perf tools: Add traceevent plugins support
v2 cover perf tools: Add traceevent plugins support
hi,
backporting traceevent plugin support from trace-cmd.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
Initial RFC is here:
http://marc.info/?l=linux-kernel&m=137803257921043&w=2
Having plugins allow traceevent to properly parse 'print fmt' line of tracepoint format and display proper/additional data in perf script command (check examples below).
Plugins backported from trace-cmd:
plugin_function.c
plugin_hrtimer.c
plugin_jbd2.c
plugin_kmem.c
plugin_kvm.c
plugin_mac80211.c
plugin_sched_switch.c
Plugins newly added:
plugin_cfg80211.c
plugin_scsi.c
plugin_xen.c
v2 changes:
- removing usage of malloc_or_die from event-pluginc.c
- using static methods in plugins where possible
- added Steven's Signed-off-by tags
- changed traceevent_host_bigendian with code suggested by Steven
- using pevent_print_func_field in hrtimer_start handler
- plugin_function.c changes:
* switching malloc_or_die into malloc
* removing index field from struct func_stack as it's not needed
* renaming get_index into add_and_get_index
* using '%*X' format string capability instead of the loop
- changing pevent_parse_format interface instead
of the pevent_parse_event as requested by Steven
- breakout LIBTRACEEVENT_FLAGS in Makefile.perf as suggested by Namhyung
- added better error handling for filename__read_str suggested by Nmahyung
- using _eprintf code in pr_stat function as suggested by Namhyung
- removing stackprotector feature check as it's not needed and is causing warnings - build failures on old gcc
Changes from RFC:
- added comment about repo/HEAD of the backport source
- omitted options plugin support
- change plugins installation path to:
$(perfexecdir)/traceevent/plugins
The code is available in here:
git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
perf/core_plugins
thanks,
jirka
Examples (diffs of 'perf script' output):
* data was generated by 'perf record -e 'cfg80211:*' -a' --- script.cfg80211.old
+++ script.cfg80211.new
- ifconfig 2705 [003] 662.896560: cfg80211:cfg80211_tx_mlme_mgmt: [FAILED TO PARSE] name=wlan0 ifindex=3 frame=ARRAY[c0, 00,
+ ifconfig 2705 [003] 662.896560: cfg80211:cfg80211_tx_mlme_mgmt: netdev:wlan0(3), ftype:0xc0
- kworker/u16:0 1697 [002] 664.808210: cfg80211:cfg80211_rx_mlme_mgmt: [FAILED TO PARSE] name=wlan0 ifindex=3 frame=ARRAY[b0, 00,
+ kworker/u16:0 1697 [002] 664.808210: cfg80211:cfg80211_rx_mlme_mgmt: netdev:wlan0(3), ftype:0xb0
* data was generated by 'perf record -e sched:sched_switch -a' --- script.sched_switch.old
+++ script.sched_switch.new
- perf 577 [002] 30965.311852: sched:sched_switch: prev_comm=perf prev_pid=577 prev_prio=120 prev_state=S ==> next_comm=sw
+ perf 577 [002] 30965.311852: sched:sched_switch: perf:577 [120] S ==> swapper/2:0 [120]
* data was generated by 'perf record -e ftrace:function ls' --- script.function.old
+++ script.function.new
- ls 10781 [001] 32667.291379: ftrace:function: ffffffff811adb80 <-- ffffffff811afc48
- ls 10781 [001] 32667.291379: ftrace:function: ffffffff811b35d0 <-- ffffffff811adb9b
- ls 10781 [001] 32667.291380: ftrace:function: ffffffff811b3520 <-- ffffffff811b35e8
- ls 10781 [001] 32667.291380: ftrace:function: ffffffff811b2720 <-- ffffffff811b3549
- ls 10781 [001] 32667.291381: ftrace:function: ffffffff81297e10 <-- ffffffff811b356c
- ls 10781 [001] 32667.291381: ftrace:function: ffffffff81298f40 <-- ffffffff81297e2c
- ls 10781 [001] 32667.291382: ftrace:function: ffffffff81076160 <-- ffffffff811afbf0
- ls 10781 [001] 32667.291383: ftrace:function: ffffffff811c3eb0 <-- ffffffff811afbfc
- ls 10781 [001] 32667.291383: ftrace:function: ffffffff8164e100 <-- ffffffff811c3ed8
- ls 10781 [001] 32667.291384: ftrace:function: ffffffff811a5d10 <-- ffffffff811c3f53
- ls 10781 [001] 32667.291384: ftrace:function: ffffffff811e8e70 <-- ffffffff811a5d58
- ls 10781 [001] 32667.291385: ftrace:function: ffffffff811f38e0 <-- ffffffff811a5d63
- ls 10781 [001] 32667.291385: ftrace:function: ffffffff811a9ff0 <-- ffffffff811a5d6b
- ls 10781 [001] 32667.291386: ftrace:function: ffffffff811a9fa0 <-- ffffffff811aa015
- ls 10781 [001] 32667.291387: ftrace:function: ffffffff810851c0 <-- ffffffff811aa053
- ls 10781 [001] 32667.291387: ftrace:function: ffffffff81090e00 <-- ffffffff81085211
+ ls 10781 [001] 32667.291379: ftrace:function: would_dump <-- setup_new_exec
+ ls 10781 [001] 32667.291379: ftrace:function: inode_permission <-- would_dump
+ ls 10781 [001] 32667.291380: ftrace:function: __inode_permission <-- inode_permission
+ ls 10781 [001] 32667.291380: ftrace:function: generic_permission <-- __inode_permission
+ ls 10781 [001] 32667.291381: ftrace:function: security_inode_permission <-- __inode_permission
+ ls 10781 [001] 32667.291381: ftrace:function: cap_inode_permission <-- security_inode_permission
+ ls 10781 [001] 32667.291382: ftrace:function: flush_signal_handlers <-- setup_new_exec
+ ls 10781 [001] 32667.291383: ftrace:function: do_close_on_exec <-- setup_new_exec
+ ls 10781 [001] 32667.291383: ftrace:function: _raw_spin_lock <-- do_close_on_exec
+ ls 10781 [001] 32667.291384: ftrace:function: filp_close <-- do_close_on_exec
+ ls 10781 [001] 32667.291384: ftrace:function: dnotify_flush <-- filp_close
+ ls 10781 [001] 32667.291385: ftrace:function: locks_remove_posix <-- filp_close
+ ls 10781 [001] 32667.291385: ftrace:function: fput <-- filp_close
+ ls 10781 [001] 32667.291386: ftrace:function: file_sb_list_del <-- fput
+ ls 10781 [001] 32667.291387: ftrace:function: task_work_add <-- fput
+ ls 10781 [001] 32667.291387: ftrace:function: kick_process <-- task_work_add
* data was generated by 'perf record -e 'kvm:*,kvmmmu:*' -a'
--- script.kvm.old
+++ script.kvm.new
qemu-system-x86 17414 [000] 6868.995053: kvm:kvm_exit: reason EPT_VIOLATION rip 0xfff0 info 184 0 qemu-system-x86 17414 [000] 6868.995109: kvm:kvm_emulate_insn: f0000:c46b:e4 71 (real)
- qemu-system-x86 3006 [002] 10562.079422: kvmmmu:kvm_mmu_get_page: [FAILED TO PARSE] mmu_valid_gen=0x2 gfn=0 role=122884 root_count=0
+ qemu-system-x86 3006 [002] 10562.079422: kvmmmu:kvm_mmu_get_page: new sp gfn 0 0/4 q0 direct --- !pge !nxe root 0 sync
- qemu-system-x86 3006 [002] 10562.080502: kvmmmu:kvm_mmu_prepare_zap_page: [FAILED TO PARSE] mmu_valid_gen=0x2 gfn=0 role=122884 root
+ qemu-system-x86 3006 [002] 10562.080502: kvmmmu:kvm_mmu_prepare_zap_page: 0/4 q0 direct --- !pge !nxe root 1 sync qemu-system-x86 3290 [002] 10708.755312: kvmmmu:fast_page_fault: [FAILED TO PARSE] vcpu_id=0 gva=4094486080 error_code=3 sptep=0xfff
- insmod 2576 [001] 781.731666: kvmmmu:kvm_mmu_sync_page: [FAILED TO PARSE] mmu_valid_gen=0x1 gfn=2 role=24624 root_count=1
+ insmod 2576 [001] 781.731666: kvmmmu:kvm_mmu_sync_page: 3/0 q0 --- !pge !nxe root 10 unsync
- insmod 2576 [001] 781.731668: kvmmmu:kvm_mmu_unsync_page: [FAILED TO PARSE] mmu_valid_gen=0x1 gfn=2 role=24624 root_count
+ insmod 2576 [001] 781.731668: kvmmmu:kvm_mmu_unsync_page: 3/0 q0 --- !pge !nxe root 10 unsync
---
Jiri Olsa (29):
perf tools: Remove stackprotector feature check
perf tools: Include test-stackprotector-all.c in test-all tools lib traceevent: Add plugin support
tools lib traceevent: Add plugin build support
tools lib traceevent: Add traceevent_host_bigendian function tools lib traceevent: Add jbd2 plugin
tools lib traceevent: Add hrtimer plugin
tools lib traceevent: Add kmem plugin
tools lib traceevent: Add kvm plugin
tools lib traceevent: Add mac80211 plugin
tools lib traceevent: Add sched_switch plugin
tools lib traceevent: Add function plugin
tools lib traceevent: Add xen plugin
tools lib traceevent: Add scsi plugin
tools lib traceevent: Add cfg80211 plugin
tools lib traceevent: Remove malloc_or_die from event-plugin.c tools lib traceevent: Use static functions in jbd2 plugin tools lib traceevent: Use pevent_print_func_field in hrtimer_start handler tools lib traceevent: Several cleanups for function plugin tools lib traceevent: Remove malloc_or_die from plugin_function.c tools lib traceevent: Update kvm plugin with is_writable_pte helper tools lib traceevent: Change pevent_parse_format to include pevent handle tools lib traceevent: Harmonize the install messages in lib-traceevent perf tools: Add build and install plugins targets
perf tools: Add udis86 disassembler feature check
perf tools: Add filename__read_str util function
perf tools: Add trace-event object
perf tools: Add trace-event global object for tracepoint interface perf tools: Overload pr_stat traceevent print function
tools/lib/traceevent/Makefile | 71 +++++++++++-- tools/lib/traceevent/event-parse.c | 56 ++++++---- tools/lib/traceevent/event-parse.h | 19 +++- tools/lib/traceevent/event-plugin.c | 215 +++++++++++++++++++++++++++++++++++++ tools/lib/traceevent/plugin_cfg80211.c | 24 +++++ tools/lib/traceevent/plugin_function.c | 160 ++++++++++++++++++++++++++++ tools/lib/traceevent/plugin_hrtimer.c | 78 ++++++++++++++ tools/lib/traceevent/plugin_jbd2.c | 68 ++++++++++++ tools/lib/traceevent/plugin_kmem.c | 72 +++++++++++++ tools/lib/traceevent/plugin_kvm.c | 436 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tools/lib/traceevent/plugin_mac80211.c | 95 +++++++++++++++++ tools/lib/traceevent/plugin_sched_switch.c | 148 ++++++++++++++++++++++++++ tools/lib/traceevent/plugin_scsi.c | 423 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tools/lib/traceevent/plugin_xen.c | 130 +++++++++++++++++++++++ tools/perf/Makefile.perf | 18 +++- tools/perf/builtin-script.c | 2 +- tools/perf/builtin-trace.c | 5 +- tools/perf/config/Makefile | 21 +++- tools/perf/config/feature-checks/Makefile | 12 +-- tools/perf/config/feature-checks/test-all.c | 10 ++ tools/perf/config/feature-checks/test-stackprotector.c | 6 -- tools/perf/config/feature-checks/test-udis86.c | 11 ++ tools/perf/util/debug.c | 30 +++++- tools/perf/util/debug.h | 2 + tools/perf/util/evsel.c | 44 +------- tools/perf/util/header.c | 8 +- tools/perf/util/python-ext-sources | 1 + tools/perf/util/session.h | 3 +- tools/perf/util/trace-event-parse.c | 13 --- tools/perf/util/trace-event-read.c | 20 ++-- tools/perf/util/trace-event.c | 82 +++++++++++++++ tools/perf/util/trace-event.h | 15 ++- tools/perf/util/util.c | 50 +++++++++ tools/perf/util/util.h | 1 + 34 files changed, 2220 insertions(+), 129 deletions(-) create mode 100644 tools/lib/traceevent/event-plugin.c create mode 100644 tools/lib/traceevent/plugin_cfg80211.c create mode 100644 tools/lib/traceevent/plugin_function.c create mode 100644 tools/lib/traceevent/plugin_hrtimer.c create mode 100644 tools/lib/traceevent/plugin_jbd2.c create mode 100644 tools/lib/traceevent/plugin_kmem.c create mode 100644 tools/lib/traceevent/plugin_kvm.c create mode 100644 tools/lib/traceevent/plugin_mac80211.c create mode 100644 tools/lib/traceevent/plugin_sched_switch.c create mode 100644 tools/lib/traceevent/plugin_scsi.c create mode 100644 tools/lib/traceevent/plugin_xen.c delete mode 100644 tools/perf/config/feature-checks/test-stackprotector.c create mode 100644 tools/perf/config/feature-checks/test-udis86.c create mode 100644 tools/perf/util/trace-event.c
PATCH 1/29 perf tools: Remove stackprotector feature check
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.
Removing stackprotector feature check which was used to enable the -Wstack-protector option.
--- tools/perf/config/Makefile | 5 ----- tools/perf/config/feature-checks/Makefile | 6 +----- tools/perf/config/feature-checks/test-stackprotector.c | 6 ------ 3 files changed, 1 insertion(+), 16 deletions(-) delete mode 100644 tools/perf/config/feature-checks/test-stackprotector.c
tools/perf/config/Makefile +0 -5
@@ -161,7 +161,6 @@ CORE_FEATURE_TESTS = \ libslang \ libunwind \ on-exit \ - stackprotector \ stackprotector-all \ timerfd
@@ -229,10 +228,6 @@ ifeq ($(feature-stackprotector-all), 1) CFLAGS += -fstack-protector-all endif -ifeq ($(feature-stackprotector), 1) - CFLAGS += -Wstack-protector -endif - ifeq ($(DEBUG),0) ifeq ($(feature-fortify-source), 1) CFLAGS += -D_FORTIFY_SOURCE=2
tools/perf/config/feature-checks/Makefile +1 -5
@@ -26,7 +26,6 @@ FILES= \ test-libunwind-debug-frame \ test-on-exit \ test-stackprotector-all \ - test-stackprotector \ test-timerfd CC := $(CC) -MD
@@ -38,7 +37,7 @@ BUILD = $(CC) $(CFLAGS) $(LDFLAGS) -o $(OUTPUT)$@ [email protected] ############################### test-all: - $(BUILD) -Werror -fstack-protector -fstack-protector-all -O2 -Werror -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell pkg-config --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl + $(BUILD) -Werror -fstack-protector-all -O2 -Werror -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell pkg-config --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl test-hello: $(BUILD)
@@ -46,9 +45,6 @@ test-hello: test-stackprotector-all: $(BUILD) -Werror -fstack-protector-all -test-stackprotector: - $(BUILD) -Werror -fstack-protector -Wstack-protector - test-fortify-source: $(BUILD) -O2 -Werror -D_FORTIFY_SOURCE=2
tools/perf/config/feature-checks/test-stackprotector.c deleted +0 -6
@@ -1,6 +0,0 @@ -#include <stdio.h> - -int main (void ) -{ -return puts ("hi" ); -}
Signature · 1 line
1.8.3.1
Re: [PATCH 01/29] perf tools: Remove stackprotector feature check
Em Thu, Nov 28, 2013 at 12:33:00PM +0100, Jiri Olsa escreveu:
Quoted 6 lines
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.
Removing stackprotector feature check which was used to enable the -Wstack-protector option.
Doesn't applies to my perf/core branch.
- Arnaldo
PATCH v3 1/29 perf tools: Remove stackprotector feature check
Quoted 9 lines · Arnaldo Carvalho de Melo
Em Thu, Nov 28, 2013 at 12:33:00PM +0100, Jiri Olsa escreveu:
Quoted 6 lines
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.
Removing stackprotector feature check which was used to enable the -Wstack-protector option.
Doesn't applies to my perf/core branch.
yep, that's that libunwind stuff messing up ;-)
attaching v3, I could resend whole rebased patchset if needed.. I also updated my perf/core_plugins branch
thanks,
jirka
---
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.
Removing stackprotector feature check which was used to enable the -Wstack-protector option.
--- tools/perf/config/Makefile | 5 ----- tools/perf/config/feature-checks/Makefile | 6 +----- tools/perf/config/feature-checks/test-stackprotector.c | 6 ------ 3 files changed, 1 insertion(+), 16 deletions(-) delete mode 100644 tools/perf/config/feature-checks/test-stackprotector.c
tools/perf/config/Makefile +0 -5
@@ -161,7 +161,6 @@ CORE_FEATURE_TESTS = \ libslang \ libunwind \ on-exit \ - stackprotector \ stackprotector-all \ timerfd
@@ -229,10 +228,6 @@ ifeq ($(feature-stackprotector-all), 1) CFLAGS += -fstack-protector-all endif -ifeq ($(feature-stackprotector), 1) - CFLAGS += -Wstack-protector -endif - ifeq ($(DEBUG),0) ifeq ($(feature-fortify-source), 1) CFLAGS += -D_FORTIFY_SOURCE=2
tools/perf/config/feature-checks/Makefile +1 -5
@@ -26,7 +26,6 @@ FILES= \ test-libunwind-debug-frame \ test-on-exit \ test-stackprotector-all \ - test-stackprotector \ test-timerfd CC := $(CC) -MD
@@ -38,7 +37,7 @@ BUILD = $(CC) $(CFLAGS) $(LDFLAGS) -o $(OUTPUT)$@ [email protected] ############################### test-all: - $(BUILD) -Werror -fstack-protector -fstack-protector-all -O2 -Werror -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell pkg-config --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl + $(BUILD) -Werror -fstack-protector-all -O2 -Werror -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell pkg-config --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl test-hello: $(BUILD)
@@ -46,9 +45,6 @@ test-hello: test-stackprotector-all: $(BUILD) -Werror -fstack-protector-all -test-stackprotector: - $(BUILD) -Werror -fstack-protector -Wstack-protector - test-fortify-source: $(BUILD) -O2 -Werror -D_FORTIFY_SOURCE=2
tools/perf/config/feature-checks/test-stackprotector.c deleted +0 -6
@@ -1,6 +0,0 @@ -#include <stdio.h> - -int main (void ) -{ -return puts ("hi" ); -}
Signature · 1 line
1.8.3.1
Re: [PATCHv3 01/29] perf tools: Remove stackprotector feature check
Quoted 38 lines · Jiri Olsa
Quoted 9 lines · Arnaldo Carvalho de Melo
Em Thu, Nov 28, 2013 at 12:33:00PM +0100, Jiri Olsa escreveu:
Quoted 6 lines
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.
Removing stackprotector feature check which was used to enable the -Wstack-protector option.
Doesn't applies to my perf/core branch.
yep, that's that libunwind stuff messing up ;-)
attaching v3, I could resend whole rebased patchset if needed.. I also updated my perf/core_plugins branch
thanks,
jirka---
We use -fstack-protector-all option to enable stack protecting for all available functions. There's no reason for enabling -Wstack-protector to get warning for unprotected functions.Removing stackprotector feature check which was used to enable the -Wstack-protector option.
Signed-off-by Jiri Olsa <[email protected]>Cc Corey Ashford <[email protected]>Cc Frederic Weisbecker <[email protected]>Cc Ingo Molnar <[email protected]>Cc Namhyung Kim <[email protected]>Cc Paul Mackerras <[email protected]>Cc Peter Zijlstra <[email protected]>Cc Arnaldo Carvalho de Melo <[email protected]>Cc Steven Rostedt <[email protected]>Cc David Ahern <[email protected]>
Quoted 6 lines
--- tools/perf/config/Makefile | 5 ----- tools/perf/config/feature-checks/Makefile | 6 +----- tools/perf/config/feature-checks/test-stackprotector.c | 6 ------ 3 files changed, 1 insertion(+), 16 deletions(-) delete mode 100644 tools/perf/config/feature-checks/test-stackprotector.c
This should also speed up the feature tests by about 10-20 msecs in the test-all case.
Thanks,
Ingo
PATCH 2/29 perf tools: Include test-stackprotector-all.c in test-all
In reply to Jiri OlsaWe're missing test-stackprotector-all.c in test-all check, adding it.
--- tools/perf/config/feature-checks/test-all.c | 5 +++++ 1 file changed, 5 insertions(+)
tools/perf/config/feature-checks/test-all.c +5 -0
@@ -85,6 +85,10 @@# include "test-timerfd.c" #undef main +#define main main_test_stackprotector_all +# include "test-stackprotector-all.c" +#undef main +int main (int argc,char *argv[]) {main_test_libpython ();
@@ -106,6 +110,7 @@ int main(int argc, char *argv[])main_test_backtrace ();main_test_libnuma ();main_test_timerfd (); +main_test_stackprotector_all ();return 0 ; }
Signature · 1 line
1.8.3.1
tip-bot for Jiri Olsa
merge bot[tip:perf/core] perf tools: Include test-stackprotector-all.c in test-all
perf tools: Include test-stackprotector-all.c in test-all
Commit-ID: bc32358b569247ceabba14dab297070a8fb255c4 Gitweb: http://git.kernel.org/tip/bc32358b569247ceabba14dab297070a8fb255c4 Author: Jiri Olsa <[email protected]> AuthorDate: Thu, 28 Nov 2013 12:33:01 +0100 Committer: Arnaldo Carvalho de Melo <[email protected]> CommitDate: Mon, 2 Dec 2013 16:19:26 -0300
perf tools: Include test-stackprotector-all.c in test-all
We're missing test-stackprotector-all.c in test-all check, adding it.
--- tools/perf/config/feature-checks/test-all.c | 5 +++++ 1 file changed, 5 insertions(+)
tools/perf/config/feature-checks/test-all.c +5 -0
@@ -85,6 +85,10 @@# include "test-timerfd.c" #undef main +#define main main_test_stackprotector_all +# include "test-stackprotector-all.c" +#undef main +int main (int argc,char *argv[]) {main_test_libpython ();
@@ -106,6 +110,7 @@ int main(int argc, char *argv[])main_test_backtrace ();main_test_libnuma ();main_test_timerfd (); +main_test_stackprotector_all ();return 0 ; }
PATCH 3/29 tools lib traceevent: Add plugin support
In reply to Jiri OlsaBackporting plugin support for traceevent lib.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
It's now possible to use following interface to load plugins (shared objects) to enhance pevent object functionality.
The plugin interface/hooks are as follows:
(taken from event-parse.h comments)
- 'pevent_plugin_loader' (required)
The function name to initialized the plugin.
int pevent_plugin_loader(struct pevent *pevent)
- 'pevent_plugin_unloader' (optional)
The function called just before unloading
int pevent_plugin_unloader(void)
- 'pevent_plugin_options' (optional)
Plugin options that can be set before loading
struct plugin_option pevent_plugin_options[] = {
{
.name = "option-name",
.plugin_alias = "overide-file-name", (optional) .description = "description of option to show users",
},
{
.name = NULL,
},
};
Array must end with .name = NULL;
The plugin_alias (below) can be used to give a shorter name to access the variable. Useful if a plugin handles more than one event.
NOTE options support is not backported yet.
- 'pevent_plugin_alias' (optional)
The name to use for finding options (uses filename if not defined)
New traceevent functions are added to search and load available plugins:
struct plugin_list* traceevent_load_plugins(struct pevent *pevent)
- loads plusing for 'struct pevent' object and returns loaded plugins list
void traceevent_unload_plugins(struct plugin_list *plugin_list);
- unload plugin list
--- tools/lib/traceevent/Makefile | 6 +- tools/lib/traceevent/event-parse.h | 5 + tools/lib/traceevent/event-plugin.c | 202 ++++++++++++++++++++++++++++++++++++ 3 files changed, 212 insertions(+), 1 deletion(-) create mode 100644 tools/lib/traceevent/event-plugin.c
tools/lib/traceevent/Makefile +5 -1
@@ -180,7 +180,11 @@ $(obj)/%.o: $(src)/%.c %.o: $(src)/%.c $(Q)$(call do_compile) -PEVENT_LIB_OBJS = event-parse.o trace-seq.o parse-filter.o parse-utils.o +PEVENT_LIB_OBJS = event-parse.o +PEVENT_LIB_OBJS += event-plugin.o +PEVENT_LIB_OBJS += trace-seq.o +PEVENT_LIB_OBJS += parse-filter.o +PEVENT_LIB_OBJS += parse-utils.o PEVENT_LIB_OBJS += kbuffer-parse.o ALL_OBJS = $(PEVENT_LIB_OBJS)
tools/lib/traceevent/event-parse.h +5 -0
@@ -377,6 +377,11 @@ enum pevent_errno { };#undef _PE +struct plugin_list ; + +struct plugin_list *traceevent_load_plugins (struct pevent *pevent); +void traceevent_unload_plugins (struct plugin_list *plugin_list); +struct cmdline ;struct cmdline_list ;struct func_map ;
tools/lib/traceevent/event-plugin.c new file +202 -0
202 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 4/29 tools lib traceevent: Add plugin build support
In reply to Jiri OlsaBackporting missing pieces of plugin building infrastructure:
- adding Makefile 'plugins' target to build all defined plugins
- adding Makefile 'install_plugins' target as 'install_lib' target dependency
- link plugin objects with shared object building
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
Plugins are by default installed into following locations:
'$(HOME)/.traceevent/plugins'
- if we are installing under $(HOME)
'$(prefix)/lib/traceevent/plugins'
- otherwise
This path is propagated to the plugin object as a plugins search path.
--- tools/lib/traceevent/Makefile | 52 ++++++++++++++++++++++++++++++++++++++----- 1 file changed, 47 insertions(+), 5 deletions(-)
tools/lib/traceevent/Makefile +47 -5
@@ -43,6 +43,30 @@ man_dir_SQ = '$(subst ','\'',$(man_dir))' export man_dir man_dir_SQ INSTALL export DESTDIR DESTDIR_SQ +set_plugin_dir := 1 + +# Set plugin_dir to preffered global plugin location +# If we install under $HOME directory we go under +# $(HOME)/.traceevent/plugins +# +# We dont set PLUGIN_DIR in case we install under $HOME +# directory, because by default the code looks under: +# $(HOME)/.traceevent/plugins by default. +# +ifeq ($(plugin_dir),) +ifeq ($(prefix),$(HOME)) +override plugin_dir = $(HOME)/.traceevent/plugins +set_plugin_dir := 0 +else +override plugin_dir = $(prefix)/lib/traceevent/plugins +endif +endif + +ifeq ($(set_plugin_dir),1) +PLUGIN_DIR = -DPLUGIN_DIR="$(DESTDIR)/$(plugin_dir)" +PLUGIN_DIR_SQ = '$(subst ','\'',$(PLUGIN_DIR))' +endif + # copy a bit from Linux kbuild ifeq ("$(origin V)", "command line")
@@ -96,6 +120,7 @@ export prefix bindir src obj # Shell quotes bindir_SQ = $(subst ','\'',$(bindir)) bindir_relative_SQ = $(subst ','\'',$(bindir_relative)) +plugin_dir_SQ = $(subst ','\'',$(plugin_dir)) LIB_FILE = libtraceevent.a libtraceevent.so
@@ -138,8 +163,8 @@ else print_app_build = echo ' BUILD '$(OBJ); print_fpic_compile = echo ' CC FPIC '$(OBJ); print_shared_lib_compile = echo ' BUILD SHARED LIB '$(OBJ); - print_plugin_obj_compile = echo ' BUILD PLUGIN OBJ '$(OBJ); - print_plugin_build = echo ' BUILD PLUGIN '$(OBJ); + print_plugin_obj_compile = echo ' CC FPIC '$(OBJ); + print_plugin_build = echo ' BUILD PLUGIN '$(OBJ); print_static_lib_build = echo ' BUILD STATIC LIB '$(OBJ); print_install = echo ' INSTALL '$1' to $(DESTDIR_SQ)$2'; endif
@@ -187,9 +212,11 @@ PEVENT_LIB_OBJS += parse-filter.o PEVENT_LIB_OBJS += parse-utils.o PEVENT_LIB_OBJS += kbuffer-parse.o -ALL_OBJS = $(PEVENT_LIB_OBJS) +PLUGINS := $(PLUGIN_OBJS:.o=.so) + +ALL_OBJS = $(PEVENT_LIB_OBJS) $(PLUGIN_OBJS) -CMD_TARGETS = $(LIB_FILE) +CMD_TARGETS = $(LIB_FILE) $(PLUGINS) TARGETS = $(CMD_TARGETS)
@@ -204,9 +231,17 @@ libtraceevent.so: $(PEVENT_LIB_OBJS) libtraceevent.a: $(PEVENT_LIB_OBJS) $(Q)$(do_build_static_lib) +plugins: $(PLUGINS) + $(PEVENT_LIB_OBJS): %.o: $(src)/%.c TRACEEVENT-CFLAGS $(Q)$(do_fpic_compile) +$(PLUGIN_OBJS): %.o : $(src)/%.c + $(Q)$(do_compile_plugin_obj) + +$(PLUGINS): %.so: %.o + $(Q)$(do_plugin_build) + define make_version.h (echo '/* This file is automatically generated. Do not modify. */'; \ echo \#define VERSION_CODE $(shell \
@@ -294,9 +329,16 @@ define do_install $(INSTALL) $1 '$(DESTDIR_SQ)$2' endef -install_lib: all_cmd +install_lib: all_cmd install_plugins $(Q)$(call do_install,$(LIB_FILE),$(bindir_SQ)) +PLUGINS_INSTALL = $(subst .so,.install,$(PLUGINS)) + +$(PLUGINS_INSTALL): %.install : %.so force + $(Q)$(call do_install,$<,$(plugin_dir_SQ)) + +install_plugins: $(PLUGINS_INSTALL) + install: install_lib clean:
Signature · 1 line
1.8.3.1
PATCH 5/29 tools lib traceevent: Add traceevent_host_bigendian function
In reply to Jiri OlsaAdding traceevent_host_bigendian function to get host endianity. It's used in following patches.
--- tools/lib/traceevent/event-parse.h | 10 ++++++++++ 1 file changed, 10 insertions(+)
tools/lib/traceevent/event-parse.h +10 -0
@@ -23,6 +23,7 @@#include <stdbool.h> #include <stdarg.h> #include <regex.h> +#include <string.h> #ifndef __maybe_unused#define __maybe_unused __attribute__((unused))
@@ -527,6 +528,15 @@ __data2host8(struct pevent *pevent, unsigned long long data)__data2host8 (pevent, __val); \ }) +static inline int traceevent_host_bigendian (void ) +{ +unsigned char str[] = {0x1 ,0x2 ,0x3 ,0x4 }; +unsigned int val; + +memcpy (&val, str,4 ); +return val ==0x01020304 ; +} +/* taken from kernel/trace/trace.h */ enum trace_flag_type {TRACE_FLAG_IRQS_OFF =0x01 ,
Signature · 1 line
1.8.3.1
PATCH 6/29 tools lib traceevent: Add jbd2 plugin
In reply to Jiri OlsaBackporting jbd2 plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds field resolving functions for following tracepoint events:
jbd2:jbd2_checkpoint_stats jbd2:jbd2_run_stats
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')
--- script.jbd2.old
+++ script.jbd2.new
- jbd2/dm-3-8 576 [000] 2983.748423: jbd2:jbd2_checkpoint_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bbb chp_time=0x0 forced_to_close=0 written=0 dropped=2 + jbd2/dm-3-8 576 [000] 2983.748423: jbd2:jbd2_checkpoint_stats: dev 253,3 tid 4139963 chp_time 0 forced_to_close 0 written 0 dropped 2 - jbd2/dm-3-8 576 [000] 2983.852789: jbd2:jbd2_run_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bc0 wait=0x0 request_delay=0x0 running=0x138d locked=0x0 flushing=0x0 logging=0x68 handle_count=2995 blocks=17 blocks_logged=18 + jbd2/dm-3-8 576 [000] 2983.852789: jbd2:jbd2_run_stats: dev 253,3 tid 4139968 wait 0 request_delay 0 running 5005 locked 0 flushing 0 logging 104 handle_count 2995 blocks 17 blocks_logged 18
This plugin also adds jbd2_dev_to_name function, which was removed from jdb2 tracepoints recently. Keeping it in for backward compatibility reasons.
--- tools/lib/traceevent/Makefile | 2 ++ tools/lib/traceevent/plugin_jbd2.c | 66 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 68 insertions(+) create mode 100644 tools/lib/traceevent/plugin_jbd2.c
tools/lib/traceevent/Makefile +2 -0
@@ -212,6 +212,8 @@ PEVENT_LIB_OBJS += parse-filter.o PEVENT_LIB_OBJS += parse-utils.o PEVENT_LIB_OBJS += kbuffer-parse.o +PLUGIN_OBJS = plugin_jbd2.o + PLUGINS := $(PLUGIN_OBJS:.o=.so) ALL_OBJS = $(PEVENT_LIB_OBJS) $(PLUGIN_OBJS)
tools/lib/traceevent/plugin_jbd2.c new file +66 -0
66 lines not on this page · raw message
Signature · 1 line
1.8.3.1
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
Em Thu, Nov 28, 2013 at 12:33:05PM +0100, Jiri Olsa escreveu:
Quoted 5 lines
Backporting jbd2 plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
Quoted 7 lines
This plugin adds field resolving functions for following tracepoint events:
jbd2:jbd2_checkpoint_stats jbd2:jbd2_run_statsThe diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')
The lines after one that starts with
---
are ignored by 'git am', please ident them with at least one space, I'm doing that this time.
Quoted 115 lines
script.jbd2.new renamed from script.jbd2.old +2
-2- jbd2/dm-3-8 576 [000] 2983.748423: jbd2:jbd2_checkpoint_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bbb chp_time=0x0 forced_to_close=0 written=0 dropped=2 + jbd2/dm-3-8 576 [000] 2983.748423: jbd2:jbd2_checkpoint_stats: dev 253,3 tid 4139963 chp_time 0 forced_to_close 0 written 0 dropped 2 - jbd2/dm-3-8 576 [000] 2983.852789: jbd2:jbd2_run_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bc0 wait=0x0 request_delay=0x0 running=0x138d locked=0x0 flushing=0x0 logging=0x68 handle_count=2995 blocks=17 blocks_logged=18 + jbd2/dm-3-8 576 [000] 2983.852789: jbd2:jbd2_run_stats: dev 253,3 tid 4139968 wait 0 request_delay 0 running 5005 locked 0 flushing 0 logging 104 handle_count 2995 blocks 17 blocks_logged 18This plugin also adds
jbd2_dev_to_namefunction, which was removed from jdb2 tracepoints recently. Keeping it in for backward compatibility reasons.Signed-off-by Steven Rostedt <[email protected]>Signed-off-by Jiri Olsa <[email protected]>Cc Corey Ashford <[email protected]>Cc Frederic Weisbecker <[email protected]>Cc Ingo Molnar <[email protected]>Cc Namhyung Kim <[email protected]>Cc Paul Mackerras <[email protected]>Cc Peter Zijlstra <[email protected]>Cc Arnaldo Carvalho de Melo <[email protected]>Cc Steven Rostedt <[email protected]>Cc David Ahern <[email protected]>--- tools/lib/traceevent/Makefile | 2 ++ tools/lib/traceevent/plugin_jbd2.c | 66 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 68 insertions(+) create mode 100644 tools/lib/traceevent/plugin_jbd2.ctools/lib/traceevent/Makefile +2
-0@@ -212,6 +212,8 @@ PEVENT_LIB_OBJS += parse-filter.o PEVENT_LIB_OBJS += parse-utils.o PEVENT_LIB_OBJS += kbuffer-parse.o +PLUGIN_OBJS = plugin_jbd2.o + PLUGINS := $(PLUGIN_OBJS:.o=.so) ALL_OBJS = $(PEVENT_LIB_OBJS) $(PLUGIN_OBJS)tools/lib/traceevent/plugin_jbd2.c new file +66
-0@@ -0,0 +1,66 @@ +/* +* Copyright (C) 2010 Red Hat Inc, Steven Rostedt <[email protected]> +* +* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU Lesser General Public +* License as published by the Free Software Foundation; +* version 2.1 of the License (not later!) +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU Lesser General Public License for more details. +* +* You should have received a copy of the GNU Lesser General Public +* License along with this program; if not, see <http://www.gnu.org/licenses> +* +* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +*/ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include "event-parse.h" + +#define MINORBITS 20 +#define MINORMASK ((1U << MINORBITS) - 1) + +#define MAJOR (dev) ((unsigned int) ((dev) >> MINORBITS)) +#define MINOR (dev) ((unsigned int) ((dev) & MINORMASK)) + +unsigned long long process_jbd2_dev_to_name (struct trace_seq *s, +unsigned long long *args) +{ +unsigned int dev = args[0 ]; + +trace_seq_printf (s,"%d:%d" ,MAJOR (dev),MINOR (dev)); +return 0 ; +} + +unsigned long long process_jiffies_to_msecs (struct trace_seq *s, +unsigned long long *args) +{ +unsigned long long jiffies = args[0 ]; + +trace_seq_printf (s,"%lld" , jiffies); +return jiffies; +} + +int PEVENT_PLUGIN_LOADER (struct pevent *pevent) +{ +pevent_register_print_function (pevent, + process_jbd2_dev_to_name, +PEVENT_FUNC_ARG_STRING , +"jbd2_dev_to_name" , +PEVENT_FUNC_ARG_INT , +PEVENT_FUNC_ARG_VOID ); + +pevent_register_print_function (pevent, + process_jiffies_to_msecs, +PEVENT_FUNC_ARG_LONG , +"jiffies_to_msecs" , +PEVENT_FUNC_ARG_LONG , +PEVENT_FUNC_ARG_VOID ); +return 0 ; +}Signature · 1 line
1.8.3.1
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
Em Fri, Nov 29, 2013 at 03:42:35PM -0300, Arnaldo Carvalho de Melo escreveu:
Quoted 10 lines
Em Thu, Nov 28, 2013 at 12:33:05PM +0100, Jiri Olsa escreveu:
Quoted 6 lines
This plugin adds field resolving functions for following tracepoint events:
jbd2:jbd2_checkpoint_stats
jbd2:jbd2_run_statsThe diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')The lines after one that starts with
Quoted 7 lines
---
are ignored by 'git am', please ident them with at least one space, I'm doing that this time.
script.jbd2.new
Quoted 2 lines
script.jbd2.new renamed from script.jbd2.old +0
-0
Also I just tried:
[acme@ssdandy linux]$ git log --oneline | head -7 b0496b02fcea tools lib traceevent: Add hrtimer plugin 1ec7a71e0b2c tools lib traceevent: Add jbd2 plugin c63100c95c6e tools lib traceevent: Add traceevent_host_bigendian function d54b335656b7 tools lib traceevent: Add plugin build support e796a0d73bcc tools lib traceevent: Add plugin support 28c803842250 perf tools: Include test-stackprotector-all.c in test-all 1a69f6ce513d perf timechart: Move wake_events list to 'struct timechart' [acme@ssdandy linux]$
[acme@ssdandy linux]$ time make O=/tmp/build/perf -C tools/perf/ install <SNIP successful build>
[acme@ssdandy linux]$ ls -la ~/.traceevent
ls: cannot access /home/acme/.traceevent: No such file or directory [acme@ssdandy linux]$
Where did the plugins go? At least two should be there by now, no?
- Arnaldo
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
Quoted 31 lines · Arnaldo Carvalho de Melo
Em Fri, Nov 29, 2013 at 03:42:35PM -0300, Arnaldo Carvalho de Melo escreveu:
Quoted 10 lines
Em Thu, Nov 28, 2013 at 12:33:05PM +0100, Jiri Olsa escreveu:
Quoted 6 lines
This plugin adds field resolving functions for following tracepoint events:
jbd2:jbd2_checkpoint_stats
jbd2:jbd2_run_statsThe diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')The lines after one that starts with
Quoted 7 lines
---
are ignored by 'git am', please ident them with at least one space, I'm doing that this time.
script.jbd2.new
Quoted 2 lines
script.jbd2.new renamed from script.jbd2.old +0
-0Also I just tried:
[acme@ssdandy linux]$ git log --oneline | head -7
b0496b02fceatools lib traceevent: Add hrtimer plugin1ec7a71e0b2ctools lib traceevent: Add jbd2 pluginc63100c95c6etools lib traceevent: Addtraceevent_host_bigendianfunctiond54b335656b7tools lib traceevent: Add plugin build supporte796a0d73bcctools lib traceevent: Add plugin support28c803842250perf tools: Includetest-stackprotector-all.cin test-all1a69f6ce513dperf timechart: Movewake_eventslist to 'struct timechart' [acme@ssdandy linux]$
Quoted 8 lines
[acme@ssdandy linux]$ time make O=/tmp/build/perf -C
tools/perf/install <SNIP successful build>[acme@ssdandy linux]$ ls -la ~/.traceevent
ls: cannot access/home/acme/.traceevent: No such file or directory [acme@ssdandy linux]$Where did the plugins go? At least two should be there by now, no?
At this stage (patches applied) you'd need to run the make for traceevent directory.
The perf installation link is done within this patch:
perf tools: Add build and install plugins targets
which I presume you haven't applied yet?
jirka
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
Em Sat, Nov 30, 2013 at 05:49:21PM +0100, Jiri Olsa escreveu:
Quoted 23 lines · Arnaldo Carvalho de Melo
Quoted 13 lines
Em Fri, Nov 29, 2013 at 03:42:35PM -0300, Arnaldo Carvalho de Melo escreveu:
Also I just tried:[acme@ssdandy linux]$ git log --oneline | head -7
b0496b02fceatools lib traceevent: Add hrtimer plugin1ec7a71e0b2ctools lib traceevent: Add jbd2 pluginc63100c95c6etools lib traceevent: Addtraceevent_host_bigendianfunctiond54b335656b7tools lib traceevent: Add plugin build supporte796a0d73bcctools lib traceevent: Add plugin support28c803842250perf tools: Includetest-stackprotector-all.cin test-all1a69f6ce513dperf timechart: Movewake_eventslist to 'struct timechart' [acme@ssdandy linux]$Quoted 8 lines
[acme@ssdandy linux]$ time make O=/tmp/build/perf -C
tools/perf/install <SNIP successful build>[acme@ssdandy linux]$ ls -la ~/.traceevent
ls: cannot access/home/acme/.traceevent: No such file or directory [acme@ssdandy linux]$Where did the plugins go? At least two should be there by now, no?
Quoted 7 lines
At this stage (patches applied) you'd need to run the make for traceevent directory.
The perf installation link is done within this patch:
perf tools: Add build and install plugins targetswhich I presume you haven't applied yet?
No I haven't, and I expected that I could test it straight away, so I suggest you reorder the patches so that, right after the first plugin is introduced a:
make -C tools/perf install
in all its variants (in place, with O=, etc) allows one to test it.
- Arnaldo
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
Quoted 32 lines · Arnaldo Carvalho de Melo
Em Sat, Nov 30, 2013 at 05:49:21PM +0100, Jiri Olsa escreveu:
Quoted 23 lines · Arnaldo Carvalho de Melo
Quoted 13 lines
Em Fri, Nov 29, 2013 at 03:42:35PM -0300, Arnaldo Carvalho de Melo escreveu:
Also I just tried:[acme@ssdandy linux]$ git log --oneline | head -7
b0496b02fceatools lib traceevent: Add hrtimer plugin1ec7a71e0b2ctools lib traceevent: Add jbd2 pluginc63100c95c6etools lib traceevent: Addtraceevent_host_bigendianfunctiond54b335656b7tools lib traceevent: Add plugin build supporte796a0d73bcctools lib traceevent: Add plugin support28c803842250perf tools: Includetest-stackprotector-all.cin test-all1a69f6ce513dperf timechart: Movewake_eventslist to 'struct timechart' [acme@ssdandy linux]$Quoted 8 lines
[acme@ssdandy linux]$ time make O=/tmp/build/perf -C
tools/perf/install <SNIP successful build>[acme@ssdandy linux]$ ls -la ~/.traceevent
ls: cannot access/home/acme/.traceevent: No such file or directory [acme@ssdandy linux]$Where did the plugins go? At least two should be there by now, no?
Quoted 7 lines
At this stage (patches applied) you'd need to run the make for traceevent directory.
The perf installation link is done within this patch:
perf tools: Add build and install plugins targetswhich I presume you haven't applied yet?
Quoted 7 lines
No I haven't, and I expected that I could test it straight away, so I suggest you reorder the patches so that, right after the first plugin is introduced a:
make -C
tools/perfinstallin all its variants (in place, with O=, etc) allows one to test it.
ok, I will send v3
jirka
Re: [PATCH 06/29] tools lib traceevent: Add jbd2 plugin
In reply to Arnaldo Carvalho de MeloQuoted 15 lines · Arnaldo Carvalho de Melo
Em Thu, Nov 28, 2013 at 12:33:05PM +0100, Jiri Olsa escreveu:
Quoted 5 lines
Backporting jbd2 plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.gitQuoted 7 lines
This plugin adds field resolving functions for following tracepoint events:
jbd2:jbd2_checkpoint_stats jbd2:jbd2_run_statsThe diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')
Quoted 6 lines
The lines after one that starts with
---
are ignored by 'git am', please ident them with at least one space, I'm doing that this time.
oops, ok
thanks,
jirka
PATCH 7/29 tools lib traceevent: Add hrtimer plugin
In reply to Jiri OlsaBackporting hrtimer plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds function field resolving for following tracepoint events:
timer:hrtimer_expire_entry timer:hrtimer_start
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'timer:hrtimer*' -a')
--- script.hrtimer.old
+++ script.hrtimer.new
- swapper 0 [000] 27405.519092: timer:hrtimer_start: [FAILED TO PARSE] hrtimer=0xffff88021e20e800 function=0xffffffff810c0e10 expires=27398383000000 softexpires=27398383000000 + swapper 0 [000] 27405.519103: timer:hrtimer_start: hrtimer=0xffff88021e20e800 function=tick_sched_timer expires=27398383000000 softexpires=27398383000000 - swapper 0 [001] 27405.519544: timer:hrtimer_expire_entry: [FAILED TO PARSE] hrtimer=0xffff880211334058 now=27398294182491 function=0xffffffff81086f20 + swapper 0 [001] 27405.519544: timer:hrtimer_expire_entry: hrtimer=0xffff880211334058 now=27398294182491 function=posix_timer_fn/0x0
Check the 'function' field is translated into the function name.
--- tools/lib/traceevent/Makefile | 3 +- tools/lib/traceevent/plugin_hrtimer.c | 93 +++++++++++++++++++++++++++++++++++ 2 files changed, 95 insertions(+), 1 deletion(-) create mode 100644 tools/lib/traceevent/plugin_hrtimer.c
tools/lib/traceevent/Makefile +2 -1
@@ -212,7 +212,8 @@ PEVENT_LIB_OBJS += parse-filter.o PEVENT_LIB_OBJS += parse-utils.o PEVENT_LIB_OBJS += kbuffer-parse.o -PLUGIN_OBJS = plugin_jbd2.o +PLUGIN_OBJS = plugin_jbd2.o +PLUGIN_OBJS += plugin_hrtimer.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_hrtimer.c new file +93 -0
93 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 8/29 tools lib traceevent: Add kmem plugin
In reply to Jiri OlsaBackporting kmem plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds call_site field resolving for following tracepoint events:
kmem:kfree
kmem:kmalloc
kmem:kmalloc_node
kmem:kmem_cache_alloc kmem:kmem_cache_alloc_node kmem:kmem_cache_free
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'kmem:*' -a')
--- script.kmem.old
+++ script.kmem.new
- perf 27846 [001] 29643.403319: kmem:kfree: call_site=ffffffff810e64f6 ptr=(nil) + perf 27846 [001] 29643.403238: kmem:kfree: (__audit_syscall_exit+0x1f6) call_site=ffffffff810e64f6 ptr=(nil) - perf 27846 [001] 29643.403337: kmem:kmem_cache_alloc: call_site=ffffffff812ff0c5 ptr=0xffff88020e155630 bytes_req=560 bytes_alloc=568 gfp_flags=GFP_KERNEL + perf 27846 [001] 29643.403337: kmem:kmem_cache_alloc: (radix_tree_preload+0x35) call_site=ffffffff812ff0c5 ptr=0xffff88020e155630 bytes_req=560 bytes_alloc=568 gfp_flags=GFP_KERNEL - perf 27846 [001] 29643.403342: kmem:kmem_cache_free: call_site=ffffffff8126ec61 ptr=0xffff88020dffe750 + perf 27846 [001] 29643.403342: kmem:kmem_cache_free: (jbd2_journal_stop+0x221) call_site=ffffffff8126ec61 ptr=0xffff88020dffe750 - firefox 954 [000] 29643.445477: kmem:kmem_cache_alloc_node: call_site=ffffffff8153c64e ptr=0xffff8801cecb4000 bytes_req=256 bytes_alloc=256 gfp_flags=GFP_KERNEL|GFP_REPEAT node=-1 + firefox 954 [000] 29643.445477: kmem:kmem_cache_alloc_node: (__alloc_skb+0x4e) call_site=ffffffff8153c64e ptr=0xffff8801cecb4000 bytes_req=256 bytes_alloc=256 gfp_flags=GFP_KERNEL|GFP_REPEAT node=-1 - perf 27846 [001] 29643.445510: kmem:kmalloc: call_site=ffffffff81250642 ptr=0xffff88020fd6c300 bytes_req=96 bytes_alloc=96 gfp_flags=GFP_NOFS|GFP_ZERO + perf 27846 [001] 29643.445510: kmem:kmalloc: (ext4_ext_find_extent+0x362) call_site=ffffffff81250642 ptr=0xffff88020fd6c300 bytes_req=96 bytes_alloc=96 gfp_flags=GFP_NOFS|GFP_ZERO - X 888 [002] 29643.445546: kmem:kmalloc_node: call_site=ffffffff8153c67e ptr=0xffff880103828e00 bytes_req=384 bytes_alloc=512 gfp_flags=GFP_KERNEL|GFP_NOWARN|GFP_REPEAT|GFP_NOMEMALLOC node=-1 + X 888 [002] 29643.445546: kmem:kmalloc_node: (__alloc_skb+0x7e) call_site=ffffffff8153c67e ptr=0xffff880103828e00 bytes_req=384 bytes_alloc=512 gfp_flags=GFP_KERNEL|GFP_NOWARN|GFP_REPEAT|GFP_NOMEMALLOC node=-1
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_kmem.c | 72 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 73 insertions(+) create mode 100644 tools/lib/traceevent/plugin_kmem.c
tools/lib/traceevent/Makefile +1 -0
@@ -214,6 +214,7 @@ PEVENT_LIB_OBJS += kbuffer-parse.o PLUGIN_OBJS = plugin_jbd2.o PLUGIN_OBJS += plugin_hrtimer.o +PLUGIN_OBJS += plugin_kmem.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_kmem.c new file +72 -0
72 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 9/29 tools lib traceevent: Add kvm plugin
In reply to Jiri OlsaBackporting kvm plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds field resolving functions for following tracepoint events:
kvm:kvm_exit
kvm:kvm_emulate_insn
kvm:kvm_nested_vmexit kvm:kvm_nested_vmexit_inject kvmmmu:kvm_mmu_get_page kvmmmu:kvm_mmu_sync_page kvmmmu:kvm_mmu_unsync_page kvmmmu:kvm_mmu_zap_page kvmmmu:kvm_mmu_prepare_zap_page
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'kvm:*,kvmmmu:*' -a')
--- script.kvm.old
+++ script.kvm.new
qemu-system-x86 17414 [000] 6868.995053: kvm:kvm_exit: reason EPT_VIOLATION rip 0xfff0 info 184 0 qemu-system-x86 17414 [000] 6868.995109: kvm:kvm_emulate_insn: f0000:c46b:e4 71 (real)
- qemu-system-x86 3006 [002] 10562.079422: kvmmmu:kvm_mmu_get_page: [FAILED TO PARSE] mmu_valid_gen=0x2 gfn=0 role=122884 root_count=0 unsync=0 created=1 + qemu-system-x86 3006 [002] 10562.079422: kvmmmu:kvm_mmu_get_page: new sp gfn 0 0/4 q0 direct --- !pge !nxe root 0 sync - qemu-system-x86 3006 [002] 10562.080502: kvmmmu:kvm_mmu_prepare_zap_page: [FAILED TO PARSE] mmu_valid_gen=0x2 gfn=0 role=122884 root_count=1 unsync=0 + qemu-system-x86 3006 [002] 10562.080502: kvmmmu:kvm_mmu_prepare_zap_page: 0/4 q0 direct --- !pge !nxe root 1 sync qemu-system-x86 3290 [002] 10708.755312: kvmmmu:fast_page_fault: [FAILED TO PARSE] vcpu_id=0 gva=4094486080 error_code=3 sptep=0xffff88019f1e3670 old_spte=336391285 new_spte=336391287 retry=1 - insmod 2576 [001] 781.731666: kvmmmu:kvm_mmu_sync_page: [FAILED TO PARSE] mmu_valid_gen=0x1 gfn=2 role=24624 root_count=10 unsync=1 + insmod 2576 [001] 781.731666: kvmmmu:kvm_mmu_sync_page: 3/0 q0 --- !pge !nxe root 10 unsync - insmod 2576 [001] 781.731668: kvmmmu:kvm_mmu_unsync_page: [FAILED TO PARSE] mmu_valid_gen=0x1 gfn=2 role=24624 root_count=10 unsync=1 + insmod 2576 [001] 781.731668: kvmmmu:kvm_mmu_unsync_page: 3/0 q0 --- !pge !nxe root 10 unsync
Note:
- kvm_mmu_zap_page is replaced by kvm_mmu_prepare_zap_page in current kernel, keeping it for backward compatibility
- some of the tracepoints keep the same output even with the plugin handling: kvm:kvm_exit, kvm:kvm_emulate_insn
- the 'kvmmmu:fast_page_fault' is still broken because of missing is_writable_pte function and is fixed in another patch
- ommited following tracepoints from backport because the output was buggy
kvm:kvm_nested_vmexit
kvm:kvm_nested_vmexit_inject
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_kvm.c | 419 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 420 insertions(+) create mode 100644 tools/lib/traceevent/plugin_kvm.c
tools/lib/traceevent/Makefile +1 -0
@@ -215,6 +215,7 @@ PEVENT_LIB_OBJS += kbuffer-parse.o PLUGIN_OBJS = plugin_jbd2.o PLUGIN_OBJS += plugin_hrtimer.o PLUGIN_OBJS += plugin_kmem.o +PLUGIN_OBJS += plugin_kvm.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_kvm.c new file +419 -0
419 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 10/29 tools lib traceevent: Add mac80211 plugin
In reply to Jiri OlsaBackporting mac80211 plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds changed field resolving for
mac80211:drv_bss_info_changed tracepoint event.
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'mac80211:drv_bss_info_changed' -a')
--- script.mac80211.old
+++ script.mac80211.new
- ifconfig 3711 [000] 1290.446492: mac80211:drv_bss_info_changed: phy0 vif:wlan0(2) changed:0x309f + ifconfig 3711 [000] 1290.446492: mac80211:drv_bss_info_changed: phy0 vif:wlan0(2) + assoc:0 aid:2 cts:0 shortpre:0 shortslot:0 dtimper:1 + bcnint:102 assoc_cap:0x431 basic_rates:0xf enable_beacon:0 + ht_operation_mode:0
Omitting the mac80211:drv_config tracepoint handling because the kernel tracepoint changed its prototype and the plugin handler is no longer working.
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_mac80211.c | 95 ++++++++++++++++++++++++++++++++++ 2 files changed, 96 insertions(+) create mode 100644 tools/lib/traceevent/plugin_mac80211.c
tools/lib/traceevent/Makefile +1 -0
@@ -216,6 +216,7 @@ PLUGIN_OBJS = plugin_jbd2.o PLUGIN_OBJS += plugin_hrtimer.o PLUGIN_OBJS += plugin_kmem.o PLUGIN_OBJS += plugin_kvm.o +PLUGIN_OBJS += plugin_mac80211.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_mac80211.c new file +95 -0
95 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 11/29 tools lib traceevent: Add sched_switch plugin
In reply to Jiri OlsaBackporting sched_switch plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds fields resolving for
sched:sched_switch tracepoint event.
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e sched:sched_switch -a')
--- script.sched_switch.old
+++ script.sched_switch.new
- perf 577 [002] 30965.311852: sched:sched_switch: prev_comm=perf prev_pid=577 prev_prio=120 prev_state=S ==> next_comm=swapper/2 next_pid=0 next_prio=120 + perf 577 [002] 30965.311852: sched:sched_switch: perf:577 [120] S ==> swapper/2:0 [120]
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_sched_switch.c | 148 +++++++++++++++++++++++++++++ 2 files changed, 149 insertions(+) create mode 100644 tools/lib/traceevent/plugin_sched_switch.c
tools/lib/traceevent/Makefile +1 -0
@@ -217,6 +217,7 @@ PLUGIN_OBJS += plugin_hrtimer.o PLUGIN_OBJS += plugin_kmem.o PLUGIN_OBJS += plugin_kvm.o PLUGIN_OBJS += plugin_mac80211.o +PLUGIN_OBJS += plugin_sched_switch.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_sched_switch.c new file +148 -0
148 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 12/29 tools lib traceevent: Add function plugin
In reply to Jiri OlsaBackporting function plugin.
Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
This plugin adds function and parent function fields resolving for ftrace:function tracepoint event.
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e ftrace:function ls')
--- script.function.old
+++ script.function.new
- ls 10781 [001] 32667.291379: ftrace:function: ffffffff811adb80 <-- ffffffff811afc48 - ls 10781 [001] 32667.291379: ftrace:function: ffffffff811b35d0 <-- ffffffff811adb9b - ls 10781 [001] 32667.291380: ftrace:function: ffffffff811b3520 <-- ffffffff811b35e8 - ls 10781 [001] 32667.291380: ftrace:function: ffffffff811b2720 <-- ffffffff811b3549 - ls 10781 [001] 32667.291381: ftrace:function: ffffffff81297e10 <-- ffffffff811b356c - ls 10781 [001] 32667.291381: ftrace:function: ffffffff81298f40 <-- ffffffff81297e2c - ls 10781 [001] 32667.291382: ftrace:function: ffffffff81076160 <-- ffffffff811afbf0 - ls 10781 [001] 32667.291383: ftrace:function: ffffffff811c3eb0 <-- ffffffff811afbfc - ls 10781 [001] 32667.291383: ftrace:function: ffffffff8164e100 <-- ffffffff811c3ed8 - ls 10781 [001] 32667.291384: ftrace:function: ffffffff811a5d10 <-- ffffffff811c3f53 - ls 10781 [001] 32667.291384: ftrace:function: ffffffff811e8e70 <-- ffffffff811a5d58 - ls 10781 [001] 32667.291385: ftrace:function: ffffffff811f38e0 <-- ffffffff811a5d63 - ls 10781 [001] 32667.291385: ftrace:function: ffffffff811a9ff0 <-- ffffffff811a5d6b - ls 10781 [001] 32667.291386: ftrace:function: ffffffff811a9fa0 <-- ffffffff811aa015 - ls 10781 [001] 32667.291387: ftrace:function: ffffffff810851c0 <-- ffffffff811aa053 - ls 10781 [001] 32667.291387: ftrace:function: ffffffff81090e00 <-- ffffffff81085211 + ls 10781 [001] 32667.291379: ftrace:function: would_dump <-- setup_new_exec + ls 10781 [001] 32667.291379: ftrace:function: inode_permission <-- would_dump + ls 10781 [001] 32667.291380: ftrace:function: __inode_permission <-- inode_permission + ls 10781 [001] 32667.291380: ftrace:function: generic_permission <-- __inode_permission + ls 10781 [001] 32667.291381: ftrace:function: security_inode_permission <-- __inode_permission + ls 10781 [001] 32667.291381: ftrace:function: cap_inode_permission <-- security_inode_permission + ls 10781 [001] 32667.291382: ftrace:function: flush_signal_handlers <-- setup_new_exec + ls 10781 [001] 32667.291383: ftrace:function: do_close_on_exec <-- setup_new_exec + ls 10781 [001] 32667.291383: ftrace:function: _raw_spin_lock <-- do_close_on_exec + ls 10781 [001] 32667.291384: ftrace:function: filp_close <-- do_close_on_exec + ls 10781 [001] 32667.291384: ftrace:function: dnotify_flush <-- filp_close + ls 10781 [001] 32667.291385: ftrace:function: locks_remove_posix <-- filp_close + ls 10781 [001] 32667.291385: ftrace:function: fput <-- filp_close + ls 10781 [001] 32667.291386: ftrace:function: file_sb_list_del <-- fput + ls 10781 [001] 32667.291387: ftrace:function: task_work_add <-- fput + ls 10781 [001] 32667.291387: ftrace:function: kick_process <-- task_work_add
Removing options support as it's not backported yet. Currently this plugin supports 2 options:
'parent' to display parent function 'indent' to show function call indents
Enabling both of them by default.
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_function.c | 151 +++++++++++++++++++++++++++++++++ 2 files changed, 152 insertions(+) create mode 100644 tools/lib/traceevent/plugin_function.c
tools/lib/traceevent/Makefile +1 -0
@@ -218,6 +218,7 @@ PLUGIN_OBJS += plugin_kmem.o PLUGIN_OBJS += plugin_kvm.o PLUGIN_OBJS += plugin_mac80211.o PLUGIN_OBJS += plugin_sched_switch.o +PLUGIN_OBJS += plugin_function.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_function.c new file +151 -0
151 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 13/29 tools lib traceevent: Add xen plugin
In reply to Jiri OlsaAdding xen plugin.
This plugin adds fields resolving for
following tracepoint events:
xen:xen_mc_entry
xen:xen_mc_extend_args
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'xen:*' ls')
--- script.xen.old
+++ script.xen.new
- swapper 0 [002] 136.267492: xen:xen_mc_entry: [FAILED TO PARSE] op=3 nargs=2 args=ARRAY[18, 00, 00, 00, 00, 00, 00, 00, 00, e0, d4, 4b, 04, 88, ff, ff, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] + swapper 0 [002] 136.267492: xen:xen_mc_entry: op 3(stack_switch) args [18, 0, 0, 0, 0, 0] - perf 1970 [008] 136.273319: xen:xen_mc_extend_args: [FAILED TO PARSE] op=1 args=16 res=1 + perf 1970 [008] 136.273319: xen:xen_mc_extend_args: extending op 1(mmu_update) by 16 bytes res ???
NOTE We still do not handle the 'sizeof' and fail
to parse following xen tracepoints:
xen:xen_mmu_set_pte
xen:xen_mmu_set_pte_atomic
xen:xen_mmu_set_domain_pte
xen:xen_mmu_set_pte_at
xen:xen_mmu_set_pmd
xen:xen_mmu_set_pud
xen:xen_mmu_set_pgd
xen:xen_mmu_ptep_modify_prot_start
xen:xen_mmu_ptep_modify_prot_commit
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_xen.c | 130 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 131 insertions(+) create mode 100644 tools/lib/traceevent/plugin_xen.c
tools/lib/traceevent/Makefile +1 -0
@@ -219,6 +219,7 @@ PLUGIN_OBJS += plugin_kvm.o PLUGIN_OBJS += plugin_mac80211.o PLUGIN_OBJS += plugin_sched_switch.o PLUGIN_OBJS += plugin_function.o +PLUGIN_OBJS += plugin_xen.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_xen.c new file +130 -0
130 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 14/29 tools lib traceevent: Add scsi plugin
In reply to Jiri OlsaAdding scsi plugin.
This plugin adds fields resolving functions for following tracepoint events:
scsi:scsi_dispatch_cmd_start scsi:scsi_dispatch_cmd_error scsi:scsi_dispatch_cmd_done scsi:scsi_dispatch_cmd_timeout
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'scsi:scsi_dispatch_cmd*' -a)
- swapper 0 [000] 6620.491019: scsi:scsi_dispatch_cmd_done: [FAILED TO PARSE] host_no=0 channel=0 id=0 lun=0 result=0 opcode=53 cmd_len=10 data_sglen=0 prot_sglen=0 prot_op=0 cmnd=5 + swapper 0 [000] 6620.491019: scsi:scsi_dispatch_cmd_done: host_no=0 channel=0 id=0 lun=0 data_sgl=0 prot_sgl=0 prot_op=SCSI_PROT_NORMAL cmnd=(SYNCHRONIZE_CACHE - raw=35 00 00 00 00 00 00 00 00 00) result=(driver=DRIVER_OK host=DID_OK message=COMMAND_COMPLETE status=SAM_STAT_GOOD) - kworker/0:0 21554 [000] 6620.491126: scsi:scsi_dispatch_cmd_start: [FAILED TO PARSE] host_no=0 channel=0 id=0 lun=0 opcode=42 cmd_len=10 data_sglen=1 prot_sglen=0 prot_op=0 cmnd=* + kworker/0:0 21554 [000] 6620.491126: scsi:scsi_dispatch_cmd_start: host_no=0 channel=0 id=0 lun=0 data_sgl=1 prot_sgl=0 prot_op=SCSI_PROT_NORMAL cmnd=(WRITE_10 lba=570899168 txlen=8 protect=0 raw=2a 00 22 07 3a e0 00 00 08 00) - jbd2/dm-3-8 593 [002] 6621.607992: scsi:scsi_dispatch_cmd_error: [FAILED TO PARSE] host_no=0 channel=0 id=0 lun=0 rtn=4182 opcode=53 cmd_len=10 data_sglen=0 prot_sglen=0 prot_op=0 cmnd=5 + jbd2/dm-3-8 593 [002] 6621.607992: scsi:scsi_dispatch_cmd_error: host_no=0 channel=0 id=0 lun=0 data_sgl=0 prot_sgl=0 prot_op=SCSI_PROT_NORMAL cmnd=(SYNCHRONIZE_CACHE - raw=35 00 00 00 00 00 00 00 00 00) rtn=4182
NOTE I couldn't generate scsi_dispatch_cmd_timeout tracepoint,
but it's similar to the rest, so I believe it's ok.
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_scsi.c | 423 +++++++++++++++++++++++++++++++++++++ 2 files changed, 424 insertions(+) create mode 100644 tools/lib/traceevent/plugin_scsi.c
tools/lib/traceevent/Makefile +1 -0
@@ -220,6 +220,7 @@ PLUGIN_OBJS += plugin_mac80211.o PLUGIN_OBJS += plugin_sched_switch.o PLUGIN_OBJS += plugin_function.o PLUGIN_OBJS += plugin_xen.o +PLUGIN_OBJS += plugin_scsi.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_scsi.c new file +423 -0
423 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 15/29 tools lib traceevent: Add cfg80211 plugin
In reply to Jiri OlsaAdding cfg80211 plugin.
This plugin adds handler for __le16_to_cpup function t properly parse following tracepoint events:
cfg80211:cfg80211_tx_mlme_mgmt cfg80211:cfg80211_rx_mlme_mgmt cfg80211:cfg80211_rx_unprot_mlme_mgmt
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'cfg80211:*' -a')
--- script.cfg80211.old
+++ script.cfg80211.new
- ifconfig 2705 [003] 662.896560: cfg80211:cfg80211_tx_mlme_mgmt: [FAILED TO PARSE] name=wlan0 ifindex=3 frame=ARRAY[c0, 00, 00, 00, 00, 3a, 98, a0, 30, 51, 10, 0b, a9, c6, f4, 74, 00, 3a, 98, a0, 30, 51, 00, 00, 03, 00] + ifconfig 2705 [003] 662.896560: cfg80211:cfg80211_tx_mlme_mgmt: netdev:wlan0(3), ftype:0xc0 - kworker/u16:0 1697 [002] 664.808210: cfg80211:cfg80211_rx_mlme_mgmt: [FAILED TO PARSE] name=wlan0 ifindex=3 frame=ARRAY[b0, 00, da, 00, 10, 0b, a9, c6, f4, 74, d8, 24, bd, a1, 26, 31, d8, 24, bd, a1, 26, 31, 10, b7, 00, 00, 02, 00, 00, 00] + kworker/u16:0 1697 [002] 664.808210: cfg80211:cfg80211_rx_mlme_mgmt: netdev:wlan0(3), ftype:0xb0
--- tools/lib/traceevent/Makefile | 1 + tools/lib/traceevent/plugin_cfg80211.c | 24 ++++++++++++++++++++++++ 2 files changed, 25 insertions(+) create mode 100644 tools/lib/traceevent/plugin_cfg80211.c
tools/lib/traceevent/Makefile +1 -0
@@ -221,6 +221,7 @@ PLUGIN_OBJS += plugin_sched_switch.o PLUGIN_OBJS += plugin_function.o PLUGIN_OBJS += plugin_xen.o PLUGIN_OBJS += plugin_scsi.o +PLUGIN_OBJS += plugin_cfg80211.o PLUGINS := $(PLUGIN_OBJS:.o=.so)
tools/lib/traceevent/plugin_cfg80211.c new file +24 -0
@@ -0,0 +1,24 @@ +#include <stdio.h> +#include <string.h> +#include <inttypes.h> +#include <endian.h> +#include "event-parse.h" + +static unsigned long long +process___le16_to_cpup (struct trace_seq *s, +unsigned long long *args) +{ +uint16_t *val = (uint16_t *) args[0 ]; +return val ? (long long )le16toh (*val) :0 ; +} + +int PEVENT_PLUGIN_LOADER (struct pevent *pevent) +{ +pevent_register_print_function (pevent, + process___le16_to_cpup, +PEVENT_FUNC_ARG_INT , +"__le16_to_cpup" , +PEVENT_FUNC_ARG_PTR , +PEVENT_FUNC_ARG_VOID ); +return 0 ; +}
Signature · 1 line
1.8.3.1
PATCH 16/29 tools lib traceevent: Remove malloc_or_die from event-plugin.c
In reply to Jiri OlsaRemoving malloc_or_die calls from event-plugin.c, replacing them with standard malloc and error path.
--- tools/lib/traceevent/event-plugin.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-)
tools/lib/traceevent/event-plugin.c +16 -3
@@ -47,7 +47,11 @@ load_plugin(struct pevent *pevent, const char *path,char *plugin;void *handle; - plugin =malloc_or_die (strlen (path) +strlen (file) +2 ); + plugin =malloc (strlen (path) +strlen (file) +2 ); +if (!plugin) { +warning ("could not allocate plugin memory\n" ); +return ; + }strcpy (plugin, path);strcat (plugin,"/" );
@@ -71,7 +75,12 @@ load_plugin(struct pevent *pevent, const char *path, goto out_free; } - list =malloc_or_die (sizeof (*list)); + list =malloc (sizeof (*list)); +if (!list) { +warning ("could not allocate plugin memory\n" ); + goto out_free; + } + list->next = *plugin_list; list->handle = handle; list->name = plugin;
@@ -163,7 +172,11 @@ load_plugins(struct pevent *pevent, const char *suffix,if (!home)return ; - path =malloc_or_die (strlen (home) +strlen (LOCAL_PLUGIN_DIR ) +2 ); + path =malloc (strlen (home) +strlen (LOCAL_PLUGIN_DIR ) +2 ); +if (!path) { +warning ("could not allocate plugin memory\n" ); +return ; + }strcpy (path, home);strcat (path,"/" );
Signature · 1 line
1.8.3.1
PATCH 17/29 tools lib traceevent: Use static functions in jbd2 plugin
In reply to Jiri OlsaThere's no need for following functions to be global:
process_jbd2_dev_to_name process_jiffies_to_msecs
Making them static.
--- tools/lib/traceevent/plugin_jbd2.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
tools/lib/traceevent/plugin_jbd2.c +6 -4
@@ -29,8 +29,9 @@#define MAJOR (dev) ((unsigned int) ((dev) >> MINORBITS))#define MINOR (dev) ((unsigned int) ((dev) & MINORMASK)) -unsigned long long process_jbd2_dev_to_name (struct trace_seq *s, -unsigned long long *args) +static unsigned long long +process_jbd2_dev_to_name (struct trace_seq *s, +unsigned long long *args) {unsigned int dev = args[0 ];
@@ -38,8 +39,9 @@ unsigned long long process_jbd2_dev_to_name(struct trace_seq *s,return 0 ; } -unsigned long long process_jiffies_to_msecs (struct trace_seq *s, -unsigned long long *args) +static unsigned long long +process_jiffies_to_msecs (struct trace_seq *s, +unsigned long long *args) {unsigned long long jiffies = args[0 ];
Signature · 1 line
1.8.3.1
PATCH 18/29 tools lib traceevent: Use pevent_print_func_field in hrtimer_start handler
In reply to Jiri OlsaThe pevent_print_func_field function encompasses all the functionality used in the hrtimer_start handler. Changing the handler to use this function.
This also unifies the function field output with the hrtimer_expire_entry handler.
--- tools/lib/traceevent/plugin_hrtimer.c | 19 ++----------------- 1 file changed, 2 insertions(+), 17 deletions(-)
tools/lib/traceevent/plugin_hrtimer.c +2 -17
@@ -48,10 +48,6 @@ static int timer_start_handler(struct trace_seq *s,struct pevent_record *record, struct event_format *event, void *context) { - struct pevent *pevent = event->pevent; -struct format_field *fn =pevent_find_field (event,"function" ); -void *data = record->data; - trace_seq_printf(s, "hrtimer="); if (pevent_print_num_field(s, "0x%llx", event, "timer",
@@ -59,19 +55,8 @@ static int timer_start_handler(struct trace_seq *s,pevent_print_num_field (s,"0x%llx" , event,"hrtimer" , record,1 ); -if (!fn) { -trace_seq_printf (s," function=MISSING" ); - }else { -unsigned long long function; -const char *func; - -if (pevent_read_number_field (fn, data, &function)) -trace_seq_printf (s," function=INVALID" ); - - func =pevent_find_function (pevent, function); - -trace_seq_printf (s," function=%s" , func); - } +pevent_print_func_field (s," function=%s" , event,"function" , + record,0 );trace_seq_printf (s," expires=" );pevent_print_num_field (s,"%llu" , event,"expires" , record,1 );
Signature · 1 line
1.8.3.1
PATCH 19/29 tools lib traceevent: Several cleanups for function plugin
In reply to Jiri OlsaSeveral cleanups suggested by Namhyung:
* removing index field from struct func_stack as it's not needed
* renaming get_index into add_and_get_index
* using '%*X' format string capability instead of the loop
--- tools/lib/traceevent/plugin_function.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
tools/lib/traceevent/plugin_function.c +4 -6
@@ -25,7 +25,6 @@#include "event-utils.h" static struct func_stack { -int index;int size;char **stack; } *fstack;
@@ -57,7 +56,7 @@ static void add_child(struct func_stack *stack, const char *child, int pos) stack->stack[pos] =strdup (child); } -static int get_index (const char *parent,const char *child,int cpu) +static int add_and_get_index (const char *parent,const char *child,int cpu) {int i;
@@ -97,7 +96,7 @@ static int function_handler(struct trace_seq *s, struct pevent_record *record,unsigned long long pfunction;const char *func;const char *parent; -int i, index; +int index;if (pevent_get_field_val (s, event,"ip" , record, &function,1 ))return trace_seq_putc (s,'!' );
@@ -109,10 +108,9 @@ static int function_handler(struct trace_seq *s, struct pevent_record *record, parent =pevent_find_function (pevent, pfunction); - index =get_index (parent, func, record->cpu); + index =add_and_get_index (parent, func, record->cpu); -for (i =0 ; i < index; i++) -trace_seq_printf (s," " ); +trace_seq_printf (s,"%*s" , index*3 ,"" );if (func)trace_seq_printf (s,"%s" , func);
Signature · 1 line
1.8.3.1
PATCH 20/29 tools lib traceevent: Remove malloc_or_die from plugin_function.c
In reply to Jiri OlsaRemoving malloc_or_die calls from plugin_function.c, replacing them with standard malloc and error path.
--- tools/lib/traceevent/plugin_function.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)
tools/lib/traceevent/plugin_function.c +13 -2
@@ -44,10 +44,16 @@ static void add_child(struct func_stack *stack, const char *child, int pos)free (stack->stack[pos]); else {if (!stack->stack) - stack->stack =malloc_or_die (sizeof (char *) *STK_BLK ); + stack->stack =malloc (sizeof (char *) *STK_BLK );else stack->stack =realloc (stack->stack,sizeof (char *) * (stack->size +STK_BLK )); + +if (!stack->stack) { +warning ("could not allocate plugin memory\n" ); +return ; + } +for (i = stack->size; i < stack->size +STK_BLK ; i++) stack->stack[i] =NULL ; stack->size +=STK_BLK ;
@@ -67,7 +73,12 @@ static int add_and_get_index(const char *parent, const char *child, int cpu)if (fstack) fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 ));else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); + fstack =malloc (sizeof (*fstack) * (cpu +1 )); + +if (!fstack) { +warning ("could not allocate plugin memory\n" ); +return 0 ; + }/* Account for holes in the cpu count */ for (i = cpus +1 ; i <= cpu; i++)
Signature · 1 line
1.8.3.1
Namhyung Kim
maintainerRe: [PATCH 20/29] tools lib traceevent: Remove malloc_or_die from plugin_function.c
Hi Jiri,
Quoted 23 lines · Jiri Olsa
Removing
malloc_or_diecalls fromplugin_function.c, replacing them with standard malloc and error path.Suggested-by Namhyung Kim <[email protected]>Signed-off-by Jiri Olsa <[email protected]>Cc Corey Ashford <[email protected]>Cc Frederic Weisbecker <[email protected]>Cc Ingo Molnar <[email protected]>Cc Namhyung Kim <[email protected]>Cc Paul Mackerras <[email protected]>Cc Peter Zijlstra <[email protected]>Cc Arnaldo Carvalho de Melo <[email protected]>Cc Steven Rostedt <[email protected]>Cc David Ahern <[email protected]>--- tools/lib/traceevent/plugin_function.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)tools/lib/traceevent/plugin_function.c +0
-0
tools/lib/traceevent/plugin_function.c static void add_child(struct func_stack *stack, const char *child, int pos) · line 50
Quoted 8 lines
tools/lib/traceevent/plugin_function.c +1
-1@@ -44,10 +44,16 @@ static void add_child(struct func_stack *stack, const char *child, int pos)free (stack->stack[pos]); else {if (!stack->stack) - stack->stack =malloc_or_die (sizeof (char *) *STK_BLK ); + stack->stack =malloc (sizeof (char *) *STK_BLK );else stack->stack =realloc (stack->stack,sizeof (char *) * (stack->size +STK_BLK ));
I think single realloc() can handle both cases. And this code has a problem that it overwrites stack->stack to NULL in case of error so that we cannot point original region anymore. You'd better to use a temp variable IMHO.
Quoted 10 lines
tools/lib/traceevent/plugin_function.c +6
-0+ +if (!stack->stack) { +warning ("could not allocate plugin memory\n" ); +return ; + } +for (i = stack->size; i < stack->size +STK_BLK ; i++) stack->stack[i] =NULL ; stack->size +=STK_BLK ;
tools/lib/traceevent/plugin_function.c static int add_and_get_index(const char *parent, const char *child, int cpu) · line 76
Quoted 5 lines
tools/lib/traceevent/plugin_function.c +1
-1@@ -67,7 +73,12 @@ static int add_and_get_index(const char *parent, const char *child, int cpu)if (fstack) fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 ));else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); + fstack =malloc (sizeof (*fstack) * (cpu +1 ));
Ditto.
Thanks,
Namhyung
Quoted 8 lines
tools/lib/traceevent/plugin_function.c +5
-0+ +if (!fstack) { +warning ("could not allocate plugin memory\n" ); +return 0 ; + }/* Account for holes in the cpu count */ for (i = cpus +1 ; i <= cpu; i++)
PATCH v3 20/29 tools lib traceevent: Remove malloc_or_die from plugin_function.c
Quoted 31 lines · Namhyung Kim
Hi Jiri,
Quoted 28 lines · Jiri Olsa
Removing
malloc_or_diecalls fromplugin_function.c, replacing them with standard malloc and error path.Suggested-by Namhyung Kim <[email protected]>Signed-off-by Jiri Olsa <[email protected]>Cc Corey Ashford <[email protected]>Cc Frederic Weisbecker <[email protected]>Cc Ingo Molnar <[email protected]>Cc Namhyung Kim <[email protected]>Cc Paul Mackerras <[email protected]>Cc Peter Zijlstra <[email protected]>Cc Arnaldo Carvalho de Melo <[email protected]>Cc Steven Rostedt <[email protected]>Cc David Ahern <[email protected]>--- tools/lib/traceevent/plugin_function.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)tools/lib/traceevent/plugin_function.c +1
-1@@ -44,10 +44,16 @@ static void add_child(struct func_stack *stack, const char *child, int pos)free (stack->stack[pos]); else {if (!stack->stack) - stack->stack =malloc_or_die (sizeof (char *) *STK_BLK ); + stack->stack =malloc (sizeof (char *) *STK_BLK );
Quoted 8 lines
tools/lib/traceevent/plugin_function.c static void add_child(struct func_stack *stack, const char *child, int pos) · line 50
Quoted 3 lines
tools/lib/traceevent/plugin_function.c +0
-0else stack->stack =realloc (stack->stack,sizeof (char *) * (stack->size +STK_BLK ));I think single
realloc()can handle both cases. And this code has a problem that it overwritesstack->stackto NULL in case of error so that we cannot point original region anymore. You'd better to use a temp variable IMHO.
right you are, please check v3 attached..
thanks,
jirka
---
Removing malloc_or_die calls from plugin_function.c, replacing them and factoring the code with standard realloc and error path.
--- tools/lib/traceevent/plugin_function.c | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-)
tools/lib/traceevent/plugin_function.c +16 -9
@@ -43,11 +43,17 @@ static void add_child(struct func_stack *stack, const char *child, int pos)if (pos < stack->size)free (stack->stack[pos]);else { -if (!stack->stack) - stack->stack =malloc_or_die (sizeof (char *) *STK_BLK ); -else - stack->stack =realloc (stack->stack,sizeof (char *) * - (stack->size +STK_BLK )); +char **ptr; + + ptr =realloc (stack->stack,sizeof (char *) * + (stack->size +STK_BLK )); +if (!ptr) { +warning ("could not allocate plugin memory\n" ); +return ; + } + + stack->stack = ptr; +for (i = stack->size; i < stack->size +STK_BLK ; i++) stack->stack[i] =NULL ; stack->size +=STK_BLK ;
@@ -64,10 +70,11 @@ static int add_and_get_index(const char *parent, const char *child, int cpu)return 0 ;if (cpu > cpus) { -if (fstack) - fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 )); -else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); + fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 )); +if (!fstack) { +warning ("could not allocate plugin memory\n" ); +return 0 ; + }/* Account for holes in the cpu count */ for (i = cpus +1 ; i <= cpu; i++)
Signature · 1 line
1.8.3.1
Namhyung Kim
maintainerRe: [PATCHv3 20/29] tools lib traceevent: Remove malloc_or_die from plugin_function.c
Quoted 3 lines
Removing
malloc_or_diecalls fromplugin_function.c, replacing them and factoring the code with standard realloc and error path.
[SNIP]
tools/lib/traceevent/plugin_function.c static int add_and_get_index(const char *parent, const char *child, int cpu) · line 73
Quoted 7 lines
if (cpu > cpus) {tools/lib/traceevent/plugin_function.c +1
-4-if (fstack) - fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 )); -else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); + fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 ));
This code also has same problem..
Thanks,
Namhyung
Quoted 10 lines
tools/lib/traceevent/plugin_function.c +4
-1+if (!fstack) { +warning ("could not allocate plugin memory\n" ); +return 0 ; + }/* Account for holes in the cpu count */ for (i = cpus +1 ; i <= cpu; i++) --1.8.3.1
PATCH v3 20/29 tools lib traceevent: Remove malloc_or_die from plugin_function.c
Quoted 6 lines · Namhyung Kim
Quoted 3 lines
Removing
malloc_or_diecalls fromplugin_function.c, replacing them and factoring the code with standard realloc and error path.[SNIP]
Quoted 1 line
Quoted 9 lines
Review
Quoted 6 lines
if (cpu > cpus) {(quoted diff) +1
-4- if (fstack) - fstack = realloc(fstack, sizeof(*fstack) * (cpu + 1)); - else - fstack = malloc_or_die(sizeof(*fstack) * (cpu + 1)); + fstack = realloc(fstack, sizeof(*fstack) * (cpu + 1));This code also has same problem..
and v3.1 ;-) thanks
jirka
---
Removing malloc_or_die calls from plugin_function.c, replacing them and factoring the code with standard realloc and error path.
--- tools/lib/traceevent/plugin_function.c | 29 ++++++++++++++++++++--------- 1 file changed, 20 insertions(+), 9 deletions(-)
tools/lib/traceevent/plugin_function.c +20 -9
@@ -43,11 +43,17 @@ static void add_child(struct func_stack *stack, const char *child, int pos)if (pos < stack->size)free (stack->stack[pos]);else { -if (!stack->stack) - stack->stack =malloc_or_die (sizeof (char *) *STK_BLK ); -else - stack->stack =realloc (stack->stack,sizeof (char *) * - (stack->size +STK_BLK )); +char **ptr; + + ptr =realloc (stack->stack,sizeof (char *) * + (stack->size +STK_BLK )); +if (!ptr) { +warning ("could not allocate plugin memory\n" ); +return ; + } + + stack->stack = ptr; +for (i = stack->size; i < stack->size +STK_BLK ; i++) stack->stack[i] =NULL ; stack->size +=STK_BLK ;
@@ -64,10 +70,15 @@ static int add_and_get_index(const char *parent, const char *child, int cpu)return 0 ;if (cpu > cpus) { -if (fstack) - fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 )); -else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); +struct func_stack *ptr; + + ptr =realloc (fstack,sizeof (*fstack) * (cpu +1 )); +if (!ptr) { +warning ("could not allocate plugin memory\n" ); +return 0 ; + } + + fstack = ptr;/* Account for holes in the cpu count */ for (i = cpus +1 ; i <= cpu; i++)
Signature · 1 line
1.8.3.1
Namhyung Kim
maintainerRe: [PATCHv3 20/29] tools lib traceevent: Remove malloc_or_die from plugin_function.c
Quoted 14 lines · Jiri Olsa
Quoted 13 lines · Namhyung Kim
Quoted 3 lines
Removing
malloc_or_diecalls fromplugin_function.c, replacing them and factoring the code with standard realloc and error path.[SNIP]
Quoted 7 lines
if (cpu > cpus) {tools/lib/traceevent/plugin_function.c +1
-4-if (fstack) - fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 )); -else - fstack =malloc_or_die (sizeof (*fstack) * (cpu +1 )); + fstack =realloc (fstack,sizeof (*fstack) * (cpu +1 ));
Quoted 4 lines
Quoted 2 lines
This code also has same problem..
and v3.1 ;-) thanks
Thank you for your effort.
Now the whole series looks all good to me (with other changes fixed).
Thanks,
Namhyung
PATCH 21/29 tools lib traceevent: Update kvm plugin with is_writable_pte helper
In reply to Jiri OlsaAdding is_writable_pte print helper function, so the kvmmmu:fast_page_fault print format gets resolved properly.
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e 'kvm:*,kvmmmu:*' -a')
--- script.kvm.old
+++ script.kvm.new
- qemu-system-x86 3290 [002] 10708.755312: kvmmmu:fast_page_fault: [FAILED TO PARSE] vcpu_id=0 gva=4094486080 error_code=3 sptep=0xffff88019f1e3670 old_spte=336391285 new_spte=336391287 retry=1
+ qemu-system-x86 3290 [002] 10708.755312: kvmmmu:fast_page_fault: vcpu 0 gva f40ce640 error_code P|W sptep 0xffff88019f1e3670 old 0x140cec75 new 140cec77 spurious 0 fixed 1
--- tools/lib/traceevent/plugin_kvm.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
tools/lib/traceevent/plugin_kvm.c +17 -0
@@ -389,6 +389,16 @@ static int kvm_mmu_get_page_handler(struct trace_seq *s,return kvm_mmu_print_role (s, record, event, context); } +#define PT_WRITABLE_SHIFT 1 +#define PT_WRITABLE_MASK (1ULL << PT_WRITABLE_SHIFT) + +static unsigned long long +process_is_writable_pte (struct trace_seq *s,unsigned long long *args) +{ +unsigned long pte = args[0 ]; +return pte &PT_WRITABLE_MASK ; +} +int PEVENT_PLUGIN_LOADER (struct pevent *pevent) {init_disassembler ();
@@ -415,5 +425,12 @@ int PEVENT_PLUGIN_LOADER(struct pevent *pevent)pevent_register_event_handler (pevent,-1 ,"kvmmmu" ,"kvm_mmu_prepare_zap_page" , kvm_mmu_print_role,NULL ); + +pevent_register_print_function (pevent, + process_is_writable_pte, +PEVENT_FUNC_ARG_INT , +"is_writable_pte" , +PEVENT_FUNC_ARG_LONG , +PEVENT_FUNC_ARG_VOID );return 0 ; }
Signature · 1 line
1.8.3.1
PATCH 22/29 tools lib traceevent: Change pevent_parse_format to include pevent handle
In reply to Jiri OlsaChanging the pevent_parse_format interface to include the pevent handle.
The goal is to always use pevent object when dealing with traceevent library. The reason is that we might need additional processing (like plugins), which is not possible otherwise.
Patches follow to make this happen completely.
--- tools/lib/traceevent/event-parse.c | 56 ++++++++++++++++++++++++-------------- tools/lib/traceevent/event-parse.h | 4 ++- tools/perf/util/evsel.c | 2 +- 3 files changed, 39 insertions(+), 23 deletions(-)
tools/lib/traceevent/event-parse.c +35 -21
@@ -5129,8 +5129,38 @@ enum pevent_errno __pevent_parse_format(struct event_format **eventp,return ret; } +static enum pevent_errno +__pevent_parse_event (struct pevent *pevent, +struct event_format **eventp, +const char *buf,unsigned long size, +const char *sys) +{ +int ret =__pevent_parse_format (eventp, pevent, buf, size, sys); +struct event_format *event = *eventp; + +if (event ==NULL ) +return ret; + +if (pevent &&add_event (pevent, event)) { + ret =PEVENT_ERRNO__MEM_ALLOC_FAILED ; + goto event_add_failed; + } + +#define PRINT_ARGS 0 +if (PRINT_ARGS && event->print_fmt.args) +print_args (event->print_fmt.args); + +return 0 ; + +event_add_failed: +pevent_free_format (event); +return ret; +} +/** * pevent_parse_format - parse the event format +* @pevent: the handle to the pevent +* @eventp: returned format * @buf: the buffer storing the event format string * @size: the size of @buf * @sys: the system the event belongs to
@@ -5142,10 +5172,12 @@ enum pevent_errno __pevent_parse_format(struct event_format **eventp,* * /sys/kernel/debug/tracing/events/.../.../format */ -enum pevent_errno pevent_parse_format (struct event_format **eventp,const char *buf, +enum pevent_errno pevent_parse_format (struct pevent *pevent, +struct event_format **eventp, +const char *buf,unsigned long size,const char *sys) { -return __pevent_parse_format (eventp,NULL , buf, size, sys); +return __pevent_parse_event (pevent, eventp, buf, size, sys); }/**
@@ -5166,25 +5198,7 @@ enum pevent_errno pevent_parse_event(struct pevent *pevent, const char *buf, unsigned long size, const char *sys) { struct event_format *event = NULL; -int ret =__pevent_parse_format (&event, pevent, buf, size, sys); - -if (event ==NULL ) -return ret; - -if (add_event (pevent, event)) { - ret =PEVENT_ERRNO__MEM_ALLOC_FAILED ; - goto event_add_failed; - } - -#define PRINT_ARGS 0 -if (PRINT_ARGS && event->print_fmt.args) -print_args (event->print_fmt.args); - -return 0 ; - -event_add_failed: -pevent_free_format (event); -return ret; +return __pevent_parse_event (pevent, &event, buf, size, sys); }#undef _PE
tools/lib/traceevent/event-parse.h +3 -1
@@ -562,7 +562,9 @@ int pevent_parse_header_page(struct pevent *pevent, char *buf, unsigned long sizenum pevent_errno pevent_parse_event (struct pevent *pevent,const char *buf,unsigned long size,const char *sys); -enum pevent_errno pevent_parse_format (struct event_format **eventp,const char *buf, +enum pevent_errno pevent_parse_format (struct pevent *pevent, +struct event_format **eventp, +const char *buf,unsigned long size,const char *sys);void pevent_free_format (struct event_format *event);
tools/perf/util/evsel.c +1 -1
@@ -210,7 +210,7 @@ struct event_format *event_format__new(const char *sys, const char *name) size += n; }while (n >0 ); -pevent_parse_format (&format, bf, size, sys); +pevent_parse_format (NULL , &format, bf, size, sys); out_free_bf:free (bf);
Signature · 1 line
1.8.3.1
PATCH 23/29 tools lib traceevent: Harmonize the install messages in lib-traceevent
In reply to Jiri OlsaRemoving the 'to ...' part out of the install message, because it does not fit to the rest of the build messages we use.
Before:
INSTALL plugin_hrtimer.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_jbd2.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_kmem.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_kvm.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_mac80211.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_sched_switch.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_function.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_xen.so to /home/jolsa/libexec/perf-core/traceevent/plugins INSTALL plugin_scsi.so to /home/jolsa/libexec/perf-core/traceevent/plugins
Now:
INSTALL plugin_jbd2.so INSTALL plugin_hrtimer.so INSTALL plugin_kmem.so INSTALL plugin_kvm.so INSTALL plugin_mac80211.so INSTALL plugin_sched_switch.so INSTALL plugin_function.so INSTALL plugin_xen.so INSTALL plugin_scsi.so
--- tools/lib/traceevent/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/lib/traceevent/Makefile +1 -1
@@ -166,7 +166,7 @@ else print_plugin_obj_compile = echo ' CC FPIC '$(OBJ); print_plugin_build = echo ' BUILD PLUGIN '$(OBJ); print_static_lib_build = echo ' BUILD STATIC LIB '$(OBJ); - print_install = echo ' INSTALL '$1' to $(DESTDIR_SQ)$2'; + print_install = echo ' INSTALL '$1; endif do_fpic_compile = \
Signature · 1 line
1.8.3.1
PATCH 24/29 perf tools: Add build and install plugins targets
In reply to Jiri OlsaAdding 'plugins' target along with the libtraceevent.a, so plugins are built together with traceevent library.
Adding 'install-traceevent-plugins' Makefile install target, instructing perf to install plugins into:
$(HOME)/.traceevent/plugins
- if installed localy under $HOME
$(DESTDIR)/$(prefix)/$(libdir)/traceevent/plugins
- if installed globally
Examples:
$ make install ... $ find ~/.traceevent/plugins/ /home/jolsa/.traceevent/plugins/ /home/jolsa/.traceevent/plugins/plugin_mac80211.so /home/jolsa/.traceevent/plugins/plugin_kvm.so /home/jolsa/.traceevent/plugins/plugin_scsi.so /home/jolsa/.traceevent/plugins/plugin_sched_switch.so /home/jolsa/.traceevent/plugins/plugin_xen.so /home/jolsa/.traceevent/plugins/plugin_cfg80211.so /home/jolsa/.traceevent/plugins/plugin_function.so /home/jolsa/.traceevent/plugins/plugin_kmem.so /home/jolsa/.traceevent/plugins/plugin_hrtimer.so /home/jolsa/.traceevent/plugins/plugin_jbd2.so
$ sudo make install DESTDIR=/opt/perf/ ... $ find /opt/perf/lib64/traceevent/plugins/ /opt/perf/lib64/traceevent/plugins/ /opt/perf/lib64/traceevent/plugins/plugin_kvm.so /opt/perf/lib64/traceevent/plugins/plugin_scsi.so /opt/perf/lib64/traceevent/plugins/plugin_mac80211.so /opt/perf/lib64/traceevent/plugins/plugin_hrtimer.so /opt/perf/lib64/traceevent/plugins/plugin_kmem.so /opt/perf/lib64/traceevent/plugins/plugin_jbd2.so /opt/perf/lib64/traceevent/plugins/plugin_sched_switch.so /opt/perf/lib64/traceevent/plugins/plugin_function.so /opt/perf/lib64/traceevent/plugins/plugin_cfg80211.so /opt/perf/lib64/traceevent/plugins/plugin_xen.so
--- tools/perf/Makefile.perf | 17 ++++++++++++----- tools/perf/config/Makefile | 8 ++++++++ 2 files changed, 20 insertions(+), 5 deletions(-)
tools/perf/Makefile.perf +12 -5
@@ -105,7 +105,7 @@ ifeq ($(config),1) include config/Makefile endif -export prefix bindir sharedir sysconfdir +export prefix bindir sharedir sysconfdir DESTDIR # sparse is architecture-neutral, which means that we need to tell it # explicitly what architecture to check for. Fix this up for yours..
@@ -710,13 +710,20 @@ $(LIB_FILE): $(LIB_OBJS) # libtraceevent.a TE_SOURCES = $(wildcard $(TRACE_EVENT_DIR)*.[ch]) -$(LIBTRACEEVENT): $(TE_SOURCES) - $(QUIET_SUBDIR0)$(TRACE_EVENT_DIR) $(QUIET_SUBDIR1) O=$(OUTPUT) CFLAGS="-g -Wall $(EXTRA_CFLAGS)" libtraceevent.a +LIBTRACEEVENT_FLAGS = $(QUIET_SUBDIR1) O=$(OUTPUT) +LIBTRACEEVENT_FLAGS += CFLAGS="-g -Wall $(EXTRA_CFLAGS)" +LIBTRACEEVENT_FLAGS += plugin_dir=$(plugindir_SQ) + +$(LIBTRACEEVENT): $(TE_SOURCES) $(OUTPUT)PERF-CFLAGS + $(QUIET_SUBDIR0)$(TRACE_EVENT_DIR) $(LIBTRACEEVENT_FLAGS) libtraceevent.a plugins $(LIBTRACEEVENT)-clean: $(call QUIET_CLEAN, libtraceevent) @$(MAKE) -C $(TRACE_EVENT_DIR) O=$(OUTPUT) clean >/dev/null +install-traceevent-plugins: + $(QUIET_SUBDIR0)$(TRACE_EVENT_DIR) $(LIBTRACEEVENT_FLAGS) install_plugins + LIBLK_SOURCES = $(wildcard $(LK_PATH)*.[ch]) # if subdir is set, we've been called from above so target has been built
@@ -785,7 +792,7 @@ cscope: ### Detect prefix changes TRACK_CFLAGS = $(subst ','\'',$(CFLAGS)):\ - $(bindir_SQ):$(perfexecdir_SQ):$(template_dir_SQ):$(prefix_SQ) + $(bindir_SQ):$(perfexecdir_SQ):$(template_dir_SQ):$(prefix_SQ):$(plugindir_SQ) $(OUTPUT)PERF-CFLAGS: .FORCE-PERF-CFLAGS @FLAGS='$(TRACK_CFLAGS)'; \
@@ -849,7 +856,7 @@ endif $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/tests/attr'; \ $(INSTALL) tests/attr/* '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/tests/attr' -install: install-bin try-install-man +install: install-bin try-install-man install-traceevent-plugins install-python_ext: $(PYTHON_WORD) util/setup.py --quiet install --root='/$(DESTDIR_SQ)'
tools/perf/config/Makefile +8 -0
@@ -597,3 +597,11 @@ else perfexec_instdir = $(prefix)/$(perfexecdir) endif perfexec_instdir_SQ = $(subst ','\'',$(perfexec_instdir)) + +# If we install to $(HOME) we keep the traceevent default: +# $(HOME)/.traceevent/plugins +# Otherwise we install plugins into the global $(libdir). +ifdef DESTDIR +plugindir=$(libdir)/traceevent/plugins +plugindir_SQ= $(subst ','\'',$(prefix)/$(plugindir)) +endif
Signature · 1 line
1.8.3.1
PATCH 25/29 perf tools: Add udis86 disassembler feature check
In reply to Jiri OlsaAdding udis86 disassembler feature check which support is needed for kvm:kvm_emulate_insn tracepoint.
The diff of 'perf script' output generated by old and new code:
(data was generated by 'perf record -e kvm:kvm_emulate_insn -a')
--- script.kvm.old
+++ script.kvm.new
- qemu-system-x86 15519 [003] 5332.470049: kvm:kvm_emulate_insn: 0:ffffffff8103c596:89 b7 00 80 5f ff (prot64)
+ qemu-system-x86 15519 [003] 5332.470049: kvm:kvm_emulate_insn: 0:ffffffff8103c596: mov %esi, -0xa08000(%rdi)
--- tools/perf/Makefile.perf | 2 +- tools/perf/config/Makefile | 8 ++++++++ tools/perf/config/feature-checks/Makefile | 6 +++++- tools/perf/config/feature-checks/test-all.c | 5 +++++ tools/perf/config/feature-checks/test-udis86.c | 11 +++++++++++ 5 files changed, 30 insertions(+), 2 deletions(-) create mode 100644 tools/perf/config/feature-checks/test-udis86.c
tools/perf/Makefile.perf +1 -1
@@ -711,7 +711,7 @@ $(LIB_FILE): $(LIB_OBJS) TE_SOURCES = $(wildcard $(TRACE_EVENT_DIR)*.[ch]) LIBTRACEEVENT_FLAGS = $(QUIET_SUBDIR1) O=$(OUTPUT) -LIBTRACEEVENT_FLAGS += CFLAGS="-g -Wall $(EXTRA_CFLAGS)" +LIBTRACEEVENT_FLAGS += CFLAGS="-g -Wall $(EXTRA_CFLAGS) $(LIBTRACEEVENT_CFLAGS)" LIBTRACEEVENT_FLAGS += plugin_dir=$(plugindir_SQ) $(LIBTRACEEVENT): $(TE_SOURCES) $(OUTPUT)PERF-CFLAGS
tools/perf/config/Makefile +8 -0
@@ -417,6 +417,14 @@ else msg := $(warning No timerfd support. Disables 'perf kvm stat live'); endif +$(call feature_check,udis86) +ifeq ($(feature-udis86), 1) + LIBTRACEEVENT_CFLAGS += -DHAVE_UDIS86 + EXTLIBS += -ludis86 +else + msg := $(warning No udis86 support.); +endif + disable-python = $(eval $(disable-python_code)) define disable-python_code CFLAGS += -DNO_LIBPYTHON
tools/perf/config/feature-checks/Makefile +5 -1
@@ -26,7 +26,8 @@ FILES= \ test-libunwind-debug-frame \ test-on-exit \ test-stackprotector-all \ - test-timerfd + test-timerfd \ + test-udis86 CC := $(CC) -MD
@@ -140,6 +141,9 @@ test-backtrace: test-timerfd: $(BUILD) +test-udis86: + $(BUILD) -DHAVE_UDIS86 -ludis86 + -include *.d ###############################
tools/perf/config/feature-checks/test-all.c +5 -0
@@ -89,6 +89,10 @@# include "test-stackprotector-all.c" #undef main +#define main main_test_udis86 +# include "test-udis86.c" +#undef main +int main (int argc,char *argv[]) {main_test_libpython ();
@@ -111,6 +115,7 @@ int main(int argc, char *argv[])main_test_libnuma ();main_test_timerfd ();main_test_stackprotector_all (); +main_test_udis86 ();return 0 ; }
tools/perf/config/feature-checks/test-udis86.c new file +11 -0
@@ -0,0 +1,11 @@ + +/* +* test for udis86 functions used by plugin_kvm.c +*/ + +#include <udis86.h> + +int main (void ) +{ +return puts ("hi" ); +}
Signature · 1 line
1.8.3.1
PATCH 26/29 perf tools: Add filename__read_str util function
In reply to Jiri OlsaAdding filename__read_str util function to read text file and return it in the char array.
The interface is:
int filename__read_str(const char *filename, char **buf, size_t *sizep)
Returns 0/-1 if the read suceeded/fail respectively.
buf - place to store the data pointer size - place to store data size
v2 change:
- better error handling suggested by Namhyung Kim.
--- tools/perf/util/util.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/util.h | 1 + 2 files changed, 51 insertions(+)
tools/perf/util/util.c +50 -0
@@ -6,6 +6,9 @@#endif #include <stdio.h> #include <stdlib.h> +#include <string.h> +#include <errno.h> +#include <linux/kernel.h> /* * XXX We need to find a better place for these things...
@@ -413,3 +416,50 @@ int filename__read_int(const char *filename, int *value)close (fd);return err; } + +int filename__read_str (const char *filename,char **buf,size_t *sizep) +{ +size_t size =0 , alloc_size =0 ; +void *bf =NULL , *nbf; +int fd, n, err =0 ; + + fd =open (filename,O_RDONLY ); +if (fd <0 ) +return -errno; + +do { +if (size == alloc_size) { + alloc_size +=BUFSIZ ; + nbf =realloc (bf, alloc_size); +if (!nbf) { + err = -ENOMEM ; +break ; + } + + bf = nbf; + } + + n =read (fd, bf + size, alloc_size - size); +if (n <0 ) { +if (size) { +pr_warning ("read failed %d: %s\n" , + errno,strerror (errno)); + err =0 ; + }else + err = -errno; + +break ; + } + + size += n; + }while (n >0 ); + +if (!err) { + *sizep = size; + *buf = bf; + }else +free (bf); + +close (fd); +return err; +}
tools/perf/util/util.h +1 -0
@@ -307,4 +307,5 @@ char *get_srcline(struct dso *dso, unsigned long addr);void free_srcline (char *srcline);int filename__read_int (const char *filename,int *value); +int filename__read_str (const char *filename,char **buf,size_t *sizep);#endif /* GIT_COMPAT_UTIL_H */
Signature · 1 line
1.8.3.1
PATCH 27/29 perf tools: Add trace-event object
In reply to Jiri OlsaAdd trace-event object to keep together 'struct pevent' object with its loaded plugins with following interface:
int trace_event__init(struct trace_event *t);
- initalizes 'struct pevent' object and loads plugins for it
void trace_event__cleanup(struct trace_event *t);
- cleanups both 'struct pevent' and plugins
--- tools/perf/Makefile.perf | 1 + tools/perf/builtin-script.c | 2 +- tools/perf/util/header.c | 8 ++++---- tools/perf/util/python-ext-sources | 1 + tools/perf/util/session.h | 3 ++- tools/perf/util/trace-event-parse.c | 13 ------------- tools/perf/util/trace-event-read.c | 20 +++++++++++--------- tools/perf/util/trace-event.c | 21 +++++++++++++++++++++ tools/perf/util/trace-event.h | 13 ++++++++++--- 9 files changed, 51 insertions(+), 31 deletions(-) create mode 100644 tools/perf/util/trace-event.c
tools/perf/Makefile.perf +1 -0
@@ -353,6 +353,7 @@ LIB_OBJS += $(OUTPUT)util/pmu-bison.o LIB_OBJS += $(OUTPUT)util/trace-event-read.o LIB_OBJS += $(OUTPUT)util/trace-event-info.o LIB_OBJS += $(OUTPUT)util/trace-event-scripting.o +LIB_OBJS += $(OUTPUT)util/trace-event.o LIB_OBJS += $(OUTPUT)util/svghelper.o LIB_OBJS += $(OUTPUT)util/sort.o LIB_OBJS += $(OUTPUT)util/hist.o
tools/perf/builtin-script.c +1 -1
@@ -1785,7 +1785,7 @@ int cmd_script(int argc, const char **argv, const char *prefix __maybe_unused)return -1 ; } - err = scripting_ops->generate_script (session->pevent, + err = scripting_ops->generate_script (session->tevent.pevent,"perf-script" ); goto out; }
tools/perf/util/header.c +4 -4
@@ -2834,11 +2834,11 @@ int perf_session__read_header(struct perf_session *session) symbol_conf.nr_events = nr_attrs; -perf_header__process_sections (header, fd, &session->pevent, +perf_header__process_sections (header, fd, &session->tevent, perf_file_section__process);if (perf_evlist__prepare_tracepoint_events (session->evlist, - session->pevent)) + session->tevent.pevent)) goto out_delete_evlist;return 0 ;
@@ -3003,7 +3003,7 @@ int perf_event__process_tracing_data(struct perf_tool *tool __maybe_unused,lseek (fd, offset +sizeof (struct tracing_data_event ),SEEK_SET ); - size_read =trace_report (fd, &session->pevent, + size_read =trace_report (fd, &session->tevent, session->repipe); padding =PERF_ALIGN (size_read,sizeof (u64)) - size_read;
@@ -3025,7 +3025,7 @@ int perf_event__process_tracing_data(struct perf_tool *tool __maybe_unused, }perf_evlist__prepare_tracepoint_events (session->evlist, - session->pevent); + session->tevent.pevent);return size_read + padding; }
tools/perf/util/python-ext-sources +1 -0
@@ -18,4 +18,5 @@ util/cgroup.c util/rblist.c util/strlist.c util/fs.c +util/trace-event.c ../../lib/rbtree.c
tools/perf/util/session.h +2 -1
@@ -1,6 +1,7 @@#ifndef __PERF_SESSION_H#define __PERF_SESSION_H +#include "trace-event.h" #include "hist.h" #include "event.h" #include "header.h"
@@ -32,7 +33,7 @@ struct perf_session {struct perf_header header;struct machines machines;struct perf_evlist *evlist; -struct pevent *pevent; +struct trace_event tevent;struct events_stats stats;bool repipe;struct ordered_samples ordered_samples;
tools/perf/util/trace-event-parse.c +0 -13
@@ -28,19 +28,6 @@#include "util.h" #include "trace-event.h" -struct pevent *read_trace_init (int file_bigendian,int host_bigendian) -{ -struct pevent *pevent =pevent_alloc (); - -if (pevent !=NULL ) { -pevent_set_flag (pevent,PEVENT_NSEC_OUTPUT ); -pevent_set_file_bigendian (pevent, file_bigendian); -pevent_set_host_bigendian (pevent, host_bigendian); - } - -return pevent; -} -static int get_common_field (struct scripting_context *context,int *offset,int *size,const char *type) {
tools/perf/util/trace-event-read.c +11 -9
@@ -343,7 +343,7 @@ static int read_event_files(struct pevent *pevent)return 0 ; } -ssize_t trace_report (int fd,struct pevent **ppevent,bool __repipe) +ssize_t trace_report (int fd,struct trace_event *tevent,bool __repipe) {char buf[BUFSIZ ];char test[] = {23 ,8 ,68 };
@@ -356,11 +356,9 @@ ssize_t trace_report(int fd, struct pevent **ppevent, bool __repipe)int host_bigendian;int file_long_size;int file_page_size; -struct pevent *pevent; +struct pevent *pevent =NULL ;int err; - *ppevent =NULL ; - repipe = __repipe; input_fd = fd;
@@ -390,12 +388,17 @@ ssize_t trace_report(int fd, struct pevent **ppevent, bool __repipe) file_bigendian = buf[0 ]; host_bigendian =bigendian (); - pevent =read_trace_init (file_bigendian, host_bigendian); -if (pevent ==NULL ) { -pr_debug ("read_trace_init failed" ); +if (trace_event__init (tevent)) { +pr_debug ("trace_event__init failed" ); goto out; } + pevent = tevent->pevent; + +pevent_set_flag (pevent,PEVENT_NSEC_OUTPUT ); +pevent_set_file_bigendian (pevent, file_bigendian); +pevent_set_host_bigendian (pevent, host_bigendian); +if (do_read (buf,1 ) <0 ) goto out; file_long_size = buf[0 ];
@@ -432,11 +435,10 @@ ssize_t trace_report(int fd, struct pevent **ppevent, bool __repipe)pevent_print_printk (pevent); } - *ppevent = pevent; pevent =NULL ; out:if (pevent) -pevent_free (pevent); +trace_event__cleanup (tevent);return size; }
tools/perf/util/trace-event.c new file +21 -0
@@ -0,0 +1,21 @@ + +#include <traceevent/event-parse.h> +#include "trace-event.h" + +int trace_event__init (struct trace_event *t) +{ +struct pevent *pevent =pevent_alloc (); + +if (pevent) { + t->plugin_list =traceevent_load_plugins (pevent); + t->pevent = pevent; + } + +return pevent ?0 :-1 ; +} + +void trace_event__cleanup (struct trace_event *t) +{ +pevent_free (t->pevent); +traceevent_unload_plugins (t->plugin_list); +}
tools/perf/util/trace-event.h +10 -3
@@ -3,17 +3,24 @@#include <traceevent/event-parse.h> #include "parse-events.h" -#include "session.h" struct machine ;struct perf_sample ;union perf_event ;struct perf_tool ;struct thread ; +struct plugin_list ; + +struct trace_event { +struct pevent *pevent; +struct plugin_list *plugin_list; +}; + +int trace_event__init (struct trace_event *t); +void trace_event__cleanup (struct trace_event *t);int bigendian (void ); -struct pevent *read_trace_init (int file_bigendian,int host_bigendian);void event_format__print (struct event_format *event,int cpu,void *data,int size);
@@ -27,7 +34,7 @@ raw_field_value(struct event_format *event, const char *name, void *data);void parse_proc_kallsyms (struct pevent *pevent,char *file,unsigned int size);void parse_ftrace_printk (struct pevent *pevent,char *file,unsigned int size); -ssize_t trace_report (int fd,struct pevent **pevent,bool repipe); +ssize_t trace_report (int fd,struct trace_event *tevent,bool repipe);struct event_format *trace_find_next_event (struct pevent *pevent,struct event_format *event);
Signature · 1 line
1.8.3.1
PATCH 28/29 perf tools: Add trace-event global object for tracepoint interface
In reply to Jiri OlsaIn order to get the proper plugins processing we need to use full trace-event interface when creating tracepoint events. So far we were using shortcut to get the parsed format.
Moving current 'event_format__new' function into trace-event object as 'trace_event__tp_format'.
This function uses properly initialized global trace-event object, ensuring proper plugins processing.
--- tools/perf/builtin-trace.c | 5 ++-- tools/perf/util/evsel.c | 44 ++----------------------------- tools/perf/util/trace-event.c | 61 +++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/trace-event.h | 2 ++ 4 files changed, 68 insertions(+), 44 deletions(-)
tools/perf/builtin-trace.c +3 -2
@@ -11,6 +11,7 @@#include "util/intlist.h" #include "util/thread_map.h" #include "util/stat.h" +#include "trace-event.h" #include <libaudit.h> #include <stdlib.h>
@@ -1430,11 +1431,11 @@ static int trace__read_syscall_info(struct trace *trace, int id) sc->fmt =syscall_fmt__find (sc->name);snprintf (tp_name,sizeof (tp_name),"sys_enter_%s" , sc->name); - sc->tp_format =event_format__new ("syscalls" , tp_name); + sc->tp_format =trace_event__tp_format ("syscalls" , tp_name);if (sc->tp_format ==NULL && sc->fmt && sc->fmt->alias) {snprintf (tp_name,sizeof (tp_name),"sys_enter_%s" , sc->fmt->alias); - sc->tp_format =event_format__new ("syscalls" , tp_name); + sc->tp_format =trace_event__tp_format ("syscalls" , tp_name); }if (sc->tp_format ==NULL )
tools/perf/util/evsel.c +2 -42
62 lines not on this page · raw message
tools/perf/util/trace-event.c +61 -0
70 lines not on this page · raw message
tools/perf/util/trace-event.h +2 -0
@@ -18,6 +18,8 @@ struct trace_event {int trace_event__init (struct trace_event *t);void trace_event__cleanup (struct trace_event *t); +struct event_format * +trace_event__tp_format (const char *sys,const char *name);int bigendian (void );
Signature · 1 line
1.8.3.1
PATCH 29/29 perf tools: Overload pr_stat traceevent print function
In reply to Jiri OlsaThe traceevent lib uses pr_stat to display all standard info. It's defined as __weak. Overloading it with perf version plugged into perf output system logic.
Displaying the pr_stat stuff under '-v' option.
--- tools/perf/util/debug.c | 30 +++++++++++++++++++++++++++--- tools/perf/util/debug.h | 2 ++ 2 files changed, 29 insertions(+), 3 deletions(-)
tools/perf/util/debug.c +27 -3
@@ -16,13 +16,11 @@int verbose;bool dump_trace = false, quiet = false; -int eprintf (int level,const char *fmt, ...) +static int _eprintf (int level,const char *fmt,va_list args) { -va_list args;int ret =0 ;if (verbose >= level) { -va_start (args, fmt);if (use_browser >=1 )ui_helpline__vshow (fmt, args);else
@@ -33,6 +31,32 @@ int eprintf(int level, const char *fmt, ...)return ret; } +int eprintf (int level,const char *fmt, ...) +{ +va_list args; +int ret; + +va_start (args, fmt); + ret =_eprintf (level, fmt, args); +va_end (args); + +return ret; +} + +/* +* Overloading libtraceevent standard info print +* function, display with -v in perf. +*/ +void pr_stat (const char *fmt, ...) +{ +va_list args; + +va_start (args, fmt); +_eprintf (1 , fmt, args); +va_end (args); +eprintf (1 ,"\n" ); +} +int dump_printf (const char *fmt, ...) {va_list args;
tools/perf/util/debug.h +2 -0
@@ -17,4 +17,6 @@ void trace_event(union perf_event *event);int ui__error (const char *format, ...) __attribute__((format (printf,1 ,2 )));int ui__warning (const char *format, ...) __attribute__((format (printf,1 ,2 ))); +void pr_stat (const char *fmt, ...); +#endif /* __PERF_DEBUG_H */
Signature · 1 line
1.8.3.1