njn | e43d3ae | 2003-05-05 13:04:49 +0000 | [diff] [blame] | 1 | |
njn | e43d3ae | 2003-05-05 13:04:49 +0000 | [diff] [blame] | 2 | Building and not installing it |
| 3 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
njn | 090dc84 | 2005-03-12 16:47:07 +0000 | [diff] [blame] | 4 | To run Valgrind without having to install it, run coregrind/valgrind |
sewardj | 45f4e7c | 2005-09-27 19:20:21 +0000 | [diff] [blame] | 5 | with the VALGRIND_LIB environment variable set, where <dir> is the root |
njn | 090dc84 | 2005-03-12 16:47:07 +0000 | [diff] [blame] | 6 | of the source tree (and must be an absolute path). Eg: |
| 7 | |
sewardj | 45f4e7c | 2005-09-27 19:20:21 +0000 | [diff] [blame] | 8 | VALGRIND_LIB=~/grind/head4/.in_place ~/grind/head4/coregrind/valgrind |
njn | e43d3ae | 2003-05-05 13:04:49 +0000 | [diff] [blame] | 9 | |
| 10 | This allows you to compile and run with "make" instead of "make install", |
| 11 | saving you time. |
| 12 | |
njn | 7bbc8d6 | 2007-02-19 04:09:24 +0000 | [diff] [blame] | 13 | Or, you can use the 'vg-in-place' script which does that for you. |
| 14 | |
njn | e43d3ae | 2003-05-05 13:04:49 +0000 | [diff] [blame] | 15 | I recommend compiling with "make --quiet" to further reduce the amount of |
| 16 | output spewed out during compilation, letting you actually see any errors, |
| 17 | warnings, etc. |
| 18 | |
| 19 | |
florian | cb8ea55 | 2011-10-19 21:56:25 +0000 | [diff] [blame] | 20 | Building a distribution tarball |
| 21 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 22 | To build a distribution tarball from the valgrind sources: |
| 23 | |
| 24 | make dist |
| 25 | |
| 26 | In addition to compiling, linking and packaging everything up, the command |
| 27 | will also build the documentation. Even if all required tools for building the |
| 28 | documentation are installed, this step may not succeed because of hidden |
| 29 | dependencies. E.g. on Ubuntu you must have "docbook-xsl" installed. |
| 30 | Additionally, specific tool versions maybe needed. |
| 31 | |
| 32 | If you only want to test whether the generated tarball is complete and runs |
| 33 | regression tests successfully, building documentation is not needed. |
| 34 | Edit docs/Makefile.am, search for BUILD_ALL_DOCS and follow instructions there. |
| 35 | |
| 36 | |
njn | e43d3ae | 2003-05-05 13:04:49 +0000 | [diff] [blame] | 37 | Running the regression tests |
| 38 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 39 | To build and run all the regression tests, run "make [--quiet] regtest". |
| 40 | |
| 41 | To run a subset of the regression tests, execute: |
| 42 | |
| 43 | perl tests/vg_regtest <name> |
| 44 | |
| 45 | where <name> is a directory (all tests within will be run) or a single |
| 46 | .vgtest test file, or the name of a program which has a like-named .vgtest |
| 47 | file. Eg: |
| 48 | |
| 49 | perl tests/vg_regtest memcheck |
| 50 | perl tests/vg_regtest memcheck/tests/badfree.vgtest |
| 51 | perl tests/vg_regtest memcheck/tests/badfree |
| 52 | |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 53 | |
njn | 6f58249 | 2006-06-02 23:59:40 +0000 | [diff] [blame] | 54 | Running the performance tests |
njn | 5359b6f | 2006-06-02 23:57:22 +0000 | [diff] [blame] | 55 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
njn | 6f58249 | 2006-06-02 23:59:40 +0000 | [diff] [blame] | 56 | To build and run all the performance tests, run "make [--quiet] perf". |
njn | 5359b6f | 2006-06-02 23:57:22 +0000 | [diff] [blame] | 57 | |
njn | 6f58249 | 2006-06-02 23:59:40 +0000 | [diff] [blame] | 58 | To run a subset of the performance suite, execute: |
njn | 5359b6f | 2006-06-02 23:57:22 +0000 | [diff] [blame] | 59 | |
| 60 | perl perf/vg_perf <name> |
| 61 | |
| 62 | where <name> is a directory (all tests within will be run) or a single |
| 63 | .vgperf test file, or the name of a program which has a like-named .vgperf |
| 64 | file. Eg: |
| 65 | |
| 66 | perl perf/vg_perf perf/ |
| 67 | perl perf/vg_perf perf/bz2.vgperf |
| 68 | perl perf/vg_perf perf/bz2 |
| 69 | |
njn | 6f58249 | 2006-06-02 23:59:40 +0000 | [diff] [blame] | 70 | To compare multiple versions of Valgrind, use the --vg= option multiple |
njn | 5359b6f | 2006-06-02 23:57:22 +0000 | [diff] [blame] | 71 | times. For example, if you have two Valgrinds next to each other, one in |
njn | 6f58249 | 2006-06-02 23:59:40 +0000 | [diff] [blame] | 72 | trunk1/ and one in trunk2/, from within either trunk1/ or trunk2/ do this to |
| 73 | compare them on all the performance tests: |
njn | 5359b6f | 2006-06-02 23:57:22 +0000 | [diff] [blame] | 74 | |
| 75 | perl perf/vg_perf --vg=../trunk1 --vg=../trunk2 perf/ |
| 76 | |
| 77 | |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 78 | Debugging Valgrind with GDB |
| 79 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
tom | 4a5223b | 2005-11-17 12:31:12 +0000 | [diff] [blame] | 80 | To debug the valgrind launcher program (<prefix>/bin/valgrind) just |
| 81 | run it under gdb in the normal way. |
nethercote | 4fffabd | 2004-11-02 09:13:12 +0000 | [diff] [blame] | 82 | |
tom | 4a5223b | 2005-11-17 12:31:12 +0000 | [diff] [blame] | 83 | Debugging the main body of the valgrind code (and/or the code for |
| 84 | a particular tool) requires a bit more trickery but can be achieved |
| 85 | without too much problem by following these steps: |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 86 | |
njn | 90c8192 | 2007-09-17 22:35:57 +0000 | [diff] [blame] | 87 | (1) Set VALGRIND_LAUNCHER to point to the valgrind executable. Eg: |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 88 | |
njn | 90c8192 | 2007-09-17 22:35:57 +0000 | [diff] [blame] | 89 | export VALGRIND_LAUNCHER=/usr/local/bin/valgrind |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 90 | |
njn | 90c8192 | 2007-09-17 22:35:57 +0000 | [diff] [blame] | 91 | or for an uninstalled version in a source directory $DIR: |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 92 | |
njn | 90c8192 | 2007-09-17 22:35:57 +0000 | [diff] [blame] | 93 | export VALGRIND_LAUNCHER=$DIR/coregrind/valgrind |
| 94 | |
| 95 | (2) Run gdb on the tool executable. Eg: |
| 96 | |
| 97 | gdb /usr/local/lib/valgrind/ppc32-linux/lackey |
| 98 | |
| 99 | or |
| 100 | |
| 101 | gdb $DIR/.in_place/x86-linux/memcheck |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 102 | |
tom | 4a5223b | 2005-11-17 12:31:12 +0000 | [diff] [blame] | 103 | (3) Do "handle SIGSEGV SIGILL nostop noprint" in GDB to prevent GDB from |
| 104 | stopping on a SIGSEGV or SIGILL: |
| 105 | |
| 106 | (gdb) handle SIGILL SIGSEGV nostop noprint |
| 107 | |
| 108 | (4) Set any breakpoints you want and proceed as normal for gdb. The |
| 109 | macro VG_(FUNC) is expanded to vgPlain_FUNC, so If you want to set |
| 110 | a breakpoint VG_(do_exec), you could do like this in GDB: |
| 111 | |
| 112 | (gdb) b vgPlain_do_exec |
| 113 | |
weidendo | 4089f32 | 2012-08-06 12:27:51 +0000 | [diff] [blame^] | 114 | (5) Run the tool with required options (the --tool option is required |
| 115 | for correct setup), e.g. |
tom | 4a5223b | 2005-11-17 12:31:12 +0000 | [diff] [blame] | 116 | |
weidendo | 4089f32 | 2012-08-06 12:27:51 +0000 | [diff] [blame^] | 117 | (gdb) run --tool=lackey pwd |
nethercote | 16b59ee | 2004-10-09 15:59:05 +0000 | [diff] [blame] | 118 | |
njn | 90c8192 | 2007-09-17 22:35:57 +0000 | [diff] [blame] | 119 | Steps (1)--(3) can be put in a .gdbinit file, but any directory names must |
| 120 | be fully expanded (ie. not an environment variable). |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 121 | |
sewardj | e089f01 | 2010-10-13 21:47:29 +0000 | [diff] [blame] | 122 | A different and possibly easier way is as follows: |
| 123 | |
| 124 | (1) Run Valgrind as normal, but add the flag --wait-for-gdb=yes. This |
| 125 | puts the tool executable into a wait loop soon after it gains |
| 126 | control. This delays startup for a few seconds. |
| 127 | |
| 128 | (2) In a different shell, do "gdb /proc/<pid>/exe <pid>", where |
| 129 | <pid> you read from the output printed by (1). This attaches |
| 130 | GDB to the tool executable, which should be in the abovementioned |
| 131 | wait loop. |
| 132 | |
| 133 | (3) Do "cont" to continue. After the loop finishes spinning, startup |
| 134 | will continue as normal. Note that comment (3) above re passing |
| 135 | signals applies here too. |
| 136 | |
njn | 4be4e2a | 2009-06-12 23:40:04 +0000 | [diff] [blame] | 137 | |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 138 | Self-hosting |
| 139 | ~~~~~~~~~~~~ |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 140 | This section explains : |
| 141 | (A) How to configure Valgrind to run under Valgrind. |
| 142 | Such a setup is called self hosting, or outer/inner setup. |
| 143 | (B) How to run Valgrind regression tests in a 'self-hosting' mode, |
| 144 | e.g. to verify Valgrind has no bugs such as memory leaks. |
| 145 | (C) How to run Valgrind performance tests in a 'self-hosting' mode, |
| 146 | to analyse and optimise the performance of Valgrind and its tools. |
| 147 | |
| 148 | (A) How to configure Valgrind to run under Valgrind: |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 149 | |
njn | 801288e | 2009-08-04 07:02:54 +0000 | [diff] [blame] | 150 | (1) Check out 2 trees, "Inner" and "Outer". Inner runs the app |
| 151 | directly. Outer runs Inner. |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 152 | |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 153 | (2) Configure inner with --enable-inner and build/install as usual. |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 154 | |
njn | 801288e | 2009-08-04 07:02:54 +0000 | [diff] [blame] | 155 | (3) Configure Outer normally and build/install as usual. |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 156 | |
| 157 | (4) Choose a very simple program (date) and try |
| 158 | |
njn | 86d6807 | 2005-11-12 19:07:45 +0000 | [diff] [blame] | 159 | outer/.../bin/valgrind --sim-hints=enable-outer --trace-children=yes \ |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 160 | --smc-check=all-non-file \ |
philippe | 277eaff | 2012-03-03 12:01:48 +0000 | [diff] [blame] | 161 | --run-libc-freeres=no --tool=cachegrind -v \ |
| 162 | inner/.../bin/valgrind --vgdb-prefix=./inner --tool=none -v prog |
njn | 585b8b3 | 2005-10-10 11:36:55 +0000 | [diff] [blame] | 163 | |
njn | 801288e | 2009-08-04 07:02:54 +0000 | [diff] [blame] | 164 | If you omit the --trace-children=yes, you'll only monitor Inner's launcher |
philippe | 277eaff | 2012-03-03 12:01:48 +0000 | [diff] [blame] | 165 | program, not its stage2. Outer needs --run-libc-freeres=no, as otherwise |
| 166 | it will try to find and run __libc_freeres in the inner, while libc is not |
| 167 | used by the inner. Inner needs --vgdb-prefix=./inner to avoid inner |
| 168 | gdbserver colliding with outer gdbserver. |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 169 | Currently, inner does *not* use the client request |
| 170 | VALGRIND_DISCARD_TRANSLATIONS for the JITted code or the code patched for |
| 171 | translation chaining. So the outer needs --smc-check=all-non-file to |
| 172 | detect the modified code. |
philippe | 277eaff | 2012-03-03 12:01:48 +0000 | [diff] [blame] | 173 | |
| 174 | Debugging the whole thing might imply to use up to 3 GDB: |
| 175 | * a GDB attached to the Outer valgrind, allowing |
| 176 | to examine the state of Outer. |
| 177 | * a GDB using Outer gdbserver, allowing to |
| 178 | examine the state of Inner. |
| 179 | * a GDB using Inner gdbserver, allowing to |
| 180 | examine the state of prog. |
njn | 7cce5b8 | 2005-11-16 20:12:22 +0000 | [diff] [blame] | 181 | |
| 182 | The whole thing is fragile, confusing and slow, but it does work well enough |
njn | 801288e | 2009-08-04 07:02:54 +0000 | [diff] [blame] | 183 | for you to get some useful performance data. Inner has most of |
njn | 7cce5b8 | 2005-11-16 20:12:22 +0000 | [diff] [blame] | 184 | its output (ie. those lines beginning with "==<pid>==") prefixed with a '>', |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 185 | which helps a lot. However, when running regression tests in an Outer/Inner |
| 186 | setup, this prefix causes the reg test diff to fail. Give |
| 187 | --sim-hints=no-inner-prefix to the Inner to disable the production |
| 188 | of the prefix in the stdout/stderr output of Inner. |
njn | 15a6563 | 2005-10-10 11:43:14 +0000 | [diff] [blame] | 189 | |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 190 | The allocator (coregrind/m_mallocfree.c) is annotated with client requests |
| 191 | so Memcheck can be used to find leaks and use after free in an Inner |
| 192 | Valgrind. |
| 193 | |
| 194 | The Valgrind "big lock" is annotated with helgrind client requests |
| 195 | so helgrind and drd can be used to find race conditions in an Inner |
| 196 | Valgrind. |
| 197 | |
| 198 | All this has not been tested much, so don't be surprised if you hit problems. |
njn | 0dc09e8 | 2005-11-03 16:24:53 +0000 | [diff] [blame] | 199 | |
njn | 801288e | 2009-08-04 07:02:54 +0000 | [diff] [blame] | 200 | When using self-hosting with an outer Callgrind tool, use '--pop-on-jump' |
| 201 | (on the outer). Otherwise, Callgrind has much higher memory requirements. |
weidendo | 10e80e3 | 2006-05-01 01:49:28 +0000 | [diff] [blame] | 202 | |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 203 | (B) Regression tests in an outer/inner setup: |
| 204 | |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 205 | To run all the regression tests with an outer memcheck, do : |
philippe | 8730e0f | 2012-05-06 21:29:59 +0000 | [diff] [blame] | 206 | perl tests/vg_regtest --outer-valgrind=../outer/.../bin/valgrind \ |
| 207 | --all |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 208 | |
| 209 | To run a specific regression tests with an outer memcheck, do: |
philippe | 8730e0f | 2012-05-06 21:29:59 +0000 | [diff] [blame] | 210 | perl tests/vg_regtest --outer-valgrind=../outer/.../bin/valgrind \ |
| 211 | none/tests/args.vgtest |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 212 | |
| 213 | To run regression tests with another outer tool: |
| 214 | perl tests/vg_regtest --outer-valgrind=../outer/.../bin/valgrind \ |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 215 | --outer-tool=helgrind --all |
philippe | 72faf10 | 2012-03-11 22:24:03 +0000 | [diff] [blame] | 216 | |
| 217 | --outer-args allows to give specific arguments to the outer tool, |
| 218 | replacing the default one provided by vg_regtest. |
| 219 | |
| 220 | When an outer valgrind runs an inner valgrind, a regression test |
| 221 | produces one additional file <testname>.outer.log which contains the |
| 222 | errors detected by the outer valgrind. E.g. for an outer memcheck, it |
| 223 | contains the leaks found in the inner, for an outer helgrind or drd, |
| 224 | it contains the detected race conditions. |
| 225 | |
| 226 | The file tests/outer_inner.supp contains suppressions for |
| 227 | the irrelevant or benign errors found in the inner. |
njn | 0b5efe7 | 2005-11-10 03:40:36 +0000 | [diff] [blame] | 228 | |
philippe | 14ab1a3 | 2012-04-08 19:52:38 +0000 | [diff] [blame] | 229 | (C) Performance tests in an outer/inner setup: |
| 230 | |
| 231 | To run all the performance tests with an outer cachegrind, do : |
| 232 | perl perf/vg_perf --outer-valgrind=../outer/.../bin/valgrind perf |
| 233 | |
| 234 | To run a specific perf test (e.g. bz2) in this setup, do : |
| 235 | perl perf/vg_perf --outer-valgrind=../outer/.../bin/valgrind perf/bz2 |
| 236 | |
| 237 | To run all the performance tests with an outer callgrind, do : |
| 238 | perl perf/vg_perf --outer-valgrind=../outer/.../bin/valgrind \ |
| 239 | --outer-tool=callgrind perf |
| 240 | |
| 241 | To compare the performance of multiple Valgrind versions, do : |
| 242 | perl perf/vg_perf --outer-valgrind=../outer/.../bin/valgrind \ |
| 243 | --vg=../inner_xxxx --vg=../inner_yyyy perf |
| 244 | (where inner_xxxx and inner_yyyy are the versions to compare). |
| 245 | Cachegrind and cg_diff are particularly handy to obtain a delta |
| 246 | between the two versions. |
| 247 | |
| 248 | When the outer tool is callgrind or cachegrind, the following |
| 249 | output files will be created for each test: |
| 250 | <outertoolname>.out.<inner_valgrind_dir>.<tt>.<perftestname>.<pid> |
| 251 | <outertoolname>.outer.log.<inner_valgrind_dir>.<tt>.<perftestname>.<pid> |
| 252 | (where tt is the two letters abbreviation for the inner tool(s) run). |
| 253 | |
| 254 | For example, the command |
| 255 | perl perf/vg_perf \ |
| 256 | --outer-valgrind=../outer_trunk/install/bin/valgrind \ |
| 257 | --outer-tool=callgrind \ |
| 258 | --vg=../inner_tchain --vg=../inner_trunk perf/many-loss-records |
| 259 | |
| 260 | produces the files |
| 261 | callgrind.out.inner_tchain.no.many-loss-records.18465 |
| 262 | callgrind.outer.log.inner_tchain.no.many-loss-records.18465 |
| 263 | callgrind.out.inner_tchain.me.many-loss-records.21899 |
| 264 | callgrind.outer.log.inner_tchain.me.many-loss-records.21899 |
| 265 | callgrind.out.inner_trunk.no.many-loss-records.21224 |
| 266 | callgrind.outer.log.inner_trunk.no.many-loss-records.21224 |
| 267 | callgrind.out.inner_trunk.me.many-loss-records.22916 |
| 268 | callgrind.outer.log.inner_trunk.me.many-loss-records.22916 |
| 269 | |
| 270 | |
njn | 0dc09e8 | 2005-11-03 16:24:53 +0000 | [diff] [blame] | 271 | Printing out problematic blocks |
| 272 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| 273 | If you want to print out a disassembly of a particular block that |
| 274 | causes a crash, do the following. |
| 275 | |
| 276 | Try running with "--vex-guest-chase-thresh=0 --trace-flags=10000000 |
| 277 | --trace-notbelow=999999". This should print one line for each block |
| 278 | translated, and that includes the address. |
| 279 | |
| 280 | Then re-run with 999999 changed to the highest bb number shown. |
| 281 | This will print the one line per block, and also will print a |
| 282 | disassembly of the block in which the fault occurred. |