- 16 Jul, 2014 11 commits
-
-
Joseph Schuchart authored
To python scripts, including pid, tid, and cpu for which the event was recorded. At the moment, the pointer to the sample struct is passed to scripts, which seems to be of little use. The patch puts this information in dictionaries for easy access by Python scripts. Signed-off-by:
Joseph Schuchart <joseph.schuchart@tu-dresden.de> Acked-by:
Thomas Ilsche <thomas.ilsche@tu-dresden.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Thomas Ilsche <thomas.ilsche@tu-dresden.de> Link: http://lkml.kernel.org/r/53BE7E20.8080500@tu-dresden.deSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Joseph Schuchart authored
This provides valuable information for tracing performance problems. Since this change alters the interface for the python scripts, also adjust the script generation and the provided scripts. Signed-off-by:
-
Joseph Schuchart authored
Signed-off-by:
-
Alexander Yarygin authored
On s390, the vmexit event has a tree-like structure: between exit_event_begin and exit_event_end several other events may happen and with each of them refining the previous ones. This patch adds a decoder for such events to the generic code and also the files <asm/kvm_perf.h> and kvm-stat.c for s390. Commands 'perf kvm stat record', 'report' and 'live' are supported. Reviewed-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
Alexander Yarygin <yarygin@linux.vnet.ibm.com> Acked-by:
Christian Borntraeger <borntraeger@de.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1404397747-20939-5-git-send-email-yarygin@linux.vnet.ibm.comSigned-off-by:
-
Alexander Yarygin authored
Current code skips output of the x86 specific HLT event in order to avoid flooding the output with enabled --duration option. The events to be skipped should be architecture dependent, though. Let's add an architecture specific array of events to be skipped and introduce a skip_event() function checking against that array. Reviewed-by:
Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by:
-
Alexander Yarygin authored
Parts of a 'perf kvm stat' code make sense only for x86. Let's move this code into the arch/x86/kvm-stat.c file and add util/kvm-stat.h for generic structure definitions. Add a global array 'kvm_reg_events_ops' for accessing the arch-specific 'kvm_events_ops' from generic code. Since the several global arrays (i.e. 'kvm_events_tp') have been moved to arch/*, we can not know their sizes and use them directly in builtin-kvm.c. This patch fixes that problem by adding trimming NULL element to each array and changing the behavior of their handlers in generic code. Reviewed-by:
-
Alexander Yarygin authored
Currently perf-kvm uses string literals for kvm event names, but it works only for x86, because other architectures may have other names for those events. To reduce dependence on architecture, we add <asm/kvm_perf.h> file with defines for: - kvm_entry and kvm_exit events, - exit reason field name in kvm_exit event, - length of exit reasons strings, - vcpu_id field name in kvm trace events, and replace literals in perf-kvm. Reviewed-by:
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf into perf/core Pull perf/core improvements and fixes from Jiri Olsa: * Add IO mode into timechart command (Stanislav Fomichev) Signed-off-by:
Jiri Olsa <jolsa@kernel.org> Signed-off-by:
Ingo Molnar <mingo@kernel.org>
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo: User visible changes: o Prep patches to support 'perf kvm stat' on s390 (Alexander Yarygin) o Add pagefault statistics in 'trace' (Stanislav Fomichev) o Add header for columns in 'top' and 'report' TUI browsers (Jiri Olsa) o Add pagefault statistics in 'trace' (Stanislav Fomichev) Build fixes: o Fix build on 32-bit systems (Arnaldo Carvalho de Melo) Cleanups: o Convert open coded equivalents to asprintf() (Andy Shevchenko) Plumbing changes: o Allow reserving a row for header purposes in the hists browser (Arnaldo Carvalho de Melo) Signed-off-by:
-
Jiri Olsa authored
The following patch added another way to get mmap name: 78d683e8 ("mm, fs: Add vm_ops->name as an alternative to arch_vma_name") The vdso vma mapping already switch to this and we no longer get vdso name via arch_vma_name function. Adding this way to the perf mmap event name retrieval code. Caught this via perf test: $ sudo ./perf test -v 7 7: Validate PERF_RECORD_* events & perf_sample fields : --- start --- SNIP PERF_RECORD_MMAP for [vdso] missing! test child finished with 255 ---- end ---- Validate PERF_RECORD_* events & perf_sample fields: FAILED! Signed-off-by:
Andy Lutomirski <luto@amacapital.net> Signed-off-by:
Peter Zijlstra <peterz@infradead.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Link: http://lkml.kernel.org/r/1405353439-14211-1-git-send-email-jolsa@kernel.orgSigned-off-by:
-
Zhouyi Zhou authored
According to Peter's advice, put the failure handling to a goto chain. Compiled in x86_64, could you check if there is anything that I missed. Signed-off-by:
Zhouyi Zhou <yizhouzhou@ict.ac.cn> Signed-off-by:
-
- 09 Jul, 2014 4 commits
-
-
Stanislav Fomichev authored
--io-skip-eagain - don't show EAGAIN errors --io-min-time - make small io bursts visible --io-merge-dist - merge adjacent events Signed-off-by:
Stanislav Fomichev <stfomichev@yandex-team.ru> Acked-by:
Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/n/1404835423-23098-5-git-send-email-stfomichev@yandex-team.ruSigned-off-by:
-
Stanislav Fomichev authored
We don't need to overwrite current task start_time on fork, so update it only if it's zero. Signed-off-by:
-
Stanislav Fomichev authored
Currently, timechart records only scheduler and CPU events (task switches, running times, CPU power states, etc); this commit adds IO mode which makes it possible to record IO (disk, network) activity. In this mode perf timechart will generate SVG with IO charts (writes, reads, tx, rx, polls). Signed-off-by:
-
Stanislav Fomichev authored
Firefox doesn't correctly handle cases where we specify number in quotes and have some padding around the number, like the following: <rect ... height=" 3.1" ...> In this case, it doesn't draw the figure. This patch removes 'field width' component from fprintf strings to fix it. Signed-off-by:
-
- 08 Jul, 2014 3 commits
-
-
Arnaldo Carvalho de Melo authored
CC /tmp/build/perf/builtin-trace.o builtin-trace.c: In function 'print_location': builtin-trace.c:1792:4: error: format '%lx' expects argument of type 'long unsigned int', but argument 4 has type 'u64' [-Werror=format] builtin-trace.c:1794:3: error: format '%lx' expects argument of type 'long unsigned int', but argument 3 has type 'u64' [-Werror=format] builtin-trace.c:1796:3: error: format '%lx' expects argument of type 'long unsigned int', but argument 3 has type 'u64' [-Werror=format] cc1: all warnings being treated as errors make[1]: *** [/tmp/build/perf/builtin-trace.o] Error 1 make[1]: *** Waiting for unfinished jobs.... make: *** [install-bin] Error 2 make: Leaving directory `/home/acme/git/linux/tools/perf' acme@linux-goap:~/git/linux> uname -a Linux linux-goap 3.7.10-1.16-desktop #1 SMP PREEMPT Fri May 31 20:21:23 UTC 2013 (97c14ba) i686 i686 i386 GNU/Linux Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stanislav Fomichev <stfomichev@yandex-team.ru> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-843p3aqbw531eqiu2hah8o9p@git.kernel.orgSigned-off-by: -
Stanislav Fomichev authored
'perf trace' can show summary of events using -S option. This commit also reports number of major/minor pagefault events in this summary. $ perf trace -s --pf all -- sleep 1 Summary of events: sleep (18604), 275 events, 99.6%, 197 minfaults, 0.000 msec syscall calls min avg max stddev (msec) (msec) (msec) (%) --------------- -------- --------- --------- --------- ------ read 2 0.000 0.001 0.002 100.00% open 3 0.004 0.005 0.007 21.13% close 3 0.001 0.001 0.001 1.37% fstat 3 0.001 0.002 0.002 10.66% mmap 8 0.002 0.004 0.006 10.69% mprotect 4 0.003 0.005 0.008 24.68% munmap 1 0.005 0.005 0.005 0.00% brk 3 0.001 0.002 0.003 28.08% access 3 0.002 0.003 0.005 24.48% nanosleep 1 1000.747 1000.747 1000.747 0.00% execve 8 0.000 0.033 0.246 91.00% arch_prctl 1 0.001 0.001 0.001 0.00% Signed-off-by: -
Arnaldo Carvalho de Melo authored
# id uid=0(root) gid=0(root) groups=0(root) # ls -la perf.data -rw-------. 1 acme acme 20720 Jul 8 11:35 perf.data Previously: # perf report file perf.data not owned by current user or root Now: # perf report File perf.data not owned by current user or root (use -f to override) Suggested-by:
-
- 07 Jul, 2014 13 commits
-
-
Andy Shevchenko authored
The following snippet V = malloc(S); if (!V) { } sprintf(V, ...) Can be easily changed to a one line: if (asprintf(&V, ...) < 0) { } Signed-off-by:Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Link: http://lkml.kernel.org/r/1404474229-15272-1-git-send-email-andriy.shevchenko@linux.intel.comSigned-off-by:
-
Alexander Yarygin authored
This patch defines CPUINFO_PROC for s390 and implements get_cpuid(). Reviewed-by:
-
Alexander Yarygin authored
cpu_isa_config() does two different things: searching for cpuid and initializing perf_kvm_stat struct with proper parameters. Let's move initialization to a separate function cpu_isa_init(), which is used to initialize all possible ISAs and can be used to init arch-depended things. Reviewed-by:
-
Alexander Yarygin authored
The perf_kvm_stat struct keeps the size of a table of exit reasons in the field 'exit_reasons_size'. The field is initialized and then used by get_exit_reason() for serial access to the table, so that the calling function does not actually need to know table size. Usage of tables with 'end of sequence' marker simplifies the get_exit_reason() function. Also the patch introduces a define_exit_reasons_table, which makes it easier to define new tables. Reviewed-by:
-
Alexander Yarygin authored
kvm stat support is currently conditional on i386/x86_64. Let's abstract this into a HAVE_KVM_STAT_SUPPORT flag, so that other architectures can support kvm stat as well. Reviewed-by:
-
Arnaldo Carvalho de Melo authored
Looks better and avoids it moving to the end of the screen as the column width changes over time in 'perf top'. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-yc144ai5jye3yl3h5yxw0scd@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Adding ui.show-headers config file option to define if the histogram entries headers will start visible or not. Currently columns headers are displayed by default, following lines in ~/.perfconfig file will disable that: [ui] show-headers = false Signed-off-by: -
Jiri Olsa authored
Displaying columns header text whenever 'H' is pressed, and hiding it on on another press. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
Open up space to show a one-line header text whenever 'H' is pressed, hide it on another key press. Follow up patch will format this line from the set of headers used. Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/n/tip-m894d6qk30h3qofw4k8neq4q@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
This requires some more work so that we can really just use the width of current entries when we want to partition the screen. Right now its just a prep patch so that we can have where to update ui_browser->rows when introducing the column headers line, that will be togglable, so we need to update it everytime we refresh the dimensions of the browser. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-ovk654rx525b4657y0mh6ku9@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
That will allow us to add a row offset to open up space for the column headers. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-otc3ployokfci5qi81o7jo22@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Some browsers, like the hist_browser, may want to be notified everytime a refresh_dimensions is needed, so that it can reset ui_browser->rows, for instance, or do some other related reaction to screen resizings. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-ielvluuemzn30bneh0zk3twi@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
The ui_browser->height is about the whole browser "window", including any header, status lines or any other space needed for some "Yes", "No", etc buttons a descendent browser, like hist_browser, may have. Since the navigation is done mostly on the ui_browser methods, it needs to know how many rows are on the screen, while details about what other components are, say, if a header (that may be composed of multiple lines, etc) is present. Besides this we'll need to add a ui_browser->refresh_dimensions() hook so that browsers like hist_browser can update ->rows in response to screen resizes, this will come in a follow up patch. This patch just adds ->rows and updates it when updating ->height, keeps using ->height for the only other widget that can come with ui_browser, the scrollbar, that goes on using all the height on the rightmost column in the screen, using ->rows for the keyboard navigation needs. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-xexmwg1mv7u03j5imn66jdak@git.kernel.orgSigned-off-by:
-
- 05 Jul, 2014 5 commits
-
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf into perf/core Pull perf/core improvements and fixes from Jiri Olsa: * Handle the num array type in python properly (Sebastian Andrzej Siewior) * Fix wrong condition for allocation failure (Jiri Olsa) * Adjust callchain based on DWARF debug info on powerpc (Sukadev Bhattiprolu) * Fix a risk for doing free on uninitialized pointer in traceevent lib (Rickard Strandqvist) Signed-off-by:
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo: New features: * Add support for pagefault tracing in 'trace', please see multiple examples in the changeset messages (Stanislav Fomichev). User visible changes: * Fallback to syscalls:* when raw_syscalls:* is not available in the perl and python perf scripts. (Daniel Bristot de Oliveira) Signed-off-by:
-
Rasmus Villemoes authored
Flipping the LSB doesn't require four lines of code. This shaves a few bytes of the generated code, including a branch. Signed-off-by:
Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Signed-off-by:
-
Jiri Olsa authored
Leftover from '8dc85d54 perf: Multiple task contexts'. Signed-off-by:
-
Vince Weaver authored
Transition to using the new generic PERF_PMU_CAP_NO_INTERRUPT method for failing a sampling event when no PMU interrupt is available. Signed-off-by:
Vince Weaver <vincent.weaver@maine.edu> Signed-off-by:
-
- 27 Jun, 2014 4 commits
-
-
Sebastian Andrzej Siewior authored
The raw_syscalls:sys_enter tracer for instance passes has one argument named 'arg' which is an array of 6 integers. Right the python scripts gets only 0 passed as an argument. The reason is that pevent_read_number() can not handle data types of 48 and returns always 0. This patch changes this by passing num array as list of nums which fit the description. As a result python will now see a list named arg which contains 6 (integer) items. Signed-off-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Acked-by:
-
Sebastian Andrzej Siewior authored
I was going to change something here and the result was so much on the right side of the screen that I decided to move that piece into its own function. This patch should make no function change except the moving the code into its own function. Signed-off-by:
-
Jiri Olsa authored
Check real allocated pointer for NULL. Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/n/tip-5rfzbalwjphmdzzil74eazyl@git.kernel.orgSigned-off-by:
-
Sukadev Bhattiprolu authored
When saving the callchain on Power, the kernel conservatively saves excess entries in the callchain. A few of these entries are needed in some cases but not others. We should use the DWARF debug information to determine when the entries are needed. Eg: the value in the link register (LR) is needed only when it holds the return address of a function. At other times it must be ignored. If the unnecessary entries are not ignored, we end up with duplicate arcs in the call-graphs. Use the DWARF debug information to determine if any callchain entries should be ignored when building call-graphs. Callgraph before the patch: 14.67% 2234 sprintft libc-2.18.so [.] __random | --- __random | |--61.12%-- __random | | | |--97.15%-- rand | | do_my_sprintf | | main | | generic_start_main.isra.0 | | __libc_start_main | | 0x0 | | | --2.85%-- do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --38.88%-- rand | |--94.01%-- rand | do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --5.99%-- do_my_sprintf main generic_start_main.isra.0 __libc_start_main 0x0 Callgraph after the patch: 14.67% 2234 sprintft libc-2.18.so [.] __random | --- __random | |--95.93%-- rand | do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --4.07%-- do_my_sprintf main generic_start_main.isra.0 __libc_start_main 0x0 TODO: For split-debug info objects like glibc, we can only determine the call-frame-address only when both .eh_frame and .debug_info sections are available. We should be able to determin the CFA even without the .eh_frame section. Fix suggested by Anton Blanchard. Thanks to valuable input on DWARF debug information from Ulrich Weigand. Reported-by:Maynard Johnson <maynard@us.ibm.com> Tested-by:
Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/20140625154903.GA29607@us.ibm.comSigned-off-by:
-
