jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 1 | |
sewardj | cd62fe7 | 2005-08-29 13:24:51 +0000 | [diff] [blame] | 2 | Release 3.0.1 (29 August 2005) |
| 3 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 4 | 3.0.1 fixes a bunch of bugs reported in 3.0.0. There is no new |
| 5 | functionality. Some of the fixed bugs are critical, so if you |
sewardj | fa23389 | 2005-08-30 12:35:36 +0000 | [diff] [blame] | 6 | use/distribute 3.0.0, an upgrade to 3.0.1 is recommended. The fixed |
sewardj | cd62fe7 | 2005-08-29 13:24:51 +0000 | [diff] [blame] | 7 | bugs are: |
| 8 | |
| 9 | (note: "n-i-bz" means "not in bugzilla" -- this bug does not have |
| 10 | a bugzilla entry). |
| 11 | |
| 12 | 109313 (== 110505) x86 cmpxchg8b |
| 13 | n-i-bz x86: track but ignore changes to %eflags.AC (alignment check) |
| 14 | 110102 dis_op2_E_G(amd64) |
| 15 | 110202 x86 sys_waitpid(#286) |
| 16 | 110203 clock_getres(,0) |
| 17 | 110208 execve fail wrong retval |
| 18 | 110274 SSE1 now mandatory for x86 |
| 19 | 110388 amd64 0xDD 0xD1 |
| 20 | 110464 amd64 0xDC 0x1D FCOMP |
| 21 | 110478 amd64 0xF 0xD PREFETCH |
| 22 | n-i-bz XML <unique> printing wrong |
| 23 | n-i-bz Dirk r4359 (amd64 syscalls from trunk) |
| 24 | 110591 amd64 and x86: rdtsc not implemented properly |
| 25 | n-i-bz Nick r4384 (stub implementations of Addrcheck and Helgrind) |
| 26 | 110652 AMD64 valgrind crashes on cwtd instruction |
| 27 | 110653 AMD64 valgrind crashes on sarb $0x4,foo(%rip) instruction |
| 28 | 110656 PATH=/usr/bin::/bin valgrind foobar stats ./fooba |
| 29 | 110657 Small test fixes |
| 30 | 110671 vex x86->IR: unhandled instruction bytes: 0xF3 0xC3 (rep ret) |
| 31 | n-i-bz Nick (Cachegrind should not assert when it encounters a client |
| 32 | request.) |
| 33 | 110685 amd64->IR: unhandled instruction bytes: 0xE1 0x56 (loope Jb) |
| 34 | 110830 configuring with --host fails to build 32 bit on 64 bit target |
| 35 | 110875 Assertion when execve fails |
| 36 | n-i-bz Updates to Memcheck manual |
| 37 | n-i-bz Fixed broken malloc_usable_size() |
| 38 | 110898 opteron instructions missing: btq btsq btrq bsfq |
| 39 | 110954 x86->IR: unhandled instruction bytes: 0xE2 0xF6 (loop Jb) |
| 40 | n-i-bz Make suppressions work for "???" lines in stacktraces. |
| 41 | 111006 bogus warnings from linuxthreads |
| 42 | 111092 x86: dis_Grp2(Reg): unhandled case(x86) |
| 43 | 111231 sctp_getladdrs() and sctp_getpaddrs() returns uninitialized |
| 44 | memory |
| 45 | 111102 (comment #4) Fixed 64-bit unclean "silly arg" message |
| 46 | n-i-bz vex x86->IR: unhandled instruction bytes: 0x14 0x0 |
| 47 | n-i-bz minor umount/fcntl wrapper fixes |
| 48 | 111090 Internal Error running Massif |
| 49 | 101204 noisy warning |
| 50 | 111513 Illegal opcode for SSE instruction (x86 movups) |
| 51 | 111555 VEX/Makefile: CC is set to gcc |
sewardj | fa23389 | 2005-08-30 12:35:36 +0000 | [diff] [blame] | 52 | n-i-bz Fix XML bugs in FAQ |
sewardj | cd62fe7 | 2005-08-29 13:24:51 +0000 | [diff] [blame] | 53 | |
sewardj | fa23389 | 2005-08-30 12:35:36 +0000 | [diff] [blame] | 54 | (3.0.1: 29 August 05, |
| 55 | vex/branches/VEX_3_0_BRANCH r1367, |
| 56 | valgrind/branches/VALGRIND_3_0_BRANCH r4574). |
sewardj | 820c772 | 2005-08-29 13:44:43 +0000 | [diff] [blame] | 57 | |
| 58 | |
sewardj | cd62fe7 | 2005-08-29 13:24:51 +0000 | [diff] [blame] | 59 | |
sewardj | ec39f33 | 2005-08-02 15:28:07 +0000 | [diff] [blame] | 60 | Release 3.0.0 (3 August 2005) |
| 61 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
sewardj | f9929e6 | 2005-08-02 21:05:58 +0000 | [diff] [blame] | 62 | 3.0.0 is a major overhaul of Valgrind. The most significant user |
| 63 | visible change is that Valgrind now supports architectures other than |
| 64 | x86. The new architectures it supports are AMD64 and PPC32, and the |
| 65 | infrastructure is present for other architectures to be added later. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 66 | |
sewardj | f9929e6 | 2005-08-02 21:05:58 +0000 | [diff] [blame] | 67 | AMD64 support works well, but has some shortcomings: |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 68 | |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 69 | - It generally won't be as solid as the x86 version. For example, |
| 70 | support for more obscure instructions and system calls may be missing. |
| 71 | We will fix these as they arise. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 72 | |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 73 | - Address space may be limited; see the point about |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 74 | position-independent executables below. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 75 | |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 76 | - If Valgrind is built on an AMD64 machine, it will only run 64-bit |
| 77 | executables. If you want to run 32-bit x86 executables under Valgrind |
| 78 | on an AMD64, you will need to build Valgrind on an x86 machine and |
| 79 | copy it to the AMD64 machine. And it probably won't work if you do |
| 80 | something tricky like exec'ing a 32-bit program from a 64-bit program |
| 81 | while using --trace-children=yes. We hope to improve this situation |
| 82 | in the future. |
| 83 | |
| 84 | The PPC32 support is very basic. It may not work reliably even for |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 85 | small programs, but it's a start. Many thanks to Paul Mackerras for |
| 86 | his great work that enabled this support. We are working to make |
| 87 | PPC32 usable as soon as possible. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 88 | |
| 89 | Other user-visible changes: |
| 90 | |
sewardj | ec39f33 | 2005-08-02 15:28:07 +0000 | [diff] [blame] | 91 | - Valgrind is no longer built by default as a position-independent |
| 92 | executable (PIE), as this caused too many problems. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 93 | |
| 94 | Without PIE enabled, AMD64 programs will only be able to access 2GB of |
| 95 | address space. We will fix this eventually, but not for the moment. |
| 96 | |
| 97 | Use --enable-pie at configure-time to turn this on. |
| 98 | |
| 99 | - Support for programs that use stack-switching has been improved. Use |
| 100 | the --max-stackframe flag for simple cases, and the |
| 101 | VALGRIND_STACK_REGISTER, VALGRIND_STACK_DEREGISTER and |
| 102 | VALGRIND_STACK_CHANGE client requests for trickier cases. |
| 103 | |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 104 | - Support for programs that use self-modifying code has been improved, |
| 105 | in particular programs that put temporary code fragments on the stack. |
| 106 | This helps for C programs compiled with GCC that use nested functions, |
njn | e772332 | 2005-07-24 17:19:09 +0000 | [diff] [blame] | 107 | and also Ada programs. This is controlled with the --smc-check |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 108 | flag, although the default setting should work in most cases. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 109 | |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 110 | - Output can now be printed in XML format. This should make it easier |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 111 | for tools such as GUI front-ends and automated error-processing |
| 112 | schemes to use Valgrind output as input. The --xml flag controls this. |
| 113 | As part of this change, ELF directory information is read from executables, |
| 114 | so absolute source file paths are available if needed. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 115 | |
| 116 | - Programs that allocate many heap blocks may run faster, due to |
| 117 | improvements in certain data structures. |
| 118 | |
njn | 7b1a6a4 | 2005-07-20 04:20:44 +0000 | [diff] [blame] | 119 | - Addrcheck is currently not working. We hope to get it working again |
| 120 | soon. Helgrind is still not working, as was the case for the 2.4.0 |
| 121 | release. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 122 | |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 123 | - The JITter has been completely rewritten, and is now in a separate |
| 124 | library, called Vex. This enabled a lot of the user-visible changes, |
| 125 | such as new architecture support. The new JIT unfortunately translates |
| 126 | more slowly than the old one, so programs may take longer to start. |
| 127 | We believe the code quality is produces is about the same, so once |
| 128 | started, programs should run at about the same speed. Feedback about |
| 129 | this would be useful. |
| 130 | |
| 131 | On the plus side, Vex and hence Memcheck tracks value flow properly |
| 132 | through floating point and vector registers, something the 2.X line |
| 133 | could not do. That means that Memcheck is much more likely to be |
| 134 | usably accurate on vectorised code. |
| 135 | |
sewardj | ec39f33 | 2005-08-02 15:28:07 +0000 | [diff] [blame] | 136 | - There is a subtle change to the way exiting of threaded programs |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 137 | is handled. In 3.0, Valgrind's final diagnostic output (leak check, |
| 138 | etc) is not printed until the last thread exits. If the last thread |
| 139 | to exit was not the original thread which started the program, any |
| 140 | other process wait()-ing on this one to exit may conclude it has |
| 141 | finished before the diagnostic output is printed. This may not be |
| 142 | what you expect. 2.X had a different scheme which avoided this |
| 143 | problem, but caused deadlocks under obscure circumstances, so we |
| 144 | are trying something different for 3.0. |
| 145 | |
| 146 | - Small changes in control log file naming which make it easier to |
sewardj | ec39f33 | 2005-08-02 15:28:07 +0000 | [diff] [blame] | 147 | use valgrind for debugging MPI-based programs. The relevant |
| 148 | new flags are --log-file-exactly= and --log-file-qualifier=. |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 149 | |
| 150 | - As part of adding AMD64 support, DWARF2 CFI-based stack unwinding |
| 151 | support was added. In principle this means Valgrind can produce |
| 152 | meaningful backtraces on x86 code compiled with -fomit-frame-pointer |
tom | 81868f0 | 2005-07-27 10:33:08 +0000 | [diff] [blame] | 153 | providing you also compile your code with -fasynchronous-unwind-tables. |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 154 | |
sewardj | ec39f33 | 2005-08-02 15:28:07 +0000 | [diff] [blame] | 155 | - The documentation build system has been completely redone. |
| 156 | The documentation masters are now in XML format, and from that |
| 157 | HTML, PostScript and PDF documentation is generated. As a result |
| 158 | the manual is now available in book form. Note that the |
| 159 | documentation in the source tarballs is pre-built, so you don't need |
| 160 | any XML processing tools to build Valgrind from a tarball. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 161 | |
| 162 | Changes that are not user-visible: |
| 163 | |
sewardj | c371bd0 | 2005-07-27 09:55:24 +0000 | [diff] [blame] | 164 | - The code has been massively overhauled in order to modularise it. |
| 165 | As a result we hope it is easier to navigate and understand. |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 166 | |
| 167 | - Lots of code has been rewritten. |
| 168 | |
sewardj | f9929e6 | 2005-08-02 21:05:58 +0000 | [diff] [blame] | 169 | BUGS FIXED: |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 170 | |
sewardj | 79bd9b6 | 2005-08-03 16:40:10 +0000 | [diff] [blame] | 171 | 110046 sz == 4 assertion failed |
| 172 | 109810 vex amd64->IR: unhandled instruction bytes: 0xA3 0x4C 0x70 0xD7 |
sewardj | f9929e6 | 2005-08-02 21:05:58 +0000 | [diff] [blame] | 173 | 109802 Add a plausible_stack_size command-line parameter ? |
| 174 | 109783 unhandled ioctl TIOCMGET (running hw detection tool discover) |
| 175 | 109780 unhandled ioctl BLKSSZGET (running fdisk -l /dev/hda) |
| 176 | 109718 vex x86->IR: unhandled instruction: ffreep |
| 177 | 109429 AMD64 unhandled syscall: 127 (sigpending) |
| 178 | 109401 false positive uninit in strchr from ld-linux.so.2 |
| 179 | 109385 "stabs" parse failure |
| 180 | 109378 amd64: unhandled instruction REP NOP |
| 181 | 109376 amd64: unhandled instruction LOOP Jb |
| 182 | 109363 AMD64 unhandled instruction bytes |
| 183 | 109362 AMD64 unhandled syscall: 24 (sched_yield) |
| 184 | 109358 fork() won't work with valgrind-3.0 SVN |
| 185 | 109332 amd64 unhandled instruction: ADC Ev, Gv |
| 186 | 109314 Bogus memcheck report on amd64 |
| 187 | 108883 Crash; vg_memory.c:905 (vgPlain_init_shadow_range): |
| 188 | Assertion `vgPlain_defined_init_shadow_page()' failed. |
| 189 | 108349 mincore syscall parameter checked incorrectly |
| 190 | 108059 build infrastructure: small update |
| 191 | 107524 epoll_ctl event parameter checked on EPOLL_CTL_DEL |
| 192 | 107123 Vex dies with unhandled instructions: 0xD9 0x31 0xF 0xAE |
| 193 | 106841 auxmap & openGL problems |
| 194 | 106713 SDL_Init causes valgrind to exit |
| 195 | 106352 setcontext and makecontext not handled correctly |
| 196 | 106293 addresses beyond initial client stack allocation |
| 197 | not checked in VALGRIND_DO_LEAK_CHECK |
| 198 | 106283 PIE client programs are loaded at address 0 |
| 199 | 105831 Assertion `vgPlain_defined_init_shadow_page()' failed. |
| 200 | 105039 long run-times probably due to memory manager |
| 201 | 104797 valgrind needs to be aware of BLKGETSIZE64 |
| 202 | 103594 unhandled instruction: FICOM |
| 203 | 103320 Valgrind 2.4.0 fails to compile with gcc 3.4.3 and -O0 |
| 204 | 103168 potentially memory leak in coregrind/ume.c |
| 205 | 102039 bad permissions for mapped region at address 0xB7C73680 |
| 206 | 101881 weird assertion problem |
| 207 | 101543 Support fadvise64 syscalls |
| 208 | 75247 x86_64/amd64 support (the biggest "bug" we have ever fixed) |
njn | 4a9f6f4 | 2005-07-19 23:46:08 +0000 | [diff] [blame] | 209 | |
sewardj | 9a91951 | 2005-08-03 17:01:51 +0000 | [diff] [blame] | 210 | (3.0RC1: 27 July 05, vex r1303, valgrind r4283). |
sewardj | 1346e13 | 2005-08-03 20:42:06 +0000 | [diff] [blame] | 211 | (3.0.0: 3 August 05, vex r1313, valgrind r4316). |
sewardj | bfa2b99 | 2005-07-27 17:49:17 +0000 | [diff] [blame] | 212 | |
| 213 | |
| 214 | |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 215 | Stable release 2.4.0 (March 2005) -- CHANGES RELATIVE TO 2.2.0 |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 216 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 217 | 2.4.0 brings many significant changes and bug fixes. The most |
| 218 | significant user-visible change is that we no longer supply our own |
| 219 | pthread implementation. Instead, Valgrind is finally capable of |
| 220 | running the native thread library, either LinuxThreads or NPTL. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 221 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 222 | This means our libpthread has gone, along with the bugs associated |
| 223 | with it. Valgrind now supports the kernel's threading syscalls, and |
| 224 | lets you use your standard system libpthread. As a result: |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 225 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 226 | * There are many fewer system dependencies and strange library-related |
| 227 | bugs. There is a small performance improvement, and a large |
| 228 | stability improvement. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 229 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 230 | * On the downside, Valgrind can no longer report misuses of the POSIX |
| 231 | PThreads API. It also means that Helgrind currently does not work. |
| 232 | We hope to fix these problems in a future release. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 233 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 234 | Note that running the native thread libraries does not mean Valgrind |
| 235 | is able to provide genuine concurrent execution on SMPs. We still |
| 236 | impose the restriction that only one thread is running at any given |
| 237 | time. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 238 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 239 | There are many other significant changes too: |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 240 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 241 | * Memcheck is (once again) the default tool. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 242 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 243 | * The default stack backtrace is now 12 call frames, rather than 4. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 244 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 245 | * Suppressions can have up to 25 call frame matches, rather than 4. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 246 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 247 | * Memcheck and Addrcheck use less memory. Under some circumstances, |
| 248 | they no longer allocate shadow memory if there are large regions of |
| 249 | memory with the same A/V states - such as an mmaped file. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 250 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 251 | * The memory-leak detector in Memcheck and Addrcheck has been |
| 252 | improved. It now reports more types of memory leak, including |
| 253 | leaked cycles. When reporting leaked memory, it can distinguish |
| 254 | between directly leaked memory (memory with no references), and |
| 255 | indirectly leaked memory (memory only referred to by other leaked |
| 256 | memory). |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 257 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 258 | * Memcheck's confusion over the effect of mprotect() has been fixed: |
| 259 | previously mprotect could erroneously mark undefined data as |
| 260 | defined. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 261 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 262 | * Signal handling is much improved and should be very close to what |
| 263 | you get when running natively. |
| 264 | |
| 265 | One result of this is that Valgrind observes changes to sigcontexts |
| 266 | passed to signal handlers. Such modifications will take effect when |
| 267 | the signal returns. You will need to run with --single-step=yes to |
| 268 | make this useful. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 269 | |
| 270 | * Valgrind is built in Position Independent Executable (PIE) format if |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 271 | your toolchain supports it. This allows it to take advantage of all |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 272 | the available address space on systems with 4Gbyte user address |
| 273 | spaces. |
| 274 | |
| 275 | * Valgrind can now run itself (requires PIE support). |
| 276 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 277 | * Syscall arguments are now checked for validity. Previously all |
| 278 | memory used by syscalls was checked, but now the actual values |
| 279 | passed are also checked. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 280 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 281 | * Syscall wrappers are more robust against bad addresses being passed |
| 282 | to syscalls: they will fail with EFAULT rather than killing Valgrind |
| 283 | with SIGSEGV. |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 284 | |
njn | 1aa5ae4 | 2005-03-26 16:04:27 +0000 | [diff] [blame] | 285 | * Because clone() is directly supported, some non-pthread uses of it |
| 286 | will work. Partial sharing (where some resources are shared, and |
| 287 | some are not) is not supported. |
| 288 | |
| 289 | * open() and readlink() on /proc/self/exe are supported. |
| 290 | |
| 291 | BUGS FIXED: |
| 292 | |
| 293 | 88520 pipe+fork+dup2 kills the main program |
| 294 | 88604 Valgrind Aborts when using $VALGRIND_OPTS and user progra... |
| 295 | 88614 valgrind: vg_libpthread.c:2323 (read): Assertion `read_pt... |
| 296 | 88703 Stabs parser fails to handle ";" |
| 297 | 88886 ioctl wrappers for TIOCMBIS and TIOCMBIC |
| 298 | 89032 valgrind pthread_cond_timedwait fails |
| 299 | 89106 the 'impossible' happened |
| 300 | 89139 Missing sched_setaffinity & sched_getaffinity |
| 301 | 89198 valgrind lacks support for SIOCSPGRP and SIOCGPGRP |
| 302 | 89263 Missing ioctl translations for scsi-generic and CD playing |
| 303 | 89440 tests/deadlock.c line endings |
| 304 | 89481 `impossible' happened: EXEC FAILED |
| 305 | 89663 valgrind 2.2.0 crash on Redhat 7.2 |
| 306 | 89792 Report pthread_mutex_lock() deadlocks instead of returnin... |
| 307 | 90111 statvfs64 gives invalid error/warning |
| 308 | 90128 crash+memory fault with stabs generated by gnat for a run... |
| 309 | 90778 VALGRIND_CHECK_DEFINED() not as documented in memcheck.h |
| 310 | 90834 cachegrind crashes at end of program without reporting re... |
| 311 | 91028 valgrind: vg_memory.c:229 (vgPlain_unmap_range): Assertio... |
| 312 | 91162 valgrind crash while debugging drivel 1.2.1 |
| 313 | 91199 Unimplemented function |
| 314 | 91325 Signal routing does not propagate the siginfo structure |
| 315 | 91599 Assertion `cv == ((void *)0)' |
| 316 | 91604 rw_lookup clears orig and sends the NULL value to rw_new |
| 317 | 91821 Small problems building valgrind with $top_builddir ne $t... |
| 318 | 91844 signal 11 (SIGSEGV) at get_tcb (libpthread.c:86) in corec... |
| 319 | 92264 UNIMPLEMENTED FUNCTION: pthread_condattr_setpshared |
| 320 | 92331 per-target flags necessitate AM_PROG_CC_C_O |
| 321 | 92420 valgrind doesn't compile with linux 2.6.8.1/9 |
| 322 | 92513 Valgrind 2.2.0 generates some warning messages |
| 323 | 92528 vg_symtab2.c:170 (addLoc): Assertion `loc->size > 0' failed. |
| 324 | 93096 unhandled ioctl 0x4B3A and 0x5601 |
| 325 | 93117 Tool and core interface versions do not match |
| 326 | 93128 Can't run valgrind --tool=memcheck because of unimplement... |
| 327 | 93174 Valgrind can crash if passed bad args to certain syscalls |
| 328 | 93309 Stack frame in new thread is badly aligned |
| 329 | 93328 Wrong types used with sys_sigprocmask() |
| 330 | 93763 /usr/include/asm/msr.h is missing |
| 331 | 93776 valgrind: vg_memory.c:508 (vgPlain_find_map_space): Asser... |
| 332 | 93810 fcntl() argument checking a bit too strict |
| 333 | 94378 Assertion `tst->sigqueue_head != tst->sigqueue_tail' failed. |
| 334 | 94429 valgrind 2.2.0 segfault with mmap64 in glibc 2.3.3 |
| 335 | 94645 Impossible happened: PINSRW mem |
| 336 | 94953 valgrind: the `impossible' happened: SIGSEGV |
| 337 | 95667 Valgrind does not work with any KDE app |
| 338 | 96243 Assertion 'res==0' failed |
| 339 | 96252 stage2 loader of valgrind fails to allocate memory |
| 340 | 96520 All programs crashing at _dl_start (in /lib/ld-2.3.3.so) ... |
| 341 | 96660 ioctl CDROMREADTOCENTRY causes bogus warnings |
| 342 | 96747 After looping in a segfault handler, the impossible happens |
| 343 | 96923 Zero sized arrays crash valgrind trace back with SIGFPE |
| 344 | 96948 valgrind stops with assertion failure regarding mmap2 |
| 345 | 96966 valgrind fails when application opens more than 16 sockets |
| 346 | 97398 valgrind: vg_libpthread.c:2667 Assertion failed |
| 347 | 97407 valgrind: vg_mylibc.c:1226 (vgPlain_safe_fd): Assertion `... |
| 348 | 97427 "Warning: invalid file descriptor -1 in syscall close()" ... |
| 349 | 97785 missing backtrace |
| 350 | 97792 build in obj dir fails - autoconf / makefile cleanup |
| 351 | 97880 pthread_mutex_lock fails from shared library (special ker... |
| 352 | 97975 program aborts without ang VG messages |
| 353 | 98129 Failed when open and close file 230000 times using stdio |
| 354 | 98175 Crashes when using valgrind-2.2.0 with a program using al... |
| 355 | 98288 Massif broken |
| 356 | 98303 UNIMPLEMENTED FUNCTION pthread_condattr_setpshared |
| 357 | 98630 failed--compilation missing warnings.pm, fails to make he... |
| 358 | 98756 Cannot valgrind signal-heavy kdrive X server |
| 359 | 98966 valgrinding the JVM fails with a sanity check assertion |
| 360 | 99035 Valgrind crashes while profiling |
| 361 | 99142 loops with message "Signal 11 being dropped from thread 0... |
| 362 | 99195 threaded apps crash on thread start (using QThread::start... |
| 363 | 99348 Assertion `vgPlain_lseek(core_fd, 0, 1) == phdrs[i].p_off... |
| 364 | 99568 False negative due to mishandling of mprotect |
| 365 | 99738 valgrind memcheck crashes on program that uses sigitimer |
| 366 | 99923 0-sized allocations are reported as leaks |
| 367 | 99949 program seg faults after exit() |
| 368 | 100036 "newSuperblock's request for 1048576 bytes failed" |
| 369 | 100116 valgrind: (pthread_cond_init): Assertion `sizeof(* cond) ... |
| 370 | 100486 memcheck reports "valgrind: the `impossible' happened: V... |
| 371 | 100833 second call to "mremap" fails with EINVAL |
| 372 | 101156 (vgPlain_find_map_space): Assertion `(addr & ((1 << 12)-1... |
| 373 | 101173 Assertion `recDepth >= 0 && recDepth < 500' failed |
| 374 | 101291 creating threads in a forked process fails |
| 375 | 101313 valgrind causes different behavior when resizing a window... |
| 376 | 101423 segfault for c++ array of floats |
| 377 | 101562 valgrind massif dies on SIGINT even with signal handler r... |
| 378 | |
njn | 8c92730 | 2005-03-12 16:45:01 +0000 | [diff] [blame] | 379 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 380 | Stable release 2.2.0 (31 August 2004) -- CHANGES RELATIVE TO 2.0.0 |
| 381 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
jseward | 7d180d9 | 2004-08-30 18:15:37 +0000 | [diff] [blame] | 382 | 2.2.0 brings nine months worth of improvements and bug fixes. We |
| 383 | believe it to be a worthy successor to 2.0.0. There are literally |
| 384 | hundreds of bug fixes and minor improvements. There are also some |
| 385 | fairly major user-visible changes: |
| 386 | |
| 387 | * A complete overhaul of handling of system calls and signals, and |
| 388 | their interaction with threads. In general, the accuracy of the |
| 389 | system call, thread and signal simulations is much improved: |
| 390 | |
| 391 | - Blocking system calls behave exactly as they do when running |
| 392 | natively (not on valgrind). That is, if a syscall blocks only the |
| 393 | calling thread when running natively, than it behaves the same on |
| 394 | valgrind. No more mysterious hangs because V doesn't know that some |
| 395 | syscall or other, should block only the calling thread. |
| 396 | |
| 397 | - Interrupted syscalls should now give more faithful results. |
| 398 | |
| 399 | - Signal contexts in signal handlers are supported. |
| 400 | |
| 401 | * Improvements to NPTL support to the extent that V now works |
| 402 | properly on NPTL-only setups. |
| 403 | |
| 404 | * Greater isolation between Valgrind and the program being run, so |
| 405 | the program is less likely to inadvertently kill Valgrind by |
| 406 | doing wild writes. |
| 407 | |
| 408 | * Massif: a new space profiling tool. Try it! It's cool, and it'll |
| 409 | tell you in detail where and when your C/C++ code is allocating heap. |
| 410 | Draws pretty .ps pictures of memory use against time. A potentially |
| 411 | powerful tool for making sense of your program's space use. |
| 412 | |
| 413 | * File descriptor leakage checks. When enabled, Valgrind will print out |
| 414 | a list of open file descriptors on exit. |
| 415 | |
| 416 | * Improved SSE2/SSE3 support. |
| 417 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 418 | * Time-stamped output; use --time-stamp=yes |
| 419 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 420 | |
| 421 | |
| 422 | Stable release 2.2.0 (31 August 2004) -- CHANGES RELATIVE TO 2.1.2 |
| 423 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 424 | 2.2.0 is not much different from 2.1.2, released seven weeks ago. |
| 425 | A number of bugs have been fixed, most notably #85658, which gave |
| 426 | problems for quite a few people. There have been many internal |
| 427 | cleanups, but those are not user visible. |
| 428 | |
| 429 | The following bugs have been fixed since 2.1.2: |
| 430 | |
| 431 | 85658 Assert in coregrind/vg_libpthread.c:2326 (open64) != |
| 432 | (void*)0 failed |
| 433 | This bug was reported multiple times, and so the following |
| 434 | duplicates of it are also fixed: 87620, 85796, 85935, 86065, |
| 435 | 86919, 86988, 87917, 88156 |
| 436 | |
| 437 | 80716 Semaphore mapping bug caused by unmap (sem_destroy) |
| 438 | (Was fixed prior to 2.1.2) |
| 439 | |
| 440 | 86987 semctl and shmctl syscalls family is not handled properly |
| 441 | |
| 442 | 86696 valgrind 2.1.2 + RH AS2.1 + librt |
| 443 | |
| 444 | 86730 valgrind locks up at end of run with assertion failure |
| 445 | in __pthread_unwind |
| 446 | |
| 447 | 86641 memcheck doesn't work with Mesa OpenGL/ATI on Suse 9.1 |
| 448 | (also fixes 74298, a duplicate of this) |
| 449 | |
| 450 | 85947 MMX/SSE unhandled instruction 'sfence' |
| 451 | |
| 452 | 84978 Wrong error "Conditional jump or move depends on |
| 453 | uninitialised value" resulting from "sbbl %reg, %reg" |
| 454 | |
| 455 | 86254 ssort() fails when signed int return type from comparison is |
| 456 | too small to handle result of unsigned int subtraction |
| 457 | |
| 458 | 87089 memalign( 4, xxx) makes valgrind assert |
| 459 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 460 | 86407 Add support for low-level parallel port driver ioctls. |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 461 | |
| 462 | 70587 Add timestamps to Valgrind output? (wishlist) |
| 463 | |
| 464 | 84937 vg_libpthread.c:2505 (se_remap): Assertion `res == 0' |
| 465 | (fixed prior to 2.1.2) |
| 466 | |
| 467 | 86317 cannot load libSDL-1.2.so.0 using valgrind |
| 468 | |
| 469 | 86989 memcpy from mac_replace_strmem.c complains about |
| 470 | uninitialized pointers passed when length to copy is zero |
| 471 | |
| 472 | 85811 gnu pascal symbol causes segmentation fault; ok in 2.0.0 |
| 473 | |
| 474 | 79138 writing to sbrk()'d memory causes segfault |
| 475 | |
| 476 | 77369 sched deadlock while signal received during pthread_join |
| 477 | and the joined thread exited |
| 478 | |
| 479 | 88115 In signal handler for SIGFPE, siginfo->si_addr is wrong |
| 480 | under Valgrind |
| 481 | |
| 482 | 78765 Massif crashes on app exit if FP exceptions are enabled |
| 483 | |
| 484 | Additionally there are the following changes, which are not |
| 485 | connected to any bug report numbers, AFAICS: |
| 486 | |
| 487 | * Fix scary bug causing mis-identification of SSE stores vs |
| 488 | loads and so causing memcheck to sometimes give nonsense results |
| 489 | on SSE code. |
| 490 | |
| 491 | * Add support for the POSIX message queue system calls. |
| 492 | |
| 493 | * Fix to allow 32-bit Valgrind to run on AMD64 boxes. Note: this does |
| 494 | NOT allow Valgrind to work with 64-bit executables - only with 32-bit |
| 495 | executables on an AMD64 box. |
| 496 | |
| 497 | * At configure time, only check whether linux/mii.h can be processed |
| 498 | so that we don't generate ugly warnings by trying to compile it. |
| 499 | |
jseward | 2eb002b | 2004-08-31 00:14:02 +0000 | [diff] [blame] | 500 | * Add support for POSIX clocks and timers. |
| 501 | |
jseward | 34ed74f | 2004-08-30 18:04:42 +0000 | [diff] [blame] | 502 | |
| 503 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 504 | Developer (cvs head) release 2.1.2 (18 July 2004) |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 505 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 506 | 2.1.2 contains four months worth of bug fixes and refinements. |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 507 | Although officially a developer release, we believe it to be stable |
| 508 | enough for widespread day-to-day use. 2.1.2 is pretty good, so try it |
| 509 | first, although there is a chance it won't work. If so then try 2.0.0 |
| 510 | and tell us what went wrong." 2.1.2 fixes a lot of problems present |
| 511 | in 2.0.0 and is generally a much better product. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 512 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 513 | Relative to 2.1.1, a large number of minor problems with 2.1.1 have |
| 514 | been fixed, and so if you use 2.1.1 you should try 2.1.2. Users of |
| 515 | 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] | 516 | |
| 517 | The following bugs, and probably many more, have been fixed. These |
| 518 | are listed at http://bugs.kde.org. Reporting a bug for valgrind in |
| 519 | the http://bugs.kde.org is much more likely to get you a fix than |
| 520 | mailing developers directly, so please continue to keep sending bugs |
| 521 | there. |
| 522 | |
| 523 | 76869 Crashes when running any tool under Fedora Core 2 test1 |
| 524 | This fixes the problem with returning from a signal handler |
jseward | 0400d4d | 2004-07-17 23:26:46 +0000 | [diff] [blame] | 525 | when VDSOs are turned off in FC2. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 526 | |
jseward | 0400d4d | 2004-07-17 23:26:46 +0000 | [diff] [blame] | 527 | 69508 java 1.4.2 client fails with erroneous "stack size too small". |
| 528 | This fix makes more of the pthread stack attribute related |
| 529 | functions work properly. Java still doesn't work though. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 530 | |
| 531 | 71906 malloc alignment should be 8, not 4 |
| 532 | All memory returned by malloc/new etc is now at least |
| 533 | 8-byte aligned. |
| 534 | |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 535 | 81970 vg_alloc_ThreadState: no free slots available |
| 536 | (closed because the workaround is simple: increase |
| 537 | VG_N_THREADS, rebuild and try again.) |
| 538 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 539 | 78514 Conditional jump or move depends on uninitialized value(s) |
| 540 | (a slight mishanding of FP code in memcheck) |
| 541 | |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 542 | 77952 pThread Support (crash) (due to initialisation-ordering probs) |
| 543 | (also 85118) |
| 544 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 545 | 80942 Addrcheck wasn't doing overlap checking as it should. |
| 546 | 78048 return NULL on malloc/new etc failure, instead of asserting |
| 547 | 73655 operator new() override in user .so files often doesn't get picked up |
| 548 | 83060 Valgrind does not handle native kernel AIO |
| 549 | 69872 Create proper coredumps after fatal signals |
| 550 | 82026 failure with new glibc versions: __libc_* functions are not exported |
| 551 | 70344 UNIMPLEMENTED FUNCTION: tcdrain |
| 552 | 81297 Cancellation of pthread_cond_wait does not require mutex |
| 553 | 82872 Using debug info from additional packages (wishlist) |
| 554 | 83025 Support for ioctls FIGETBSZ and FIBMAP |
| 555 | 83340 Support for ioctl HDIO_GET_IDENTITY |
| 556 | 79714 Support for the semtimedop system call. |
| 557 | 77022 Support for ioctls FBIOGET_VSCREENINFO and FBIOGET_FSCREENINFO |
| 558 | 82098 hp2ps ansification (wishlist) |
| 559 | 83573 Valgrind SIGSEGV on execve |
| 560 | 82999 show which cmdline option was erroneous (wishlist) |
| 561 | 83040 make valgrind VPATH and distcheck-clean (wishlist) |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 562 | 83998 Assertion `newfd > vgPlain_max_fd' failed (see below) |
| 563 | 82722 Unchecked mmap in as_pad leads to mysterious failures later |
| 564 | 78958 memcheck seg faults while running Mozilla |
jseward | a2c0800 | 2004-07-18 09:24:57 +0000 | [diff] [blame] | 565 | 85416 Arguments with colon (e.g. --logsocket) ignored |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 566 | |
| 567 | |
| 568 | Additionally there are the following changes, which are not |
| 569 | connected to any bug report numbers, AFAICS: |
| 570 | |
| 571 | * Rearranged address space layout relative to 2.1.1, so that |
| 572 | Valgrind/tools will run out of memory later than currently in many |
| 573 | circumstances. This is good news esp. for Calltree. It should |
| 574 | be possible for client programs to allocate over 800MB of |
| 575 | memory when using memcheck now. |
| 576 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 577 | * Improved checking when laying out memory. Should hopefully avoid |
| 578 | the random segmentation faults that 2.1.1 sometimes caused. |
| 579 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 580 | * Support for Fedora Core 2 and SuSE 9.1. Improvements to NPTL |
| 581 | support to the extent that V now works properly on NPTL-only setups. |
| 582 | |
| 583 | * Renamed the following options: |
| 584 | --logfile-fd --> --log-fd |
| 585 | --logfile --> --log-file |
| 586 | --logsocket --> --log-socket |
| 587 | to be consistent with each other and other options (esp. --input-fd). |
| 588 | |
| 589 | * Add support for SIOCGMIIPHY, SIOCGMIIREG and SIOCSMIIREG ioctls and |
| 590 | improve the checking of other interface related ioctls. |
| 591 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 592 | * Fix building with gcc-3.4.1. |
| 593 | |
| 594 | * Remove limit on number of semaphores supported. |
| 595 | |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 596 | * Add support for syscalls: set_tid_address (258), acct (51). |
| 597 | |
| 598 | * Support instruction "repne movs" -- not official but seems to occur. |
| 599 | |
| 600 | * Implement an emulated soft limit for file descriptors in addition to |
| 601 | the current reserved area, which effectively acts as a hard limit. The |
| 602 | setrlimit system call now simply updates the emulated limits as best |
| 603 | as possible - the hard limit is not allowed to move at all and just |
| 604 | returns EPERM if you try and change it. This should stop reductions |
| 605 | in the soft limit causing assertions when valgrind tries to allocate |
| 606 | descriptors from the reserved area. |
jseward | a6017a4 | 2004-07-17 23:44:35 +0000 | [diff] [blame] | 607 | (This actually came from bug #83998). |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 608 | |
nethercote | 76e8fd9 | 2004-07-21 10:37:31 +0000 | [diff] [blame] | 609 | * Major overhaul of Cachegrind implementation. First user-visible change |
nethercote | 5e63bfb | 2004-07-17 12:49:53 +0000 | [diff] [blame] | 610 | is that cachegrind.out files are now typically 90% smaller than they |
nethercote | 76e8fd9 | 2004-07-21 10:37:31 +0000 | [diff] [blame] | 611 | used to be; code annotation times are correspondingly much smaller. |
| 612 | Second user-visible change is that hit/miss counts for code that is |
| 613 | unloaded at run-time is no longer dumped into a single "discard" pile, |
| 614 | but accurately preserved. |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 615 | |
| 616 | * Client requests for telling valgrind about memory pools. |
| 617 | |
| 618 | |
| 619 | |
jseward | 9cd978c | 2004-07-17 13:38:12 +0000 | [diff] [blame] | 620 | Developer (cvs head) release 2.1.1 (12 March 2004) |
| 621 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
jseward | bd0a6e4 | 2004-07-17 12:19:43 +0000 | [diff] [blame] | 622 | 2.1.1 contains some internal structural changes needed for V's |
jseward | e3e7704 | 2004-03-12 21:07:05 +0000 | [diff] [blame] | 623 | long-term future. These don't affect end-users. Most notable |
| 624 | user-visible changes are: |
| 625 | |
| 626 | * Greater isolation between Valgrind and the program being run, so |
| 627 | the program is less likely to inadvertently kill Valgrind by |
| 628 | doing wild writes. |
| 629 | |
| 630 | * Massif: a new space profiling tool. Try it! It's cool, and it'll |
| 631 | tell you in detail where and when your C/C++ code is allocating heap. |
| 632 | Draws pretty .ps pictures of memory use against time. A potentially |
| 633 | powerful tool for making sense of your program's space use. |
| 634 | |
| 635 | * Fixes for many bugs, including support for more SSE2/SSE3 instructions, |
| 636 | various signal/syscall things, and various problems with debug |
| 637 | info readers. |
| 638 | |
| 639 | * Support for glibc-2.3.3 based systems. |
| 640 | |
| 641 | We are now doing automatic overnight build-and-test runs on a variety |
| 642 | of distros. As a result, we believe 2.1.1 builds and runs on: |
| 643 | Red Hat 7.2, 7.3, 8.0, 9, Fedora Core 1, SuSE 8.2, SuSE 9. |
| 644 | |
| 645 | |
| 646 | The following bugs, and probably many more, have been fixed. These |
| 647 | are listed at http://bugs.kde.org. Reporting a bug for valgrind in |
| 648 | the http://bugs.kde.org is much more likely to get you a fix than |
| 649 | mailing developers directly, so please continue to keep sending bugs |
| 650 | there. |
| 651 | |
| 652 | 69616 glibc 2.3.2 w/NPTL is massively different than what valgrind expects |
| 653 | 69856 I don't know how to instrument MMXish stuff (Helgrind) |
| 654 | 73892 valgrind segfaults starting with Objective-C debug info |
| 655 | (fix for S-type stabs) |
| 656 | 73145 Valgrind complains too much about close(<reserved fd>) |
| 657 | 73902 Shadow memory allocation seems to fail on RedHat 8.0 |
| 658 | 68633 VG_N_SEMAPHORES too low (V itself was leaking semaphores) |
| 659 | 75099 impossible to trace multiprocess programs |
| 660 | 76839 the `impossible' happened: disInstr: INT but not 0x80 ! |
| 661 | 76762 vg_to_ucode.c:3748 (dis_push_segreg): Assertion `sz == 4' failed. |
| 662 | 76747 cannot include valgrind.h in c++ program |
| 663 | 76223 parsing B(3,10) gave NULL type => impossible happens |
| 664 | 75604 shmdt handling problem |
| 665 | 76416 Problems with gcc 3.4 snap 20040225 |
| 666 | 75614 using -gstabs when building your programs the `impossible' happened |
| 667 | 75787 Patch for some CDROM ioctls CDORM_GET_MCN, CDROM_SEND_PACKET, |
| 668 | 75294 gcc 3.4 snapshot's libstdc++ have unsupported instructions. |
| 669 | (REP RET) |
| 670 | 73326 vg_symtab2.c:272 (addScopeRange): Assertion `range->size > 0' failed. |
| 671 | 72596 not recognizing __libc_malloc |
| 672 | 69489 Would like to attach ddd to running program |
| 673 | 72781 Cachegrind crashes with kde programs |
| 674 | 73055 Illegal operand at DXTCV11CompressBlockSSE2 (more SSE opcodes) |
| 675 | 73026 Descriptor leak check reports port numbers wrongly |
| 676 | 71705 README_MISSING_SYSCALL_OR_IOCTL out of date |
| 677 | 72643 Improve support for SSE/SSE2 instructions |
| 678 | 72484 valgrind leaves it's own signal mask in place when execing |
| 679 | 72650 Signal Handling always seems to restart system calls |
| 680 | 72006 The mmap system call turns all errors in ENOMEM |
| 681 | 71781 gdb attach is pretty useless |
| 682 | 71180 unhandled instruction bytes: 0xF 0xAE 0x85 0xE8 |
| 683 | 69886 writes to zero page cause valgrind to assert on exit |
| 684 | 71791 crash when valgrinding gimp 1.3 (stabs reader problem) |
| 685 | 69783 unhandled syscall: 218 |
| 686 | 69782 unhandled instruction bytes: 0x66 0xF 0x2B 0x80 |
| 687 | 70385 valgrind fails if the soft file descriptor limit is less |
| 688 | than about 828 |
| 689 | 69529 "rep; nop" should do a yield |
| 690 | 70827 programs with lots of shared libraries report "mmap failed" |
| 691 | for some of them when reading symbols |
| 692 | 71028 glibc's strnlen is optimised enough to confuse valgrind |
| 693 | |
| 694 | |
| 695 | |
| 696 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 697 | Unstable (cvs head) release 2.1.0 (15 December 2003) |
| 698 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 699 | For whatever it's worth, 2.1.0 actually seems pretty darn stable to me |
| 700 | (Julian). It looks eminently usable, and given that it fixes some |
| 701 | significant bugs, may well be worth using on a day-to-day basis. |
| 702 | 2.1.0 is known to build and pass regression tests on: SuSE 9, SuSE |
| 703 | 8.2, RedHat 8. |
| 704 | |
| 705 | 2.1.0 most notably includes Jeremy Fitzhardinge's complete overhaul of |
| 706 | handling of system calls and signals, and their interaction with |
| 707 | threads. In general, the accuracy of the system call, thread and |
| 708 | signal simulations is much improved. Specifically: |
| 709 | |
| 710 | - Blocking system calls behave exactly as they do when running |
| 711 | natively (not on valgrind). That is, if a syscall blocks only the |
| 712 | calling thread when running natively, than it behaves the same on |
| 713 | valgrind. No more mysterious hangs because V doesn't know that some |
| 714 | syscall or other, should block only the calling thread. |
| 715 | |
| 716 | - Interrupted syscalls should now give more faithful results. |
| 717 | |
| 718 | - Finally, signal contexts in signal handlers are supported. As a |
| 719 | result, konqueror on SuSE 9 no longer segfaults when notified of |
| 720 | file changes in directories it is watching. |
| 721 | |
| 722 | Other changes: |
| 723 | |
| 724 | - Robert Walsh's file descriptor leakage checks. When enabled, |
| 725 | Valgrind will print out a list of open file descriptors on |
| 726 | exit. Along with each file descriptor, Valgrind prints out a stack |
| 727 | backtrace of where the file was opened and any details relating to the |
| 728 | file descriptor such as the file name or socket details. |
| 729 | To use, give: --track-fds=yes |
| 730 | |
| 731 | - Implemented a few more SSE/SSE2 instructions. |
| 732 | |
| 733 | - Less crud on the stack when you do 'where' inside a GDB attach. |
| 734 | |
| 735 | - Fixed the following bugs: |
| 736 | 68360: Valgrind does not compile against 2.6.0-testX kernels |
| 737 | 68525: CVS head doesn't compile on C90 compilers |
| 738 | 68566: pkgconfig support (wishlist) |
| 739 | 68588: Assertion `sz == 4' failed in vg_to_ucode.c (disInstr) |
| 740 | 69140: valgrind not able to explicitly specify a path to a binary. |
| 741 | 69432: helgrind asserts encountering a MutexErr when there are |
| 742 | EraserErr suppressions |
| 743 | |
jseward | 0b12daf | 2003-12-14 14:35:28 +0000 | [diff] [blame] | 744 | - Increase the max size of the translation cache from 200k average bbs |
| 745 | to 300k average bbs. Programs on the size of OOo (680m17) are |
| 746 | thrashing the cache at the smaller size, creating large numbers of |
| 747 | retranslations and wasting significant time as a result. |
| 748 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 749 | |
| 750 | |
| 751 | Stable release 2.0.0 (5 Nov 2003) |
| 752 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 753 | |
| 754 | 2.0.0 improves SSE/SSE2 support, fixes some minor bugs, and |
| 755 | improves support for SuSE 9 and the Red Hat "Severn" beta. |
| 756 | |
| 757 | - Further improvements to SSE/SSE2 support. The entire test suite of |
| 758 | the GNU Scientific Library (gsl-1.4) compiled with Intel Icc 7.1 |
| 759 | 20030307Z '-g -O -xW' now works. I think this gives pretty good |
| 760 | coverage of SSE/SSE2 floating point instructions, or at least the |
| 761 | subset emitted by Icc. |
| 762 | |
| 763 | - Also added support for the following instructions: |
| 764 | MOVNTDQ UCOMISD UNPCKLPS UNPCKHPS SQRTSS |
| 765 | PUSH/POP %{FS,GS}, and PUSH %CS (Nb: there is no POP %CS). |
| 766 | |
| 767 | - CFI support for GDB version 6. Needed to enable newer GDBs |
| 768 | to figure out where they are when using --gdb-attach=yes. |
| 769 | |
| 770 | - Fix this: |
| 771 | mc_translate.c:1091 (memcheck_instrument): Assertion |
| 772 | `u_in->size == 4 || u_in->size == 16' failed. |
| 773 | |
| 774 | - Return an error rather than panicing when given a bad socketcall. |
| 775 | |
| 776 | - Fix checking of syscall rt_sigtimedwait(). |
| 777 | |
| 778 | - Implement __NR_clock_gettime (syscall 265). Needed on Red Hat Severn. |
| 779 | |
| 780 | - Fixed bug in overlap check in strncpy() -- it was assuming the src was 'n' |
| 781 | bytes long, when it could be shorter, which could cause false |
| 782 | positives. |
| 783 | |
| 784 | - Support use of select() for very large numbers of file descriptors. |
| 785 | |
| 786 | - Don't fail silently if the executable is statically linked, or is |
| 787 | setuid/setgid. Print an error message instead. |
| 788 | |
| 789 | - Support for old DWARF-1 format line number info. |
| 790 | |
| 791 | |
| 792 | |
| 793 | Snapshot 20031012 (12 October 2003) |
| 794 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 795 | |
| 796 | Three months worth of bug fixes, roughly. Most significant single |
| 797 | change is improved SSE/SSE2 support, mostly thanks to Dirk Mueller. |
| 798 | |
| 799 | 20031012 builds on Red Hat Fedora ("Severn") but doesn't really work |
| 800 | (curiosly, mozilla runs OK, but a modest "ls -l" bombs). I hope to |
| 801 | get a working version out soon. It may or may not work ok on the |
| 802 | forthcoming SuSE 9; I hear positive noises about it but haven't been |
| 803 | able to verify this myself (not until I get hold of a copy of 9). |
| 804 | |
| 805 | A detailed list of changes, in no particular order: |
| 806 | |
| 807 | - Describe --gen-suppressions in the FAQ. |
| 808 | |
| 809 | - Syscall __NR_waitpid supported. |
| 810 | |
| 811 | - Minor MMX bug fix. |
| 812 | |
| 813 | - -v prints program's argv[] at startup. |
| 814 | |
| 815 | - More glibc-2.3 suppressions. |
| 816 | |
| 817 | - Suppressions for stack underrun bug(s) in the c++ support library |
| 818 | distributed with Intel Icc 7.0. |
| 819 | |
| 820 | - Fix problems reading /proc/self/maps. |
| 821 | |
| 822 | - Fix a couple of messages that should have been suppressed by -q, |
| 823 | but weren't. |
| 824 | |
| 825 | - Make Addrcheck understand "Overlap" suppressions. |
| 826 | |
| 827 | - At startup, check if program is statically linked and bail out if so. |
| 828 | |
| 829 | - Cachegrind: Auto-detect Intel Pentium-M, also VIA Nehemiah |
| 830 | |
| 831 | - Memcheck/addrcheck: minor speed optimisations |
| 832 | |
| 833 | - Handle syscall __NR_brk more correctly than before. |
| 834 | |
| 835 | - Fixed incorrect allocate/free mismatch errors when using |
| 836 | operator new(unsigned, std::nothrow_t const&) |
| 837 | operator new[](unsigned, std::nothrow_t const&) |
| 838 | |
| 839 | - Support POSIX pthread spinlocks. |
| 840 | |
| 841 | - Fixups for clean compilation with gcc-3.3.1. |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 842 | |
| 843 | - Implemented more opcodes: |
| 844 | - push %es |
| 845 | - push %ds |
| 846 | - pop %es |
| 847 | - pop %ds |
| 848 | - movntq |
| 849 | - sfence |
| 850 | - pshufw |
| 851 | - pavgb |
| 852 | - ucomiss |
| 853 | - enter |
daywalker | b18d253 | 2003-09-27 20:15:01 +0000 | [diff] [blame] | 854 | - mov imm32, %esp |
| 855 | - all "in" and "out" opcodes |
daywalker | 79aad84 | 2003-09-30 22:58:12 +0000 | [diff] [blame] | 856 | - inc/dec %esp |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 857 | - A whole bunch of SSE/SSE2 instructions |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 858 | |
jseward | 4323065 | 2003-12-14 00:14:54 +0000 | [diff] [blame] | 859 | - Memcheck: don't bomb on SSE/SSE2 code. |
daywalker | 5d945de | 2003-09-26 00:32:53 +0000 | [diff] [blame] | 860 | |
sewardj | 9023879 | 2003-05-05 00:23:42 +0000 | [diff] [blame] | 861 | |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 862 | Snapshot 20030725 (25 July 2003) |
| 863 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 864 | |
| 865 | Fixes some minor problems in 20030716. |
| 866 | |
| 867 | - Fix bugs in overlap checking for strcpy/memcpy etc. |
| 868 | |
| 869 | - Do overlap checking with Addrcheck as well as Memcheck. |
| 870 | |
| 871 | - Fix this: |
| 872 | Memcheck: the `impossible' happened: |
| 873 | get_error_name: unexpected type |
| 874 | |
| 875 | - Install headers needed to compile new skins. |
| 876 | |
sewardj | e3dd2e0 | 2003-07-27 08:04:48 +0000 | [diff] [blame] | 877 | - Remove leading spaces and colon in the LD_LIBRARY_PATH / LD_PRELOAD |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 878 | passed to non-traced children. |
| 879 | |
sewardj | b9eda95 | 2003-07-26 21:39:05 +0000 | [diff] [blame] | 880 | - Fix file descriptor leak in valgrind-listener. |
| 881 | |
sewardj | e3dd2e0 | 2003-07-27 08:04:48 +0000 | [diff] [blame] | 882 | - Fix longstanding bug in which the allocation point of a |
| 883 | block resized by realloc was not correctly set. This may |
| 884 | have caused confusing error messages. |
sewardj | 945f39f | 2003-07-25 21:11:40 +0000 | [diff] [blame] | 885 | |
| 886 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 887 | Snapshot 20030716 (16 July 2003) |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 888 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 889 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 890 | 20030716 is a snapshot of our current CVS head (development) branch. |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 891 | This is the branch which will become valgrind-2.0. It contains |
| 892 | significant enhancements over the 1.9.X branch. |
| 893 | |
| 894 | Despite this being a snapshot of the CVS head, it is believed to be |
| 895 | quite stable -- at least as stable as 1.9.6 or 1.0.4, if not more so |
| 896 | -- and therefore suitable for widespread use. Please let us know asap |
| 897 | if it causes problems for you. |
| 898 | |
| 899 | Two reasons for releasing a snapshot now are: |
| 900 | |
| 901 | - It's been a while since 1.9.6, and this snapshot fixes |
| 902 | various problems that 1.9.6 has with threaded programs |
| 903 | on glibc-2.3.X based systems. |
| 904 | |
| 905 | - So as to make available improvements in the 2.0 line. |
| 906 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 907 | Major changes in 20030716, as compared to 1.9.6: |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 908 | |
| 909 | - More fixes to threading support on glibc-2.3.1 and 2.3.2-based |
| 910 | systems (SuSE 8.2, Red Hat 9). If you have had problems |
| 911 | with inconsistent/illogical behaviour of errno, h_errno or the DNS |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 912 | resolver functions in threaded programs, 20030716 should improve |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 913 | matters. This snapshot seems stable enough to run OpenOffice.org |
| 914 | 1.1rc on Red Hat 7.3, SuSE 8.2 and Red Hat 9, and that's a big |
| 915 | threaded app if ever I saw one. |
| 916 | |
| 917 | - Automatic generation of suppression records; you no longer |
| 918 | need to write them by hand. Use --gen-suppressions=yes. |
| 919 | |
sewardj | 2151180 | 2003-07-22 17:47:42 +0000 | [diff] [blame] | 920 | - strcpy/memcpy/etc check their arguments for overlaps, when |
| 921 | running with the Memcheck or Addrcheck skins. |
| 922 | |
| 923 | - malloc_usable_size() is now supported. |
| 924 | |
| 925 | - new client requests: |
| 926 | - VALGRIND_COUNT_ERRORS, VALGRIND_COUNT_LEAKS: |
| 927 | useful with regression testing |
| 928 | - VALGRIND_NON_SIMD_CALL[0123]: for running arbitrary functions |
| 929 | on real CPU (use with caution!) |
| 930 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 931 | - The GDB attach mechanism is more flexible. Allow the GDB to |
| 932 | be run to be specified by --gdb-path=/path/to/gdb, and specify |
| 933 | which file descriptor V will read its input from with |
| 934 | --input-fd=<number>. |
| 935 | |
sewardj | 2151180 | 2003-07-22 17:47:42 +0000 | [diff] [blame] | 936 | - Cachegrind gives more accurate results (wasn't tracking instructions in |
| 937 | malloc() and friends previously, is now). |
| 938 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 939 | - Complete support for the MMX instruction set. |
| 940 | |
| 941 | - Partial support for the SSE and SSE2 instruction sets. Work for this |
| 942 | is ongoing. About half the SSE/SSE2 instructions are done, so |
| 943 | some SSE based programs may work. Currently you need to specify |
| 944 | --skin=addrcheck. Basically not suitable for real use yet. |
| 945 | |
| 946 | - Significant speedups (10%-20%) for standard memory checking. |
| 947 | |
| 948 | - Fix assertion failure in pthread_once(). |
| 949 | |
| 950 | - Fix this: |
| 951 | valgrind: vg_intercept.c:598 (vgAllRoadsLeadToRome_select): |
| 952 | Assertion `ms_end >= ms_now' failed. |
| 953 | |
| 954 | - Implement pthread_mutexattr_setpshared. |
| 955 | |
| 956 | - Understand Pentium 4 branch hints. Also implemented a couple more |
| 957 | obscure x86 instructions. |
| 958 | |
| 959 | - Lots of other minor bug fixes. |
| 960 | |
sewardj | 626fd89 | 2003-07-16 20:10:26 +0000 | [diff] [blame] | 961 | - We have a decent regression test system, for the first time. |
| 962 | This doesn't help you directly, but it does make it a lot easier |
| 963 | for us to track the quality of the system, especially across |
| 964 | multiple linux distributions. |
| 965 | |
| 966 | You can run the regression tests with 'make regtest' after 'make |
| 967 | install' completes. On SuSE 8.2 and Red Hat 9 I get this: |
| 968 | |
| 969 | == 84 tests, 0 stderr failures, 0 stdout failures == |
| 970 | |
| 971 | On Red Hat 8, I get this: |
| 972 | |
| 973 | == 84 tests, 2 stderr failures, 1 stdout failure == |
| 974 | corecheck/tests/res_search (stdout) |
| 975 | memcheck/tests/sigaltstack (stderr) |
| 976 | |
| 977 | sigaltstack is probably harmless. res_search doesn't work |
| 978 | on R H 8 even running natively, so I'm not too worried. |
| 979 | |
| 980 | On Red Hat 7.3, a glibc-2.2.5 system, I get these harmless failures: |
| 981 | |
| 982 | == 84 tests, 2 stderr failures, 1 stdout failure == |
| 983 | corecheck/tests/pth_atfork1 (stdout) |
| 984 | corecheck/tests/pth_atfork1 (stderr) |
| 985 | memcheck/tests/sigaltstack (stderr) |
| 986 | |
| 987 | You need to run on a PII system, at least, since some tests |
| 988 | contain P6-specific instructions, and the test machine needs |
| 989 | access to the internet so that corecheck/tests/res_search |
| 990 | (a test that the DNS resolver works) can function. |
| 991 | |
sewardj | 9d916ed | 2003-07-14 23:38:40 +0000 | [diff] [blame] | 992 | As ever, thanks for the vast amount of feedback :) and bug reports :( |
| 993 | We may not answer all messages, but we do at least look at all of |
| 994 | them, and tend to fix the most frequently reported bugs. |
| 995 | |
| 996 | |
| 997 | |
sewardj | 3791882 | 2003-05-05 01:05:09 +0000 | [diff] [blame] | 998 | Version 1.9.6 (7 May 2003 or thereabouts) |
| 999 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 1000 | |
| 1001 | Major changes in 1.9.6: |
| 1002 | |
| 1003 | - Improved threading support for glibc >= 2.3.2 (SuSE 8.2, |
| 1004 | RedHat 9, to name but two ...) It turned out that 1.9.5 |
| 1005 | had problems with threading support on glibc >= 2.3.2, |
| 1006 | usually manifested by threaded programs deadlocking in system calls, |
| 1007 | or running unbelievably slowly. Hopefully these are fixed now. 1.9.6 |
| 1008 | is the first valgrind which gives reasonable support for |
| 1009 | glibc-2.3.2. Also fixed a 2.3.2 problem with pthread_atfork(). |
| 1010 | |
| 1011 | - Majorly expanded FAQ.txt. We've added workarounds for all |
| 1012 | common problems for which a workaround is known. |
| 1013 | |
| 1014 | Minor changes in 1.9.6: |
| 1015 | |
| 1016 | - Fix identification of the main thread's stack. Incorrect |
| 1017 | identification of it was causing some on-stack addresses to not get |
| 1018 | identified as such. This only affected the usefulness of some error |
| 1019 | messages; the correctness of the checks made is unchanged. |
| 1020 | |
| 1021 | - Support for kernels >= 2.5.68. |
| 1022 | |
| 1023 | - Dummy implementations of __libc_current_sigrtmin, |
| 1024 | __libc_current_sigrtmax and __libc_allocate_rtsig, hopefully |
| 1025 | good enough to keep alive programs which previously died for lack of |
| 1026 | them. |
| 1027 | |
| 1028 | - Fix bug in the VALGRIND_DISCARD_TRANSLATIONS client request. |
| 1029 | |
sewardj | 3d47b79 | 2003-05-05 22:15:35 +0000 | [diff] [blame] | 1030 | - Fix bug in the DWARF2 debug line info loader, when instructions |
| 1031 | following each other have source lines far from each other |
| 1032 | (e.g. with inlined functions). |
| 1033 | |
sewardj | 3791882 | 2003-05-05 01:05:09 +0000 | [diff] [blame] | 1034 | - Debug info reading: read symbols from both "symtab" and "dynsym" |
| 1035 | sections, rather than merely from the one that comes last in the |
| 1036 | file. |
| 1037 | |
| 1038 | - New syscall support: prctl(), creat(), lookup_dcookie(). |
| 1039 | |
| 1040 | - When checking calls to accept(), recvfrom(), getsocketopt(), |
| 1041 | don't complain if buffer values are NULL. |
| 1042 | |
| 1043 | - Try and avoid assertion failures in |
| 1044 | mash_LD_PRELOAD_and_LD_LIBRARY_PATH. |
| 1045 | |
| 1046 | - Minor bug fixes in cg_annotate. |
| 1047 | |
| 1048 | |
| 1049 | |
sewardj | 9023879 | 2003-05-05 00:23:42 +0000 | [diff] [blame] | 1050 | Version 1.9.5 (7 April 2003) |
| 1051 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 1052 | |
| 1053 | It occurs to me that it would be helpful for valgrind users to record |
| 1054 | in the source distribution the changes in each release. So I now |
| 1055 | attempt to mend my errant ways :-) Changes in this and future releases |
| 1056 | will be documented in the NEWS file in the source distribution. |
| 1057 | |
| 1058 | Major changes in 1.9.5: |
| 1059 | |
| 1060 | - (Critical bug fix): Fix a bug in the FPU simulation. This was |
| 1061 | causing some floating point conditional tests not to work right. |
| 1062 | Several people reported this. If you had floating point code which |
| 1063 | didn't work right on 1.9.1 to 1.9.4, it's worth trying 1.9.5. |
| 1064 | |
| 1065 | - Partial support for Red Hat 9. RH9 uses the new Native Posix |
| 1066 | Threads Library (NPTL), instead of the older LinuxThreads. |
| 1067 | This potentially causes problems with V which will take some |
| 1068 | time to correct. In the meantime we have partially worked around |
| 1069 | this, and so 1.9.5 works on RH9. Threaded programs still work, |
| 1070 | but they may deadlock, because some system calls (accept, read, |
| 1071 | write, etc) which should be nonblocking, in fact do block. This |
| 1072 | is a known bug which we are looking into. |
| 1073 | |
| 1074 | If you can, your best bet (unfortunately) is to avoid using |
| 1075 | 1.9.5 on a Red Hat 9 system, or on any NPTL-based distribution. |
| 1076 | If your glibc is 2.3.1 or earlier, you're almost certainly OK. |
| 1077 | |
| 1078 | Minor changes in 1.9.5: |
| 1079 | |
| 1080 | - Added some #errors to valgrind.h to ensure people don't include |
| 1081 | it accidentally in their sources. This is a change from 1.0.X |
| 1082 | which was never properly documented. The right thing to include |
| 1083 | is now memcheck.h. Some people reported problems and strange |
| 1084 | behaviour when (incorrectly) including valgrind.h in code with |
| 1085 | 1.9.1 -- 1.9.4. This is no longer possible. |
| 1086 | |
| 1087 | - Add some __extension__ bits and pieces so that gcc configured |
| 1088 | for valgrind-checking compiles even with -Werror. If you |
| 1089 | don't understand this, ignore it. Of interest to gcc developers |
| 1090 | only. |
| 1091 | |
| 1092 | - Removed a pointless check which caused problems interworking |
| 1093 | with Clearcase. V would complain about shared objects whose |
| 1094 | names did not end ".so", and refuse to run. This is now fixed. |
| 1095 | In fact it was fixed in 1.9.4 but not documented. |
| 1096 | |
| 1097 | - Fixed a bug causing an assertion failure of "waiters == 1" |
| 1098 | somewhere in vg_scheduler.c, when running large threaded apps, |
| 1099 | notably MySQL. |
| 1100 | |
| 1101 | - Add support for the munlock system call (124). |
| 1102 | |
| 1103 | Some comments about future releases: |
| 1104 | |
| 1105 | 1.9.5 is, we hope, the most stable Valgrind so far. It pretty much |
| 1106 | supersedes the 1.0.X branch. If you are a valgrind packager, please |
| 1107 | consider making 1.9.5 available to your users. You can regard the |
| 1108 | 1.0.X branch as obsolete: 1.9.5 is stable and vastly superior. There |
| 1109 | are no plans at all for further releases of the 1.0.X branch. |
| 1110 | |
| 1111 | If you want a leading-edge valgrind, consider building the cvs head |
| 1112 | (from SourceForge), or getting a snapshot of it. Current cool stuff |
| 1113 | going in includes MMX support (done); SSE/SSE2 support (in progress), |
| 1114 | a significant (10-20%) performance improvement (done), and the usual |
| 1115 | large collection of minor changes. Hopefully we will be able to |
| 1116 | improve our NPTL support, but no promises. |
| 1117 | |