summaryrefslogtreecommitdiffstats
path: root/tools
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'for-linus' of ↵Linus Torvalds2011-07-262-149/+963
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-ktest * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-ktest: ktest: Fix bug when ADD_CONFIG is set but MIN_CONFIG is not ktest: Keep fonud configs separate from default configs ktest: Add prompt to use OUTPUT_MIN_CONFIG ktest: Use Kconfig dependencies to shorten time to make min_config ktest: Add test type make_min_config ktest: Require one TEST_START in config file ktest: Add helper function to avoid duplicate code ktest: Add IGNORE_WARNINGS to ignore warnings in some patches ktest: Fix tar extracting of modules to target ktest: Have the testing tmp dir include machine name ktest: Add POST/PRE_BUILD options ktest: Allow initrd processing without modules defined ktest: Have LOG_FILE evaluate options as well ktest: Have wait on stdio honor bug timeout ktest: Implement our own force min config ktest: Add TEST_NAME option ktest: Add CONFIG_BISECT_GOOD option ktest: Add detection of triple faults ktest: Notify reason to break out of monitoring boot
| * ktest: Fix bug when ADD_CONFIG is set but MIN_CONFIG is notSteven Rostedt2011-07-151-12/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The MIN_CONFIG is a single config that is considered to have all the configs that are required to boot the box. ADD_CONFIG is a list of configs that we add that may contain configs known to be broken (set off) or just configs that we want every box to have and this can include shared configs. If a config has no MIN_CONFIG defined, but has multiple files defined for the ADD_CONFIG, the test will die, because the MIN_CONFIG will default to ADD_CONFIG. The problem is the code to open MIN_CONFIG expects a string of one file, not multiple, and the open will fail. Since the real minconfig that is used is a concatination of MIN_CONFIG and ADD_CONFIG files, we change the code to open that instead of whatever MIN_CONFIG defaults to. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Keep fonud configs separate from default configsSteven Rostedt2011-07-151-4/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The IGNORE_CONFIG file holds the configs that we don't want to change (with their proper settings). But on start up, the make noconfig is executed, and the configs that are on are also put into the ignore config category. But these are configs that were forced on by the kconfig scripts and not something that we found must be enabled to boot our machine. By keeping the configs that are forced on by default, separate from the configs we found that are required to boot the box, we can get a much more interesting IGNORE_CONFIG. In fact, the IGNORE_CONFIG can usually end up being the must have configs to boot, and only have 6 or 7 configs set. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add prompt to use OUTPUT_MIN_CONFIGSteven Rostedt2011-07-152-14/+59
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If the defined OUTPUT_MIN_CONFIG in the make_min_config test exists, then give a prompt to ask the user if they want to use that config instead, as it is very often the case, especially when the test has been interrupted. The OUTPUT_MIN_CONFIG is usually the config that one wants to use to continue the test where they left off. But if START_MIN_CONFIG is defined (thus the MIN_CONFIG is not the default), then do not prompt, as it will be annoying if the user has this as one of many tests, and the test pauses waiting for input, while the user is sleeping. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Use Kconfig dependencies to shorten time to make min_configSteven Rostedt2011-07-152-22/+273
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | To save time, the test does not just grab any option and test it. The Kconfig files are examined to determine the dependencies of the configs. If a config is chosen that depends on another config, that config will be checked first. By checking the parents first, we can eliminate whole groups of configs that may have been enabled. For example, if a USB device config is chosen and depends on CONFIG_USB, the CONFIG_USB will be tested before the device. If CONFIG_USB is found not to be needed, it, as well as all configs that depend on it, will be disabled and removed from the current min_config. Note, the code from streamline_config (make localmodconfig) was copied and used to find the dependencies in the Kconfig file. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add test type make_min_configSteven Rostedt2011-07-152-4/+301
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | After doing a make localyesconfig, your kernel configuration may not be the most useful minimum configuration. Having a true minimum config that you can use against other configs is very useful if someone else has a config that breaks on your code. By only forcing those configurations that are truly required to boot your machine will give you less of a chance that one of your set configurations will make the bug go away. This will give you a better chance to be able to reproduce the reported bug matching the broken config. Note, this does take some time, and may require you to run the test over night, or perhaps over the weekend. But it also allows you to interrupt it, and gives you the current minimum config that was found till that time. Note, this test automatically assumes a BUILD_TYPE of oldconfig and its test type acts like boot. TODO: add a test version that makes the config do more than just boot, like having network access. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Require one TEST_START in config fileSteven Rostedt2011-06-141-0/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | There has been too many times that I put in one too many SKIP TEST_STARTs and start the test with the default randconfig by accident that I added this to have ktest ask the user for which test they want to run if no TEST_START is specified. Now if I accidently start the test with all TEST_STARTs skipped, ktest asks what test do I want to run, and I now have a chance to kill it before it does a make mrproper on my build directory. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add helper function to avoid duplicate codeSteven Rostedt2011-06-141-26/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Several places had the following code: get_grub_index; get_version; install; start_monitor; return monitor; Creating a function "start_monitor_and_boot()" replaces these mulitple uses with a single call. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add IGNORE_WARNINGS to ignore warnings in some patchesSteven Rostedt2011-06-142-2/+20
| | | | | | | | | | | | | | | | | | Doing a patchcheck test, there may be warnings that gcc produces which may be OK, and the test should not fail on that commit. By adding a IGNORE_WARNINGS option to list a space delimited SHA1s that are ignored lets the user avoid having the test fail on certain commits. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Fix tar extracting of modules to targetSteven Rostedt2011-06-141-1/+1
| | | | | | | | | | | | | | | | The tar command to create the module directory is cjf, but the extraction only had xf. This works on most versions of tar, but some versions of tar require xjf for extraction as well. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Have the testing tmp dir include machine nameSteven Rostedt2011-06-142-3/+3
| | | | | | | | | | | | | | | | As multiple tests may be executed by the same server, have the test machine name add uniqueness to the value of the temp directory. Otherwise the temp directories may overwrite each other's tests. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add POST/PRE_BUILD optionsSteven Rostedt2011-06-142-3/+60
| | | | | | | | | | | | | | | | There are some cases that a patch may be needed to apply to the kernel in patchcheck or bisect tests. Adding a PRE_BUILD option to apply the patch and POST_BUILD to remove it, allows for this to be done easily. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Allow initrd processing without modules definedSteven Rostedt2011-06-131-6/+12
| | | | | | | | | | | | | | | | | | | | | | When a config is set with CONFIG_MODULES=n, it does not mean that the kernel does not need an initrd to boot. For systems that depend on LVM and such, an initrd must run first. If POST_INSTALL is defined, then run the post install regardless if modules are needed or not. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Have LOG_FILE evaluate options as wellSteven Rostedt2011-06-131-58/+68
| | | | | | | | | | | | The LOG_FILE variable needs to evaluate the $ options as well. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Have wait on stdio honor bug timeoutSteven Rostedt2011-06-131-1/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | After a bug is found, the STOP_AFTER_FAILURE timeout is used to determine how much output should be printed before breaking out of the monitor loop. This is to get things like call traces and enough infromation about the bug to help determine what caused it. The STOP_AFTER_FAILURE is usually much shorter than the TIMEOUT that is used to determine when to quit after no more stdio is given. But since the stdio read uses a wait on I/O, the STOP_AFTER_FAILURE is only checked after we get something from I/O. But if the I/O does not return any more data, we wait the TIMEOUT period instead, even though we already triggered a bug report. The wait on I/O should honor the STOP_AFTER_FAILURE time if a bug has been found. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Implement our own force min configSteven Rostedt2011-06-131-13/+61
| | | | | | | | | | | | | | | | | | | | | | | | | | Using the build KCONFIG_ALLCONFIG environment variable to force the min config may not always work properly. Since ktest is written in perl, it is trivial to read and replace the current config with the configs specified by the min config. Now the min config (and add configs) are read by perl and before a make is done, these configs in the .config file are replaced by the version in the min config. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add TEST_NAME optionSteven Rostedt2011-06-132-2/+22
| | | | | | | | | | | | | | | | Searching through several tests, it gets confusing which test result is for which test. By adding the TEST_NAME option, the user can tell which test result belongs to which test. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add CONFIG_BISECT_GOOD optionSteven Rostedt2011-06-132-7/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently the config_bisect compares the min config with the CONFIG_BISECT config. There may be another config that we know is good that we want to ignore configs on. By passing in this config it will ignore the options that are set in the good config. Note: This only ignores the config, it does not (yet) handle options that are different between the two configs. If the good config has "SLAB" set and the bad config has "SLUB" it will not find the bug if the bug had to do with changing these two options. This is something that I intend to implement in the future. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Add detection of triple faultsSteven Rostedt2011-06-132-0/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When a triple fault happens in a test, no call trace nor panic is displayed. Instead, the system reboots to the good kernel. Since the good kernel may display a boot prompt that matches the success string, ktest may think that the test succeeded, when it did not. Detecting triple faults is tricky because it is hard to generalize what a reboot looks like. The best that we can come up with for now is to examine the Linux banner. If we detect that the Linux banner matches the test we want to test, then look to see if we hit another Linux banner with a different kernel is booted. This can be assumed to be a triple fault. We can't just check for two Linux banners because things like early printk may cause the Linux banner to be displayed twice. Checking for different kernel versions should be the safe bet. If this for some reason detects a false triple boot. A new ktest config option is also created: DETECT_TRIPLE_FAULT This can be set to 0 to disable this checking. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * ktest: Notify reason to break out of monitoring bootSteven Rostedt2011-06-131-2/+10
| | | | | | | | | | | | | | | | | | Different timeouts can cause the ktest monitor to break out of the loop. It becomes annoying that one does not know the reason why it exited the monitor loop. Display the cause of the reason why the loop was exited. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| |
| \
| \
| \
| \
| \
*-----. \ Merge branches 'x86-urgent-for-linus', 'core-debug-for-linus', ↵Linus Torvalds2011-07-231-1/+4
|\ \ \ \ \ | | | |_|/ | | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 'irq-core-for-linus' and 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: um: Make rwsem.S depend on CONFIG_RWSEM_XCHGADD_ALGORITHM * 'core-debug-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: debug: Make CONFIG_EXPERT select CONFIG_DEBUG_KERNEL to unhide debug options * 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: genirq: Remove unused CHECK_IRQ_PER_CPU() * 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: perf tools, x86: Fix 32-bit compile on 64-bit system
| | | | * perf tools, x86: Fix 32-bit compile on 64-bit systemDavid Ahern2011-07-211-1/+4
| | | |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Builds for 32-bit perf binaries on a 64-bit host currently fail with this error: [...] bench/../../../arch/x86/lib/memcpy_64.S: Assembler messages: bench/../../../arch/x86/lib/memcpy_64.S:29: Error: bad register name `%rdi' bench/../../../arch/x86/lib/memcpy_64.S:34: Error: invalid instruction suffix for `movs' bench/../../../arch/x86/lib/memcpy_64.S:50: Error: bad register name `%rdi' bench/../../../arch/x86/lib/memcpy_64.S:61: Error: bad register name `%rdi' ... The problem is the detection of the host arch without considering passed in flags. This change fixes 32-bit builds via: make EXTRA_CFLAGS=-m32 and 64-bit builds still reference the memcpy_64.S. Signed-off-by: David Ahern <dsahern@gmail.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: <stable@kernel.org> Link: http://lkml.kernel.org/r/1310420304-21452-1-git-send-email-dsahern@gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf tools: Make test use the preset debugfs pathJiri Olsa2011-07-211-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Use preset debugfs path instead of hardcoded one. Signed-off-by: Jiri Olsa <jolsa@redhat.com> Cc: acme@redhat.com Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org Link: http://lkml.kernel.org/r/1310635534-4013-4-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf tools: Add automated tests for events parsingJiri Olsa2011-07-211-0/+245
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Adding builtin test for parse_events function, which is responsible for parsing/processing "-e" option for stat/top/record commands. This new test will run within the builtin test command suite (perf test). One or several tests were added for each type of event. More tests could be added easily if needed. Signed-off-by: Jiri Olsa <jolsa@redhat.com> Cc: acme@redhat.com Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org Link: http://lkml.kernel.org/r/1310635534-4013-3-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf tools: De-opt the parse_events functionJiri Olsa2011-07-215-14/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Moving out the option parameter from parse_events function, and adding new parse_events_option function instead. The option parameter is used only to carry "struct perf_evlist" pointer for chaining new events. Putting it away, enable us to call parse_events from other places without using the option parameter. Signed-off-by: Jiri Olsa <jolsa@redhat.com> Cc: acme@redhat.com Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org Link: http://lkml.kernel.org/r/1310635534-4013-2-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf script: Fix display of IP address for non-callchain pathDavid Ahern2011-07-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Non-callchain path is using al.addr which prints as: openssl 14564 17672.003587: 7862d _x86_64_AES_encrypt_compact This should be sample->ip to print as: openssl 14564 17672.003587: 3f7867862d _x86_64_AES_encrypt_compact Signed-off-by: David Ahern <dsahern@gmail.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: acme@ghostprotocols.net Cc: peterz@infradead.org Cc: paulus@samba.org Link: http://lkml.kernel.org/r/1306768587-15376-1-git-send-email-dsahern@gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf tools: Fix endian conversion reading event attr from file headerDavid Ahern2011-07-213-13/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The perf_event_attr struct has two __u32's at the top and they need to be swapped individually. With this change I was able to analyze a perf.data collected in a 32-bit PPC VM on an x86 system. I tested both 32-bit and 64-bit binaries for the Intel analysis side; both read the PPC perf.data file correctly. -v2: - changed the existing perf_event__attr_swap() to swap only elements of perf_event_attr and exported it for use in swapping the attributes in the file header - updated swap_ops used for processing events Signed-off-by: David Ahern <dsahern@gmail.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: acme@ghostprotocols.net Cc: peterz@infradead.org Cc: paulus@samba.org Cc: <stable@kernel.org> Link: http://lkml.kernel.org/r/1310754849-12474-1-git-send-email-dsahern@gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf tools: Add missing 'node' alias to the hw_cache[] arrayJiri Olsa2011-07-211-5/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add "node" as a simple alias for NODE cache events. The addition of NODE cache events broke the parse_alias function, so any mismatched event caused the segfault, like: # ./perf stat -e krava ls The hw_cache/hw_cache_op/hw_cache_result arrays needs to follow PERF_COUNT_HW_CACHE_*MAX enums. Adding those MAXs to be size of those arrays, so possible ommision in future wil not lead to segfault. Adding read/write/prefetch as allowed operations for node cache event. Signed-off-by: Jiri Olsa <jolsa@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: acme@redhat.com Link: http://lkml.kernel.org/r/20110713205818.GB7827@jolsa.brq.redhat.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf probe: Support adding probes on offline kernel modulesMasami Hiramatsu2011-07-153-12/+49
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Support adding probes on offline kernel modules. This enables perf-probe to trace kernel-module init functions via perf-probe. If user gives the path of module with -m option, perf-probe expects the module is offline. This feature works with --add, --funcs, and --vars. E.g) # perf probe -m /lib/modules/`uname -r`/kernel/fs/btrfs/btrfs.ko \ -a "extent_io_init:5 extent_state_cache" Add new events: probe:extent_io_init (on extent_io_init:5 with extent_state_cache) probe:extent_io_init_1 (on extent_io_init:5 with extent_state_cache) You can now use it on all perf tools, such as: perf record -e probe:extent_io_init_1 -aR sleep 1 Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072751.6528.10230.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf probe: Add probed module in front of functionMasami Hiramatsu2011-07-152-9/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add probed module name and ":" in front of function name if -m module option is given. In the result, the symbol name passed to kprobe-tracer becomes MODULE:FUNCTION, so that kallsyms can solve it as a symbol in the module correctly. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072745.6528.26416.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf probe: Introduce debuginfo to encapsulate dwarf informationMasami Hiramatsu2011-07-153-136/+182
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduce debuginfo to encapsulate dwarf information. This new object allows us to reuse and expand debuginfo easily. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072739.6528.12438.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf-probe: Move dwarf library routines to dwarf-aux.{c, h}Masami Hiramatsu2011-07-155-523/+768
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Move dwarf library related routines to dwarf-aux.{c,h}. This includes several minor changes. - Add simple documents for each API. - Rename die_find_real_subprogram() to die_find_realfunc() - Rename line_walk_handler_t to line_walk_callback_t. - Minor cleanups. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072727.6528.57647.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf probe: Remove redundant dwarf functionsMasami Hiramatsu2011-07-151-37/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since there are dwarf_bitsize, dwarf_bitoffset and dwarf_bytesize defined in libdw, we don't need die_get_bit_size, die_get_bit_offset and die_get_byte_size anymore. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072721.6528.2747.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf probe: Move strtailcmp to string.cMasami Hiramatsu2011-07-153-15/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since strtailcmp() is enough generic, it should be defined in string.c. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/20110627072715.6528.10677.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf probe: Rename DIE_FIND_CB_FOUND to DIE_FIND_CB_ENDMasami Hiramatsu2011-07-151-6/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since die_find/walk* callbacks use DIE_FIND_CB_FOUND for both of failed and found cases, it should be "END" instead "FOUND" for avoiding confusion. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reported-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Link: http://lkml.kernel.org/r/20110627072709.6528.45706.stgit@fedora15 Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf: Robustify proc and debugfs file recordingSonny Rao2011-07-141-91/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | While attempting to create a timechart of boot up I found perf didn't tolerate modules being loaded/unloaded. This patch fixes this by reading the file once and then writing the size read at the correct point in the file. It also simplifies the code somewhat. Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Signed-off-by: Sonny Rao <sonnyrao@chromium.org> Signed-off-by: Michael Neuling <mikey@neuling.org> Link: http://lkml.kernel.org/r/10011.1310614483@neuling.org Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | | perf report/annotate/script: Add option to specify a CPU rangeAnton Blanchard2011-07-058-0/+102
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add an option to perf report/annotate/script to specify which CPUs to operate on. This enables us to take a single system wide profile and analyse each CPU (or group of CPUs) in isolation. This was useful when profiling a multiprocess workload where the bottleneck was on one CPU but this was hidden in the overall profile. Per process and per thread breakdowns didn't help because multiple processes were running on each CPU and no single process consumed an entire CPU. The patch converts the list of CPUs returned by cpu_map__new into a bitmap for fast lookup. I wanted to use -C to be consistent with perf top/record/stat, but unfortunately perf report already uses -C <comms>. v2: Incorporate suggestions from David Ahern: - Added -c to perf script - Check that SAMPLE_CPU is set when -c is used - Update documentation v3: Create perf_session__cpu_bitmap() Signed-off-by: Anton Blanchard <anton@samba.org> Acked-by: David Ahern <dsahern@gmail.com> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Link: http://lkml.kernel.org/r/20110704215750.11647eb9@kryten Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | perf stat: Add noise output for csv modeZhengyu He2011-07-011-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, when you want perf-stat to output the statistics in csv mode, no information of the noise will be printed out. For example right now we output this --repeat information: ./perf stat -r3 -x, sleep 1 1.164789,task-clock 8,context-switches 0,CPU-migrations 219,page-faults 3337800,cycles With this patch, the output will be appended with an additional entry for the noise value: ./perf stat -r3 -x, sleep 1 1.164789,task-clock,3.75% 8,context-switches,75.00% 0,CPU-migrations,100.00% 219,page-faults,0.00% 3337800,cycles,3.36% Signed-off-by: Zhengyu He <zhengyuh@google.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Cc: Stephane Eranian <eranian@google.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Link: http://lkml.kernel.org/r/1308861942-4945-1-git-send-email-zhengyuh@google.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | | | Merge branch 'perf/core' of ↵Ingo Molnar2011-07-017-144/+169
|\ \ \ \ | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing into perf/core
| * | | | perf tools: Only display parent field if explictly sortedFrederic Weisbecker2011-06-301-1/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We don't need to display the parent field if the parent sorting machinery is only used for parent filtering (as in "-p foo"). However if parent filtering is used in combination with explicit parent sorting ( -s parent), we want to display it. Result with: perf report -p kernel_thread -s parent Before: # Overhead Parent symbol # ........ ............. # 0.07% | --- ioread8 ata_sff_check_status ata_sff_tf_load ata_sff_qc_issue ata_bmdma_qc_issue ata_qc_issue ata_scsi_translate ata_scsi_queuecmd scsi_dispatch_cmd scsi_request_fn __blk_run_queue __make_request generic_make_request submit_bio submit_bh journal_submit_commit_record jbd2_journal_commit_transaction kjournald2 kthread kernel_thread_helpe After: # Overhead Parent symbol # ........ ............. # 0.07% kernel_thread_helper | --- ioread8 ata_sff_check_status ata_sff_tf_load ata_sff_qc_issue ata_bmdma_qc_issue ata_qc_issue ata_scsi_translate ata_scsi_queuecmd scsi_dispatch_cmd scsi_request_fn __blk_run_queue __make_request generic_make_request submit_bio submit_bh journal_submit_commit_record jbd2_journal_commit_transaction kjournald2 kthread kernel_thread_helper Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Cc: Sam Liao <phyomh@gmail.com>
| * | | | perf tools: Allow sort dimensions to be registered more than onceFrederic Weisbecker2011-06-301-6/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | So that the parent sort dimension can be registered twice: once if we add it as an explicit sort dimension (-s parent) and twice if we request a parent filter (-p foo). We'll have only one parent sort dimension in the end but this allows to override the default parent filter with we gave in "-p" option. The goal of this is to prepare to allow the use of "-s parent" and "-p foo" at the same time, ie: sort by filtered parent. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Cc: Sam Liao <phyomh@gmail.com>
| * | | | perf tools: Don't display ignored entries on stdio uiFrederic Weisbecker2011-06-301-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As for newt ui, don't display entries that have been marked as ignored. The practical current effect of this is to make parent filtering really working. Before, entries that were ignored were given a null parent but were still displayed. This resulted in some weird effects: # Overhead Command Shared Object Symbol # ........ ........... ................. ............ # ^A | --- __lock_acquire | |--95.97%-- lock_acquire | | | |--30.75%-- _raw_spin_lock Discard these from the stdio display. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Cc: Sam Liao <phyomh@gmail.com>
| * | | | perf tools: Remove sort print helpers declarationsFrederic Weisbecker2011-06-301-6/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These are probably some old leftovers. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Cc: Sam Liao <phyomh@gmail.com>
| * | | | perf tools: Make sort operations staticFrederic Weisbecker2011-06-302-120/+99
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These don't need to be globally visible. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Cc: Sam Liao <phyomh@gmail.com>
| * | | | perf tools: Add inverted call graph report support.Sam Liao2011-06-305-11/+53
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add "caller/callee" option to support inverted butterfly report, in the inverted report (with caller option), the call graph start from the callee's ancestor. Users can use such view to catch system's performance bottleneck from a sysprof like view. Using this option with specified sort order like pid gives us high level view of call graph statistics. Also add "-G" alias for inverted call graph. Signed-off-by: Sam Liao <phyomh@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Stephane Eranian <eranian@google.com> Cc: David Ahern <dsahern@gmail.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
* | | | | Merge commit 'v3.0-rc5' into perf/coreIngo Molnar2011-07-013-2/+3
|\ \ \ \ \ | |/ / / / |/| | | / | | |_|/ | |/| | | | | | Merge reason: Pick up the latest fixes. Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | | |
| | \ \
| | \ \
| | \ \
| *---. \ \ Merge branches 'perf-urgent-for-linus', 'sched-urgent-for-linus', ↵Linus Torvalds2011-06-191-1/+1
| |\ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 'timers-urgent-for-linus' and 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: tools/perf: Fix static build of perf tool tracing: Fix regression in printk_formats file * 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: generic-ipi: Fix kexec boot crash by initializing call_single_queue before enabling interrupts * 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: clocksource: Make watchdog robust vs. interruption timerfd: Fix wakeup of processes when timer is cancelled on clock change * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: x86, MAINTAINERS: Add x86 MCE people x86, efi: Do not reserve boot services regions within reserved areas
| | * | | | | tools/perf: Fix static build of perf toolMathias Krause2011-06-161-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | To build a statically linked version of the perf tool all needed libraries must be added in the correct order to get the symbols resolved. Currently this is broken when, e.g. python or newt support is enabled -- libpython needs libpthread which is an unconditional link dependency of the perf tool; libslang needs libm, another unconditional dependency. To solve the problem in the long run without the need to keep track of transitive library dependencies, simply make the linker look at the EXTLIBS multiple times until it has all symbols resolved. Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Link: http://lkml.kernel.org/r/1308171818-20370-1-git-send-email-minipli@googlemail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * | | | | | Merge branch 'core-urgent-for-linus' of ↵Linus Torvalds2011-06-191-0/+1
| |\ \ \ \ \ \ | | |_|/ / / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: rcu: Move RCU_BOOST #ifdefs to header file rcu: use softirq instead of kthreads except when RCU_BOOST=y rcu: Use softirq to address performance regression rcu: Simplify curing of load woes
| | * | | | | rcu: Use softirq to address performance regressionShaohua Li2011-06-141-0/+1
| | | |_|_|/ | | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit a26ac2455ffcf3(rcu: move TREE_RCU from softirq to kthread) introduced performance regression. In an AIM7 test, this commit degraded performance by about 40%. The commit runs rcu callbacks in a kthread instead of softirq. We observed high rate of context switch which is caused by this. Out test system has 64 CPUs and HZ is 1000, so we saw more than 64k context switch per second which is caused by RCU's per-CPU kthread. A trace showed that most of the time the RCU per-CPU kthread doesn't actually handle any callbacks, but instead just does a very small amount of work handling grace periods. This means that RCU's per-CPU kthreads are making the scheduler do quite a bit of work in order to allow a very small amount of RCU-related processing to be done. Alex Shi's analysis determined that this slowdown is due to lock contention within the scheduler. Unfortunately, as Peter Zijlstra points out, the scheduler's real-time semantics require global action, which means that this contention is inherent in real-time scheduling. (Yes, perhaps someone will come up with a workaround -- otherwise, -rt is not going to do well on large SMP systems -- but this patch will work around this issue in the meantime. And "the meantime" might well be forever.) This patch therefore re-introduces softirq processing to RCU, but only for core RCU work. RCU callbacks are still executed in kthread context, so that only a small amount of RCU work runs in softirq context in the common case. This should minimize ksoftirqd execution, allowing us to skip boosting of ksoftirqd for CONFIG_RCU_BOOST=y kernels. Signed-off-by: Shaohua Li <shaohua.li@intel.com> Tested-by: "Alex,Shi" <alex.shi@intel.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
OpenPOWER on IntegriCloud