kernel/tools/perf/util, branch linux-5.1.y

perf stat: Fix group lookup for metric group

2019-07-26T07:12:48Z

[ Upstream commit 2f87f33f4226523df9c9cc28f9874ea02fcc3d3f ] The metric group code tries to find a group it added earlier in the evlist. Fix the lookup to handle groups with partially overlaps correctly. When a sub string match fails and we reset the match, we have to compare the first element again. I also renamed the find_evsel function to find_evsel_group to make its purpose clearer. With the earlier changes this fixes: Before: % perf stat -M UPI,IPC sleep 1 ... 1,032,922 uops_retired.retire_slots # 1.1 UPI 1,896,096 inst_retired.any 1,896,096 inst_retired.any 1,177,254 cpu_clk_unhalted.thread After: % perf stat -M UPI,IPC sleep 1 ... 1,013,193 uops_retired.retire_slots # 1.1 UPI 932,033 inst_retired.any 932,033 inst_retired.any # 0.9 IPC 1,091,245 cpu_clk_unhalted.thread Signed-off-by: Andi Kleen Acked-by: Jiri Olsa Cc: Kan Liang Fixes: b18f3e365019 ("perf stat: Support JSON metrics in perf stat") Link: http://lkml.kernel.org/r/20190624193711.35241-4-andi@firstfloor.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf stat: Don't merge events in the same PMU

2019-07-26T07:12:48Z

[ Upstream commit 6c5f4e5cb35b4694dc035d91092d23f596ecd06a ] Event merging is mainly to collapse similar events in lots of different duplicated PMUs. It can break metric displaying. It's possible for two metrics to have the same event, and when the two events happen in a row the second wouldn't be displayed. This would also not show the second metric. To avoid this don't merge events in the same PMU. This makes sense, if we have multiple events in the same PMU there is likely some reason for it (e.g. using multiple groups) and we better not merge them. While in theory it would be possible to construct metrics that have events with the same name in different PMU no current metrics have this problem. This is the fix for perf stat -M UPI,IPC (needs also another bug fix to completely work) Signed-off-by: Andi Kleen Acked-by: Jiri Olsa Cc: Kan Liang Fixes: 430daf2dc7af ("perf stat: Collapse identically named events") Link: http://lkml.kernel.org/r/20190624193711.35241-3-andi@firstfloor.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf stat: Fix metrics with --no-merge

2019-07-26T07:12:48Z

[ Upstream commit e3a9427323a53ceee540276a74af7706f350d052 ] Since Fixes: 8c5421c016a4 ("perf pmu: Display pmu name when printing unmerged events in stat") using --no-merge adds the PMU name to the evsel name. This breaks the metric value lookup because the parser doesn't know about this. Remove the extra postfixes for the metric evaluation. Signed-off-by: Andi Kleen Acked-by: Jiri Olsa Cc: Agustin Vega-Frias Cc: Kan Liang Fixes: 8c5421c016a4 ("perf pmu: Display pmu name when printing unmerged events in stat") Link: http://lkml.kernel.org/r/20190624193711.35241-5-andi@firstfloor.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf stat: Make metric event lookup more robust

2019-07-26T07:12:48Z

[ Upstream commit 145c407c808352acd625be793396fd4f33c794f8 ] After setting up metric groups through the event parser, the metricgroup code looks them up again in the event list. Make sure we only look up events that haven't been used by some other metric. The data structures currently cannot handle more than one metric per event. This avoids problems with multiple events partially overlapping. Signed-off-by: Andi Kleen Acked-by: Jiri Olsa Cc: Kan Liang Link: http://lkml.kernel.org/r/20190624193711.35241-2-andi@firstfloor.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf tools: Increase MAX_NR_CPUS and MAX_CACHES

2019-07-26T07:12:43Z

[ Upstream commit 9f94c7f947e919c343b30f080285af53d0fa9902 ] Attempting to profile 1024 or more CPUs with perf causes two errors: perf record -a [ perf record: Woken up X times to write data ] way too many cpu caches.. [ perf record: Captured and wrote X MB perf.data (X samples) ] perf report -C 1024 Error: failed to set cpu bitmap Requested CPU 1024 too large. Consider raising MAX_NR_CPUS Increasing MAX_NR_CPUS from 1024 to 2048 and redefining MAX_CACHES as MAX_NR_CPUS * 4 returns normal functionality to perf: perf record -a [ perf record: Woken up X times to write data ] [ perf record: Captured and wrote X MB perf.data (X samples) ] perf report -C 1024 ... Signed-off-by: Kyle Meyer Cc: Alexander Shishkin Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20190620193630.154025-1-meyerk@stormcage.eag.rdlabs.hpecorp.net Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf evsel: Make perf_evsel__name() accept a NULL argument

2019-07-26T07:12:36Z

[ Upstream commit fdbdd7e8580eac9bdafa532746c865644d125e34 ] In which case it simply returns "unknown", like when it can't figure out the evsel->name value. This makes this code more robust and fixes a problem in 'perf trace' where a NULL evsel was being passed to a routine that only used the evsel for printing its name when a invalid syscall id was passed. Reported-by: Leo Yan Cc: Adrian Hunter Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lkml.kernel.org/n/tip-f30ztaasku3z935cn3ak3h53@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf report: Fix OOM error in TUI mode on s390

2019-07-26T07:12:33Z

[ Upstream commit 8a07aa4e9b7b0222129c07afff81634a884b2866 ] Debugging a OOM error using the TUI interface revealed this issue on s390: [tmricht@m83lp54 perf]$ cat /proc/kallsyms |sort .... 00000001119b7158 B radix_tree_node_cachep 00000001119b8000 B __bss_stop 00000001119b8000 B _end 000003ff80002850 t autofs_mount [autofs4] 000003ff80002868 t autofs_show_options [autofs4] 000003ff80002a98 t autofs_evict_inode [autofs4] .... There is a huge gap between the last kernel symbol __bss_stop/_end and the first kernel module symbol autofs_mount (from autofs4 module). After reading the kernel symbol table via functions: dso__load() +--> dso__load_kernel_sym() +--> dso__load_kallsyms() +--> __dso_load_kallsyms() +--> symbols__fixup_end() the symbol __bss_stop has a start address of 1119b8000 and an end address of 3ff80002850, as can be seen by this debug statement: symbols__fixup_end __bss_stop start:0x1119b8000 end:0x3ff80002850 The size of symbol __bss_stop is 0x3fe6e64a850 bytes! It is the last kernel symbol and fills up the space until the first kernel module symbol. This size kills the TUI interface when executing the following code: process_sample_event() hist_entry_iter__add() hist_iter__report_callback() hist_entry__inc_addr_samples() symbol__inc_addr_samples(symbol = __bss_stop) symbol__cycles_hist() annotated_source__alloc_histograms(..., symbol__size(sym), ...) This function allocates memory to save sample histograms. The symbol_size() marco is defined as sym->end - sym->start, which results in above value of 0x3fe6e64a850 bytes and the call to calloc() in annotated_source__alloc_histograms() fails. The histgram memory allocation might fail, make this failure no-fatal and continue processing. Output before: [tmricht@m83lp54 perf]$ ./perf --debug stderr=1 report -vvvvv \ -i ~/slow.data 2>/tmp/2 [tmricht@m83lp54 perf]$ tail -5 /tmp/2 __symbol__inc_addr_samples(875): ENOMEM! sym->name=__bss_stop, start=0x1119b8000, addr=0x2aa0005eb08, end=0x3ff80002850, func: 0 problem adding hist entry, skipping event 0x938b8 [0x8]: failed to process type: 68 [Cannot allocate memory] [tmricht@m83lp54 perf]$ Output after: [tmricht@m83lp54 perf]$ ./perf --debug stderr=1 report -vvvvv \ -i ~/slow.data 2>/tmp/2 [tmricht@m83lp54 perf]$ tail -5 /tmp/2 symbol__inc_addr_samples map:0x1597830 start:0x110730000 end:0x3ff80002850 symbol__hists notes->src:0x2aa2a70 nr_hists:1 symbol__inc_addr_samples sym:unlink_anon_vmas src:0x2aa2a70 __symbol__inc_addr_samples: addr=0x11094c69e 0x11094c670 unlink_anon_vmas: period++ [addr: 0x11094c69e, 0x2e, evidx=0] => nr_samples: 1, period: 526008 [tmricht@m83lp54 perf]$ There is no error about failed memory allocation and the TUI interface shows all entries. Signed-off-by: Thomas Richter Reviewed-by: Hendrik Brueckner Cc: Heiko Carstens Cc: Hendrik Brueckner Link: http://lkml.kernel.org/r/90cb5607-3e12-5167-682d-978eba7dafa8@linux.ibm.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf header: Assign proper ff->ph in perf_event__synthesize_features()

2019-07-14T06:09:48Z

commit c952b35f4b15dd1b83e952718dec3307256383ef upstream. bpf/btf write_* functions need ff->ph->env. With this missing, pipe-mode (perf record -o -) would crash like: Program terminated with signal SIGSEGV, Segmentation fault. This patch assign proper ph value to ff. Committer testing: (gdb) run record -o - Starting program: /root/bin/perf record -o - PERFILE2 Thread 1 "perf" received signal SIGSEGV, Segmentation fault. __do_write_buf (size=4, buf=0x160, ff=0x7fffffff8f80) at util/header.c:126 126 memcpy(ff->buf + ff->offset, buf, size); (gdb) bt #0 __do_write_buf (size=4, buf=0x160, ff=0x7fffffff8f80) at util/header.c:126 #1 do_write (ff=ff@entry=0x7fffffff8f80, buf=buf@entry=0x160, size=4) at util/header.c:137 #2 0x00000000004eddba in write_bpf_prog_info (ff=0x7fffffff8f80, evlist=) at util/header.c:912 #3 0x00000000004f69d7 in perf_event__synthesize_features (tool=tool@entry=0x97cc00 , session=session@entry=0x7fffe9c6d010, evlist=0x7fffe9cae010, process=process@entry=0x4435d0 ) at util/header.c:3695 #4 0x0000000000443c79 in record__synthesize (tail=tail@entry=false, rec=0x97cc00 ) at builtin-record.c:1214 #5 0x0000000000444ec9 in __cmd_record (rec=0x97cc00 , argv=, argc=0) at builtin-record.c:1435 #6 cmd_record (argc=0, argv=) at builtin-record.c:2450 #7 0x00000000004ae3e9 in run_builtin (p=p@entry=0x98e058 , argc=argc@entry=3, argv=0x7fffffffd670) at perf.c:304 #8 0x000000000042eded in handle_internal_command (argv=, argc=) at perf.c:356 #9 run_argv (argcp=, argv=) at perf.c:400 #10 main (argc=3, argv=) at perf.c:522 (gdb) After the patch the SEGSEGV is gone. Reported-by: David Carrillo Cisneros Signed-off-by: Song Liu Tested-by: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Namhyung Kim Cc: kernel-team@fb.com Cc: stable@vger.kernel.org # v5.1+ Fixes: 606f972b1361 ("perf bpf: Save bpf_prog_info information as headers to perf.data") Link: http://lkml.kernel.org/r/20190620010453.4118689-1-songliubraving@fb.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman

perf thread-stack: Fix thread stack return from kernel for kernel-only case

2019-07-14T06:09:48Z

commit 97860b483c5597663a174ff7405be957b4838391 upstream. Commit f08046cb3082 ("perf thread-stack: Represent jmps to the start of a different symbol") had the side-effect of introducing more stack entries before return from kernel space. When user space is also traced, those entries are popped before entry to user space, but when user space is not traced, they get stuck at the bottom of the stack, making the stack grow progressively larger. Fix by detecting a return-from-kernel branch type, and popping kernel addresses from the stack then. Note, the problem and fix affect the exported Call Graph / Tree but not the callindent option used by "perf script --call-trace". Example: perf-with-kcore record example -e intel_pt//k -- ls perf-with-kcore script example --itrace=bep -s ~/libexec/perf-core/scripts/python/export-to-sqlite.py example.db branches calls ~/libexec/perf-core/scripts/python/exported-sql-viewer.py example.db Menu option: Reports -> Context-Sensitive Call Graph Before: (showing Call Path column only) Call Path ▶ perf ▼ ls ▼ 12111:12111 ▶ setup_new_exec ▶ __task_pid_nr_ns ▶ perf_event_pid_type ▶ perf_event_comm_output ▶ perf_iterate_ctx ▶ perf_iterate_sb ▶ perf_event_comm ▶ __set_task_comm ▶ load_elf_binary ▶ search_binary_handler ▶ __do_execve_file.isra.41 ▶ __x64_sys_execve ▶ do_syscall_64 ▼ entry_SYSCALL_64_after_hwframe ▼ swapgs_restore_regs_and_return_to_usermode ▼ native_iret ▶ error_entry ▶ do_page_fault ▼ error_exit ▼ retint_user ▶ prepare_exit_to_usermode ▼ native_iret ▶ error_entry ▶ do_page_fault ▼ error_exit ▼ retint_user ▶ prepare_exit_to_usermode ▼ native_iret ▶ error_entry ▶ do_page_fault ▼ error_exit ▼ retint_user ▶ prepare_exit_to_usermode ▶ native_iret After: (showing Call Path column only) Call Path ▶ perf ▼ ls ▼ 12111:12111 ▶ setup_new_exec ▶ __task_pid_nr_ns ▶ perf_event_pid_type ▶ perf_event_comm_output ▶ perf_iterate_ctx ▶ perf_iterate_sb ▶ perf_event_comm ▶ __set_task_comm ▶ load_elf_binary ▶ search_binary_handler ▶ __do_execve_file.isra.41 ▶ __x64_sys_execve ▶ do_syscall_64 ▶ entry_SYSCALL_64_after_hwframe ▶ page_fault ▼ entry_SYSCALL_64 ▼ do_syscall_64 ▶ __x64_sys_brk ▶ __x64_sys_access ▶ __x64_sys_openat ▶ __x64_sys_newfstat ▶ __x64_sys_mmap ▶ __x64_sys_close ▶ __x64_sys_read ▶ __x64_sys_mprotect ▶ __x64_sys_arch_prctl ▶ __x64_sys_munmap ▶ exit_to_usermode_loop ▶ __x64_sys_set_tid_address ▶ __x64_sys_set_robust_list ▶ __x64_sys_rt_sigaction ▶ __x64_sys_rt_sigprocmask ▶ __x64_sys_prlimit64 ▶ __x64_sys_statfs ▶ __x64_sys_ioctl ▶ __x64_sys_getdents64 ▶ __x64_sys_write ▶ __x64_sys_exit_group Committer notes: The first arg to the perf-with-kcore needs to be the same for the 'record' and 'script' lines, otherwise we'll record the perf.data file and kcore_dir/ files in one directory ('example') to then try to use it from the 'bep' directory, fix the instructions above it so that both use 'example'. Signed-off-by: Adrian Hunter Tested-by: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: stable@vger.kernel.org Fixes: f08046cb3082 ("perf thread-stack: Represent jmps to the start of a different symbol") Link: http://lkml.kernel.org/r/20190619064429.14940-2-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman

perf pmu: Fix uncore PMU alias list for ARM64

2019-07-14T06:09:48Z

commit 599ee18f0740d7661b8711249096db94c09bc508 upstream. In commit 292c34c10249 ("perf pmu: Fix core PMU alias list for X86 platform"), we fixed the issue of CPU events being aliased to uncore events. Fix this same issue for ARM64, since the said commit left the (broken) behaviour untouched for ARM64. Signed-off-by: John Garry Cc: Alexander Shishkin Cc: Ben Hutchings Cc: Hendrik Brueckner Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Mathieu Poirier Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Shaokun Zhang Cc: Thomas Richter Cc: Will Deacon Cc: linux-arm-kernel@lists.infradead.org Cc: linuxarm@huawei.com Cc: stable@vger.kernel.org Fixes: 292c34c10249 ("perf pmu: Fix core PMU alias list for X86 platform") Link: http://lkml.kernel.org/r/1560521283-73314-2-git-send-email-john.garry@huawei.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman