 +0000 | [diff] [blame] | 669 | - Fix longstanding bug in which the allocation point of a |
| 670 | block resized by realloc was not correctly set. This may |
| 671 | have caused confusing error messages. |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 672 | |
| 673 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 674 | Snapshot 20030716 (16 July 2003) |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 675 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 676 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 677 | 20030716 is a snapshot of our current CVS head (development) branch. |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 678 | This is the branch which will become valgrind-2.0. It contains |
| 679 | significant enhancements over the 1.9.X branch. |
| 680 | |
| 681 | Despite this being a snapshot of the CVS head, it is believed to be |
| 682 | quite stable -- at least as stable as 1.9.6 or 1.0.4, if not more so |
| 683 | -- and therefore suitable for widespread use. Please let us know asap |
| 684 | if it causes problems for you. |
| 685 | |
| 686 | Two reasons for releasing a snapshot now are: |
| 687 | |
| 688 | - It's been a while since 1.9.6, and this snapshot fixes |
| 689 | various problems that 1.9.6 has with threaded programs |
| 690 | on glibc-2.3.X based systems. |
| 691 | |
| 692 | - So as to make available improvements in the 2.0 line. |
| 693 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 694 | Major changes in 20030716, as compared to 1.9.6: |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 695 | |
| 696 | - More fixes to threading support on glibc-2.3.1 and 2.3.2-based |
| 697 | systems (SuSE 8.2, Red Hat 9). If you have had problems |
| 698 | with inconsistent/illogical behaviour of errno, h_errno or the DNS |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 699 | resolver functions in threaded programs, 20030716 should improve |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 700 | matters. This snapshot seems stable enough to run OpenOffice.org |
| 701 | 1.1rc on Red Hat 7.3, SuSE 8.2 and Red Hat 9, and that's a big |
| 702 | threaded app if ever I saw one. |
| 703 | |
| 704 | - Automatic generation of suppression records; you no longer |
| 705 | need to write them by hand. Use --gen-suppressions=yes. |
| 706 | |
sewardj | 2151180 | 2003-07-22 17:47:42 +0000 | [diff] [blame] | 707 | - strcpy/memcpy/etc check their arguments for overlaps, when |
| 708 | running with the Memcheck or Addrcheck skins. |
| 709 | |
| 710 | - malloc_usable_size() is now supported. |
| 711 | |
| 712 | - new client requests: |
| 713 | - VALGRIND_COUNT_ERRORS, VALGRIND_COUNT_LEAKS: |
| 714 | useful with regression testing |
| 715 | - VALGRIND_NON_SIMD_CALL[0123]: for running arbitrary functions |
| 716 | on real CPU (use with caution!) |
| 717 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 718 | - The GDB attach mechanism is more flexible. Allow the GDB to |
| 719 | be run to be specified by --gdb-path=/path/to/gdb, and specify |
| 720 | which file descriptor V will read its input from with |
| 721 | --input-fd=<number>. |
| 722 | |
sewardj | 2151180 | 2003-07-22 17:47:42 +0000 | [diff] [blame] | 723 | - Cachegrind gives more accurate results (wasn't tracking instructions in |
| 724 | malloc() and friends previously, is now). |
| 725 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 726 | - Complete support for the MMX instruction set. |
| 727 | |
| 728 | - Partial support for the SSE and SSE2 instruction sets. Work for this |
| 729 | is ongoing. About half the SSE/SSE2 instructions are done, so |
| 730 | some SSE based programs may work. Currently you need to specify |
| 731 | --skin=addrcheck. Basically not suitable for real use yet. |
| 732 | |
| 733 | - Significant speedups (10%-20%) for standard memory checking. |
| 734 | |
| 735 | - Fix assertion failure in pthread_once(). |
| 736 | |
| 737 | - Fix this: |
| 738 | valgrind: vg_intercept.c:598 (vgAllRoadsLeadToRome_select): |
| 739 | Assertion `ms_end >= ms_now' failed. |
| 740 | |
| 741 | - Implement pthread_mutexattr_setpshared. |
| 742 | |
| 743 | - Understand Pentium 4 branch hints. Also implemented a couple more |
| 744 | obscure x86 instructions. |
| 745 | |
| 746 | - Lots of other minor bug fixes. |
| 747 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 748 | - We have a decent regression test system, for the first time. |
| 749 | This doesn't help you directly, but it does make it a lot easier |
| 750 | for us to track the quality of the system, especially across |
| 751 | multiple linux distributions. |
| 752 | |
| 753 | You can run the regression tests with 'make regtest' after 'make |
| 754 | install' completes. On SuSE 8.2 and Red Hat 9 I get this: |
| 755 | |
| 756 | == 84 tests, 0 stderr failures, 0 stdout failures == |
| 757 | |
| 758 | On Red Hat 8, I get this: |
| 759 | |
| 760 | == 84 tests, 2 stderr failures, 1 stdout failure == |
| 761 | corecheck/tests/res_search (stdout) |
| 762 | memcheck/tests/sigaltstack (stderr) |
| 763 | |
| 764 | sigaltstack is probably harmless. res_search doesn't work |
| 765 | on R H 8 even running natively, so I'm not too worried. |
| 766 | |
| 767 | On Red Hat 7.3, a glibc-2.2.5 system, I get these harmless failures: |
| 768 | |
| 769 | == 84 tests, 2 stderr failures, 1 stdout failure == |
| 770 | corecheck/tests/pth_atfork1 (stdout) |
| 771 | corecheck/tests/pth_atfork1 (stderr) |
| 772 | memcheck/tests/sigaltstack (stderr) |
| 773 | |
| 774 | You need to run on a PII system, at least, since some tests |
| 775 | contain P6-specific instructions, and the test machine needs |
| 776 | access to the internet so that corecheck/tests/res_search |
| 777 | (a test that the DNS resolver works) can function. |
| 778 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 779 | As ever, thanks for the vast amount of feedback :) and bug reports :( |
| 780 | We may not answer all messages, but we do at least look at all of |
| 781 | them, and tend to fix the most frequently reported bugs. |
| 782 | |
| 783 | |
| 784 | |
sewardj | 3791882 | 2003-05-05 01:05:09 +0000 | [diff] [blame] | 785 | Version 1.9.6 (7 May 2003 or thereabouts) |
| 786 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 787 | |
| 788 | Major changes in 1.9.6: |
| 789 | |
| 790 | - Improved threading support for glibc >= 2.3.2 (SuSE 8.2, |
| 791 | RedHat 9, to name but two ...) It turned out that 1.9.5 |
| 792 | had problems with threading support on glibc >= 2.3.2, |
| 793 | usually manifested by threaded programs deadlocking in system calls, |
| 794 | or running unbelievably slowly. Hopefully these are fixed now. 1.9.6 |
| 795 | is the first valgrind which gives reasonable support for |
| 796 | glibc-2.3.2. Also fixed a 2.3.2 problem with pthread_atfork(). |
| 797 | |
| 798 | - Majorly expanded FAQ.txt. We've added workarounds for all |
| 799 | common problems for which a workaround is known. |
| 800 | |
| 801 | Minor changes in 1.9.6: |
| 802 | |
| 803 | - Fix identification of the main thread's stack. Incorrect |
| 804 | identification of it was causing some on-stack addresses to not get |
| 805 | identified as such. This only affected the usefulness of some error |
| 806 | messages; the correctness of the checks made is unchanged. |
| 807 | |
| 808 | - Support for kernels >= 2.5.68. |
| 809 | |
| 810 | - Dummy implementations of __libc_current_sigrtmin, |
| 811 | __libc_current_sigrtmax and __libc_allocate_rtsig, hopefully |
| 812 | good enough to keep alive programs which previously died for lack of |
| 813 | them. |
| 814 | |
| 815 | - Fix bug in the VALGRIND_DISCARD_TRANSLATIONS client request. |
| 816 | |
sewardj | 3d47b79 | 2003-05-05 22:15:35 +0000 | [diff] [blame] | 817 | - Fix bug in the DWARF2 debug line info loader, when instructions |
| 818 | following each other have source lines far from each other |
| 819 | (e.g. with inlined functions). |
| 820 | |
sewardj | 3791882 | 2003-05-05 01:05:09 +0000 | [diff] [blame] | 821 | - Debug info reading: read symbols from both "symtab" and "dynsym" |
| 822 | sections, rather than merely from the one that comes last in the |
| 823 | file. |
| 824 | |
| 825 | - New syscall support: prctl(), creat(), lookup_dcookie(). |
| 826 | |
| 827 | - When checking calls to accept(), recvfrom(), getsocketopt(), |
| 828 | don't complain if buffer values are NULL. |
| 829 | |
| 830 | - Try and avoid assertion failures in |
| 831 | mash_LD_PRELOAD_and_LD_LIBRARY_PATH. |
| 832 | |
| 833 | - Minor bug fixes in cg_annotate. |
| 834 | |
| 835 | |
| 836 | |
sewardj | 9023879 | 2003-05-05 00:23:42 +0000 | [diff] [blame] | 837 | Version 1.9.5 (7 April 2003) |
| 838 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 839 | |
| 840 | It occurs to me that it would be helpful for valgrind users to record |
| 841 | in the source distribution the changes in each release. So I now |
| 842 | attempt to mend my errant ways :-) Changes in this and future releases |
| 843 | will be documented in the NEWS file in the source distribution. |
| 844 | |
| 845 | Major changes in 1.9.5: |
| 846 | |
| 847 | - (Critical bug fix): Fix a bug in the FPU simulation. This was |
| 848 | causing some floating point conditional tests not to work right. |
| 849 | Several people reported this. If you had floating point code which |
| 850 | didn't work right on 1.9.1 to 1.9.4, it's worth trying 1.9.5. |
| 851 | |
| 852 | - Partial support for Red Hat 9. RH9 uses the new Native Posix |
| 853 | Threads Library (NPTL), instead of the older LinuxThreads. |
| 854 | This potentially causes problems with V which will take some |
| 855 | time to correct. In the meantime we have partially worked around |
| 856 | this, and so 1.9.5 works on RH9. Threaded programs still work, |
| 857 | but they may deadlock, because some system calls (accept, read, |
| 858 | write, etc) which should be nonblocking, in fact do block. This |
| 859 | is a known bug which we are looking into. |
| 860 | |
| 861 | If you can, your best bet (unfortunately) is to avoid using |
| 862 | 1.9.5 on a Red Hat 9 system, or on any NPTL-based distribution. |
| 863 | If your glibc is 2.3.1 or earlier, you're almost certainly OK. |
| 864 | |
| 865 | Minor changes in 1.9.5: |
| 866 | |
| 867 | - Added some #errors to valgrind.h to ensure people don't include |
| 868 | it accidentally in their sources. This is a change from 1.0.X |
| 869 | which was never properly documented. The right thing to include |
| 870 | is now memcheck.h. Some people reported problems and strange |
| 871 | behaviour when (incorrectly) including valgrind.h in code with |
| 872 | 1.9.1 -- 1.9.4. This is no longer possible. |
| 873 | |
| 874 | - Add some __extension__ bits and pieces so that gcc configured |
| 875 | for valgrind-checking compiles even with -Werror. If you |
| 876 | don't understand this, ignore it. Of interest to gcc developers |
| 877 | only. |
| 878 | |
| 879 | - Removed a pointless check which caused problems interworking |
| 880 | with Clearcase. V would complain about shared objects whose |
| 881 | names did not end ".so", and refuse to run. This is now fixed. |
| 882 | In fact it was fixed in 1.9.4 but not documented. |
| 883 | |
| 884 | - Fixed a bug causing an assertion failure of "waiters == 1" |
| 885 | somewhere in vg_scheduler.c, when running large threaded apps, |
| 886 | notably MySQL. |
| 887 | |
| 888 | - Add support for the munlock system call (124). |
| 889 | |
| 890 | Some comments about future releases: |
| 891 | |
| 892 | 1.9.5 is, we hope, the most stable Valgrind so far. It pretty much |
| 893 | supersedes the 1.0.X branch. If you are a valgrind packager, please |
| 894 | consider making 1.9.5 available to your users. You can regard the |
| 895 | 1.0.X branch as obsolete: 1.9.5 is stable and vastly superior. There |
| 896 | are no plans at all for further releases of the 1.0.X branch. |
| 897 | |
| 898 | If you want a leading-edge valgrind, consider building the cvs head |
| 899 | (from SourceForge), or getting a snapshot of it. Current cool stuff |
| 900 | going in includes MMX support (done); SSE/SSE2 support (in progress), |
| 901 | a significant (10-20%) performance improvement (done), and the usual |
| 902 | large collection of minor changes. Hopefully we will be able to |
| 903 | improve our NPTL support, but no promises. |
| 904 | |