From 91557e847dbc715acf2d847a1ffec63f71b00b65 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 30 May 2012 12:25:52 -0300 Subject: perf report: Use the right symbol for annotation In non symbolic views, i.e. --sort without "symbol", as in: perf report --sort comm We're segfaulting in the --tui because we're testing the symbol resolved and then trying to use the symbol on the histogram entry where we're coalescing all hits for a COMM, and the first hist_entry for a comm may have a NULL symbol, i.e. the RIP didn't resolve to any symbol. In this case we're segfaulting, fix it by testing against the symbol in the histogram entry. Reported-by: William Cohen Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-8ylwubbcmu27ucc9ffrku3yv@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-report.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 8c767c6bca91..2400e009f149 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -162,7 +162,7 @@ static int perf_evsel__add_hist_entry(struct perf_evsel *evsel, * so we don't allocated the extra space needed because the stdio * code will not use it. */ - if (al->sym != NULL && use_browser > 0) { + if (he->ms.sym != NULL && use_browser > 0) { struct annotation *notes = symbol__annotation(he->ms.sym); assert(evsel != NULL); -- cgit From 107baecaca0b2843f1a0464701b253e51ef6f0e2 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 30 May 2012 12:31:44 -0300 Subject: perf annotate browser: Fix help window entry for navigating to hottest line Its 'H', not 'h'. The later is for getting to the help window. Reported-by: Ingo Molnar Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-7zvwphhm815y2zczoxgstzuf@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/annotate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c index 4deea6aaf927..34b1c46eaf42 100644 --- a/tools/perf/ui/browsers/annotate.c +++ b/tools/perf/ui/browsers/annotate.c @@ -668,7 +668,7 @@ static int annotate_browser__run(struct annotate_browser *browser, int evidx, "q/ESC/CTRL+C Exit\n\n" "-> Go to target\n" "<- Exit\n" - "h Cycle thru hottest instructions\n" + "H Cycle thru hottest instructions\n" "j Toggle showing jump to target arrows\n" "J Toggle showing number of jump sources on targets\n" "n Search next string\n" -- cgit From 79695e1bb65ba0e21488c360a1bed6e358354aaa Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 30 May 2012 13:53:54 -0300 Subject: perf stat: Initialize default events wrt exclude_{guest,host} When no event is specified the tools use perf_evlist__add_default(), that will call event_attr_init to initialize the KVM exclusion bits. When the change was made to the tools so that by default guest samples would be excluded, the changes were made just to the parsing routines and to perf_evlist__add_default(), not to perf_evlist__add_attrs, that is used so far just by perf stat to add multiple events, according to the level of detail specified. Recently the tools were changed to reconstruct the event name from all the details in perf_event_attr, not just from .type and .config, but taking into account all the feature bits (.exclude_{guest,host,user,kernel,etc}, .precise_ip, etc). That is when we noticed that the default for perf stat wasn't the one for the rest of the tools, i.e. the .exclude_guest bit wasn't being set. I.e. the default, that doesn't call event_attr_init was showing the :HG modifier: $ perf stat usleep 1 Performance counter stats for 'usleep 1': 0.942119 task-clock # 0.454 CPUs utilized 1 context-switches # 0.001 M/sec 0 CPU-migrations # 0.000 K/sec 126 page-faults # 0.134 M/sec 693,193 cycles:HG # 0.736 GHz [40.11%] 407,461 stalled-cycles-frontend:HG # 58.78% frontend cycles idle [72.29%] 365,403 stalled-cycles-backend:HG # 52.71% backend cycles idle 465,982 instructions:HG # 0.67 insns per cycle # 0.87 stalled cycles per insn 89,760 branches:HG # 95.275 M/sec 6,178 branch-misses:HG # 6.88% of all branches 0.002077228 seconds time elapsed While if one explicitely specifies the same events, which will make the parsing code to be called and thus event_attr_init is called: $ perf stat -e task-clock,context-switches,migrations,page-faults,cycles,stalled-cycles-frontend,stalled-cycles-backend,instructions,branches,branch-misses usleep 1 Performance counter stats for 'usleep 1': 1.040349 task-clock # 0.500 CPUs utilized 2 context-switches # 0.002 M/sec 0 CPU-migrations # 0.000 K/sec 127 page-faults # 0.122 M/sec 587,966 cycles # 0.565 GHz [13.18%] 459,167 stalled-cycles-frontend # 78.09% frontend cycles idle 390,249 stalled-cycles-backend # 66.37% backend cycles idle 504,006 instructions # 0.86 insns per cycle # 0.91 stalled cycles per insn 96,455 branches # 92.714 M/sec 6,522 branch-misses # 6.76% of all branches [96.12%] 0.002078681 seconds time elapsed Fix it by introducing a perf_evlist__add_default_attrs method that will call evlist_attr_init in all the perf_event_attr entries before adding the events. Reported-by: Ingo Molnar Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-4eysr236r0pgiyum9epwxw7s@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-stat.c | 8 ++++---- tools/perf/util/evlist.c | 11 +++++++++++ tools/perf/util/evlist.h | 4 ++++ 3 files changed, 19 insertions(+), 4 deletions(-) diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 62ae30d34fa6..262589991ea4 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -1129,7 +1129,7 @@ static int add_default_attributes(void) return 0; if (!evsel_list->nr_entries) { - if (perf_evlist__add_attrs_array(evsel_list, default_attrs) < 0) + if (perf_evlist__add_default_attrs(evsel_list, default_attrs) < 0) return -1; } @@ -1139,21 +1139,21 @@ static int add_default_attributes(void) return 0; /* Append detailed run extra attributes: */ - if (perf_evlist__add_attrs_array(evsel_list, detailed_attrs) < 0) + if (perf_evlist__add_default_attrs(evsel_list, detailed_attrs) < 0) return -1; if (detailed_run < 2) return 0; /* Append very detailed run extra attributes: */ - if (perf_evlist__add_attrs_array(evsel_list, very_detailed_attrs) < 0) + if (perf_evlist__add_default_attrs(evsel_list, very_detailed_attrs) < 0) return -1; if (detailed_run < 3) return 0; /* Append very, very detailed run extra attributes: */ - return perf_evlist__add_attrs_array(evsel_list, very_very_detailed_attrs); + return perf_evlist__add_default_attrs(evsel_list, very_very_detailed_attrs); } int cmd_stat(int argc, const char **argv, const char *prefix __used) diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index 4ac5f5ae4ce9..ed277e5627cf 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -159,6 +159,17 @@ out_delete_partial_list: return -1; } +int __perf_evlist__add_default_attrs(struct perf_evlist *evlist, + struct perf_event_attr *attrs, size_t nr_attrs) +{ + size_t i; + + for (i = 0; i < nr_attrs; i++) + event_attr_init(attrs + i); + + return perf_evlist__add_attrs(evlist, attrs, nr_attrs); +} + static int trace_event__id(const char *evname) { char *filename, *colon; diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h index 58abb63ac13a..989bee9624c2 100644 --- a/tools/perf/util/evlist.h +++ b/tools/perf/util/evlist.h @@ -54,6 +54,8 @@ void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry); int perf_evlist__add_default(struct perf_evlist *evlist); int perf_evlist__add_attrs(struct perf_evlist *evlist, struct perf_event_attr *attrs, size_t nr_attrs); +int __perf_evlist__add_default_attrs(struct perf_evlist *evlist, + struct perf_event_attr *attrs, size_t nr_attrs); int perf_evlist__add_tracepoints(struct perf_evlist *evlist, const char *tracepoints[], size_t nr_tracepoints); int perf_evlist__set_tracepoints_handlers(struct perf_evlist *evlist, @@ -62,6 +64,8 @@ int perf_evlist__set_tracepoints_handlers(struct perf_evlist *evlist, #define perf_evlist__add_attrs_array(evlist, array) \ perf_evlist__add_attrs(evlist, array, ARRAY_SIZE(array)) +#define perf_evlist__add_default_attrs(evlist, array) \ + __perf_evlist__add_default_attrs(evlist, array, ARRAY_SIZE(array)) #define perf_evlist__add_tracepoints_array(evlist, array) \ perf_evlist__add_tracepoints(evlist, array, ARRAY_SIZE(array)) -- cgit From 52deff71bc2b2c24587ab71f588ff5e4c9279349 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Tue, 29 May 2012 22:58:26 -0600 Subject: perf script: Fix regression in callchain dso name $ perf script -i /tmp/perf.data ... gcc 13623 544315.062858: context-switches: ffffffff815f65c9 __schedule ([kernel.kallsyms]) ffffffff81087cea __cond_resched ([kernel.kallsyms]) ffffffff815f6b92 _cond_resched ([kernel.kallsyms]) ffffffff815fb87a do_page_fault ([kernel.kallsyms]) ffffffff815f8465 page_fault ([kernel.kallsyms]) 2b7a71ea0303 _dl_lookup_symbol_x ([kernel.kallsyms]) 2b7a71ea1eb5 _dl_relocate_object ([kernel.kallsyms]) 2b7a71e99b2e dl_main ([kernel.kallsyms]) 2b7a71eab7f4 _dl_sysdep_start ([kernel.kallsyms]) All DSO's in a callchain are printed as [kernel.kallsyms]. git bisect chased it to: 547a92e0aedb88129e7fbd804697a11949de2e5a is the first bad commit commit 547a92e0aedb88129e7fbd804697a11949de2e5a Author: Akihiro Nagai Date: Mon Jan 30 13:42:57 2012 +0900 perf script: Unify the expressions indicating "unknown" The perf script command uses various expressions to indicate "unknown". It is unfriendly for user scripts to parse it. So, this patch unifies the expressions to "[unknown]". Looks like a copy-paste in that the other references use al.map but this one should be node->map. With this patch you get: $ perf script -i /tmp/perf.data ... gcc 13623 544315.062858: context-switches: ffffffff815f65c9 __schedule ([kernel.kallsyms]) ffffffff81087cea __cond_resched ([kernel.kallsyms]) ffffffff815f6b92 _cond_resched ([kernel.kallsyms]) ffffffff815fb87a do_page_fault ([kernel.kallsyms]) ffffffff815f8465 page_fault ([kernel.kallsyms]) 2b7a71ea0303 _dl_lookup_symbol_x (/lib64/ld-2.14.90.so) 2b7a71ea1eb5 _dl_relocate_object (/lib64/ld-2.14.90.so) 2b7a71e99b2e dl_main (/lib64/ld-2.14.90.so) 2b7a71eab7f4 _dl_sysdep_start (/lib64/ld-2.14.90.so) Signed-off-by: David Ahern Cc: Akihiro Nagai Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1338353906-60706-1-git-send-email-dsahern@gmail.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 93d355d27109..48206144758e 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1460,7 +1460,7 @@ void perf_event__print_ip(union perf_event *event, struct perf_sample *sample, } if (print_dso) { printf(" ("); - map__fprintf_dsoname(al.map, stdout); + map__fprintf_dsoname(node->map, stdout); printf(")"); } printf("\n"); -- cgit From f1439c315b328bbb3f7f6fdd814a7057f9e130d5 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 30 May 2012 15:02:42 -0300 Subject: perf tools: Fix make tarballs The patch series that introduced the top level tools/ makefile and the libtraceevent broke this feature where files needed to build in a detached tarball were not included in the MANIFEST file and thus not included in the tarball. Fix it by adding the relevant files to the MANIFEST. Cc: Borislav Petkov Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Steven Rostedt Link: http://lkml.kernel.org/n/tip-z3mjj74927xvqwhlmu18kj80@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/MANIFEST | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST index 5476bc0a1eac..b4b572e8c100 100644 --- a/tools/perf/MANIFEST +++ b/tools/perf/MANIFEST @@ -1,4 +1,6 @@ tools/perf +tools/scripts +tools/lib/traceevent include/linux/const.h include/linux/perf_event.h include/linux/rbtree.h -- cgit From ea1b3ebac9a1c72c4362c784b4ed069a938a4ddb Mon Sep 17 00:00:00 2001 From: Avik Sil Date: Tue, 29 May 2012 16:05:25 +0530 Subject: perf tools: Fix pager on minimal-install embedded systems Some Distributions may lack "less" package being included by default, e.g., Linaro nano rootfs. In those cases use the portable "pager" command instead of "less". Signed-off-by: Avik Sil Acked-by: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338287725-26382-1-git-send-email-avik.sil@linaro.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/pager.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tools/perf/util/pager.c b/tools/perf/util/pager.c index 1915de20dcac..3322b8446e89 100644 --- a/tools/perf/util/pager.c +++ b/tools/perf/util/pager.c @@ -57,6 +57,10 @@ void setup_pager(void) } if (!pager) pager = getenv("PAGER"); + if (!pager) { + if (!access("/usr/bin/pager", X_OK)) + pager = "/usr/bin/pager"; + } if (!pager) pager = "less"; else if (!*pager || !strcmp(pager, "cat")) -- cgit From 472606458f3e1ced5fe3cc5f04e90a6b5a4732cf Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:43:26 +0900 Subject: perf callchain: Make callchain cursors TLS perf top -G has a race on callchain cursor between main thread and display thread. Since the callchain cursors are used locally make them thread-local data would solve the problem. Signed-off-by: Namhyung Kim Reported-by: Sunjin Yang Suggested-by: Arnaldo Carvalho de Melo Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sunjin Yang Link: http://lkml.kernel.org/r/1338443007-24857-1-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-report.c | 2 +- tools/perf/builtin-top.c | 2 +- tools/perf/util/callchain.c | 2 ++ tools/perf/util/callchain.h | 2 ++ tools/perf/util/hist.c | 7 ++++--- tools/perf/util/hist.h | 2 -- tools/perf/util/session.c | 14 +++++++------- 7 files changed, 17 insertions(+), 14 deletions(-) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 2400e009f149..25249f76329d 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -152,7 +152,7 @@ static int perf_evsel__add_hist_entry(struct perf_evsel *evsel, if (symbol_conf.use_callchain) { err = callchain_append(he->callchain, - &evsel->hists.callchain_cursor, + &callchain_cursor, sample->period); if (err) return err; diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 871b540293e1..6bb0277b7dfe 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -787,7 +787,7 @@ static void perf_event__process_sample(struct perf_tool *tool, } if (symbol_conf.use_callchain) { - err = callchain_append(he->callchain, &evsel->hists.callchain_cursor, + err = callchain_append(he->callchain, &callchain_cursor, sample->period); if (err) return; diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 9f7106a8d9a4..3a6bff47614f 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -18,6 +18,8 @@ #include "util.h" #include "callchain.h" +__thread struct callchain_cursor callchain_cursor; + bool ip_callchain__valid(struct ip_callchain *chain, const union perf_event *event) { diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index 7f9c0f1ae3a9..3bdb407f9cd9 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -76,6 +76,8 @@ struct callchain_cursor { struct callchain_cursor_node *curr; }; +extern __thread struct callchain_cursor callchain_cursor; + static inline void callchain_init(struct callchain_root *root) { INIT_LIST_HEAD(&root->node.siblings); diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 1293b5ebea4d..514e2a4b367d 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -378,7 +378,7 @@ void hist_entry__free(struct hist_entry *he) * collapse the histogram */ -static bool hists__collapse_insert_entry(struct hists *hists, +static bool hists__collapse_insert_entry(struct hists *hists __used, struct rb_root *root, struct hist_entry *he) { @@ -397,8 +397,9 @@ static bool hists__collapse_insert_entry(struct hists *hists, iter->period += he->period; iter->nr_events += he->nr_events; if (symbol_conf.use_callchain) { - callchain_cursor_reset(&hists->callchain_cursor); - callchain_merge(&hists->callchain_cursor, iter->callchain, + callchain_cursor_reset(&callchain_cursor); + callchain_merge(&callchain_cursor, + iter->callchain, he->callchain); } hist_entry__free(he); diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index cfc64e293f90..34bb556d6219 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -67,8 +67,6 @@ struct hists { struct events_stats stats; u64 event_stream; u16 col_len[HISTC_NR_COLS]; - /* Best would be to reuse the session callchain cursor */ - struct callchain_cursor callchain_cursor; }; struct hist_entry *__hists__add_entry(struct hists *self, diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 48206144758e..3b6f8e460a31 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -288,7 +288,8 @@ struct branch_info *machine__resolve_bstack(struct machine *self, return bi; } -int machine__resolve_callchain(struct machine *self, struct perf_evsel *evsel, +int machine__resolve_callchain(struct machine *self, + struct perf_evsel *evsel __used, struct thread *thread, struct ip_callchain *chain, struct symbol **parent) @@ -297,7 +298,7 @@ int machine__resolve_callchain(struct machine *self, struct perf_evsel *evsel, unsigned int i; int err; - callchain_cursor_reset(&evsel->hists.callchain_cursor); + callchain_cursor_reset(&callchain_cursor); for (i = 0; i < chain->nr; i++) { u64 ip; @@ -333,7 +334,7 @@ int machine__resolve_callchain(struct machine *self, struct perf_evsel *evsel, break; } - err = callchain_cursor_append(&evsel->hists.callchain_cursor, + err = callchain_cursor_append(&callchain_cursor, ip, al.map, al.sym); if (err) return err; @@ -1428,7 +1429,6 @@ void perf_event__print_ip(union perf_event *event, struct perf_sample *sample, int print_sym, int print_dso, int print_symoffset) { struct addr_location al; - struct callchain_cursor *cursor = &evsel->hists.callchain_cursor; struct callchain_cursor_node *node; if (perf_event__preprocess_sample(event, machine, &al, sample, @@ -1446,10 +1446,10 @@ void perf_event__print_ip(union perf_event *event, struct perf_sample *sample, error("Failed to resolve callchain. Skipping\n"); return; } - callchain_cursor_commit(cursor); + callchain_cursor_commit(&callchain_cursor); while (1) { - node = callchain_cursor_current(cursor); + node = callchain_cursor_current(&callchain_cursor); if (!node) break; @@ -1465,7 +1465,7 @@ void perf_event__print_ip(union perf_event *event, struct perf_sample *sample, } printf("\n"); - callchain_cursor_advance(cursor); + callchain_cursor_advance(&callchain_cursor); } } else { -- cgit From 114067b69e7b2c691faace0e33db2f04096f668d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:43:27 +0900 Subject: perf tools: Check if callchain is corrupted We faced segmentation fault on perf top -G at very high sampling rate due to a corrupted callchain. While the root cause was not revealed (I failed to figure it out), this patch tries to protect us from the segfault on such cases. Reported-by: Arnaldo Carvalho de Melo Signed-off-by: Namhyung Kim Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sunjin Yang Link: http://lkml.kernel.org/r/1338443007-24857-2-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- include/linux/perf_event.h | 4 ++-- tools/perf/util/session.c | 14 +++++++++++++- 2 files changed, 15 insertions(+), 3 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index f32578634d9d..1817d4015e5f 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -555,6 +555,8 @@ enum perf_event_type { PERF_RECORD_MAX, /* non-ABI */ }; +#define PERF_MAX_STACK_DEPTH 255 + enum perf_callchain_context { PERF_CONTEXT_HV = (__u64)-32, PERF_CONTEXT_KERNEL = (__u64)-128, @@ -609,8 +611,6 @@ struct perf_guest_info_callbacks { #include #include -#define PERF_MAX_STACK_DEPTH 255 - struct perf_callchain_entry { __u64 nr; __u64 ip[PERF_MAX_STACK_DEPTH]; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 3b6f8e460a31..04d1e33f4592 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -300,6 +300,11 @@ int machine__resolve_callchain(struct machine *self, callchain_cursor_reset(&callchain_cursor); + if (chain->nr > PERF_MAX_STACK_DEPTH) { + pr_warning("corrupted callchain. skipping...\n"); + return 0; + } + for (i = 0; i < chain->nr; i++) { u64 ip; struct addr_location al; @@ -318,7 +323,14 @@ int machine__resolve_callchain(struct machine *self, case PERF_CONTEXT_USER: cpumode = PERF_RECORD_MISC_USER; break; default: - break; + pr_debug("invalid callchain context: " + "%"PRId64"\n", (s64) ip); + /* + * It seems the callchain is corrupted. + * Discard all. + */ + callchain_cursor_reset(&callchain_cursor); + return 0; } continue; } -- cgit From aa5cdd308ddbb08959534be408eba8ef040b5989 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 31 May 2012 11:17:34 -0300 Subject: perf tools: Make --version show kernel version instead of pull req tag Before: $ perf --version perf version perf.urgent.for.mingo.5.g37da28 After: $ perf --version perf version 3.4.8941.g37da28.dirty Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-vc9b4e6023iegz9kabr3yvyv@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/PERF-VERSION-GEN | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/perf/util/PERF-VERSION-GEN b/tools/perf/util/PERF-VERSION-GEN index ad73300f7bac..95264f304179 100755 --- a/tools/perf/util/PERF-VERSION-GEN +++ b/tools/perf/util/PERF-VERSION-GEN @@ -12,7 +12,7 @@ LF=' # First check if there is a .git to get the version from git describe # otherwise try to get the version from the kernel makefile if test -d ../../.git -o -f ../../.git && - VN=$(git describe --abbrev=4 HEAD 2>/dev/null) && + VN=$(git describe --match 'v[0-9].[0-9]*' --abbrev=4 HEAD 2>/dev/null) && case "$VN" in *$LF*) (exit 1) ;; v[0-9]*) -- cgit From a59e64a13a927fb7530bef39e9f5e7de8268137e Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:51:45 +0900 Subject: perf tools: Update ioctl documentation for PERF_IOC_FLAG_GROUP The ioctl interface of perf event fd receives 3 arguments to control event group behavior but it lacked documentation. Signed-off-by: Namhyung Kim Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338443506-25009-2-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/design.txt | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/tools/perf/design.txt b/tools/perf/design.txt index bd0bb1b1279b..67e5d0cace85 100644 --- a/tools/perf/design.txt +++ b/tools/perf/design.txt @@ -409,14 +409,15 @@ Counters can be enabled and disabled in two ways: via ioctl and via prctl. When a counter is disabled, it doesn't count or generate events but does continue to exist and maintain its count value. -An individual counter or counter group can be enabled with +An individual counter can be enabled with - ioctl(fd, PERF_EVENT_IOC_ENABLE); + ioctl(fd, PERF_EVENT_IOC_ENABLE, 0); or disabled with - ioctl(fd, PERF_EVENT_IOC_DISABLE); + ioctl(fd, PERF_EVENT_IOC_DISABLE, 0); +For a counter group, pass PERF_IOC_FLAG_GROUP as the third argument. Enabling or disabling the leader of a group enables or disables the whole group; that is, while the group leader is disabled, none of the counters in the group will count. Enabling or disabling a member of a -- cgit From 55da80059de6c7533724fcd95f16c5d5618ecf4d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:51:46 +0900 Subject: perf evlist: Pass third argument to ioctl explicitly The ioctl on perf event fd wants 3 arguments but we only passed 2. As the only user of the functions is perf record and it calls them for every event (regardless of group setting), just pass 0 for now. Signed-off-by: Namhyung Kim Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338443506-25009-3-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evlist.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index ed277e5627cf..7400fb3fc50c 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -274,7 +274,8 @@ void perf_evlist__disable(struct perf_evlist *evlist) for (cpu = 0; cpu < evlist->cpus->nr; cpu++) { list_for_each_entry(pos, &evlist->entries, node) { for (thread = 0; thread < evlist->threads->nr; thread++) - ioctl(FD(pos, cpu, thread), PERF_EVENT_IOC_DISABLE); + ioctl(FD(pos, cpu, thread), + PERF_EVENT_IOC_DISABLE, 0); } } } @@ -287,7 +288,8 @@ void perf_evlist__enable(struct perf_evlist *evlist) for (cpu = 0; cpu < evlist->cpus->nr; cpu++) { list_for_each_entry(pos, &evlist->entries, node) { for (thread = 0; thread < evlist->threads->nr; thread++) - ioctl(FD(pos, cpu, thread), PERF_EVENT_IOC_ENABLE); + ioctl(FD(pos, cpu, thread), + PERF_EVENT_IOC_ENABLE, 0); } } } -- cgit From 8db4841fc72acc58254029f050226ea5f8103854 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 30 May 2012 14:23:42 +0200 Subject: perf symbols: Handle different endians properly during symbol load Currently we dont care about the file object's endianness. It's possible we read buildid file object from different architecture than we are currentlly running on. So we need to care about properly reading such object's data - handle different endianness properly. Adding: needs_swap DSO field dso__swap_init function to initialize DSO's needs_swap DSO__SWAP to read the data with proper swaps Together with other endianity patches, this change fixies perf report discrepancies on origin and target systems as described in test 1 below, e.g. following perf report diff: ... 0.12% ps [kernel.kallsyms] [k] clear_page - 0.12% awk bash [.] alloc_word_desc + 0.12% awk bash [.] yyparse 0.11% beah-rhts-task libpython2.6.so.1.0 [.] 0x5560e 0.10% perf libc-2.12.so [.] __ctype_toupper_loc - 0.09% rhts-test-runne bash [.] maybe_make_export_env + 0.09% rhts-test-runne bash [.] 0x385a0 0.09% ps [kernel.kallsyms] [k] page_fault ... Note, running following to test perf endianity handling: test 1) - origin system: # perf record -a -- sleep 10 (any perf record will do) # perf report > report.origin # perf archive perf.data - copy the perf.data, report.origin and perf.data.tar.bz2 to a target system and run: # tar xjvf perf.data.tar.bz2 -C ~/.debug # perf report > report.target # diff -u report.origin report.target - the diff should produce no output (besides some white space stuff and possibly different date/TZ output) test 1) - origin system: # perf record -ag -fo /tmp/perf.data -- sleep 1 - mount origin system root to the target system on /mnt/origin - target system: # perf script --symfs /mnt/origin -I -i /mnt/origin/tmp/perf.data \ --kallsyms /mnt/origin/proc/kallsyms - complete perf.data header is displayed Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338380624-7443-2-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 33 ++++++++++++++++++++++++++++++++- tools/perf/util/symbol.h | 30 ++++++++++++++++++++++++++++++ 2 files changed, 62 insertions(+), 1 deletion(-) diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index e2ba8858f3e1..9d04dcd91815 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -323,6 +323,7 @@ struct dso *dso__new(const char *name) dso->sorted_by_name = 0; dso->has_build_id = 0; dso->kernel = DSO_TYPE_USER; + dso->needs_swap = DSO_SWAP__UNSET; INIT_LIST_HEAD(&dso->node); } @@ -1156,6 +1157,33 @@ static size_t elf_addr_to_index(Elf *elf, GElf_Addr addr) return -1; } +static int dso__swap_init(struct dso *dso, unsigned char eidata) +{ + static unsigned int const endian = 1; + + dso->needs_swap = DSO_SWAP__NO; + + switch (eidata) { + case ELFDATA2LSB: + /* We are big endian, DSO is little endian. */ + if (*(unsigned char const *)&endian != 1) + dso->needs_swap = DSO_SWAP__YES; + break; + + case ELFDATA2MSB: + /* We are little endian, DSO is big endian. */ + if (*(unsigned char const *)&endian != 0) + dso->needs_swap = DSO_SWAP__YES; + break; + + default: + pr_err("unrecognized DSO data encoding %d\n", eidata); + return -EINVAL; + } + + return 0; +} + static int dso__load_sym(struct dso *dso, struct map *map, const char *name, int fd, symbol_filter_t filter, int kmodule, int want_symtab) @@ -1187,6 +1215,9 @@ static int dso__load_sym(struct dso *dso, struct map *map, const char *name, goto out_elf_end; } + if (dso__swap_init(dso, ehdr.e_ident[EI_DATA])) + goto out_elf_end; + /* Always reject images with a mismatched build-id: */ if (dso->has_build_id) { u8 build_id[BUILD_ID_SIZE]; @@ -1272,7 +1303,7 @@ static int dso__load_sym(struct dso *dso, struct map *map, const char *name, if (opdsec && sym.st_shndx == opdidx) { u32 offset = sym.st_value - opdshdr.sh_addr; u64 *opd = opddata->d_buf + offset; - sym.st_value = *opd; + sym.st_value = DSO__SWAP(dso, u64, *opd); sym.st_shndx = elf_addr_to_index(elf, sym.st_value); } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 5649d63798cb..af0752b1aca1 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -9,6 +9,7 @@ #include #include #include +#include #ifdef HAVE_CPLUS_DEMANGLE extern char *cplus_demangle(const char *, int); @@ -160,11 +161,18 @@ enum dso_kernel_type { DSO_TYPE_GUEST_KERNEL }; +enum dso_swap_type { + DSO_SWAP__UNSET, + DSO_SWAP__NO, + DSO_SWAP__YES, +}; + struct dso { struct list_head node; struct rb_root symbols[MAP__NR_TYPES]; struct rb_root symbol_names[MAP__NR_TYPES]; enum dso_kernel_type kernel; + enum dso_swap_type needs_swap; u8 adjust_symbols:1; u8 has_build_id:1; u8 hit:1; @@ -182,6 +190,28 @@ struct dso { char name[0]; }; +#define DSO__SWAP(dso, type, val) \ +({ \ + type ____r = val; \ + BUG_ON(dso->needs_swap == DSO_SWAP__UNSET); \ + if (dso->needs_swap == DSO_SWAP__YES) { \ + switch (sizeof(____r)) { \ + case 2: \ + ____r = bswap_16(val); \ + break; \ + case 4: \ + ____r = bswap_32(val); \ + break; \ + case 8: \ + ____r = bswap_64(val); \ + break; \ + default: \ + BUG_ON(1); \ + } \ + } \ + ____r; \ +}) + struct dso *dso__new(const char *name); void dso__delete(struct dso *dso); -- cgit From 268fb20f832e1eb4afd5113ee31fef9332986b13 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 30 May 2012 14:23:43 +0200 Subject: perf session: Handle endianity swap on sample_id_all header data Adding endianity swapping for event header attached via sample_id_all. Currently we dont do that and it's causing wrong data to be read when running report on architecture with different endianity than the record. The perf is currently able to process 32-bit PPC samples on 32-bit and 64-bit x86. Together with other endianity patches, this change fixies perf report discrepancies on origin and target systems as described in test 1 below, e.g. following perf report diff: ... 0.12% ps [kernel.kallsyms] [k] clear_page - 0.12% awk bash [.] alloc_word_desc + 0.12% awk bash [.] yyparse 0.11% beah-rhts-task libpython2.6.so.1.0 [.] 0x5560e 0.10% perf libc-2.12.so [.] __ctype_toupper_loc - 0.09% rhts-test-runne bash [.] maybe_make_export_env + 0.09% rhts-test-runne bash [.] 0x385a0 0.09% ps [kernel.kallsyms] [k] page_fault ... Note, running following to test perf endianity handling: test 1) - origin system: # perf record -a -- sleep 10 (any perf record will do) # perf report > report.origin # perf archive perf.data - copy the perf.data, report.origin and perf.data.tar.bz2 to a target system and run: # tar xjvf perf.data.tar.bz2 -C ~/.debug # perf report > report.target # diff -u report.origin report.target - the diff should produce no output (besides some white space stuff and possibly different date/TZ output) test 2) - origin system: # perf record -ag -fo /tmp/perf.data -- sleep 1 - mount origin system root to the target system on /mnt/origin - target system: # perf script --symfs /mnt/origin -I -i /mnt/origin/tmp/perf.data \ --kallsyms /mnt/origin/proc/kallsyms - complete perf.data header is displayed Signed-off-by: Jiri Olsa Reviewed-by: David Ahern Tested-by: David Ahern Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338380624-7443-3-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 67 ++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 55 insertions(+), 12 deletions(-) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 04d1e33f4592..2600916efa83 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -454,37 +454,65 @@ void mem_bswap_64(void *src, int byte_size) } } -static void perf_event__all64_swap(union perf_event *event) +static void swap_sample_id_all(union perf_event *event, void *data) +{ + void *end = (void *) event + event->header.size; + int size = end - data; + + BUG_ON(size % sizeof(u64)); + mem_bswap_64(data, size); +} + +static void perf_event__all64_swap(union perf_event *event, + bool sample_id_all __used) { struct perf_event_header *hdr = &event->header; mem_bswap_64(hdr + 1, event->header.size - sizeof(*hdr)); } -static void perf_event__comm_swap(union perf_event *event) +static void perf_event__comm_swap(union perf_event *event, bool sample_id_all) { event->comm.pid = bswap_32(event->comm.pid); event->comm.tid = bswap_32(event->comm.tid); + + if (sample_id_all) { + void *data = &event->comm.comm; + + data += ALIGN(strlen(data) + 1, sizeof(u64)); + swap_sample_id_all(event, data); + } } -static void perf_event__mmap_swap(union perf_event *event) +static void perf_event__mmap_swap(union perf_event *event, + bool sample_id_all) { event->mmap.pid = bswap_32(event->mmap.pid); event->mmap.tid = bswap_32(event->mmap.tid); event->mmap.start = bswap_64(event->mmap.start); event->mmap.len = bswap_64(event->mmap.len); event->mmap.pgoff = bswap_64(event->mmap.pgoff); + + if (sample_id_all) { + void *data = &event->mmap.filename; + + data += ALIGN(strlen(data) + 1, sizeof(u64)); + swap_sample_id_all(event, data); + } } -static void perf_event__task_swap(union perf_event *event) +static void perf_event__task_swap(union perf_event *event, bool sample_id_all) { event->fork.pid = bswap_32(event->fork.pid); event->fork.tid = bswap_32(event->fork.tid); event->fork.ppid = bswap_32(event->fork.ppid); event->fork.ptid = bswap_32(event->fork.ptid); event->fork.time = bswap_64(event->fork.time); + + if (sample_id_all) + swap_sample_id_all(event, &event->fork + 1); } -static void perf_event__read_swap(union perf_event *event) +static void perf_event__read_swap(union perf_event *event, bool sample_id_all) { event->read.pid = bswap_32(event->read.pid); event->read.tid = bswap_32(event->read.tid); @@ -492,6 +520,9 @@ static void perf_event__read_swap(union perf_event *event) event->read.time_enabled = bswap_64(event->read.time_enabled); event->read.time_running = bswap_64(event->read.time_running); event->read.id = bswap_64(event->read.id); + + if (sample_id_all) + swap_sample_id_all(event, &event->read + 1); } static u8 revbyte(u8 b) @@ -543,7 +574,8 @@ void perf_event__attr_swap(struct perf_event_attr *attr) swap_bitfield((u8 *) (&attr->read_format + 1), sizeof(u64)); } -static void perf_event__hdr_attr_swap(union perf_event *event) +static void perf_event__hdr_attr_swap(union perf_event *event, + bool sample_id_all __used) { size_t size; @@ -554,18 +586,21 @@ static void perf_event__hdr_attr_swap(union perf_event *event) mem_bswap_64(event->attr.id, size); } -static void perf_event__event_type_swap(union perf_event *event) +static void perf_event__event_type_swap(union perf_event *event, + bool sample_id_all __used) { event->event_type.event_type.event_id = bswap_64(event->event_type.event_type.event_id); } -static void perf_event__tracing_data_swap(union perf_event *event) +static void perf_event__tracing_data_swap(union perf_event *event, + bool sample_id_all __used) { event->tracing_data.size = bswap_32(event->tracing_data.size); } -typedef void (*perf_event__swap_op)(union perf_event *event); +typedef void (*perf_event__swap_op)(union perf_event *event, + bool sample_id_all); static perf_event__swap_op perf_event__swap_ops[] = { [PERF_RECORD_MMAP] = perf_event__mmap_swap, @@ -999,6 +1034,15 @@ static int perf_session__process_user_event(struct perf_session *session, union } } +static void event_swap(union perf_event *event, bool sample_id_all) +{ + perf_event__swap_op swap; + + swap = perf_event__swap_ops[event->header.type]; + if (swap) + swap(event, sample_id_all); +} + static int perf_session__process_event(struct perf_session *session, union perf_event *event, struct perf_tool *tool, @@ -1007,9 +1051,8 @@ static int perf_session__process_event(struct perf_session *session, struct perf_sample sample; int ret; - if (session->header.needs_swap && - perf_event__swap_ops[event->header.type]) - perf_event__swap_ops[event->header.type](event); + if (session->header.needs_swap) + event_swap(event, session->sample_id_all); if (event->header.type >= PERF_RECORD_HEADER_MAX) return -EINVAL; -- cgit From 37073f9e449bc430e6c40b9cffc2558002a0256a Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 30 May 2012 14:23:44 +0200 Subject: perf evsel: Fix 32 bit values endianity swap for sample_id_all header We swap the sample_id_all header by u64 pointers. Some members of the header happen to be 32 bit values. We need to handle them separatelly. Together with other endianity patches, this change fixies perf report discrepancies on origin and target systems as described in test 1 below, e.g. following perf report diff: ... 0.12% ps [kernel.kallsyms] [k] clear_page - 0.12% awk bash [.] alloc_word_desc + 0.12% awk bash [.] yyparse 0.11% beah-rhts-task libpython2.6.so.1.0 [.] 0x5560e 0.10% perf libc-2.12.so [.] __ctype_toupper_loc - 0.09% rhts-test-runne bash [.] maybe_make_export_env + 0.09% rhts-test-runne bash [.] 0x385a0 0.09% ps [kernel.kallsyms] [k] page_fault ... Note, running following to test perf endianity handling: test 1) - origin system: # perf record -a -- sleep 10 (any perf record will do) # perf report > report.origin # perf archive perf.data - copy the perf.data, report.origin and perf.data.tar.bz2 to a target system and run: # tar xjvf perf.data.tar.bz2 -C ~/.debug # perf report > report.target # diff -u report.origin report.target - the diff should produce no output (besides some white space stuff and possibly different date/TZ output) test 2) - origin system: # perf record -ag -fo /tmp/perf.data -- sleep 1 - mount origin system root to the target system on /mnt/origin - target system: # perf script --symfs /mnt/origin -I -i /mnt/origin/tmp/perf.data \ --kallsyms /mnt/origin/proc/kallsyms - complete perf.data header is displayed Signed-off-by: Jiri Olsa Reviewed-by: David Ahern Tested-by: David Ahern Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338380624-7443-4-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evsel.c | 29 ++++++++++++++++++++++------- 1 file changed, 22 insertions(+), 7 deletions(-) diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 91d19138f3ec..9f6cebd798ee 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -494,16 +494,24 @@ int perf_evsel__open_per_thread(struct perf_evsel *evsel, } static int perf_event__parse_id_sample(const union perf_event *event, u64 type, - struct perf_sample *sample) + struct perf_sample *sample, + bool swapped) { const u64 *array = event->sample.array; + union u64_swap u; array += ((event->header.size - sizeof(event->header)) / sizeof(u64)) - 1; if (type & PERF_SAMPLE_CPU) { - u32 *p = (u32 *)array; - sample->cpu = *p; + u.val64 = *array; + if (swapped) { + /* undo swap of u64, then swap on individual u32s */ + u.val64 = bswap_64(u.val64); + u.val32[0] = bswap_32(u.val32[0]); + } + + sample->cpu = u.val32[0]; array--; } @@ -523,9 +531,16 @@ static int perf_event__parse_id_sample(const union perf_event *event, u64 type, } if (type & PERF_SAMPLE_TID) { - u32 *p = (u32 *)array; - sample->pid = p[0]; - sample->tid = p[1]; + u.val64 = *array; + if (swapped) { + /* undo swap of u64, then swap on individual u32s */ + u.val64 = bswap_64(u.val64); + u.val32[0] = bswap_32(u.val32[0]); + u.val32[1] = bswap_32(u.val32[1]); + } + + sample->pid = u.val32[0]; + sample->tid = u.val32[1]; } return 0; @@ -562,7 +577,7 @@ int perf_event__parse_sample(const union perf_event *event, u64 type, if (event->header.type != PERF_RECORD_SAMPLE) { if (!sample_id_all) return 0; - return perf_event__parse_id_sample(event, type, data); + return perf_event__parse_id_sample(event, type, data, swapped); } array = event->sample.array; -- cgit From 378474e4b23183002f1791b294a4440b6b7df36a Mon Sep 17 00:00:00 2001 From: Srikar Dronamraju Date: Thu, 31 May 2012 17:16:56 +0530 Subject: perf symbols: Check for valid dso before creating map dso__new() can return NULL. Hence verify dso before creating a new map. Signed-off-by: Srikar Dronamraju Suggested-by: Arnaldo Carvalho de Melo Cc: Ananth N Mavinakayanahalli Cc: Anton Arapov Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Oleg Nesterov Cc: Peter Zijlstra Cc: Steven Rostedt Link: http://lkml.kernel.org/r/20120531114656.23691.54223.sendpatchset@srdronam.in.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 9d04dcd91815..3e2e5ea0f03f 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -2817,8 +2817,11 @@ int machine__load_vmlinux_path(struct machine *machine, enum map_type type, struct map *dso__new_map(const char *name) { + struct map *map = NULL; struct dso *dso = dso__new(name); - struct map *map = map__new2(0, dso, MAP__FUNCTION); + + if (dso) + map = map__new2(0, dso, MAP__FUNCTION); return map; } -- cgit From a23c4dc422cff7bd30f40f5dc7c9ac4a6339005a Mon Sep 17 00:00:00 2001 From: Srikar Dronamraju Date: Thu, 31 May 2012 17:16:43 +0530 Subject: perf uprobes: Remove unnecessary check before strlist__delete Since strlist__delete() itself checks, the additional check before calling strlist__delete() is redundant. No Functional change. Signed-off-by: Srikar Dronamraju Suggested-by: Arnaldo Carvalho de Melo Cc: Ananth N Mavinakayanahalli Cc: Anton Arapov Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Oleg Nesterov Cc: Peter Zijlstra Cc: Steven Rostedt Link: http://lkml.kernel.org/r/20120531114643.23691.38666.sendpatchset@srdronam.in.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-event.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 59dccc98b554..0dda25d82d06 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -2164,16 +2164,12 @@ int del_perf_probe_events(struct strlist *dellist) error: if (kfd >= 0) { - if (namelist) - strlist__delete(namelist); - + strlist__delete(namelist); close(kfd); } if (ufd >= 0) { - if (unamelist) - strlist__delete(unamelist); - + strlist__delete(unamelist); close(ufd); } -- cgit From cb7225feec627e91d598198996429e9ee6804f8d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:51:44 +0900 Subject: perf: Remove duplicate invocation on perf_event_for_each The @func callback was invoked twice for group leader when perf_event_for_each() called. It seems the commit 75f937f24bd9 ("perf_counter: Fix ctx->mutex vs counter ->mutex inversion") made the mistake during the change. Signed-off-by: Namhyung Kim Acked-by: Peter Zijlstra Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338443506-25009-1-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index 5b06cbbf6931..f85c0154b333 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3181,7 +3181,6 @@ static void perf_event_for_each(struct perf_event *event, event = event->group_leader; perf_event_for_each_child(event, func); - func(event); list_for_each_entry(sibling, &event->sibling_list, group_entry) perf_event_for_each_child(sibling, func); mutex_unlock(&ctx->mutex); -- cgit From 436d03faf6961b30e13b2d0967aea9d772d6cf44 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jun 2012 00:09:11 +0900 Subject: x86/decoder: Fix bsr/bsf/jmpe decoding with operand-size prefix Fix the x86 instruction decoder to decode bsr/bsf/jmpe with operand-size prefix (66h). This fixes the test case failure reported by Linus, attached below. bsf/bsr/jmpe have a special encoding. Opcode map in Intel Software Developers Manual vol2 says they have TZCNT/LZCNT variants if it has F3h prefix. However, there is no information if it has other 66h or F2h prefixes. Current instruction decoder supposes that those are bad instructions, but it actually accepts at least operand-size prefixes. H. Peter Anvin further explains: " TZCNT/LZCNT are F3 + BSF/BSR exactly because the F2 and F3 prefixes have historically been no-ops with most instructions. This allows software to unconditionally use the prefixed versions and get TZCNT/LZCNT on the processors that have them if they don't care about the difference. " This fixes errors reported by test_get_len: Warning: arch/x86/tools/test_get_len found difference at :ffffffff81036d87 Warning: ffffffff81036de5: 66 0f bc c2 bsf %dx,%ax Warning: objdump says 4 bytes, but insn_get_length() says 3 Warning: arch/x86/tools/test_get_len found difference at :ffffffff81036ea6 Warning: ffffffff81036f04: 66 0f bd c2 bsr %dx,%ax Warning: objdump says 4 bytes, but insn_get_length() says 3 Warning: decoded and checked 13298882 instructions with 2 warnings Reported-by: Linus Torvalds Reported-by: Pekka Enberg Signed-off-by: Masami Hiramatsu Cc: "H. Peter Anvin" Cc: Link: http://lkml.kernel.org/r/20120604150911.22338.43296.stgit@localhost.localdomain Signed-off-by: Ingo Molnar --- arch/x86/lib/x86-opcode-map.txt | 8 ++++---- arch/x86/tools/gen-insn-attr-x86.awk | 14 +++++++++----- 2 files changed, 13 insertions(+), 9 deletions(-) diff --git a/arch/x86/lib/x86-opcode-map.txt b/arch/x86/lib/x86-opcode-map.txt index 819137904428..5d7e51f3fd28 100644 --- a/arch/x86/lib/x86-opcode-map.txt +++ b/arch/x86/lib/x86-opcode-map.txt @@ -28,7 +28,7 @@ # - (66): the last prefix is 0x66 # - (F3): the last prefix is 0xF3 # - (F2): the last prefix is 0xF2 -# +# - (!F3) : the last prefix is not 0xF3 (including non-last prefix case) Table: one byte opcode Referrer: @@ -515,12 +515,12 @@ b4: LFS Gv,Mp b5: LGS Gv,Mp b6: MOVZX Gv,Eb b7: MOVZX Gv,Ew -b8: JMPE | POPCNT Gv,Ev (F3) +b8: JMPE (!F3) | POPCNT Gv,Ev (F3) b9: Grp10 (1A) ba: Grp8 Ev,Ib (1A) bb: BTC Ev,Gv -bc: BSF Gv,Ev | TZCNT Gv,Ev (F3) -bd: BSR Gv,Ev | LZCNT Gv,Ev (F3) +bc: BSF Gv,Ev (!F3) | TZCNT Gv,Ev (F3) +bd: BSR Gv,Ev (!F3) | LZCNT Gv,Ev (F3) be: MOVSX Gv,Eb bf: MOVSX Gv,Ew # 0x0f 0xc0-0xcf diff --git a/arch/x86/tools/gen-insn-attr-x86.awk b/arch/x86/tools/gen-insn-attr-x86.awk index 5f6a5b6c3a15..ddcf39b1a18d 100644 --- a/arch/x86/tools/gen-insn-attr-x86.awk +++ b/arch/x86/tools/gen-insn-attr-x86.awk @@ -66,9 +66,10 @@ BEGIN { rex_expr = "^REX(\\.[XRWB]+)*" fpu_expr = "^ESC" # TODO - lprefix1_expr = "\\(66\\)" + lprefix1_expr = "\\((66|!F3)\\)" lprefix2_expr = "\\(F3\\)" - lprefix3_expr = "\\(F2\\)" + lprefix3_expr = "\\((F2|!F3)\\)" + lprefix_expr = "\\((66|F2|F3)\\)" max_lprefix = 4 # All opcodes starting with lower-case 'v' or with (v1) superscript @@ -333,13 +334,16 @@ function convert_operands(count,opnd, i,j,imm,mod) if (match(ext, lprefix1_expr)) { lptable1[idx] = add_flags(lptable1[idx],flags) variant = "INAT_VARIANT" - } else if (match(ext, lprefix2_expr)) { + } + if (match(ext, lprefix2_expr)) { lptable2[idx] = add_flags(lptable2[idx],flags) variant = "INAT_VARIANT" - } else if (match(ext, lprefix3_expr)) { + } + if (match(ext, lprefix3_expr)) { lptable3[idx] = add_flags(lptable3[idx],flags) variant = "INAT_VARIANT" - } else { + } + if (!match(ext, lprefix_expr)){ table[idx] = add_flags(table[idx],flags) } } -- cgit From b430f7c4706aeba4270c7ab7744fc504b9315e1c Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 5 Jun 2012 15:30:31 +0200 Subject: perf/x86: Fix Intel shared extra MSR allocation Zheng Yan reported that event group validation can wreck event state when Intel extra_reg allocation changes event state. Validation shouldn't change any persistent state. Cloning events in validate_{event,group}() isn't really pretty either, so add a few special cases to avoid modifying the event state. The code is restructured to minimize the special case impact. Reported-by: Zheng Yan Acked-by: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1338903031.28282.175.camel@twins Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 1 + arch/x86/kernel/cpu/perf_event.h | 1 + arch/x86/kernel/cpu/perf_event_intel.c | 92 +++++++++++++++++++++++----------- 3 files changed, 66 insertions(+), 28 deletions(-) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index e049d6da0183..cb608383e4f6 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1496,6 +1496,7 @@ static struct cpu_hw_events *allocate_fake_cpuc(void) if (!cpuc->shared_regs) goto error; } + cpuc->is_fake = 1; return cpuc; error: free_fake_cpuc(cpuc); diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h index 6638aaf54493..83794d8e6af0 100644 --- a/arch/x86/kernel/cpu/perf_event.h +++ b/arch/x86/kernel/cpu/perf_event.h @@ -117,6 +117,7 @@ struct cpu_hw_events { struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ unsigned int group_flag; + int is_fake; /* * Intel DebugStore bits diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 166546ec6aef..965baa2fa790 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1119,27 +1119,33 @@ intel_bts_constraints(struct perf_event *event) return NULL; } -static bool intel_try_alt_er(struct perf_event *event, int orig_idx) +static int intel_alt_er(int idx) { if (!(x86_pmu.er_flags & ERF_HAS_RSP_1)) - return false; + return idx; - if (event->hw.extra_reg.idx == EXTRA_REG_RSP_0) { - event->hw.config &= ~INTEL_ARCH_EVENT_MASK; - event->hw.config |= 0x01bb; - event->hw.extra_reg.idx = EXTRA_REG_RSP_1; - event->hw.extra_reg.reg = MSR_OFFCORE_RSP_1; - } else if (event->hw.extra_reg.idx == EXTRA_REG_RSP_1) { + if (idx == EXTRA_REG_RSP_0) + return EXTRA_REG_RSP_1; + + if (idx == EXTRA_REG_RSP_1) + return EXTRA_REG_RSP_0; + + return idx; +} + +static void intel_fixup_er(struct perf_event *event, int idx) +{ + event->hw.extra_reg.idx = idx; + + if (idx == EXTRA_REG_RSP_0) { event->hw.config &= ~INTEL_ARCH_EVENT_MASK; event->hw.config |= 0x01b7; - event->hw.extra_reg.idx = EXTRA_REG_RSP_0; event->hw.extra_reg.reg = MSR_OFFCORE_RSP_0; + } else if (idx == EXTRA_REG_RSP_1) { + event->hw.config &= ~INTEL_ARCH_EVENT_MASK; + event->hw.config |= 0x01bb; + event->hw.extra_reg.reg = MSR_OFFCORE_RSP_1; } - - if (event->hw.extra_reg.idx == orig_idx) - return false; - - return true; } /* @@ -1157,14 +1163,18 @@ __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, struct event_constraint *c = &emptyconstraint; struct er_account *era; unsigned long flags; - int orig_idx = reg->idx; + int idx = reg->idx; - /* already allocated shared msr */ - if (reg->alloc) + /* + * reg->alloc can be set due to existing state, so for fake cpuc we + * need to ignore this, otherwise we might fail to allocate proper fake + * state for this extra reg constraint. Also see the comment below. + */ + if (reg->alloc && !cpuc->is_fake) return NULL; /* call x86_get_event_constraint() */ again: - era = &cpuc->shared_regs->regs[reg->idx]; + era = &cpuc->shared_regs->regs[idx]; /* * we use spin_lock_irqsave() to avoid lockdep issues when * passing a fake cpuc @@ -1173,6 +1183,29 @@ again: if (!atomic_read(&era->ref) || era->config == reg->config) { + /* + * If its a fake cpuc -- as per validate_{group,event}() we + * shouldn't touch event state and we can avoid doing so + * since both will only call get_event_constraints() once + * on each event, this avoids the need for reg->alloc. + * + * Not doing the ER fixup will only result in era->reg being + * wrong, but since we won't actually try and program hardware + * this isn't a problem either. + */ + if (!cpuc->is_fake) { + if (idx != reg->idx) + intel_fixup_er(event, idx); + + /* + * x86_schedule_events() can call get_event_constraints() + * multiple times on events in the case of incremental + * scheduling(). reg->alloc ensures we only do the ER + * allocation once. + */ + reg->alloc = 1; + } + /* lock in msr value */ era->config = reg->config; era->reg = reg->reg; @@ -1180,17 +1213,17 @@ again: /* one more user */ atomic_inc(&era->ref); - /* no need to reallocate during incremental event scheduling */ - reg->alloc = 1; - /* * need to call x86_get_event_constraint() * to check if associated event has constraints */ c = NULL; - } else if (intel_try_alt_er(event, orig_idx)) { - raw_spin_unlock_irqrestore(&era->lock, flags); - goto again; + } else { + idx = intel_alt_er(idx); + if (idx != reg->idx) { + raw_spin_unlock_irqrestore(&era->lock, flags); + goto again; + } } raw_spin_unlock_irqrestore(&era->lock, flags); @@ -1204,11 +1237,14 @@ __intel_shared_reg_put_constraints(struct cpu_hw_events *cpuc, struct er_account *era; /* - * only put constraint if extra reg was actually - * allocated. Also takes care of event which do - * not use an extra shared reg + * Only put constraint if extra reg was actually allocated. Also takes + * care of event which do not use an extra shared reg. + * + * Also, if this is a fake cpuc we shouldn't touch any event state + * (reg->alloc) and we don't care about leaving inconsistent cpuc state + * either since it'll be thrown out. */ - if (!reg->alloc) + if (!reg->alloc || cpuc->is_fake) return; era = &cpuc->shared_regs->regs[reg->idx]; -- cgit From cccb9ba9e4ee0d750265f53de9258df69655c40b Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 5 Jun 2012 10:26:43 +0200 Subject: perf/x86: Implement cycles:p for SNB/IVB Now that there's finally a chip with working PEBS (IvyBridge), we can enable the hardware and implement cycles:p for SNB/IVB. Cc: Stephane Eranian Requested-and-tested-by: Linus Torvalds Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1338884803.28282.153.camel@twins Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.h | 1 + arch/x86/kernel/cpu/perf_event_intel.c | 50 ++++++++++++++++++++++++++++------ 2 files changed, 43 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h index 83794d8e6af0..7241e2fc3c17 100644 --- a/arch/x86/kernel/cpu/perf_event.h +++ b/arch/x86/kernel/cpu/perf_event.h @@ -365,6 +365,7 @@ struct x86_pmu { int pebs_record_size; void (*drain_pebs)(struct pt_regs *regs); struct event_constraint *pebs_constraints; + void (*pebs_aliases)(struct perf_event *event); /* * Intel LBR diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 965baa2fa790..2312c1ff1b19 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1336,15 +1336,9 @@ static void intel_put_event_constraints(struct cpu_hw_events *cpuc, intel_put_shared_regs_event_constraints(cpuc, event); } -static int intel_pmu_hw_config(struct perf_event *event) +static void intel_pebs_aliases_core2(struct perf_event *event) { - int ret = x86_pmu_hw_config(event); - - if (ret) - return ret; - - if (event->attr.precise_ip && - (event->hw.config & X86_RAW_EVENT_MASK) == 0x003c) { + if ((event->hw.config & X86_RAW_EVENT_MASK) == 0x003c) { /* * Use an alternative encoding for CPU_CLK_UNHALTED.THREAD_P * (0x003c) so that we can use it with PEBS. @@ -1365,10 +1359,48 @@ static int intel_pmu_hw_config(struct perf_event *event) */ u64 alt_config = X86_CONFIG(.event=0xc0, .inv=1, .cmask=16); + alt_config |= (event->hw.config & ~X86_RAW_EVENT_MASK); + event->hw.config = alt_config; + } +} + +static void intel_pebs_aliases_snb(struct perf_event *event) +{ + if ((event->hw.config & X86_RAW_EVENT_MASK) == 0x003c) { + /* + * Use an alternative encoding for CPU_CLK_UNHALTED.THREAD_P + * (0x003c) so that we can use it with PEBS. + * + * The regular CPU_CLK_UNHALTED.THREAD_P event (0x003c) isn't + * PEBS capable. However we can use UOPS_RETIRED.ALL + * (0x01c2), which is a PEBS capable event, to get the same + * count. + * + * UOPS_RETIRED.ALL counts the number of cycles that retires + * CNTMASK micro-ops. By setting CNTMASK to a value (16) + * larger than the maximum number of micro-ops that can be + * retired per cycle (4) and then inverting the condition, we + * count all cycles that retire 16 or less micro-ops, which + * is every cycle. + * + * Thereby we gain a PEBS capable cycle counter. + */ + u64 alt_config = X86_CONFIG(.event=0xc2, .umask=0x01, .inv=1, .cmask=16); alt_config |= (event->hw.config & ~X86_RAW_EVENT_MASK); event->hw.config = alt_config; } +} + +static int intel_pmu_hw_config(struct perf_event *event) +{ + int ret = x86_pmu_hw_config(event); + + if (ret) + return ret; + + if (event->attr.precise_ip && x86_pmu.pebs_aliases) + x86_pmu.pebs_aliases(event); if (intel_pmu_needs_lbr_smpl(event)) { ret = intel_pmu_setup_lbr_filter(event); @@ -1643,6 +1675,7 @@ static __initconst const struct x86_pmu intel_pmu = { .max_period = (1ULL << 31) - 1, .get_event_constraints = intel_get_event_constraints, .put_event_constraints = intel_put_event_constraints, + .pebs_aliases = intel_pebs_aliases_core2, .format_attrs = intel_arch3_formats_attr, @@ -1885,6 +1918,7 @@ __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_snb_event_constraints; x86_pmu.pebs_constraints = intel_snb_pebs_event_constraints; + x86_pmu.pebs_aliases = intel_pebs_aliases_snb; x86_pmu.extra_regs = intel_snb_extra_regs; /* all extra regs are per-cpu when HT is on */ x86_pmu.er_flags |= ERF_HAS_RSP_1; -- cgit From b6db437ba8322f5cee0bd355ad2ef9f73c413754 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 5 Jun 2012 10:26:43 +0200 Subject: perf/x86: Enable/Add IvyBridge hardware support Implement rudimentary IVB perf support. The SDM states its identical to SNB with exception of the exact event tables, but a quick look suggests they're similar enough. Also mark SNB-EP as broken for now. Requested-and-tested-by: Linus Torvalds Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1338884803.28282.153.camel@twins Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event_intel.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 2312c1ff1b19..187c294bc658 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1909,8 +1909,9 @@ __init int intel_pmu_init(void) break; case 42: /* SandyBridge */ - x86_add_quirk(intel_sandybridge_quirk); case 45: /* SandyBridge, "Romely-EP" */ + x86_add_quirk(intel_sandybridge_quirk); + case 58: /* IvyBridge */ memcpy(hw_cache_event_ids, snb_hw_cache_event_ids, sizeof(hw_cache_event_ids)); -- cgit From 8440ccb43fc0ecffcf1acee0273d766e6a8cd51d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 5 Jun 2012 10:26:43 +0200 Subject: perf/x86: Update SNB PEBS constraints Afaict there's no need to (incompletely) iterate the MEM_UOPS_RETIRED.* umask state. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1338884803.28282.153.camel@twins Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event_intel_ds.c | 9 +-------- 1 file changed, 1 insertion(+), 8 deletions(-) diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c index 5a3edc27f6e5..35e2192df9f4 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -400,14 +400,7 @@ struct event_constraint intel_snb_pebs_event_constraints[] = { INTEL_EVENT_CONSTRAINT(0xc4, 0xf), /* BR_INST_RETIRED.* */ INTEL_EVENT_CONSTRAINT(0xc5, 0xf), /* BR_MISP_RETIRED.* */ INTEL_EVENT_CONSTRAINT(0xcd, 0x8), /* MEM_TRANS_RETIRED.* */ - INTEL_UEVENT_CONSTRAINT(0x11d0, 0xf), /* MEM_UOP_RETIRED.STLB_MISS_LOADS */ - INTEL_UEVENT_CONSTRAINT(0x12d0, 0xf), /* MEM_UOP_RETIRED.STLB_MISS_STORES */ - INTEL_UEVENT_CONSTRAINT(0x21d0, 0xf), /* MEM_UOP_RETIRED.LOCK_LOADS */ - INTEL_UEVENT_CONSTRAINT(0x22d0, 0xf), /* MEM_UOP_RETIRED.LOCK_STORES */ - INTEL_UEVENT_CONSTRAINT(0x41d0, 0xf), /* MEM_UOP_RETIRED.SPLIT_LOADS */ - INTEL_UEVENT_CONSTRAINT(0x42d0, 0xf), /* MEM_UOP_RETIRED.SPLIT_STORES */ - INTEL_UEVENT_CONSTRAINT(0x81d0, 0xf), /* MEM_UOP_RETIRED.ANY_LOADS */ - INTEL_UEVENT_CONSTRAINT(0x82d0, 0xf), /* MEM_UOP_RETIRED.ANY_STORES */ + INTEL_EVENT_CONSTRAINT(0xd0, 0xf), /* MEM_UOP_RETIRED.* */ INTEL_EVENT_CONSTRAINT(0xd1, 0xf), /* MEM_LOAD_UOPS_RETIRED.* */ INTEL_EVENT_CONSTRAINT(0xd2, 0xf), /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */ INTEL_UEVENT_CONSTRAINT(0x02d4, 0xf), /* MEM_LOAD_UOPS_MISC_RETIRED.LLC_MISS */ -- cgit From 302fa4b58ac754a6da13f4f5546f710fecc3b945 Mon Sep 17 00:00:00 2001 From: Arun Sharma Date: Fri, 20 Apr 2012 15:41:33 -0700 Subject: perf/x86: Allow multiple stacks Without this patch, applications with two different stack regions (eg: native stack vs JIT stack) get truncated callchains even when RBP chaining is present. GDB shows proper stack traces and the frame pointer chaining is intact. This patch disables the (fp < RSP) check, hoping that other checks in the code save the day for us. In our limited testing, this didn't seem to break anything. In the long term, we could potentially have userspace advise the kernel on the range of valid stack addresses, so we don't spend a lot of time unwinding from bogus addresses. Signed-off-by: Arun Sharma CC: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Mike Galbraith Cc: Paul Mackerras Cc: Stephane Eranian Cc: Namhyung Kim Cc: Tom Zanussi Cc: linux-kernel@vger.kernel.org Cc: linux-perf-users@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1334961696-19580-2-git-send-email-asharma@fb.com Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index cb608383e4f6..e78bc256aea8 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1781,9 +1781,6 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) if (bytes != sizeof(frame)) break; - if (fp < compat_ptr(regs->sp)) - break; - perf_callchain_store(entry, frame.return_address); fp = compat_ptr(frame.next_frame); } @@ -1827,9 +1824,6 @@ perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) if (bytes != sizeof(frame)) break; - if ((unsigned long)fp < regs->sp) - break; - perf_callchain_store(entry, frame.return_address); fp = frame.next_frame; } -- cgit From 0b0d9cf6ec7bab91977da2d71c09157f110f7c2e Mon Sep 17 00:00:00 2001 From: Arun Sharma Date: Fri, 20 Apr 2012 15:41:34 -0700 Subject: perf: Limit callchains to 127 Stack depth of 255 seems excessive, given that copy_from_user_nmi() could be slow. Signed-off-by: Arun Sharma Cc: Linus Torvalds Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1334961696-19580-3-git-send-email-asharma@fb.com Signed-off-by: Ingo Molnar --- include/linux/perf_event.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 1817d4015e5f..45db49f64bb4 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -555,7 +555,7 @@ enum perf_event_type { PERF_RECORD_MAX, /* non-ABI */ }; -#define PERF_MAX_STACK_DEPTH 255 +#define PERF_MAX_STACK_DEPTH 127 enum perf_callchain_context { PERF_CONTEXT_HV = (__u64)-32, -- cgit From bc6ca7b342d5ae15c3ba3081fd40271b8039fb25 Mon Sep 17 00:00:00 2001 From: Arun Sharma Date: Fri, 20 Apr 2012 15:41:35 -0700 Subject: perf/x86: Check if user fp is valid Signed-off-by: Arun Sharma Cc: Linus Torvalds Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1334961696-19580-4-git-send-email-asharma@fb.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uaccess.h | 12 ++++++------ arch/x86/kernel/cpu/perf_event.c | 12 ++++++++++++ 2 files changed, 18 insertions(+), 6 deletions(-) diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h index 04cd6882308e..e1f3a17034fc 100644 --- a/arch/x86/include/asm/uaccess.h +++ b/arch/x86/include/asm/uaccess.h @@ -33,9 +33,8 @@ #define segment_eq(a, b) ((a).seg == (b).seg) #define user_addr_max() (current_thread_info()->addr_limit.seg) -#define __addr_ok(addr) \ - ((unsigned long __force)(addr) < \ - (current_thread_info()->addr_limit.seg)) +#define __addr_ok(addr) \ + ((unsigned long __force)(addr) < user_addr_max()) /* * Test whether a block of memory is a valid user space address. @@ -47,14 +46,14 @@ * This needs 33-bit (65-bit for x86_64) arithmetic. We have a carry... */ -#define __range_not_ok(addr, size) \ +#define __range_not_ok(addr, size, limit) \ ({ \ unsigned long flag, roksum; \ __chk_user_ptr(addr); \ asm("add %3,%1 ; sbb %0,%0 ; cmp %1,%4 ; sbb $0,%0" \ : "=&r" (flag), "=r" (roksum) \ : "1" (addr), "g" ((long)(size)), \ - "rm" (current_thread_info()->addr_limit.seg)); \ + "rm" (limit)); \ flag; \ }) @@ -77,7 +76,8 @@ * checks that the pointer is in the user space range - after calling * this function, memory access functions may still return -EFAULT. */ -#define access_ok(type, addr, size) (likely(__range_not_ok(addr, size) == 0)) +#define access_ok(type, addr, size) \ + (likely(__range_not_ok(addr, size, user_addr_max()) == 0)) /* * The exception table consists of pairs of addresses relative to the diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index e78bc256aea8..c4706cf9c011 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1757,6 +1757,12 @@ perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) dump_trace(NULL, regs, NULL, 0, &backtrace_ops, entry); } +static inline int +valid_user_frame(const void __user *fp, unsigned long size) +{ + return (__range_not_ok(fp, size, TASK_SIZE) == 0); +} + #ifdef CONFIG_COMPAT #include @@ -1781,6 +1787,9 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) if (bytes != sizeof(frame)) break; + if (!valid_user_frame(fp, sizeof(frame))) + break; + perf_callchain_store(entry, frame.return_address); fp = compat_ptr(frame.next_frame); } @@ -1824,6 +1833,9 @@ perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) if (bytes != sizeof(frame)) break; + if (!valid_user_frame(fp, sizeof(frame))) + break; + perf_callchain_store(entry, frame.return_address); fp = frame.next_frame; } -- cgit From db0dc75d6403b6663c0eab4c6ccb672eb9b2ed72 Mon Sep 17 00:00:00 2001 From: Arun Sharma Date: Fri, 20 Apr 2012 15:41:36 -0700 Subject: perf/x86: Check user address explicitly in copy_from_user_nmi() Signed-off-by: Arun Sharma Cc: Linus Torvalds Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1334961696-19580-5-git-send-email-asharma@fb.com Signed-off-by: Ingo Molnar --- arch/x86/lib/usercopy.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c index f61ee67ec00f..677b1ed184c9 100644 --- a/arch/x86/lib/usercopy.c +++ b/arch/x86/lib/usercopy.c @@ -8,6 +8,7 @@ #include #include +#include /* * best effort, GUP based copy_from_user() that is NMI-safe @@ -21,6 +22,9 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n) void *map; int ret; + if (__range_not_ok(from, n, TASK_SIZE) == 0) + return len; + do { ret = __get_user_pages_fast(addr, 1, 0, &page); if (!ret) -- cgit