- 16 Nov, 2017 20 commits
-
-
Arnaldo Carvalho de Melo authored
Two more, that were just in perf/core and thus weren't covered by Ingo's latest headers synch, kcmp.h and prctl.h, silencing this: Warning: Kernel ABI header at 'tools/include/uapi/linux/kcmp.h' differs from latest version at 'include/uapi/linux/kcmp.h' Warning: Kernel ABI header at 'tools/include/uapi/linux/prctl.h' differs from latest version at 'include/uapi/linux/prctl.h' Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: http://lkml.kernel.org/n/tip-2a0r7iybyqpkftllyy5t9hfk@git.kernel.orgSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Jiri Olsa authored
Align source with offset lines, which are more advanced, because of the address column. Before: : static void *worker_thread(void *__tdata) : { 0.00 : 48a971: push %rbp 0.00 : 48a972: mov %rsp,%rbp 0.00 : 48a975: sub $0x30,%rsp 0.00 : 48a979: mov %rdi,-0x28(%rbp) 0.00 : 48a97d: mov %fs:0x28,%rax 0.00 : 48a986: mov %rax,-0x8(%rbp) 0.00 : 48a98a: xor %eax,%eax : struct thread_data *td = __tdata; 0.00 : 48a98c: mov -0x28(%rbp),%rax 0.00 : 48a990: mov %rax,-0x10(%rbp) : int m = 0, i; 0.00 : 48a994: movl $0x0,-0x1c(%rbp) : int ret; : : for (i = 0; i < loops; i++) { 0.00 : 48a99b: movl $0x0,-0x18(%rbp) After: : static void *worker_thread(void *__tdata) : { 0.00 : 48a971: push %rbp 0.00 : 48a972: mov %rsp,%rbp 0.00 : 48a975: sub $0x30,%rsp 0.00 : 48a979: mov %rdi,-0x28(%rbp) 0.00 : 48a97d: mov %fs:0x28,%rax 0.00 : 48a986: mov %rax,-0x8(%rbp) 0.00 : 48a98a: xor %eax,%eax : struct thread_data *td = __tdata; 0.00 : 48a98c: mov -0x28(%rbp),%rax 0.00 : 48a990: mov %rax,-0x10(%rbp) : int m = 0, i; 0.00 : 48a994: movl $0x0,-0x1c(%rbp) : int ret; : : for (i = 0; i < loops; i++) { 0.00 : 48a99b: movl $0x0,-0x18(%rbp) It makes bigger different when displaying script sources, where the comment lines looks oddly shifted from the lines which actually hold code. I'll send script support separately. Committer note: Do not use a fixed column width for the addresses, as kernel ones se more than 10 columns, look at the last offset and get the right width. Signed-off-by:Jiri Olsa <jolsa@kernel.org> Cc: Andi Kleen <andi@firstfloor.org> Cc: David Ahern <dsahern@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20171011150158.11895-36-jolsa@kernel.orgSigned-off-by:
-
Jiri Olsa authored
Factor disasm_line__write function from annotate_browser__write, which now keeps only generic display code. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line in browser::b::top. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line in find functions: annotate_browser__find_string annotate_browser__find_string_reverse Signed-off-by:
-
Jiri Olsa authored
Using struct annotation_line arg in browser_line function to make it generic. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line as a browser::offsets array entry. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line as a browser::selection. We want to be able to use the annotate_browser for all sorts of source data, so it needs to be able to work over the generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Rename disasm_line__browser function to browser_line, because the browser got generic and is no longer disasm specific. Signed-off-by:
-
Jiri Olsa authored
Rename struct browser_disasm_line to browser_line, because the browser operates now on generic lines and no longer on disasm lines. Signed-off-by:
-
Jiri Olsa authored
We now keep samples_nr in struct annotation_line, so there's no need to pass nr_events to disasm_rb_tree__insert function. Signed-off-by:
-
Jiri Olsa authored
We now carry the data in 'struct annotation_line', so using it instead of samples from 'struct browser_disasm_line' and removing it and its setup. Signed-off-by:
-
Jiri Olsa authored
Move generic annotation line display code into annotation_line__print function. Signed-off-by:
-
Jiri Olsa authored
Separating struct annotation_line display function, it will hold the generic line display code. Signed-off-by:
-
Jiri Olsa authored
Remove struct source_line*, no longer needed. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() function, because it's no longer needed. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() from annotate_browser__calc_percent(), because we already have the data calculated in struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() from disasm_line__print(), because we already have the data calculated in struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Replace symbol__get_source_line() with symbol__calc_lines(), which calculates the source line tree over the struct annotation_line. This will allow us to remove redundant struct source_line in following patches. Signed-off-by:
-
Jiri Olsa authored
Add symbol__calc_percent function, that calculates annotation data for symbol and put the data in the struct annotation_line::samples array. Committer notes: Made symbol__calc_percent non static to be used in the next two patches, which will get some fixups from jolsa, doing it this way to keep this bisectable. Signed-off-by:
-
- 13 Nov, 2017 20 commits
-
-
Jiri Olsa authored
Add samples array into struct annotation_line to hold the annotation data. The data is populated in the following patches. Signed-off-by:
-
Jiri Olsa authored
Mov disasm__purge() to annotated_source__purge() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Changing the way the annotation lines are allocated and adding annotation_line__(new|delete) functions to deal with this. Before the allocation schema was as follows: ----------------------------------------------------------- struct disasm_line | struct annotation_line | private space ----------------------------------------------------------- Where the private space is used in TUI code to store computed annotation data for events. The stdio code computes the data on the fly. The goal is to compute and store annotation line's data directly in the struct annotation_line itself, so this patch changes the line allocation schema as follows: ------------------------------------------------------------ privsize space | struct disasm_line | struct annotation_line ------------------------------------------------------------ Moving struct annotation_line to the end, because in following changes we will move here the non-fixed length event's data. Signed-off-by:
-
Jiri Olsa authored
Move rb_node to struct annotation_line to make struct annotation_line the rb tree node for sorted lines used in both stdio and TUI code. This way we can unite the sorted lines lines codes for both TUI and stdio in the following patches. Signed-off-by:
-
Jiri Olsa authored
Rename disasm__add() into annotation_line__add() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Rename disasm__get_next_ip_line() to annotation_line__next() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Add evsel into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. This change also allow us to move the arch name initialization under symbol__annotate function. Link: http://lkml.kernel.org/n/tip-a9ok53rrgt1s5e8uglyvy6qt@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Add offset/line/line_nr into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. Signed-off-by:
-
Jiri Olsa authored
Add map into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. Signed-off-by:
-
Jiri Olsa authored
Add arch into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. Signed-off-by:
-
Jiri Olsa authored
Adding struct annotate_args to reduce the number of arguments, that need to travel all the way to line allocation. This makes the code easier to read and ease up the changes for following patches. Signed-off-by:
-
Jiri Olsa authored
Add symbol__annotate function to have generic annotation function to be called for all annotation sources. It calls the generic annotation init and then the specific annotation data retrieval function. Signed-off-by:
-
Jiri Olsa authored
Move ipc/cycles into annotation_line struct to be used as generic members for any annotation source. Signed-off-by:
-
Jiri Olsa authored
Move the line/line_nr/offset menbers to the annotation_line struct to be used as generic members for any annotation source. Signed-off-by:
-
Jiri Olsa authored
In order to make the annotation support generic, addadding 'struct annotation_line', which will hold generic data common to annotation sources (such as the one for python scripts, coming on upcoming patches). Having this, we can add different annotation line support other than objdump disasm. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
When we use an initial delay, e.g.: 'perf record --delay 1000', we do not enable the events until that delay has passed after we started the workload, including the tracking event, i.e. the one for which we have attr.mmap, etc, enabled to ask the kernel to generate the PERF_RECORD_{MMAP,COMM,EXEC} metadata events that will then allow us to resolve addresses in samples to the map, dso and symbol. There will be a shadow that even synthesizing samples won't cover, i.e. the workload that we start and other processes forking while we wait for the initial delay to expire. So use a dummy event to be the tracking one and make it be enabled on exec. Before: # perf record --delay 1000 stress --cpu 1 --timeout 5 stress: info: [9029] dispatching hogs: 1 cpu, 0 io, 0 vm, 0 hdd stress: info: [9029] successful run completed in 5s [ perf record: Woken up 3 times to write data ] [ perf record: Captured and wrote 0.624 MB perf.data (15908 samples) ] # perf script | head :9031 9031 32001.826888: 1 cycles:ppp: ffffffff831aa30d event_function (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826893: 1 cycles:ppp: ffffffff8300d1a0 intel_bts_enable_local (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826895: 7 cycles:ppp: ffffffff83023870 sched_clock (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826897: 103 cycles:ppp: ffffffff8300c331 intel_pmu_handle_irq (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826899: 1615 cycles:ppp: ffffffff830231f8 native_sched_clock (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826902: 26724 cycles:ppp: ffffffff8384c6a7 native_irq_return_iret (/lib/modules/4.14.0-rc6+/build/vmlinux) :9031 9031 32001.826913: 329739 cycles:ppp: 7fb2a5410932 [unknown] ([unknown]) :9031 9031 32001.827033: 1225451 cycles:ppp: 7fb2a5410930 [unknown] ([unknown]) :9031 9031 32001.827474: 1391725 cycles:ppp: 7fb2a5410930 [unknown] ([unknown]) :9031 9031 32001.827978: 1233697 cycles:ppp: 7fb2a5410928 [unknown] ([unknown]) # After: # perf record --delay 1000 stress --cpu 1 --timeout 5 stress: info: [9741] dispatching hogs: 1 cpu, 0 io, 0 vm, 0 hdd stress: info: [9741] successful run completed in 5s [ perf record: Woken up 3 times to write data ] [ perf record: Captured and wrote 0.751 MB perf.data (15976 samples) ] # perf script | head stress 9742 32110.959106: 1 cycles:ppp: ffffffff831b26f6 __perf_event_task_sched_in (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959110: 1 cycles:ppp: ffffffff8300c2e9 intel_pmu_handle_irq (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959112: 7 cycles:ppp: ffffffff830231e0 native_sched_clock (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959115: 101 cycles:ppp: ffffffff83023870 sched_clock (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959117: 1533 cycles:ppp: ffffffff830231f8 native_sched_clock (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959119: 23992 cycles:ppp: ffffffff831b0900 ctx_sched_in (/lib/modules/4.14.0-rc6+/build/vmlinux) stress 9742 32110.959129: 329406 cycles:ppp: 7f4b1b661930 __random_r (/usr/lib64/libc-2.25.so) stress 9742 32110.959249: 1288322 cycles:ppp: 5566e1e7cbc9 hogcpu (/usr/bin/stress) stress 9742 32110.959712: 1464046 cycles:ppp: 7f4b1b66179e __random (/usr/lib64/libc-2.25.so) stress 9742 32110.960241: 1266918 cycles:ppp: 7f4b1b66195b __random_r (/usr/lib64/libc-2.25.so) # Reported-by:Bram Stolk <b.stolk@gmail.com> Tested-by:
-
Arnaldo Carvalho de Melo authored
The evsel->idx field is used mainly to access the right bucket in per-event arrays such as the annotation ones, but also to set evsel->tracking, that in turn will decide what of the events will ask for PERF_RECORD_{MMAP,COMM,EXEC} to be generated, i.e. which perf_event_attr will have its mmap, etc fields set. When we were adding the "dummy" event using perf_evlist__add_dummy() we were not setting it correctly, which could result in multiple tracking events. Now that I'll try using a dummy event to be the tracking one when using 'perf record --delay', i.e. when we process the --delay setting we may already have the evlist set up, like with: perf record -e cycles,instructions --delay 1000 ./workload We will need to add a "dummy" event, then reset evsel->tracking for the first event, "cycles", and set it instead to the dummy one, and also setting its attr.enable_on_exec, so that we get the PERF_RECORD_MMAP, etc metadata events while waiting to enable the explicitely requested events, so lets get this straight and set the right evsel->idx. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Bram Stolk <b.stolk@gmail.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: http://lkml.kernel.org/n/tip-nrdfchshqxf7diszhxcecqb9@git.kernel.orgSigned-off-by: -
Arnaldo Carvalho de Melo authored
To pick up fixes. Signed-off-by: -
Ingo Molnar authored
The jprobes APIs are deprecated - but are still in occasional use for code that few people seem to care about, so stop generating deprecation warnings. Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by:Ingo Molnar <mingo@kernel.org>
-
Linus Torvalds authored
The (alleged) users of the module addresses are the same: kernel profiling. So just expose the same helper and format macros, and unify the logic. Signed-off-by:Linus Torvalds <torvalds@linux-foundation.org>
-
