From 20d3cbd75ef17b96dfb4f5f1600c0278d330f5ca Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Thu, 12 Jun 2014 22:10:03 -0400 Subject: tools lib traceevent: Report unknown VMX exit reasons with code Allows to parse the result even if the KVM plugin does not yet understand a specific exit code. Link: http://lkml.kernel.org/r/5207446F.1090703@web.de Acked-by: Namhyung Kim Signed-off-by: Jan Kiszka Signed-off-by: Steven Rostedt Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_kvm.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_kvm.c b/tools/lib/traceevent/plugin_kvm.c index 9e0e8c6..3e61220 100644 --- a/tools/lib/traceevent/plugin_kvm.c +++ b/tools/lib/traceevent/plugin_kvm.c @@ -240,9 +240,8 @@ static const char *find_exit_reason(unsigned isa, int val) for (i = 0; strings[i].val >= 0; i++) if (strings[i].val == val) break; - if (strings[i].str) - return strings[i].str; - return "UNKNOWN"; + + return strings[i].str; } static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, @@ -251,6 +250,7 @@ static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, unsigned long long isa; unsigned long long val; unsigned long long info1 = 0, info2 = 0; + const char *reason; if (pevent_get_field_val(s, event, "exit_reason", record, &val, 1) < 0) return -1; @@ -258,7 +258,11 @@ static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, if (pevent_get_field_val(s, event, "isa", record, &isa, 0) < 0) isa = 1; - trace_seq_printf(s, "reason %s", find_exit_reason(isa, val)); + reason = find_exit_reason(isa, val); + if (reason) + trace_seq_printf(s, "reason %s", reason); + else + trace_seq_printf(s, "reason UNKNOWN (%llu)", val); pevent_print_num_field(s, " rip 0x%lx", event, "guest_rip", record, 1); -- cgit v1.1 From 6f21037b3295d4ba20ad2cc7cd4073ec64440f8f Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Thu, 12 Jun 2014 22:10:04 -0400 Subject: tools lib traceevent: Factor out print_exit_reason in kvm plugin We will reuse it for nested vmexit tracepoints. Link: http://lkml.kernel.org/r/619c418c8af87f03027b8c8013b0443996605700.1388855989.git.jan.kiszka@web.de Acked-by: Namhyung Kim Signed-off-by: Jan Kiszka Signed-off-by: Steven Rostedt Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_kvm.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_kvm.c b/tools/lib/traceevent/plugin_kvm.c index 3e61220..2d7d1d7 100644 --- a/tools/lib/traceevent/plugin_kvm.c +++ b/tools/lib/traceevent/plugin_kvm.c @@ -244,15 +244,14 @@ static const char *find_exit_reason(unsigned isa, int val) return strings[i].str; } -static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, - struct event_format *event, void *context) +static int print_exit_reason(struct trace_seq *s, struct pevent_record *record, + struct event_format *event, const char *field) { unsigned long long isa; unsigned long long val; - unsigned long long info1 = 0, info2 = 0; const char *reason; - if (pevent_get_field_val(s, event, "exit_reason", record, &val, 1) < 0) + if (pevent_get_field_val(s, event, field, record, &val, 1) < 0) return -1; if (pevent_get_field_val(s, event, "isa", record, &isa, 0) < 0) @@ -263,6 +262,16 @@ static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, trace_seq_printf(s, "reason %s", reason); else trace_seq_printf(s, "reason UNKNOWN (%llu)", val); + return 0; +} + +static int kvm_exit_handler(struct trace_seq *s, struct pevent_record *record, + struct event_format *event, void *context) +{ + unsigned long long info1 = 0, info2 = 0; + + if (print_exit_reason(s, record, event, "exit_reason") < 0) + return -1; pevent_print_num_field(s, " rip 0x%lx", event, "guest_rip", record, 1); -- cgit v1.1 From ea092aeb6d4725048c5a46d1c9cbb4fea49b80b0 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Thu, 12 Jun 2014 22:10:05 -0400 Subject: tools lib traceevent: Add back in kvm plugins nested_vmexit events The nested vmexit events were removed from the backport from trace-cmd because they were considered buggy. They have since been updated in trace-cmd but are still missing from the traceevent library. Add back in the buggy version to be able to backport the fixes. Acked-by: Namhyung Kim Signed-off-by: Steven Rostedt Link: http://lkml.kernel.org/r/20140613021157.291421941@goodmis.org Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_kvm.c | 41 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 41 insertions(+) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_kvm.c b/tools/lib/traceevent/plugin_kvm.c index 2d7d1d7..0d43783 100644 --- a/tools/lib/traceevent/plugin_kvm.c +++ b/tools/lib/traceevent/plugin_kvm.c @@ -326,6 +326,35 @@ static int kvm_emulate_insn_handler(struct trace_seq *s, return 0; } + +static int kvm_nested_vmexit_inject_handler(struct trace_seq *s, struct pevent_record *record, + struct event_format *event, void *context) +{ + unsigned long long val; + + pevent_print_num_field(s, " rip %0x016llx", event, "rip", record, 1); + + if (pevent_get_field_val(s, event, "exit_code", record, &val, 1) < 0) + return -1; + + trace_seq_printf(s, "reason %s", find_exit_reason(2, val)); + + pevent_print_num_field(s, " ext_inf1: %0x016llx", event, "exit_info1", record, 1); + pevent_print_num_field(s, " ext_inf2: %0x016llx", event, "exit_info2", record, 1); + pevent_print_num_field(s, " ext_int: %0x016llx", event, "exit_int_info", record, 1); + pevent_print_num_field(s, " ext_int_err: %0x016llx", event, "exit_int_info_err", record, 1); + + return 0; +} + +static int kvm_nested_vmexit_handler(struct trace_seq *s, struct pevent_record *record, + struct event_format *event, void *context) +{ + pevent_print_num_field(s, " rip %0x016llx", event, "rip", record, 1); + + return kvm_nested_vmexit_inject_handler(s, record, event, context); +} + union kvm_mmu_page_role { unsigned word; struct { @@ -422,6 +451,12 @@ int PEVENT_PLUGIN_LOADER(struct pevent *pevent) pevent_register_event_handler(pevent, -1, "kvm", "kvm_emulate_insn", kvm_emulate_insn_handler, NULL); + pevent_register_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit", + kvm_nested_vmexit_handler, NULL); + + pevent_register_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit_inject", + kvm_nested_vmexit_inject_handler, NULL); + pevent_register_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_get_page", kvm_mmu_get_page_handler, NULL); @@ -456,6 +491,12 @@ void PEVENT_PLUGIN_UNLOADER(struct pevent *pevent) pevent_unregister_event_handler(pevent, -1, "kvm", "kvm_emulate_insn", kvm_emulate_insn_handler, NULL); + pevent_unregister_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit", + kvm_nested_vmexit_handler, NULL); + + pevent_unregister_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit_inject", + kvm_nested_vmexit_inject_handler, NULL); + pevent_unregister_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_get_page", kvm_mmu_get_page_handler, NULL); -- cgit v1.1 From 7f6e3635db39fb2400dc515192125e7b73258000 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Thu, 12 Jun 2014 22:10:06 -0400 Subject: tools lib traceevent: Fix and cleanup kvm_nested_vmexit tracepoints Fix several issues of kvm_nested_vmexit[_inject]: field width aren't supported with pevent_print, rip was printed twice/incorrectly, SVM ISA was hard-coded, we don't use ':' to separate field names. Link: http://lkml.kernel.org/r/8e6c02b22ea8136c139a91c69d6cc73b8c5c184b.1388855989.git.jan.kiszka@web.de Acked-by: Namhyung Kim Signed-off-by: Jan Kiszka Signed-off-by: Steven Rostedt Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_kvm.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_kvm.c b/tools/lib/traceevent/plugin_kvm.c index 0d43783..0575e59 100644 --- a/tools/lib/traceevent/plugin_kvm.c +++ b/tools/lib/traceevent/plugin_kvm.c @@ -330,19 +330,13 @@ static int kvm_emulate_insn_handler(struct trace_seq *s, static int kvm_nested_vmexit_inject_handler(struct trace_seq *s, struct pevent_record *record, struct event_format *event, void *context) { - unsigned long long val; - - pevent_print_num_field(s, " rip %0x016llx", event, "rip", record, 1); - - if (pevent_get_field_val(s, event, "exit_code", record, &val, 1) < 0) + if (print_exit_reason(s, record, event, "exit_code") < 0) return -1; - trace_seq_printf(s, "reason %s", find_exit_reason(2, val)); - - pevent_print_num_field(s, " ext_inf1: %0x016llx", event, "exit_info1", record, 1); - pevent_print_num_field(s, " ext_inf2: %0x016llx", event, "exit_info2", record, 1); - pevent_print_num_field(s, " ext_int: %0x016llx", event, "exit_int_info", record, 1); - pevent_print_num_field(s, " ext_int_err: %0x016llx", event, "exit_int_info_err", record, 1); + pevent_print_num_field(s, " info1 %llx", event, "exit_info1", record, 1); + pevent_print_num_field(s, " info2 %llx", event, "exit_info2", record, 1); + pevent_print_num_field(s, " int_info %llx", event, "exit_int_info", record, 1); + pevent_print_num_field(s, " int_info_err %llx", event, "exit_int_info_err", record, 1); return 0; } @@ -350,7 +344,7 @@ static int kvm_nested_vmexit_inject_handler(struct trace_seq *s, struct pevent_r static int kvm_nested_vmexit_handler(struct trace_seq *s, struct pevent_record *record, struct event_format *event, void *context) { - pevent_print_num_field(s, " rip %0x016llx", event, "rip", record, 1); + pevent_print_num_field(s, "rip %lx ", event, "rip", record, 1); return kvm_nested_vmexit_inject_handler(s, record, event, context); } -- cgit v1.1 From a21e3a34bef9e9c177be61111ced23e71e25fd8a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Jun 2014 10:31:27 -0400 Subject: tools lib traceevent: Fix format in plugin_kvm The format field argument passed to the format in pevent_print_num_field() will be of type long long. That means that %ll must be used instead of %l. Acked-by: Namhyung Kim Reported-by: Namhyung Kim Signed-off-by: Steven Rostedt Link: http://lkml.kernel.org/r/20140613103127.1a9bdee7@gandalf.local.home Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_kvm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_kvm.c b/tools/lib/traceevent/plugin_kvm.c index 0575e59..88fe83d 100644 --- a/tools/lib/traceevent/plugin_kvm.c +++ b/tools/lib/traceevent/plugin_kvm.c @@ -344,7 +344,7 @@ static int kvm_nested_vmexit_inject_handler(struct trace_seq *s, struct pevent_r static int kvm_nested_vmexit_handler(struct trace_seq *s, struct pevent_record *record, struct event_format *event, void *context) { - pevent_print_num_field(s, "rip %lx ", event, "rip", record, 1); + pevent_print_num_field(s, "rip %llx ", event, "rip", record, 1); return kvm_nested_vmexit_inject_handler(s, record, event, context); } -- cgit v1.1 From 24eda087ccbb32f556cb3f9f78be152312bf6cc4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Jun 2014 19:44:20 -0400 Subject: tools lib traceevent: Clean up format of args in cfg80211 plugin While synchronizing what's in trace-cmd vs what's in perf, I came across a change that was made when entering the cfg80211 plugin into the tools/lib/traceevent directory. The function prototype went from: static unsigned long long process___le16_to_cpup(struct trace_seq *s, unsigned long long *args) to: static unsigned long long process___le16_to_cpup(struct trace_seq *s, unsigned long long *args) I can understand the line break after the long long, but there's no reason to keep args on a separate line. Signed-off-by: Steven Rostedt Link: http://lkml.kernel.org/r/20140612194420.24073744@gandalf.local.home Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_cfg80211.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_cfg80211.c b/tools/lib/traceevent/plugin_cfg80211.c index c066b25..4592d84 100644 --- a/tools/lib/traceevent/plugin_cfg80211.c +++ b/tools/lib/traceevent/plugin_cfg80211.c @@ -5,8 +5,7 @@ #include "event-parse.h" static unsigned long long -process___le16_to_cpup(struct trace_seq *s, - unsigned long long *args) +process___le16_to_cpup(struct trace_seq *s, unsigned long long *args) { uint16_t *val = (uint16_t *) (unsigned long) args[0]; return val ? (long long) le16toh(*val) : 0; -- cgit v1.1 From 1545d8aca9ac1cb3f503fb9c29543d539d99c7af Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Jun 2014 20:41:44 -0400 Subject: tools lib traceevent: Clean up format of args in jbd2 plugin While synchronizing what's in trace-cmd vs what's in perf, I came across a change that was made when entering the jbd2 plugin into the tools/lib/traceevent directory. For example, one of the function prototypes went from: unsigned long long process_jbd2_dev_to_name(struct trace_seq *s, unsigned long long *args) to: static unsigned long long process_jbd2_dev_to_name(struct trace_seq *s, unsigned long long *args) I can understand the line break after the long long, but there's no reason to keep args on a separate line. Signed-off-by: Steven Rostedt Link: http://lkml.kernel.org/r/20140612204144.018410d4@gandalf.local.home Signed-off-by: Jiri Olsa --- tools/lib/traceevent/plugin_jbd2.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/plugin_jbd2.c b/tools/lib/traceevent/plugin_jbd2.c index 0db714c..5c23d5b 100644 --- a/tools/lib/traceevent/plugin_jbd2.c +++ b/tools/lib/traceevent/plugin_jbd2.c @@ -30,8 +30,7 @@ #define MINOR(dev) ((unsigned int) ((dev) & MINORMASK)) static unsigned long long -process_jbd2_dev_to_name(struct trace_seq *s, - unsigned long long *args) +process_jbd2_dev_to_name(struct trace_seq *s, unsigned long long *args) { unsigned int dev = args[0]; @@ -40,8 +39,7 @@ process_jbd2_dev_to_name(struct trace_seq *s, } static unsigned long long -process_jiffies_to_msecs(struct trace_seq *s, - unsigned long long *args) +process_jiffies_to_msecs(struct trace_seq *s, unsigned long long *args) { unsigned long long jiffies = args[0]; -- cgit v1.1 From cdcd1e6bd8a92f8353fc2f37003c6eae2d1e6903 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 10 Jun 2014 16:00:18 -0300 Subject: perf trace: Fix up fd -> pathname resolution There was a brown paper bag bug in the patch that introduced a reference implementation on using 'perf probe' made wannabe tracepoints that broke fd -> pathname resolution, fix it: [root@zoo ~]# perf probe 'vfs_getname=getname_flags:65 pathname=result->name:string' Added new event: probe:vfs_getname (on getname_flags:65 with pathname=result->name:string) You can now use it in all perf tools, such as: perf record -e probe:vfs_getname -aR sleep 1 [root@zoo ~] Before: [acme@zoo linux]$ trace touch -e open,fstat /tmp/b 1.159 ( 0.007 ms): open(filename: 0x7fd73f2fe088, flags: CLOEXEC ) = 3 1.163 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e610 ) = 0 1.192 ( 0.009 ms): open(filename: 0x7fd73f4fedb8, flags: CLOEXEC ) = 3 1.201 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff1b25e660 ) = 0 1.501 ( 0.013 ms): open(filename: 0x7fd73f0a1610, flags: CLOEXEC ) = 3 1.505 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fd73f2ddb60 ) = 0 1.581 ( 0.011 ms): open(filename: 0x7fff1b2603da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3 [acme@zoo linux]$ After: [acme@zoo linux]$ trace touch -e open,fstat,dup2,mmap,close /tmp/b 1.105 ( 0.004 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fbf000 1.136 ( 0.008 ms): open(filename: 0x7f8902dbc088, flags: CLOEXEC ) = 3 1.140 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff19889ef0 ) = 0 1.146 ( 0.004 ms): mmap(len: 86079, prot: READ, flags: PRIVATE, fd: 3 ) = 0x2fa9000 1.149 ( 0.001 ms): close(fd: 3 ) = 0 1.170 ( 0.010 ms): open(filename: 0x7f8902fbcdb8, flags: CLOEXEC ) = 3 1.178 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7fff19889f40 ) = 0 1.188 ( 0.006 ms): mmap(len: 3924576, prot: EXEC|READ, flags: PRIVATE|DENYWRITE, fd: 3) = 0x29e2000 1.207 ( 0.007 ms): mmap(addr: 0x7f8902d96000, len: 24576, prot: READ|WRITE, flags: PRIVATE|DENYWRITE|FIXED, fd: 3, off: 1785856) = 0x2d96000 1.217 ( 0.004 ms): mmap(addr: 0x7f8902d9c000, len: 16992, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS|FIXED, fd: -1) = 0x2d9c000 1.228 ( 0.002 ms): close(fd: 3 ) = 0 1.243 ( 0.003 ms): mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa8000 1.250 ( 0.003 ms): mmap(len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS, fd: -1 ) = 0x2fa6000 1.452 ( 0.010 ms): open(filename: 0x7f8902b5f610, flags: CLOEXEC ) = 3 1.455 ( 0.002 ms): fstat(fd: 3, statbuf: 0x7f8902d9bb60 ) = 0 1.461 ( 0.004 ms): mmap(len: 106070960, prot: READ, flags: PRIVATE, fd: 3) = 0xfc4b9000 1.469 ( 0.002 ms): close(fd: 3 ) = 0 1.528 ( 0.010 ms): open(filename: 0x7fff1988c3da, flags: CREAT|NOCTTY|NONBLOCK|WRONLY, mode: 438) = 3 1.532 ( 0.002 ms): dup2(oldfd: 3 ) = 0 1.535 ( 0.001 ms): close(fd: 3 ) = 0 1.544 ( 0.001 ms): close( ) = 0 1.555 ( 0.001 ms): close(fd: 1 ) = 0 1.558 ( 0.001 ms): close(fd: 2 ) = 0 [acme@zoo linux]$ Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-vcm22xpjxc3j4hbyuzjzf7ik@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index f954c26..5ab2f67 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1276,11 +1276,11 @@ static const char *thread__fd_path(struct thread *thread, int fd, if (fd < 0) return NULL; - if ((fd > ttrace->paths.max || ttrace->paths.table[fd] == NULL)) + if ((fd > ttrace->paths.max || ttrace->paths.table[fd] == NULL)) { if (!trace->live) return NULL; ++trace->stats.proc_getname; - if (thread__read_fd_path(thread, fd)) { + if (thread__read_fd_path(thread, fd)) return NULL; } -- cgit v1.1 From 5229e366ee6baeb58b77e09643d2e11cbbd29950 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 10 Jun 2014 17:18:54 -0300 Subject: perf evlist: Add suggestion of how to set perf_event_paranoid sysctl Minor hint to speed up problem resolution and get 'trace' working for non root users. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-abdqi8km4fj9osrn70q2zj9v@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evlist.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index 59ef280..c51223a 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -1214,10 +1214,11 @@ int perf_evlist__strerror_open(struct perf_evlist *evlist __maybe_unused, "For your workloads it needs to be <= 1\nHint:\t"); } printed += scnprintf(buf + printed, size - printed, - "For system wide tracing it needs to be set to -1"); + "For system wide tracing it needs to be set to -1.\n"); printed += scnprintf(buf + printed, size - printed, - ".\nHint:\tThe current value is %d.", value); + "Hint:\tTry: 'sudo sh -c \"echo -1 > /proc/sys/kernel/perf_event_paranoid\"'\n" + "Hint:\tThe current value is %d.", value); break; default: scnprintf(buf, size, "%s", emsg); -- cgit v1.1 From 774135344fa8aa044290d030068f92e9a3aab8cc Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 16 Jun 2014 17:21:59 -0300 Subject: perf trace: Remove needless reassignments The thread->priv value is already obtained a few lines earlier from the thread__trace() call. Leftovers from before thread__trace(). Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-96laa634vzfwlwxurevo40wp@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 5ab2f67..28c86e2 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1629,7 +1629,6 @@ static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, return -1; args = perf_evsel__sc_tp_ptr(evsel, args, sample); - ttrace = thread->priv; if (ttrace->entry_str == NULL) { ttrace->entry_str = malloc(1024); @@ -1687,8 +1686,6 @@ static int trace__sys_exit(struct trace *trace, struct perf_evsel *evsel, ++trace->stats.vfs_getname; } - ttrace = thread->priv; - ttrace->exit_time = sample->time; if (ttrace->entry_time) { -- cgit v1.1 From 5089f20ee7104bd219cafefa62d83f53e75cc44c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 17 Jun 2014 14:29:24 -0300 Subject: perf trace: Cache the is_exit syscall test No need to use two strcmp calls per syscall entry, do it just once, when reading the per syscall info. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-lymtxhz0mg3adyt5e2pssn8f@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 28c86e2..5549cee 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1108,6 +1108,7 @@ struct syscall { struct event_format *tp_format; const char *name; bool filtered; + bool is_exit; struct syscall_fmt *fmt; size_t (**arg_scnprintf)(char *bf, size_t size, struct syscall_arg *arg); void **arg_parm; @@ -1473,6 +1474,8 @@ static int trace__read_syscall_info(struct trace *trace, int id) if (sc->tp_format == NULL) return -1; + sc->is_exit = !strcmp(name, "exit_group") || !strcmp(name, "exit"); + return syscall__set_arg_fmts(sc); } @@ -1643,7 +1646,7 @@ static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, printed += syscall__scnprintf_args(sc, msg + printed, 1024 - printed, args, trace, thread); - if (!strcmp(sc->name, "exit_group") || !strcmp(sc->name, "exit")) { + if (sc->is_exit) { if (!trace->duration_filter && !trace->summary_only) { trace__fprintf_entry_head(trace, thread, 1, sample->time, trace->output); fprintf(trace->output, "%-70s\n", ttrace->entry_str); -- cgit v1.1 From dd00d486ddb7f181cf9487f6aceb1066bc6b0b6a Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Thu, 19 Jun 2014 13:41:13 +0200 Subject: perf hists browser: Remove ev_name argument from perf_evsel__hists_browse Removing ev_name argument from perf_evsel__hists_browse function, because it's not needed. We can get the name out of the 'struct perf_evsel' which is passed as argument as well. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403178076-14072-3-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 25 +++++++++++-------------- 1 file changed, 11 insertions(+), 14 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 52c03fb..1bd35e8 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -32,8 +32,7 @@ struct hist_browser { extern void hist_browser__init_hpp(void); -static int hists__browser_title(struct hists *hists, char *bf, size_t size, - const char *ev_name); +static int hists__browser_title(struct hists *hists, char *bf, size_t size); static void hist_browser__update_nr_entries(struct hist_browser *hb); static struct rb_node *hists__filter_entries(struct rb_node *nd, @@ -345,7 +344,7 @@ static void ui_browser__warn_lost_events(struct ui_browser *browser) "Or reduce the sampling frequency."); } -static int hist_browser__run(struct hist_browser *browser, const char *ev_name, +static int hist_browser__run(struct hist_browser *browser, struct hist_browser_timer *hbt) { int key; @@ -356,7 +355,7 @@ static int hist_browser__run(struct hist_browser *browser, const char *ev_name, browser->b.nr_entries = hist_browser__nr_entries(browser); hist_browser__refresh_dimensions(browser); - hists__browser_title(browser->hists, title, sizeof(title), ev_name); + hists__browser_title(browser->hists, title, sizeof(title)); if (ui_browser__show(&browser->b, title, "Press '?' for help on key bindings") < 0) @@ -383,7 +382,7 @@ static int hist_browser__run(struct hist_browser *browser, const char *ev_name, ui_browser__warn_lost_events(&browser->b); } - hists__browser_title(browser->hists, title, sizeof(title), ev_name); + hists__browser_title(browser->hists, title, sizeof(title)); ui_browser__show_title(&browser->b, title); continue; } @@ -1212,8 +1211,7 @@ static struct thread *hist_browser__selected_thread(struct hist_browser *browser return browser->he_selection->thread; } -static int hists__browser_title(struct hists *hists, char *bf, size_t size, - const char *ev_name) +static int hists__browser_title(struct hists *hists, char *bf, size_t size) { char unit; int printed; @@ -1222,6 +1220,7 @@ static int hists__browser_title(struct hists *hists, char *bf, size_t size, unsigned long nr_samples = hists->stats.nr_events[PERF_RECORD_SAMPLE]; u64 nr_events = hists->stats.total_period; struct perf_evsel *evsel = hists_to_evsel(hists); + const char *ev_name = perf_evsel__name(evsel); char buf[512]; size_t buflen = sizeof(buf); @@ -1389,7 +1388,7 @@ static void hist_browser__update_nr_entries(struct hist_browser *hb) } static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events, - const char *helpline, const char *ev_name, + const char *helpline, bool left_exits, struct hist_browser_timer *hbt, float min_pcnt, @@ -1464,7 +1463,7 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events, nr_options = 0; - key = hist_browser__run(browser, ev_name, hbt); + key = hist_browser__run(browser, hbt); if (browser->he_selection != NULL) { thread = hist_browser__selected_thread(browser); @@ -1832,7 +1831,7 @@ static int perf_evsel_menu__run(struct perf_evsel_menu *menu, { struct perf_evlist *evlist = menu->b.priv; struct perf_evsel *pos; - const char *ev_name, *title = "Available samples"; + const char *title = "Available samples"; int delay_secs = hbt ? hbt->refresh : 0; int key; @@ -1865,9 +1864,8 @@ browse_hists: */ if (hbt) hbt->timer(hbt->arg); - ev_name = perf_evsel__name(pos); key = perf_evsel__hists_browse(pos, nr_events, help, - ev_name, true, hbt, + true, hbt, menu->min_pcnt, menu->env); ui_browser__show_title(&menu->b, title); @@ -1971,10 +1969,9 @@ int perf_evlist__tui_browse_hists(struct perf_evlist *evlist, const char *help, single_entry: if (nr_entries == 1) { struct perf_evsel *first = perf_evlist__first(evlist); - const char *ev_name = perf_evsel__name(first); return perf_evsel__hists_browse(first, nr_entries, help, - ev_name, false, hbt, min_pcnt, + false, hbt, min_pcnt, env); } -- cgit v1.1 From 89632972e2c56356d1e227aac151cf4e7c2f30d6 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Thu, 19 Jun 2014 13:41:14 +0200 Subject: perf ui browser: Fix scrollbar refresh row index The ui_browser__gotorc function needs offset from 'y' member, so the row index has to begin with 0, which happens by accident in current code, because we display only one header line. The bug shows when we want to display more than 1 header lines like columns headers in following patches. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403178076-14072-4-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browser.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/ui/browser.c b/tools/perf/ui/browser.c index 3ccf6e1..9d2294e 100644 --- a/tools/perf/ui/browser.c +++ b/tools/perf/ui/browser.c @@ -279,7 +279,7 @@ static void ui_browser__scrollbar_set(struct ui_browser *browser) { int height = browser->height, h = 0, pct = 0, col = browser->width, - row = browser->y - 1; + row = 0; if (browser->nr_entries > 1) { pct = ((browser->index * (browser->height - 1)) / -- cgit v1.1 From b094c99e8e284cff839400a3b61fda1fa53962fc Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Mon, 16 Jun 2014 11:14:22 -0700 Subject: perf bench sched-messaging: Plug memleak Explicitly free the thread array ('pth_tab'). Signed-off-by: Davidlohr Bueso Cc: Aswin Chandramouleeswaran Cc: Hitoshi Mitake Cc: Jiri Olsa Link: http://lkml.kernel.org/r/1402942467-10671-5-git-send-email-davidlohr@hp.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/bench/sched-messaging.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools') diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c index cc1190a..fc4fe91 100644 --- a/tools/perf/bench/sched-messaging.c +++ b/tools/perf/bench/sched-messaging.c @@ -332,5 +332,7 @@ int bench_sched_messaging(int argc, const char **argv, break; } + free(pth_tab); + return 0; } -- cgit v1.1 From b6f0629a94f7ed6089560be7f0561be19f934fc4 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Mon, 16 Jun 2014 11:14:19 -0700 Subject: perf bench: Add --repeat option There are a number of benchmarks that do single runs and as a result does not really help users gain a general idea of how the workload performs. So the user must either manually do multiple runs or just use single bogus results. This option will enable users to specify the amount of runs (arbitrarily defaulted to 10, to use the existing benchmarks default) through the '--repeat' option. Add it to perf-bench instead of implementing it always in each specific benchmark. Signed-off-by: Davidlohr Bueso Cc: Aswin Chandramouleeswaran Cc: Hitoshi Mitake Cc: Jiri Olsa Link: http://lkml.kernel.org/r/1402942467-10671-2-git-send-email-davidlohr@hp.com [ Kept the existing default of 10, changing it to something else should be done on separate patch ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-bench.txt | 4 ++++ tools/perf/bench/bench.h | 1 + tools/perf/builtin-bench.c | 7 +++++++ 3 files changed, 12 insertions(+) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-bench.txt b/tools/perf/Documentation/perf-bench.txt index 4464ad7..f6480cb 100644 --- a/tools/perf/Documentation/perf-bench.txt +++ b/tools/perf/Documentation/perf-bench.txt @@ -16,6 +16,10 @@ This 'perf bench' command is a general framework for benchmark suites. COMMON OPTIONS -------------- +-r:: +--repeat=:: +Specify amount of times to repeat the run (default 10). + -f:: --format=:: Specify format style. diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h index eba46709..3c4dd44 100644 --- a/tools/perf/bench/bench.h +++ b/tools/perf/bench/bench.h @@ -43,5 +43,6 @@ extern int bench_futex_requeue(int argc, const char **argv, const char *prefix); #define BENCH_FORMAT_UNKNOWN -1 extern int bench_format; +extern unsigned int bench_repeat; #endif diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c index 1e6e777..b9a56fa 100644 --- a/tools/perf/builtin-bench.c +++ b/tools/perf/builtin-bench.c @@ -104,9 +104,11 @@ static const char *bench_format_str; /* Output/formatting style, exported to benchmark modules: */ int bench_format = BENCH_FORMAT_DEFAULT; +unsigned int bench_repeat = 10; /* default number of times to repeat the run */ static const struct option bench_options[] = { OPT_STRING('f', "format", &bench_format_str, "default", "Specify format style"), + OPT_UINTEGER('r', "repeat", &bench_repeat, "Specify amount of times to repeat the run"), OPT_END() }; @@ -226,6 +228,11 @@ int cmd_bench(int argc, const char **argv, const char *prefix __maybe_unused) goto end; } + if (bench_repeat == 0) { + printf("Invalid repeat option: Must specify a positive value\n"); + goto end; + } + if (argc < 1) { print_usage(); goto end; -- cgit v1.1 From d9de84afd1f3a464135abe2b26e66aa86be5af8d Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Mon, 16 Jun 2014 11:14:23 -0700 Subject: perf bench futex: Use global --repeat option This option is available through perf-bench, use it instead and free the local option. Signed-off-by: Davidlohr Bueso Cc: Aswin Chandramouleeswaran Cc: Hitoshi Mitake Cc: Jiri Olsa Link: http://lkml.kernel.org/r/1402942467-10671-6-git-send-email-davidlohr@hp.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/bench/futex-requeue.c | 10 +--------- tools/perf/bench/futex-wake.c | 12 ++---------- 2 files changed, 3 insertions(+), 19 deletions(-) (limited to 'tools') diff --git a/tools/perf/bench/futex-requeue.c b/tools/perf/bench/futex-requeue.c index a1625587..732403b 100644 --- a/tools/perf/bench/futex-requeue.c +++ b/tools/perf/bench/futex-requeue.c @@ -29,13 +29,6 @@ static u_int32_t futex1 = 0, futex2 = 0; */ static unsigned int nrequeue = 1; -/* - * There can be significant variance from run to run, - * the more repeats, the more exact the overall avg and - * the better idea of the futex latency. - */ -static unsigned int repeat = 10; - static pthread_t *worker; static bool done = 0, silent = 0; static pthread_mutex_t thread_lock; @@ -46,7 +39,6 @@ static unsigned int ncpus, threads_starting, nthreads = 0; static const struct option options[] = { OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"), OPT_UINTEGER('q', "nrequeue", &nrequeue, "Specify amount of threads to requeue at once"), - OPT_UINTEGER('r', "repeat", &repeat, "Specify amount of times to repeat the run"), OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"), OPT_END() }; @@ -146,7 +138,7 @@ int bench_futex_requeue(int argc, const char **argv, pthread_cond_init(&thread_parent, NULL); pthread_cond_init(&thread_worker, NULL); - for (j = 0; j < repeat && !done; j++) { + for (j = 0; j < bench_repeat && !done; j++) { unsigned int nrequeued = 0; struct timeval start, end, runtime; diff --git a/tools/perf/bench/futex-wake.c b/tools/perf/bench/futex-wake.c index d096169..50022cb 100644 --- a/tools/perf/bench/futex-wake.c +++ b/tools/perf/bench/futex-wake.c @@ -30,15 +30,8 @@ static u_int32_t futex1 = 0; */ static unsigned int nwakes = 1; -/* - * There can be significant variance from run to run, - * the more repeats, the more exact the overall avg and - * the better idea of the futex latency. - */ -static unsigned int repeat = 10; - pthread_t *worker; -static bool done = 0, silent = 0; +static bool done = false, silent = false; static pthread_mutex_t thread_lock; static pthread_cond_t thread_parent, thread_worker; static struct stats waketime_stats, wakeup_stats; @@ -47,7 +40,6 @@ static unsigned int ncpus, threads_starting, nthreads = 0; static const struct option options[] = { OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"), OPT_UINTEGER('w', "nwakes", &nwakes, "Specify amount of threads to wake at once"), - OPT_UINTEGER('r', "repeat", &repeat, "Specify amount of times to repeat the run"), OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"), OPT_END() }; @@ -149,7 +141,7 @@ int bench_futex_wake(int argc, const char **argv, pthread_cond_init(&thread_parent, NULL); pthread_cond_init(&thread_worker, NULL); - for (j = 0; j < repeat && !done; j++) { + for (j = 0; j < bench_repeat && !done; j++) { unsigned int nwoken = 0; struct timeval start, end, runtime; -- cgit v1.1 From 424e9634887842ac59c1d06d3264aaeb18853c0b Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Mon, 16 Jun 2014 11:14:25 -0700 Subject: perf bench mem: The -o and -n options are mutually exclusive -o, --only-prefault Show only the result with page faults before mem* -n, --no-prefault Show only the result without page faults before mem* Makes no sense to call together. Applies to both memset and memcpy. Signed-off-by: Davidlohr Bueso Cc: Aswin Chandramouleeswaran Cc: Hitoshi Mitake Cc: Jiri Olsa Link: http://lkml.kernel.org/r/1402942467-10671-8-git-send-email-davidlohr@hp.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/bench/mem-memcpy.c | 5 +++++ tools/perf/bench/mem-memset.c | 5 +++++ 2 files changed, 10 insertions(+) (limited to 'tools') diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c index 5ce71d3..e622c3e 100644 --- a/tools/perf/bench/mem-memcpy.c +++ b/tools/perf/bench/mem-memcpy.c @@ -189,6 +189,11 @@ int bench_mem_memcpy(int argc, const char **argv, argc = parse_options(argc, argv, options, bench_mem_memcpy_usage, 0); + if (no_prefault && only_prefault) { + fprintf(stderr, "Invalid options: -o and -n are mutually exclusive\n"); + return 1; + } + if (use_cycle) init_cycle(); diff --git a/tools/perf/bench/mem-memset.c b/tools/perf/bench/mem-memset.c index 9af79d2..2a65468 100644 --- a/tools/perf/bench/mem-memset.c +++ b/tools/perf/bench/mem-memset.c @@ -181,6 +181,11 @@ int bench_mem_memset(int argc, const char **argv, argc = parse_options(argc, argv, options, bench_mem_memset_usage, 0); + if (no_prefault && only_prefault) { + fprintf(stderr, "Invalid options: -o and -n are mutually exclusive\n"); + return 1; + } + if (use_cycle) init_cycle(); -- cgit v1.1 From ecdac96899e3db3f428e4d2e978f25e3f8d35a6c Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Mon, 16 Jun 2014 11:14:26 -0700 Subject: perf bench sched-messaging: Drop barf() Instead of reinventing the wheel, we can use err(2) when dealing with fatal errors. Exit code is now always EXIT_FAILURE (1). Signed-off-by: Davidlohr Bueso Cc: Aswin Chandramouleeswaran Cc: Hitoshi Mitake Cc: Jiri Olsa Link: http://lkml.kernel.org/r/1402942467-10671-9-git-send-email-davidlohr@hp.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/bench/sched-messaging.c | 45 ++++++++++++++++---------------------- 1 file changed, 19 insertions(+), 26 deletions(-) (limited to 'tools') diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c index fc4fe91..52a5659 100644 --- a/tools/perf/bench/sched-messaging.c +++ b/tools/perf/bench/sched-messaging.c @@ -28,6 +28,7 @@ #include #include #include +#include #define DATASIZE 100 @@ -50,12 +51,6 @@ struct receiver_context { int wakefd; }; -static void barf(const char *msg) -{ - fprintf(stderr, "%s (error: %s)\n", msg, strerror(errno)); - exit(1); -} - static void fdpair(int fds[2]) { if (use_pipes) { @@ -66,7 +61,7 @@ static void fdpair(int fds[2]) return; } - barf(use_pipes ? "pipe()" : "socketpair()"); + err(EXIT_FAILURE, use_pipes ? "pipe()" : "socketpair()"); } /* Block until we're ready to go */ @@ -77,11 +72,11 @@ static void ready(int ready_out, int wakefd) /* Tell them we're ready. */ if (write(ready_out, &dummy, 1) != 1) - barf("CLIENT: ready write"); + err(EXIT_FAILURE, "CLIENT: ready write"); /* Wait for "GO" signal */ if (poll(&pollfd, 1, -1) != 1) - barf("poll"); + err(EXIT_FAILURE, "poll"); } /* Sender sprays loops messages down each file descriptor */ @@ -101,7 +96,7 @@ again: ret = write(ctx->out_fds[j], data + done, sizeof(data)-done); if (ret < 0) - barf("SENDER: write"); + err(EXIT_FAILURE, "SENDER: write"); done += ret; if (done < DATASIZE) goto again; @@ -131,7 +126,7 @@ static void *receiver(struct receiver_context* ctx) again: ret = read(ctx->in_fds[0], data + done, DATASIZE - done); if (ret < 0) - barf("SERVER: read"); + err(EXIT_FAILURE, "SERVER: read"); done += ret; if (done < DATASIZE) goto again; @@ -144,14 +139,14 @@ static pthread_t create_worker(void *ctx, void *(*func)(void *)) { pthread_attr_t attr; pthread_t childid; - int err; + int ret; if (!thread_mode) { /* process mode */ /* Fork the receiver. */ switch (fork()) { case -1: - barf("fork()"); + err(EXIT_FAILURE, "fork()"); break; case 0: (*func) (ctx); @@ -165,19 +160,17 @@ static pthread_t create_worker(void *ctx, void *(*func)(void *)) } if (pthread_attr_init(&attr) != 0) - barf("pthread_attr_init:"); + err(EXIT_FAILURE, "pthread_attr_init:"); #ifndef __ia64__ if (pthread_attr_setstacksize(&attr, PTHREAD_STACK_MIN) != 0) - barf("pthread_attr_setstacksize"); + err(EXIT_FAILURE, "pthread_attr_setstacksize"); #endif - err = pthread_create(&childid, &attr, func, ctx); - if (err != 0) { - fprintf(stderr, "pthread_create failed: %s (%d)\n", - strerror(err), err); - exit(-1); - } + ret = pthread_create(&childid, &attr, func, ctx); + if (ret != 0) + err(EXIT_FAILURE, "pthread_create failed"); + return childid; } @@ -207,14 +200,14 @@ static unsigned int group(pthread_t *pth, + num_fds * sizeof(int)); if (!snd_ctx) - barf("malloc()"); + err(EXIT_FAILURE, "malloc()"); for (i = 0; i < num_fds; i++) { int fds[2]; struct receiver_context *ctx = malloc(sizeof(*ctx)); if (!ctx) - barf("malloc()"); + err(EXIT_FAILURE, "malloc()"); /* Create the pipe between client and server */ @@ -281,7 +274,7 @@ int bench_sched_messaging(int argc, const char **argv, pth_tab = malloc(num_fds * 2 * num_groups * sizeof(pthread_t)); if (!pth_tab) - barf("main:malloc()"); + err(EXIT_FAILURE, "main:malloc()"); fdpair(readyfds); fdpair(wakefds); @@ -294,13 +287,13 @@ int bench_sched_messaging(int argc, const char **argv, /* Wait for everyone to be ready */ for (i = 0; i < total_children; i++) if (read(readyfds[0], &dummy, 1) != 1) - barf("Reading for readyfds"); + err(EXIT_FAILURE, "Reading for readyfds"); gettimeofday(&start, NULL); /* Kick them off */ if (write(wakefds[1], &dummy, 1) != 1) - barf("Writing to start them"); + err(EXIT_FAILURE, "Writing to start them"); /* Reap them all */ for (i = 0; i < total_children; i++) -- cgit v1.1 From 07100877ea8fd9b2feabb4dd78f3322892f6bd77 Mon Sep 17 00:00:00 2001 From: Daniel Bristot de Oliveira Date: Wed, 11 Jun 2014 16:09:08 -0300 Subject: perf scripts: Fallback to syscalls:* when raw_syscalls:* is not available Older kernels (e.g., RHEL6) do system call tracing via the syscalls:sys_{enter,exit} tracepoints rather than using raw_syscalls:*. Update perf python and perl scripts to fallback to syscalls:* when raw_syscalls:* isn't available. Signed-off-by: Daniel Bristot de Oliveira Cc: Ingo Molnar Cc: Luis Claudio R. Goncalves Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/5a6c64081a3375bc3bc66351b14559678ef4d71e.1402507908.git.bristot@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/scripts/perl/bin/failed-syscalls-record | 3 ++- tools/perf/scripts/perl/failed-syscalls.pl | 5 +++++ tools/perf/scripts/python/bin/failed-syscalls-by-pid-record | 3 ++- tools/perf/scripts/python/bin/sctop-record | 3 ++- tools/perf/scripts/python/bin/syscall-counts-by-pid-record | 3 ++- tools/perf/scripts/python/bin/syscall-counts-record | 3 ++- tools/perf/scripts/python/failed-syscalls-by-pid.py | 5 +++++ tools/perf/scripts/python/sctop.py | 5 +++++ tools/perf/scripts/python/syscall-counts-by-pid.py | 5 +++++ tools/perf/scripts/python/syscall-counts.py | 5 +++++ 10 files changed, 35 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/scripts/perl/bin/failed-syscalls-record b/tools/perf/scripts/perl/bin/failed-syscalls-record index 8104895..74685f3 100644 --- a/tools/perf/scripts/perl/bin/failed-syscalls-record +++ b/tools/perf/scripts/perl/bin/failed-syscalls-record @@ -1,2 +1,3 @@ #!/bin/bash -perf record -e raw_syscalls:sys_exit $@ +(perf record -e raw_syscalls:sys_exit $@ || \ + perf record -e syscalls:sys_exit $@) 2> /dev/null diff --git a/tools/perf/scripts/perl/failed-syscalls.pl b/tools/perf/scripts/perl/failed-syscalls.pl index 94bc25a..55e7ae4 100644 --- a/tools/perf/scripts/perl/failed-syscalls.pl +++ b/tools/perf/scripts/perl/failed-syscalls.pl @@ -26,6 +26,11 @@ sub raw_syscalls::sys_exit } } +sub syscalls::sys_exit +{ + raw_syscalls::sys_exit(@_) +} + sub trace_end { printf("\nfailed syscalls by comm:\n\n"); diff --git a/tools/perf/scripts/python/bin/failed-syscalls-by-pid-record b/tools/perf/scripts/python/bin/failed-syscalls-by-pid-record index 8104895..74685f3 100644 --- a/tools/perf/scripts/python/bin/failed-syscalls-by-pid-record +++ b/tools/perf/scripts/python/bin/failed-syscalls-by-pid-record @@ -1,2 +1,3 @@ #!/bin/bash -perf record -e raw_syscalls:sys_exit $@ +(perf record -e raw_syscalls:sys_exit $@ || \ + perf record -e syscalls:sys_exit $@) 2> /dev/null diff --git a/tools/perf/scripts/python/bin/sctop-record b/tools/perf/scripts/python/bin/sctop-record index 4efbfaa..d694084 100644 --- a/tools/perf/scripts/python/bin/sctop-record +++ b/tools/perf/scripts/python/bin/sctop-record @@ -1,2 +1,3 @@ #!/bin/bash -perf record -e raw_syscalls:sys_enter $@ +(perf record -e raw_syscalls:sys_enter $@ || \ + perf record -e syscalls:sys_enter $@) 2> /dev/null diff --git a/tools/perf/scripts/python/bin/syscall-counts-by-pid-record b/tools/perf/scripts/python/bin/syscall-counts-by-pid-record index 4efbfaa..d694084 100644 --- a/tools/perf/scripts/python/bin/syscall-counts-by-pid-record +++ b/tools/perf/scripts/python/bin/syscall-counts-by-pid-record @@ -1,2 +1,3 @@ #!/bin/bash -perf record -e raw_syscalls:sys_enter $@ +(perf record -e raw_syscalls:sys_enter $@ || \ + perf record -e syscalls:sys_enter $@) 2> /dev/null diff --git a/tools/perf/scripts/python/bin/syscall-counts-record b/tools/perf/scripts/python/bin/syscall-counts-record index 4efbfaa..d694084 100644 --- a/tools/perf/scripts/python/bin/syscall-counts-record +++ b/tools/perf/scripts/python/bin/syscall-counts-record @@ -1,2 +1,3 @@ #!/bin/bash -perf record -e raw_syscalls:sys_enter $@ +(perf record -e raw_syscalls:sys_enter $@ || \ + perf record -e syscalls:sys_enter $@) 2> /dev/null diff --git a/tools/perf/scripts/python/failed-syscalls-by-pid.py b/tools/perf/scripts/python/failed-syscalls-by-pid.py index 85805fa..266a836 100644 --- a/tools/perf/scripts/python/failed-syscalls-by-pid.py +++ b/tools/perf/scripts/python/failed-syscalls-by-pid.py @@ -50,6 +50,11 @@ def raw_syscalls__sys_exit(event_name, context, common_cpu, except TypeError: syscalls[common_comm][common_pid][id][ret] = 1 +def syscalls__sys_exit(event_name, context, common_cpu, + common_secs, common_nsecs, common_pid, common_comm, + id, ret): + raw_syscalls__sys_exit(**locals()) + def print_error_totals(): if for_comm is not None: print "\nsyscall errors for %s:\n\n" % (for_comm), diff --git a/tools/perf/scripts/python/sctop.py b/tools/perf/scripts/python/sctop.py index 42c267e..c9f3058 100644 --- a/tools/perf/scripts/python/sctop.py +++ b/tools/perf/scripts/python/sctop.py @@ -53,6 +53,11 @@ def raw_syscalls__sys_enter(event_name, context, common_cpu, except TypeError: syscalls[id] = 1 +def syscalls__sys_enter(event_name, context, common_cpu, + common_secs, common_nsecs, common_pid, common_comm, + id, args): + raw_syscalls__sys_enter(**locals()) + def print_syscall_totals(interval): while 1: clear_term() diff --git a/tools/perf/scripts/python/syscall-counts-by-pid.py b/tools/perf/scripts/python/syscall-counts-by-pid.py index c64d1c5..cf2054c 100644 --- a/tools/perf/scripts/python/syscall-counts-by-pid.py +++ b/tools/perf/scripts/python/syscall-counts-by-pid.py @@ -48,6 +48,11 @@ def raw_syscalls__sys_enter(event_name, context, common_cpu, except TypeError: syscalls[common_comm][common_pid][id] = 1 +def syscalls__sys_enter(event_name, context, common_cpu, + common_secs, common_nsecs, common_pid, common_comm, + id, args): + raw_syscalls__sys_enter(**locals()) + def print_syscall_totals(): if for_comm is not None: print "\nsyscall events for %s:\n\n" % (for_comm), diff --git a/tools/perf/scripts/python/syscall-counts.py b/tools/perf/scripts/python/syscall-counts.py index b435d3f..92b2938 100644 --- a/tools/perf/scripts/python/syscall-counts.py +++ b/tools/perf/scripts/python/syscall-counts.py @@ -44,6 +44,11 @@ def raw_syscalls__sys_enter(event_name, context, common_cpu, except TypeError: syscalls[id] = 1 +def syscalls__sys_enter(event_name, context, common_cpu, + common_secs, common_nsecs, common_pid, common_comm, + id, args): + raw_syscalls__sys_enter(**locals()) + def print_syscall_totals(): if for_comm is not None: print "\nsyscall events for %s:\n\n" % (for_comm), -- cgit v1.1 From 0c82adcf141935b6312593a53f87342dbb12b704 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Thu, 26 Jun 2014 20:14:24 +0400 Subject: perf trace: Add perf_event parameter to tracepoint_handler It will be used by next pagefault tracing patches in the series. Signed-off-by: Stanislav Fomichev Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403799268-1367-2-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 5549cee..4a9e26b 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1538,6 +1538,7 @@ static size_t syscall__scnprintf_args(struct syscall *sc, char *bf, size_t size, } typedef int (*tracepoint_handler)(struct trace *trace, struct perf_evsel *evsel, + union perf_event *event, struct perf_sample *sample); static struct syscall *trace__syscall_info(struct trace *trace, @@ -1610,6 +1611,7 @@ static void thread__update_stats(struct thread_trace *ttrace, } static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, + union perf_event *event __maybe_unused, struct perf_sample *sample) { char *msg; @@ -1658,6 +1660,7 @@ static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, } static int trace__sys_exit(struct trace *trace, struct perf_evsel *evsel, + union perf_event *event __maybe_unused, struct perf_sample *sample) { int ret; @@ -1735,6 +1738,7 @@ out: } static int trace__vfs_getname(struct trace *trace, struct perf_evsel *evsel, + union perf_event *event __maybe_unused, struct perf_sample *sample) { trace->last_vfs_getname = perf_evsel__rawptr(evsel, sample, "pathname"); @@ -1742,6 +1746,7 @@ static int trace__vfs_getname(struct trace *trace, struct perf_evsel *evsel, } static int trace__sched_stat_runtime(struct trace *trace, struct perf_evsel *evsel, + union perf_event *event __maybe_unused, struct perf_sample *sample) { u64 runtime = perf_evsel__intval(evsel, sample, "runtime"); @@ -1781,7 +1786,7 @@ static bool skip_sample(struct trace *trace, struct perf_sample *sample) } static int trace__process_sample(struct perf_tool *tool, - union perf_event *event __maybe_unused, + union perf_event *event, struct perf_sample *sample, struct perf_evsel *evsel, struct machine *machine __maybe_unused) @@ -1799,7 +1804,7 @@ static int trace__process_sample(struct perf_tool *tool, if (handler) { ++trace->nr_events; - handler(trace, evsel, sample); + handler(trace, evsel, event, sample); } return err; @@ -1990,7 +1995,7 @@ again: } handler = evsel->handler; - handler(trace, evsel, &sample); + handler(trace, evsel, event, &sample); next_event: perf_evlist__mmap_consume(evlist, i); -- cgit v1.1 From 598d02c5a07b60e5c824184cdaf697b70f3c452a Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Thu, 26 Jun 2014 20:14:25 +0400 Subject: perf trace: Add support for pagefault tracing This patch adds optional pagefault tracing support to 'perf trace'. Using -F/--pf option user can specify whether he wants minor, major or all pagefault events to be traced. This patch adds only live mode, record and replace will come in a separate patch. Example output: 1756272.905 ( 0.000 ms): curl/5937 majfault [0x7fa7261978b6] => /usr/lib/x86_64-linux-gnu/libkrb5.so.26.0.0@0x85288 (d.) 1862866.036 ( 0.000 ms): wget/8460 majfault [__clear_user+0x3f] => 0x659cb4 (?k) Signed-off-by: Stanislav Fomichev Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403799268-1367-3-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-trace.txt | 39 ++++++++++ tools/perf/builtin-trace.c | 125 +++++++++++++++++++++++++++++++- 2 files changed, 163 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-trace.txt b/tools/perf/Documentation/perf-trace.txt index fae38d9..72397d9 100644 --- a/tools/perf/Documentation/perf-trace.txt +++ b/tools/perf/Documentation/perf-trace.txt @@ -107,6 +107,45 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs. Show tool stats such as number of times fd->pathname was discovered thru hooking the open syscall return + vfs_getname or via reading /proc/pid/fd, etc. +-F=[all|min|maj]:: +--pf=[all|min|maj]:: + Trace pagefaults. Optionally, you can specify whether you want minor, + major or all pagefaults. Default value is maj. + +PAGEFAULTS +---------- + +When tracing pagefaults, the format of the trace is as follows: + +fault [+] => (). + +- min/maj indicates whether fault event is minor or major; +- ip.symbol shows symbol for instruction pointer (the code that generated the + fault); if no debug symbols available, perf trace will print raw IP; +- addr.dso shows DSO for the faulted address; +- map type is either 'd' for non-executable maps or 'x' for executable maps; +- addr level is either 'k' for kernel dso or '.' for user dso. + +For symbols resolution you may need to install debugging symbols. + +Please be aware that duration is currently always 0 and doesn't reflect actual +time it took for fault to be handled! + +When --verbose specified, perf trace tries to print all available information +for both IP and fault address in the form of dso@symbol+offset. + +EXAMPLES +-------- + +Trace syscalls, major and minor pagefaults: + + $ perf trace -F all + + 1416.547 ( 0.000 ms): python/20235 majfault [CRYPTO_push_info_+0x0] => /lib/x86_64-linux-gnu/libcrypto.so.1.0.0@0x61be0 (x.) + + As you can see, there was major pagefault in python process, from + CRYPTO_push_info_ routine which faulted somewhere in libcrypto.so. + SEE ALSO -------- linkperf:perf-record[1], linkperf:perf-script[1] diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 4a9e26b..1985c3b 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1178,6 +1178,9 @@ fail: return NULL; } +#define TRACE_PFMAJ (1 << 0) +#define TRACE_PFMIN (1 << 1) + struct trace { struct perf_tool tool; struct { @@ -1212,6 +1215,7 @@ struct trace { bool summary_only; bool show_comm; bool show_tool_stats; + int trace_pgfaults; }; static int trace__set_fd_pathname(struct thread *thread, int fd, const char *pathname) @@ -1773,6 +1777,68 @@ out_dump: return 0; } +static void print_location(FILE *f, struct perf_sample *sample, + struct addr_location *al, + bool print_dso, bool print_sym) +{ + + if ((verbose || print_dso) && al->map) + fprintf(f, "%s@", al->map->dso->long_name); + + if ((verbose || print_sym) && al->sym) + fprintf(f, "%s+0x%lx", al->sym->name, + al->addr - al->sym->start); + else if (al->map) + fprintf(f, "0x%lx", al->addr); + else + fprintf(f, "0x%lx", sample->addr); +} + +static int trace__pgfault(struct trace *trace, + struct perf_evsel *evsel, + union perf_event *event, + struct perf_sample *sample) +{ + struct thread *thread; + u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + struct addr_location al; + char map_type = 'd'; + + thread = machine__findnew_thread(trace->host, sample->pid, sample->tid); + + thread__find_addr_location(thread, trace->host, cpumode, MAP__FUNCTION, + sample->ip, &al); + + trace__fprintf_entry_head(trace, thread, 0, sample->time, trace->output); + + fprintf(trace->output, "%sfault [", + evsel->attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ ? + "maj" : "min"); + + print_location(trace->output, sample, &al, false, true); + + fprintf(trace->output, "] => "); + + thread__find_addr_location(thread, trace->host, cpumode, MAP__VARIABLE, + sample->addr, &al); + + if (!al.map) { + thread__find_addr_location(thread, trace->host, cpumode, + MAP__FUNCTION, sample->addr, &al); + + if (al.map) + map_type = 'x'; + else + map_type = '?'; + } + + print_location(trace->output, sample, &al, true, false); + + fprintf(trace->output, " (%c%c)\n", map_type, al.level); + + return 0; +} + static bool skip_sample(struct trace *trace, struct perf_sample *sample) { if ((trace->pid_list && intlist__find(trace->pid_list, sample->pid)) || @@ -1887,6 +1953,30 @@ static void perf_evlist__add_vfs_getname(struct perf_evlist *evlist) perf_evlist__add(evlist, evsel); } +static int perf_evlist__add_pgfault(struct perf_evlist *evlist, + u64 config) +{ + struct perf_evsel *evsel; + struct perf_event_attr attr = { + .type = PERF_TYPE_SOFTWARE, + .mmap_data = 1, + .sample_period = 1, + }; + + attr.config = config; + + event_attr_init(&attr); + + evsel = perf_evsel__new(&attr); + if (!evsel) + return -ENOMEM; + + evsel->handler = trace__pgfault; + perf_evlist__add(evlist, evsel); + + return 0; +} + static int trace__run(struct trace *trace, int argc, const char **argv) { struct perf_evlist *evlist = perf_evlist__new(); @@ -1907,6 +1997,14 @@ static int trace__run(struct trace *trace, int argc, const char **argv) perf_evlist__add_vfs_getname(evlist); + if ((trace->trace_pgfaults & TRACE_PFMAJ) && + perf_evlist__add_pgfault(evlist, PERF_COUNT_SW_PAGE_FAULTS_MAJ)) + goto out_error_tp; + + if ((trace->trace_pgfaults & TRACE_PFMIN) && + perf_evlist__add_pgfault(evlist, PERF_COUNT_SW_PAGE_FAULTS_MIN)) + goto out_error_tp; + if (trace->sched && perf_evlist__add_newtp(evlist, "sched", "sched_stat_runtime", trace__sched_stat_runtime)) @@ -1987,7 +2085,8 @@ again: goto next_event; } - if (sample.raw_data == NULL) { + if (evsel->attr.type == PERF_TYPE_TRACEPOINT && + sample.raw_data == NULL) { fprintf(trace->output, "%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n", perf_evsel__name(evsel), sample.tid, sample.cpu, sample.raw_size); @@ -2269,6 +2368,23 @@ static int trace__open_output(struct trace *trace, const char *filename) return trace->output == NULL ? -errno : 0; } +static int parse_pagefaults(const struct option *opt, const char *str, + int unset __maybe_unused) +{ + int *trace_pgfaults = opt->value; + + if (strcmp(str, "all") == 0) + *trace_pgfaults |= TRACE_PFMAJ | TRACE_PFMIN; + else if (strcmp(str, "maj") == 0) + *trace_pgfaults |= TRACE_PFMAJ; + else if (strcmp(str, "min") == 0) + *trace_pgfaults |= TRACE_PFMIN; + else + return -1; + + return 0; +} + int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) { const char * const trace_usage[] = { @@ -2335,6 +2451,8 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) "Show only syscall summary with statistics"), OPT_BOOLEAN('S', "with-summary", &trace.summary, "Show all syscalls and summary with statistics"), + OPT_CALLBACK_DEFAULT('F', "pf", &trace.trace_pgfaults, "all|maj|min", + "Trace pagefaults", parse_pagefaults, "maj"), OPT_END() }; int err; @@ -2349,6 +2467,11 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) if (trace.summary_only) trace.summary = trace.summary_only; + if (trace.trace_pgfaults) { + trace.opts.sample_address = true; + trace.opts.sample_time = true; + } + if (output_name != NULL) { err = trace__open_output(&trace, output_name); if (err < 0) { -- cgit v1.1 From 1e28fe0a4ff8680d5a0fb84995fd2444dac19cc4 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Thu, 26 Jun 2014 20:14:26 +0400 Subject: perf trace: Add pagefaults record and replay support Previous commit added live pagefault trace support, this one adds record and replay support. Example: [root@zoo /]# echo 1 > /proc/sys/vm/drop_caches ; trace -F all record -a sleep 10 [ perf record: Woken up 0 times to write data ] [ perf record: Captured and wrote 1029.722 MB perf.data (~44989242 samples) ] [root@zoo /]# ls -la perf.data -rw-------. 1 root root 1083921722 Jun 26 17:44 perf.data [root@zoo /]# perf evlist raw_syscalls:sys_enter raw_syscalls:sys_exit major-faults minor-faults [root@zoo /]# trace -i perf.data | grep -v trace\/ | tail -15 156.137 ( 0.000 ms): perl/18476 minfault [0xb4243] => 0x0 (?.) 156.139 ( 0.000 ms): perl/18476 minfault [Perl_sv_clear+0x123] => 0x0 (?.) 156.140 ( 0.000 ms): perl/18476 minfault [Perl_sv_clear+0xc4] => 0x0 (?.) 156.144 ( 0.000 ms): perl/18476 minfault [_int_free+0xda] => 0x0 (?.) 156.151 ( 0.000 ms): perl/18476 minfault [_int_free+0x1df] => 0x0 (?.) 156.158 ( 0.000 ms): perl/18476 minfault [0xb4243] => 0x0 (?.) 156.161 ( 0.000 ms): perl/18476 minfault [0xb4243] => 0x0 (?.) 156.168 ( 0.000 ms): perl/18476 minfault [0xb4243] => 0x0 (?.) 156.172 ( 0.000 ms): perl/18476 minfault [0xb4243] => 0x0 (?.) 156.173 ( 0.000 ms): perl/18476 minfault [_int_free+0xda] => 0x0 (?.) 156.183 ( 0.000 ms): perl/18476 minfault [Perl_hfree_next_entry+0xb4] => 0x0 (?.) 156.197 ( 0.000 ms): perl/18476 minfault [_int_free+0x1df] => 0x0 (?.) 156.216 ( 0.000 ms): perl/18476 minfault [Perl_sv_clear+0x123] => 0x0 (?.) 156.221 ( 0.000 ms): perl/18476 minfault [Perl_sv_clear+0x123] => 0x0 (?.) [root@zoo /]# Signed-off-by: Stanislav Fomichev Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403799268-1367-4-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 63 +++++++++++++++++++++++++++++++++------------- 1 file changed, 45 insertions(+), 18 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 1985c3b..0b58e24 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1897,7 +1897,7 @@ static int parse_target_str(struct trace *trace) return 0; } -static int trace__record(int argc, const char **argv) +static int trace__record(struct trace *trace, int argc, const char **argv) { unsigned int rec_argc, i, j; const char **rec_argv; @@ -1906,34 +1906,52 @@ static int trace__record(int argc, const char **argv) "-R", "-m", "1024", "-c", "1", - "-e", }; + const char * const sc_args[] = { "-e", }; + unsigned int sc_args_nr = ARRAY_SIZE(sc_args); + const char * const majpf_args[] = { "-e", "major-faults" }; + unsigned int majpf_args_nr = ARRAY_SIZE(majpf_args); + const char * const minpf_args[] = { "-e", "minor-faults" }; + unsigned int minpf_args_nr = ARRAY_SIZE(minpf_args); + /* +1 is for the event string below */ - rec_argc = ARRAY_SIZE(record_args) + 1 + argc; + rec_argc = ARRAY_SIZE(record_args) + sc_args_nr + 1 + + majpf_args_nr + minpf_args_nr + argc; rec_argv = calloc(rec_argc + 1, sizeof(char *)); if (rec_argv == NULL) return -ENOMEM; + j = 0; for (i = 0; i < ARRAY_SIZE(record_args); i++) - rec_argv[i] = record_args[i]; + rec_argv[j++] = record_args[i]; + + for (i = 0; i < sc_args_nr; i++) + rec_argv[j++] = sc_args[i]; /* event string may be different for older kernels - e.g., RHEL6 */ if (is_valid_tracepoint("raw_syscalls:sys_enter")) - rec_argv[i] = "raw_syscalls:sys_enter,raw_syscalls:sys_exit"; + rec_argv[j++] = "raw_syscalls:sys_enter,raw_syscalls:sys_exit"; else if (is_valid_tracepoint("syscalls:sys_enter")) - rec_argv[i] = "syscalls:sys_enter,syscalls:sys_exit"; + rec_argv[j++] = "syscalls:sys_enter,syscalls:sys_exit"; else { pr_err("Neither raw_syscalls nor syscalls events exist.\n"); return -1; } - i++; - for (j = 0; j < (unsigned int)argc; j++, i++) - rec_argv[i] = argv[j]; + if (trace->trace_pgfaults & TRACE_PFMAJ) + for (i = 0; i < majpf_args_nr; i++) + rec_argv[j++] = majpf_args[i]; + + if (trace->trace_pgfaults & TRACE_PFMIN) + for (i = 0; i < minpf_args_nr; i++) + rec_argv[j++] = minpf_args[i]; + + for (i = 0; i < (unsigned int)argc; i++) + rec_argv[j++] = argv[i]; - return cmd_record(i, rec_argv, NULL); + return cmd_record(j, rec_argv, NULL); } static size_t trace__fprintf_thread_summary(struct trace *trace, FILE *fp); @@ -2224,6 +2242,14 @@ static int trace__replay(struct trace *trace) goto out; } + evlist__for_each(session->evlist, evsel) { + if (evsel->attr.type == PERF_TYPE_SOFTWARE && + (evsel->attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ || + evsel->attr.config == PERF_COUNT_SW_PAGE_FAULTS_MIN || + evsel->attr.config == PERF_COUNT_SW_PAGE_FAULTS)) + evsel->handler = trace__pgfault; + } + err = parse_target_str(trace); if (err != 0) goto out; @@ -2458,20 +2484,21 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) int err; char bf[BUFSIZ]; - if ((argc > 1) && (strcmp(argv[1], "record") == 0)) - return trace__record(argc-2, &argv[2]); - - argc = parse_options(argc, argv, trace_options, trace_usage, 0); - - /* summary_only implies summary option, but don't overwrite summary if set */ - if (trace.summary_only) - trace.summary = trace.summary_only; + argc = parse_options(argc, argv, trace_options, trace_usage, + PARSE_OPT_STOP_AT_NON_OPTION); if (trace.trace_pgfaults) { trace.opts.sample_address = true; trace.opts.sample_time = true; } + if ((argc >= 1) && (strcmp(argv[0], "record") == 0)) + return trace__record(&trace, argc-1, &argv[1]); + + /* summary_only implies summary option, but don't overwrite summary if set */ + if (trace.summary_only) + trace.summary = trace.summary_only; + if (output_name != NULL) { err = trace__open_output(&trace, output_name); if (err < 0) { -- cgit v1.1 From e281a9606d7073c517f2571e83faaff029ddc1cf Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Thu, 26 Jun 2014 20:14:28 +0400 Subject: perf trace: Add possibility to switch off syscall events Currently, we may either trace syscalls or syscalls+pagefaults. We'd like to be able to trace *only* pagefaults and this commit implements this feature. Example: [root@zoo /]# echo 1 > /proc/sys/vm/drop_caches ; trace --no-syscalls -F -p `pidof xchat` 0.000 ( 0.000 ms): xchat/4574 majfault [g_unichar_get_script+0x11] => /usr/lib64/libglib-2.0.so.0.3800.2@0xc403b (x.) 0.202 ( 0.000 ms): xchat/4574 majfault [_cairo_hash_table_lookup+0x53] => 0x2280ff0 (?.) 20.854 ( 0.000 ms): xchat/4574 majfault [gdk_cairo_set_source_pixbuf+0x110] => /usr/bin/xchat@0x6da1f (x.) 1022.000 ( 0.000 ms): xchat/4574 majfault [__memcpy_sse2_unaligned+0x29] => 0x7ff5a8ca0400 (?.) ^C[root@zoo /]# Below we can see malloc calls, 'trace' reading symbol tables in libraries to resolve symbols, etc. [root@zoo /]# echo 1 > /proc/sys/vm/drop_caches ; trace --no-syscalls -F all --cpu 1 sleep 10 0.000 ( 0.000 ms): chrome/26589 minfault [0x1b53129] => /tmp/perf-26589.map@0x33cbcbf7f000 (x.) 96.477 ( 0.000 ms): libvirtd/947 minfault [copy_user_enhanced_fast_string+0x5] => 0x7f7685bba000 (?k) 113.164 ( 0.000 ms): Xorg/1063 minfault [0x786da] => 0x7fce52882a3c (?.) 7162.801 ( 0.000 ms): chrome/3747 minfault [0x8e1a89] => 0xfcaefed0008 (?.) 7773.138 ( 0.000 ms): chrome/3886 minfault [0x8e1a89] => 0xfcb0ce28008 (?.) 7992.022 ( 0.000 ms): chrome/26574 minfault [0x1b5a708] => 0x3de7b5fc5000 (?.) 8108.949 ( 0.000 ms): qemu-system-x8/4537 majfault [_int_malloc+0xee] => 0x7faffc466d60 (?.) 8108.975 ( 0.000 ms): qemu-system-x8/4537 minfault [_int_malloc+0x102] => 0x7faffc466d60 (?.) 8148.174 ( 0.000 ms): qemu-system-x8/4537 minfault [_int_malloc+0x102] => 0x7faffc4eb500 (?.) 8270.855 ( 0.000 ms): chrome/26245 minfault [do_bo_emit_reloc+0xdb] => 0x45d092bc004 (?.) 8270.869 ( 0.000 ms): chrome/26245 minfault [do_bo_emit_reloc+0x108] => 0x45d09150000 (?.) no symbols found in /usr/lib64/libspice-server.so.1.9.0, maybe install a debug package? 8273.831 ( 0.000 ms): trace/20198 majfault [__memcmp_sse4_1+0xbc6] => /usr/lib64/libspice-server.so.1.9.0@0xdf000 (d.) 8275.121 ( 0.000 ms): trace/20198 minfault [dso__load+0x38] => 0x14fe756 (?.) no symbols found in /usr/lib64/libelf-0.158.so, maybe install a debug package? 8275.142 ( 0.000 ms): trace/20198 minfault [__memcmp_sse4_1+0xbc6] => /usr/lib64/libelf-0.158.so@0x0 (d.) [root@zoo /]# Signed-off-by: Stanislav Fomichev Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403799268-1367-6-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-trace.txt | 7 ++++ tools/perf/builtin-trace.c | 58 ++++++++++++++++++--------------- 2 files changed, 39 insertions(+), 26 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-trace.txt b/tools/perf/Documentation/perf-trace.txt index 72397d9..02aac83 100644 --- a/tools/perf/Documentation/perf-trace.txt +++ b/tools/perf/Documentation/perf-trace.txt @@ -112,6 +112,9 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs. Trace pagefaults. Optionally, you can specify whether you want minor, major or all pagefaults. Default value is maj. +--syscalls:: + Trace system calls. This options is enabled by default. + PAGEFAULTS ---------- @@ -137,6 +140,10 @@ for both IP and fault address in the form of dso@symbol+offset. EXAMPLES -------- +Trace only major pagefaults: + + $ perf trace --no-syscalls -F + Trace syscalls, major and minor pagefaults: $ perf trace -F all diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 0b58e24..dc7a694 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1215,6 +1215,7 @@ struct trace { bool summary_only; bool show_comm; bool show_tool_stats; + bool trace_syscalls; int trace_pgfaults; }; @@ -1927,17 +1928,19 @@ static int trace__record(struct trace *trace, int argc, const char **argv) for (i = 0; i < ARRAY_SIZE(record_args); i++) rec_argv[j++] = record_args[i]; - for (i = 0; i < sc_args_nr; i++) - rec_argv[j++] = sc_args[i]; - - /* event string may be different for older kernels - e.g., RHEL6 */ - if (is_valid_tracepoint("raw_syscalls:sys_enter")) - rec_argv[j++] = "raw_syscalls:sys_enter,raw_syscalls:sys_exit"; - else if (is_valid_tracepoint("syscalls:sys_enter")) - rec_argv[j++] = "syscalls:sys_enter,syscalls:sys_exit"; - else { - pr_err("Neither raw_syscalls nor syscalls events exist.\n"); - return -1; + if (trace->trace_syscalls) { + for (i = 0; i < sc_args_nr; i++) + rec_argv[j++] = sc_args[i]; + + /* event string may be different for older kernels - e.g., RHEL6 */ + if (is_valid_tracepoint("raw_syscalls:sys_enter")) + rec_argv[j++] = "raw_syscalls:sys_enter,raw_syscalls:sys_exit"; + else if (is_valid_tracepoint("syscalls:sys_enter")) + rec_argv[j++] = "syscalls:sys_enter,syscalls:sys_exit"; + else { + pr_err("Neither raw_syscalls nor syscalls events exist.\n"); + return -1; + } } if (trace->trace_pgfaults & TRACE_PFMAJ) @@ -2010,10 +2013,13 @@ static int trace__run(struct trace *trace, int argc, const char **argv) goto out; } - if (perf_evlist__add_syscall_newtp(evlist, trace__sys_enter, trace__sys_exit)) + if (trace->trace_syscalls && + perf_evlist__add_syscall_newtp(evlist, trace__sys_enter, + trace__sys_exit)) goto out_error_tp; - perf_evlist__add_vfs_getname(evlist); + if (trace->trace_syscalls) + perf_evlist__add_vfs_getname(evlist); if ((trace->trace_pgfaults & TRACE_PFMAJ) && perf_evlist__add_pgfault(evlist, PERF_COUNT_SW_PAGE_FAULTS_MAJ)) @@ -2215,13 +2221,10 @@ static int trace__replay(struct trace *trace) if (evsel == NULL) evsel = perf_evlist__find_tracepoint_by_name(session->evlist, "syscalls:sys_enter"); - if (evsel == NULL) { - pr_err("Data file does not have raw_syscalls:sys_enter event\n"); - goto out; - } - if (perf_evsel__init_syscall_tp(evsel, trace__sys_enter) < 0 || - perf_evsel__init_sc_tp_ptr_field(evsel, args)) { + if (evsel && + (perf_evsel__init_syscall_tp(evsel, trace__sys_enter) < 0 || + perf_evsel__init_sc_tp_ptr_field(evsel, args))) { pr_err("Error during initialize raw_syscalls:sys_enter event\n"); goto out; } @@ -2231,13 +2234,9 @@ static int trace__replay(struct trace *trace) if (evsel == NULL) evsel = perf_evlist__find_tracepoint_by_name(session->evlist, "syscalls:sys_exit"); - if (evsel == NULL) { - pr_err("Data file does not have raw_syscalls:sys_exit event\n"); - goto out; - } - - if (perf_evsel__init_syscall_tp(evsel, trace__sys_exit) < 0 || - perf_evsel__init_sc_tp_uint_field(evsel, ret)) { + if (evsel && + (perf_evsel__init_syscall_tp(evsel, trace__sys_exit) < 0 || + perf_evsel__init_sc_tp_uint_field(evsel, ret))) { pr_err("Error during initialize raw_syscalls:sys_exit event\n"); goto out; } @@ -2440,6 +2439,7 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) }, .output = stdout, .show_comm = true, + .trace_syscalls = true, }; const char *output_name = NULL; const char *ev_qualifier_str = NULL; @@ -2479,6 +2479,7 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) "Show all syscalls and summary with statistics"), OPT_CALLBACK_DEFAULT('F', "pf", &trace.trace_pgfaults, "all|maj|min", "Trace pagefaults", parse_pagefaults, "maj"), + OPT_BOOLEAN(0, "syscalls", &trace.trace_syscalls, "Trace syscalls"), OPT_END() }; int err; @@ -2499,6 +2500,11 @@ int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused) if (trace.summary_only) trace.summary = trace.summary_only; + if (!trace.trace_syscalls && !trace.trace_pgfaults) { + pr_err("Please specify something to trace.\n"); + return -1; + } + if (output_name != NULL) { err = trace__open_output(&trace, output_name); if (err < 0) { -- cgit v1.1 From 21da83fb6c3ff67e969b8770bbb33138c9ede88e Mon Sep 17 00:00:00 2001 From: Rickard Strandqvist Date: Tue, 24 Jun 2014 13:09:10 +0200 Subject: tools lib traceevent: Fix a risk for doing free on uninitialized pointer Fix a risk of doing free on an uninitialized pointer. This was found using a static code analysis program called cppcheck. Acked-by: Namhyung Kim Signed-off-by: Rickard Strandqvist Link: http://lkml.kernel.org/r/1403608150-13037-1-git-send-email-rickard_strandqvist@spectrumdigital.se Signed-off-by: Jiri Olsa --- tools/lib/traceevent/event-parse.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/lib/traceevent/event-parse.c b/tools/lib/traceevent/event-parse.c index 93825a1..cf3a44b 100644 --- a/tools/lib/traceevent/event-parse.c +++ b/tools/lib/traceevent/event-parse.c @@ -2395,7 +2395,7 @@ process_flags(struct event_format *event, struct print_arg *arg, char **tok) { struct print_arg *field; enum event_type type; - char *token; + char *token = NULL; memset(arg, 0, sizeof(*arg)); arg->type = PRINT_FLAGS; @@ -2448,7 +2448,7 @@ process_symbols(struct event_format *event, struct print_arg *arg, char **tok) { struct print_arg *field; enum event_type type; - char *token; + char *token = NULL; memset(arg, 0, sizeof(*arg)); arg->type = PRINT_SYMBOL; @@ -2487,7 +2487,7 @@ process_hex(struct event_format *event, struct print_arg *arg, char **tok) { struct print_arg *field; enum event_type type; - char *token; + char *token = NULL; memset(arg, 0, sizeof(*arg)); arg->type = PRINT_HEX; -- cgit v1.1 From a60335ba32981db5bc057b35782644e9e2436407 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Wed, 25 Jun 2014 08:49:03 -0700 Subject: perf tools powerpc: Adjust callchain based on DWARF debug info When saving the callchain on Power, the kernel conservatively saves excess entries in the callchain. A few of these entries are needed in some cases but not others. We should use the DWARF debug information to determine when the entries are needed. Eg: the value in the link register (LR) is needed only when it holds the return address of a function. At other times it must be ignored. If the unnecessary entries are not ignored, we end up with duplicate arcs in the call-graphs. Use the DWARF debug information to determine if any callchain entries should be ignored when building call-graphs. Callgraph before the patch: 14.67% 2234 sprintft libc-2.18.so [.] __random | --- __random | |--61.12%-- __random | | | |--97.15%-- rand | | do_my_sprintf | | main | | generic_start_main.isra.0 | | __libc_start_main | | 0x0 | | | --2.85%-- do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --38.88%-- rand | |--94.01%-- rand | do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --5.99%-- do_my_sprintf main generic_start_main.isra.0 __libc_start_main 0x0 Callgraph after the patch: 14.67% 2234 sprintft libc-2.18.so [.] __random | --- __random | |--95.93%-- rand | do_my_sprintf | main | generic_start_main.isra.0 | __libc_start_main | 0x0 | --4.07%-- do_my_sprintf main generic_start_main.isra.0 __libc_start_main 0x0 TODO: For split-debug info objects like glibc, we can only determine the call-frame-address only when both .eh_frame and .debug_info sections are available. We should be able to determin the CFA even without the .eh_frame section. Fix suggested by Anton Blanchard. Thanks to valuable input on DWARF debug information from Ulrich Weigand. Reported-by: Maynard Johnson Tested-by: Maynard Johnson Signed-off-by: Sukadev Bhattiprolu Link: http://lkml.kernel.org/r/20140625154903.GA29607@us.ibm.com Signed-off-by: Jiri Olsa --- tools/perf/arch/powerpc/Makefile | 1 + tools/perf/arch/powerpc/util/skip-callchain-idx.c | 266 ++++++++++++++++++++++ tools/perf/config/Makefile | 4 + tools/perf/util/callchain.h | 13 ++ tools/perf/util/machine.c | 18 +- 5 files changed, 300 insertions(+), 2 deletions(-) create mode 100644 tools/perf/arch/powerpc/util/skip-callchain-idx.c (limited to 'tools') diff --git a/tools/perf/arch/powerpc/Makefile b/tools/perf/arch/powerpc/Makefile index 744e629..b92219b 100644 --- a/tools/perf/arch/powerpc/Makefile +++ b/tools/perf/arch/powerpc/Makefile @@ -3,3 +3,4 @@ PERF_HAVE_DWARF_REGS := 1 LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/dwarf-regs.o endif LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/header.o +LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/skip-callchain-idx.o diff --git a/tools/perf/arch/powerpc/util/skip-callchain-idx.c b/tools/perf/arch/powerpc/util/skip-callchain-idx.c new file mode 100644 index 0000000..a7c23a4 --- /dev/null +++ b/tools/perf/arch/powerpc/util/skip-callchain-idx.c @@ -0,0 +1,266 @@ +/* + * Use DWARF Debug information to skip unnecessary callchain entries. + * + * Copyright (C) 2014 Sukadev Bhattiprolu, IBM Corporation. + * Copyright (C) 2014 Ulrich Weigand, IBM Corporation. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ +#include +#include +#include + +#include "util/thread.h" +#include "util/callchain.h" + +/* + * When saving the callchain on Power, the kernel conservatively saves + * excess entries in the callchain. A few of these entries are needed + * in some cases but not others. If the unnecessary entries are not + * ignored, we end up with duplicate arcs in the call-graphs. Use + * DWARF debug information to skip over any unnecessary callchain + * entries. + * + * See function header for arch_adjust_callchain() below for more details. + * + * The libdwfl code in this file is based on code from elfutils + * (libdwfl/argp-std.c, libdwfl/tests/addrcfi.c, etc). + */ +static char *debuginfo_path; + +static const Dwfl_Callbacks offline_callbacks = { + .debuginfo_path = &debuginfo_path, + .find_debuginfo = dwfl_standard_find_debuginfo, + .section_address = dwfl_offline_section_address, +}; + + +/* + * Use the DWARF expression for the Call-frame-address and determine + * if return address is in LR and if a new frame was allocated. + */ +static int check_return_reg(int ra_regno, Dwarf_Frame *frame) +{ + Dwarf_Op ops_mem[2]; + Dwarf_Op dummy; + Dwarf_Op *ops = &dummy; + size_t nops; + int result; + + result = dwarf_frame_register(frame, ra_regno, ops_mem, &ops, &nops); + if (result < 0) { + pr_debug("dwarf_frame_register() %s\n", dwarf_errmsg(-1)); + return -1; + } + + /* + * Check if return address is on the stack. + */ + if (nops != 0 || ops != NULL) + return 0; + + /* + * Return address is in LR. Check if a frame was allocated + * but not-yet used. + */ + result = dwarf_frame_cfa(frame, &ops, &nops); + if (result < 0) { + pr_debug("dwarf_frame_cfa() returns %d, %s\n", result, + dwarf_errmsg(-1)); + return -1; + } + + /* + * If call frame address is in r1, no new frame was allocated. + */ + if (nops == 1 && ops[0].atom == DW_OP_bregx && ops[0].number == 1 && + ops[0].number2 == 0) + return 1; + + /* + * A new frame was allocated but has not yet been used. + */ + return 2; +} + +/* + * Get the DWARF frame from the .eh_frame section. + */ +static Dwarf_Frame *get_eh_frame(Dwfl_Module *mod, Dwarf_Addr pc) +{ + int result; + Dwarf_Addr bias; + Dwarf_CFI *cfi; + Dwarf_Frame *frame; + + cfi = dwfl_module_eh_cfi(mod, &bias); + if (!cfi) { + pr_debug("%s(): no CFI - %s\n", __func__, dwfl_errmsg(-1)); + return NULL; + } + + result = dwarf_cfi_addrframe(cfi, pc, &frame); + if (result) { + pr_debug("%s(): %s\n", __func__, dwfl_errmsg(-1)); + return NULL; + } + + return frame; +} + +/* + * Get the DWARF frame from the .debug_frame section. + */ +static Dwarf_Frame *get_dwarf_frame(Dwfl_Module *mod, Dwarf_Addr pc) +{ + Dwarf_CFI *cfi; + Dwarf_Addr bias; + Dwarf_Frame *frame; + int result; + + cfi = dwfl_module_dwarf_cfi(mod, &bias); + if (!cfi) { + pr_debug("%s(): no CFI - %s\n", __func__, dwfl_errmsg(-1)); + return NULL; + } + + result = dwarf_cfi_addrframe(cfi, pc, &frame); + if (result) { + pr_debug("%s(): %s\n", __func__, dwfl_errmsg(-1)); + return NULL; + } + + return frame; +} + +/* + * Return: + * 0 if return address for the program counter @pc is on stack + * 1 if return address is in LR and no new stack frame was allocated + * 2 if return address is in LR and a new frame was allocated (but not + * yet used) + * -1 in case of errors + */ +static int check_return_addr(const char *exec_file, Dwarf_Addr pc) +{ + int rc = -1; + Dwfl *dwfl; + Dwfl_Module *mod; + Dwarf_Frame *frame; + int ra_regno; + Dwarf_Addr start = pc; + Dwarf_Addr end = pc; + bool signalp; + + dwfl = dwfl_begin(&offline_callbacks); + if (!dwfl) { + pr_debug("dwfl_begin() failed: %s\n", dwarf_errmsg(-1)); + return -1; + } + + if (dwfl_report_offline(dwfl, "", exec_file, -1) == NULL) { + pr_debug("dwfl_report_offline() failed %s\n", dwarf_errmsg(-1)); + goto out; + } + + mod = dwfl_addrmodule(dwfl, pc); + if (!mod) { + pr_debug("dwfl_addrmodule() failed, %s\n", dwarf_errmsg(-1)); + goto out; + } + + /* + * To work with split debug info files (eg: glibc), check both + * .eh_frame and .debug_frame sections of the ELF header. + */ + frame = get_eh_frame(mod, pc); + if (!frame) { + frame = get_dwarf_frame(mod, pc); + if (!frame) + goto out; + } + + ra_regno = dwarf_frame_info(frame, &start, &end, &signalp); + if (ra_regno < 0) { + pr_debug("Return address register unavailable: %s\n", + dwarf_errmsg(-1)); + goto out; + } + + rc = check_return_reg(ra_regno, frame); + +out: + dwfl_end(dwfl); + return rc; +} + +/* + * The callchain saved by the kernel always includes the link register (LR). + * + * 0: PERF_CONTEXT_USER + * 1: Program counter (Next instruction pointer) + * 2: LR value + * 3: Caller's caller + * 4: ... + * + * The value in LR is only needed when it holds a return address. If the + * return address is on the stack, we should ignore the LR value. + * + * Further, when the return address is in the LR, if a new frame was just + * allocated but the LR was not saved into it, then the LR contains the + * caller, slot 4: contains the caller's caller and the contents of slot 3: + * (chain->ips[3]) is undefined and must be ignored. + * + * Use DWARF debug information to determine if any entries need to be skipped. + * + * Return: + * index: of callchain entry that needs to be ignored (if any) + * -1 if no entry needs to be ignored or in case of errors + */ +int arch_skip_callchain_idx(struct machine *machine, struct thread *thread, + struct ip_callchain *chain) +{ + struct addr_location al; + struct dso *dso = NULL; + int rc; + u64 ip; + u64 skip_slot = -1; + + if (chain->nr < 3) + return skip_slot; + + ip = chain->ips[2]; + + thread__find_addr_location(thread, machine, PERF_RECORD_MISC_USER, + MAP__FUNCTION, ip, &al); + + if (al.map) + dso = al.map->dso; + + if (!dso) { + pr_debug("%" PRIx64 " dso is NULL\n", ip); + return skip_slot; + } + + rc = check_return_addr(dso->long_name, ip); + + pr_debug("DSO %s, nr %" PRIx64 ", ip 0x%" PRIx64 "rc %d\n", + dso->long_name, chain->nr, ip, rc); + + if (rc == 0) { + /* + * Return address on stack. Ignore LR value in callchain + */ + skip_slot = 2; + } else if (rc == 2) { + /* + * New frame allocated but return address still in LR. + * Ignore the caller's caller entry in callchain. + */ + skip_slot = 3; + } + return skip_slot; +} diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile index f30ac5e..346bdb6 100644 --- a/tools/perf/config/Makefile +++ b/tools/perf/config/Makefile @@ -48,6 +48,10 @@ ifneq ($(ARCH),$(filter $(ARCH),x86 arm)) NO_LIBDW_DWARF_UNWIND := 1 endif +ifeq ($(ARCH),powerpc) + CFLAGS += -DHAVE_SKIP_CALLCHAIN_IDX +endif + ifeq ($(LIBUNWIND_LIBS),) NO_LIBUNWIND := 1 else diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index 8f84423..da43619 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -176,4 +176,17 @@ static inline void callchain_cursor_snapshot(struct callchain_cursor *dest, dest->first = src->curr; dest->nr -= src->pos; } + +#ifdef HAVE_SKIP_CALLCHAIN_IDX +extern int arch_skip_callchain_idx(struct machine *machine, + struct thread *thread, struct ip_callchain *chain); +#else +static inline int arch_skip_callchain_idx(struct machine *machine __maybe_unused, + struct thread *thread __maybe_unused, + struct ip_callchain *chain __maybe_unused) +{ + return -1; +} +#endif + #endif /* __PERF_CALLCHAIN_H */ diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index c73e1fc..e9b943a 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1281,7 +1281,9 @@ static int machine__resolve_callchain_sample(struct machine *machine, u8 cpumode = PERF_RECORD_MISC_USER; int chain_nr = min(max_stack, (int)chain->nr); int i; + int j; int err; + int skip_idx __maybe_unused; callchain_cursor_reset(&callchain_cursor); @@ -1290,14 +1292,26 @@ static int machine__resolve_callchain_sample(struct machine *machine, return 0; } + /* + * Based on DWARF debug information, some architectures skip + * a callchain entry saved by the kernel. + */ + skip_idx = arch_skip_callchain_idx(machine, thread, chain); + for (i = 0; i < chain_nr; i++) { u64 ip; struct addr_location al; if (callchain_param.order == ORDER_CALLEE) - ip = chain->ips[i]; + j = i; else - ip = chain->ips[chain->nr - i - 1]; + j = chain->nr - i - 1; + +#ifdef HAVE_SKIP_CALLCHAIN_IDX + if (j == skip_idx) + continue; +#endif + ip = chain->ips[j]; if (ip >= PERF_CONTEXT_MAX) { switch (ip) { -- cgit v1.1 From d180ac14a95d738c2d2622db82c2212a8f998200 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 20 Jun 2014 10:56:34 +0200 Subject: perf tools: Fix wrong condition for allocation failure Check real allocated pointer for NULL. Cc: Arnaldo Carvalho de Melo Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-5rfzbalwjphmdzzil74eazyl@git.kernel.org Signed-off-by: Jiri Olsa --- tools/perf/builtin-stat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 65a151e..3e80aa1 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -184,7 +184,7 @@ static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel) static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel) { evsel->priv = zalloc(sizeof(struct perf_stat)); - if (evsel == NULL) + if (evsel->priv == NULL) return -ENOMEM; perf_evsel__reset_stat_priv(evsel); return 0; -- cgit v1.1 From 33058b948e545a911e388e69b8be7274da158fb6 Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior Date: Tue, 27 May 2014 18:14:33 +0200 Subject: perf script: Move the number processing into its own function I was going to change something here and the result was so much on the right side of the screen that I decided to move that piece into its own function. This patch should make no function change except the moving the code into its own function. Signed-off-by: Sebastian Andrzej Siewior Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1401207274-8170-1-git-send-email-bigeasy@linutronix.de Signed-off-by: Jiri Olsa --- .../util/scripting-engines/trace-event-python.c | 38 +++++++++++++--------- 1 file changed, 23 insertions(+), 15 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index 1c41932..99c2536 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -231,6 +231,28 @@ static inline struct event_format *find_cache_event(struct perf_evsel *evsel) return event; } +static PyObject *get_field_numeric_entry(struct event_format *event, + struct format_field *field, void *data) +{ + PyObject *obj; + unsigned long long val; + + val = read_size(event, data + field->offset, field->size); + if (field->flags & FIELD_IS_SIGNED) { + if ((long long)val >= LONG_MIN && + (long long)val <= LONG_MAX) + obj = PyInt_FromLong(val); + else + obj = PyLong_FromLongLong(val); + } else { + if (val <= LONG_MAX) + obj = PyInt_FromLong(val); + else + obj = PyLong_FromUnsignedLongLong(val); + } + return obj; +} + static void python_process_tracepoint(struct perf_sample *sample, struct perf_evsel *evsel, struct thread *thread, @@ -239,7 +261,6 @@ static void python_process_tracepoint(struct perf_sample *sample, PyObject *handler, *retval, *context, *t, *obj, *dict = NULL; static char handler_name[256]; struct format_field *field; - unsigned long long val; unsigned long s, ns; struct event_format *event; unsigned n = 0; @@ -303,20 +324,7 @@ static void python_process_tracepoint(struct perf_sample *sample, offset = field->offset; obj = PyString_FromString((char *)data + offset); } else { /* FIELD_IS_NUMERIC */ - val = read_size(event, data + field->offset, - field->size); - if (field->flags & FIELD_IS_SIGNED) { - if ((long long)val >= LONG_MIN && - (long long)val <= LONG_MAX) - obj = PyInt_FromLong(val); - else - obj = PyLong_FromLongLong(val); - } else { - if (val <= LONG_MAX) - obj = PyInt_FromLong(val); - else - obj = PyLong_FromUnsignedLongLong(val); - } + obj = get_field_numeric_entry(event, field, data); } if (handler) PyTuple_SetItem(t, n++, obj); -- cgit v1.1 From 8ac631cd502d6b31fd29f6d019305305b479fa3e Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior Date: Tue, 27 May 2014 18:14:34 +0200 Subject: perf script: Handle the num array type in python properly The raw_syscalls:sys_enter tracer for instance passes has one argument named 'arg' which is an array of 6 integers. Right the python scripts gets only 0 passed as an argument. The reason is that pevent_read_number() can not handle data types of 48 and returns always 0. This patch changes this by passing num array as list of nums which fit the description. As a result python will now see a list named arg which contains 6 (integer) items. Signed-off-by: Sebastian Andrzej Siewior Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1401207274-8170-2-git-send-email-bigeasy@linutronix.de Signed-off-by: Jiri Olsa --- .../util/scripting-engines/trace-event-python.c | 43 ++++++++++++++++------ 1 file changed, 31 insertions(+), 12 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index 99c2536..e55b65a 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -234,22 +234,41 @@ static inline struct event_format *find_cache_event(struct perf_evsel *evsel) static PyObject *get_field_numeric_entry(struct event_format *event, struct format_field *field, void *data) { - PyObject *obj; + bool is_array = field->flags & FIELD_IS_ARRAY; + PyObject *obj, *list = NULL; unsigned long long val; + unsigned int item_size, n_items, i; - val = read_size(event, data + field->offset, field->size); - if (field->flags & FIELD_IS_SIGNED) { - if ((long long)val >= LONG_MIN && - (long long)val <= LONG_MAX) - obj = PyInt_FromLong(val); - else - obj = PyLong_FromLongLong(val); + if (is_array) { + list = PyList_New(field->arraylen); + item_size = field->size / field->arraylen; + n_items = field->arraylen; } else { - if (val <= LONG_MAX) - obj = PyInt_FromLong(val); - else - obj = PyLong_FromUnsignedLongLong(val); + item_size = field->size; + n_items = 1; + } + + for (i = 0; i < n_items; i++) { + + val = read_size(event, data + field->offset + i * item_size, + item_size); + if (field->flags & FIELD_IS_SIGNED) { + if ((long long)val >= LONG_MIN && + (long long)val <= LONG_MAX) + obj = PyInt_FromLong(val); + else + obj = PyLong_FromLongLong(val); + } else { + if (val <= LONG_MAX) + obj = PyInt_FromLong(val); + else + obj = PyLong_FromUnsignedLongLong(val); + } + if (is_array) + PyList_SET_ITEM(list, i, obj); } + if (is_array) + obj = list; return obj; } -- cgit v1.1 From 62c95ae33f64ecf398f490886f026a6132fb1520 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 1 Jul 2014 11:07:54 -0300 Subject: perf ui browser: Add ->rows to disambiguate from ->height The ui_browser->height is about the whole browser "window", including any header, status lines or any other space needed for some "Yes", "No", etc buttons a descendent browser, like hist_browser, may have. Since the navigation is done mostly on the ui_browser methods, it needs to know how many rows are on the screen, while details about what other components are, say, if a header (that may be composed of multiple lines, etc) is present. Besides this we'll need to add a ui_browser->refresh_dimensions() hook so that browsers like hist_browser can update ->rows in response to screen resizes, this will come in a follow up patch. This patch just adds ->rows and updates it when updating ->height, keeps using ->height for the only other widget that can come with ui_browser, the scrollbar, that goes on using all the height on the rightmost column in the screen, using ->rows for the keyboard navigation needs. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-xexmwg1mv7u03j5imn66jdak@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browser.c | 30 +++++++++++++++--------------- tools/perf/ui/browser.h | 2 +- tools/perf/ui/browsers/hists.c | 16 ++++++++-------- 3 files changed, 24 insertions(+), 24 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browser.c b/tools/perf/ui/browser.c index 9d2294e..adb294a 100644 --- a/tools/perf/ui/browser.c +++ b/tools/perf/ui/browser.c @@ -150,7 +150,7 @@ unsigned int ui_browser__rb_tree_refresh(struct ui_browser *browser) while (nd != NULL) { ui_browser__gotorc(browser, row, 0); browser->write(browser, nd, row); - if (++row == browser->height) + if (++row == browser->rows) break; nd = rb_next(nd); } @@ -166,7 +166,7 @@ bool ui_browser__is_current_entry(struct ui_browser *browser, unsigned row) void ui_browser__refresh_dimensions(struct ui_browser *browser) { browser->width = SLtt_Screen_Cols - 1; - browser->height = SLtt_Screen_Rows - 2; + browser->height = browser->rows = SLtt_Screen_Rows - 2; browser->y = 1; browser->x = 0; } @@ -389,7 +389,7 @@ int ui_browser__run(struct ui_browser *browser, int delay_secs) if (browser->index == browser->nr_entries - 1) break; ++browser->index; - if (browser->index == browser->top_idx + browser->height) { + if (browser->index == browser->top_idx + browser->rows) { ++browser->top_idx; browser->seek(browser, +1, SEEK_CUR); } @@ -405,10 +405,10 @@ int ui_browser__run(struct ui_browser *browser, int delay_secs) break; case K_PGDN: case ' ': - if (browser->top_idx + browser->height > browser->nr_entries - 1) + if (browser->top_idx + browser->rows > browser->nr_entries - 1) break; - offset = browser->height; + offset = browser->rows; if (browser->index + offset > browser->nr_entries - 1) offset = browser->nr_entries - 1 - browser->index; browser->index += offset; @@ -419,10 +419,10 @@ int ui_browser__run(struct ui_browser *browser, int delay_secs) if (browser->top_idx == 0) break; - if (browser->top_idx < browser->height) + if (browser->top_idx < browser->rows) offset = browser->top_idx; else - offset = browser->height; + offset = browser->rows; browser->index -= offset; browser->top_idx -= offset; @@ -432,7 +432,7 @@ int ui_browser__run(struct ui_browser *browser, int delay_secs) ui_browser__reset_index(browser); break; case K_END: - offset = browser->height - 1; + offset = browser->rows - 1; if (offset >= browser->nr_entries) offset = browser->nr_entries - 1; @@ -462,7 +462,7 @@ unsigned int ui_browser__list_head_refresh(struct ui_browser *browser) if (!browser->filter || !browser->filter(browser, pos)) { ui_browser__gotorc(browser, row, 0); browser->write(browser, pos, row); - if (++row == browser->height) + if (++row == browser->rows) break; } } @@ -587,7 +587,7 @@ unsigned int ui_browser__argv_refresh(struct ui_browser *browser) if (!browser->filter || !browser->filter(browser, *pos)) { ui_browser__gotorc(browser, row, 0); browser->write(browser, pos, row); - if (++row == browser->height) + if (++row == browser->rows) break; } @@ -623,7 +623,7 @@ static void __ui_browser__line_arrow_up(struct ui_browser *browser, SLsmg_set_char_set(1); - if (start < browser->top_idx + browser->height) { + if (start < browser->top_idx + browser->rows) { row = start - browser->top_idx; ui_browser__gotorc(browser, row, column); SLsmg_write_char(SLSMG_LLCORN_CHAR); @@ -633,7 +633,7 @@ static void __ui_browser__line_arrow_up(struct ui_browser *browser, if (row-- == 0) goto out; } else - row = browser->height - 1; + row = browser->rows - 1; if (end > browser->top_idx) end_row = end - browser->top_idx; @@ -675,8 +675,8 @@ static void __ui_browser__line_arrow_down(struct ui_browser *browser, } else row = 0; - if (end >= browser->top_idx + browser->height) - end_row = browser->height - 1; + if (end >= browser->top_idx + browser->rows) + end_row = browser->rows - 1; else end_row = end - browser->top_idx; @@ -684,7 +684,7 @@ static void __ui_browser__line_arrow_down(struct ui_browser *browser, SLsmg_draw_vline(end_row - row + 1); ui_browser__gotorc(browser, end_row, column); - if (end < browser->top_idx + browser->height) { + if (end < browser->top_idx + browser->rows) { SLsmg_write_char(SLSMG_LLCORN_CHAR); ui_browser__gotorc(browser, end_row, column + 1); SLsmg_write_char(SLSMG_HLINE_CHAR); diff --git a/tools/perf/ui/browser.h b/tools/perf/ui/browser.h index 03d4d62..cb8f39a 100644 --- a/tools/perf/ui/browser.h +++ b/tools/perf/ui/browser.h @@ -14,7 +14,7 @@ struct ui_browser { u64 index, top_idx; void *top, *entries; - u16 y, x, width, height; + u16 y, x, width, height, rows; int current_color; void *priv; const char *title; diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 2185091..e16aff4 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -392,10 +392,10 @@ static int hist_browser__run(struct hist_browser *browser, struct hist_entry *h = rb_entry(browser->b.top, struct hist_entry, rb_node); ui_helpline__pop(); - ui_helpline__fpush("%d: nr_ent=(%d,%d), height=%d, idx=%d, fve: idx=%d, row_off=%d, nrows=%d", + ui_helpline__fpush("%d: nr_ent=(%d,%d), rows=%d, idx=%d, fve: idx=%d, row_off=%d, nrows=%d", seq++, browser->b.nr_entries, browser->hists->nr_entries, - browser->b.height, + browser->b.rows, browser->b.index, browser->b.top_idx, h->row_offset, h->nr_rows); @@ -514,7 +514,7 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *browse slsmg_write_nstring(str, width); free(alloc_str); - if (++row == browser->b.height) + if (++row == browser->b.rows) goto out; do_next: if (folded_sign == '+') @@ -527,7 +527,7 @@ do_next: new_level, row, row_offset, is_current_entry); } - if (row == browser->b.height) + if (row == browser->b.rows) goto out; node = next; } @@ -573,7 +573,7 @@ static int hist_browser__show_callchain_node(struct hist_browser *browser, slsmg_printf("%c ", folded_sign); slsmg_write_nstring(s, width - 2); - if (++row == browser->b.height) + if (++row == browser->b.rows) goto out; } @@ -602,7 +602,7 @@ static int hist_browser__show_callchain(struct hist_browser *browser, row += hist_browser__show_callchain_node(browser, node, level, row, row_offset, is_current_entry); - if (row == browser->b.height) + if (row == browser->b.rows) break; } @@ -776,7 +776,7 @@ static int hist_browser__show_entry(struct hist_browser *browser, } else --row_offset; - if (folded_sign == '-' && row != browser->b.height) { + if (folded_sign == '-' && row != browser->b.rows) { printed += hist_browser__show_callchain(browser, &entry->sorted_chain, 1, row, &row_offset, ¤t_entry); @@ -817,7 +817,7 @@ static unsigned int hist_browser__refresh(struct ui_browser *browser) continue; row += hist_browser__show_entry(hb, h, row); - if (row == browser->height) + if (row == browser->rows) break; } -- cgit v1.1 From fa70b5d66d6e9de3331f57e434a16bc0eee980a8 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 1 Jul 2014 16:34:42 -0300 Subject: perf ui browser: Allow overriding refresh_dimensions method Some browsers, like the hist_browser, may want to be notified everytime a refresh_dimensions is needed, so that it can reset ui_browser->rows, for instance, or do some other related reaction to screen resizings. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-ielvluuemzn30bneh0zk3twi@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browser.c | 7 +++++-- tools/perf/ui/browser.h | 1 + 2 files changed, 6 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browser.c b/tools/perf/ui/browser.c index adb294a..6680fa5 100644 --- a/tools/perf/ui/browser.c +++ b/tools/perf/ui/browser.c @@ -250,7 +250,10 @@ int ui_browser__show(struct ui_browser *browser, const char *title, int err; va_list ap; - ui_browser__refresh_dimensions(browser); + if (browser->refresh_dimensions == NULL) + browser->refresh_dimensions = ui_browser__refresh_dimensions; + + browser->refresh_dimensions(browser); pthread_mutex_lock(&ui__lock); __ui_browser__show_title(browser, title); @@ -367,7 +370,7 @@ int ui_browser__run(struct ui_browser *browser, int delay_secs) if (key == K_RESIZE) { ui__refresh_dimensions(false); - ui_browser__refresh_dimensions(browser); + browser->refresh_dimensions(browser); __ui_browser__show_title(browser, browser->title); ui_helpline__puts(browser->helpline); continue; diff --git a/tools/perf/ui/browser.h b/tools/perf/ui/browser.h index cb8f39a..92ae721 100644 --- a/tools/perf/ui/browser.h +++ b/tools/perf/ui/browser.h @@ -19,6 +19,7 @@ struct ui_browser { void *priv; const char *title; char *helpline; + void (*refresh_dimensions)(struct ui_browser *browser); unsigned int (*refresh)(struct ui_browser *browser); void (*write)(struct ui_browser *browser, void *entry, int row); void (*seek)(struct ui_browser *browser, off_t offset, int whence); -- cgit v1.1 From ca3ff33b5c2dc67daa8d4d130e3bd6231fcb3d14 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 1 Jul 2014 16:42:24 -0300 Subject: perf hists browser: Introduce gotorc method That will allow us to add a row offset to open up space for the column headers. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-otc3ployokfci5qi81o7jo22@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index e16aff4..ca63564 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -63,6 +63,11 @@ static void hist_browser__refresh_dimensions(struct hist_browser *browser) sizeof("[k]")); } +static void hist_browser__gotorc(struct hist_browser *browser, int row, int column) +{ + ui_browser__gotorc(&browser->b, row, column); +} + static void hist_browser__reset(struct hist_browser *browser) { /* @@ -508,7 +513,7 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *browse } ui_browser__set_color(&browser->b, color); - ui_browser__gotorc(&browser->b, row, 0); + hist_browser__gotorc(browser, row, 0); slsmg_write_nstring(" ", offset + extra_offset); slsmg_printf("%c ", folded_sign); slsmg_write_nstring(str, width); @@ -567,7 +572,7 @@ static int hist_browser__show_callchain_node(struct hist_browser *browser, s = callchain_list__sym_name(chain, bf, sizeof(bf), browser->show_dso); - ui_browser__gotorc(&browser->b, row, 0); + hist_browser__gotorc(browser, row, 0); ui_browser__set_color(&browser->b, color); slsmg_write_nstring(" ", offset); slsmg_printf("%c ", folded_sign); @@ -732,7 +737,7 @@ static int hist_browser__show_entry(struct hist_browser *browser, .ptr = &arg, }; - ui_browser__gotorc(&browser->b, row, 0); + hist_browser__gotorc(browser, row, 0); perf_hpp__for_each_format(fmt) { if (perf_hpp__should_skip(fmt)) -- cgit v1.1 From 357cfff1c2173b7aedc0e521a2146b78396171e0 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 1 Jul 2014 17:01:01 -0300 Subject: perf hists browser: Override ui_browser refresh_dimensions method This requires some more work so that we can really just use the width of current entries when we want to partition the screen. Right now its just a prep patch so that we can have where to update ui_browser->rows when introducing the column headers line, that will be togglable, so we need to update it everytime we refresh the dimensions of the browser. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-ovk654rx525b4657y0mh6ku9@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index ca63564..ef7abf8 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -56,11 +56,19 @@ static u32 hist_browser__nr_entries(struct hist_browser *hb) return nr_entries + hb->nr_callchain_rows; } -static void hist_browser__refresh_dimensions(struct hist_browser *browser) +static void hist_browser__refresh_dimensions(struct ui_browser *browser) { + struct hist_browser *hb = container_of(browser, struct hist_browser, b); + /* 3 == +/- toggle symbol before actual hist_entry rendering */ - browser->b.width = 3 + (hists__sort_list_width(browser->hists) + - sizeof("[k]")); + browser->width = 3 + (hists__sort_list_width(hb->hists) + sizeof("[k]")); + /* + * FIXME: Just keeping existing behaviour, but this really should be + * before updating browser->width, as it will invalidate the + * calculation above. Fix this and the fallout in another + * changeset. + */ + ui_browser__refresh_dimensions(browser); } static void hist_browser__gotorc(struct hist_browser *browser, int row, int column) @@ -78,7 +86,7 @@ static void hist_browser__reset(struct hist_browser *browser) hist_browser__update_nr_entries(browser); browser->b.nr_entries = hist_browser__nr_entries(browser); - hist_browser__refresh_dimensions(browser); + hist_browser__refresh_dimensions(&browser->b); ui_browser__reset_index(&browser->b); } @@ -360,7 +368,6 @@ static int hist_browser__run(struct hist_browser *browser, browser->b.entries = &browser->hists->entries; browser->b.nr_entries = hist_browser__nr_entries(browser); - hist_browser__refresh_dimensions(browser); hists__browser_title(browser->hists, title, sizeof(title)); if (ui_browser__show(&browser->b, title, @@ -1195,6 +1202,7 @@ static struct hist_browser *hist_browser__new(struct hists *hists) if (browser) { browser->hists = hists; browser->b.refresh = hist_browser__refresh; + browser->b.refresh_dimensions = hist_browser__refresh_dimensions; browser->b.seek = ui_browser__hists_seek; browser->b.use_navkeypressed = true; } -- cgit v1.1 From 025bf7ea0aca26df4a4b28f0b12afe1aaf91e4bd Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sat, 14 Jun 2014 15:44:52 +0200 Subject: perf hists browser: Add support for showing columns header Open up space to show a one-line header text whenever 'H' is pressed, hide it on another key press. Follow up patch will format this line from the set of headers used. Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-m894d6qk30h3qofw4k8neq4q@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 42 ++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 40 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index ef7abf8..2be71bf 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -26,6 +26,7 @@ struct hist_browser { struct map_symbol *selection; int print_seq; bool show_dso; + bool show_headers; float min_pcnt; u64 nr_non_filtered_entries; u64 nr_callchain_rows; @@ -56,6 +57,21 @@ static u32 hist_browser__nr_entries(struct hist_browser *hb) return nr_entries + hb->nr_callchain_rows; } +static void hist_browser__update_rows(struct hist_browser *hb) +{ + struct ui_browser *browser = &hb->b; + u16 header_offset = hb->show_headers ? 1 : 0, index_row; + + browser->rows = browser->height - header_offset; + /* + * Verify if we were at the last line and that line isn't + * visibe because we now show the header line(s). + */ + index_row = browser->index - browser->top_idx; + if (index_row >= browser->rows) + browser->index -= index_row - browser->rows + 1; +} + static void hist_browser__refresh_dimensions(struct ui_browser *browser) { struct hist_browser *hb = container_of(browser, struct hist_browser, b); @@ -69,11 +85,14 @@ static void hist_browser__refresh_dimensions(struct ui_browser *browser) * changeset. */ ui_browser__refresh_dimensions(browser); + hist_browser__update_rows(hb); } static void hist_browser__gotorc(struct hist_browser *browser, int row, int column) { - ui_browser__gotorc(&browser->b, row, column); + u16 header_offset = browser->show_headers ? 1 : 0; + + ui_browser__gotorc(&browser->b, row + header_offset, column); } static void hist_browser__reset(struct hist_browser *browser) @@ -421,6 +440,10 @@ static int hist_browser__run(struct hist_browser *browser, /* Expand the whole world. */ hist_browser__set_folding(browser, true); break; + case 'H': + browser->show_headers = !browser->show_headers; + hist_browser__update_rows(browser); + break; case K_ENTER: if (hist_browser__toggle_fold(browser)) break; @@ -799,6 +822,13 @@ static int hist_browser__show_entry(struct hist_browser *browser, return printed; } +static void hist_browser__show_headers(struct hist_browser *browser) +{ + ui_browser__gotorc(&browser->b, 0, 0); + ui_browser__set_color(&browser->b, HE_COLORSET_ROOT); + slsmg_write_nstring(" ", browser->b.width + 1); +} + static void ui_browser__hists_init_top(struct ui_browser *browser) { if (browser->top == NULL) { @@ -812,9 +842,15 @@ static void ui_browser__hists_init_top(struct ui_browser *browser) static unsigned int hist_browser__refresh(struct ui_browser *browser) { unsigned row = 0; + u16 header_offset = 0; struct rb_node *nd; struct hist_browser *hb = container_of(browser, struct hist_browser, b); + if (hb->show_headers) { + hist_browser__show_headers(hb); + header_offset = 1; + } + ui_browser__hists_init_top(browser); for (nd = browser->top; nd; nd = rb_next(nd)) { @@ -833,7 +869,7 @@ static unsigned int hist_browser__refresh(struct ui_browser *browser) break; } - return row; + return row + header_offset; } static struct rb_node *hists__filter_entries(struct rb_node *nd, @@ -1205,6 +1241,7 @@ static struct hist_browser *hist_browser__new(struct hists *hists) browser->b.refresh_dimensions = hist_browser__refresh_dimensions; browser->b.seek = ui_browser__hists_seek; browser->b.use_navkeypressed = true; + browser->show_headers = false; } return browser; @@ -1434,6 +1471,7 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events, "d Zoom into current DSO\n" \ "E Expand all callchains\n" \ "F Toggle percentage of filtered entries\n" \ + "H Display column headers\n" \ /* help messages are sorted by lexical order of the hotkey */ const char report_help[] = HIST_BROWSER_HELP_COMMON -- cgit v1.1 From 81a888fea2cfd727052926e95510c11981d9b1c2 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Sat, 14 Jun 2014 15:44:52 +0200 Subject: perf hists browser: Display columns header text on 'H' press Displaying columns header text whenever 'H' is pressed, and hiding it on on another press. Signed-off-by: Jiri Olsa Cc: Arnaldo Carvalho de Melo Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-w9pcqpum5erza2a05ysvollz@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 47 ++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 45 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 2be71bf..5fa2e18 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -822,11 +822,54 @@ static int hist_browser__show_entry(struct hist_browser *browser, return printed; } +static int advance_hpp_check(struct perf_hpp *hpp, int inc) +{ + advance_hpp(hpp, inc); + return hpp->size <= 0; +} + +static int hists__scnprintf_headers(char *buf, size_t size, struct hists *hists) +{ + struct perf_hpp dummy_hpp = { + .buf = buf, + .size = size, + }; + struct perf_hpp_fmt *fmt; + size_t ret = 0; + + if (symbol_conf.use_callchain) { + ret = scnprintf(buf, size, " "); + if (advance_hpp_check(&dummy_hpp, ret)) + return ret; + } + + perf_hpp__for_each_format(fmt) { + if (perf_hpp__should_skip(fmt)) + continue; + + /* We need to add the length of the columns header. */ + perf_hpp__reset_width(fmt, hists); + + ret = fmt->header(fmt, &dummy_hpp, hists_to_evsel(hists)); + if (advance_hpp_check(&dummy_hpp, ret)) + break; + + ret = scnprintf(dummy_hpp.buf, dummy_hpp.size, " "); + if (advance_hpp_check(&dummy_hpp, ret)) + break; + } + + return ret; +} + static void hist_browser__show_headers(struct hist_browser *browser) { + char headers[1024]; + + hists__scnprintf_headers(headers, sizeof(headers), browser->hists); ui_browser__gotorc(&browser->b, 0, 0); ui_browser__set_color(&browser->b, HE_COLORSET_ROOT); - slsmg_write_nstring(" ", browser->b.width + 1); + slsmg_write_nstring(headers, browser->b.width + 1); } static void ui_browser__hists_init_top(struct ui_browser *browser) @@ -1241,7 +1284,7 @@ static struct hist_browser *hist_browser__new(struct hists *hists) browser->b.refresh_dimensions = hist_browser__refresh_dimensions; browser->b.seek = ui_browser__hists_seek; browser->b.use_navkeypressed = true; - browser->show_headers = false; + browser->show_headers = true; } return browser; -- cgit v1.1 From c83023676dc34f1b4422b842e1e2dc5c21bfc4dc Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 27 Jun 2014 18:26:58 +0200 Subject: perf hists browser: Add ui.show-headers config file option Adding ui.show-headers config file option to define if the histogram entries headers will start visible or not. Currently columns headers are displayed by default, following lines in ~/.perfconfig file will disable that: [ui] show-headers = false Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1403886418-5556-4-git-send-email-jolsa@kernel.org [ renamed symbol_conf.show_headers to .show_hist_headers ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browsers/hists.c | 2 +- tools/perf/util/config.c | 13 +++++++++++++ tools/perf/util/symbol.c | 1 + tools/perf/util/symbol.h | 3 ++- 4 files changed, 17 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 5fa2e18..a94b11fc 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -1284,7 +1284,7 @@ static struct hist_browser *hist_browser__new(struct hists *hists) browser->b.refresh_dimensions = hist_browser__refresh_dimensions; browser->b.seek = ui_browser__hists_seek; browser->b.use_navkeypressed = true; - browser->show_headers = true; + browser->show_headers = symbol_conf.show_hist_headers; } return browser; diff --git a/tools/perf/util/config.c b/tools/perf/util/config.c index 24519e1..1e5e2e5 100644 --- a/tools/perf/util/config.c +++ b/tools/perf/util/config.c @@ -350,6 +350,16 @@ static int perf_default_core_config(const char *var __maybe_unused, return 0; } +static int perf_ui_config(const char *var, const char *value) +{ + /* Add other config variables here. */ + if (!strcmp(var, "ui.show-headers")) { + symbol_conf.show_hist_headers = perf_config_bool(var, value); + return 0; + } + return 0; +} + int perf_default_config(const char *var, const char *value, void *dummy __maybe_unused) { @@ -359,6 +369,9 @@ int perf_default_config(const char *var, const char *value, if (!prefixcmp(var, "hist.")) return perf_hist_config(var, value); + if (!prefixcmp(var, "ui.")) + return perf_ui_config(var, value); + /* Add other config variables here. */ return 0; } diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 7b9096f..2e6a2e2 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -34,6 +34,7 @@ struct symbol_conf symbol_conf = { .annotate_src = true, .demangle = true, .cumulate_callchain = true, + .show_hist_headers = true, .symfs = "", }; diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 615c752..a81877b 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -118,7 +118,8 @@ struct symbol_conf { annotate_src, event_group, demangle, - filter_relative; + filter_relative, + show_hist_headers; const char *vmlinux_name, *kallsyms_name, *source_prefix, -- cgit v1.1 From 1561880d7b114a65e307d1ac9a0991dc715355b3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 7 Jul 2014 15:23:44 -0300 Subject: perf hists browser: Left justify column headers Looks better and avoids it moving to the end of the screen as the column width changes over time in 'perf top'. Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-yc144ai5jye3yl3h5yxw0scd@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/sort.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index 1ec57dd..14e5a03 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -1215,7 +1215,7 @@ static int __sort__hpp_header(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp, hse = container_of(fmt, struct hpp_sort_entry, hpp); len = hists__col_len(&evsel->hists, hse->se->se_width_idx); - return scnprintf(hpp->buf, hpp->size, "%*s", len, hse->se->se_header); + return scnprintf(hpp->buf, hpp->size, "%-*s", len, hse->se->se_header); } static int __sort__hpp_width(struct perf_hpp_fmt *fmt, -- cgit v1.1 From da50ad69723111e0292f1943dad77b41bb9a25b0 Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 17:59:49 +0400 Subject: perf kvm: Introduce HAVE_KVM_STAT_SUPPORT flag kvm stat support is currently conditional on i386/x86_64. Let's abstract this into a HAVE_KVM_STAT_SUPPORT flag, so that other architectures can support kvm stat as well. Reviewed-by: Cornelia Huck Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404395992-17095-2-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/x86/Makefile | 1 + tools/perf/builtin-kvm.c | 6 +++--- tools/perf/config/Makefile | 4 ++++ 3 files changed, 8 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/arch/x86/Makefile b/tools/perf/arch/x86/Makefile index 1641542..d393901 100644 --- a/tools/perf/arch/x86/Makefile +++ b/tools/perf/arch/x86/Makefile @@ -15,3 +15,4 @@ endif LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/header.o LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/tsc.o LIB_H += arch/$(ARCH)/util/tsc.h +HAVE_KVM_STAT_SUPPORT := 1 diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 0f1e5a2..214ec0e 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -29,7 +29,7 @@ #include #include -#if defined(__i386__) || defined(__x86_64__) +#ifdef HAVE_KVM_STAT_SUPPORT #include #include #include @@ -1609,7 +1609,7 @@ static int kvm_cmd_stat(const char *file_name, int argc, const char **argv) perf_stat: return cmd_stat(argc, argv, NULL); } -#endif +#endif /* HAVE_KVM_STAT_SUPPORT */ static int __cmd_record(const char *file_name, int argc, const char **argv) { @@ -1726,7 +1726,7 @@ int cmd_kvm(int argc, const char **argv, const char *prefix __maybe_unused) return cmd_top(argc, argv, NULL); else if (!strncmp(argv[0], "buildid-list", 12)) return __cmd_buildid_list(file_name, argc, argv); -#if defined(__i386__) || defined(__x86_64__) +#ifdef HAVE_KVM_STAT_SUPPORT else if (!strncmp(argv[0], "stat", 4)) return kvm_cmd_stat(file_name, argc, argv); #endif diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile index 346bdb6..b7f42d5 100644 --- a/tools/perf/config/Makefile +++ b/tools/perf/config/Makefile @@ -594,6 +594,10 @@ ifndef NO_LIBNUMA endif endif +ifdef HAVE_KVM_STAT_SUPPORT + CFLAGS += -DHAVE_KVM_STAT_SUPPORT +endif + # Among the variables below, these: # perfexecdir # template_dir -- cgit v1.1 From df74c13b6c53c97576652f7b2840764ad7d5f949 Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 17:59:50 +0400 Subject: perf kvm: Simplify exit reasons tables definitions The perf_kvm_stat struct keeps the size of a table of exit reasons in the field 'exit_reasons_size'. The field is initialized and then used by get_exit_reason() for serial access to the table, so that the calling function does not actually need to know table size. Usage of tables with 'end of sequence' marker simplifies the get_exit_reason() function. Also the patch introduces a define_exit_reasons_table, which makes it easier to define new tables. Reviewed-by: Cornelia Huck Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404395992-17095-3-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-kvm.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 214ec0e..75f3544 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -99,7 +99,6 @@ struct perf_kvm_stat { int trace_vcpu; struct exit_reasons_table *exit_reasons; - int exit_reasons_size; const char *exit_reasons_isa; struct kvm_events_ops *events_ops; @@ -158,20 +157,19 @@ static bool exit_event_end(struct perf_evsel *evsel, return kvm_entry_event(evsel); } -static struct exit_reasons_table vmx_exit_reasons[] = { - VMX_EXIT_REASONS -}; +#define define_exit_reasons_table(name, symbols) \ + static struct exit_reasons_table name[] = { \ + symbols, { -1, NULL } \ + } -static struct exit_reasons_table svm_exit_reasons[] = { - SVM_EXIT_REASONS -}; +define_exit_reasons_table(vmx_exit_reasons, VMX_EXIT_REASONS); +define_exit_reasons_table(svm_exit_reasons, SVM_EXIT_REASONS); -static const char *get_exit_reason(struct perf_kvm_stat *kvm, u64 exit_code) +static const char *get_exit_reason(struct perf_kvm_stat *kvm, + struct exit_reasons_table *tbl, + u64 exit_code) { - int i = kvm->exit_reasons_size; - struct exit_reasons_table *tbl = kvm->exit_reasons; - - while (i--) { + while (tbl->reason != NULL) { if (tbl->exit_code == exit_code) return tbl->reason; tbl++; @@ -186,7 +184,8 @@ static void exit_event_decode_key(struct perf_kvm_stat *kvm, struct event_key *key, char decode[20]) { - const char *exit_reason = get_exit_reason(kvm, key->key); + const char *exit_reason = get_exit_reason(kvm, kvm->exit_reasons, + key->key); scnprintf(decode, 20, "%s", exit_reason); } @@ -862,7 +861,6 @@ static int cpu_isa_config(struct perf_kvm_stat *kvm) if (isa == 1) { kvm->exit_reasons = vmx_exit_reasons; - kvm->exit_reasons_size = ARRAY_SIZE(vmx_exit_reasons); kvm->exit_reasons_isa = "VMX"; } @@ -1586,7 +1584,6 @@ static int kvm_cmd_stat(const char *file_name, int argc, const char **argv) .sort_key = "sample", .exit_reasons = svm_exit_reasons, - .exit_reasons_size = ARRAY_SIZE(svm_exit_reasons), .exit_reasons_isa = "SVM", }; -- cgit v1.1 From 65c647a65c155e69bd5765d5e454982566ac1c62 Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 17:59:51 +0400 Subject: perf kvm: Refactoring of cpu_isa_config() cpu_isa_config() does two different things: searching for cpuid and initializing perf_kvm_stat struct with proper parameters. Let's move initialization to a separate function cpu_isa_init(), which is used to initialize all possible ISAs and can be used to init arch-depended things. Reviewed-by: Cornelia Huck Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404395992-17095-4-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-kvm.c | 39 +++++++++++++++++++++++---------------- 1 file changed, 23 insertions(+), 16 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 75f3544..41dbeaf 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -835,36 +835,45 @@ static int process_sample_event(struct perf_tool *tool, return 0; } +static int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid) +{ + if (strstr(cpuid, "Intel")) { + kvm->exit_reasons = vmx_exit_reasons; + kvm->exit_reasons_isa = "VMX"; + } else if (strstr(cpuid, "AMD")) { + kvm->exit_reasons = svm_exit_reasons; + kvm->exit_reasons_isa = "SVM"; + } else + return -ENOTSUP; + + return 0; +} + static int cpu_isa_config(struct perf_kvm_stat *kvm) { char buf[64], *cpuid; - int err, isa; + int err; if (kvm->live) { err = get_cpuid(buf, sizeof(buf)); if (err != 0) { - pr_err("Failed to look up CPU type (Intel or AMD)\n"); + pr_err("Failed to look up CPU type\n"); return err; } cpuid = buf; } else cpuid = kvm->session->header.env.cpuid; - if (strstr(cpuid, "Intel")) - isa = 1; - else if (strstr(cpuid, "AMD")) - isa = 0; - else { - pr_err("CPU %s is not supported.\n", cpuid); - return -ENOTSUP; + if (!cpuid) { + pr_err("Failed to look up CPU type\n"); + return -EINVAL; } - if (isa == 1) { - kvm->exit_reasons = vmx_exit_reasons; - kvm->exit_reasons_isa = "VMX"; - } + err = cpu_isa_init(kvm, cpuid); + if (err == -ENOTSUP) + pr_err("CPU %s is not supported.\n", cpuid); - return 0; + return err; } static bool verify_vcpu(int vcpu) @@ -1583,8 +1592,6 @@ static int kvm_cmd_stat(const char *file_name, int argc, const char **argv) .report_event = "vmexit", .sort_key = "sample", - .exit_reasons = svm_exit_reasons, - .exit_reasons_isa = "SVM", }; if (argc == 1) { -- cgit v1.1 From f8d9ccde2be4c24bfb8787af36a59a981b00a02d Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 18:08:22 +0400 Subject: perf tools: Allow to use cpuinfo on s390 This patch defines CPUINFO_PROC for s390 and implements get_cpuid(). Reviewed-by: Cornelia Huck Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/87ioneo7qh.wl%yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/s390/Makefile | 1 + tools/perf/arch/s390/util/header.c | 28 ++++++++++++++++++++++++++++ tools/perf/perf-sys.h | 1 + 3 files changed, 30 insertions(+) create mode 100644 tools/perf/arch/s390/util/header.c (limited to 'tools') diff --git a/tools/perf/arch/s390/Makefile b/tools/perf/arch/s390/Makefile index 15130b50..744e629 100644 --- a/tools/perf/arch/s390/Makefile +++ b/tools/perf/arch/s390/Makefile @@ -2,3 +2,4 @@ ifndef NO_DWARF PERF_HAVE_DWARF_REGS := 1 LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/dwarf-regs.o endif +LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/header.o diff --git a/tools/perf/arch/s390/util/header.c b/tools/perf/arch/s390/util/header.c new file mode 100644 index 0000000..9fa6c3e --- /dev/null +++ b/tools/perf/arch/s390/util/header.c @@ -0,0 +1,28 @@ +/* + * Implementation of get_cpuid(). + * + * Copyright 2014 IBM Corp. + * Author(s): Alexander Yarygin + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License (version 2 only) + * as published by the Free Software Foundation. + */ + +#include +#include +#include +#include + +#include "../../util/header.h" + +int get_cpuid(char *buffer, size_t sz) +{ + const char *cpuid = "IBM/S390"; + + if (strlen(cpuid) + 1 > sz) + return -1; + + strcpy(buffer, cpuid); + return 0; +} diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h index 5268a14..937e432 100644 --- a/tools/perf/perf-sys.h +++ b/tools/perf/perf-sys.h @@ -54,6 +54,7 @@ #define mb() asm volatile("bcr 15,0" ::: "memory") #define wmb() asm volatile("bcr 15,0" ::: "memory") #define rmb() asm volatile("bcr 15,0" ::: "memory") +#define CPUINFO_PROC "vendor_id" #endif #ifdef __sh__ -- cgit v1.1 From d400a68d1f1e7a1fc7c5caf9024d4e67b218558d Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Fri, 4 Jul 2014 14:43:48 +0300 Subject: perf tools: Convert open coded equivalents to asprintf() The following snippet V = malloc(S); if (!V) { } sprintf(V, ...) Can be easily changed to a one line: if (asprintf(&V, ...) < 0) { } Signed-off-by: Andy Shevchenko Cc: Adrian Hunter Link: http://lkml.kernel.org/r/1404474229-15272-1-git-send-email-andriy.shevchenko@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/trace-event-info.c | 12 +++--------- tools/perf/util/util.c | 9 ++------- 2 files changed, 5 insertions(+), 16 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index 7e6fcfe..c3bba88 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -191,12 +191,10 @@ static int copy_event_system(const char *sys, struct tracepoint_path *tps) strcmp(dent->d_name, "..") == 0 || !name_in_tp_list(dent->d_name, tps)) continue; - format = malloc(strlen(sys) + strlen(dent->d_name) + 10); - if (!format) { + if (asprintf(&format, "%s/%s/format", sys, dent->d_name) < 0) { err = -ENOMEM; goto out; } - sprintf(format, "%s/%s/format", sys, dent->d_name); ret = stat(format, &st); free(format); if (ret < 0) @@ -217,12 +215,10 @@ static int copy_event_system(const char *sys, struct tracepoint_path *tps) strcmp(dent->d_name, "..") == 0 || !name_in_tp_list(dent->d_name, tps)) continue; - format = malloc(strlen(sys) + strlen(dent->d_name) + 10); - if (!format) { + if (asprintf(&format, "%s/%s/format", sys, dent->d_name) < 0) { err = -ENOMEM; goto out; } - sprintf(format, "%s/%s/format", sys, dent->d_name); ret = stat(format, &st); if (ret >= 0) { @@ -317,12 +313,10 @@ static int record_event_files(struct tracepoint_path *tps) strcmp(dent->d_name, "ftrace") == 0 || !system_in_tp_list(dent->d_name, tps)) continue; - sys = malloc(strlen(path) + strlen(dent->d_name) + 2); - if (!sys) { + if (asprintf(&sys, "%s/%s", path, dent->d_name) < 0) { err = -ENOMEM; goto out; } - sprintf(sys, "%s/%s", path, dent->d_name); ret = stat(sys, &st); if (ret >= 0) { ssize_t size = strlen(dent->d_name) + 1; diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c index 95aefa7..e4132ae 100644 --- a/tools/perf/util/util.c +++ b/tools/perf/util/util.c @@ -333,12 +333,9 @@ const char *find_tracing_dir(void) if (!debugfs) return NULL; - tracing = malloc(strlen(debugfs) + 9); - if (!tracing) + if (asprintf(&tracing, "%s/tracing", debugfs) < 0) return NULL; - sprintf(tracing, "%s/tracing", debugfs); - tracing_found = 1; return tracing; } @@ -352,11 +349,9 @@ char *get_tracing_file(const char *name) if (!tracing) return NULL; - file = malloc(strlen(tracing) + strlen(name) + 2); - if (!file) + if (asprintf(&file, "%s/%s", tracing, name) < 0) return NULL; - sprintf(file, "%s/%s", tracing, name); return file; } -- cgit v1.1 From 071266bf5b6a4f41250d2b26b000f1d2f9620afe Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 8 Jul 2014 12:40:11 -0300 Subject: perf tools: Suggest using -f to override perf.data file ownership message # id uid=0(root) gid=0(root) groups=0(root) # ls -la perf.data -rw-------. 1 acme acme 20720 Jul 8 11:35 perf.data Previously: # perf report file perf.data not owned by current user or root Now: # perf report File perf.data not owned by current user or root (use -f to override) Suggested-by: Ingo Molnar Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-0j2wuuegnhv3gljbil8ld6kx@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/data.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c index 55de44e..ee370a7f 100644 --- a/tools/perf/util/data.c +++ b/tools/perf/util/data.c @@ -65,7 +65,7 @@ static int open_file_read(struct perf_data_file *file) goto out_close; if (!file->force && st.st_uid && (st.st_uid != geteuid())) { - pr_err("file %s not owned by current user or root\n", + pr_err("File %s not owned by current user or root (use -f to override)\n", file->path); goto out_close; } -- cgit v1.1 From a2ea67d7b5618c868c1fe15a50af71b858d36887 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 8 Jul 2014 22:05:16 +0400 Subject: perf trace: Add pagefault statistics 'perf trace' can show summary of events using -S option. This commit also reports number of major/minor pagefault events in this summary. $ perf trace -s --pf all -- sleep 1 Summary of events: sleep (18604), 275 events, 99.6%, 197 minfaults, 0.000 msec syscall calls min avg max stddev (msec) (msec) (msec) (%) --------------- -------- --------- --------- --------- ------ read 2 0.000 0.001 0.002 100.00% open 3 0.004 0.005 0.007 21.13% close 3 0.001 0.001 0.001 1.37% fstat 3 0.001 0.002 0.002 10.66% mmap 8 0.002 0.004 0.006 10.69% mprotect 4 0.003 0.005 0.008 24.68% munmap 1 0.005 0.005 0.005 0.00% brk 3 0.001 0.002 0.003 28.08% access 3 0.002 0.003 0.005 24.48% nanosleep 1 1000.747 1000.747 1000.747 0.00% execve 8 0.000 0.033 0.246 91.00% arch_prctl 1 0.001 0.001 0.001 0.00% Signed-off-by: Stanislav Fomichev Cc: David Ahern Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404842716-19190-1-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index dc7a694..b94dffc 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1133,6 +1133,7 @@ struct thread_trace { u64 exit_time; bool entry_pending; unsigned long nr_events; + unsigned long pfmaj, pfmin; char *entry_str; double runtime_ms; struct { @@ -1804,8 +1805,20 @@ static int trace__pgfault(struct trace *trace, u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; struct addr_location al; char map_type = 'd'; + struct thread_trace *ttrace; thread = machine__findnew_thread(trace->host, sample->pid, sample->tid); + ttrace = thread__trace(thread, trace->output); + if (ttrace == NULL) + return -1; + + if (evsel->attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ) + ttrace->pfmaj++; + else + ttrace->pfmin++; + + if (trace->summary_only) + return 0; thread__find_addr_location(thread, trace->host, cpumode, MAP__FUNCTION, sample->ip, &al); @@ -2346,6 +2359,10 @@ static int trace__fprintf_one_thread(struct thread *thread, void *priv) printed += fprintf(fp, " %s (%d), ", thread__comm_str(thread), thread->tid); printed += fprintf(fp, "%lu events, ", ttrace->nr_events); printed += fprintf(fp, "%.1f%%", ratio); + if (ttrace->pfmaj) + printed += fprintf(fp, ", %lu majfaults", ttrace->pfmaj); + if (ttrace->pfmin) + printed += fprintf(fp, ", %lu minfaults", ttrace->pfmin); printed += fprintf(fp, ", %.3f msec\n", ttrace->runtime_ms); printed += thread__dump_stats(ttrace, trace, fp); -- cgit v1.1 From 4414a3c51028aea2ae2fe06c0377490eaa6abbfd Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 8 Jul 2014 15:39:21 -0300 Subject: perf trace: Fix build on 32-bit systems CC /tmp/build/perf/builtin-trace.o builtin-trace.c: In function 'print_location': builtin-trace.c:1792:4: error: format '%lx' expects argument of type 'long unsigned int', but argument 4 has type 'u64' [-Werror=format] builtin-trace.c:1794:3: error: format '%lx' expects argument of type 'long unsigned int', but argument 3 has type 'u64' [-Werror=format] builtin-trace.c:1796:3: error: format '%lx' expects argument of type 'long unsigned int', but argument 3 has type 'u64' [-Werror=format] cc1: all warnings being treated as errors make[1]: *** [/tmp/build/perf/builtin-trace.o] Error 1 make[1]: *** Waiting for unfinished jobs.... make: *** [install-bin] Error 2 make: Leaving directory `/home/acme/git/linux/tools/perf' acme@linux-goap:~/git/linux> uname -a Linux linux-goap 3.7.10-1.16-desktop #1 SMP PREEMPT Fri May 31 20:21:23 UTC 2013 (97c14ba) i686 i686 i386 GNU/Linux Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stanislav Fomichev Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-843p3aqbw531eqiu2hah8o9p@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index b94dffc..c4a5a7d 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1788,12 +1788,12 @@ static void print_location(FILE *f, struct perf_sample *sample, fprintf(f, "%s@", al->map->dso->long_name); if ((verbose || print_sym) && al->sym) - fprintf(f, "%s+0x%lx", al->sym->name, + fprintf(f, "%s+0x%" PRIx64, al->sym->name, al->addr - al->sym->start); else if (al->map) - fprintf(f, "0x%lx", al->addr); + fprintf(f, "0x%" PRIx64, al->addr); else - fprintf(f, "0x%lx", sample->addr); + fprintf(f, "0x%" PRIx64, sample->addr); } static int trace__pgfault(struct trace *trace, -- cgit v1.1 From f8dda74fb4139f026571b46d8d17f67e75aa157c Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 8 Jul 2014 20:03:40 +0400 Subject: perf timechart: Fix rendering in Firefox Firefox doesn't correctly handle cases where we specify number in quotes and have some padding around the number, like the following: In this case, it doesn't draw the figure. This patch removes 'field width' component from fprintf strings to fix it. Signed-off-by: Stanislav Fomichev Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1404835423-23098-2-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Jiri Olsa --- tools/perf/util/svghelper.c | 72 ++++++++++++++++++++++----------------------- 1 file changed, 36 insertions(+), 36 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/svghelper.c b/tools/perf/util/svghelper.c index 6a0a13d..27caf0f 100644 --- a/tools/perf/util/svghelper.c +++ b/tools/perf/util/svghelper.c @@ -137,7 +137,7 @@ void svg_box(int Yslot, u64 start, u64 end, const char *type) if (!svgfile) return; - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), time2pixels(end)-time2pixels(start), Yslot * SLOT_MULT, SLOT_HEIGHT, type); } @@ -174,7 +174,7 @@ void svg_running(int Yslot, int cpu, u64 start, u64 end, const char *backtrace) cpu, time_to_string(end - start)); if (backtrace) fprintf(svgfile, "Switched because:\n%s\n", backtrace); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), time2pixels(end)-time2pixels(start), Yslot * SLOT_MULT, SLOT_HEIGHT, type); @@ -186,7 +186,7 @@ void svg_running(int Yslot, int cpu, u64 start, u64 end, const char *backtrace) text_size = round_text_size(text_size); if (text_size > MIN_TEXT_SIZE) - fprintf(svgfile, "%i\n", + fprintf(svgfile, "%i\n", time2pixels(start), Yslot * SLOT_MULT + SLOT_HEIGHT - 1, text_size, cpu + 1); fprintf(svgfile, "\n"); @@ -202,10 +202,10 @@ static char *time_to_string(u64 duration) return text; if (duration < 1000 * 1000) { /* less than 1 msec */ - sprintf(text, "%4.1f us", duration / 1000.0); + sprintf(text, "%.1f us", duration / 1000.0); return text; } - sprintf(text, "%4.1f ms", duration / 1000.0 / 1000); + sprintf(text, "%.1f ms", duration / 1000.0 / 1000); return text; } @@ -233,14 +233,14 @@ void svg_waiting(int Yslot, int cpu, u64 start, u64 end, const char *backtrace) font_size = round_text_size(font_size); - fprintf(svgfile, "\n", time2pixels(start), Yslot * SLOT_MULT); + fprintf(svgfile, "\n", time2pixels(start), Yslot * SLOT_MULT); fprintf(svgfile, "#%d waiting %s\n", cpu, time_to_string(end - start)); if (backtrace) fprintf(svgfile, "Waiting on:\n%s\n", backtrace); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(end)-time2pixels(start), SLOT_HEIGHT, style); if (font_size > MIN_TEXT_SIZE) - fprintf(svgfile, " %s\n", + fprintf(svgfile, " %s\n", font_size, text); fprintf(svgfile, "\n"); } @@ -289,16 +289,16 @@ void svg_cpu_box(int cpu, u64 __max_freq, u64 __turbo_freq) fprintf(svgfile, "\n"); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(first_time), time2pixels(last_time)-time2pixels(first_time), cpu2y(cpu), SLOT_MULT+SLOT_HEIGHT); sprintf(cpu_string, "CPU %i", (int)cpu); - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", 10+time2pixels(first_time), cpu2y(cpu) + SLOT_HEIGHT/2, cpu_string); - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", 10+time2pixels(first_time), cpu2y(cpu) + SLOT_MULT + SLOT_HEIGHT - 4, cpu_model()); fprintf(svgfile, "\n"); @@ -319,11 +319,11 @@ void svg_process(int cpu, u64 start, u64 end, int pid, const char *name, const c else type = "sample"; - fprintf(svgfile, "\n", time2pixels(start), cpu2y(cpu)); + fprintf(svgfile, "\n", time2pixels(start), cpu2y(cpu)); fprintf(svgfile, "%d %s running %s\n", pid, name, time_to_string(end - start)); if (backtrace) fprintf(svgfile, "Switched because:\n%s\n", backtrace); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(end)-time2pixels(start), SLOT_MULT+SLOT_HEIGHT, type); width = time2pixels(end)-time2pixels(start); if (width > 6) @@ -332,7 +332,7 @@ void svg_process(int cpu, u64 start, u64 end, int pid, const char *name, const c width = round_text_size(width); if (width > MIN_TEXT_SIZE) - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", width, name); fprintf(svgfile, "\n"); @@ -353,7 +353,7 @@ void svg_cstate(int cpu, u64 start, u64 end, int type) type = 6; sprintf(style, "c%i", type); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", style, time2pixels(start), time2pixels(end)-time2pixels(start), cpu2y(cpu), SLOT_MULT+SLOT_HEIGHT); @@ -365,7 +365,7 @@ void svg_cstate(int cpu, u64 start, u64 end, int type) width = round_text_size(width); if (width > MIN_TEXT_SIZE) - fprintf(svgfile, "C%i\n", + fprintf(svgfile, "C%i\n", time2pixels(start), cpu2y(cpu)+width, width, type); fprintf(svgfile, "\n"); @@ -407,9 +407,9 @@ void svg_pstate(int cpu, u64 start, u64 end, u64 freq) if (max_freq) height = freq * 1.0 / max_freq * (SLOT_HEIGHT + SLOT_MULT); height = 1 + cpu2y(cpu) + SLOT_MULT + SLOT_HEIGHT - height; - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), time2pixels(end), height, height); - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", time2pixels(start), height+0.9, HzToHuman(freq)); fprintf(svgfile, "\n"); @@ -435,32 +435,32 @@ void svg_partial_wakeline(u64 start, int row1, char *desc1, int row2, char *desc if (row1 < row2) { if (row1) { - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row1 * SLOT_MULT + SLOT_HEIGHT, time2pixels(start), row1 * SLOT_MULT + SLOT_HEIGHT + SLOT_MULT/32); if (desc2) - fprintf(svgfile, "%s >\n", + fprintf(svgfile, "%s >\n", time2pixels(start), row1 * SLOT_MULT + SLOT_HEIGHT + SLOT_HEIGHT/48, desc2); } if (row2) { - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row2 * SLOT_MULT - SLOT_MULT/32, time2pixels(start), row2 * SLOT_MULT); if (desc1) - fprintf(svgfile, "%s >\n", + fprintf(svgfile, "%s >\n", time2pixels(start), row2 * SLOT_MULT - SLOT_MULT/32, desc1); } } else { if (row2) { - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row2 * SLOT_MULT + SLOT_HEIGHT, time2pixels(start), row2 * SLOT_MULT + SLOT_HEIGHT + SLOT_MULT/32); if (desc1) - fprintf(svgfile, "%s <\n", + fprintf(svgfile, "%s <\n", time2pixels(start), row2 * SLOT_MULT + SLOT_HEIGHT + SLOT_MULT/48, desc1); } if (row1) { - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row1 * SLOT_MULT - SLOT_MULT/32, time2pixels(start), row1 * SLOT_MULT); if (desc2) - fprintf(svgfile, "%s <\n", + fprintf(svgfile, "%s <\n", time2pixels(start), row1 * SLOT_MULT - SLOT_HEIGHT/32, desc2); } } @@ -468,7 +468,7 @@ void svg_partial_wakeline(u64 start, int row1, char *desc1, int row2, char *desc if (row2 > row1) height += SLOT_HEIGHT; if (row1) - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), height); fprintf(svgfile, "\n"); @@ -488,16 +488,16 @@ void svg_wakeline(u64 start, int row1, int row2, const char *backtrace) fprintf(svgfile, "%s\n", backtrace); if (row1 < row2) - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row1 * SLOT_MULT + SLOT_HEIGHT, time2pixels(start), row2 * SLOT_MULT); else - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row2 * SLOT_MULT + SLOT_HEIGHT, time2pixels(start), row1 * SLOT_MULT); height = row1 * SLOT_MULT; if (row2 > row1) height += SLOT_HEIGHT; - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), height); fprintf(svgfile, "\n"); @@ -515,9 +515,9 @@ void svg_interrupt(u64 start, int row, const char *backtrace) if (backtrace) fprintf(svgfile, "%s\n", backtrace); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row * SLOT_MULT); - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(start), row * SLOT_MULT + SLOT_HEIGHT); fprintf(svgfile, "\n"); @@ -528,7 +528,7 @@ void svg_text(int Yslot, u64 start, const char *text) if (!svgfile) return; - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", time2pixels(start), Yslot * SLOT_MULT+SLOT_HEIGHT/2, text); } @@ -537,9 +537,9 @@ static void svg_legenda_box(int X, const char *text, const char *style) double boxsize; boxsize = SLOT_HEIGHT / 2; - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", X, boxsize, boxsize, style); - fprintf(svgfile, "%s\n", + fprintf(svgfile, "%s\n", X + boxsize + 5, boxsize, 0.8 * boxsize, text); } @@ -579,7 +579,7 @@ void svg_time_grid(void) color = 128; } - fprintf(svgfile, "\n", + fprintf(svgfile, "\n", time2pixels(i), SLOT_MULT/2, time2pixels(i), total_height, color, color, color, thickness); i += 10000000; -- cgit v1.1 From b97b59b93d10a54022afb06d5725d7aa55d98dd7 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 8 Jul 2014 20:03:41 +0400 Subject: perf timechart: Implement IO mode Currently, timechart records only scheduler and CPU events (task switches, running times, CPU power states, etc); this commit adds IO mode which makes it possible to record IO (disk, network) activity. In this mode perf timechart will generate SVG with IO charts (writes, reads, tx, rx, polls). Signed-off-by: Stanislav Fomichev Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1404835423-23098-3-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Jiri Olsa --- tools/perf/Documentation/perf-timechart.txt | 25 +- tools/perf/builtin-timechart.c | 620 +++++++++++++++++++++++++++- tools/perf/util/svghelper.c | 98 ++++- tools/perf/util/svghelper.h | 6 +- 4 files changed, 725 insertions(+), 24 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-timechart.txt b/tools/perf/Documentation/perf-timechart.txt index 5e0f986..ec6b46c 100644 --- a/tools/perf/Documentation/perf-timechart.txt +++ b/tools/perf/Documentation/perf-timechart.txt @@ -15,10 +15,20 @@ DESCRIPTION There are two variants of perf timechart: 'perf timechart record ' to record the system level events - of an arbitrary workload. + of an arbitrary workload. By default timechart records only scheduler + and CPU events (task switches, running times, CPU power states, etc), + but it's possible to record IO (disk, network) activity using -I argument. 'perf timechart' to turn a trace into a Scalable Vector Graphics file, - that can be viewed with popular SVG viewers such as 'Inkscape'. + that can be viewed with popular SVG viewers such as 'Inkscape'. Depending + on the events in the perf.data file, timechart will contain scheduler/cpu + events or IO events. + + In IO mode, every bar has two charts: upper and lower. + Upper bar shows incoming events (disk reads, ingress network packets). + Lower bar shows outgoing events (disk writes, egress network packets). + There are also poll bars which show how much time application spent + in poll/epoll/select syscalls. TIMECHART OPTIONS ----------------- @@ -63,6 +73,9 @@ RECORD OPTIONS -T:: --tasks-only:: Record only tasks-related events +-I:: +--io-only:: + Record only io-related events -g:: --callchain:: Do call-graph (stack chain/backtrace) recording @@ -87,6 +100,14 @@ Record system-wide timechart: $ perf timechart --highlight gcc +Record system-wide IO events: + + $ perf timechart record -I + + then generate timechart: + + $ perf timechart + SEE ALSO -------- linkperf:perf-record[1] diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 74db256..4079062 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -60,10 +60,14 @@ struct timechart { tasks_only, with_backtrace, topology; + /* IO related settings */ + u64 io_events; + bool io_only; }; struct per_pidcomm; struct cpu_sample; +struct io_sample; /* * Datastructure layout: @@ -84,6 +88,7 @@ struct per_pid { u64 start_time; u64 end_time; u64 total_time; + u64 total_bytes; int display; struct per_pidcomm *all; @@ -97,6 +102,8 @@ struct per_pidcomm { u64 start_time; u64 end_time; u64 total_time; + u64 max_bytes; + u64 total_bytes; int Y; int display; @@ -107,6 +114,7 @@ struct per_pidcomm { char *comm; struct cpu_sample *samples; + struct io_sample *io_samples; }; struct sample_wrapper { @@ -131,6 +139,27 @@ struct cpu_sample { const char *backtrace; }; +enum { + IOTYPE_READ, + IOTYPE_WRITE, + IOTYPE_SYNC, + IOTYPE_TX, + IOTYPE_RX, + IOTYPE_POLL, +}; + +struct io_sample { + struct io_sample *next; + + u64 start_time; + u64 end_time; + u64 bytes; + int type; + int fd; + int err; + int merges; +}; + #define CSTATE 1 #define PSTATE 2 @@ -682,6 +711,219 @@ static void end_sample_processing(struct timechart *tchart) } } +static int pid_begin_io_sample(struct timechart *tchart, int pid, int type, + u64 start, int fd) +{ + struct per_pid *p = find_create_pid(tchart, pid); + struct per_pidcomm *c = p->current; + struct io_sample *sample; + struct io_sample *prev; + + if (!c) { + c = zalloc(sizeof(*c)); + if (!c) + return -ENOMEM; + p->current = c; + c->next = p->all; + p->all = c; + } + + prev = c->io_samples; + + if (prev && prev->start_time && !prev->end_time) { + pr_warning("Skip invalid start event: " + "previous event already started!\n"); + + /* remove previous event that has been started, + * we are not sure we will ever get an end for it */ + c->io_samples = prev->next; + free(prev); + return 0; + } + + sample = zalloc(sizeof(*sample)); + if (!sample) + return -ENOMEM; + sample->start_time = start; + sample->type = type; + sample->fd = fd; + sample->next = c->io_samples; + c->io_samples = sample; + + if (c->start_time == 0 || c->start_time > start) + c->start_time = start; + + return 0; +} + +static int pid_end_io_sample(struct timechart *tchart, int pid, int type, + u64 end, long ret) +{ + struct per_pid *p = find_create_pid(tchart, pid); + struct per_pidcomm *c = p->current; + struct io_sample *sample; + + if (!c) { + pr_warning("Invalid pidcomm!\n"); + return -1; + } + + sample = c->io_samples; + + if (!sample) /* skip partially captured events */ + return 0; + + if (sample->end_time) { + pr_warning("Skip invalid end event: " + "previous event already ended!\n"); + return 0; + } + + if (sample->type != type) { + pr_warning("Skip invalid end event: invalid event type!\n"); + return 0; + } + + sample->end_time = end; + + if (ret < 0) { + sample->err = ret; + } else if (type == IOTYPE_READ || type == IOTYPE_WRITE || + type == IOTYPE_TX || type == IOTYPE_RX) { + + if ((u64)ret > c->max_bytes) + c->max_bytes = ret; + + c->total_bytes += ret; + p->total_bytes += ret; + sample->bytes = ret; + } + + tchart->io_events++; + + return 0; +} + +static int +process_enter_read(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_READ, + sample->time, fd); +} + +static int +process_exit_read(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_READ, + sample->time, ret); +} + +static int +process_enter_write(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_WRITE, + sample->time, fd); +} + +static int +process_exit_write(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_WRITE, + sample->time, ret); +} + +static int +process_enter_sync(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_SYNC, + sample->time, fd); +} + +static int +process_exit_sync(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_SYNC, + sample->time, ret); +} + +static int +process_enter_tx(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_TX, + sample->time, fd); +} + +static int +process_exit_tx(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_TX, + sample->time, ret); +} + +static int +process_enter_rx(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_RX, + sample->time, fd); +} + +static int +process_exit_rx(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_RX, + sample->time, ret); +} + +static int +process_enter_poll(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long fd = perf_evsel__intval(evsel, sample, "fd"); + return pid_begin_io_sample(tchart, sample->tid, IOTYPE_POLL, + sample->time, fd); +} + +static int +process_exit_poll(struct timechart *tchart, + struct perf_evsel *evsel, + struct perf_sample *sample) +{ + long ret = perf_evsel__intval(evsel, sample, "ret"); + return pid_end_io_sample(tchart, sample->tid, IOTYPE_POLL, + sample->time, ret); +} + /* * Sort the pid datastructure */ @@ -852,6 +1094,117 @@ static void draw_cpu_usage(struct timechart *tchart) } } +static void draw_io_bars(struct timechart *tchart) +{ + const char *suf; + double bytes; + char comm[256]; + struct per_pid *p; + struct per_pidcomm *c; + struct io_sample *sample; + int Y = 1; + + p = tchart->all_data; + while (p) { + c = p->all; + while (c) { + if (!c->display) { + c->Y = 0; + c = c->next; + continue; + } + + svg_box(Y, c->start_time, c->end_time, "process3"); + sample = c->io_samples; + for (sample = c->io_samples; sample; sample = sample->next) { + double h = (double)sample->bytes / c->max_bytes; + + if (sample->err) + h = 1; + + if (sample->type == IOTYPE_SYNC) + svg_fbox(Y, + sample->start_time, + sample->end_time, + 1, + sample->err ? "error" : "sync", + sample->fd, + sample->err, + sample->merges); + else if (sample->type == IOTYPE_POLL) + svg_fbox(Y, + sample->start_time, + sample->end_time, + 1, + sample->err ? "error" : "poll", + sample->fd, + sample->err, + sample->merges); + else if (sample->type == IOTYPE_READ) + svg_ubox(Y, + sample->start_time, + sample->end_time, + h, + sample->err ? "error" : "disk", + sample->fd, + sample->err, + sample->merges); + else if (sample->type == IOTYPE_WRITE) + svg_lbox(Y, + sample->start_time, + sample->end_time, + h, + sample->err ? "error" : "disk", + sample->fd, + sample->err, + sample->merges); + else if (sample->type == IOTYPE_RX) + svg_ubox(Y, + sample->start_time, + sample->end_time, + h, + sample->err ? "error" : "net", + sample->fd, + sample->err, + sample->merges); + else if (sample->type == IOTYPE_TX) + svg_lbox(Y, + sample->start_time, + sample->end_time, + h, + sample->err ? "error" : "net", + sample->fd, + sample->err, + sample->merges); + } + + suf = ""; + bytes = c->total_bytes; + if (bytes > 1024) { + bytes = bytes / 1024; + suf = "K"; + } + if (bytes > 1024) { + bytes = bytes / 1024; + suf = "M"; + } + if (bytes > 1024) { + bytes = bytes / 1024; + suf = "G"; + } + + + sprintf(comm, "%s:%i (%3.1f %sbytes)", c->comm ?: "", p->pid, bytes, suf); + svg_text(Y, c->start_time, comm); + + c->Y = Y; + Y++; + c = c->next; + } + p = p->next; + } +} + static void draw_process_bars(struct timechart *tchart) { struct per_pid *p; @@ -987,9 +1340,6 @@ static int determine_display_tasks(struct timechart *tchart, u64 threshold) struct per_pidcomm *c; int count = 0; - if (process_filter) - return determine_display_tasks_filtered(tchart); - p = tchart->all_data; while (p) { p->display = 0; @@ -1025,15 +1375,46 @@ static int determine_display_tasks(struct timechart *tchart, u64 threshold) return count; } +static int determine_display_io_tasks(struct timechart *timechart, u64 threshold) +{ + struct per_pid *p; + struct per_pidcomm *c; + int count = 0; + + p = timechart->all_data; + while (p) { + /* no exit marker, task kept running to the end */ + if (p->end_time == 0) + p->end_time = timechart->last_time; + + c = p->all; + + while (c) { + c->display = 0; + + if (c->total_bytes >= threshold) { + c->display = 1; + count++; + } + if (c->end_time == 0) + c->end_time = timechart->last_time; + + c = c->next; + } + p = p->next; + } + return count; +} +#define BYTES_THRESH (1 * 1024 * 1024) #define TIME_THRESH 10000000 static void write_svg_file(struct timechart *tchart, const char *filename) { u64 i; int count; - int thresh = TIME_THRESH; + int thresh = tchart->io_events ? BYTES_THRESH : TIME_THRESH; if (tchart->power_only) tchart->proc_num = 0; @@ -1041,28 +1422,43 @@ static void write_svg_file(struct timechart *tchart, const char *filename) /* We'd like to show at least proc_num tasks; * be less picky if we have fewer */ do { - count = determine_display_tasks(tchart, thresh); + if (process_filter) + count = determine_display_tasks_filtered(tchart); + else if (tchart->io_events) + count = determine_display_io_tasks(tchart, thresh); + else + count = determine_display_tasks(tchart, thresh); thresh /= 10; } while (!process_filter && thresh && count < tchart->proc_num); if (!tchart->proc_num) count = 0; - open_svg(filename, tchart->numcpus, count, tchart->first_time, tchart->last_time); + if (tchart->io_events) { + open_svg(filename, 0, count, tchart->first_time, tchart->last_time); - svg_time_grid(); - svg_legenda(); + svg_time_grid(0.5); + svg_io_legenda(); - for (i = 0; i < tchart->numcpus; i++) - svg_cpu_box(i, tchart->max_freq, tchart->turbo_frequency); + draw_io_bars(tchart); + } else { + open_svg(filename, tchart->numcpus, count, tchart->first_time, tchart->last_time); - draw_cpu_usage(tchart); - if (tchart->proc_num) - draw_process_bars(tchart); - if (!tchart->tasks_only) - draw_c_p_states(tchart); - if (tchart->proc_num) - draw_wakeups(tchart); + svg_time_grid(0); + + svg_legenda(); + + for (i = 0; i < tchart->numcpus; i++) + svg_cpu_box(i, tchart->max_freq, tchart->turbo_frequency); + + draw_cpu_usage(tchart); + if (tchart->proc_num) + draw_process_bars(tchart); + if (!tchart->tasks_only) + draw_c_p_states(tchart); + if (tchart->proc_num) + draw_wakeups(tchart); + } svg_close(); } @@ -1110,6 +1506,56 @@ static int __cmd_timechart(struct timechart *tchart, const char *output_name) { "power:power_end", process_sample_power_end }, { "power:power_frequency", process_sample_power_frequency }, #endif + + { "syscalls:sys_enter_read", process_enter_read }, + { "syscalls:sys_enter_pread64", process_enter_read }, + { "syscalls:sys_enter_readv", process_enter_read }, + { "syscalls:sys_enter_preadv", process_enter_read }, + { "syscalls:sys_enter_write", process_enter_write }, + { "syscalls:sys_enter_pwrite64", process_enter_write }, + { "syscalls:sys_enter_writev", process_enter_write }, + { "syscalls:sys_enter_pwritev", process_enter_write }, + { "syscalls:sys_enter_sync", process_enter_sync }, + { "syscalls:sys_enter_sync_file_range", process_enter_sync }, + { "syscalls:sys_enter_fsync", process_enter_sync }, + { "syscalls:sys_enter_msync", process_enter_sync }, + { "syscalls:sys_enter_recvfrom", process_enter_rx }, + { "syscalls:sys_enter_recvmmsg", process_enter_rx }, + { "syscalls:sys_enter_recvmsg", process_enter_rx }, + { "syscalls:sys_enter_sendto", process_enter_tx }, + { "syscalls:sys_enter_sendmsg", process_enter_tx }, + { "syscalls:sys_enter_sendmmsg", process_enter_tx }, + { "syscalls:sys_enter_epoll_pwait", process_enter_poll }, + { "syscalls:sys_enter_epoll_wait", process_enter_poll }, + { "syscalls:sys_enter_poll", process_enter_poll }, + { "syscalls:sys_enter_ppoll", process_enter_poll }, + { "syscalls:sys_enter_pselect6", process_enter_poll }, + { "syscalls:sys_enter_select", process_enter_poll }, + + { "syscalls:sys_exit_read", process_exit_read }, + { "syscalls:sys_exit_pread64", process_exit_read }, + { "syscalls:sys_exit_readv", process_exit_read }, + { "syscalls:sys_exit_preadv", process_exit_read }, + { "syscalls:sys_exit_write", process_exit_write }, + { "syscalls:sys_exit_pwrite64", process_exit_write }, + { "syscalls:sys_exit_writev", process_exit_write }, + { "syscalls:sys_exit_pwritev", process_exit_write }, + { "syscalls:sys_exit_sync", process_exit_sync }, + { "syscalls:sys_exit_sync_file_range", process_exit_sync }, + { "syscalls:sys_exit_fsync", process_exit_sync }, + { "syscalls:sys_exit_msync", process_exit_sync }, + { "syscalls:sys_exit_recvfrom", process_exit_rx }, + { "syscalls:sys_exit_recvmmsg", process_exit_rx }, + { "syscalls:sys_exit_recvmsg", process_exit_rx }, + { "syscalls:sys_exit_sendto", process_exit_tx }, + { "syscalls:sys_exit_sendmsg", process_exit_tx }, + { "syscalls:sys_exit_sendmmsg", process_exit_tx }, + { "syscalls:sys_exit_epoll_pwait", process_exit_poll }, + { "syscalls:sys_exit_epoll_wait", process_exit_poll }, + { "syscalls:sys_exit_poll", process_exit_poll }, + { "syscalls:sys_exit_ppoll", process_exit_poll }, + { "syscalls:sys_exit_pselect6", process_exit_poll }, + { "syscalls:sys_exit_select", process_exit_poll }, }; struct perf_data_file file = { .path = input_name, @@ -1154,6 +1600,139 @@ out_delete: return ret; } +static int timechart__io_record(int argc, const char **argv) +{ + unsigned int rec_argc, i; + const char **rec_argv; + const char **p; + char *filter = NULL; + + const char * const common_args[] = { + "record", "-a", "-R", "-c", "1", + }; + unsigned int common_args_nr = ARRAY_SIZE(common_args); + + const char * const disk_events[] = { + "syscalls:sys_enter_read", + "syscalls:sys_enter_pread64", + "syscalls:sys_enter_readv", + "syscalls:sys_enter_preadv", + "syscalls:sys_enter_write", + "syscalls:sys_enter_pwrite64", + "syscalls:sys_enter_writev", + "syscalls:sys_enter_pwritev", + "syscalls:sys_enter_sync", + "syscalls:sys_enter_sync_file_range", + "syscalls:sys_enter_fsync", + "syscalls:sys_enter_msync", + + "syscalls:sys_exit_read", + "syscalls:sys_exit_pread64", + "syscalls:sys_exit_readv", + "syscalls:sys_exit_preadv", + "syscalls:sys_exit_write", + "syscalls:sys_exit_pwrite64", + "syscalls:sys_exit_writev", + "syscalls:sys_exit_pwritev", + "syscalls:sys_exit_sync", + "syscalls:sys_exit_sync_file_range", + "syscalls:sys_exit_fsync", + "syscalls:sys_exit_msync", + }; + unsigned int disk_events_nr = ARRAY_SIZE(disk_events); + + const char * const net_events[] = { + "syscalls:sys_enter_recvfrom", + "syscalls:sys_enter_recvmmsg", + "syscalls:sys_enter_recvmsg", + "syscalls:sys_enter_sendto", + "syscalls:sys_enter_sendmsg", + "syscalls:sys_enter_sendmmsg", + + "syscalls:sys_exit_recvfrom", + "syscalls:sys_exit_recvmmsg", + "syscalls:sys_exit_recvmsg", + "syscalls:sys_exit_sendto", + "syscalls:sys_exit_sendmsg", + "syscalls:sys_exit_sendmmsg", + }; + unsigned int net_events_nr = ARRAY_SIZE(net_events); + + const char * const poll_events[] = { + "syscalls:sys_enter_epoll_pwait", + "syscalls:sys_enter_epoll_wait", + "syscalls:sys_enter_poll", + "syscalls:sys_enter_ppoll", + "syscalls:sys_enter_pselect6", + "syscalls:sys_enter_select", + + "syscalls:sys_exit_epoll_pwait", + "syscalls:sys_exit_epoll_wait", + "syscalls:sys_exit_poll", + "syscalls:sys_exit_ppoll", + "syscalls:sys_exit_pselect6", + "syscalls:sys_exit_select", + }; + unsigned int poll_events_nr = ARRAY_SIZE(poll_events); + + rec_argc = common_args_nr + + disk_events_nr * 4 + + net_events_nr * 4 + + poll_events_nr * 4 + + argc; + rec_argv = calloc(rec_argc + 1, sizeof(char *)); + + if (rec_argv == NULL) + return -ENOMEM; + + if (asprintf(&filter, "common_pid != %d", getpid()) < 0) + return -ENOMEM; + + p = rec_argv; + for (i = 0; i < common_args_nr; i++) + *p++ = strdup(common_args[i]); + + for (i = 0; i < disk_events_nr; i++) { + if (!is_valid_tracepoint(disk_events[i])) { + rec_argc -= 4; + continue; + } + + *p++ = "-e"; + *p++ = strdup(disk_events[i]); + *p++ = "--filter"; + *p++ = filter; + } + for (i = 0; i < net_events_nr; i++) { + if (!is_valid_tracepoint(net_events[i])) { + rec_argc -= 4; + continue; + } + + *p++ = "-e"; + *p++ = strdup(net_events[i]); + *p++ = "--filter"; + *p++ = filter; + } + for (i = 0; i < poll_events_nr; i++) { + if (!is_valid_tracepoint(poll_events[i])) { + rec_argc -= 4; + continue; + } + + *p++ = "-e"; + *p++ = strdup(poll_events[i]); + *p++ = "--filter"; + *p++ = filter; + } + + for (i = 0; i < (unsigned int)argc; i++) + *p++ = argv[i]; + + return cmd_record(rec_argc, rec_argv, NULL); +} + + static int timechart__record(struct timechart *tchart, int argc, const char **argv) { unsigned int rec_argc, i, j; @@ -1314,6 +1893,8 @@ int cmd_timechart(int argc, const char **argv, OPT_BOOLEAN('P', "power-only", &tchart.power_only, "output power data only"), OPT_BOOLEAN('T', "tasks-only", &tchart.tasks_only, "output processes data only"), + OPT_BOOLEAN('I', "io-only", &tchart.io_only, + "record only IO data"), OPT_BOOLEAN('g', "callchain", &tchart.with_backtrace, "record callchain"), OPT_END() }; @@ -1340,7 +1921,10 @@ int cmd_timechart(int argc, const char **argv, return -1; } - return timechart__record(&tchart, argc, argv); + if (tchart.io_only) + return timechart__io_record(argc, argv); + else + return timechart__record(&tchart, argc, argv); } else if (argc) usage_with_options(timechart_usage, timechart_options); diff --git a/tools/perf/util/svghelper.c b/tools/perf/util/svghelper.c index 27caf0f..283d3e7 100644 --- a/tools/perf/util/svghelper.c +++ b/tools/perf/util/svghelper.c @@ -30,6 +30,7 @@ static u64 turbo_frequency, max_freq; #define SLOT_MULT 30.0 #define SLOT_HEIGHT 25.0 +#define SLOT_HALF (SLOT_HEIGHT / 2) int svg_page_width = 1000; u64 svg_highlight; @@ -114,8 +115,14 @@ void open_svg(const char *filename, int cpus, int rows, u64 start, u64 end) fprintf(svgfile, " rect { stroke-width: 1; }\n"); fprintf(svgfile, " rect.process { fill:rgb(180,180,180); fill-opacity:0.9; stroke-width:1; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.process2 { fill:rgb(180,180,180); fill-opacity:0.9; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.process3 { fill:rgb(180,180,180); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.sample { fill:rgb( 0, 0,255); fill-opacity:0.8; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.sample_hi{ fill:rgb(255,128, 0); fill-opacity:0.8; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.error { fill:rgb(255, 0, 0); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.net { fill:rgb( 0,128, 0); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.disk { fill:rgb( 0, 0,255); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.sync { fill:rgb(128,128, 0); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); + fprintf(svgfile, " rect.poll { fill:rgb( 0,128,128); fill-opacity:0.2; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.blocked { fill:rgb(255, 0, 0); fill-opacity:0.5; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.waiting { fill:rgb(224,214, 0); fill-opacity:0.8; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); fprintf(svgfile, " rect.WAITING { fill:rgb(255,214, 48); fill-opacity:0.6; stroke-width:0; stroke:rgb( 0, 0, 0); } \n"); @@ -132,6 +139,75 @@ void open_svg(const char *filename, int cpus, int rows, u64 start, u64 end) fprintf(svgfile, " ]]>\n \n\n"); } +static double normalize_height(double height) +{ + if (height < 0.25) + return 0.25; + else if (height < 0.50) + return 0.50; + else if (height < 0.75) + return 0.75; + else + return 0.100; +} + +void svg_ubox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges) +{ + double w = time2pixels(end) - time2pixels(start); + height = normalize_height(height); + + if (!svgfile) + return; + + fprintf(svgfile, "\n"); + fprintf(svgfile, "fd=%d error=%d merges=%d\n", fd, err, merges); + fprintf(svgfile, "\n", + time2pixels(start), + w, + Yslot * SLOT_MULT, + SLOT_HALF * height, + type); + fprintf(svgfile, "\n"); +} + +void svg_lbox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges) +{ + double w = time2pixels(end) - time2pixels(start); + height = normalize_height(height); + + if (!svgfile) + return; + + fprintf(svgfile, "\n"); + fprintf(svgfile, "fd=%d error=%d merges=%d\n", fd, err, merges); + fprintf(svgfile, "\n", + time2pixels(start), + w, + Yslot * SLOT_MULT + SLOT_HEIGHT - SLOT_HALF * height, + SLOT_HALF * height, + type); + fprintf(svgfile, "\n"); +} + +void svg_fbox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges) +{ + double w = time2pixels(end) - time2pixels(start); + height = normalize_height(height); + + if (!svgfile) + return; + + fprintf(svgfile, "\n"); + fprintf(svgfile, "fd=%d error=%d merges=%d\n", fd, err, merges); + fprintf(svgfile, "\n", + time2pixels(start), + w, + Yslot * SLOT_MULT + SLOT_HEIGHT - SLOT_HEIGHT * height, + SLOT_HEIGHT * height, + type); + fprintf(svgfile, "\n"); +} + void svg_box(int Yslot, u64 start, u64 end, const char *type) { if (!svgfile) @@ -543,6 +619,20 @@ static void svg_legenda_box(int X, const char *text, const char *style) X + boxsize + 5, boxsize, 0.8 * boxsize, text); } +void svg_io_legenda(void) +{ + if (!svgfile) + return; + + fprintf(svgfile, "\n"); + svg_legenda_box(0, "Disk", "disk"); + svg_legenda_box(100, "Network", "net"); + svg_legenda_box(200, "Sync", "sync"); + svg_legenda_box(300, "Poll", "poll"); + svg_legenda_box(400, "Error", "error"); + fprintf(svgfile, "\n"); +} + void svg_legenda(void) { if (!svgfile) @@ -559,7 +649,7 @@ void svg_legenda(void) fprintf(svgfile, "\n"); } -void svg_time_grid(void) +void svg_time_grid(double min_thickness) { u64 i; @@ -579,8 +669,10 @@ void svg_time_grid(void) color = 128; } - fprintf(svgfile, "\n", - time2pixels(i), SLOT_MULT/2, time2pixels(i), total_height, color, color, color, thickness); + if (thickness >= min_thickness) + fprintf(svgfile, "\n", + time2pixels(i), SLOT_MULT/2, time2pixels(i), + total_height, color, color, color, thickness); i += 10000000; } diff --git a/tools/perf/util/svghelper.h b/tools/perf/util/svghelper.h index e3aff53..9292a52 100644 --- a/tools/perf/util/svghelper.h +++ b/tools/perf/util/svghelper.h @@ -4,6 +4,9 @@ #include extern void open_svg(const char *filename, int cpus, int rows, u64 start, u64 end); +extern void svg_ubox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges); +extern void svg_lbox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges); +extern void svg_fbox(int Yslot, u64 start, u64 end, double height, const char *type, int fd, int err, int merges); extern void svg_box(int Yslot, u64 start, u64 end, const char *type); extern void svg_blocked(int Yslot, int cpu, u64 start, u64 end, const char *backtrace); extern void svg_running(int Yslot, int cpu, u64 start, u64 end, const char *backtrace); @@ -16,7 +19,8 @@ extern void svg_cstate(int cpu, u64 start, u64 end, int type); extern void svg_pstate(int cpu, u64 start, u64 end, u64 freq); -extern void svg_time_grid(void); +extern void svg_time_grid(double min_thickness); +extern void svg_io_legenda(void); extern void svg_legenda(void); extern void svg_wakeline(u64 start, int row1, int row2, const char *backtrace); extern void svg_partial_wakeline(u64 start, int row1, char *desc1, int row2, char *desc2, const char *backtrace); -- cgit v1.1 From 962e310af5dff432745ae7494096ed54b752e63e Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 8 Jul 2014 20:03:42 +0400 Subject: perf timechart: Conditionally update start_time on fork We don't need to overwrite current task start_time on fork, so update it only if it's zero. Signed-off-by: Stanislav Fomichev Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1404835423-23098-4-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Jiri Olsa --- tools/perf/builtin-timechart.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 4079062..37bf1eb 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -242,7 +242,7 @@ static void pid_fork(struct timechart *tchart, int pid, int ppid, u64 timestamp) pid_set_comm(tchart, pid, pp->current->comm); p->start_time = timestamp; - if (p->current) { + if (p->current && !p->current->start_time) { p->current->start_time = timestamp; p->current->state_since = timestamp; } -- cgit v1.1 From d243144af0b52fc5164a0823194f29a5979e236c Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 8 Jul 2014 20:03:43 +0400 Subject: perf timechart: Add more options to IO mode --io-skip-eagain - don't show EAGAIN errors --io-min-time - make small io bursts visible --io-merge-dist - merge adjacent events Signed-off-by: Stanislav Fomichev Acked-by: Namhyung Kim Link: http://lkml.kernel.org/n/1404835423-23098-5-git-send-email-stfomichev@yandex-team.ru Signed-off-by: Jiri Olsa --- tools/perf/Documentation/perf-timechart.txt | 13 +++++ tools/perf/builtin-timechart.c | 75 ++++++++++++++++++++++++++++- 2 files changed, 86 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-timechart.txt b/tools/perf/Documentation/perf-timechart.txt index ec6b46c..df98d1c 100644 --- a/tools/perf/Documentation/perf-timechart.txt +++ b/tools/perf/Documentation/perf-timechart.txt @@ -64,6 +64,19 @@ TIMECHART OPTIONS duration or tasks with given name. If number is given it's interpreted as number of nanoseconds. If non-numeric string is given it's interpreted as task name. +--io-skip-eagain:: + Don't draw EAGAIN IO events. +--io-min-time=:: + Draw small events as if they lasted min-time. Useful when you need + to see very small and fast IO. It's possible to specify ms or us + suffix to specify time in milliseconds or microseconds. + Default value is 1ms. +--io-merge-dist=:: + Merge events that are merge-dist nanoseconds apart. + Reduces number of figures on the SVG and makes it more render-friendly. + It's possible to specify ms or us suffix to specify time in + milliseconds or microseconds. + Default value is 1us. RECORD OPTIONS -------------- diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 37bf1eb..04c9c53 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -62,7 +62,10 @@ struct timechart { topology; /* IO related settings */ u64 io_events; - bool io_only; + bool io_only, + skip_eagain; + u64 min_time, + merge_dist; }; struct per_pidcomm; @@ -761,7 +764,7 @@ static int pid_end_io_sample(struct timechart *tchart, int pid, int type, { struct per_pid *p = find_create_pid(tchart, pid); struct per_pidcomm *c = p->current; - struct io_sample *sample; + struct io_sample *sample, *prev; if (!c) { pr_warning("Invalid pidcomm!\n"); @@ -785,6 +788,18 @@ static int pid_end_io_sample(struct timechart *tchart, int pid, int type, } sample->end_time = end; + prev = sample->next; + + /* we want to be able to see small and fast transfers, so make them + * at least min_time long, but don't overlap them */ + if (sample->end_time - sample->start_time < tchart->min_time) + sample->end_time = sample->start_time + tchart->min_time; + if (prev && sample->start_time < prev->end_time) { + if (prev->err) /* try to make errors more visible */ + sample->start_time = prev->end_time; + else + prev->end_time = sample->start_time; + } if (ret < 0) { sample->err = ret; @@ -799,6 +814,24 @@ static int pid_end_io_sample(struct timechart *tchart, int pid, int type, sample->bytes = ret; } + /* merge two requests to make svg smaller and render-friendly */ + if (prev && + prev->type == sample->type && + prev->err == sample->err && + prev->fd == sample->fd && + prev->end_time + tchart->merge_dist >= sample->start_time) { + + sample->bytes += prev->bytes; + sample->merges += prev->merges + 1; + + sample->start_time = prev->start_time; + sample->next = prev->next; + free(prev); + + if (!sample->err && sample->bytes > c->max_bytes) + c->max_bytes = sample->bytes; + } + tchart->io_events++; return 0; @@ -1119,6 +1152,10 @@ static void draw_io_bars(struct timechart *tchart) for (sample = c->io_samples; sample; sample = sample->next) { double h = (double)sample->bytes / c->max_bytes; + if (tchart->skip_eagain && + sample->err == -EAGAIN) + continue; + if (sample->err) h = 1; @@ -1849,6 +1886,30 @@ parse_highlight(const struct option *opt __maybe_unused, const char *arg, return 0; } +static int +parse_time(const struct option *opt, const char *arg, int __maybe_unused unset) +{ + char unit = 'n'; + u64 *value = opt->value; + + if (sscanf(arg, "%" PRIu64 "%cs", value, &unit) > 0) { + switch (unit) { + case 'm': + *value *= 1000000; + break; + case 'u': + *value *= 1000; + break; + case 'n': + break; + default: + return -1; + } + } + + return 0; +} + int cmd_timechart(int argc, const char **argv, const char *prefix __maybe_unused) { @@ -1861,6 +1922,8 @@ int cmd_timechart(int argc, const char **argv, .ordered_samples = true, }, .proc_num = 15, + .min_time = 1000000, + .merge_dist = 1000, }; const char *output_name = "output.svg"; const struct option timechart_options[] = { @@ -1882,6 +1945,14 @@ int cmd_timechart(int argc, const char **argv, "min. number of tasks to print"), OPT_BOOLEAN('t', "topology", &tchart.topology, "sort CPUs according to topology"), + OPT_BOOLEAN(0, "io-skip-eagain", &tchart.skip_eagain, + "skip EAGAIN errors"), + OPT_CALLBACK(0, "io-min-time", &tchart.min_time, "time", + "all IO faster than min-time will visually appear longer", + parse_time), + OPT_CALLBACK(0, "io-merge-dist", &tchart.merge_dist, "time", + "merge events that are merge-dist us apart", + parse_time), OPT_END() }; const char * const timechart_usage[] = { -- cgit v1.1 From 44b3802122174ba499613bac3aab2e66e948ce1e Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 18:29:04 +0400 Subject: perf kvm: Use defines of kvm events Currently perf-kvm uses string literals for kvm event names, but it works only for x86, because other architectures may have other names for those events. To reduce dependence on architecture, we add file with defines for: - kvm_entry and kvm_exit events, - exit reason field name in kvm_exit event, - length of exit reasons strings, - vcpu_id field name in kvm trace events, and replace literals in perf-kvm. Reviewed-by: Cornelia Huck Reviewed-by David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404397747-20939-2-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/MANIFEST | 1 + tools/perf/builtin-kvm.c | 34 ++++++++++++++++------------------ 2 files changed, 17 insertions(+), 18 deletions(-) (limited to 'tools') diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST index 45da209..02b485d 100644 --- a/tools/perf/MANIFEST +++ b/tools/perf/MANIFEST @@ -37,3 +37,4 @@ arch/x86/include/asm/kvm_host.h arch/x86/include/uapi/asm/svm.h arch/x86/include/uapi/asm/vmx.h arch/x86/include/uapi/asm/kvm.h +arch/x86/include/uapi/asm/kvm_perf.h diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 41dbeaf..6d73346 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -30,9 +30,7 @@ #include #ifdef HAVE_KVM_STAT_SUPPORT -#include -#include -#include +#include struct event_key { #define INVALID_KEY (~0ULL) @@ -75,7 +73,7 @@ struct kvm_events_ops { bool (*is_end_event)(struct perf_evsel *evsel, struct perf_sample *sample, struct event_key *key); void (*decode_key)(struct perf_kvm_stat *kvm, struct event_key *key, - char decode[20]); + char *decode); const char *name; }; @@ -126,12 +124,12 @@ static void exit_event_get_key(struct perf_evsel *evsel, struct event_key *key) { key->info = 0; - key->key = perf_evsel__intval(evsel, sample, "exit_reason"); + key->key = perf_evsel__intval(evsel, sample, KVM_EXIT_REASON); } static bool kvm_exit_event(struct perf_evsel *evsel) { - return !strcmp(evsel->name, "kvm:kvm_exit"); + return !strcmp(evsel->name, KVM_EXIT_TRACE); } static bool exit_event_begin(struct perf_evsel *evsel, @@ -147,7 +145,7 @@ static bool exit_event_begin(struct perf_evsel *evsel, static bool kvm_entry_event(struct perf_evsel *evsel) { - return !strcmp(evsel->name, "kvm:kvm_entry"); + return !strcmp(evsel->name, KVM_ENTRY_TRACE); } static bool exit_event_end(struct perf_evsel *evsel, @@ -182,12 +180,12 @@ static const char *get_exit_reason(struct perf_kvm_stat *kvm, static void exit_event_decode_key(struct perf_kvm_stat *kvm, struct event_key *key, - char decode[20]) + char *decode) { const char *exit_reason = get_exit_reason(kvm, kvm->exit_reasons, key->key); - scnprintf(decode, 20, "%s", exit_reason); + scnprintf(decode, DECODE_STR_LEN, "%s", exit_reason); } static struct kvm_events_ops exit_events = { @@ -249,9 +247,9 @@ static bool mmio_event_end(struct perf_evsel *evsel, struct perf_sample *sample, static void mmio_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, struct event_key *key, - char decode[20]) + char *decode) { - scnprintf(decode, 20, "%#lx:%s", (unsigned long)key->key, + scnprintf(decode, DECODE_STR_LEN, "%#lx:%s", (unsigned long)key->key, key->info == KVM_TRACE_MMIO_WRITE ? "W" : "R"); } @@ -292,9 +290,9 @@ static bool ioport_event_end(struct perf_evsel *evsel, static void ioport_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, struct event_key *key, - char decode[20]) + char *decode) { - scnprintf(decode, 20, "%#llx:%s", (unsigned long long)key->key, + scnprintf(decode, DECODE_STR_LEN, "%#llx:%s", (unsigned long long)key->key, key->info ? "POUT" : "PIN"); } @@ -524,7 +522,7 @@ static bool handle_end_event(struct perf_kvm_stat *kvm, time_diff = sample->time - time_begin; if (kvm->duration && time_diff > kvm->duration) { - char decode[32]; + char decode[DECODE_STR_LEN]; kvm->events_ops->decode_key(kvm, &event->key, decode); if (strcmp(decode, "HLT")) { @@ -552,7 +550,7 @@ struct vcpu_event_record *per_vcpu_record(struct thread *thread, return NULL; } - vcpu_record->vcpu_id = perf_evsel__intval(evsel, sample, "vcpu_id"); + vcpu_record->vcpu_id = perf_evsel__intval(evsel, sample, VCPU_ID); thread->priv = vcpu_record; } @@ -739,7 +737,7 @@ static void show_timeofday(void) static void print_result(struct perf_kvm_stat *kvm) { - char decode[20]; + char decode[DECODE_STR_LEN]; struct kvm_event *event; int vcpu = kvm->trace_vcpu; @@ -750,7 +748,7 @@ static void print_result(struct perf_kvm_stat *kvm) pr_info("\n\n"); print_vcpu_info(kvm); - pr_info("%20s ", kvm->events_ops->name); + pr_info("%*s ", DECODE_STR_LEN, kvm->events_ops->name); pr_info("%10s ", "Samples"); pr_info("%9s ", "Samples%"); @@ -769,7 +767,7 @@ static void print_result(struct perf_kvm_stat *kvm) min = get_event_min(event, vcpu); kvm->events_ops->decode_key(kvm, &event->key, decode); - pr_info("%20s ", decode); + pr_info("%*s ", DECODE_STR_LEN, decode); pr_info("%10llu ", (unsigned long long)ecount); pr_info("%8.2f%% ", (double)ecount / kvm->total_count * 100); pr_info("%8.2f%% ", (double)etime / kvm->total_time * 100); -- cgit v1.1 From 9daa81239e60c162153fb2a365b8492c9a9bf632 Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 18:29:05 +0400 Subject: perf kvm: Move arch specific code into arch/ Parts of a 'perf kvm stat' code make sense only for x86. Let's move this code into the arch/x86/kvm-stat.c file and add util/kvm-stat.h for generic structure definitions. Add a global array 'kvm_reg_events_ops' for accessing the arch-specific 'kvm_events_ops' from generic code. Since the several global arrays (i.e. 'kvm_events_tp') have been moved to arch/*, we can not know their sizes and use them directly in builtin-kvm.c. This patch fixes that problem by adding trimming NULL element to each array and changing the behavior of their handlers in generic code. Reviewed-by: David Ahern Reviewed-by: Cornelia Huck Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404397747-20939-3-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Makefile.perf | 1 + tools/perf/arch/x86/Makefile | 1 + tools/perf/arch/x86/util/kvm-stat.c | 151 ++++++++++++++++++ tools/perf/builtin-kvm.c | 297 +++++------------------------------- tools/perf/util/kvm-stat.h | 130 ++++++++++++++++ 5 files changed, 317 insertions(+), 263 deletions(-) create mode 100644 tools/perf/arch/x86/util/kvm-stat.c create mode 100644 tools/perf/util/kvm-stat.h (limited to 'tools') diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf index 9670a16..90c4983 100644 --- a/tools/perf/Makefile.perf +++ b/tools/perf/Makefile.perf @@ -300,6 +300,7 @@ LIB_H += ui/progress.h LIB_H += ui/util.h LIB_H += ui/ui.h LIB_H += util/data.h +LIB_H += util/kvm-stat.h LIB_OBJS += $(OUTPUT)util/abspath.o LIB_OBJS += $(OUTPUT)util/alias.o diff --git a/tools/perf/arch/x86/Makefile b/tools/perf/arch/x86/Makefile index d393901..9b21881 100644 --- a/tools/perf/arch/x86/Makefile +++ b/tools/perf/arch/x86/Makefile @@ -16,3 +16,4 @@ LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/header.o LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/tsc.o LIB_H += arch/$(ARCH)/util/tsc.h HAVE_KVM_STAT_SUPPORT := 1 +LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/kvm-stat.o diff --git a/tools/perf/arch/x86/util/kvm-stat.c b/tools/perf/arch/x86/util/kvm-stat.c new file mode 100644 index 0000000..2f8d2c1 --- /dev/null +++ b/tools/perf/arch/x86/util/kvm-stat.c @@ -0,0 +1,151 @@ +#include "../../util/kvm-stat.h" +#include + +define_exit_reasons_table(vmx_exit_reasons, VMX_EXIT_REASONS); +define_exit_reasons_table(svm_exit_reasons, SVM_EXIT_REASONS); + +static struct kvm_events_ops exit_events = { + .is_begin_event = exit_event_begin, + .is_end_event = exit_event_end, + .decode_key = exit_event_decode_key, + .name = "VM-EXIT" +}; + +/* + * For the mmio events, we treat: + * the time of MMIO write: kvm_mmio(KVM_TRACE_MMIO_WRITE...) -> kvm_entry + * the time of MMIO read: kvm_exit -> kvm_mmio(KVM_TRACE_MMIO_READ...). + */ +static void mmio_event_get_key(struct perf_evsel *evsel, struct perf_sample *sample, + struct event_key *key) +{ + key->key = perf_evsel__intval(evsel, sample, "gpa"); + key->info = perf_evsel__intval(evsel, sample, "type"); +} + +#define KVM_TRACE_MMIO_READ_UNSATISFIED 0 +#define KVM_TRACE_MMIO_READ 1 +#define KVM_TRACE_MMIO_WRITE 2 + +static bool mmio_event_begin(struct perf_evsel *evsel, + struct perf_sample *sample, struct event_key *key) +{ + /* MMIO read begin event in kernel. */ + if (kvm_exit_event(evsel)) + return true; + + /* MMIO write begin event in kernel. */ + if (!strcmp(evsel->name, "kvm:kvm_mmio") && + perf_evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_WRITE) { + mmio_event_get_key(evsel, sample, key); + return true; + } + + return false; +} + +static bool mmio_event_end(struct perf_evsel *evsel, struct perf_sample *sample, + struct event_key *key) +{ + /* MMIO write end event in kernel. */ + if (kvm_entry_event(evsel)) + return true; + + /* MMIO read end event in kernel.*/ + if (!strcmp(evsel->name, "kvm:kvm_mmio") && + perf_evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_READ) { + mmio_event_get_key(evsel, sample, key); + return true; + } + + return false; +} + +static void mmio_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, + struct event_key *key, + char *decode) +{ + scnprintf(decode, DECODE_STR_LEN, "%#lx:%s", + (unsigned long)key->key, + key->info == KVM_TRACE_MMIO_WRITE ? "W" : "R"); +} + +static struct kvm_events_ops mmio_events = { + .is_begin_event = mmio_event_begin, + .is_end_event = mmio_event_end, + .decode_key = mmio_event_decode_key, + .name = "MMIO Access" +}; + + /* The time of emulation pio access is from kvm_pio to kvm_entry. */ +static void ioport_event_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + key->key = perf_evsel__intval(evsel, sample, "port"); + key->info = perf_evsel__intval(evsel, sample, "rw"); +} + +static bool ioport_event_begin(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + if (!strcmp(evsel->name, "kvm:kvm_pio")) { + ioport_event_get_key(evsel, sample, key); + return true; + } + + return false; +} + +static bool ioport_event_end(struct perf_evsel *evsel, + struct perf_sample *sample __maybe_unused, + struct event_key *key __maybe_unused) +{ + return kvm_entry_event(evsel); +} + +static void ioport_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, + struct event_key *key, + char *decode) +{ + scnprintf(decode, DECODE_STR_LEN, "%#llx:%s", + (unsigned long long)key->key, + key->info ? "POUT" : "PIN"); +} + +static struct kvm_events_ops ioport_events = { + .is_begin_event = ioport_event_begin, + .is_end_event = ioport_event_end, + .decode_key = ioport_event_decode_key, + .name = "IO Port Access" +}; + +const char * const kvm_events_tp[] = { + "kvm:kvm_entry", + "kvm:kvm_exit", + "kvm:kvm_mmio", + "kvm:kvm_pio", + NULL, +}; + +struct kvm_reg_events_ops kvm_reg_events_ops[] = { + { .name = "vmexit", .ops = &exit_events }, + { .name = "mmio", .ops = &mmio_events }, + { .name = "ioport", .ops = &ioport_events }, + { NULL, NULL }, +}; + +int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid) +{ + if (strstr(cpuid, "Intel")) { + kvm->exit_reasons = vmx_exit_reasons; + kvm->exit_reasons_isa = "VMX"; + } else if (strstr(cpuid, "AMD")) { + kvm->exit_reasons = svm_exit_reasons; + kvm->exit_reasons_isa = "SVM"; + } else + return -ENOTSUP; + + return 0; +} diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 6d73346..75ee8c1 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -31,109 +31,23 @@ #ifdef HAVE_KVM_STAT_SUPPORT #include +#include "util/kvm-stat.h" -struct event_key { - #define INVALID_KEY (~0ULL) - u64 key; - int info; -}; - -struct kvm_event_stats { - u64 time; - struct stats stats; -}; - -struct kvm_event { - struct list_head hash_entry; - struct rb_node rb; - - struct event_key key; - - struct kvm_event_stats total; - - #define DEFAULT_VCPU_NUM 8 - int max_vcpu; - struct kvm_event_stats *vcpu; -}; - -typedef int (*key_cmp_fun)(struct kvm_event*, struct kvm_event*, int); - -struct kvm_event_key { - const char *name; - key_cmp_fun key; -}; - - -struct perf_kvm_stat; - -struct kvm_events_ops { - bool (*is_begin_event)(struct perf_evsel *evsel, - struct perf_sample *sample, - struct event_key *key); - bool (*is_end_event)(struct perf_evsel *evsel, - struct perf_sample *sample, struct event_key *key); - void (*decode_key)(struct perf_kvm_stat *kvm, struct event_key *key, - char *decode); - const char *name; -}; - -struct exit_reasons_table { - unsigned long exit_code; - const char *reason; -}; - -#define EVENTS_BITS 12 -#define EVENTS_CACHE_SIZE (1UL << EVENTS_BITS) - -struct perf_kvm_stat { - struct perf_tool tool; - struct record_opts opts; - struct perf_evlist *evlist; - struct perf_session *session; - - const char *file_name; - const char *report_event; - const char *sort_key; - int trace_vcpu; - - struct exit_reasons_table *exit_reasons; - const char *exit_reasons_isa; - - struct kvm_events_ops *events_ops; - key_cmp_fun compare; - struct list_head kvm_events_cache[EVENTS_CACHE_SIZE]; - - u64 total_time; - u64 total_count; - u64 lost_events; - u64 duration; - - const char *pid_str; - struct intlist *pid_list; - - struct rb_root result; - - int timerfd; - unsigned int display_time; - bool live; -}; - - -static void exit_event_get_key(struct perf_evsel *evsel, - struct perf_sample *sample, - struct event_key *key) +void exit_event_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) { key->info = 0; key->key = perf_evsel__intval(evsel, sample, KVM_EXIT_REASON); } -static bool kvm_exit_event(struct perf_evsel *evsel) +bool kvm_exit_event(struct perf_evsel *evsel) { return !strcmp(evsel->name, KVM_EXIT_TRACE); } -static bool exit_event_begin(struct perf_evsel *evsel, - struct perf_sample *sample, struct event_key *key) +bool exit_event_begin(struct perf_evsel *evsel, + struct perf_sample *sample, struct event_key *key) { if (kvm_exit_event(evsel)) { exit_event_get_key(evsel, sample, key); @@ -143,26 +57,18 @@ static bool exit_event_begin(struct perf_evsel *evsel, return false; } -static bool kvm_entry_event(struct perf_evsel *evsel) +bool kvm_entry_event(struct perf_evsel *evsel) { return !strcmp(evsel->name, KVM_ENTRY_TRACE); } -static bool exit_event_end(struct perf_evsel *evsel, - struct perf_sample *sample __maybe_unused, - struct event_key *key __maybe_unused) +bool exit_event_end(struct perf_evsel *evsel, + struct perf_sample *sample __maybe_unused, + struct event_key *key __maybe_unused) { return kvm_entry_event(evsel); } -#define define_exit_reasons_table(name, symbols) \ - static struct exit_reasons_table name[] = { \ - symbols, { -1, NULL } \ - } - -define_exit_reasons_table(vmx_exit_reasons, VMX_EXIT_REASONS); -define_exit_reasons_table(svm_exit_reasons, SVM_EXIT_REASONS); - static const char *get_exit_reason(struct perf_kvm_stat *kvm, struct exit_reasons_table *tbl, u64 exit_code) @@ -178,9 +84,9 @@ static const char *get_exit_reason(struct perf_kvm_stat *kvm, return "UNKNOWN"; } -static void exit_event_decode_key(struct perf_kvm_stat *kvm, - struct event_key *key, - char *decode) +void exit_event_decode_key(struct perf_kvm_stat *kvm, + struct event_key *key, + char *decode) { const char *exit_reason = get_exit_reason(kvm, kvm->exit_reasons, key->key); @@ -188,139 +94,20 @@ static void exit_event_decode_key(struct perf_kvm_stat *kvm, scnprintf(decode, DECODE_STR_LEN, "%s", exit_reason); } -static struct kvm_events_ops exit_events = { - .is_begin_event = exit_event_begin, - .is_end_event = exit_event_end, - .decode_key = exit_event_decode_key, - .name = "VM-EXIT" -}; - -/* - * For the mmio events, we treat: - * the time of MMIO write: kvm_mmio(KVM_TRACE_MMIO_WRITE...) -> kvm_entry - * the time of MMIO read: kvm_exit -> kvm_mmio(KVM_TRACE_MMIO_READ...). - */ -static void mmio_event_get_key(struct perf_evsel *evsel, struct perf_sample *sample, - struct event_key *key) -{ - key->key = perf_evsel__intval(evsel, sample, "gpa"); - key->info = perf_evsel__intval(evsel, sample, "type"); -} - -#define KVM_TRACE_MMIO_READ_UNSATISFIED 0 -#define KVM_TRACE_MMIO_READ 1 -#define KVM_TRACE_MMIO_WRITE 2 - -static bool mmio_event_begin(struct perf_evsel *evsel, - struct perf_sample *sample, struct event_key *key) -{ - /* MMIO read begin event in kernel. */ - if (kvm_exit_event(evsel)) - return true; - - /* MMIO write begin event in kernel. */ - if (!strcmp(evsel->name, "kvm:kvm_mmio") && - perf_evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_WRITE) { - mmio_event_get_key(evsel, sample, key); - return true; - } - - return false; -} - -static bool mmio_event_end(struct perf_evsel *evsel, struct perf_sample *sample, - struct event_key *key) -{ - /* MMIO write end event in kernel. */ - if (kvm_entry_event(evsel)) - return true; - - /* MMIO read end event in kernel.*/ - if (!strcmp(evsel->name, "kvm:kvm_mmio") && - perf_evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_READ) { - mmio_event_get_key(evsel, sample, key); - return true; - } - - return false; -} - -static void mmio_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, - struct event_key *key, - char *decode) -{ - scnprintf(decode, DECODE_STR_LEN, "%#lx:%s", (unsigned long)key->key, - key->info == KVM_TRACE_MMIO_WRITE ? "W" : "R"); -} - -static struct kvm_events_ops mmio_events = { - .is_begin_event = mmio_event_begin, - .is_end_event = mmio_event_end, - .decode_key = mmio_event_decode_key, - .name = "MMIO Access" -}; - - /* The time of emulation pio access is from kvm_pio to kvm_entry. */ -static void ioport_event_get_key(struct perf_evsel *evsel, - struct perf_sample *sample, - struct event_key *key) +static bool register_kvm_events_ops(struct perf_kvm_stat *kvm) { - key->key = perf_evsel__intval(evsel, sample, "port"); - key->info = perf_evsel__intval(evsel, sample, "rw"); -} + struct kvm_reg_events_ops *events_ops = kvm_reg_events_ops; -static bool ioport_event_begin(struct perf_evsel *evsel, - struct perf_sample *sample, - struct event_key *key) -{ - if (!strcmp(evsel->name, "kvm:kvm_pio")) { - ioport_event_get_key(evsel, sample, key); - return true; + for (events_ops = kvm_reg_events_ops; events_ops->name; events_ops++) { + if (!strcmp(events_ops->name, kvm->report_event)) { + kvm->events_ops = events_ops->ops; + return true; + } } return false; } -static bool ioport_event_end(struct perf_evsel *evsel, - struct perf_sample *sample __maybe_unused, - struct event_key *key __maybe_unused) -{ - return kvm_entry_event(evsel); -} - -static void ioport_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused, - struct event_key *key, - char *decode) -{ - scnprintf(decode, DECODE_STR_LEN, "%#llx:%s", (unsigned long long)key->key, - key->info ? "POUT" : "PIN"); -} - -static struct kvm_events_ops ioport_events = { - .is_begin_event = ioport_event_begin, - .is_end_event = ioport_event_end, - .decode_key = ioport_event_decode_key, - .name = "IO Port Access" -}; - -static bool register_kvm_events_ops(struct perf_kvm_stat *kvm) -{ - bool ret = true; - - if (!strcmp(kvm->report_event, "vmexit")) - kvm->events_ops = &exit_events; - else if (!strcmp(kvm->report_event, "mmio")) - kvm->events_ops = &mmio_events; - else if (!strcmp(kvm->report_event, "ioport")) - kvm->events_ops = &ioport_events; - else { - pr_err("Unknown report event:%s\n", kvm->report_event); - ret = false; - } - - return ret; -} - struct vcpu_event_record { int vcpu_id; u64 start_time; @@ -833,20 +620,6 @@ static int process_sample_event(struct perf_tool *tool, return 0; } -static int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid) -{ - if (strstr(cpuid, "Intel")) { - kvm->exit_reasons = vmx_exit_reasons; - kvm->exit_reasons_isa = "VMX"; - } else if (strstr(cpuid, "AMD")) { - kvm->exit_reasons = svm_exit_reasons; - kvm->exit_reasons_isa = "SVM"; - } else - return -ENOTSUP; - - return 0; -} - static int cpu_isa_config(struct perf_kvm_stat *kvm) { char buf[64], *cpuid; @@ -1305,13 +1078,6 @@ exit: return ret; } -static const char * const kvm_events_tp[] = { - "kvm:kvm_entry", - "kvm:kvm_exit", - "kvm:kvm_mmio", - "kvm:kvm_pio", -}; - #define STRDUP_FAIL_EXIT(s) \ ({ char *_p; \ _p = strdup(s); \ @@ -1323,7 +1089,7 @@ static const char * const kvm_events_tp[] = { static int kvm_events_record(struct perf_kvm_stat *kvm, int argc, const char **argv) { - unsigned int rec_argc, i, j; + unsigned int rec_argc, i, j, events_tp_size; const char **rec_argv; const char * const record_args[] = { "record", @@ -1331,9 +1097,14 @@ kvm_events_record(struct perf_kvm_stat *kvm, int argc, const char **argv) "-m", "1024", "-c", "1", }; + const char * const *events_tp; + events_tp_size = 0; + + for (events_tp = kvm_events_tp; *events_tp; events_tp++) + events_tp_size++; rec_argc = ARRAY_SIZE(record_args) + argc + 2 + - 2 * ARRAY_SIZE(kvm_events_tp); + 2 * events_tp_size; rec_argv = calloc(rec_argc + 1, sizeof(char *)); if (rec_argv == NULL) @@ -1342,7 +1113,7 @@ kvm_events_record(struct perf_kvm_stat *kvm, int argc, const char **argv) for (i = 0; i < ARRAY_SIZE(record_args); i++) rec_argv[i] = STRDUP_FAIL_EXIT(record_args[i]); - for (j = 0; j < ARRAY_SIZE(kvm_events_tp); j++) { + for (j = 0; j < events_tp_size; j++) { rec_argv[i++] = "-e"; rec_argv[i++] = STRDUP_FAIL_EXIT(kvm_events_tp[j]); } @@ -1396,16 +1167,16 @@ static struct perf_evlist *kvm_live_event_list(void) { struct perf_evlist *evlist; char *tp, *name, *sys; - unsigned int j; int err = -1; + const char * const *events_tp; evlist = perf_evlist__new(); if (evlist == NULL) return NULL; - for (j = 0; j < ARRAY_SIZE(kvm_events_tp); j++) { + for (events_tp = kvm_events_tp; *events_tp; events_tp++) { - tp = strdup(kvm_events_tp[j]); + tp = strdup(*events_tp); if (tp == NULL) goto out; @@ -1414,7 +1185,7 @@ static struct perf_evlist *kvm_live_event_list(void) name = strchr(tp, ':'); if (name == NULL) { pr_err("Error parsing %s tracepoint: subsystem delimiter not found\n", - kvm_events_tp[j]); + *events_tp); free(tp); goto out; } @@ -1422,7 +1193,7 @@ static struct perf_evlist *kvm_live_event_list(void) name++; if (perf_evlist__add_newtp(evlist, sys, name, NULL)) { - pr_err("Failed to add %s tracepoint to the list\n", kvm_events_tp[j]); + pr_err("Failed to add %s tracepoint to the list\n", *events_tp); free(tp); goto out; } diff --git a/tools/perf/util/kvm-stat.h b/tools/perf/util/kvm-stat.h new file mode 100644 index 0000000..d0d9fb1 --- /dev/null +++ b/tools/perf/util/kvm-stat.h @@ -0,0 +1,130 @@ +#ifndef __PERF_KVM_STAT_H +#define __PERF_KVM_STAT_H + +#include "../perf.h" +#include "evsel.h" +#include "evlist.h" +#include "session.h" +#include "tool.h" +#include "stat.h" + +struct event_key { + #define INVALID_KEY (~0ULL) + u64 key; + int info; +}; + +struct kvm_event_stats { + u64 time; + struct stats stats; +}; + +struct kvm_event { + struct list_head hash_entry; + struct rb_node rb; + + struct event_key key; + + struct kvm_event_stats total; + + #define DEFAULT_VCPU_NUM 8 + int max_vcpu; + struct kvm_event_stats *vcpu; +}; + +typedef int (*key_cmp_fun)(struct kvm_event*, struct kvm_event*, int); + +struct kvm_event_key { + const char *name; + key_cmp_fun key; +}; + +struct perf_kvm_stat; + +struct kvm_events_ops { + bool (*is_begin_event)(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key); + bool (*is_end_event)(struct perf_evsel *evsel, + struct perf_sample *sample, struct event_key *key); + void (*decode_key)(struct perf_kvm_stat *kvm, struct event_key *key, + char *decode); + const char *name; +}; + +struct exit_reasons_table { + unsigned long exit_code; + const char *reason; +}; + +#define EVENTS_BITS 12 +#define EVENTS_CACHE_SIZE (1UL << EVENTS_BITS) + +struct perf_kvm_stat { + struct perf_tool tool; + struct record_opts opts; + struct perf_evlist *evlist; + struct perf_session *session; + + const char *file_name; + const char *report_event; + const char *sort_key; + int trace_vcpu; + + struct exit_reasons_table *exit_reasons; + const char *exit_reasons_isa; + + struct kvm_events_ops *events_ops; + key_cmp_fun compare; + struct list_head kvm_events_cache[EVENTS_CACHE_SIZE]; + + u64 total_time; + u64 total_count; + u64 lost_events; + u64 duration; + + const char *pid_str; + struct intlist *pid_list; + + struct rb_root result; + + int timerfd; + unsigned int display_time; + bool live; +}; + +struct kvm_reg_events_ops { + const char *name; + struct kvm_events_ops *ops; +}; + +void exit_event_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key); +bool exit_event_begin(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key); +bool exit_event_end(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key); +void exit_event_decode_key(struct perf_kvm_stat *kvm, + struct event_key *key, + char *decode); + +bool kvm_exit_event(struct perf_evsel *evsel); +bool kvm_entry_event(struct perf_evsel *evsel); + +#define define_exit_reasons_table(name, symbols) \ + static struct exit_reasons_table name[] = { \ + symbols, { -1, NULL } \ + } + +/* + * arch specific callbacks and data structures + */ +int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid); + +extern const char * const kvm_events_tp[]; +extern struct kvm_reg_events_ops kvm_reg_events_ops[]; + +#endif /* __PERF_KVM_STAT_H */ -- cgit v1.1 From 54c801ff71ba9c9ae41871e226b9d846ff9c6bab Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 18:29:06 +0400 Subject: perf kvm: Add skip_event() for --duration option Current code skips output of the x86 specific HLT event in order to avoid flooding the output with enabled --duration option. The events to be skipped should be architecture dependent, though. Let's add an architecture specific array of events to be skipped and introduce a skip_event() function checking against that array. Reviewed-by: Christian Borntraeger Reviewed-by: Cornelia Huck Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404397747-20939-4-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/x86/util/kvm-stat.c | 5 +++++ tools/perf/builtin-kvm.c | 13 ++++++++++++- tools/perf/util/kvm-stat.h | 1 + 3 files changed, 18 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/arch/x86/util/kvm-stat.c b/tools/perf/arch/x86/util/kvm-stat.c index 2f8d2c1..14e4e66 100644 --- a/tools/perf/arch/x86/util/kvm-stat.c +++ b/tools/perf/arch/x86/util/kvm-stat.c @@ -136,6 +136,11 @@ struct kvm_reg_events_ops kvm_reg_events_ops[] = { { NULL, NULL }, }; +const char * const kvm_skip_events[] = { + "HLT", + NULL, +}; + int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid) { if (strstr(cpuid, "Intel")) { diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 75ee8c1..fc2d63d 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -261,6 +261,17 @@ static bool update_kvm_event(struct kvm_event *event, int vcpu_id, return true; } +static bool skip_event(const char *event) +{ + const char * const *skip_events; + + for (skip_events = kvm_skip_events; *skip_events; skip_events++) + if (!strcmp(event, *skip_events)) + return true; + + return false; +} + static bool handle_end_event(struct perf_kvm_stat *kvm, struct vcpu_event_record *vcpu_record, struct event_key *key, @@ -312,7 +323,7 @@ static bool handle_end_event(struct perf_kvm_stat *kvm, char decode[DECODE_STR_LEN]; kvm->events_ops->decode_key(kvm, &event->key, decode); - if (strcmp(decode, "HLT")) { + if (!skip_event(decode)) { pr_info("%" PRIu64 " VM %d, vcpu %d: %s event took %" PRIu64 "usec\n", sample->time, sample->pid, vcpu_record->vcpu_id, decode, time_diff/1000); diff --git a/tools/perf/util/kvm-stat.h b/tools/perf/util/kvm-stat.h index d0d9fb1..ba937ca 100644 --- a/tools/perf/util/kvm-stat.h +++ b/tools/perf/util/kvm-stat.h @@ -126,5 +126,6 @@ int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid); extern const char * const kvm_events_tp[]; extern struct kvm_reg_events_ops kvm_reg_events_ops[]; +extern const char * const kvm_skip_events[]; #endif /* __PERF_KVM_STAT_H */ -- cgit v1.1 From 3be8e2a0a53c3179a44a933614f6a893da0b5c19 Mon Sep 17 00:00:00 2001 From: Alexander Yarygin Date: Thu, 3 Jul 2014 18:29:07 +0400 Subject: perf kvm: Add stat support on s390 On s390, the vmexit event has a tree-like structure: between exit_event_begin and exit_event_end several other events may happen and with each of them refining the previous ones. This patch adds a decoder for such events to the generic code and also the files and kvm-stat.c for s390. Commands 'perf kvm stat record', 'report' and 'live' are supported. Reviewed-by: David Ahern Signed-off-by: Alexander Yarygin Acked-by: Christian Borntraeger Cc: Christian Borntraeger Cc: Cornelia Huck Cc: David Ahern Cc: Ingo Molnar Cc: Jiri Olsa Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1404397747-20939-5-git-send-email-yarygin@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-kvm.txt | 16 +++--- tools/perf/MANIFEST | 2 + tools/perf/arch/s390/Makefile | 2 + tools/perf/arch/s390/util/kvm-stat.c | 105 ++++++++++++++++++++++++++++++++++ tools/perf/builtin-kvm.c | 52 +++++++++++++++-- tools/perf/util/kvm-stat.h | 9 +++ 6 files changed, 175 insertions(+), 11 deletions(-) create mode 100644 tools/perf/arch/s390/util/kvm-stat.c (limited to 'tools') diff --git a/tools/perf/Documentation/perf-kvm.txt b/tools/perf/Documentation/perf-kvm.txt index 52276a6..6e689dc 100644 --- a/tools/perf/Documentation/perf-kvm.txt +++ b/tools/perf/Documentation/perf-kvm.txt @@ -51,9 +51,9 @@ There are a couple of variants of perf kvm: 'perf kvm stat ' to run a command and gather performance counter statistics. Especially, perf 'kvm stat record/report' generates a statistical analysis - of KVM events. Currently, vmexit, mmio and ioport events are supported. - 'perf kvm stat record ' records kvm events and the events between - start and end . + of KVM events. Currently, vmexit, mmio (x86 only) and ioport (x86 only) + events are supported. 'perf kvm stat record ' records kvm events + and the events between start and end . And this command produces a file which contains tracing results of kvm events. @@ -103,8 +103,8 @@ STAT REPORT OPTIONS analyze events which occures on this vcpu. (default: all vcpus) --event=:: - event to be analyzed. Possible values: vmexit, mmio, ioport. - (default: vmexit) + event to be analyzed. Possible values: vmexit, mmio (x86 only), + ioport (x86 only). (default: vmexit) -k:: --key=:: Sorting key. Possible values: sample (default, sort by samples @@ -138,7 +138,8 @@ STAT LIVE OPTIONS --event=:: - event to be analyzed. Possible values: vmexit, mmio, ioport. + event to be analyzed. Possible values: vmexit, + mmio (x86 only), ioport (x86 only). (default: vmexit) -k:: @@ -147,7 +148,8 @@ STAT LIVE OPTIONS number), time (sort by average time). --duration=:: - Show events other than HLT that take longer than duration usecs. + Show events other than HLT (x86 only) or Wait state (s390 only) + that take longer than duration usecs. SEE ALSO -------- diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST index 02b485d..344c4d3 100644 --- a/tools/perf/MANIFEST +++ b/tools/perf/MANIFEST @@ -38,3 +38,5 @@ arch/x86/include/uapi/asm/svm.h arch/x86/include/uapi/asm/vmx.h arch/x86/include/uapi/asm/kvm.h arch/x86/include/uapi/asm/kvm_perf.h +arch/s390/include/uapi/asm/sie.h +arch/s390/include/uapi/asm/kvm_perf.h diff --git a/tools/perf/arch/s390/Makefile b/tools/perf/arch/s390/Makefile index 744e629..798ac73 100644 --- a/tools/perf/arch/s390/Makefile +++ b/tools/perf/arch/s390/Makefile @@ -3,3 +3,5 @@ PERF_HAVE_DWARF_REGS := 1 LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/dwarf-regs.o endif LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/header.o +HAVE_KVM_STAT_SUPPORT := 1 +LIB_OBJS += $(OUTPUT)arch/$(ARCH)/util/kvm-stat.o diff --git a/tools/perf/arch/s390/util/kvm-stat.c b/tools/perf/arch/s390/util/kvm-stat.c new file mode 100644 index 0000000..a5dbc07 --- /dev/null +++ b/tools/perf/arch/s390/util/kvm-stat.c @@ -0,0 +1,105 @@ +/* + * Arch specific functions for perf kvm stat. + * + * Copyright 2014 IBM Corp. + * Author(s): Alexander Yarygin + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License (version 2 only) + * as published by the Free Software Foundation. + */ + +#include "../../util/kvm-stat.h" +#include + +define_exit_reasons_table(sie_exit_reasons, sie_intercept_code); +define_exit_reasons_table(sie_icpt_insn_codes, icpt_insn_codes); +define_exit_reasons_table(sie_sigp_order_codes, sigp_order_codes); +define_exit_reasons_table(sie_diagnose_codes, diagnose_codes); +define_exit_reasons_table(sie_icpt_prog_codes, icpt_prog_codes); + +static void event_icpt_insn_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + unsigned long insn; + + insn = perf_evsel__intval(evsel, sample, "instruction"); + key->key = icpt_insn_decoder(insn); + key->exit_reasons = sie_icpt_insn_codes; +} + +static void event_sigp_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + key->key = perf_evsel__intval(evsel, sample, "order_code"); + key->exit_reasons = sie_sigp_order_codes; +} + +static void event_diag_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + key->key = perf_evsel__intval(evsel, sample, "code"); + key->exit_reasons = sie_diagnose_codes; +} + +static void event_icpt_prog_get_key(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + key->key = perf_evsel__intval(evsel, sample, "code"); + key->exit_reasons = sie_icpt_prog_codes; +} + +static struct child_event_ops child_events[] = { + { .name = "kvm:kvm_s390_intercept_instruction", + .get_key = event_icpt_insn_get_key }, + { .name = "kvm:kvm_s390_handle_sigp", + .get_key = event_sigp_get_key }, + { .name = "kvm:kvm_s390_handle_diag", + .get_key = event_diag_get_key }, + { .name = "kvm:kvm_s390_intercept_prog", + .get_key = event_icpt_prog_get_key }, + { NULL, NULL }, +}; + +static struct kvm_events_ops exit_events = { + .is_begin_event = exit_event_begin, + .is_end_event = exit_event_end, + .child_ops = child_events, + .decode_key = exit_event_decode_key, + .name = "VM-EXIT" +}; + +const char * const kvm_events_tp[] = { + "kvm:kvm_s390_sie_enter", + "kvm:kvm_s390_sie_exit", + "kvm:kvm_s390_intercept_instruction", + "kvm:kvm_s390_handle_sigp", + "kvm:kvm_s390_handle_diag", + "kvm:kvm_s390_intercept_prog", + NULL, +}; + +struct kvm_reg_events_ops kvm_reg_events_ops[] = { + { .name = "vmexit", .ops = &exit_events }, + { NULL, NULL }, +}; + +const char * const kvm_skip_events[] = { + "Wait state", + NULL, +}; + +int cpu_isa_init(struct perf_kvm_stat *kvm, const char *cpuid) +{ + if (strstr(cpuid, "IBM/S390")) { + kvm->exit_reasons = sie_exit_reasons; + kvm->exit_reasons_isa = "SIE"; + } else + return -ENOTSUP; + + return 0; +} diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index fc2d63d..43367eb 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -88,7 +88,7 @@ void exit_event_decode_key(struct perf_kvm_stat *kvm, struct event_key *key, char *decode) { - const char *exit_reason = get_exit_reason(kvm, kvm->exit_reasons, + const char *exit_reason = get_exit_reason(kvm, key->exit_reasons, key->key); scnprintf(decode, DECODE_STR_LEN, "%s", exit_reason); @@ -261,6 +261,43 @@ static bool update_kvm_event(struct kvm_event *event, int vcpu_id, return true; } +static bool is_child_event(struct perf_kvm_stat *kvm, + struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key) +{ + struct child_event_ops *child_ops; + + child_ops = kvm->events_ops->child_ops; + + if (!child_ops) + return false; + + for (; child_ops->name; child_ops++) { + if (!strcmp(evsel->name, child_ops->name)) { + child_ops->get_key(evsel, sample, key); + return true; + } + } + + return false; +} + +static bool handle_child_event(struct perf_kvm_stat *kvm, + struct vcpu_event_record *vcpu_record, + struct event_key *key, + struct perf_sample *sample __maybe_unused) +{ + struct kvm_event *event = NULL; + + if (key->key != INVALID_KEY) + event = find_create_kvm_event(kvm, key); + + vcpu_record->last_event = event; + + return true; +} + static bool skip_event(const char *event) { const char * const *skip_events; @@ -361,7 +398,8 @@ static bool handle_kvm_event(struct perf_kvm_stat *kvm, struct perf_sample *sample) { struct vcpu_event_record *vcpu_record; - struct event_key key = {.key = INVALID_KEY}; + struct event_key key = { .key = INVALID_KEY, + .exit_reasons = kvm->exit_reasons }; vcpu_record = per_vcpu_record(thread, evsel, sample); if (!vcpu_record) @@ -375,6 +413,9 @@ static bool handle_kvm_event(struct perf_kvm_stat *kvm, if (kvm->events_ops->is_begin_event(evsel, sample, &key)) return handle_begin_event(kvm, vcpu_record, &key, sample->time); + if (is_child_event(kvm, evsel, sample, &key)) + return handle_child_event(kvm, vcpu_record, &key, sample); + if (kvm->events_ops->is_end_event(evsel, sample, &key)) return handle_end_event(kvm, vcpu_record, &key, sample); @@ -1143,7 +1184,8 @@ kvm_events_report(struct perf_kvm_stat *kvm, int argc, const char **argv) { const struct option kvm_events_report_options[] = { OPT_STRING(0, "event", &kvm->report_event, "report event", - "event for reporting: vmexit, mmio, ioport"), + "event for reporting: vmexit, " + "mmio (x86 only), ioport (x86 only)"), OPT_INTEGER(0, "vcpu", &kvm->trace_vcpu, "vcpu id to report"), OPT_STRING('k', "key", &kvm->sort_key, "sort-key", @@ -1249,7 +1291,9 @@ static int kvm_events_live(struct perf_kvm_stat *kvm, "key for sorting: sample(sort by samples number)" " time (sort by avg time)"), OPT_U64(0, "duration", &kvm->duration, - "show events other than HALT that take longer than duration usecs"), + "show events other than" + " HLT (x86 only) or Wait state (s390 only)" + " that take longer than duration usecs"), OPT_END() }; const char * const live_usage[] = { diff --git a/tools/perf/util/kvm-stat.h b/tools/perf/util/kvm-stat.h index ba937ca..0b5a8cd 100644 --- a/tools/perf/util/kvm-stat.h +++ b/tools/perf/util/kvm-stat.h @@ -12,6 +12,7 @@ struct event_key { #define INVALID_KEY (~0ULL) u64 key; int info; + struct exit_reasons_table *exit_reasons; }; struct kvm_event_stats { @@ -41,12 +42,20 @@ struct kvm_event_key { struct perf_kvm_stat; +struct child_event_ops { + void (*get_key)(struct perf_evsel *evsel, + struct perf_sample *sample, + struct event_key *key); + const char *name; +}; + struct kvm_events_ops { bool (*is_begin_event)(struct perf_evsel *evsel, struct perf_sample *sample, struct event_key *key); bool (*is_end_event)(struct perf_evsel *evsel, struct perf_sample *sample, struct event_key *key); + struct child_event_ops *child_ops; void (*decode_key)(struct perf_kvm_stat *kvm, struct event_key *key, char *decode); const char *name; -- cgit v1.1 From 05f832e3a267d6e45d092595bdf9339d127ea137 Mon Sep 17 00:00:00 2001 From: Joseph Schuchart Date: Wed, 9 Jul 2014 16:16:31 +0200 Subject: perf script: Add missing calls to Py_DECREF for return values Signed-off-by: Joseph Schuchart Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Ilsche Link: http://lkml.kernel.org/r/53BD4EBF.5050407@tu-dresden.de Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/scripting-engines/trace-event-python.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index e55b65a..b6c1a69 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -50,10 +50,14 @@ static int zero_flag_atom; static PyObject *main_module, *main_dict; +static void handler_call_die(const char *handler_name) NORETURN; static void handler_call_die(const char *handler_name) { PyErr_Print(); Py_FatalError("problem in Python trace event handler"); + // Py_FatalError does not return + // but we have to make the compiler happy + abort(); } /* @@ -97,6 +101,7 @@ static void define_value(enum print_arg_type field_type, retval = PyObject_CallObject(handler, t); if (retval == NULL) handler_call_die(handler_name); + Py_DECREF(retval); } Py_DECREF(t); @@ -143,6 +148,7 @@ static void define_field(enum print_arg_type field_type, retval = PyObject_CallObject(handler, t); if (retval == NULL) handler_call_die(handler_name); + Py_DECREF(retval); } Py_DECREF(t); @@ -361,6 +367,7 @@ static void python_process_tracepoint(struct perf_sample *sample, retval = PyObject_CallObject(handler, t); if (retval == NULL) handler_call_die(handler_name); + Py_DECREF(retval); } else { handler = PyDict_GetItemString(main_dict, "trace_unhandled"); if (handler && PyCallable_Check(handler)) { @@ -368,6 +375,7 @@ static void python_process_tracepoint(struct perf_sample *sample, retval = PyObject_CallObject(handler, t); if (retval == NULL) handler_call_die("trace_unhandled"); + Py_DECREF(retval); } Py_DECREF(dict); } @@ -427,6 +435,7 @@ static void python_process_general_event(struct perf_sample *sample, retval = PyObject_CallObject(handler, t); if (retval == NULL) handler_call_die(handler_name); + Py_DECREF(retval); exit: Py_DECREF(dict); Py_DECREF(t); @@ -548,8 +557,7 @@ static int python_stop_script(void) retval = PyObject_CallObject(handler, NULL); if (retval == NULL) handler_call_die("trace_end"); - else - Py_DECREF(retval); + Py_DECREF(retval); out: Py_XDECREF(main_dict); Py_XDECREF(main_module); -- cgit v1.1 From 0f5f5bcd112292f14b75750dde7461463bb1c7bb Mon Sep 17 00:00:00 2001 From: Joseph Schuchart Date: Thu, 10 Jul 2014 13:50:51 +0200 Subject: perf script: Add callchain to generic and tracepoint events This provides valuable information for tracing performance problems. Since this change alters the interface for the python scripts, also adjust the script generation and the provided scripts. Signed-off-by: Joseph Schuchart Acked-by: Thomas Ilsche Cc: Ingo Molnar Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Ilsche Link: http://lkml.kernel.org/r/53BE7E1B.10503@tu-dresden.de Signed-off-by: Arnaldo Carvalho de Melo --- .../python/Perf-Trace-Util/lib/Perf/Trace/Core.py | 3 +- tools/perf/scripts/python/check-perf-trace.py | 4 +- .../perf/scripts/python/failed-syscalls-by-pid.py | 2 +- tools/perf/scripts/python/futex-contention.py | 4 +- tools/perf/scripts/python/net_dropmonitor.py | 2 +- tools/perf/scripts/python/netdev-times.py | 26 ++--- tools/perf/scripts/python/sched-migration.py | 41 ++++---- tools/perf/scripts/python/sctop.py | 2 +- tools/perf/scripts/python/syscall-counts-by-pid.py | 2 +- tools/perf/scripts/python/syscall-counts.py | 2 +- .../util/scripting-engines/trace-event-python.c | 106 ++++++++++++++++++++- 11 files changed, 146 insertions(+), 48 deletions(-) (limited to 'tools') diff --git a/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Core.py b/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Core.py index de7211e..38dfb72 100644 --- a/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Core.py +++ b/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Core.py @@ -107,12 +107,13 @@ def taskState(state): class EventHeaders: def __init__(self, common_cpu, common_secs, common_nsecs, - common_pid, common_comm): + common_pid, common_comm, common_callchain): self.cpu = common_cpu self.secs = common_secs self.nsecs = common_nsecs self.pid = common_pid self.comm = common_comm + self.callchain = common_callchain def ts(self): return (self.secs * (10 ** 9)) + self.nsecs diff --git a/tools/perf/scripts/python/check-perf-trace.py b/tools/perf/scripts/python/check-perf-trace.py index 4647a76..334599c 100644 --- a/tools/perf/scripts/python/check-perf-trace.py +++ b/tools/perf/scripts/python/check-perf-trace.py @@ -27,7 +27,7 @@ def trace_end(): def irq__softirq_entry(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - vec): + common_callchain, vec): print_header(event_name, common_cpu, common_secs, common_nsecs, common_pid, common_comm) @@ -38,7 +38,7 @@ def irq__softirq_entry(event_name, context, common_cpu, def kmem__kmalloc(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - call_site, ptr, bytes_req, bytes_alloc, + common_callchain, call_site, ptr, bytes_req, bytes_alloc, gfp_flags): print_header(event_name, common_cpu, common_secs, common_nsecs, common_pid, common_comm) diff --git a/tools/perf/scripts/python/failed-syscalls-by-pid.py b/tools/perf/scripts/python/failed-syscalls-by-pid.py index 266a836..cafeff3 100644 --- a/tools/perf/scripts/python/failed-syscalls-by-pid.py +++ b/tools/perf/scripts/python/failed-syscalls-by-pid.py @@ -39,7 +39,7 @@ def trace_end(): def raw_syscalls__sys_exit(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - id, ret): + common_callchain, id, ret): if (for_comm and common_comm != for_comm) or \ (for_pid and common_pid != for_pid ): return diff --git a/tools/perf/scripts/python/futex-contention.py b/tools/perf/scripts/python/futex-contention.py index 11e70a3..0f5cf43 100644 --- a/tools/perf/scripts/python/futex-contention.py +++ b/tools/perf/scripts/python/futex-contention.py @@ -21,7 +21,7 @@ thread_blocktime = {} lock_waits = {} # long-lived stats on (tid,lock) blockage elapsed time process_names = {} # long-lived pid-to-execname mapping -def syscalls__sys_enter_futex(event, ctxt, cpu, s, ns, tid, comm, +def syscalls__sys_enter_futex(event, ctxt, cpu, s, ns, tid, comm, callchain, nr, uaddr, op, val, utime, uaddr2, val3): cmd = op & FUTEX_CMD_MASK if cmd != FUTEX_WAIT: @@ -31,7 +31,7 @@ def syscalls__sys_enter_futex(event, ctxt, cpu, s, ns, tid, comm, thread_thislock[tid] = uaddr thread_blocktime[tid] = nsecs(s, ns) -def syscalls__sys_exit_futex(event, ctxt, cpu, s, ns, tid, comm, +def syscalls__sys_exit_futex(event, ctxt, cpu, s, ns, tid, comm, callchain, nr, ret): if thread_blocktime.has_key(tid): elapsed = nsecs(s, ns) - thread_blocktime[tid] diff --git a/tools/perf/scripts/python/net_dropmonitor.py b/tools/perf/scripts/python/net_dropmonitor.py index b574059..0b6ce8c 100755 --- a/tools/perf/scripts/python/net_dropmonitor.py +++ b/tools/perf/scripts/python/net_dropmonitor.py @@ -66,7 +66,7 @@ def trace_end(): print_drop_table() # called from perf, when it finds a correspoinding event -def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm, +def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, location, protocol): slocation = str(location) try: diff --git a/tools/perf/scripts/python/netdev-times.py b/tools/perf/scripts/python/netdev-times.py index 9aa0a32..4d21ef2 100644 --- a/tools/perf/scripts/python/netdev-times.py +++ b/tools/perf/scripts/python/netdev-times.py @@ -224,75 +224,75 @@ def trace_end(): (len(rx_skb_list), of_count_rx_skb_list) # called from perf, when it finds a correspoinding event -def irq__softirq_entry(name, context, cpu, sec, nsec, pid, comm, vec): +def irq__softirq_entry(name, context, cpu, sec, nsec, pid, comm, callchain, vec): if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": return event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) all_event_list.append(event_info) -def irq__softirq_exit(name, context, cpu, sec, nsec, pid, comm, vec): +def irq__softirq_exit(name, context, cpu, sec, nsec, pid, comm, callchain, vec): if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": return event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) all_event_list.append(event_info) -def irq__softirq_raise(name, context, cpu, sec, nsec, pid, comm, vec): +def irq__softirq_raise(name, context, cpu, sec, nsec, pid, comm, callchain, vec): if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": return event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) all_event_list.append(event_info) def irq__irq_handler_entry(name, context, cpu, sec, nsec, pid, comm, - irq, irq_name): + callchain, irq, irq_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, irq, irq_name) all_event_list.append(event_info) -def irq__irq_handler_exit(name, context, cpu, sec, nsec, pid, comm, irq, ret): +def irq__irq_handler_exit(name, context, cpu, sec, nsec, pid, comm, callchain, irq, ret): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, irq, ret) all_event_list.append(event_info) -def napi__napi_poll(name, context, cpu, sec, nsec, pid, comm, napi, dev_name): +def napi__napi_poll(name, context, cpu, sec, nsec, pid, comm, callchain, napi, dev_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, napi, dev_name) all_event_list.append(event_info) -def net__netif_receive_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr, +def net__netif_receive_skb(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, skblen, dev_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, skblen, dev_name) all_event_list.append(event_info) -def net__netif_rx(name, context, cpu, sec, nsec, pid, comm, skbaddr, +def net__netif_rx(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, skblen, dev_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, skblen, dev_name) all_event_list.append(event_info) -def net__net_dev_queue(name, context, cpu, sec, nsec, pid, comm, +def net__net_dev_queue(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, skblen, dev_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, skblen, dev_name) all_event_list.append(event_info) -def net__net_dev_xmit(name, context, cpu, sec, nsec, pid, comm, +def net__net_dev_xmit(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, skblen, rc, dev_name): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, skblen, rc ,dev_name) all_event_list.append(event_info) -def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm, +def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, protocol, location): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, protocol, location) all_event_list.append(event_info) -def skb__consume_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr): +def skb__consume_skb(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr) all_event_list.append(event_info) -def skb__skb_copy_datagram_iovec(name, context, cpu, sec, nsec, pid, comm, +def skb__skb_copy_datagram_iovec(name, context, cpu, sec, nsec, pid, comm, callchain, skbaddr, skblen): event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, skbaddr, skblen) diff --git a/tools/perf/scripts/python/sched-migration.py b/tools/perf/scripts/python/sched-migration.py index 74d55ec..de66cb3 100644 --- a/tools/perf/scripts/python/sched-migration.py +++ b/tools/perf/scripts/python/sched-migration.py @@ -369,93 +369,92 @@ def trace_end(): def sched__sched_stat_runtime(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, runtime, vruntime): + common_callchain, comm, pid, runtime, vruntime): pass def sched__sched_stat_iowait(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, delay): + common_callchain, comm, pid, delay): pass def sched__sched_stat_sleep(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, delay): + common_callchain, comm, pid, delay): pass def sched__sched_stat_wait(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, delay): + common_callchain, comm, pid, delay): pass def sched__sched_process_fork(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - parent_comm, parent_pid, child_comm, child_pid): + common_callchain, parent_comm, parent_pid, child_comm, child_pid): pass def sched__sched_process_wait(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio): + common_callchain, comm, pid, prio): pass def sched__sched_process_exit(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio): + common_callchain, comm, pid, prio): pass def sched__sched_process_free(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio): + common_callchain, comm, pid, prio): pass def sched__sched_migrate_task(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio, orig_cpu, + common_callchain, comm, pid, prio, orig_cpu, dest_cpu): headers = EventHeaders(common_cpu, common_secs, common_nsecs, - common_pid, common_comm) + common_pid, common_comm, common_callchain) parser.migrate(headers, pid, prio, orig_cpu, dest_cpu) def sched__sched_switch(event_name, context, common_cpu, - common_secs, common_nsecs, common_pid, common_comm, + common_secs, common_nsecs, common_pid, common_comm, common_callchain, prev_comm, prev_pid, prev_prio, prev_state, next_comm, next_pid, next_prio): headers = EventHeaders(common_cpu, common_secs, common_nsecs, - common_pid, common_comm) + common_pid, common_comm, common_callchain) parser.sched_switch(headers, prev_comm, prev_pid, prev_prio, prev_state, next_comm, next_pid, next_prio) def sched__sched_wakeup_new(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio, success, + common_callchain, comm, pid, prio, success, target_cpu): headers = EventHeaders(common_cpu, common_secs, common_nsecs, - common_pid, common_comm) + common_pid, common_comm, common_callchain) parser.wake_up(headers, comm, pid, success, target_cpu, 1) def sched__sched_wakeup(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio, success, + common_callchain, comm, pid, prio, success, target_cpu): headers = EventHeaders(common_cpu, common_secs, common_nsecs, - common_pid, common_comm) + common_pid, common_comm, common_callchain) parser.wake_up(headers, comm, pid, success, target_cpu, 0) def sched__sched_wait_task(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid, prio): + common_callchain, comm, pid, prio): pass def sched__sched_kthread_stop_ret(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - ret): + common_callchain, ret): pass def sched__sched_kthread_stop(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - comm, pid): + common_callchain, comm, pid): pass -def trace_unhandled(event_name, context, common_cpu, common_secs, common_nsecs, - common_pid, common_comm): +def trace_unhandled(event_name, context, event_fields_dict): pass diff --git a/tools/perf/scripts/python/sctop.py b/tools/perf/scripts/python/sctop.py index c9f3058..61621b9 100644 --- a/tools/perf/scripts/python/sctop.py +++ b/tools/perf/scripts/python/sctop.py @@ -44,7 +44,7 @@ def trace_begin(): def raw_syscalls__sys_enter(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - id, args): + common_callchain, id, args): if for_comm is not None: if common_comm != for_comm: return diff --git a/tools/perf/scripts/python/syscall-counts-by-pid.py b/tools/perf/scripts/python/syscall-counts-by-pid.py index cf2054c..daf314c 100644 --- a/tools/perf/scripts/python/syscall-counts-by-pid.py +++ b/tools/perf/scripts/python/syscall-counts-by-pid.py @@ -38,7 +38,7 @@ def trace_end(): def raw_syscalls__sys_enter(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - id, args): + common_callchain, id, args): if (for_comm and common_comm != for_comm) or \ (for_pid and common_pid != for_pid ): diff --git a/tools/perf/scripts/python/syscall-counts.py b/tools/perf/scripts/python/syscall-counts.py index 92b2938..e66a773 100644 --- a/tools/perf/scripts/python/syscall-counts.py +++ b/tools/perf/scripts/python/syscall-counts.py @@ -35,7 +35,7 @@ def trace_end(): def raw_syscalls__sys_enter(event_name, context, common_cpu, common_secs, common_nsecs, common_pid, common_comm, - id, args): + common_callchain, id, args): if for_comm is not None: if common_comm != for_comm: return diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index b6c1a69..cf65404 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -32,6 +32,7 @@ #include "../event.h" #include "../thread.h" #include "../trace-event.h" +#include "../machine.h" PyMODINIT_FUNC initperf_trace_context(void); @@ -278,12 +279,90 @@ static PyObject *get_field_numeric_entry(struct event_format *event, return obj; } + +static PyObject *python_process_callchain(struct perf_sample *sample, + struct perf_evsel *evsel, + struct addr_location *al) +{ + PyObject *pylist; + + pylist = PyList_New(0); + if (!pylist) + Py_FatalError("couldn't create Python list"); + + if (!symbol_conf.use_callchain || !sample->callchain) + goto exit; + + if (machine__resolve_callchain(al->machine, evsel, al->thread, + sample, NULL, NULL, + PERF_MAX_STACK_DEPTH) != 0) { + pr_err("Failed to resolve callchain. Skipping\n"); + goto exit; + } + callchain_cursor_commit(&callchain_cursor); + + + while (1) { + PyObject *pyelem; + struct callchain_cursor_node *node; + node = callchain_cursor_current(&callchain_cursor); + if (!node) + break; + + pyelem = PyDict_New(); + if (!pyelem) + Py_FatalError("couldn't create Python dictionary"); + + + pydict_set_item_string_decref(pyelem, "ip", + PyLong_FromUnsignedLongLong(node->ip)); + + if (node->sym) { + PyObject *pysym = PyDict_New(); + if (!pysym) + Py_FatalError("couldn't create Python dictionary"); + pydict_set_item_string_decref(pysym, "start", + PyLong_FromUnsignedLongLong(node->sym->start)); + pydict_set_item_string_decref(pysym, "end", + PyLong_FromUnsignedLongLong(node->sym->end)); + pydict_set_item_string_decref(pysym, "binding", + PyInt_FromLong(node->sym->binding)); + pydict_set_item_string_decref(pysym, "name", + PyString_FromStringAndSize(node->sym->name, + node->sym->namelen)); + pydict_set_item_string_decref(pyelem, "sym", pysym); + } + + if (node->map) { + struct map *map = node->map; + const char *dsoname = "[unknown]"; + if (map && map->dso && (map->dso->name || map->dso->long_name)) { + if (symbol_conf.show_kernel_path && map->dso->long_name) + dsoname = map->dso->long_name; + else if (map->dso->name) + dsoname = map->dso->name; + } + pydict_set_item_string_decref(pyelem, "dso", + PyString_FromString(dsoname)); + } + + callchain_cursor_advance(&callchain_cursor); + PyList_Append(pylist, pyelem); + Py_DECREF(pyelem); + } + +exit: + return pylist; +} + + static void python_process_tracepoint(struct perf_sample *sample, struct perf_evsel *evsel, struct thread *thread, struct addr_location *al) { - PyObject *handler, *retval, *context, *t, *obj, *dict = NULL; + PyObject *handler, *retval, *context, *t, *obj, *callchain; + PyObject *dict = NULL; static char handler_name[256]; struct format_field *field; unsigned long s, ns; @@ -326,18 +405,23 @@ static void python_process_tracepoint(struct perf_sample *sample, PyTuple_SetItem(t, n++, PyString_FromString(handler_name)); PyTuple_SetItem(t, n++, context); + /* ip unwinding */ + callchain = python_process_callchain(sample, evsel, al); + if (handler) { PyTuple_SetItem(t, n++, PyInt_FromLong(cpu)); PyTuple_SetItem(t, n++, PyInt_FromLong(s)); PyTuple_SetItem(t, n++, PyInt_FromLong(ns)); PyTuple_SetItem(t, n++, PyInt_FromLong(pid)); PyTuple_SetItem(t, n++, PyString_FromString(comm)); + PyTuple_SetItem(t, n++, callchain); } else { pydict_set_item_string_decref(dict, "common_cpu", PyInt_FromLong(cpu)); pydict_set_item_string_decref(dict, "common_s", PyInt_FromLong(s)); pydict_set_item_string_decref(dict, "common_ns", PyInt_FromLong(ns)); pydict_set_item_string_decref(dict, "common_pid", PyInt_FromLong(pid)); pydict_set_item_string_decref(dict, "common_comm", PyString_FromString(comm)); + pydict_set_item_string_decref(dict, "common_callchain", callchain); } for (field = event->format.fields; field; field = field->next) { if (field->flags & FIELD_IS_STRING) { @@ -357,6 +441,7 @@ static void python_process_tracepoint(struct perf_sample *sample, pydict_set_item_string_decref(dict, field->name, obj); } + if (!handler) PyTuple_SetItem(t, n++, dict); @@ -388,7 +473,7 @@ static void python_process_general_event(struct perf_sample *sample, struct thread *thread, struct addr_location *al) { - PyObject *handler, *retval, *t, *dict; + PyObject *handler, *retval, *t, *dict, *callchain; static char handler_name[64]; unsigned n = 0; @@ -428,6 +513,10 @@ static void python_process_general_event(struct perf_sample *sample, PyString_FromString(al->sym->name)); } + /* ip unwinding */ + callchain = python_process_callchain(sample, evsel, al); + pydict_set_item_string_decref(dict, "callchain", callchain); + PyTuple_SetItem(t, n++, dict); if (_PyTuple_Resize(&t, n) == -1) Py_FatalError("error resizing Python tuple"); @@ -624,6 +713,7 @@ static int python_generate_script(struct pevent *pevent, const char *outfile) fprintf(ofp, "common_nsecs, "); fprintf(ofp, "common_pid, "); fprintf(ofp, "common_comm,\n\t"); + fprintf(ofp, "common_callchain, "); not_first = 0; count = 0; @@ -667,7 +757,7 @@ static int python_generate_script(struct pevent *pevent, const char *outfile) fprintf(ofp, "%%u"); } - fprintf(ofp, "\\n\" %% \\\n\t\t("); + fprintf(ofp, "\" %% \\\n\t\t("); not_first = 0; count = 0; @@ -703,7 +793,15 @@ static int python_generate_script(struct pevent *pevent, const char *outfile) fprintf(ofp, "%s", f->name); } - fprintf(ofp, "),\n\n"); + fprintf(ofp, ")\n\n"); + + fprintf(ofp, "\t\tfor node in common_callchain:"); + fprintf(ofp, "\n\t\t\tif 'sym' in node:"); + fprintf(ofp, "\n\t\t\t\tprint \"\\t[%%x] %%s\" %% (node['ip'], node['sym']['name'])"); + fprintf(ofp, "\n\t\t\telse:"); + fprintf(ofp, "\n\t\t\t\tprint \"\t[%%x]\" %% (node['ip'])\n\n"); + fprintf(ofp, "\t\tprint \"\\n\"\n\n"); + } fprintf(ofp, "def trace_unhandled(event_name, context, " -- cgit v1.1 From 57608cfd8827a74237d264a197722e2c99f72da4 Mon Sep 17 00:00:00 2001 From: Joseph Schuchart Date: Thu, 10 Jul 2014 13:50:56 +0200 Subject: perf script: Provide additional sample information on generic events To python scripts, including pid, tid, and cpu for which the event was recorded. At the moment, the pointer to the sample struct is passed to scripts, which seems to be of little use. The patch puts this information in dictionaries for easy access by Python scripts. Signed-off-by: Joseph Schuchart Acked-by: Thomas Ilsche Cc: Ingo Molnar Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Ilsche Link: http://lkml.kernel.org/r/53BE7E20.8080500@tu-dresden.de Signed-off-by: Arnaldo Carvalho de Melo --- .../util/scripting-engines/trace-event-python.c | 23 +++++++++++++++++++--- 1 file changed, 20 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index cf65404..b366b48 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -473,7 +473,7 @@ static void python_process_general_event(struct perf_sample *sample, struct thread *thread, struct addr_location *al) { - PyObject *handler, *retval, *t, *dict, *callchain; + PyObject *handler, *retval, *t, *dict, *callchain, *dict_sample; static char handler_name[64]; unsigned n = 0; @@ -489,6 +489,10 @@ static void python_process_general_event(struct perf_sample *sample, if (!dict) Py_FatalError("couldn't create Python dictionary"); + dict_sample = PyDict_New(); + if (!dict_sample) + Py_FatalError("couldn't create Python dictionary"); + snprintf(handler_name, sizeof(handler_name), "%s", "process_event"); handler = PyDict_GetItemString(main_dict, handler_name); @@ -498,8 +502,21 @@ static void python_process_general_event(struct perf_sample *sample, pydict_set_item_string_decref(dict, "ev_name", PyString_FromString(perf_evsel__name(evsel))); pydict_set_item_string_decref(dict, "attr", PyString_FromStringAndSize( (const char *)&evsel->attr, sizeof(evsel->attr))); - pydict_set_item_string_decref(dict, "sample", PyString_FromStringAndSize( - (const char *)sample, sizeof(*sample))); + + pydict_set_item_string_decref(dict_sample, "pid", + PyInt_FromLong(sample->pid)); + pydict_set_item_string_decref(dict_sample, "tid", + PyInt_FromLong(sample->tid)); + pydict_set_item_string_decref(dict_sample, "cpu", + PyInt_FromLong(sample->cpu)); + pydict_set_item_string_decref(dict_sample, "ip", + PyLong_FromUnsignedLongLong(sample->ip)); + pydict_set_item_string_decref(dict_sample, "time", + PyLong_FromUnsignedLongLong(sample->time)); + pydict_set_item_string_decref(dict_sample, "period", + PyLong_FromUnsignedLongLong(sample->period)); + pydict_set_item_string_decref(dict, "sample", dict_sample); + pydict_set_item_string_decref(dict, "raw_buf", PyString_FromStringAndSize( (const char *)sample->raw_data, sample->raw_size)); pydict_set_item_string_decref(dict, "comm", -- cgit v1.1 From 1fcb8768636d38cb6fdfeef83a5ee596c4bd9c56 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:25 +0300 Subject: perf machine: Fix the value used for unknown pids The value used for unknown pids cannot be zero because that is used by the "idle" task. Use -1 instead. Also handle the unknown pid case when creating map groups. Note that, threads with an unknown pid should not occur because fork (or synthesized) events precede the thread's existence. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-sched.c | 12 ++++++------ tools/perf/util/machine.c | 6 +++--- tools/perf/util/session.c | 5 +++-- tools/perf/util/thread.c | 2 +- 4 files changed, 13 insertions(+), 12 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index c38d06c..b7f555a 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -935,8 +935,8 @@ static int latency_switch_event(struct perf_sched *sched, return -1; } - sched_out = machine__findnew_thread(machine, 0, prev_pid); - sched_in = machine__findnew_thread(machine, 0, next_pid); + sched_out = machine__findnew_thread(machine, -1, prev_pid); + sched_in = machine__findnew_thread(machine, -1, next_pid); out_events = thread_atoms_search(&sched->atom_root, sched_out, &sched->cmp_pid); if (!out_events) { @@ -979,7 +979,7 @@ static int latency_runtime_event(struct perf_sched *sched, { const u32 pid = perf_evsel__intval(evsel, sample, "pid"); const u64 runtime = perf_evsel__intval(evsel, sample, "runtime"); - struct thread *thread = machine__findnew_thread(machine, 0, pid); + struct thread *thread = machine__findnew_thread(machine, -1, pid); struct work_atoms *atoms = thread_atoms_search(&sched->atom_root, thread, &sched->cmp_pid); u64 timestamp = sample->time; int cpu = sample->cpu; @@ -1012,7 +1012,7 @@ static int latency_wakeup_event(struct perf_sched *sched, struct thread *wakee; u64 timestamp = sample->time; - wakee = machine__findnew_thread(machine, 0, pid); + wakee = machine__findnew_thread(machine, -1, pid); atoms = thread_atoms_search(&sched->atom_root, wakee, &sched->cmp_pid); if (!atoms) { if (thread_atoms_insert(sched, wakee)) @@ -1072,7 +1072,7 @@ static int latency_migrate_task_event(struct perf_sched *sched, if (sched->profile_cpu == -1) return 0; - migrant = machine__findnew_thread(machine, 0, pid); + migrant = machine__findnew_thread(machine, -1, pid); atoms = thread_atoms_search(&sched->atom_root, migrant, &sched->cmp_pid); if (!atoms) { if (thread_atoms_insert(sched, migrant)) @@ -1290,7 +1290,7 @@ static int map_switch_event(struct perf_sched *sched, struct perf_evsel *evsel, return -1; } - sched_in = machine__findnew_thread(machine, 0, next_pid); + sched_in = machine__findnew_thread(machine, -1, next_pid); sched->curr_thread[this_cpu] = sched_in; diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index e9b943a..5b80877 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -34,7 +34,7 @@ int machine__init(struct machine *machine, const char *root_dir, pid_t pid) return -ENOMEM; if (pid != HOST_KERNEL_ID) { - struct thread *thread = machine__findnew_thread(machine, 0, + struct thread *thread = machine__findnew_thread(machine, -1, pid); char comm[64]; @@ -286,7 +286,7 @@ static struct thread *__machine__findnew_thread(struct machine *machine, * the full rbtree: */ if (machine->last_match && machine->last_match->tid == tid) { - if (pid && pid != machine->last_match->pid_) + if (pid != -1 && pid != machine->last_match->pid_) machine->last_match->pid_ = pid; return machine->last_match; } @@ -297,7 +297,7 @@ static struct thread *__machine__findnew_thread(struct machine *machine, if (th->tid == tid) { machine->last_match = th; - if (pid && pid != th->pid_) + if (pid != -1 && pid != th->pid_) th->pid_ = pid; return th; } diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 64a186e..c2f4ca9 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1083,13 +1083,14 @@ void perf_event_header__bswap(struct perf_event_header *hdr) struct thread *perf_session__findnew(struct perf_session *session, pid_t pid) { - return machine__findnew_thread(&session->machines.host, 0, pid); + return machine__findnew_thread(&session->machines.host, -1, pid); } static struct thread *perf_session__register_idle_thread(struct perf_session *session) { - struct thread *thread = perf_session__findnew(session, 0); + struct thread *thread; + thread = machine__findnew_thread(&session->machines.host, 0, 0); if (thread == NULL || thread__set_comm(thread, "swapper", 0)) { pr_err("problem inserting idle task.\n"); thread = NULL; diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 2fde0d5..7a32f44 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -13,7 +13,7 @@ int thread__init_map_groups(struct thread *thread, struct machine *machine) struct thread *leader; pid_t pid = thread->pid_; - if (pid == thread->tid) { + if (pid == thread->tid || pid == -1) { thread->mg = map_groups__new(); } else { leader = machine__findnew_thread(machine, pid, pid); -- cgit v1.1 From 022c50d09c2c2bc31506ad16c4bcba7fb418ce34 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:27 +0300 Subject: perf script: Display PERF_RECORD_MISC_COMM_EXEC flag Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-4-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/event.c | 9 ++++++++- tools/perf/util/evsel.c | 3 ++- 2 files changed, 10 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index d0281bd..198c4cc 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -603,7 +603,14 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, size_t perf_event__fprintf_comm(union perf_event *event, FILE *fp) { - return fprintf(fp, ": %s:%d\n", event->comm.comm, event->comm.tid); + const char *s; + + if (event->header.misc & PERF_RECORD_MISC_COMM_EXEC) + s = " exec"; + else + s = ""; + + return fprintf(fp, "%s: %s:%d\n", s, event->comm.comm, event->comm.tid); } int perf_event__process_comm(struct perf_tool *tool __maybe_unused, diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 8606175..b760d32 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -960,6 +960,7 @@ static size_t perf_event_attr__fprintf(struct perf_event_attr *attr, FILE *fp) ret += PRINT_ATTR2(exclude_user, exclude_kernel); ret += PRINT_ATTR2(exclude_hv, exclude_idle); ret += PRINT_ATTR2(mmap, comm); + ret += PRINT_ATTR2(mmap2, comm_exec); ret += PRINT_ATTR2(freq, inherit_stat); ret += PRINT_ATTR2(enable_on_exec, task); ret += PRINT_ATTR2(watermark, precise_ip); @@ -967,7 +968,6 @@ static size_t perf_event_attr__fprintf(struct perf_event_attr *attr, FILE *fp) ret += PRINT_ATTR2(exclude_host, exclude_guest); ret += PRINT_ATTR2N("excl.callchain_kern", exclude_callchain_kernel, "excl.callchain_user", exclude_callchain_user); - ret += PRINT_ATTR_U32(mmap2); ret += PRINT_ATTR_U32(wakeup_events); ret += PRINT_ATTR_U32(wakeup_watermark); @@ -1940,6 +1940,7 @@ int perf_evsel__fprintf(struct perf_evsel *evsel, if_print(mmap); if_print(mmap2); if_print(comm); + if_print(comm_exec); if_print(freq); if_print(inherit_stat); if_print(enable_on_exec); -- cgit v1.1 From 39e09d40bea440d9cfe645b55aff251294318669 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:28 +0300 Subject: perf record: Select comm_exec flag if supported The comm_exec flag on the attribute can later be found in the perf.data file allowing a tool to know in advance if the captured data has the flag. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-5-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/record.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/record.c b/tools/perf/util/record.c index 049e0a0..1657231 100644 --- a/tools/perf/util/record.c +++ b/tools/perf/util/record.c @@ -69,15 +69,26 @@ static void perf_probe_sample_identifier(struct perf_evsel *evsel) evsel->attr.sample_type |= PERF_SAMPLE_IDENTIFIER; } +static void perf_probe_comm_exec(struct perf_evsel *evsel) +{ + evsel->attr.comm_exec = 1; +} + bool perf_can_sample_identifier(void) { return perf_probe_api(perf_probe_sample_identifier); } +static bool perf_can_comm_exec(void) +{ + return perf_probe_api(perf_probe_comm_exec); +} + void perf_evlist__config(struct perf_evlist *evlist, struct record_opts *opts) { struct perf_evsel *evsel; bool use_sample_identifier = false; + bool use_comm_exec; /* * Set the evsel leader links before we configure attributes, @@ -89,8 +100,13 @@ void perf_evlist__config(struct perf_evlist *evlist, struct record_opts *opts) if (evlist->cpus->map[0] < 0) opts->no_inherit = true; - evlist__for_each(evlist, evsel) + use_comm_exec = perf_can_comm_exec(); + + evlist__for_each(evlist, evsel) { perf_evsel__config(evsel, opts); + if (!evsel->idx && use_comm_exec) + evsel->attr.comm_exec = 1; + } if (evlist->nr_entries > 1) { struct perf_evsel *first = perf_evlist__first(evlist); -- cgit v1.1 From 1f2a7069b6e895487b1d9229f5f62799cc4ea0aa Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:31 +0300 Subject: perf tools: Fix missing kernel map load thread__find_addr_map() falls back to trying the kernel maps if the address is negative and is not found in userspace maps. As commented in the code, the kernel maps must be "loaded" before use. This patch ensures that happens under the fallback condition also. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-8-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/event.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools') diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 198c4cc..7e0e8ae 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -788,6 +788,7 @@ try_again: cpumode == PERF_RECORD_MISC_USER && machine && mg != &machine->kmaps) { mg = &machine->kmaps; + load_map = true; goto try_again; } } else { -- cgit v1.1 From a2f3b6bf0adadcb5f9383c60aa1355e0f9cba3da Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:33 +0300 Subject: perf symbols: Fix missing GNU IFUNC symbols Symbols of type STT_GNU_IFUNC are functions so accept them as such. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-10-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 6864661..dce5ccf 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -49,7 +49,8 @@ static inline uint8_t elf_sym__type(const GElf_Sym *sym) static inline int elf_sym__is_function(const GElf_Sym *sym) { - return elf_sym__type(sym) == STT_FUNC && + return (elf_sym__type(sym) == STT_FUNC || + elf_sym__type(sym) == STT_GNU_IFUNC) && sym->st_name != 0 && sym->st_shndx != SHN_UNDEF; } -- cgit v1.1 From e38b43c3f3fd8ebe6f558400d1647a923bc19d44 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:34 +0300 Subject: perf inject: Fix build id injection Build Ids won't be injected unless the build id feature flag is set. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-11-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-inject.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools') diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index 16c7c11..cf6a605 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -389,6 +389,9 @@ static int __cmd_inject(struct perf_inject *inject) ret = perf_session__process_events(session, &inject->tool); if (!file_out->is_pipe) { + if (inject->build_ids) + perf_header__set_feat(&session->header, + HEADER_BUILD_ID); session->header.data_size = inject->bytes_written; perf_session__write_header(session, session->evlist, file_out->fd, true); } -- cgit v1.1 From 4d40b051b1ac41ecbc818deed27750b4c1697520 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:35 +0300 Subject: perf callchain: Fix appending a callchain from a previous sample hist_entry__append_callchain() must check if the sample has a callcahin or it will append the callchain from a previous sample. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-12-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/callchain.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index 48b6d3f..437ee09 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -626,7 +626,7 @@ int sample__resolve_callchain(struct perf_sample *sample, struct symbol **parent int hist_entry__append_callchain(struct hist_entry *he, struct perf_sample *sample) { - if (!symbol_conf.use_callchain) + if (!symbol_conf.use_callchain || sample->callchain == NULL) return 0; return callchain_append(he->callchain, &callchain_cursor, sample->period); } -- cgit v1.1 From 5173fbb8a11b2857aeec9e5f4e9568d4e1b84dbd Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:38 +0300 Subject: perf buildid-cache: Apply force option to copying kcore Currently a copy of kcore is not made if there is one already with the same modules at the same addresses. Change this to make a copy anyway if the force (-f) option is also used. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-15-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-buildid-cache.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-buildid-cache.c b/tools/perf/builtin-buildid-cache.c index b22dbb1..2a2c78f 100644 --- a/tools/perf/builtin-buildid-cache.c +++ b/tools/perf/builtin-buildid-cache.c @@ -125,7 +125,8 @@ static int build_id_cache__kcore_existing(const char *from_dir, char *to_dir, return ret; } -static int build_id_cache__add_kcore(const char *filename, const char *debugdir) +static int build_id_cache__add_kcore(const char *filename, const char *debugdir, + bool force) { char dir[32], sbuildid[BUILD_ID_SIZE * 2 + 1]; char from_dir[PATH_MAX], to_dir[PATH_MAX]; @@ -144,7 +145,8 @@ static int build_id_cache__add_kcore(const char *filename, const char *debugdir) scnprintf(to_dir, sizeof(to_dir), "%s/[kernel.kcore]/%s", debugdir, sbuildid); - if (!build_id_cache__kcore_existing(from_dir, to_dir, sizeof(to_dir))) { + if (!force && + !build_id_cache__kcore_existing(from_dir, to_dir, sizeof(to_dir))) { pr_debug("same kcore found in %s\n", to_dir); return 0; } @@ -389,7 +391,7 @@ int cmd_buildid_cache(int argc, const char **argv, } if (kcore_filename && - build_id_cache__add_kcore(kcore_filename, debugdir)) + build_id_cache__add_kcore(kcore_filename, debugdir, force)) pr_warning("Couldn't add %s\n", kcore_filename); return ret; -- cgit v1.1 From c6d8f2a4a0c5e366330a6a2a94c06b652f4ca554 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:41 +0300 Subject: perf symbols: Record whether a dso is 64-bit Add a flag to 'struct dso' to record if the dso is 64-bit or not. Update the flag when reading the ELF. This is needed for instruction decoding. For example, x86 instruction decoding depends on whether or not the 64-bit instruction set is used. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-18-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 1 + tools/perf/util/dso.h | 1 + tools/perf/util/symbol-elf.c | 3 +++ tools/perf/util/symbol-minimal.c | 22 ++++++++++++++++++++++ tools/perf/util/symbol.c | 1 + tools/perf/util/symbol.h | 1 + 6 files changed, 29 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 819f104..fc006fe 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -703,6 +703,7 @@ struct dso *dso__new(const char *name) dso->data.fd = -1; dso->symtab_type = DSO_BINARY_TYPE__NOT_FOUND; dso->binary_type = DSO_BINARY_TYPE__NOT_FOUND; + dso->is_64_bit = (sizeof(void *) == 8); dso->loaded = 0; dso->rel = 0; dso->sorted_by_name = 0; diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index ad553ba..c239e86 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -90,6 +90,7 @@ struct dso { u8 annotate_warned:1; u8 short_name_allocated:1; u8 long_name_allocated:1; + u8 is_64_bit:1; u8 sorted_by_name; u8 loaded; u8 rel; diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index dce5ccf..cef8f42 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -599,6 +599,8 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, goto out_elf_end; } + ss->is_64_bit = (gelf_getclass(elf) == ELFCLASS64); + ss->symtab = elf_section_by_name(elf, &ehdr, &ss->symshdr, ".symtab", NULL); if (ss->symshdr.sh_type != SHT_SYMTAB) @@ -699,6 +701,7 @@ int dso__load_sym(struct dso *dso, struct map *map, bool remap_kernel = false, adjust_kernel_syms = false; dso->symtab_type = syms_ss->type; + dso->is_64_bit = syms_ss->is_64_bit; dso->rel = syms_ss->ehdr.e_type == ET_REL; /* diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c index bd15f49..101f55d 100644 --- a/tools/perf/util/symbol-minimal.c +++ b/tools/perf/util/symbol-minimal.c @@ -288,6 +288,23 @@ int dso__synthesize_plt_symbols(struct dso *dso __maybe_unused, return 0; } +static int fd__is_64_bit(int fd) +{ + u8 e_ident[EI_NIDENT]; + + if (lseek(fd, 0, SEEK_SET)) + return -1; + + if (readn(fd, e_ident, sizeof(e_ident)) != sizeof(e_ident)) + return -1; + + if (memcmp(e_ident, ELFMAG, SELFMAG) || + e_ident[EI_VERSION] != EV_CURRENT) + return -1; + + return e_ident[EI_CLASS] == ELFCLASS64; +} + int dso__load_sym(struct dso *dso, struct map *map __maybe_unused, struct symsrc *ss, struct symsrc *runtime_ss __maybe_unused, @@ -295,6 +312,11 @@ int dso__load_sym(struct dso *dso, struct map *map __maybe_unused, int kmodule __maybe_unused) { unsigned char *build_id[BUILD_ID_SIZE]; + int ret; + + ret = fd__is_64_bit(ss->fd); + if (ret >= 0) + dso->is_64_bit = ret; if (filename__read_build_id(ss->name, build_id, BUILD_ID_SIZE) > 0) { dso__set_build_id(dso, build_id); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 2e6a2e2..ae2e446 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1065,6 +1065,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map, &is_64_bit); if (err) goto out_err; + dso->is_64_bit = is_64_bit; if (list_empty(&md.maps)) { err = -EINVAL; diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index a81877b..436169d 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -216,6 +216,7 @@ struct symsrc { GElf_Shdr dynshdr; bool adjust_symbols; + bool is_64_bit; #endif }; -- cgit v1.1 From bdac0bcf779250e89b96d4a3f381ebaf02c2f4a9 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:43 +0300 Subject: perf symbols: Do not attempt to read data from kallsyms Record kallsyms binary type so that tools will not attempt to read binary data from it. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-20-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index ae2e446..156ae36 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1664,6 +1664,7 @@ do_kallsyms: free(kallsyms_allocated_filename); if (err > 0 && !dso__is_kcore(dso)) { + dso->binary_type = DSO_BINARY_TYPE__KALLSYMS; dso__set_long_name(dso, "[kernel.kallsyms]", false); map__fixup_start(map); map__fixup_end(map); @@ -1711,6 +1712,7 @@ static int dso__load_guest_kernel_sym(struct dso *dso, struct map *map, if (err > 0) pr_debug("Using %s for symbols\n", kallsyms_filename); if (err > 0 && !dso__is_kcore(dso)) { + dso->binary_type = DSO_BINARY_TYPE__GUEST_KALLSYMS; machine__mmap_name(machine, path, sizeof(path)); dso__set_long_name(dso, strdup(path), true); map__fixup_start(map); -- cgit v1.1 From 9c00a81b6aafc4ed375a43e7a54e6cf2d720c7c6 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:50 +0300 Subject: perf symbols: Add ability to iterate over a dso's symbols Expose dso__first_symbol() and dso__next_symbol() to make it possible to iterate over a dso's symbols. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-27-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 17 ++++++++++++++++- tools/perf/util/symbol.h | 3 +++ 2 files changed, 19 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 156ae36..eb06746 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -342,6 +342,16 @@ static struct symbol *symbols__first(struct rb_root *symbols) return NULL; } +static struct symbol *symbols__next(struct symbol *sym) +{ + struct rb_node *n = rb_next(&sym->rb_node); + + if (n) + return rb_entry(n, struct symbol, rb_node); + + return NULL; +} + struct symbol_name_rb_node { struct rb_node rb_node; struct symbol sym; @@ -412,11 +422,16 @@ struct symbol *dso__find_symbol(struct dso *dso, return symbols__find(&dso->symbols[type], addr); } -static struct symbol *dso__first_symbol(struct dso *dso, enum map_type type) +struct symbol *dso__first_symbol(struct dso *dso, enum map_type type) { return symbols__first(&dso->symbols[type]); } +struct symbol *dso__next_symbol(struct symbol *sym) +{ + return symbols__next(sym); +} + struct symbol *dso__find_symbol_by_name(struct dso *dso, enum map_type type, const char *name) { diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 436169d..ee2d3cc 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -240,6 +240,9 @@ struct symbol *dso__find_symbol(struct dso *dso, enum map_type type, struct symbol *dso__find_symbol_by_name(struct dso *dso, enum map_type type, const char *name); +struct symbol *dso__first_symbol(struct dso *dso, enum map_type type); +struct symbol *dso__next_symbol(struct symbol *sym); + int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); int modules__parse(const char *filename, void *arg, -- cgit v1.1 From 919d86d3a3109d7d4f0d7347f34711ee2f8e6609 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:51 +0300 Subject: perf session: Flag if the event stream is entirely in memory Flag if the event stream is a file that has been mmapped in one go. This is useful, for example, if a tool needs to keep an event for later reference. If the new flag is set, a pointer to the event can be retained, otherwise the event must be copied. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-28-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 9 ++++++++- tools/perf/util/session.h | 3 +++ 2 files changed, 11 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index c2f4ca9..eac14ce 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1297,8 +1297,10 @@ int __perf_session__process_events(struct perf_session *session, ui_progress__init(&prog, file_size, "Processing events..."); mmap_size = MMAP_SIZE; - if (mmap_size > file_size) + if (mmap_size > file_size) { mmap_size = file_size; + session->one_mmap = true; + } memset(mmaps, 0, sizeof(mmaps)); @@ -1320,6 +1322,10 @@ remap: mmaps[map_idx] = buf; map_idx = (map_idx + 1) & (ARRAY_SIZE(mmaps) - 1); file_pos = file_offset + head; + if (session->one_mmap) { + session->one_mmap_addr = buf; + session->one_mmap_offset = file_offset; + } more: event = fetch_mmaped_event(session, head, mmap_size, buf); @@ -1365,6 +1371,7 @@ out_err: ui_progress__finish(); perf_session__warn_about_errors(session, tool); perf_session_free_sample_buffers(session); + session->one_mmap = false; return err; } diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 3140f8a..0321013 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -36,6 +36,9 @@ struct perf_session { struct trace_event tevent; struct events_stats stats; bool repipe; + bool one_mmap; + void *one_mmap_addr; + u64 one_mmap_offset; struct ordered_samples ordered_samples; struct perf_data_file *file; }; -- cgit v1.1 From a8a8f3eb5de55aeaf007c18572668e8ec463547b Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:52 +0300 Subject: perf evlist: Pass mmap parameters in a struct In preparation for adding more mmap parameters, pass existing parameters in a struct. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-29-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evlist.c | 46 ++++++++++++++++++++++++++-------------------- 1 file changed, 26 insertions(+), 20 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index c51223a..814e954 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -606,12 +606,17 @@ static int perf_evlist__alloc_mmap(struct perf_evlist *evlist) return evlist->mmap != NULL ? 0 : -ENOMEM; } -static int __perf_evlist__mmap(struct perf_evlist *evlist, - int idx, int prot, int mask, int fd) +struct mmap_params { + int prot; + int mask; +}; + +static int __perf_evlist__mmap(struct perf_evlist *evlist, int idx, + struct mmap_params *mp, int fd) { evlist->mmap[idx].prev = 0; - evlist->mmap[idx].mask = mask; - evlist->mmap[idx].base = mmap(NULL, evlist->mmap_len, prot, + evlist->mmap[idx].mask = mp->mask; + evlist->mmap[idx].base = mmap(NULL, evlist->mmap_len, mp->prot, MAP_SHARED, fd, 0); if (evlist->mmap[idx].base == MAP_FAILED) { pr_debug2("failed to mmap perf event ring buffer, error %d\n", @@ -625,8 +630,8 @@ static int __perf_evlist__mmap(struct perf_evlist *evlist, } static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx, - int prot, int mask, int cpu, int thread, - int *output) + struct mmap_params *mp, int cpu, + int thread, int *output) { struct perf_evsel *evsel; @@ -635,8 +640,7 @@ static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx, if (*output == -1) { *output = fd; - if (__perf_evlist__mmap(evlist, idx, prot, mask, - *output) < 0) + if (__perf_evlist__mmap(evlist, idx, mp, *output) < 0) return -1; } else { if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, *output) != 0) @@ -651,8 +655,8 @@ static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx, return 0; } -static int perf_evlist__mmap_per_cpu(struct perf_evlist *evlist, int prot, - int mask) +static int perf_evlist__mmap_per_cpu(struct perf_evlist *evlist, + struct mmap_params *mp) { int cpu, thread; int nr_cpus = cpu_map__nr(evlist->cpus); @@ -663,8 +667,8 @@ static int perf_evlist__mmap_per_cpu(struct perf_evlist *evlist, int prot, int output = -1; for (thread = 0; thread < nr_threads; thread++) { - if (perf_evlist__mmap_per_evsel(evlist, cpu, prot, mask, - cpu, thread, &output)) + if (perf_evlist__mmap_per_evsel(evlist, cpu, mp, cpu, + thread, &output)) goto out_unmap; } } @@ -677,8 +681,8 @@ out_unmap: return -1; } -static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist, int prot, - int mask) +static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist, + struct mmap_params *mp) { int thread; int nr_threads = thread_map__nr(evlist->threads); @@ -687,8 +691,8 @@ static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist, int prot, for (thread = 0; thread < nr_threads; thread++) { int output = -1; - if (perf_evlist__mmap_per_evsel(evlist, thread, prot, mask, 0, - thread, &output)) + if (perf_evlist__mmap_per_evsel(evlist, thread, mp, 0, thread, + &output)) goto out_unmap; } @@ -793,7 +797,9 @@ int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages, struct perf_evsel *evsel; const struct cpu_map *cpus = evlist->cpus; const struct thread_map *threads = evlist->threads; - int prot = PROT_READ | (overwrite ? 0 : PROT_WRITE), mask; + struct mmap_params mp = { + .prot = PROT_READ | (overwrite ? 0 : PROT_WRITE), + }; if (evlist->mmap == NULL && perf_evlist__alloc_mmap(evlist) < 0) return -ENOMEM; @@ -804,7 +810,7 @@ int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages, evlist->overwrite = overwrite; evlist->mmap_len = perf_evlist__mmap_size(pages); pr_debug("mmap size %zuB\n", evlist->mmap_len); - mask = evlist->mmap_len - page_size - 1; + mp.mask = evlist->mmap_len - page_size - 1; evlist__for_each(evlist, evsel) { if ((evsel->attr.read_format & PERF_FORMAT_ID) && @@ -814,9 +820,9 @@ int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages, } if (cpu_map__empty(cpus)) - return perf_evlist__mmap_per_thread(evlist, prot, mask); + return perf_evlist__mmap_per_thread(evlist, &mp); - return perf_evlist__mmap_per_cpu(evlist, prot, mask); + return perf_evlist__mmap_per_cpu(evlist, &mp); } int perf_evlist__create_maps(struct perf_evlist *evlist, struct target *target) -- cgit v1.1 From f6d313699a9612a30fabe05bf2c9302c1408b5cf Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:53 +0300 Subject: perf tools: Add feature test for __sync_val_compare_and_swap Add a feature test for __sync_val_compare_and_swap() and __sync_bool_compare_and_swap() Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-30-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/config/Makefile | 6 ++++++ tools/perf/config/feature-checks/Makefile | 4 ++++ tools/perf/config/feature-checks/test-all.c | 5 +++++ .../config/feature-checks/test-sync-compare-and-swap.c | 14 ++++++++++++++ 4 files changed, 29 insertions(+) create mode 100644 tools/perf/config/feature-checks/test-sync-compare-and-swap.c (limited to 'tools') diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile index b7f42d5..1f67aa0 100644 --- a/tools/perf/config/Makefile +++ b/tools/perf/config/Makefile @@ -164,6 +164,7 @@ CORE_FEATURE_TESTS = \ backtrace \ dwarf \ fortify-source \ + sync-compare-and-swap \ glibc \ gtk2 \ gtk2-infobar \ @@ -199,6 +200,7 @@ LIB_FEATURE_TESTS = \ VF_FEATURE_TESTS = \ backtrace \ fortify-source \ + sync-compare-and-swap \ gtk2-infobar \ libelf-getphdrnum \ libelf-mmap \ @@ -272,6 +274,10 @@ CFLAGS += -I$(LIB_INCLUDE) CFLAGS += -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE +ifeq ($(feature-sync-compare-and-swap), 1) + CFLAGS += -DHAVE_SYNC_COMPARE_AND_SWAP_SUPPORT +endif + ifndef NO_BIONIC $(call feature_check,bionic) ifeq ($(feature-bionic), 1) diff --git a/tools/perf/config/feature-checks/Makefile b/tools/perf/config/feature-checks/Makefile index 64c84e5..6088f8d 100644 --- a/tools/perf/config/feature-checks/Makefile +++ b/tools/perf/config/feature-checks/Makefile @@ -5,6 +5,7 @@ FILES= \ test-bionic.bin \ test-dwarf.bin \ test-fortify-source.bin \ + test-sync-compare-and-swap.bin \ test-glibc.bin \ test-gtk2.bin \ test-gtk2-infobar.bin \ @@ -141,6 +142,9 @@ test-timerfd.bin: test-libdw-dwarf-unwind.bin: $(BUILD) +test-sync-compare-and-swap.bin: + $(BUILD) -Werror + -include *.d ############################### diff --git a/tools/perf/config/feature-checks/test-all.c b/tools/perf/config/feature-checks/test-all.c index fe5c1e5..a7d022e 100644 --- a/tools/perf/config/feature-checks/test-all.c +++ b/tools/perf/config/feature-checks/test-all.c @@ -89,6 +89,10 @@ # include "test-libdw-dwarf-unwind.c" #undef main +#define main main_test_sync_compare_and_swap +# include "test-sync-compare-and-swap.c" +#undef main + int main(int argc, char *argv[]) { main_test_libpython(); @@ -111,6 +115,7 @@ int main(int argc, char *argv[]) main_test_timerfd(); main_test_stackprotector_all(); main_test_libdw_dwarf_unwind(); + main_test_sync_compare_and_swap(argc, argv); return 0; } diff --git a/tools/perf/config/feature-checks/test-sync-compare-and-swap.c b/tools/perf/config/feature-checks/test-sync-compare-and-swap.c new file mode 100644 index 0000000..c34d4ca --- /dev/null +++ b/tools/perf/config/feature-checks/test-sync-compare-and-swap.c @@ -0,0 +1,14 @@ +#include + +volatile uint64_t x; + +int main(int argc, char *argv[]) +{ + uint64_t old, new = argc; + + argv = argv; + do { + old = __sync_val_compare_and_swap(&x, 0, 0); + } while (!__sync_bool_compare_and_swap(&x, old, new)); + return old == new; +} -- cgit v1.1 From ea8e08a16a1e6566be3f775c0bd351fa52ab6b9d Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:54 +0300 Subject: perf tools: Add option macro OPT_CALLBACK_OPTARG Add an option macro that is the same as OPT_CALLBACK except that the argument is optional and it is possible to associate additional data with it. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-31-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-options.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/parse-options.h b/tools/perf/util/parse-options.h index d8dac8a..b59ba85 100644 --- a/tools/perf/util/parse-options.h +++ b/tools/perf/util/parse-options.h @@ -98,6 +98,7 @@ struct option { parse_opt_cb *callback; intptr_t defval; bool *set; + void *data; }; #define check_vtype(v, type) ( BUILD_BUG_ON_ZERO(!__builtin_types_compatible_p(typeof(v), type)) + v ) @@ -131,6 +132,10 @@ struct option { { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l),\ .value = (v), (a), .help = (h), .callback = (f), .defval = (intptr_t)d,\ .flags = PARSE_OPT_LASTARG_DEFAULT | PARSE_OPT_NOARG} +#define OPT_CALLBACK_OPTARG(s, l, v, d, a, h, f) \ + { .type = OPTION_CALLBACK, .short_name = (s), .long_name = (l), \ + .value = (v), (a), .help = (h), .callback = (f), \ + .flags = PARSE_OPT_OPTARG, .data = (d) } /* parse_options() will filter out the processed options and leave the * non-option argments in argv[]. -- cgit v1.1 From 6ff1ce763921f605aaf98c7a828b7df24d6923dc Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:56 +0300 Subject: perf evsel: Add 'no_aux_samples' option Add an option to prevent additional samples being added to a selected event by perf_evsel__config(). This is needed when using the sched_switch tracepoint to follow object code execution. Since sched_switch will be used only for switch information, additional sampling is wasteful. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-33-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evsel.c | 6 +++--- tools/perf/util/evsel.h | 1 + 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index b760d32..40626e5 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -623,7 +623,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts) attr->mmap_data = track; } - if (opts->call_graph_enabled) + if (opts->call_graph_enabled && !evsel->no_aux_samples) perf_evsel__config_callgraph(evsel, opts); if (target__has_cpu(&opts->target)) @@ -637,7 +637,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts) target__has_cpu(&opts->target) || per_cpu)) perf_evsel__set_sample_bit(evsel, TIME); - if (opts->raw_samples) { + if (opts->raw_samples && !evsel->no_aux_samples) { perf_evsel__set_sample_bit(evsel, TIME); perf_evsel__set_sample_bit(evsel, RAW); perf_evsel__set_sample_bit(evsel, CPU); @@ -650,7 +650,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts) attr->watermark = 0; attr->wakeup_events = 1; } - if (opts->branch_stack) { + if (opts->branch_stack && !evsel->no_aux_samples) { perf_evsel__set_sample_bit(evsel, BRANCH_STACK); attr->branch_sample_type = opts->branch_stack; } diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index a52e9a5..8dfec05 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -83,6 +83,7 @@ struct perf_evsel { int is_pos; bool supported; bool needs_swap; + bool no_aux_samples; /* parse modifier helper */ int exclude_GH; int nr_members; -- cgit v1.1 From 2afd2bcfc3a026d6e4c2184bf41ccd74eb05758b Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:02:57 +0300 Subject: perf evsel: Add 'immediate' option Add an option to cause a selected event to be enabled immediately when configured by perf_evsel__config(). This is needed when using the sched_switch tracepoint to follow object code execution. By having sched_switch enabled immediately the first sched_switch event precedes the start of other tracing. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-34-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evsel.c | 5 +++++ tools/perf/util/evsel.h | 1 + 2 files changed, 6 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 40626e5..90f58cd 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -681,6 +681,11 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts) if (target__none(&opts->target) && perf_evsel__is_group_leader(evsel) && !opts->initial_delay) attr->enable_on_exec = 1; + + if (evsel->immediate) { + attr->disabled = 0; + attr->enable_on_exec = 0; + } } int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads) diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index 8dfec05..d7f93ce 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -84,6 +84,7 @@ struct perf_evsel { bool supported; bool needs_swap; bool no_aux_samples; + bool immediate; /* parse modifier helper */ int exclude_GH; int nr_members; -- cgit v1.1 From 29ce36121e6738012aaf00d983d25260627f2b0d Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 16 Jul 2014 11:07:13 +0300 Subject: perf machine: Fix map groups of threads with unknown pids Events like sched_switch do not provide a pid (tgid) which can result in threads with an unknown pid. If the pid is later discovered, join the map groups. Note the thread's map groups should be empty because they are populated by MMAP events which do provide the pid and tid. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405498033-23817-1-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 57 ++++++++++++++++++++++++++++++++++++++++++----- tools/perf/util/map.c | 14 ++++++++++++ tools/perf/util/map.h | 1 + 3 files changed, 66 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 5b80877..5484fa4 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -272,6 +272,52 @@ void machines__set_id_hdr_size(struct machines *machines, u16 id_hdr_size) return; } +static void machine__update_thread_pid(struct machine *machine, + struct thread *th, pid_t pid) +{ + struct thread *leader; + + if (pid == th->pid_ || pid == -1 || th->pid_ != -1) + return; + + th->pid_ = pid; + + if (th->pid_ == th->tid) + return; + + leader = machine__findnew_thread(machine, th->pid_, th->pid_); + if (!leader) + goto out_err; + + if (!leader->mg) + leader->mg = map_groups__new(); + + if (!leader->mg) + goto out_err; + + if (th->mg == leader->mg) + return; + + if (th->mg) { + /* + * Maps are created from MMAP events which provide the pid and + * tid. Consequently there never should be any maps on a thread + * with an unknown pid. Just print an error if there are. + */ + if (!map_groups__empty(th->mg)) + pr_err("Discarding thread maps for %d:%d\n", + th->pid_, th->tid); + map_groups__delete(th->mg); + } + + th->mg = map_groups__get(leader->mg); + + return; + +out_err: + pr_err("Failed to join map groups for %d:%d\n", th->pid_, th->tid); +} + static struct thread *__machine__findnew_thread(struct machine *machine, pid_t pid, pid_t tid, bool create) @@ -285,10 +331,10 @@ static struct thread *__machine__findnew_thread(struct machine *machine, * so most of the time we dont have to look up * the full rbtree: */ - if (machine->last_match && machine->last_match->tid == tid) { - if (pid != -1 && pid != machine->last_match->pid_) - machine->last_match->pid_ = pid; - return machine->last_match; + th = machine->last_match; + if (th && th->tid == tid) { + machine__update_thread_pid(machine, th, pid); + return th; } while (*p != NULL) { @@ -297,8 +343,7 @@ static struct thread *__machine__findnew_thread(struct machine *machine, if (th->tid == tid) { machine->last_match = th; - if (pid != -1 && pid != th->pid_) - th->pid_ = pid; + machine__update_thread_pid(machine, th, pid); return th; } diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 25c571f..7af1480 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -454,6 +454,20 @@ void map_groups__exit(struct map_groups *mg) } } +bool map_groups__empty(struct map_groups *mg) +{ + int i; + + for (i = 0; i < MAP__NR_TYPES; ++i) { + if (maps__first(&mg->maps[i])) + return false; + if (!list_empty(&mg->removed_maps[i])) + return false; + } + + return true; +} + struct map_groups *map_groups__new(void) { struct map_groups *mg = malloc(sizeof(*mg)); diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 7758c72..5806a90 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -66,6 +66,7 @@ struct map_groups { struct map_groups *map_groups__new(void); void map_groups__delete(struct map_groups *mg); +bool map_groups__empty(struct map_groups *mg); static inline struct map_groups *map_groups__get(struct map_groups *mg) { -- cgit v1.1 From 9608b84e4dd95341b88cad646b114811f5bccbba Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 16 Jul 2014 10:19:43 +0300 Subject: perf thread: Allow deletion of a thread with no map groups It needs to be possible to call thread__delete() on a thread with no map groups. This is needed for a subsequent patch which deletes a thread on the error path before map groups have been attached. Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405495184-20441-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Adrian Hunter Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/thread.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 7a32f44..b9c36ef 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -60,8 +60,10 @@ void thread__delete(struct thread *thread) { struct comm *comm, *tmp; - map_groups__put(thread->mg); - thread->mg = NULL; + if (thread->mg) { + map_groups__put(thread->mg); + thread->mg = NULL; + } list_for_each_entry_safe(comm, tmp, &thread->comm_list, list) { list_del(&comm->list); comm__free(comm); -- cgit v1.1 From 418029b7324f8b90ac1dfbc8a44555d6905be761 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 16 Jul 2014 10:19:44 +0300 Subject: perf machine: Fix leak of 'struct thread' on error path __machine__findnew_thread() creates a 'struct thread' but does not free it on the error path. Fix it. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405495184-20441-3-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 5484fa4..93c8b6f 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -370,8 +370,10 @@ static struct thread *__machine__findnew_thread(struct machine *machine, * within thread__init_map_groups to find the thread * leader and that would screwed the rb tree. */ - if (thread__init_map_groups(th, machine)) + if (thread__init_map_groups(th, machine)) { + thread__delete(th); return NULL; + } } return th; -- cgit v1.1 From acebd408bef17169fbf79079b96f0264b535916c Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 14 Jul 2014 23:46:47 +0200 Subject: perf tools: Remove verbose from functions prototypes And use verbose as an global object in following functions: __map_groups__fprintf_maps __map_groups__fprintf_removed_maps map_groups__fprintf_maps map_groups__fprintf Also making map_groups__fprintf_maps static. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1405374411-29012-2-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/stdio/hist.c | 2 +- tools/perf/util/map.c | 24 ++++++++++++------------ tools/perf/util/map.h | 9 ++++----- tools/perf/util/thread.c | 4 ++-- 4 files changed, 19 insertions(+), 20 deletions(-) (limited to 'tools') diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c index 90122ab..40af0ac 100644 --- a/tools/perf/ui/stdio/hist.c +++ b/tools/perf/ui/stdio/hist.c @@ -479,7 +479,7 @@ print_entries: if (h->ms.map == NULL && verbose > 1) { __map_groups__fprintf_maps(h->thread->mg, - MAP__FUNCTION, verbose, fp); + MAP__FUNCTION, fp); fprintf(fp, "%.10s end\n", graph_dotted_line); } } diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 7af1480..845f627 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -12,6 +12,7 @@ #include "vdso.h" #include "build-id.h" #include "util.h" +#include "debug.h" #include const char *map_type__name[MAP__NR_TYPES] = { @@ -568,8 +569,8 @@ int map_groups__find_ams(struct addr_map_symbol *ams, symbol_filter_t filter) return ams->sym ? 0 : -1; } -size_t __map_groups__fprintf_maps(struct map_groups *mg, - enum map_type type, int verbose, FILE *fp) +size_t __map_groups__fprintf_maps(struct map_groups *mg, enum map_type type, + FILE *fp) { size_t printed = fprintf(fp, "%s:\n", map_type__name[type]); struct rb_node *nd; @@ -587,17 +588,16 @@ size_t __map_groups__fprintf_maps(struct map_groups *mg, return printed; } -size_t map_groups__fprintf_maps(struct map_groups *mg, int verbose, FILE *fp) +static size_t map_groups__fprintf_maps(struct map_groups *mg, FILE *fp) { size_t printed = 0, i; for (i = 0; i < MAP__NR_TYPES; ++i) - printed += __map_groups__fprintf_maps(mg, i, verbose, fp); + printed += __map_groups__fprintf_maps(mg, i, fp); return printed; } static size_t __map_groups__fprintf_removed_maps(struct map_groups *mg, - enum map_type type, - int verbose, FILE *fp) + enum map_type type, FILE *fp) { struct map *pos; size_t printed = 0; @@ -614,23 +614,23 @@ static size_t __map_groups__fprintf_removed_maps(struct map_groups *mg, } static size_t map_groups__fprintf_removed_maps(struct map_groups *mg, - int verbose, FILE *fp) + FILE *fp) { size_t printed = 0, i; for (i = 0; i < MAP__NR_TYPES; ++i) - printed += __map_groups__fprintf_removed_maps(mg, i, verbose, fp); + printed += __map_groups__fprintf_removed_maps(mg, i, fp); return printed; } -size_t map_groups__fprintf(struct map_groups *mg, int verbose, FILE *fp) +size_t map_groups__fprintf(struct map_groups *mg, FILE *fp) { - size_t printed = map_groups__fprintf_maps(mg, verbose, fp); + size_t printed = map_groups__fprintf_maps(mg, fp); printed += fprintf(fp, "Removed maps:\n"); - return printed + map_groups__fprintf_removed_maps(mg, verbose, fp); + return printed + map_groups__fprintf_removed_maps(mg, fp); } int map_groups__fixup_overlappings(struct map_groups *mg, struct map *map, - int verbose, FILE *fp) + FILE *fp) { struct rb_root *root = &mg->maps[map->type]; struct rb_node *next = rb_first(root); diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 5806a90..22d13a2 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -142,8 +142,8 @@ void map__fixup_end(struct map *map); void map__reloc_vmlinux(struct map *map); -size_t __map_groups__fprintf_maps(struct map_groups *mg, - enum map_type type, int verbose, FILE *fp); +size_t __map_groups__fprintf_maps(struct map_groups *mg, enum map_type type, + FILE *fp); void maps__insert(struct rb_root *maps, struct map *map); void maps__remove(struct rb_root *maps, struct map *map); struct map *maps__find(struct rb_root *maps, u64 addr); @@ -153,8 +153,7 @@ void map_groups__init(struct map_groups *mg); void map_groups__exit(struct map_groups *mg); int map_groups__clone(struct map_groups *mg, struct map_groups *parent, enum map_type type); -size_t map_groups__fprintf(struct map_groups *mg, int verbose, FILE *fp); -size_t map_groups__fprintf_maps(struct map_groups *mg, int verbose, FILE *fp); +size_t map_groups__fprintf(struct map_groups *mg, FILE *fp); int maps__set_kallsyms_ref_reloc_sym(struct map **maps, const char *symbol_name, u64 addr); @@ -211,7 +210,7 @@ struct symbol *map_groups__find_function_by_name(struct map_groups *mg, } int map_groups__fixup_overlappings(struct map_groups *mg, struct map *map, - int verbose, FILE *fp); + FILE *fp); struct map *map_groups__find_by_name(struct map_groups *mg, enum map_type type, const char *name); diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index b9c36ef..9692c06 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -129,12 +129,12 @@ int thread__comm_len(struct thread *thread) size_t thread__fprintf(struct thread *thread, FILE *fp) { return fprintf(fp, "Thread %d %s\n", thread->tid, thread__comm_str(thread)) + - map_groups__fprintf(thread->mg, verbose, fp); + map_groups__fprintf(thread->mg, fp); } void thread__insert_map(struct thread *thread, struct map *map) { - map_groups__fixup_overlappings(thread->mg, map, verbose, stderr); + map_groups__fixup_overlappings(thread->mg, map, stderr); map_groups__insert(thread->mg, map); } -- cgit v1.1 From 84f5d36f486609277801e827241396334185d11c Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 14 Jul 2014 23:46:48 +0200 Subject: perf tools: Move pr_* debug macros into debug object Moving pr_* debug macros to have it with in same object as debug variables, becase we will change them to use verbose variable in next patch. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1405374411-29012-3-git-send-email-jolsa@kernel.org [ Add missing debug.h include in python scripting glue and in the libdw unwind lib ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/x86/tests/dwarf-unwind.c | 1 + tools/perf/arch/x86/util/unwind-libunwind.c | 1 + tools/perf/builtin-evlist.c | 1 + tools/perf/builtin-help.c | 1 + tools/perf/builtin-timechart.c | 1 + tools/perf/tests/dso-data.c | 1 + tools/perf/tests/evsel-roundtrip-name.c | 1 + tools/perf/tests/evsel-tp-sched.c | 1 + tools/perf/tests/open-syscall-tp-fields.c | 1 + tools/perf/tests/parse-events.c | 1 + tools/perf/tests/parse-no-sample-id-all.c | 1 + tools/perf/tests/sample-parsing.c | 1 + tools/perf/tests/thread-mg-share.c | 1 + tools/perf/util/data.c | 1 + tools/perf/util/debug.h | 20 ++++++++++++++++++++ tools/perf/util/include/linux/kernel.h | 21 --------------------- tools/perf/util/pstack.c | 1 + .../perf/util/scripting-engines/trace-event-perl.c | 1 + .../util/scripting-engines/trace-event-python.c | 1 + tools/perf/util/trace-event-info.c | 1 + tools/perf/util/trace-event-read.c | 1 + tools/perf/util/unwind-libdw.c | 1 + tools/perf/util/unwind-libunwind.c | 1 + tools/perf/util/util.c | 1 + tools/perf/util/vdso.c | 1 + 25 files changed, 43 insertions(+), 21 deletions(-) (limited to 'tools') diff --git a/tools/perf/arch/x86/tests/dwarf-unwind.c b/tools/perf/arch/x86/tests/dwarf-unwind.c index 9f89f89..d8bbf7a 100644 --- a/tools/perf/arch/x86/tests/dwarf-unwind.c +++ b/tools/perf/arch/x86/tests/dwarf-unwind.c @@ -3,6 +3,7 @@ #include "thread.h" #include "map.h" #include "event.h" +#include "debug.h" #include "tests/tests.h" #define STACK_SIZE 8192 diff --git a/tools/perf/arch/x86/util/unwind-libunwind.c b/tools/perf/arch/x86/util/unwind-libunwind.c index 3261f68..db25e93 100644 --- a/tools/perf/arch/x86/util/unwind-libunwind.c +++ b/tools/perf/arch/x86/util/unwind-libunwind.c @@ -3,6 +3,7 @@ #include #include "perf_regs.h" #include "../../util/unwind.h" +#include "../../util/debug.h" #ifdef HAVE_ARCH_X86_64_SUPPORT int libunwind__arch_reg_id(int regnum) diff --git a/tools/perf/builtin-evlist.c b/tools/perf/builtin-evlist.c index c99e0de..66e12f5 100644 --- a/tools/perf/builtin-evlist.c +++ b/tools/perf/builtin-evlist.c @@ -15,6 +15,7 @@ #include "util/parse-options.h" #include "util/session.h" #include "util/data.h" +#include "util/debug.h" static int __cmd_evlist(const char *file_name, struct perf_attr_details *details) { diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index 178b88a..0384d93 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -11,6 +11,7 @@ #include "util/parse-options.h" #include "util/run-command.h" #include "util/help.h" +#include "util/debug.h" static struct man_viewer_list { struct man_viewer_list *next; diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 04c9c53..2f1a522 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -37,6 +37,7 @@ #include "util/svghelper.h" #include "util/tool.h" #include "util/data.h" +#include "util/debug.h" #define SUPPORT_OLD_POWER_EVENTS 1 #define PWR_EVENT_EXIT -1 diff --git a/tools/perf/tests/dso-data.c b/tools/perf/tests/dso-data.c index 630808c..caaf37f 100644 --- a/tools/perf/tests/dso-data.c +++ b/tools/perf/tests/dso-data.c @@ -10,6 +10,7 @@ #include "machine.h" #include "symbol.h" #include "tests.h" +#include "debug.h" static char *test_file(int size) { diff --git a/tools/perf/tests/evsel-roundtrip-name.c b/tools/perf/tests/evsel-roundtrip-name.c index 465cdbc..b8d8341 100644 --- a/tools/perf/tests/evsel-roundtrip-name.c +++ b/tools/perf/tests/evsel-roundtrip-name.c @@ -2,6 +2,7 @@ #include "evsel.h" #include "parse-events.h" #include "tests.h" +#include "debug.h" static int perf_evsel__roundtrip_cache_name_test(void) { diff --git a/tools/perf/tests/evsel-tp-sched.c b/tools/perf/tests/evsel-tp-sched.c index 35d7fdb..5216242 100644 --- a/tools/perf/tests/evsel-tp-sched.c +++ b/tools/perf/tests/evsel-tp-sched.c @@ -1,6 +1,7 @@ #include #include "evsel.h" #include "tests.h" +#include "debug.h" static int perf_evsel__test_field(struct perf_evsel *evsel, const char *name, int size, bool should_be_signed) diff --git a/tools/perf/tests/open-syscall-tp-fields.c b/tools/perf/tests/open-syscall-tp-fields.c index c505ef2..0785b64 100644 --- a/tools/perf/tests/open-syscall-tp-fields.c +++ b/tools/perf/tests/open-syscall-tp-fields.c @@ -3,6 +3,7 @@ #include "evsel.h" #include "thread_map.h" #include "tests.h" +#include "debug.h" int test__syscall_open_tp_fields(void) { diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c index deba669..5941927 100644 --- a/tools/perf/tests/parse-events.c +++ b/tools/perf/tests/parse-events.c @@ -5,6 +5,7 @@ #include #include #include "tests.h" +#include "debug.h" #include #define PERF_TP_SAMPLE_TYPE (PERF_SAMPLE_RAW | PERF_SAMPLE_TIME | \ diff --git a/tools/perf/tests/parse-no-sample-id-all.c b/tools/perf/tests/parse-no-sample-id-all.c index 905019f..2c63ea6 100644 --- a/tools/perf/tests/parse-no-sample-id-all.c +++ b/tools/perf/tests/parse-no-sample-id-all.c @@ -7,6 +7,7 @@ #include "evlist.h" #include "header.h" #include "util.h" +#include "debug.h" static int process_event(struct perf_evlist **pevlist, union perf_event *event) { diff --git a/tools/perf/tests/sample-parsing.c b/tools/perf/tests/sample-parsing.c index 7ae8d17..ca292f9 100644 --- a/tools/perf/tests/sample-parsing.c +++ b/tools/perf/tests/sample-parsing.c @@ -4,6 +4,7 @@ #include "util.h" #include "event.h" #include "evsel.h" +#include "debug.h" #include "tests.h" diff --git a/tools/perf/tests/thread-mg-share.c b/tools/perf/tests/thread-mg-share.c index 2b2e0db..b028499 100644 --- a/tools/perf/tests/thread-mg-share.c +++ b/tools/perf/tests/thread-mg-share.c @@ -2,6 +2,7 @@ #include "machine.h" #include "thread.h" #include "map.h" +#include "debug.h" int test__thread_mg_share(void) { diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c index ee370a7f..29d720c 100644 --- a/tools/perf/util/data.c +++ b/tools/perf/util/data.c @@ -7,6 +7,7 @@ #include "data.h" #include "util.h" +#include "debug.h" static bool check_pipe(struct perf_data_file *file) { diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h index 443694c..8a8ceb3 100644 --- a/tools/perf/util/debug.h +++ b/tools/perf/util/debug.h @@ -11,6 +11,24 @@ extern int verbose; extern bool quiet, dump_trace; +#ifndef pr_fmt +#define pr_fmt(fmt) fmt +#endif + +#define pr_err(fmt, ...) \ + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_warning(fmt, ...) \ + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_info(fmt, ...) \ + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debug(fmt, ...) \ + eprintf(1, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debugN(n, fmt, ...) \ + eprintf(n, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debug2(fmt, ...) pr_debugN(2, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debug3(fmt, ...) pr_debugN(3, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debug4(fmt, ...) pr_debugN(4, pr_fmt(fmt), ##__VA_ARGS__) + int dump_printf(const char *fmt, ...) __attribute__((format(printf, 1, 2))); void trace_event(union perf_event *event); @@ -19,4 +37,6 @@ int ui__warning(const char *format, ...) __attribute__((format(printf, 1, 2))); void pr_stat(const char *fmt, ...); +int eprintf(int level, const char *fmt, ...) __attribute__((format(printf, 2, 3))); + #endif /* __PERF_DEBUG_H */ diff --git a/tools/perf/util/include/linux/kernel.h b/tools/perf/util/include/linux/kernel.h index 9844c31..09e8e7a 100644 --- a/tools/perf/util/include/linux/kernel.h +++ b/tools/perf/util/include/linux/kernel.h @@ -94,27 +94,6 @@ static inline int scnprintf(char * buf, size_t size, const char * fmt, ...) return (i >= ssize) ? (ssize - 1) : i; } -int eprintf(int level, - const char *fmt, ...) __attribute__((format(printf, 2, 3))); - -#ifndef pr_fmt -#define pr_fmt(fmt) fmt -#endif - -#define pr_err(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_warning(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_info(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_debug(fmt, ...) \ - eprintf(1, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_debugN(n, fmt, ...) \ - eprintf(n, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_debug2(fmt, ...) pr_debugN(2, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_debug3(fmt, ...) pr_debugN(3, pr_fmt(fmt), ##__VA_ARGS__) -#define pr_debug4(fmt, ...) pr_debugN(4, pr_fmt(fmt), ##__VA_ARGS__) - /* * This looks more complex than it should be. But we need to * get the type for the ~ right in round_down (it needs to be diff --git a/tools/perf/util/pstack.c b/tools/perf/util/pstack.c index daa17ae..a126e6c 100644 --- a/tools/perf/util/pstack.c +++ b/tools/perf/util/pstack.c @@ -6,6 +6,7 @@ #include "util.h" #include "pstack.h" +#include "debug.h" #include #include diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c index af7da56..b2dba9c 100644 --- a/tools/perf/util/scripting-engines/trace-event-perl.c +++ b/tools/perf/util/scripting-engines/trace-event-perl.c @@ -34,6 +34,7 @@ #include "../event.h" #include "../trace-event.h" #include "../evsel.h" +#include "../debug.h" void boot_Perf__Trace__Context(pTHX_ CV *cv); void boot_DynaLoader(pTHX_ CV *cv); diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c index b366b48..cbce254 100644 --- a/tools/perf/util/scripting-engines/trace-event-python.c +++ b/tools/perf/util/scripting-engines/trace-event-python.c @@ -27,6 +27,7 @@ #include #include "../../perf.h" +#include "../debug.h" #include "../evsel.h" #include "../util.h" #include "../event.h" diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index c3bba88..eb72716 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -40,6 +40,7 @@ #include "trace-event.h" #include #include "evsel.h" +#include "debug.h" #define VERSION "0.5" diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index e113e18..ea3fd7f 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -36,6 +36,7 @@ #include "../perf.h" #include "util.h" #include "trace-event.h" +#include "debug.h" static int input_fd; diff --git a/tools/perf/util/unwind-libdw.c b/tools/perf/util/unwind-libdw.c index 5ec80a5..7419768 100644 --- a/tools/perf/util/unwind-libdw.c +++ b/tools/perf/util/unwind-libdw.c @@ -3,6 +3,7 @@ #include #include #include +#include "debug.h" #include "unwind.h" #include "unwind-libdw.h" #include "machine.h" diff --git a/tools/perf/util/unwind-libunwind.c b/tools/perf/util/unwind-libunwind.c index 25578b9..92b56db 100644 --- a/tools/perf/util/unwind-libunwind.c +++ b/tools/perf/util/unwind-libunwind.c @@ -30,6 +30,7 @@ #include "unwind.h" #include "symbol.h" #include "util.h" +#include "debug.h" extern int UNW_OBJ(dwarf_search_unwind_table) (unw_addr_space_t as, diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c index e4132ae..e52e746 100644 --- a/tools/perf/util/util.c +++ b/tools/perf/util/util.c @@ -1,5 +1,6 @@ #include "../perf.h" #include "util.h" +#include "debug.h" #include #include #ifdef HAVE_BACKTRACE_SUPPORT diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index 0ddb3b8..2905824 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -12,6 +12,7 @@ #include "util.h" #include "symbol.h" #include "linux/string.h" +#include "debug.h" static bool vdso_found; static char vdso_file[] = "/tmp/perf-vdso.so-XXXXXX"; -- cgit v1.1 From c95688aac7723c17b2badc23233706b2f02e58ed Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 14 Jul 2014 23:46:49 +0200 Subject: perf tools: Factor eprintf to allow different debug variables This way we can easily reuse current debug functions for another debug variables other than verbose. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1405374411-29012-4-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/debug.c | 12 ++++++------ tools/perf/util/debug.h | 12 ++++++------ tools/perf/util/python.c | 4 ++-- 3 files changed, 14 insertions(+), 14 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c index 299b555..c208d6f 100644 --- a/tools/perf/util/debug.c +++ b/tools/perf/util/debug.c @@ -16,11 +16,11 @@ int verbose; bool dump_trace = false, quiet = false; -static int _eprintf(int level, const char *fmt, va_list args) +static int _eprintf(int level, int var, const char *fmt, va_list args) { int ret = 0; - if (verbose >= level) { + if (var >= level) { if (use_browser >= 1) ui_helpline__vshow(fmt, args); else @@ -30,13 +30,13 @@ static int _eprintf(int level, const char *fmt, va_list args) return ret; } -int eprintf(int level, const char *fmt, ...) +int eprintf(int level, int var, const char *fmt, ...) { va_list args; int ret; va_start(args, fmt); - ret = _eprintf(level, fmt, args); + ret = _eprintf(level, var, fmt, args); va_end(args); return ret; @@ -51,9 +51,9 @@ void pr_stat(const char *fmt, ...) va_list args; va_start(args, fmt); - _eprintf(1, fmt, args); + _eprintf(1, verbose, fmt, args); va_end(args); - eprintf(1, "\n"); + eprintf(1, verbose, "\n"); } int dump_printf(const char *fmt, ...) diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h index 8a8ceb3..1cb8081 100644 --- a/tools/perf/util/debug.h +++ b/tools/perf/util/debug.h @@ -16,15 +16,15 @@ extern bool quiet, dump_trace; #endif #define pr_err(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) + eprintf(0, verbose, pr_fmt(fmt), ##__VA_ARGS__) #define pr_warning(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) + eprintf(0, verbose, pr_fmt(fmt), ##__VA_ARGS__) #define pr_info(fmt, ...) \ - eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) + eprintf(0, verbose, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug(fmt, ...) \ - eprintf(1, pr_fmt(fmt), ##__VA_ARGS__) + eprintf(1, verbose, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debugN(n, fmt, ...) \ - eprintf(n, pr_fmt(fmt), ##__VA_ARGS__) + eprintf(n, verbose, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug2(fmt, ...) pr_debugN(2, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug3(fmt, ...) pr_debugN(3, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug4(fmt, ...) pr_debugN(4, pr_fmt(fmt), ##__VA_ARGS__) @@ -37,6 +37,6 @@ int ui__warning(const char *format, ...) __attribute__((format(printf, 1, 2))); void pr_stat(const char *fmt, ...); -int eprintf(int level, const char *fmt, ...) __attribute__((format(printf, 2, 3))); +int eprintf(int level, int var, const char *fmt, ...) __attribute__((format(printf, 3, 4))); #endif /* __PERF_DEBUG_H */ diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c index 122669c..12aa9b0 100644 --- a/tools/perf/util/python.c +++ b/tools/perf/util/python.c @@ -14,12 +14,12 @@ */ int verbose; -int eprintf(int level, const char *fmt, ...) +int eprintf(int level, int var, const char *fmt, ...) { va_list args; int ret = 0; - if (verbose >= level) { + if (var >= level) { va_start(args, fmt); ret = vfprintf(stderr, fmt, args); va_end(args); -- cgit v1.1 From bbb2cea7e8dd496b41558df1a0ec9205497b7ebf Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Thu, 17 Jul 2014 12:55:00 +0200 Subject: perf tools: Add --debug optionto set debug variable Adding --debug option as a way to setup debug variables. Starting with support for verbose, more will come. It's possible to use it now with report command: $ perf --debug verbose ... $ perf --debug verbose=2 ... I'll need this support to add separated debug variable for ordered events change in order to separate debug output out of standard verbose stream. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Namhyung Kim Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20140717105500.GG516@krava.redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf.txt | 10 ++++++++- tools/perf/perf.c | 13 +++++++++++- tools/perf/util/debug.c | 44 +++++++++++++++++++++++++++++++++++++++ tools/perf/util/debug.h | 2 ++ 4 files changed, 67 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf.txt b/tools/perf/Documentation/perf.txt index 0eeb247..d240bb2 100644 --- a/tools/perf/Documentation/perf.txt +++ b/tools/perf/Documentation/perf.txt @@ -8,7 +8,15 @@ perf - Performance analysis tools for Linux SYNOPSIS -------- [verse] -'perf' [--version] [--help] COMMAND [ARGS] +'perf' [--version] [--help] [OPTIONS] COMMAND [ARGS] + +OPTIONS +------- +--debug:: + Setup debug variable (just verbose for now) in value + range (0, 10). Use like: + --debug verbose # sets verbose = 1 + --debug verbose=2 # sets verbose = 2 DESCRIPTION ----------- diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 95c58fc..eed3fb2 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -13,11 +13,12 @@ #include "util/quote.h" #include "util/run-command.h" #include "util/parse-events.h" +#include "util/debug.h" #include #include const char perf_usage_string[] = - "perf [--version] [--help] COMMAND [ARGS]"; + "perf [--version] [--debug variable[=VALUE]] [--help] COMMAND [ARGS]"; const char perf_more_info_string[] = "See 'perf help COMMAND' for more information on a specific command."; @@ -212,6 +213,16 @@ static int handle_options(const char ***argv, int *argc, int *envchanged) printf("%s ", p->cmd); } exit(0); + } else if (!strcmp(cmd, "--debug")) { + if (*argc < 2) { + fprintf(stderr, "No variable specified for --debug.\n"); + usage(perf_usage_string); + } + if (perf_debug_option((*argv)[1])) + usage(perf_usage_string); + + (*argv)++; + (*argc)--; } else { fprintf(stderr, "Unknown option: %s\n", cmd); usage(perf_usage_string); diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c index c208d6f..71d4193 100644 --- a/tools/perf/util/debug.c +++ b/tools/perf/util/debug.c @@ -105,3 +105,47 @@ void trace_event(union perf_event *event) } printf(".\n"); } + +static struct debug_variable { + const char *name; + int *ptr; +} debug_variables[] = { + { .name = "verbose", .ptr = &verbose }, + { .name = NULL, } +}; + +int perf_debug_option(const char *str) +{ + struct debug_variable *var = &debug_variables[0]; + char *vstr, *s = strdup(str); + int v = 1; + + vstr = strchr(s, '='); + if (vstr) + *vstr++ = 0; + + while (var->name) { + if (!strcmp(s, var->name)) + break; + var++; + } + + if (!var->name) { + pr_err("Unknown debug variable name '%s'\n", s); + free(s); + return -1; + } + + if (vstr) { + v = atoi(vstr); + /* + * Allow only values in range (0, 10), + * otherwise set 0. + */ + v = (v < 0) || (v > 10) ? 0 : v; + } + + *var->ptr = v; + free(s); + return 0; +} diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h index 1cb8081..89fb6b0 100644 --- a/tools/perf/util/debug.h +++ b/tools/perf/util/debug.h @@ -39,4 +39,6 @@ void pr_stat(const char *fmt, ...); int eprintf(int level, int var, const char *fmt, ...) __attribute__((format(printf, 3, 4))); +int perf_debug_option(const char *str); + #endif /* __PERF_DEBUG_H */ -- cgit v1.1 From ff527bccd469067a64f4ae9747b9045914667d34 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 14 Jul 2014 23:46:51 +0200 Subject: perf tools: Remove needless getopt.h includes We don't use getopt.h interfaces. Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1405374411-29012-6-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-finder.c | 1 - tools/perf/util/trace-event-read.c | 1 - 2 files changed, 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index 98e3047..dca9145 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -26,7 +26,6 @@ #include #include #include -#include #include #include #include diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index ea3fd7f..54d9e9b 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -22,7 +22,6 @@ #include #include #include -#include #include #include #include -- cgit v1.1 From 0b437860818dc717f6a9e8a5089223a8414f5fff Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 14 Jul 2014 13:03:03 +0300 Subject: perf tools: Allow TSC conversion on any arch It is possible to record a perf.data file on one architecture and process it on another. Consequently, TSC conversion functions need to be moved out of the arch directory. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405332185-4050-40-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Makefile.perf | 2 ++ tools/perf/arch/x86/util/tsc.c | 22 +--------------------- tools/perf/arch/x86/util/tsc.h | 3 --- tools/perf/tests/perf-time-to-tsc.c | 3 +-- tools/perf/util/tsc.c | 25 +++++++++++++++++++++++++ tools/perf/util/tsc.h | 11 +++++++++++ 6 files changed, 40 insertions(+), 26 deletions(-) create mode 100644 tools/perf/util/tsc.c create mode 100644 tools/perf/util/tsc.h (limited to 'tools') diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf index 90c4983..3308b22 100644 --- a/tools/perf/Makefile.perf +++ b/tools/perf/Makefile.perf @@ -295,6 +295,7 @@ LIB_H += util/intlist.h LIB_H += util/perf_regs.h LIB_H += util/unwind.h LIB_H += util/vdso.h +LIB_H += util/tsc.h LIB_H += ui/helpline.h LIB_H += ui/progress.h LIB_H += ui/util.h @@ -374,6 +375,7 @@ LIB_OBJS += $(OUTPUT)util/stat.o LIB_OBJS += $(OUTPUT)util/record.o LIB_OBJS += $(OUTPUT)util/srcline.o LIB_OBJS += $(OUTPUT)util/data.o +LIB_OBJS += $(OUTPUT)util/tsc.o LIB_OBJS += $(OUTPUT)ui/setup.o LIB_OBJS += $(OUTPUT)ui/helpline.o diff --git a/tools/perf/arch/x86/util/tsc.c b/tools/perf/arch/x86/util/tsc.c index 40021fa..3655f24 100644 --- a/tools/perf/arch/x86/util/tsc.c +++ b/tools/perf/arch/x86/util/tsc.c @@ -6,29 +6,9 @@ #include "../../perf.h" #include #include "../../util/debug.h" +#include "../../util/tsc.h" #include "tsc.h" -u64 perf_time_to_tsc(u64 ns, struct perf_tsc_conversion *tc) -{ - u64 t, quot, rem; - - t = ns - tc->time_zero; - quot = t / tc->time_mult; - rem = t % tc->time_mult; - return (quot << tc->time_shift) + - (rem << tc->time_shift) / tc->time_mult; -} - -u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc) -{ - u64 quot, rem; - - quot = cyc >> tc->time_shift; - rem = cyc & ((1 << tc->time_shift) - 1); - return tc->time_zero + quot * tc->time_mult + - ((rem * tc->time_mult) >> tc->time_shift); -} - int perf_read_tsc_conversion(const struct perf_event_mmap_page *pc, struct perf_tsc_conversion *tc) { diff --git a/tools/perf/arch/x86/util/tsc.h b/tools/perf/arch/x86/util/tsc.h index 2affe03..2edc4d3 100644 --- a/tools/perf/arch/x86/util/tsc.h +++ b/tools/perf/arch/x86/util/tsc.h @@ -14,7 +14,4 @@ struct perf_event_mmap_page; int perf_read_tsc_conversion(const struct perf_event_mmap_page *pc, struct perf_tsc_conversion *tc); -u64 perf_time_to_tsc(u64 ns, struct perf_tsc_conversion *tc); -u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc); - #endif /* TOOLS_PERF_ARCH_X86_UTIL_TSC_H__ */ diff --git a/tools/perf/tests/perf-time-to-tsc.c b/tools/perf/tests/perf-time-to-tsc.c index 3b7cd4d..0372f6e 100644 --- a/tools/perf/tests/perf-time-to-tsc.c +++ b/tools/perf/tests/perf-time-to-tsc.c @@ -8,10 +8,9 @@ #include "evsel.h" #include "thread_map.h" #include "cpumap.h" +#include "tsc.h" #include "tests.h" -#include "../arch/x86/util/tsc.h" - #define CHECK__(x) { \ while ((x) < 0) { \ pr_debug(#x " failed!\n"); \ diff --git a/tools/perf/util/tsc.c b/tools/perf/util/tsc.c new file mode 100644 index 0000000..ef4749836 --- /dev/null +++ b/tools/perf/util/tsc.c @@ -0,0 +1,25 @@ +#include +#include + +#include "tsc.h" + +u64 perf_time_to_tsc(u64 ns, struct perf_tsc_conversion *tc) +{ + u64 t, quot, rem; + + t = ns - tc->time_zero; + quot = t / tc->time_mult; + rem = t % tc->time_mult; + return (quot << tc->time_shift) + + (rem << tc->time_shift) / tc->time_mult; +} + +u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc) +{ + u64 quot, rem; + + quot = cyc >> tc->time_shift; + rem = cyc & ((1 << tc->time_shift) - 1); + return tc->time_zero + quot * tc->time_mult + + ((rem * tc->time_mult) >> tc->time_shift); +} diff --git a/tools/perf/util/tsc.h b/tools/perf/util/tsc.h new file mode 100644 index 0000000..4eca848 --- /dev/null +++ b/tools/perf/util/tsc.h @@ -0,0 +1,11 @@ +#ifndef __PERF_TSC_H +#define __PERF_TSC_H + +#include + +#include "../arch/x86/util/tsc.h" + +u64 perf_time_to_tsc(u64 ns, struct perf_tsc_conversion *tc); +u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc); + +#endif -- cgit v1.1 From 57480d2cd93579d665e57e144e1e63f7f02ef058 Mon Sep 17 00:00:00 2001 From: Yann Droneaud Date: Mon, 30 Jun 2014 22:28:47 +0200 Subject: perf tools: Enable close-on-exec flag on perf file descriptor In commit a21b0b354d4a ('perf: Introduce a flag to enable close-on-exec in perf_event_open()'), flag PERF_FLAG_FD_CLOEXEC was added to perf_event_open(2) syscall to allows userspace to atomically enable close-on-exec behavor when creating the file descriptor. This patch makes perf tools use the new flag if supported by the kernel, so that the event file descriptors got automatically closed if perf tool exec a sub-command. Signed-off-by: Yann Droneaud Cc: Andi Kleen Cc: Arnaldo Carvalho de Melo Cc: Ingo Molnar Cc: Jiri Olsa Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/1404160127-7475-1-git-send-email-ydroneaud@opteya.com Signed-off-by: Jiri Olsa --- tools/perf/Makefile.perf | 1 + tools/perf/bench/mem-memcpy.c | 4 ++- tools/perf/bench/mem-memset.c | 4 ++- tools/perf/builtin-sched.c | 4 ++- tools/perf/tests/bp_signal.c | 4 ++- tools/perf/tests/bp_signal_overflow.c | 4 ++- tools/perf/tests/rdpmc.c | 4 ++- tools/perf/util/cloexec.c | 57 +++++++++++++++++++++++++++++++++++ tools/perf/util/cloexec.h | 6 ++++ tools/perf/util/evsel.c | 12 ++++++-- tools/perf/util/record.c | 9 ++++-- 11 files changed, 97 insertions(+), 12 deletions(-) create mode 100644 tools/perf/util/cloexec.c create mode 100644 tools/perf/util/cloexec.h (limited to 'tools') diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf index 3308b22..2240974 100644 --- a/tools/perf/Makefile.perf +++ b/tools/perf/Makefile.perf @@ -376,6 +376,7 @@ LIB_OBJS += $(OUTPUT)util/record.o LIB_OBJS += $(OUTPUT)util/srcline.o LIB_OBJS += $(OUTPUT)util/data.o LIB_OBJS += $(OUTPUT)util/tsc.o +LIB_OBJS += $(OUTPUT)util/cloexec.o LIB_OBJS += $(OUTPUT)ui/setup.o LIB_OBJS += $(OUTPUT)ui/helpline.o diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c index e622c3e..2465141 100644 --- a/tools/perf/bench/mem-memcpy.c +++ b/tools/perf/bench/mem-memcpy.c @@ -10,6 +10,7 @@ #include "../util/util.h" #include "../util/parse-options.h" #include "../util/header.h" +#include "../util/cloexec.h" #include "bench.h" #include "mem-memcpy-arch.h" @@ -83,7 +84,8 @@ static struct perf_event_attr cycle_attr = { static void init_cycle(void) { - cycle_fd = sys_perf_event_open(&cycle_attr, getpid(), -1, -1, 0); + cycle_fd = sys_perf_event_open(&cycle_attr, getpid(), -1, -1, + perf_event_open_cloexec_flag()); if (cycle_fd < 0 && errno == ENOSYS) die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); diff --git a/tools/perf/bench/mem-memset.c b/tools/perf/bench/mem-memset.c index 2a65468..75fc3e6 100644 --- a/tools/perf/bench/mem-memset.c +++ b/tools/perf/bench/mem-memset.c @@ -10,6 +10,7 @@ #include "../util/util.h" #include "../util/parse-options.h" #include "../util/header.h" +#include "../util/cloexec.h" #include "bench.h" #include "mem-memset-arch.h" @@ -83,7 +84,8 @@ static struct perf_event_attr cycle_attr = { static void init_cycle(void) { - cycle_fd = sys_perf_event_open(&cycle_attr, getpid(), -1, -1, 0); + cycle_fd = sys_perf_event_open(&cycle_attr, getpid(), -1, -1, + perf_event_open_cloexec_flag()); if (cycle_fd < 0 && errno == ENOSYS) die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index b7f555a..f83c08c 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -10,6 +10,7 @@ #include "util/header.h" #include "util/session.h" #include "util/tool.h" +#include "util/cloexec.h" #include "util/parse-options.h" #include "util/trace-event.h" @@ -434,7 +435,8 @@ static int self_open_counters(void) attr.type = PERF_TYPE_SOFTWARE; attr.config = PERF_COUNT_SW_TASK_CLOCK; - fd = sys_perf_event_open(&attr, 0, -1, -1, 0); + fd = sys_perf_event_open(&attr, 0, -1, -1, + perf_event_open_cloexec_flag()); if (fd < 0) pr_err("Error: sys_perf_event_open() syscall returned " diff --git a/tools/perf/tests/bp_signal.c b/tools/perf/tests/bp_signal.c index aba0954..a02b035 100644 --- a/tools/perf/tests/bp_signal.c +++ b/tools/perf/tests/bp_signal.c @@ -25,6 +25,7 @@ #include "tests.h" #include "debug.h" #include "perf.h" +#include "cloexec.h" static int fd1; static int fd2; @@ -78,7 +79,8 @@ static int bp_event(void *fn, int setup_signal) pe.exclude_kernel = 1; pe.exclude_hv = 1; - fd = sys_perf_event_open(&pe, 0, -1, -1, 0); + fd = sys_perf_event_open(&pe, 0, -1, -1, + perf_event_open_cloexec_flag()); if (fd < 0) { pr_debug("failed opening event %llx\n", pe.config); return TEST_FAIL; diff --git a/tools/perf/tests/bp_signal_overflow.c b/tools/perf/tests/bp_signal_overflow.c index 44ac821..e765377 100644 --- a/tools/perf/tests/bp_signal_overflow.c +++ b/tools/perf/tests/bp_signal_overflow.c @@ -24,6 +24,7 @@ #include "tests.h" #include "debug.h" #include "perf.h" +#include "cloexec.h" static int overflows; @@ -91,7 +92,8 @@ int test__bp_signal_overflow(void) pe.exclude_kernel = 1; pe.exclude_hv = 1; - fd = sys_perf_event_open(&pe, 0, -1, -1, 0); + fd = sys_perf_event_open(&pe, 0, -1, -1, + perf_event_open_cloexec_flag()); if (fd < 0) { pr_debug("failed opening event %llx\n", pe.config); return TEST_FAIL; diff --git a/tools/perf/tests/rdpmc.c b/tools/perf/tests/rdpmc.c index e59143f..c04d1f2 100644 --- a/tools/perf/tests/rdpmc.c +++ b/tools/perf/tests/rdpmc.c @@ -6,6 +6,7 @@ #include "perf.h" #include "debug.h" #include "tests.h" +#include "cloexec.h" #if defined(__x86_64__) || defined(__i386__) @@ -104,7 +105,8 @@ static int __test__rdpmc(void) sa.sa_sigaction = segfault_handler; sigaction(SIGSEGV, &sa, NULL); - fd = sys_perf_event_open(&attr, 0, -1, -1, 0); + fd = sys_perf_event_open(&attr, 0, -1, -1, + perf_event_open_cloexec_flag()); if (fd < 0) { pr_err("Error: sys_perf_event_open() syscall returned " "with %d (%s)\n", fd, strerror(errno)); diff --git a/tools/perf/util/cloexec.c b/tools/perf/util/cloexec.c new file mode 100644 index 0000000..c5d05ec --- /dev/null +++ b/tools/perf/util/cloexec.c @@ -0,0 +1,57 @@ +#include "util.h" +#include "../perf.h" +#include "cloexec.h" +#include "asm/bug.h" + +static unsigned long flag = PERF_FLAG_FD_CLOEXEC; + +static int perf_flag_probe(void) +{ + /* use 'safest' configuration as used in perf_evsel__fallback() */ + struct perf_event_attr attr = { + .type = PERF_COUNT_SW_CPU_CLOCK, + .config = PERF_COUNT_SW_CPU_CLOCK, + }; + int fd; + int err; + + /* check cloexec flag */ + fd = sys_perf_event_open(&attr, 0, -1, -1, + PERF_FLAG_FD_CLOEXEC); + err = errno; + + if (fd >= 0) { + close(fd); + return 1; + } + + WARN_ONCE(err != EINVAL, + "perf_event_open(..., PERF_FLAG_FD_CLOEXEC) failed with unexpected error %d (%s)\n", + err, strerror(err)); + + /* not supported, confirm error related to PERF_FLAG_FD_CLOEXEC */ + fd = sys_perf_event_open(&attr, 0, -1, -1, 0); + err = errno; + + if (WARN_ONCE(fd < 0, + "perf_event_open(..., 0) failed unexpectedly with error %d (%s)\n", + err, strerror(err))) + return -1; + + close(fd); + + return 0; +} + +unsigned long perf_event_open_cloexec_flag(void) +{ + static bool probed; + + if (!probed) { + if (perf_flag_probe() <= 0) + flag = 0; + probed = true; + } + + return flag; +} diff --git a/tools/perf/util/cloexec.h b/tools/perf/util/cloexec.h new file mode 100644 index 0000000..94a5a7d --- /dev/null +++ b/tools/perf/util/cloexec.h @@ -0,0 +1,6 @@ +#ifndef __PERF_CLOEXEC_H +#define __PERF_CLOEXEC_H + +unsigned long perf_event_open_cloexec_flag(void); + +#endif /* __PERF_CLOEXEC_H */ diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 90f58cd..21a373e 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -29,6 +29,7 @@ static struct { bool sample_id_all; bool exclude_guest; bool mmap2; + bool cloexec; } perf_missing_features; #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y)) @@ -994,7 +995,7 @@ static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, struct thread_map *threads) { int cpu, thread; - unsigned long flags = 0; + unsigned long flags = PERF_FLAG_FD_CLOEXEC; int pid = -1, err; enum { NO_CHANGE, SET_TO_MAX, INCREASED_MAX } set_rlimit = NO_CHANGE; @@ -1003,11 +1004,13 @@ static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, return -ENOMEM; if (evsel->cgrp) { - flags = PERF_FLAG_PID_CGROUP; + flags |= PERF_FLAG_PID_CGROUP; pid = evsel->cgrp->fd; } fallback_missing_features: + if (perf_missing_features.cloexec) + flags &= ~(unsigned long)PERF_FLAG_FD_CLOEXEC; if (perf_missing_features.mmap2) evsel->attr.mmap2 = 0; if (perf_missing_features.exclude_guest) @@ -1076,7 +1079,10 @@ try_fallback: if (err != -EINVAL || cpu > 0 || thread > 0) goto out_close; - if (!perf_missing_features.mmap2 && evsel->attr.mmap2) { + if (!perf_missing_features.cloexec && (flags & PERF_FLAG_FD_CLOEXEC)) { + perf_missing_features.cloexec = true; + goto fallback_missing_features; + } else if (!perf_missing_features.mmap2 && evsel->attr.mmap2) { perf_missing_features.mmap2 = true; goto fallback_missing_features; } else if (!perf_missing_features.exclude_guest && diff --git a/tools/perf/util/record.c b/tools/perf/util/record.c index 1657231..fe8079e 100644 --- a/tools/perf/util/record.c +++ b/tools/perf/util/record.c @@ -4,6 +4,7 @@ #include "parse-events.h" #include #include "util.h" +#include "cloexec.h" typedef void (*setup_probe_fn_t)(struct perf_evsel *evsel); @@ -11,6 +12,7 @@ static int perf_do_probe_api(setup_probe_fn_t fn, int cpu, const char *str) { struct perf_evlist *evlist; struct perf_evsel *evsel; + unsigned long flags = perf_event_open_cloexec_flag(); int err = -EAGAIN, fd; evlist = perf_evlist__new(); @@ -22,14 +24,14 @@ static int perf_do_probe_api(setup_probe_fn_t fn, int cpu, const char *str) evsel = perf_evlist__first(evlist); - fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, 0); + fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, flags); if (fd < 0) goto out_delete; close(fd); fn(evsel); - fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, 0); + fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, flags); if (fd < 0) { if (errno == EINVAL) err = -EINVAL; @@ -219,7 +221,8 @@ bool perf_evlist__can_select_event(struct perf_evlist *evlist, const char *str) cpu = evlist->cpus->map[0]; } - fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, 0); + fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, + perf_event_open_cloexec_flag()); if (fd >= 0) { close(fd); ret = true; -- cgit v1.1 From fb578179a54553716afb0666a4d8c194acd313ba Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 14 Jul 2014 08:50:03 +0200 Subject: perf tests: Update attr test with PERF_FLAG_FD_CLOEXEC flag We not support PERF_FLAG_FD_CLOEXEC flag, updating attr tests accordingly. Cc: Arnaldo Carvalho de Melo Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-2mgzn5p9u74drmdmxo5unar0@git.kernel.org Signed-off-by: Jiri Olsa --- tools/perf/tests/attr/base-record | 3 ++- tools/perf/tests/attr/base-stat | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/tests/attr/base-record b/tools/perf/tests/attr/base-record index e9bd639..f710b92 100644 --- a/tools/perf/tests/attr/base-record +++ b/tools/perf/tests/attr/base-record @@ -1,7 +1,8 @@ [event] fd=1 group_fd=-1 -flags=0 +# 0 or PERF_FLAG_FD_CLOEXEC flag +flags=0|8 cpu=* type=0|1 size=96 diff --git a/tools/perf/tests/attr/base-stat b/tools/perf/tests/attr/base-stat index 91cd48b..dc3ada2 100644 --- a/tools/perf/tests/attr/base-stat +++ b/tools/perf/tests/attr/base-stat @@ -1,7 +1,8 @@ [event] fd=1 group_fd=-1 -flags=0 +# 0 or PERF_FLAG_FD_CLOEXEC flag +flags=0|8 cpu=* type=0 size=96 -- cgit v1.1 From a6f6ae99f12e8154d5cde6e8366fc228f68d6ae5 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Thu, 17 Jul 2014 11:43:09 +0300 Subject: perf tools: Fix incorrect fd error comparison Zero is a valid fd. Error comparison should check for negative fd. Signed-off-by: Adrian Hunter Acked-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1405586590-13657-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index fc006fe..28cf747 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -216,7 +216,7 @@ static int open_dso(struct dso *dso, struct machine *machine) { int fd = __open_dso(dso, machine); - if (fd > 0) { + if (fd >= 0) { dso__list_add(dso); /* * Check if we crossed the allowed number -- cgit v1.1 From deaff8b659cf4d34181c087b8cdf74f1eb17b02b Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:10 +0300 Subject: perf tools: Fix jump label always changing during tracing Intel PT decoding walks the object code to reconstruct the trace. A jump label change during tracing causes decoding errors. The "Enable close-on-exec flag on perf file descriptor" patch caused there to be always a jump label change. It was found that using a per-cpu context instead of a per-thread context for the probe of the close-on-exec feature, made the problem go away. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/cloexec.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/cloexec.c b/tools/perf/util/cloexec.c index c5d05ec..6a37be5 100644 --- a/tools/perf/util/cloexec.c +++ b/tools/perf/util/cloexec.c @@ -1,3 +1,4 @@ +#include #include "util.h" #include "../perf.h" #include "cloexec.h" @@ -14,9 +15,13 @@ static int perf_flag_probe(void) }; int fd; int err; + int cpu = sched_getcpu(); + + if (cpu < 0) + cpu = 0; /* check cloexec flag */ - fd = sys_perf_event_open(&attr, 0, -1, -1, + fd = sys_perf_event_open(&attr, -1, cpu, -1, PERF_FLAG_FD_CLOEXEC); err = errno; @@ -30,7 +35,7 @@ static int perf_flag_probe(void) err, strerror(err)); /* not supported, confirm error related to PERF_FLAG_FD_CLOEXEC */ - fd = sys_perf_event_open(&attr, 0, -1, -1, 0); + fd = sys_perf_event_open(&attr, -1, cpu, -1, 0); err = errno; if (WARN_ONCE(fd < 0, -- cgit v1.1 From 8066be5fe755e95ee6a858bbcf4e6b9e933e1866 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:15 +0300 Subject: perf script: Improve srcline display for BTS Change the order of the output to put the srcline last. It puts the branch 'from address' and 'to address' on the same line, which is how it would be without the source line reference. So it makes it consistent and much easier to read. E.g. old format: 4028fc main+0x2c (/bin/ls) /build/buildd/coreutils-8.20/src/ls.c:1269 => 40d8a0 set_program_name+0x0 (/bin/ls) new format: 4028fc main+0x2c (/bin/ls) => 40d8a0 set_program_name+0x0 (/bin/ls) /build/buildd/coreutils-8.20/src/ls.c:1269 Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-7-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 9e9c91f..333b15e 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -427,15 +427,22 @@ static void print_sample_bts(union perf_event *event, struct addr_location *al) { struct perf_event_attr *attr = &evsel->attr; + bool print_srcline_last = false; /* print branch_from information */ if (PRINT_FIELD(IP)) { - if (!symbol_conf.use_callchain) - printf(" "); - else + unsigned int print_opts = output[attr->type].print_ip_opts; + + if (symbol_conf.use_callchain && sample->callchain) { printf("\n"); - perf_evsel__print_ip(evsel, sample, al, - output[attr->type].print_ip_opts, + } else { + printf(" "); + if (print_opts & PRINT_IP_OPT_SRCLINE) { + print_srcline_last = true; + print_opts &= ~PRINT_IP_OPT_SRCLINE; + } + } + perf_evsel__print_ip(evsel, sample, al, print_opts, PERF_MAX_STACK_DEPTH); } @@ -447,6 +454,9 @@ static void print_sample_bts(union perf_event *event, !output[attr->type].user_set)) print_sample_addr(event, sample, al->machine, thread, attr); + if (print_srcline_last) + map__fprintf_srcline(al->map, al->addr, "\n ", stdout); + printf("\n"); } -- cgit v1.1 From 578bea40058b68f83658827c2b68a596b08419fc Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:16 +0300 Subject: perf script: Do not print dangling '=>' for BTS When using: perf record -e branches:u -c1 usleep 1 perf script -f ip lines are displayed like: ffffffff813b23d5 => Change so that the dangling '=>' does not appear. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-8-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 333b15e..582da97 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -446,13 +446,13 @@ static void print_sample_bts(union perf_event *event, PERF_MAX_STACK_DEPTH); } - printf(" => "); - /* print branch_to information */ if (PRINT_FIELD(ADDR) || ((evsel->attr.sample_type & PERF_SAMPLE_ADDR) && - !output[attr->type].user_set)) + !output[attr->type].user_set)) { + printf(" => "); print_sample_addr(event, sample, al->machine, thread, attr); + } if (print_srcline_last) map__fprintf_srcline(al->map, al->addr, "\n ", stdout); -- cgit v1.1 From c27697d6dee02ef2389b6701c792a075bc9873dc Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:18 +0300 Subject: perf tools: Record whether a dso has data Add 'data.status' to record whether a dso has data (i.e. an object file). This is used to avoid repeatedly creating the file name and attempting to open a file that is not present. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-10-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 23 +++++++++++++++-------- tools/perf/util/dso.h | 7 +++++++ 2 files changed, 22 insertions(+), 8 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 28cf747..8827db3 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -331,26 +331,32 @@ int dso__data_fd(struct dso *dso, struct machine *machine) }; int i = 0; + if (dso->data.status == DSO_DATA_STATUS_ERROR) + return -1; + if (dso->data.fd >= 0) - return dso->data.fd; + goto out; if (dso->binary_type != DSO_BINARY_TYPE__NOT_FOUND) { dso->data.fd = open_dso(dso, machine); - return dso->data.fd; + goto out; } do { - int fd; - dso->binary_type = binary_type_data[i++]; - fd = open_dso(dso, machine); - if (fd >= 0) - return dso->data.fd = fd; + dso->data.fd = open_dso(dso, machine); + if (dso->data.fd >= 0) + goto out; } while (dso->binary_type != DSO_BINARY_TYPE__NOT_FOUND); +out: + if (dso->data.fd >= 0) + dso->data.status = DSO_DATA_STATUS_OK; + else + dso->data.status = DSO_DATA_STATUS_ERROR; - return -EINVAL; + return dso->data.fd; } static void @@ -701,6 +707,7 @@ struct dso *dso__new(const char *name) dso->symbols[i] = dso->symbol_names[i] = RB_ROOT; dso->data.cache = RB_ROOT; dso->data.fd = -1; + dso->data.status = DSO_DATA_STATUS_UNKNOWN; dso->symtab_type = DSO_BINARY_TYPE__NOT_FOUND; dso->binary_type = DSO_BINARY_TYPE__NOT_FOUND; dso->is_64_bit = (sizeof(void *) == 8); diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index c239e86..aeb7bcb 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -40,6 +40,12 @@ enum dso_swap_type { DSO_SWAP__YES, }; +enum dso_data_status { + DSO_DATA_STATUS_ERROR = -1, + DSO_DATA_STATUS_UNKNOWN = 0, + DSO_DATA_STATUS_OK = 1, +}; + #define DSO__SWAP(dso, type, val) \ ({ \ type ____r = val; \ @@ -104,6 +110,7 @@ struct dso { struct { struct rb_root cache; int fd; + int status; size_t file_size; struct list_head open_entry; } data; -- cgit v1.1 From 288be943b5024729cd6809b61b62f727960178f3 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:19 +0300 Subject: perf tools: Add dso__data_status_seen() Add a function to track whether a caller has seen the data status of a dso. This is needed to enable callers to report the error exactly once only per dso. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-11-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 12 ++++++++++++ tools/perf/util/dso.h | 6 ++++++ 2 files changed, 18 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 8827db3..84e4a05 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -359,6 +359,18 @@ out: return dso->data.fd; } +bool dso__data_status_seen(struct dso *dso, enum dso_data_status_seen by) +{ + u32 flag = 1 << by; + + if (dso->data.status_seen & flag) + return true; + + dso->data.status_seen |= flag; + + return false; +} + static void dso_cache__free(struct rb_root *root) { diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index aeb7bcb..c805f60 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -46,6 +46,10 @@ enum dso_data_status { DSO_DATA_STATUS_OK = 1, }; +enum dso_data_status_seen { + DSO_DATA_STATUS_SEEN_ITRACE, +}; + #define DSO__SWAP(dso, type, val) \ ({ \ type ____r = val; \ @@ -111,6 +115,7 @@ struct dso { struct rb_root cache; int fd; int status; + u32 status_seen; size_t file_size; struct list_head open_entry; } data; @@ -203,6 +208,7 @@ ssize_t dso__data_read_offset(struct dso *dso, struct machine *machine, ssize_t dso__data_read_addr(struct dso *dso, struct map *map, struct machine *machine, u64 addr, u8 *data, ssize_t size); +bool dso__data_status_seen(struct dso *dso, enum dso_data_status_seen by); struct map *dso__new_map(const char *name); struct dso *dso__kernel_findnew(struct machine *machine, const char *name, -- cgit v1.1 From 1f625b0b3dd641b74ba21640a1fea19a3dd893a2 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:23 +0300 Subject: perf tools: Add dsos__hit_all() Add ability to mark all dsos as hit. This is needed in the case of Instruction Tracing. It takes so long to decode an Instruction Trace that it is not worth doing just to determine which dsos are hit. A later patch takes this into use. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-15-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/header.c | 41 +++++++++++++++++++++++++++++++++++++++++ tools/perf/util/header.h | 2 ++ 2 files changed, 43 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 893f8e2..a588a3e 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -200,6 +200,47 @@ static int write_buildid(const char *name, size_t name_len, u8 *build_id, return write_padded(fd, name, name_len + 1, len); } +static int __dsos__hit_all(struct list_head *head) +{ + struct dso *pos; + + list_for_each_entry(pos, head, node) + pos->hit = true; + + return 0; +} + +static int machine__hit_all_dsos(struct machine *machine) +{ + int err; + + err = __dsos__hit_all(&machine->kernel_dsos); + if (err) + return err; + + return __dsos__hit_all(&machine->user_dsos); +} + +int dsos__hit_all(struct perf_session *session) +{ + struct rb_node *nd; + int err; + + err = machine__hit_all_dsos(&session->machines.host); + if (err) + return err; + + for (nd = rb_first(&session->machines.guests); nd; nd = rb_next(nd)) { + struct machine *pos = rb_entry(nd, struct machine, rb_node); + + err = machine__hit_all_dsos(pos); + if (err) + return err; + } + + return 0; +} + static int __dsos__write_buildid_table(struct list_head *head, struct machine *machine, pid_t pid, u16 misc, int fd) diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index d08cfe4..8f5cbae 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -151,6 +151,8 @@ int perf_event__process_build_id(struct perf_tool *tool, struct perf_session *session); bool is_perf_magic(u64 magic); +int dsos__hit_all(struct perf_session *session); + /* * arch specific callback */ -- cgit v1.1 From bf49c35f630452d85c9cd7205a72df841e8d99b9 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:24 +0300 Subject: perf tools: Add cpu to struct thread Tools may wish to track on which cpu a thread is running. Add 'cpu' to struct thread for that purpose. This will be used to determine the cpu when decoding a per-thread Instruction Trace. E.g: Intel PT decoding uses sched_switch events to determine which task is running on which cpu. The Intel PT data comes straight from the hardware which doesn't know about linux threads. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-16-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/thread.c | 1 + tools/perf/util/thread.h | 1 + 2 files changed, 2 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 9692c06..12c7a25 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -34,6 +34,7 @@ struct thread *thread__new(pid_t pid, pid_t tid) thread->pid_ = pid; thread->tid = tid; thread->ppid = -1; + thread->cpu = -1; INIT_LIST_HEAD(&thread->comm_list); comm_str = malloc(32); diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index 3c0c272..716b772 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -17,6 +17,7 @@ struct thread { pid_t pid_; /* Not all tools update this */ pid_t tid; pid_t ppid; + int cpu; char shortname[3]; bool comm_set; bool dead; /* if set thread has exited */ -- cgit v1.1 From b9d266baac0429f70df3f9cf751b045730d612e3 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:25 +0300 Subject: perf machine: Add ability to record the current tid for each cpu Add an array to struct machine to store the current tid running on each cpu. Add machine functions to get / set the tid for a cpu. This will be used to determine the tid when decoding a per-cpu Instruction Trace. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-17-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/machine.h | 5 +++++ 2 files changed, 51 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 93c8b6f..cfc6910 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -45,6 +45,8 @@ int machine__init(struct machine *machine, const char *root_dir, pid_t pid) thread__set_comm(thread, comm, 0); } + machine->current_tid = NULL; + return 0; } @@ -104,6 +106,7 @@ void machine__exit(struct machine *machine) dsos__delete(&machine->user_dsos); dsos__delete(&machine->kernel_dsos); zfree(&machine->root_dir); + zfree(&machine->current_tid); } void machine__delete(struct machine *machine) @@ -1481,3 +1484,46 @@ int __machine__synthesize_threads(struct machine *machine, struct perf_tool *too /* command specified */ return 0; } + +pid_t machine__get_current_tid(struct machine *machine, int cpu) +{ + if (cpu < 0 || cpu >= MAX_NR_CPUS || !machine->current_tid) + return -1; + + return machine->current_tid[cpu]; +} + +int machine__set_current_tid(struct machine *machine, int cpu, pid_t pid, + pid_t tid) +{ + struct thread *thread; + + if (cpu < 0) + return -EINVAL; + + if (!machine->current_tid) { + int i; + + machine->current_tid = calloc(MAX_NR_CPUS, sizeof(pid_t)); + if (!machine->current_tid) + return -ENOMEM; + for (i = 0; i < MAX_NR_CPUS; i++) + machine->current_tid[i] = -1; + } + + if (cpu >= MAX_NR_CPUS) { + pr_err("Requested CPU %d too large. ", cpu); + pr_err("Consider raising MAX_NR_CPUS\n"); + return -EINVAL; + } + + machine->current_tid[cpu] = tid; + + thread = machine__findnew_thread(machine, pid, tid); + if (!thread) + return -ENOMEM; + + thread->cpu = cpu; + + return 0; +} diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h index c8c74a1..8771d0c 100644 --- a/tools/perf/util/machine.h +++ b/tools/perf/util/machine.h @@ -33,6 +33,7 @@ struct machine { struct map_groups kmaps; struct map *vmlinux_maps[MAP__NR_TYPES]; symbol_filter_t symbol_filter; + pid_t *current_tid; }; static inline @@ -191,4 +192,8 @@ int machine__synthesize_threads(struct machine *machine, struct target *target, perf_event__process, data_mmap); } +pid_t machine__get_current_tid(struct machine *machine, int cpu); +int machine__set_current_tid(struct machine *machine, int cpu, pid_t pid, + pid_t tid); + #endif /* __PERF_MACHINE_H */ -- cgit v1.1 From a6a69db4b686e51045771945669bba6578af67c1 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:32 +0300 Subject: perf tools: Move rdtsc() function Move the rdtsc() function so it can be reusued. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-24-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/x86/util/tsc.c | 9 +++++++++ tools/perf/tests/perf-time-to-tsc.c | 9 --------- tools/perf/util/tsc.c | 5 +++++ tools/perf/util/tsc.h | 1 + 4 files changed, 15 insertions(+), 9 deletions(-) (limited to 'tools') diff --git a/tools/perf/arch/x86/util/tsc.c b/tools/perf/arch/x86/util/tsc.c index 3655f24..fd28684 100644 --- a/tools/perf/arch/x86/util/tsc.c +++ b/tools/perf/arch/x86/util/tsc.c @@ -37,3 +37,12 @@ int perf_read_tsc_conversion(const struct perf_event_mmap_page *pc, return 0; } + +u64 rdtsc(void) +{ + unsigned int low, high; + + asm volatile("rdtsc" : "=a" (low), "=d" (high)); + + return low | ((u64)high) << 32; +} diff --git a/tools/perf/tests/perf-time-to-tsc.c b/tools/perf/tests/perf-time-to-tsc.c index 0372f6e..f238442 100644 --- a/tools/perf/tests/perf-time-to-tsc.c +++ b/tools/perf/tests/perf-time-to-tsc.c @@ -25,15 +25,6 @@ } \ } -static u64 rdtsc(void) -{ - unsigned int low, high; - - asm volatile("rdtsc" : "=a" (low), "=d" (high)); - - return low | ((u64)high) << 32; -} - /** * test__perf_time_to_tsc - test converting perf time to TSC. * diff --git a/tools/perf/util/tsc.c b/tools/perf/util/tsc.c index ef4749836..4d4210d 100644 --- a/tools/perf/util/tsc.c +++ b/tools/perf/util/tsc.c @@ -23,3 +23,8 @@ u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc) return tc->time_zero + quot * tc->time_mult + ((rem * tc->time_mult) >> tc->time_shift); } + +u64 __weak rdtsc(void) +{ + return 0; +} diff --git a/tools/perf/util/tsc.h b/tools/perf/util/tsc.h index 4eca848..a8b78f1 100644 --- a/tools/perf/util/tsc.h +++ b/tools/perf/util/tsc.h @@ -7,5 +7,6 @@ u64 perf_time_to_tsc(u64 ns, struct perf_tsc_conversion *tc); u64 tsc_to_perf_time(u64 cyc, struct perf_tsc_conversion *tc); +u64 rdtsc(void); #endif -- cgit v1.1 From 6d363459d7e6e05f0afeb8e36e6de262a4964694 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:35 +0300 Subject: perf tools: Add dso__data_size() Add a function to return the dso data size, for use in estimating the size an instruction cache. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-27-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 22 ++++++++++++++++++++++ tools/perf/util/dso.h | 2 ++ 2 files changed, 24 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index 84e4a05..e657d86 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -544,6 +544,28 @@ static int data_file_size(struct dso *dso) return 0; } +/** + * dso__data_size - Return dso data size + * @dso: dso object + * @machine: machine object + * + * Return: dso data size + */ +off_t dso__data_size(struct dso *dso, struct machine *machine) +{ + int fd; + + fd = dso__data_fd(dso, machine); + if (fd < 0) + return fd; + + if (data_file_size(dso)) + return -1; + + /* For now just estimate dso data size is close to file size */ + return dso->data.file_size; +} + static ssize_t data_read_offset(struct dso *dso, u64 offset, u8 *data, ssize_t size) { diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index c805f60..af1c256 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -166,6 +166,7 @@ int dso__read_binary_type_filename(const struct dso *dso, enum dso_binary_type t * The dso__data_* external interface provides following functions: * dso__data_fd * dso__data_close + * dso__data_size * dso__data_read_offset * dso__data_read_addr * @@ -203,6 +204,7 @@ int dso__read_binary_type_filename(const struct dso *dso, enum dso_binary_type t int dso__data_fd(struct dso *dso, struct machine *machine); void dso__data_close(struct dso *dso); +off_t dso__data_size(struct dso *dso, struct machine *machine); ssize_t dso__data_read_offset(struct dso *dso, struct machine *machine, u64 offset, u8 *data, ssize_t size); ssize_t dso__data_read_addr(struct dso *dso, struct map *map, -- cgit v1.1 From 2a03068c5cfa104768703cbefa2e23a6353f8de5 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:53 +0300 Subject: perf tools: Pass machine to vdso__dso_findnew() This is preparation for removing the global variables used in vdso.c and thereby fixing the lifetime of the VDSO temporary file. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-45-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 4 ++-- tools/perf/util/map.c | 7 ++++--- tools/perf/util/map.h | 2 +- tools/perf/util/vdso.c | 7 ++++--- tools/perf/util/vdso.h | 4 +++- 5 files changed, 14 insertions(+), 10 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index cfc6910..a25f3ee 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1095,7 +1095,7 @@ int machine__process_mmap2_event(struct machine *machine, else type = MAP__FUNCTION; - map = map__new(&machine->user_dsos, event->mmap2.start, + map = map__new(machine, event->mmap2.start, event->mmap2.len, event->mmap2.pgoff, event->mmap2.pid, event->mmap2.maj, event->mmap2.min, event->mmap2.ino, @@ -1145,7 +1145,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event else type = MAP__FUNCTION; - map = map__new(&machine->user_dsos, event->mmap.start, + map = map__new(machine, event->mmap.start, event->mmap.len, event->mmap.pgoff, event->mmap.pid, 0, 0, 0, 0, 0, 0, event->mmap.filename, diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 845f627..dffc8dc 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -13,6 +13,7 @@ #include "build-id.h" #include "util.h" #include "debug.h" +#include "machine.h" #include const char *map_type__name[MAP__NR_TYPES] = { @@ -137,7 +138,7 @@ void map__init(struct map *map, enum map_type type, map->erange_warned = false; } -struct map *map__new(struct list_head *dsos__list, u64 start, u64 len, +struct map *map__new(struct machine *machine, u64 start, u64 len, u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino, u64 ino_gen, u32 prot, u32 flags, char *filename, enum map_type type) @@ -173,9 +174,9 @@ struct map *map__new(struct list_head *dsos__list, u64 start, u64 len, if (vdso) { pgoff = 0; - dso = vdso__dso_findnew(dsos__list); + dso = vdso__dso_findnew(machine); } else - dso = __dsos__findnew(dsos__list, filename); + dso = __dsos__findnew(&machine->user_dsos, filename); if (dso == NULL) goto out_delete; diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 22d13a2..a95e677 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -119,7 +119,7 @@ typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym); void map__init(struct map *map, enum map_type type, u64 start, u64 end, u64 pgoff, struct dso *dso); -struct map *map__new(struct list_head *dsos__list, u64 start, u64 len, +struct map *map__new(struct machine *machine, u64 start, u64 len, u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino, u64 ino_gen, u32 prot, u32 flags, char *filename, enum map_type type); diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index 2905824..da5ba4d 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -11,6 +11,7 @@ #include "vdso.h" #include "util.h" #include "symbol.h" +#include "machine.h" #include "linux/string.h" #include "debug.h" @@ -90,9 +91,9 @@ void vdso__exit(void) unlink(vdso_file); } -struct dso *vdso__dso_findnew(struct list_head *head) +struct dso *vdso__dso_findnew(struct machine *machine) { - struct dso *dso = dsos__find(head, VDSO__MAP_NAME, true); + struct dso *dso = dsos__find(&machine->user_dsos, VDSO__MAP_NAME, true); if (!dso) { char *file; @@ -103,7 +104,7 @@ struct dso *vdso__dso_findnew(struct list_head *head) dso = dso__new(VDSO__MAP_NAME); if (dso != NULL) { - dsos__add(head, dso); + dsos__add(&machine->user_dsos, dso); dso__set_long_name(dso, file, false); } } diff --git a/tools/perf/util/vdso.h b/tools/perf/util/vdso.h index 0f76e7c..9ab0738 100644 --- a/tools/perf/util/vdso.h +++ b/tools/perf/util/vdso.h @@ -12,7 +12,9 @@ static inline bool is_vdso_map(const char *filename) return !strcmp(filename, VDSO__MAP_NAME); } -struct dso *vdso__dso_findnew(struct list_head *head); +struct machine; + +struct dso *vdso__dso_findnew(struct machine *machine); void vdso__exit(void); #endif /* __PERF_VDSO__ */ -- cgit v1.1 From 6f917c70998d22bda6d06cb168665068c6066cd4 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 23 Jul 2014 22:19:57 +0300 Subject: perf session: Add ability to 'skip' a non-piped event stream A piped event stream may contain arbitary sized tracepoint information following a PERF_RECORD_HEADER_TRACING_DATA event. The position in the stream has to be 'skipped' to match the start of the next event. Provide the same ability to a non-piped event stream to allow for Instruction Trace data that may also be in a non-piped event stream. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406143198-20732-1-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index eac14ce..f4399b2 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1284,6 +1284,7 @@ int __perf_session__process_events(struct perf_session *session, union perf_event *event; uint32_t size; struct ui_progress prog; + int skip; perf_tool__fill_defaults(tool); @@ -1344,7 +1345,8 @@ more: size = event->header.size; if (size < sizeof(struct perf_event_header) || - perf_session__process_event(session, event, tool, file_pos) < 0) { + (skip = perf_session__process_event(session, event, tool, file_pos)) + < 0) { pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n", file_offset + head, event->header.size, event->header.type); @@ -1352,6 +1354,9 @@ more: goto out_err; } + if (skip) + size += skip; + head += size; file_pos += size; -- cgit v1.1 From d5652d865ea734a13a16cf563937291a84d4364d Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 23 Jul 2014 22:19:58 +0300 Subject: perf session: Add ability to skip 4GiB or more A session can be made to skip portions of the input file. Do not limit that size to 32-bits. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406143198-20732-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index f4399b2..d3da105 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -994,8 +994,10 @@ static int perf_session_deliver_event(struct perf_session *session, } } -static int perf_session__process_user_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool, u64 file_offset) +static s64 perf_session__process_user_event(struct perf_session *session, + union perf_event *event, + struct perf_tool *tool, + u64 file_offset) { int fd = perf_data_file__fd(session->file); int err; @@ -1037,7 +1039,7 @@ static void event_swap(union perf_event *event, bool sample_id_all) swap(event, sample_id_all); } -static int perf_session__process_event(struct perf_session *session, +static s64 perf_session__process_event(struct perf_session *session, union perf_event *event, struct perf_tool *tool, u64 file_offset) @@ -1148,7 +1150,7 @@ static int __perf_session__process_pipe_events(struct perf_session *session, union perf_event *event; uint32_t size, cur_size = 0; void *buf = NULL; - int skip = 0; + s64 skip = 0; u64 head; ssize_t err; void *p; @@ -1277,14 +1279,13 @@ int __perf_session__process_events(struct perf_session *session, u64 file_size, struct perf_tool *tool) { int fd = perf_data_file__fd(session->file); - u64 head, page_offset, file_offset, file_pos; + u64 head, page_offset, file_offset, file_pos, size; int err, mmap_prot, mmap_flags, map_idx = 0; size_t mmap_size; char *buf, *mmaps[NUM_MMAPS]; union perf_event *event; - uint32_t size; struct ui_progress prog; - int skip; + s64 skip; perf_tool__fill_defaults(tool); -- cgit v1.1 From 30f4f815a45d0b148d17afb0d5a5575ae2ba4309 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:54 +0300 Subject: perf tools: Group VDSO global variables into a structure This is preparation for removing the global variables used in vdso.c and thereby fixing the lifetime of the VDSO temporary file. Also allowance is made for the later addition of support for compat VDSOs. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-46-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/vdso.c | 44 ++++++++++++++++++++++++++++++++------------ 1 file changed, 32 insertions(+), 12 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index da5ba4d..75245f0 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -15,8 +15,27 @@ #include "linux/string.h" #include "debug.h" -static bool vdso_found; -static char vdso_file[] = "/tmp/perf-vdso.so-XXXXXX"; +#define VDSO__TEMP_FILE_NAME "/tmp/perf-vdso.so-XXXXXX" + +struct vdso_file { + bool found; + bool error; + char temp_file_name[sizeof(VDSO__TEMP_FILE_NAME)]; + const char *dso_name; +}; + +struct vdso_info { + struct vdso_file vdso; +}; + +static struct vdso_info vdso_info_ = { + .vdso = { + .temp_file_name = VDSO__TEMP_FILE_NAME, + .dso_name = VDSO__MAP_NAME, + }, +}; + +static struct vdso_info *vdso_info = &vdso_info_; static int find_vdso_map(void **start, void **end) { @@ -49,7 +68,7 @@ static int find_vdso_map(void **start, void **end) return !found; } -static char *get_file(void) +static char *get_file(struct vdso_file *vdso_file) { char *vdso = NULL; char *buf = NULL; @@ -57,10 +76,10 @@ static char *get_file(void) size_t size; int fd; - if (vdso_found) - return vdso_file; + if (vdso_file->found) + return vdso_file->temp_file_name; - if (find_vdso_map(&start, &end)) + if (vdso_file->error || find_vdso_map(&start, &end)) return NULL; size = end - start; @@ -69,26 +88,27 @@ static char *get_file(void) if (!buf) return NULL; - fd = mkstemp(vdso_file); + fd = mkstemp(vdso_file->temp_file_name); if (fd < 0) goto out; if (size == (size_t) write(fd, buf, size)) - vdso = vdso_file; + vdso = vdso_file->temp_file_name; close(fd); out: free(buf); - vdso_found = (vdso != NULL); + vdso_file->found = (vdso != NULL); + vdso_file->error = !vdso_file->found; return vdso; } void vdso__exit(void) { - if (vdso_found) - unlink(vdso_file); + if (vdso_info->vdso.found) + unlink(vdso_info->vdso.temp_file_name); } struct dso *vdso__dso_findnew(struct machine *machine) @@ -98,7 +118,7 @@ struct dso *vdso__dso_findnew(struct machine *machine) if (!dso) { char *file; - file = get_file(); + file = get_file(&vdso_info->vdso); if (!file) return NULL; -- cgit v1.1 From d027b64001b21328cc92d35c6444e1a7a926ea76 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 23 Jul 2014 14:23:00 +0300 Subject: perf machine: Fix the lifetime of the VDSO temporary file The VDSO temporary file is unlinked when a session is deleted. That precludes the possibilities that there is no session or there is more than one session. Correctly the vdso belongs to the machine so put the information on 'struct machine' and get rid of the global variables. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/53CF9B14.7040408@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 4 ++++ tools/perf/util/machine.h | 3 +++ tools/perf/util/session.c | 2 -- tools/perf/util/vdso.c | 39 +++++++++++++++++++++++++++++---------- tools/perf/util/vdso.h | 2 +- 5 files changed, 37 insertions(+), 13 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index a25f3ee..65269b8 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -8,6 +8,7 @@ #include "sort.h" #include "strlist.h" #include "thread.h" +#include "vdso.h" #include #include #include "unwind.h" @@ -23,6 +24,8 @@ int machine__init(struct machine *machine, const char *root_dir, pid_t pid) INIT_LIST_HEAD(&machine->dead_threads); machine->last_match = NULL; + machine->vdso_info = NULL; + machine->kmaps.machine = machine; machine->pid = pid; @@ -105,6 +108,7 @@ void machine__exit(struct machine *machine) map_groups__exit(&machine->kmaps); dsos__delete(&machine->user_dsos); dsos__delete(&machine->kernel_dsos); + vdso__exit(machine); zfree(&machine->root_dir); zfree(&machine->current_tid); } diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h index 8771d0c..b972824 100644 --- a/tools/perf/util/machine.h +++ b/tools/perf/util/machine.h @@ -20,6 +20,8 @@ union perf_event; extern const char *ref_reloc_sym_names[]; +struct vdso_info; + struct machine { struct rb_node rb_node; pid_t pid; @@ -28,6 +30,7 @@ struct machine { struct rb_root threads; struct list_head dead_threads; struct thread *last_match; + struct vdso_info *vdso_info; struct list_head user_dsos; struct list_head kernel_dsos; struct map_groups kmaps; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index d3da105..fab5838 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -14,7 +14,6 @@ #include "util.h" #include "cpumap.h" #include "perf_regs.h" -#include "vdso.h" static int perf_session__open(struct perf_session *session) { @@ -156,7 +155,6 @@ void perf_session__delete(struct perf_session *session) if (session->file) perf_data_file__close(session->file); free(session); - vdso__exit(); } static int process_event_synth_tracing_data_stub(struct perf_tool *tool diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index 75245f0..fdaccaf 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -28,14 +28,17 @@ struct vdso_info { struct vdso_file vdso; }; -static struct vdso_info vdso_info_ = { - .vdso = { - .temp_file_name = VDSO__TEMP_FILE_NAME, - .dso_name = VDSO__MAP_NAME, - }, -}; - -static struct vdso_info *vdso_info = &vdso_info_; +static struct vdso_info *vdso_info__new(void) +{ + static const struct vdso_info vdso_info_init = { + .vdso = { + .temp_file_name = VDSO__TEMP_FILE_NAME, + .dso_name = VDSO__MAP_NAME, + }, + }; + + return memdup(&vdso_info_init, sizeof(vdso_info_init)); +} static int find_vdso_map(void **start, void **end) { @@ -105,16 +108,32 @@ static char *get_file(struct vdso_file *vdso_file) return vdso; } -void vdso__exit(void) +void vdso__exit(struct machine *machine) { + struct vdso_info *vdso_info = machine->vdso_info; + + if (!vdso_info) + return; + if (vdso_info->vdso.found) unlink(vdso_info->vdso.temp_file_name); + + zfree(&machine->vdso_info); } struct dso *vdso__dso_findnew(struct machine *machine) { - struct dso *dso = dsos__find(&machine->user_dsos, VDSO__MAP_NAME, true); + struct vdso_info *vdso_info; + struct dso *dso; + + if (!machine->vdso_info) + machine->vdso_info = vdso_info__new(); + + vdso_info = machine->vdso_info; + if (!vdso_info) + return NULL; + dso = dsos__find(&machine->user_dsos, VDSO__MAP_NAME, true); if (!dso) { char *file; diff --git a/tools/perf/util/vdso.h b/tools/perf/util/vdso.h index 9ab0738..7cf1576 100644 --- a/tools/perf/util/vdso.h +++ b/tools/perf/util/vdso.h @@ -15,6 +15,6 @@ static inline bool is_vdso_map(const char *filename) struct machine; struct dso *vdso__dso_findnew(struct machine *machine); -void vdso__exit(void); +void vdso__exit(struct machine *machine); #endif /* __PERF_VDSO__ */ -- cgit v1.1 From 4f71f2a0abe8b2a76669a4608a62f5020a64205a Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:56 +0300 Subject: perf tools: Add vdso__new() This is preparation for adding support for compat VDSOs. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-48-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/vdso.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index fdaccaf..946d927 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -121,6 +121,20 @@ void vdso__exit(struct machine *machine) zfree(&machine->vdso_info); } +static struct dso *vdso__new(struct machine *machine, const char *short_name, + const char *long_name) +{ + struct dso *dso; + + dso = dso__new(short_name); + if (dso != NULL) { + dsos__add(&machine->user_dsos, dso); + dso__set_long_name(dso, long_name, false); + } + + return dso; +} + struct dso *vdso__dso_findnew(struct machine *machine) { struct vdso_info *vdso_info; @@ -141,11 +155,7 @@ struct dso *vdso__dso_findnew(struct machine *machine) if (!file) return NULL; - dso = dso__new(VDSO__MAP_NAME); - if (dso != NULL) { - dsos__add(&machine->user_dsos, dso); - dso__set_long_name(dso, file, false); - } + dso = vdso__new(machine, VDSO__MAP_NAME, file); } return dso; -- cgit v1.1 From 51682dc744c3db89e515ac47a4c1f7003fd81d20 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:57 +0300 Subject: perf tools: Separate the VDSO map name from the VDSO dso name This is in preparation for supporting 32-bit compatibility VDSOs. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-49-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/header.c | 10 +++++----- tools/perf/util/symbol-elf.c | 2 +- tools/perf/util/vdso.c | 11 ++++++++--- tools/perf/util/vdso.h | 6 ++++++ 4 files changed, 20 insertions(+), 9 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index a588a3e..158c787 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -256,9 +256,9 @@ static int __dsos__write_buildid_table(struct list_head *head, if (!pos->hit) continue; - if (is_vdso_map(pos->short_name)) { - name = (char *) VDSO__MAP_NAME; - name_len = sizeof(VDSO__MAP_NAME) + 1; + if (dso__is_vdso(pos)) { + name = pos->short_name; + name_len = pos->short_name_len + 1; } else if (dso__is_kcore(pos)) { machine__mmap_name(machine, nm, sizeof(nm)); name = nm; @@ -339,7 +339,7 @@ int build_id_cache__add_s(const char *sbuild_id, const char *debugdir, len = scnprintf(filename, size, "%s%s%s", debugdir, slash ? "/" : "", - is_vdso ? VDSO__MAP_NAME : realname); + is_vdso ? DSO__NAME_VDSO : realname); if (mkdir_p(filename, 0755)) goto out_free; @@ -427,7 +427,7 @@ static int dso__cache_build_id(struct dso *dso, struct machine *machine, const char *debugdir) { bool is_kallsyms = dso->kernel && dso->long_name[0] != '/'; - bool is_vdso = is_vdso_map(dso->short_name); + bool is_vdso = dso__is_vdso(dso); const char *name = dso->long_name; char nm[PATH_MAX]; diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index cef8f42..61b9cd4 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -622,7 +622,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, GElf_Shdr shdr; ss->adjust_symbols = (ehdr.e_type == ET_EXEC || ehdr.e_type == ET_REL || - is_vdso_map(dso->short_name) || + dso__is_vdso(dso) || elf_section_by_name(elf, &ehdr, &shdr, ".gnu.prelink_undo", NULL) != NULL); diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index 946d927..a9300f8 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -33,7 +33,7 @@ static struct vdso_info *vdso_info__new(void) static const struct vdso_info vdso_info_init = { .vdso = { .temp_file_name = VDSO__TEMP_FILE_NAME, - .dso_name = VDSO__MAP_NAME, + .dso_name = DSO__NAME_VDSO, }, }; @@ -147,7 +147,7 @@ struct dso *vdso__dso_findnew(struct machine *machine) if (!vdso_info) return NULL; - dso = dsos__find(&machine->user_dsos, VDSO__MAP_NAME, true); + dso = dsos__find(&machine->user_dsos, DSO__NAME_VDSO, true); if (!dso) { char *file; @@ -155,8 +155,13 @@ struct dso *vdso__dso_findnew(struct machine *machine) if (!file) return NULL; - dso = vdso__new(machine, VDSO__MAP_NAME, file); + dso = vdso__new(machine, DSO__NAME_VDSO, file); } return dso; } + +bool dso__is_vdso(struct dso *dso) +{ + return !strcmp(dso->short_name, DSO__NAME_VDSO); +} diff --git a/tools/perf/util/vdso.h b/tools/perf/util/vdso.h index 7cf1576..be3eb43 100644 --- a/tools/perf/util/vdso.h +++ b/tools/perf/util/vdso.h @@ -7,11 +7,17 @@ #define VDSO__MAP_NAME "[vdso]" +#define DSO__NAME_VDSO "[vdso]" + static inline bool is_vdso_map(const char *filename) { return !strcmp(filename, VDSO__MAP_NAME); } +struct dso; + +bool dso__is_vdso(struct dso *dso); + struct machine; struct dso *vdso__dso_findnew(struct machine *machine); -- cgit v1.1 From 2b5b8bb27b9ec899cfba686dabec113ea0d4cbda Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:59 +0300 Subject: perf tools: Add dso__type() dso__type() determines wheather a dso is 32-bit, x32 (32-bit with 64-bit registers) or 64-bit. dso__type() will be used to determine the VDSO a program maps. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-51-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/dso.c | 11 +++++++++++ tools/perf/util/dso.h | 10 ++++++++++ tools/perf/util/symbol-elf.c | 33 +++++++++++++++++++++++++++++++++ tools/perf/util/symbol-minimal.c | 21 +++++++++++++++++++++ tools/perf/util/symbol.h | 2 ++ 5 files changed, 77 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c index e657d86..90d02c66 100644 --- a/tools/perf/util/dso.c +++ b/tools/perf/util/dso.c @@ -940,3 +940,14 @@ size_t dso__fprintf(struct dso *dso, enum map_type type, FILE *fp) return ret; } + +enum dso_type dso__type(struct dso *dso, struct machine *machine) +{ + int fd; + + fd = dso__data_fd(dso, machine); + if (fd < 0) + return DSO__TYPE_UNKNOWN; + + return dso__type_fd(fd); +} diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h index af1c256..5e463c0 100644 --- a/tools/perf/util/dso.h +++ b/tools/perf/util/dso.h @@ -5,6 +5,7 @@ #include #include #include +#include #include "map.h" #include "build-id.h" @@ -50,6 +51,13 @@ enum dso_data_status_seen { DSO_DATA_STATUS_SEEN_ITRACE, }; +enum dso_type { + DSO__TYPE_UNKNOWN, + DSO__TYPE_64BIT, + DSO__TYPE_32BIT, + DSO__TYPE_X32BIT, +}; + #define DSO__SWAP(dso, type, val) \ ({ \ type ____r = val; \ @@ -245,4 +253,6 @@ static inline bool dso__is_kcore(struct dso *dso) void dso__free_a2l(struct dso *dso); +enum dso_type dso__type(struct dso *dso, struct machine *machine); + #endif /* __PERF_DSO */ diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 61b9cd4..d753499 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -1028,6 +1028,39 @@ int file__read_maps(int fd, bool exe, mapfn_t mapfn, void *data, return err; } +enum dso_type dso__type_fd(int fd) +{ + enum dso_type dso_type = DSO__TYPE_UNKNOWN; + GElf_Ehdr ehdr; + Elf_Kind ek; + Elf *elf; + + elf = elf_begin(fd, PERF_ELF_C_READ_MMAP, NULL); + if (elf == NULL) + goto out; + + ek = elf_kind(elf); + if (ek != ELF_K_ELF) + goto out_end; + + if (gelf_getclass(elf) == ELFCLASS64) { + dso_type = DSO__TYPE_64BIT; + goto out_end; + } + + if (gelf_getehdr(elf, &ehdr) == NULL) + goto out_end; + + if (ehdr.e_machine == EM_X86_64) + dso_type = DSO__TYPE_X32BIT; + else + dso_type = DSO__TYPE_32BIT; +out_end: + elf_end(elf); +out: + return dso_type; +} + static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len) { ssize_t r; diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c index 101f55d..c9541fe 100644 --- a/tools/perf/util/symbol-minimal.c +++ b/tools/perf/util/symbol-minimal.c @@ -305,6 +305,27 @@ static int fd__is_64_bit(int fd) return e_ident[EI_CLASS] == ELFCLASS64; } +enum dso_type dso__type_fd(int fd) +{ + Elf64_Ehdr ehdr; + int ret; + + ret = fd__is_64_bit(fd); + if (ret < 0) + return DSO__TYPE_UNKNOWN; + + if (ret) + return DSO__TYPE_64BIT; + + if (readn(fd, &ehdr, sizeof(ehdr)) != sizeof(ehdr)) + return DSO__TYPE_UNKNOWN; + + if (ehdr.e_machine == EM_X86_64) + return DSO__TYPE_X32BIT; + + return DSO__TYPE_32BIT; +} + int dso__load_sym(struct dso *dso, struct map *map __maybe_unused, struct symsrc *ss, struct symsrc *runtime_ss __maybe_unused, diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index ee2d3cc..e7295e9 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -243,6 +243,8 @@ struct symbol *dso__find_symbol_by_name(struct dso *dso, enum map_type type, struct symbol *dso__first_symbol(struct dso *dso, enum map_type type); struct symbol *dso__next_symbol(struct symbol *sym); +enum dso_type dso__type_fd(int fd); + int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); int modules__parse(const char *filename, void *arg, -- cgit v1.1 From 5835eddab6f162b38e9a6a5447a2c3a128637956 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:18:00 +0300 Subject: perf tools: Add thread parameter to vdso__dso_findnew() The thread will be needed to determine the VDSO type. Reviewed-by: Jiri Olsa Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-52-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 4 ++-- tools/perf/util/map.c | 4 ++-- tools/perf/util/map.h | 3 ++- tools/perf/util/vdso.c | 3 ++- tools/perf/util/vdso.h | 3 ++- 5 files changed, 10 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 65269b8..16bba9f 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1106,7 +1106,7 @@ int machine__process_mmap2_event(struct machine *machine, event->mmap2.ino_generation, event->mmap2.prot, event->mmap2.flags, - event->mmap2.filename, type); + event->mmap2.filename, type, thread); if (map == NULL) goto out_problem; @@ -1153,7 +1153,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event event->mmap.len, event->mmap.pgoff, event->mmap.pid, 0, 0, 0, 0, 0, 0, event->mmap.filename, - type); + type, thread); if (map == NULL) goto out_problem; diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index dffc8dc..31b8905 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -141,7 +141,7 @@ void map__init(struct map *map, enum map_type type, struct map *map__new(struct machine *machine, u64 start, u64 len, u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino, u64 ino_gen, u32 prot, u32 flags, char *filename, - enum map_type type) + enum map_type type, struct thread *thread) { struct map *map = malloc(sizeof(*map)); @@ -174,7 +174,7 @@ struct map *map__new(struct machine *machine, u64 start, u64 len, if (vdso) { pgoff = 0; - dso = vdso__dso_findnew(machine); + dso = vdso__dso_findnew(machine, thread); } else dso = __dsos__findnew(&machine->user_dsos, filename); diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index a95e677..2f83954 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -104,6 +104,7 @@ u64 map__rip_2objdump(struct map *map, u64 rip); u64 map__objdump_2mem(struct map *map, u64 ip); struct symbol; +struct thread; /* map__for_each_symbol - iterate over the symbols in the given map * @@ -122,7 +123,7 @@ void map__init(struct map *map, enum map_type type, struct map *map__new(struct machine *machine, u64 start, u64 len, u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino, u64 ino_gen, u32 prot, u32 flags, - char *filename, enum map_type type); + char *filename, enum map_type type, struct thread *thread); struct map *map__new2(u64 start, struct dso *dso, enum map_type type); void map__delete(struct map *map); struct map *map__clone(struct map *map); diff --git a/tools/perf/util/vdso.c b/tools/perf/util/vdso.c index a9300f8..adca693 100644 --- a/tools/perf/util/vdso.c +++ b/tools/perf/util/vdso.c @@ -135,7 +135,8 @@ static struct dso *vdso__new(struct machine *machine, const char *short_name, return dso; } -struct dso *vdso__dso_findnew(struct machine *machine) +struct dso *vdso__dso_findnew(struct machine *machine, + struct thread *thread __maybe_unused) { struct vdso_info *vdso_info; struct dso *dso; diff --git a/tools/perf/util/vdso.h b/tools/perf/util/vdso.h index be3eb43..af9d692 100644 --- a/tools/perf/util/vdso.h +++ b/tools/perf/util/vdso.h @@ -19,8 +19,9 @@ struct dso; bool dso__is_vdso(struct dso *dso); struct machine; +struct thread; -struct dso *vdso__dso_findnew(struct machine *machine); +struct dso *vdso__dso_findnew(struct machine *machine, struct thread *thread); void vdso__exit(struct machine *machine); #endif /* __PERF_VDSO__ */ -- cgit v1.1 From 0524798cc3d74546e83ebd795e9a5ed44481eaa6 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 23 Jul 2014 18:15:09 -0300 Subject: perf tools: Fix build on gcc 4.4.7 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit [acme@sandy linux]$ gcc --version gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-3) Copyright (C) 2010 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. [acme@sandy linux]$ make O=/tmp/build/perf -C tools/perf install-bin CC /tmp/build/perf/builtin-trace.o builtin-trace.c: In function ‘perf_evlist__add_pgfault’: builtin-trace.c:1997: error: unknown field ‘sample_period’ specified in initializer make[1]: *** [/tmp/build/perf/builtin-trace.o] Error 1 make: *** [install-bin] Error 2 make: Leaving directory `/home/acme/git/linux/tools/perf' [acme@sandy linux]$ make O=/tmp/build/perf -C tools/perf install-bin make O=/tmp/build/perf -C tools/perf install-bin make: Entering directory `/home/acme/git/linux/tools/perf' Cc: Adrian Hunter Cc: David Ahern Cc: Don Zickus 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-qt7h2g5fcf42qiw5hv7mgpjk@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index c4a5a7d..a6c3752 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1994,10 +1994,10 @@ static int perf_evlist__add_pgfault(struct perf_evlist *evlist, struct perf_event_attr attr = { .type = PERF_TYPE_SOFTWARE, .mmap_data = 1, - .sample_period = 1, }; attr.config = config; + attr.sample_period = 1; event_attr_init(&attr); -- cgit v1.1 From c94b93cbca59435dfc0f2a838fea55bd632145d3 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Jul 2014 00:47:18 -0700 Subject: perf powerpc: Include util/util.h and remove stringify macros The stringify macros are defined in tools/perf/util/util.h and don't need to be redfined specfiically for powerpc. Signed-off-by: Sukadev Bhattiprolu Cc: Andi Kleen Cc: Jiri Olsa Cc: Michael Ellerman Link: http://lkml.kernel.org/r/20140724074718.GB18829@us.ibm.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/arch/powerpc/util/header.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/arch/powerpc/util/header.c b/tools/perf/arch/powerpc/util/header.c index 2f7073d..6c1b8a7 100644 --- a/tools/perf/arch/powerpc/util/header.c +++ b/tools/perf/arch/powerpc/util/header.c @@ -5,9 +5,7 @@ #include #include "../../util/header.h" - -#define __stringify_1(x) #x -#define __stringify(x) __stringify_1(x) +#include "../../util/util.h" #define mfspr(rn) ({unsigned long rval; \ asm volatile("mfspr %0," __stringify(rn) \ -- cgit v1.1 From f1dd1460a40894b00bbeacd753025e9251ec11bd Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Sun, 20 Jul 2014 23:55:45 +0200 Subject: perf session: Fix accounting of ordered samples queue Properly account flushed samples within the ordered samples queue. Signed-off-by: Jiri Olsa Acked-by: Adrian Hunter Acked-by: David Ahern Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jean Pihet Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1405893363-21967-2-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index fab5838..88dfef7 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -509,6 +509,7 @@ static int flush_sample_queue(struct perf_session *s, os->last_flush = iter->timestamp; list_del(&iter->list); list_add(&iter->list, &os->sample_cache); + os->nr_samples--; if (show_progress) ui_progress__update(&prog, 1); @@ -521,8 +522,6 @@ static int flush_sample_queue(struct perf_session *s, list_entry(head->prev, struct sample_queue, list); } - os->nr_samples = 0; - return 0; } -- cgit v1.1 From 9b0d2d875d57d85fdfb35ac27f89951520a8b473 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:39 +0300 Subject: perf tools: Expose 'addr' functions so they can be reused Move some functions and functionality related to the use of 'addr' out of builtin-script so they can be reused. The moved functions are: is_bts_event() and sample_addr_correlates_sym() and a new function perf_event__preprocess_sample_addr() is created from bits of print_sample_addr(). perf_event__preprocess_sample_addr() is the equivalent of perf_event__preprocess_sample() but for 'addr' instead of 'ip'. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-31-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 34 +--------------------------------- tools/perf/util/event.c | 42 ++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/event.h | 10 ++++++++++ 3 files changed, 53 insertions(+), 33 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 582da97..f57035b 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -358,27 +358,6 @@ static void print_sample_start(struct perf_sample *sample, } } -static bool is_bts_event(struct perf_event_attr *attr) -{ - return ((attr->type == PERF_TYPE_HARDWARE) && - (attr->config & PERF_COUNT_HW_BRANCH_INSTRUCTIONS) && - (attr->sample_period == 1)); -} - -static bool sample_addr_correlates_sym(struct perf_event_attr *attr) -{ - if ((attr->type == PERF_TYPE_SOFTWARE) && - ((attr->config == PERF_COUNT_SW_PAGE_FAULTS) || - (attr->config == PERF_COUNT_SW_PAGE_FAULTS_MIN) || - (attr->config == PERF_COUNT_SW_PAGE_FAULTS_MAJ))) - return true; - - if (is_bts_event(attr)) - return true; - - return false; -} - static void print_sample_addr(union perf_event *event, struct perf_sample *sample, struct machine *machine, @@ -386,24 +365,13 @@ static void print_sample_addr(union perf_event *event, struct perf_event_attr *attr) { struct addr_location al; - u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; printf("%16" PRIx64, sample->addr); if (!sample_addr_correlates_sym(attr)) return; - thread__find_addr_map(thread, machine, cpumode, MAP__FUNCTION, - sample->addr, &al); - if (!al.map) - thread__find_addr_map(thread, machine, cpumode, MAP__VARIABLE, - sample->addr, &al); - - al.cpu = sample->cpu; - al.sym = NULL; - - if (al.map) - al.sym = map__find_symbol(al.map, al.addr, NULL); + perf_event__preprocess_sample_addr(event, sample, machine, thread, &al); if (PRINT_FIELD(SYM)) { printf(" "); diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 7e0e8ae..1398c83 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -874,3 +874,45 @@ int perf_event__preprocess_sample(const union perf_event *event, return 0; } + +bool is_bts_event(struct perf_event_attr *attr) +{ + return attr->type == PERF_TYPE_HARDWARE && + (attr->config & PERF_COUNT_HW_BRANCH_INSTRUCTIONS) && + attr->sample_period == 1; +} + +bool sample_addr_correlates_sym(struct perf_event_attr *attr) +{ + if (attr->type == PERF_TYPE_SOFTWARE && + (attr->config == PERF_COUNT_SW_PAGE_FAULTS || + attr->config == PERF_COUNT_SW_PAGE_FAULTS_MIN || + attr->config == PERF_COUNT_SW_PAGE_FAULTS_MAJ)) + return true; + + if (is_bts_event(attr)) + return true; + + return false; +} + +void perf_event__preprocess_sample_addr(union perf_event *event, + struct perf_sample *sample, + struct machine *machine, + struct thread *thread, + struct addr_location *al) +{ + u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + + thread__find_addr_map(thread, machine, cpumode, MAP__FUNCTION, + sample->addr, al); + if (!al->map) + thread__find_addr_map(thread, machine, cpumode, MAP__VARIABLE, + sample->addr, al); + + al->cpu = sample->cpu; + al->sym = NULL; + + if (al->map) + al->sym = map__find_symbol(al->map, al->addr, NULL); +} diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index e5dd40a..94d6976 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -288,6 +288,16 @@ int perf_event__preprocess_sample(const union perf_event *event, struct addr_location *al, struct perf_sample *sample); +struct thread; + +bool is_bts_event(struct perf_event_attr *attr); +bool sample_addr_correlates_sym(struct perf_event_attr *attr); +void perf_event__preprocess_sample_addr(union perf_event *event, + struct perf_sample *sample, + struct machine *machine, + struct thread *thread, + struct addr_location *al); + const char *perf_event__name(unsigned int id); size_t perf_event__sample_event_size(const struct perf_sample *sample, u64 type, -- cgit v1.1 From a7a2b8b4ce9e0bfd085c5797d535487594a71882 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 22 Jul 2014 16:17:38 +0300 Subject: perf inject: Add --kallsyms parameter Let perf inject take --kallsyms parameter the same as perf script and perf report do. That is needed for decoding Instruction Trace data using a copy of /proc/kcore for the kernel object because the kallsyms path is used to locate that copy. Signed-off-by: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1406035081-14301-30-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/Documentation/perf-inject.txt | 3 +++ tools/perf/builtin-inject.c | 2 ++ 2 files changed, 5 insertions(+) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-inject.txt b/tools/perf/Documentation/perf-inject.txt index a00a342..dc7442c 100644 --- a/tools/perf/Documentation/perf-inject.txt +++ b/tools/perf/Documentation/perf-inject.txt @@ -41,6 +41,9 @@ OPTIONS tasks slept. sched_switch contains a callchain where a task slept and sched_stat contains a timeslice how long a task slept. +--kallsyms=:: + kallsyms pathname + SEE ALSO -------- linkperf:perf-record[1], linkperf:perf-report[1], linkperf:perf-archive[1] diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index cf6a605..9a02807 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -439,6 +439,8 @@ int cmd_inject(int argc, const char **argv, const char *prefix __maybe_unused) "where and how long tasks slept"), OPT_INCR('v', "verbose", &verbose, "be more verbose (show build ids, etc)"), + OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name, "file", + "kallsyms pathname"), OPT_END() }; const char * const inject_usage[] = { -- cgit v1.1 From 33bf7481971a622a2b8c8aaa0e5e61a6eaeecd71 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 25 Jul 2014 16:56:15 +0200 Subject: perf record: Always force PERF_RECORD_FINISHED_ROUND event The PERF_RECORD_FINISHED_ROUND synthetic record governs queue flushing in reporting, so it needs to be stored for any kind of event. The lack of such periodic flushing made the tools use more memory than needed, as the reordering was being done only after processing all events. This was the case when no tracepoints were in the mix. Forcing the PERF_RECORD_FINISHED_ROUND event to be stored for all event types. Signed-off-by: Jiri Olsa Acked-by: Adrian Hunter Acked-by: David Ahern Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jean Pihet Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1406300177-31805-18-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-record.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 378b85b..4a1a542 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -250,8 +250,7 @@ static int record__mmap_read_all(struct record *rec) } } - if (perf_header__has_feat(&rec->session->header, HEADER_TRACING_DATA)) - rc = record__write(rec, &finished_round_event, sizeof(finished_round_event)); + rc = record__write(rec, &finished_round_event, sizeof(finished_round_event)); out: return rc; -- cgit v1.1 From dcabb507fd3a2b19aed6b4068e2a954f5fd8de45 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 25 Jul 2014 16:56:16 +0200 Subject: perf record: Store PERF_RECORD_FINISHED_ROUND only for nonempty rounds Currently we store PERF_RECORD_FINISHED_ROUND event each time we go throught mmap buffers no matter if it contains any data, which is useless. Forcing the PERF_RECORD_FINISHED_ROUND event to be stored any time we finished the round AND wrote at least one event. Signed-off-by: Jiri Olsa Acked-by: Adrian Hunter Acked-by: David Ahern Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jean Pihet Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1406300177-31805-19-git-send-email-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-record.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 4a1a542..4869050 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -238,6 +238,7 @@ static struct perf_event_header finished_round_event = { static int record__mmap_read_all(struct record *rec) { + u64 bytes_written = rec->bytes_written; int i; int rc = 0; @@ -250,7 +251,12 @@ static int record__mmap_read_all(struct record *rec) } } - rc = record__write(rec, &finished_round_event, sizeof(finished_round_event)); + /* + * Mark the round finished in case we wrote + * at least one event. + */ + if (bytes_written != rec->bytes_written) + rc = record__write(rec, &finished_round_event, sizeof(finished_round_event)); out: return rc; -- cgit v1.1 From 78a1b503c8d4115cc608208cc767f57c91381165 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Fri, 18 Jul 2014 09:11:30 +0200 Subject: perf tools: Fix perf usage string leftover Fixing perf usage string leftover pointed out by Namhyung. Reported-by: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-vkcf53oultknsh3ue9fhin94@git.kernel.org Signed-off-by: Jiri Olsa --- tools/perf/perf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/perf.c b/tools/perf/perf.c index eed3fb2..2282d41 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -18,7 +18,7 @@ #include const char perf_usage_string[] = - "perf [--version] [--debug variable[=VALUE]] [--help] COMMAND [ARGS]"; + "perf [--version] [--help] [OPTIONS] COMMAND [ARGS]"; const char perf_more_info_string[] = "See 'perf help COMMAND' for more information on a specific command."; -- cgit v1.1 From 8578b891ec38c39abd5d6376afd05fce78e2628c Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 28 Jul 2014 15:58:47 +0200 Subject: Revert "perf tools: Fix jump label always changing during tracing" This reverts commit deaff8b659cf4d34181c087b8cdf74f1eb17b02b. This commit makes CLOEXEC feature undetected for normal users, because per-cpu events are priviledged. Reported-by: Peter Zijlstra Cc: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/20140728065844.GK6758@twins.programming.kicks-ass.net Signed-off-by: Jiri Olsa --- tools/perf/util/cloexec.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/cloexec.c b/tools/perf/util/cloexec.c index 6a37be5..c5d05ec 100644 --- a/tools/perf/util/cloexec.c +++ b/tools/perf/util/cloexec.c @@ -1,4 +1,3 @@ -#include #include "util.h" #include "../perf.h" #include "cloexec.h" @@ -15,13 +14,9 @@ static int perf_flag_probe(void) }; int fd; int err; - int cpu = sched_getcpu(); - - if (cpu < 0) - cpu = 0; /* check cloexec flag */ - fd = sys_perf_event_open(&attr, -1, cpu, -1, + fd = sys_perf_event_open(&attr, 0, -1, -1, PERF_FLAG_FD_CLOEXEC); err = errno; @@ -35,7 +30,7 @@ static int perf_flag_probe(void) err, strerror(err)); /* not supported, confirm error related to PERF_FLAG_FD_CLOEXEC */ - fd = sys_perf_event_open(&attr, -1, cpu, -1, 0); + fd = sys_perf_event_open(&attr, 0, -1, -1, 0); err = errno; if (WARN_ONCE(fd < 0, -- cgit v1.1