- 07 Dec, 2015 14 commits
-
-
Russell King authored
Add basic support to parse ARM assembly. This: * enables perf to correctly show the disassembly, rather than chopping some constants off at the '#' (which is not a comment character on ARM). * allows perf to identify ARM instructions that branch to other parts within the same function, thereby properly annotating them. * allows perf to identify function calls, allowing called functions to be followed in the annotated view. Signed-off-by:
Russell King <rmk+kernel@arm.linux.org.uk> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Will Deacon <will.deacon@arm.com> Link: http://lkml.kernel.org/n/tip-owp1uj0nmcgfrlppfyeetuyf@git.kernel.orgSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Jiri Olsa authored
It's more readable this way and we can save one perf_evsel__is_group_leader condition in current code. Signed-off-by:
Jiri Olsa <jolsa@kernel.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1449133606-14429-7-git-send-email-jolsa@kernel.orgSigned-off-by:
-
Jiri Olsa authored
Currently we have 2 kinds of stat counters based on when the event is enabled: 1) tracee command events, which are enable once the tracee executes exec syscall (enable_on_exec bit) 2) all other events which get alive within the perf_event_open syscall And 2) case could raise a problem in case we want additional filter to be attached for event. In this case we want the event to be enabled after it's configured with filter. Changing the behaviour of 2) events, so they all are created as disabled (disabled bit). Adding extra enable call to make them alive once they finish setup. Signed-off-by: -
Jiri Olsa authored
No need to mimic the behaviour of perf_evlist__enable, we can use it directly. Signed-off-by:
-
Jiri Olsa authored
Use perf_evsel__(enable|disable) functions in perf_evlist__(enable|disable) functions in order to centralize ioctl enable/disable calls. This way we eliminate 2 places calling directly ioctl. Signed-off-by:
-
Jiri Olsa authored
Adding perf_evsel__disable function to have complement for perf_evsel__enable function. Both will be used in following patch to factor perf_evlist__(enable|disable). Signed-off-by:
-
Jiri Olsa authored
All events now share proper cpu and thread maps. There's no need to pass those maps from evlist, it's safe to use evsel maps for enabling event. Signed-off-by:
-
Jiri Olsa authored
It fixes segfault within machine__exit, that's caused but not creating kernel maps for machine.. We're calling machine__destroy_kernel_maps in machine__exit since commit: ebe9729c perf machine: Fix to destroy kernel maps when machine exits Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Tested-by:
-
Jiri Olsa authored
The bpf-output is added under software events, but is not parse-able within parse_events, which is what round trip test is expecting. Checking software events only until dummy event. Signed-off-by:
-
Jiri Olsa authored
In error path to try user space event, both cpus and threads map now owned by evlist and freed by perf_evlist__set_maps call. Getting reference to keep them alive. Signed-off-by:
-
Jiri Olsa authored
This is more straightforward than what we have now. It also fixes a segfault within machine__exit, that's caused by not creating kernel maps for machine.. We're calling machine__destroy_kernel_maps in machine__exit since commit: ebe9729c perf machine: Fix to destroy kernel maps when machine exits Signed-off-by:
-
Jiri Olsa authored
This is more straightforward than what we have now. It also fixes a segfault within machine__exit, that's caused by not creating kernel maps for machine.. We're calling machine__destroy_kernel_maps in machine__exit since commit: ebe9729c perf machine: Fix to destroy kernel maps when machine exits Signed-off-by:
-
Jiri Olsa authored
This is more straightforward than what we have now. It also fixes a segfault within machine__exit, that's caused by not creating kernel maps for machine.. We're calling machine__destroy_kernel_maps in machine__exit since commit: ebe9729c perf machine: Fix to destroy kernel maps when machine exits Signed-off-by:
-
Wang Nan authored
There's a mistake in dso__adjust_kmod_long_name() that it use strdup() to dup the new long_name of a dso, but passes the original string to dso__set_long_name(). Which causes random crash during cleanup. Signed-off-by:
Wang Nan <wangnan0@huawei.com> Reviewed-by:
Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Zefan Li <lizefan@huawei.com> Cc: pi3orama@163.com Fixes: c03d5184 ("perf machine: Adjust dso->long_name for offline module") Link: http://lkml.kernel.org/r/1449455785-42020-1-git-send-email-wangnan0@huawei.comSigned-off-by:
-
- 06 Dec, 2015 10 commits
-
-
Andi Kleen authored
Now that we have generic MSR trace points we can remove the old hackish perf MSR read tracing code. Signed-off-by:
Andi Kleen <ak@linux.intel.com> Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vince Weaver <vincent.weaver@maine.edu> Cc: rostedt@goodmis.org Link: http://lkml.kernel.org/r/1449018060-1742-4-git-send-email-andi@firstfloor.orgSigned-off-by:
Ingo Molnar <mingo@kernel.org>
-
Andi Kleen authored
For debugging low level code interacting with the CPU it is often useful to trace the MSR read/writes. This gives a concise summary of PMU and other operations. perf has an ad-hoc way to do this using trace_printk, but it's somewhat limited (and also now spews ugly boot messages when enabled) Instead define real trace points for all MSR accesses. This adds three new trace points: read_msr and write_msr and rdpmc. They also report if the access faulted (if *_safe is used) This allows filtering and triggering on specific MSR values, which allows various more advanced debugging techniques. All the values are well defined in the CPU documentation. The trace can be post processed with Documentation/trace/postprocess/decode_msr.py to add symbolic MSR names to the trace. I only added it to native MSR accesses in C, not paravirtualized or in entry*.S (which is not too interesting) Originally the patch kit moved the MSRs out of line. This uses an alternative approach recommended by Steven Rostedt of only moving the trace calls out of line, but open coding the access to the jump label. Signed-off-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vince Weaver <vincent.weaver@maine.edu> Link: http://lkml.kernel.org/r/1449018060-1742-3-git-send-email-andi@firstfloor.orgSigned-off-by:
-
Andi Kleen authored
Steven recommended open coding access to tracepoint->key to add trace points to headers. Unfortunately this is difficult for some headers (such as x86 asm/msr.h) because including tracepoint.h includes so many other headers that it causes include loops. The main problem is the include of linux/rcupdate.h, which pulls in a lot of other headers. The rcu header is only needed when actually defining trace points. Move the struct tracepoint into a separate tracepoint-defs.h header that can be included without pulling in all of RCU. Signed-off-by:
-
Andi Kleen authored
asm/atomic.h doesn't really need asm/processor.h anymore. Everything it uses has moved to other header files. So remove that include. processor.h is a nasty header that includes lots of other headers and makes it prone to include loops. Removing the include here makes asm/atomic.h a "leaf" header that can be safely included in most other headers. The only fallout is in the lib/atomic tester which relied on this implicit include. Give it an explicit include. (the include is in ifdef because the user is also in ifdef) Signed-off-by:
-
Andi Kleen authored
Fix a definition in the perf rapl driver. __initconst must be applied to a const object, but to declare a const pointer you need to use * const ..., not const ... * This fixes a section attribute conflict with LTO builds. Signed-off-by:
-
Peter Zijlstra authored
Various functions implement the same pattern to send IPIs to an event's CPU. Collapse the easy ones in a common helper function to reduce duplication. Signed-off-by:
-
Ingo Molnar authored
Signed-off-by: -
Jiri Olsa authored
In case we monitor events system wide, we get EXIT event (when configured) twice for each task that exited. Note doubled lines with same pid/tid in following example: $ sudo ./perf record -a ^C[ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.480 MB perf.data (2518 samples) ] $ sudo ./perf report -D | grep EXIT 0 60290687567581 0x59910 [0x38]: PERF_RECORD_EXIT(1250:1250):(1250:1250) 0 60290687568354 0x59948 [0x38]: PERF_RECORD_EXIT(1250:1250):(1250:1250) 0 60290687988744 0x59ad8 [0x38]: PERF_RECORD_EXIT(1250:1250):(1250:1250) 0 60290687989198 0x59b10 [0x38]: PERF_RECORD_EXIT(1250:1250):(1250:1250) 1 60290692567895 0x62af0 [0x38]: PERF_RECORD_EXIT(1253:1253):(1253:1253) 1 60290692568322 0x62b28 [0x38]: PERF_RECORD_EXIT(1253:1253):(1253:1253) 2 60290692739276 0x69a18 [0x38]: PERF_RECORD_EXIT(1252:1252):(1252:1252) 2 60290692739910 0x69a50 [0x38]: PERF_RECORD_EXIT(1252:1252):(1252:1252) The reason is that the cpu contexts are processes each time we call perf_event_task. I'm changing the perf_event_aux logic to serve task_ctx and cpu contexts separately, which ensure we don't get EXIT event generated twice on same cpu context. This does not affect other auxiliary events, as they don't use task_ctx at all. Signed-off-by:
-
Jiri Olsa authored
We need to add rest of the flags to the constraint mask instead of another INTEL_ARCH_EVENT_MASK, fixing a typo. Signed-off-by:
-
Yuanfang Chen authored
There was a mistake in the Haswell constraints table. Signed-off-by:
Yuanfang Chen <cheny@udel.edu> Signed-off-by:
-
- 04 Dec, 2015 1 commit
-
-
Peter Zijlstra authored
Dmitry reported a fairly silly recursive lock deadlock for PERF_EVENT_IOC_PERIOD, fix this by explicitly doing the inactive part of __perf_event_period() instead of calling that function. Reported-by:
Dmitry Vyukov <dvyukov@google.com> Signed-off-by:
-
- 28 Nov, 2015 6 commits
-
-
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: - Fix 'perf list' segfault due to lack of support for PERF_CONF_SW_BPF_OUTPUT in an array used just for printing available events, robustify the code involved (Arnaldo Carvalho de Melo) - 'perf test unwind' should create kernel maps, now that entry works and the test passes (Jiri Olsa) - Fix showing the running kernel build id in 'perf buildid-list' (Michael Petlan) - Fix command line symbol filtering in 'perf report' (Namhyung Kim) Infrastructure changes: - Extract and collect map info from BPF object files in libbpf (Wang Nan) Signed-off-by:
-
Wang Nan authored
Following patches are going to introduce BPF object level configuration to enable setting values into BPF maps. To avoid confusion, this patch renames existing 'config' in bpf-loader.c to 'program config'. Following patches would introduce 'object config'. Signed-off-by:
-
Wang Nan authored
This patch collects name of maps in BPF object files and saves them into 'maps' field in 'struct bpf_object'. 'bpf_object__get_map_by_name' is introduced to retrive fd and definitions of a map through its name. Signed-off-by:
He Kuang <hekuang@huawei.com> Cc: Alexei Starovoitov <ast@kernel.org> Cc: He Kuang <hekuang@huawei.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Zefan Li <lizefan@huawei.com> Cc: pi3orama@163.com Link: http://lkml.kernel.org/r/1448614067-197576-3-git-send-email-wangnan0@huawei.comSigned-off-by:
-
Wang Nan authored
This patch collects more information from maps sections in BPF object files into 'struct bpf_object', enables later patches access those information (such as the type and size of the map). In this patch, a new handler 'struct bpf_map' is extracted in parallel with bpf_object and bpf_program. Its iterator and accessor is also created. Signed-off-by:
-
Namhyung Kim authored
When perf report on TUI was called with -S symbol filter, it should update nr entries even if min_pcnt is 0. IIRC the reason was to update nr entries after applying minimum percent threshold. But if symbol filter was given on command line (with -S option), it should use hists->nr_non_filtered_entries instead of hists->nr_entries. So this patch fixes a bug of navigating hists browser that the cursor goes beyond the number of entries when -S (or similar) option is used. Signed-off-by:
-
Namhyung Kim authored
If user gives a filter, perf marks the corresponding column elided and omits the output. But it should process and aggregates samples using the field, otherwise samples will be aggregated as if the column was not there resulted in incorrect output. For example, I'd like to set a filter on native_write_msr_safe. The original overhead of the function is negligible. $ perf report | grep native_write_msr_safe 0.00% swapper [kernel.vmlinux] native_write_msr_safe 0.00% perf [kernel.vmlinux] native_write_msr_safe However adding -S option gives different output. $ perf report -S native_write_msr_safe --percentage absolute | \ > grep -e swapper -e perf 51.47% swapper [kernel.vmlinux] 4.14% perf [kernel.vmlinux] Since it aggregated samples using comm and dso only. In fact, the above values are same when it sorts with -s comm,dso. $ perf report -s comm,dso | grep -e swapper -e perf 51.47% swapper [kernel.vmlinux] 4.14% perf [kernel.vmlinux] This resulted in TUI failure with -ERANGE since it tries to increase sample hit count for annotation with wrong symbols due to incorrect aggregation. This patch fixes it not to skip elided fields when comparing samples in order to insert them to the hists. Commiter note: After the patch, with a different workloads: # perf report --show-total-period -S native_write_msr_safe --stdio # # symbol: native_write_msr_safe # # Samples: 455 of event 'cycles:pp' # Event count (approx.): 134787489 # # Overhead Period Command Shared Object # ........ ...... ............... ................ # 0.22% 293081 qemu-system-x86 [vmlinux] 0.19% 255914 swapper [vmlinux] 0.00% 2054 Timer [vmlinux] 0.00% 1021 firefox [vmlinux] 0.00% 2 perf [vmlinux] # perf report --show-total-period | grep native_write_msr_safe Failed to open /tmp/perf-14838.map, continuing without symbols 0.22% 293081 qemu-system-x86 [vmlinux] [k] native_write_msr_safe 0.19% 255914 swapper [vmlinux] [k] native_write_msr_safe 0.00% 2054 Timer [vmlinux] [k] native_write_msr_safe 0.00% 1021 firefox [vmlinux] [k] native_write_msr_safe 0.00% 2 perf [vmlinux] [k] native_write_msr_safe # Reported-by:
-
- 27 Nov, 2015 6 commits
-
-
Namhyung Kim authored
Currently when perf fails to process samples for some reason, it doesn't show any message about the failure. This is very inconvenient for users especially on TUI as screen is reset after the failure. Reported-by:
-
Arnaldo Carvalho de Melo authored
When a43eec30 ("bpf: introduce bpf_perf_event_output() helper") added PERF_COUNT_SW_BPF_OUTPUT we ended up with a new entry in the event_symbols_sw array that wasn't initialized, thus set to NULL, fix print_symbol_events() to check for that case so that we don't crash if this happens again. (gdb) bt #0 __match_glob (ignore_space=false, pat=<optimized out>, str=<optimized out>) at util/string.c:198 #1 strglobmatch (str=<optimized out>, pat=pat@entry=0x7fffffffe61d "stall") at util/string.c:252 #2 0x00000000004993a5 in print_symbol_events (type=1, syms=0x872880 <event_symbols_sw+160>, max=11, name_only=false, event_glob=0x7fffffffe61d "stall") at util/parse-events.c:1615 #3 print_events (event_glob=event_glob@entry=0x7fffffffe61d "stall", name_only=false) at util/parse-events.c:1675 #4 0x000000000042c79e in cmd_list (argc=1, argv=0x7fffffffe390, prefix=<optimized out>) at builtin-list.c:68 #5 0x00000000004788a5 in run_builtin (p=p@entry=0x871758 <commands+120>, argc=argc@entry=2, argv=argv@entry=0x7fffffffe390) at perf.c:370 #6 0x0000000000420ab0 in handle_internal_command (argv=0x7fffffffe390, argc=2) at perf.c:429 #7 run_argv (argv=0x7fffffffe110, argcp=0x7fffffffe11c) at perf.c:473 #8 main (argc=2, argv=0x7fffffffe390) at perf.c:588 (gdb) p event_symbols_sw[PERF_COUNT_SW_BPF_OUTPUT] $4 = {symbol = 0x0, alias = 0x0} (gdb) A patch to robustify perf to not segfault when the next counter gets added in the kernel will follow this one. Reported-by:
-
Arnaldo Carvalho de Melo authored
When PERF_COUNT_SW_BPF_OUTPUT was added to the kernel we should've added it to tools/perf, where it is used just to list events. This ended up causing a segfault in commands like "perf list stall". Fix it by adding that new software counter. A patch to robustify perf to not segfault when the next counter gets added in the kernel will follow this one. Reported-by:
-
Jiri Olsa authored
The 'perf test unwind' is failing because it forgot to create the kernel maps, fix it. After the patch: # perf test unwind 40: Test dwarf unwind : Ok Reported-and-Tested-by:
Jiri Olsa <jolsa@redhat.com> Cc: David Ahern <dsahern@gmail.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/r/20151127082121.GA24503@krava.brq.redhat.comSigned-off-by:
-
Michael Petlan authored
The --kernel option of perf buildid-list tool should show the running kernel buildid. The functionality has been lost during other changes of the related code. The build_id__sprintf() function should return length of the build-id string, but it was the length of the build-id raw data instead. Due to that, some return value checking caused that the final string was not printed out. With this patch the build_id__sprintf() returns the correct value, so the --kernel option works again. Before: # perf buildid-list --kernel # After: # perf buildid-list --kernel 972c1edab5bdc06cc224af45d510af662a3c6972 # Signed-off-by:
Michael Petlan <mpetlan@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> LPU-Reference: 1448632089.24573.114.camel@redhat.com 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: User visible changes: - Add 'vmlinux.debug' to the vmlinux search path (Ekaterina Tumanova) - Do not show sample_(type|period) in the perf_event_attr dump when using -v with 'perf stat' (Jiri Olsa) - Display the WEIGHT sample bit, when set, in 'perf evlist -v' (Jiri Olsa) - Honour --hide-unresolved in 'report', will honour it as well in 'top' when --hide-unresolved gets supported in that tool (Namhyung Kim) - Fix freeze wit h--call-graph 'flat/folded' due to not properly reinitializing the callchain rb_tree (Namhyumg Kim) - Set dso->long_name when a module name is passed as a parameter to tools like 'perf probe' but the 'struct dso' associated to that module still doesn't have the full path for the module, just the '[name]' one obtained from /proc/modules (Wang Nan) - Fix anon_hugepage mmaps detection using scanf on /proc/PID/smaps (Yannick Brosseau) Infrastructure changes: - Add helper function for updating bpf maps elements (He Kuang) - Fix traceevents plugins build race (Jiri Olsa) - Add the $OUTPUT path prefix with 'fixdep' (Jiri Olsa) Signed-off-by:
-
- 26 Nov, 2015 3 commits
-
-
He Kuang authored
Add bpf_map_update_elem() helper function which calls the sys_bpf syscall to update elements in bpf maps. Upcoming patches will use it to adjust data in map through the perf command line. Signed-off-by:
-
Jiri Olsa authored
Adding WIEGHT bit_name call to display sample_type properly. $ perf evlist -v cpu/mem-loads/pp: ...SNIP... sample_type: IP|TID|TIME|ADDR|ID|CPU|DATA_SRC|WEIGHT ... Signed-off-by:
-
Jiri Olsa authored
Clear sample_(type|period) for counting, as it only confuses debug output with unwanted sampling details: Before: $ sudo perf stat -e 'raw_syscalls:sys_enter' -vv ls ------------------------------------------------------------ perf_event_attr: type 2 size 112 config 0x11 { sample_period, sample_freq } 1 sample_type TIME|CPU|PERIOD|RAW read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 enable_on_exec 1 exclude_guest 1 ... After: $ sudo perf stat -e 'raw_syscalls:sys_enter' -vv ls ------------------------------------------------------------ perf_event_attr: type 2 size 112 config 0x11 read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 enable_on_exec 1 exclude_guest 1 ... Signed-off-by:
-
