1. 02 Jun, 2009 3 commits
    • Steven Rostedt's avatar
      tracing: combine the default tracers into one config · 897f17a6
      Steven Rostedt authored
      Both event tracer and sched switch plugin are selected by default
      by all generic tracers. But if no generic tracer is enabled, their options
      appear. But ether one of them will select the other, thus it only
      makes sense to have the default tracers be selected by one option.
      [ Impact: clean up kconfig menu ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
    • Steven Rostedt's avatar
      tracing: fix config options to not show when automatically selected · 5e0a0939
      Steven Rostedt authored
      There are two options that are selected by all tracers, but we want
      to have those options available when no tracer is selected. These are
       The event tracer and sched switch tracer.
      The are enabled by all tracers, but if a tracer is not selected we want
      the options to appear. All tracers including them select TRACING.
      Thus what we would like to do is:
        config EVENT_TRACER
      	bool "prompt"
      	depends on TRACING
      	select TRACING
      But that gives us a bug in the kbuild system since we just created a
      circular dependency. We only want the prompt to show when TRACING is off.
      This patch adds GENERIC_TRACER that all tracers will select instead of
      TRACING. The two options (sched switch and event tracer) will select
      TRACING directly and depend on !GENERIC_TRACER. This solves the cicular
      [ Impact: hide options that are selected by default ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
    • Steven Rostedt's avatar
      ftrace: add kernel command line function filtering · 2af15d6a
      Steven Rostedt authored
      When using ftrace=function on the command line to trace functions
      on boot up, one can not filter out functions that are commonly called.
      This patch adds two new ftrace command line commands.
      Where function-list is a comma separated list of functions to filter.
      The ftrace_notrace will make the functions listed not be included
      in the function tracing, and ftrace_filter will only trace the functions
      These two act the same as the debugfs/tracing/set_ftrace_notrace and
      debugfs/tracing/set_ftrace_filter respectively.
      The simple glob expressions that are allowed by the filter files can also
      be used by the command line interface.
      Will not trace any function that starts with rcu, ends with lock, or has
      the word spin in it.
      Note, if the self tests are enabled, they may interfere with the filtering
      set by the command lines.
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
  2. 27 May, 2009 2 commits
    • Zhaolei's avatar
      ftrace: don't convert function's local variable name in macro · f2aebaee
      Zhaolei authored
      "call" is an argument of macro, but it is also used as a local
      variable name of function in macro.
      We should keep this local variable name distinct from any
      CPP macro parameter name if both are in the same macro scope,
      although it hasn't caused any problem yet.
      [ Impact: robustify macro ]
      Signed-off-by: default avatarZhao Lei <zhaolei@cn.fujitsu.com>
      Acked-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Heiko Carstens's avatar
      trace: disable preemption before taking raw spinlocks · 5b6045a9
      Heiko Carstens authored
      s390 code uses smp_processor_id() in __raw_spin_lock() code which
      reveals that a (raw) spinlock is taken without preemption disabled.
      This can potentially deadlock.
      To fix this explicitly disable and enable preemption.
      BUG: using smp_processor_id() in preemptible [00000000] code: cat/2278
      caller is trace_find_cmdline+0x40/0xfc
      CPU: 0 Not tainted 2.6.30-rc7-dirty #39
      Process cat (pid: 2278, task: 000000003faedb68, ksp: 000000003b33b988)
      000000003b33b988 000000003b33bae0 0000000000000002 0000000000000000
             000000003b33bb80 000000003b33baf8 000000003b33baf8 00000000000175d6
             0000000000000001 000000003b33b988 000000003f9b0000 000000000000000b
             000000000000000c 000000003b33bb40 000000003b33bae0 0000000000000000
             0000000000000000 00000000000175d6 000000003b33bae0 000000003b33bb28
      Call Trace:
      ([<00000000000174b2>] show_trace+0x112/0x170)
       [<0000000000017582>] show_stack+0x72/0x100
       [<0000000000441538>] dump_stack+0xc8/0xd8
       [<000000000025c350>] debug_smp_processor_id+0x114/0x130
       [<00000000000bf0e4>] trace_find_cmdline+0x40/0xfc
       [<00000000000c35d4>] trace_print_context+0x58/0xac
       [<00000000000bb676>] print_trace_line+0x416/0x470
       [<00000000000bc8fe>] s_show+0x4e/0x428
       [<000000000013834e>] seq_read+0x36a/0x5d4
       [<0000000000112a78>] vfs_read+0xc8/0x174
       [<0000000000112c58>] SyS_read+0x74/0xc4
       [<000000000002c7ae>] sysc_noemu+0x10/0x16
       [<000002000012436c>] 0x2000012436c
      1 lock held by cat/2278:
       #0:  (&p->lock){+.+.+.}, at: [<0000000000138056>] seq_read+0x72/0x5d4
      [ Impact: fix preempt-unsafe raw spinlock ]
      Signed-off-by: default avatarHeiko Carstens <heiko.carstens@de.ibm.com>
      Acked-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
  3. 26 May, 2009 8 commits
    • Steven Rostedt's avatar
      tracing: convert irq events to use __print_symbolic · c2adae09
      Steven Rostedt authored
      The recording of the names at trace time is inefficient. This patch
      implements the softirq event recording to only record the vector
      and then use the __print_symbolic interface to print out the names.
      [ Impact: faster recording of softirq events ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Steven Rostedt's avatar
      tracing: add __print_symbolic to trace events · 0f4fc29d
      Steven Rostedt authored
      This patch adds __print_symbolic which is similar to __print_flags but
      works for an enumeration type instead. That is, there is only a one to one
      mapping between the values and the symbols. When a match is made, then
      it is printed, otherwise the hex value is outputed.
      [ Impact: add interface for showing symbol names in events ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Steven Rostedt's avatar
      tracing: add flag output for kmem events · 62ba180e
      Steven Rostedt authored
      This patch changes the output for gfp_flags from being a simple hex value
      to the actual names.
        gfp_flags=GFP_ATOMIC  instead of gfp_flags=00000020
      And even
        gfp_flags=GFP_KERNEL instead of gfp_flags=000000d0
      (Thanks to Frederic Weisbecker for pointing out that the first version
       had a bad order of GFP masks)
      [ Impact: more human readable output from tracer ]
      Acked-by: default avatarEduard - Gabriel Munteanu <eduard.munteanu@linux360.ro>
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Steven Rostedt's avatar
      tracing: add previous task state info to sched switch event · 937cdb9d
      Steven Rostedt authored
      It is useful to see the state of a task that is being switched out.
      This patch adds the output of the state of the previous task in
      the context switch event.
      [ Impact: see state of switched out task in context switch ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Steven Rostedt's avatar
      tracing: add __print_flags for events · be74b73a
      Steven Rostedt authored
      Developers have been asking for the ability in the ftrace event tracer
      to display names of bits in a flags variable.
      Instead of printing out c2, it would be easier to read FOO|BAR|GOO,
      assuming that FOO is bit 1, BAR is bit 6 and GOO is bit 7.
      Some examples where this would be useful are the state flags in a context
      switch, kmalloc flags, and even permision flags in accessing files.
        v2 changes include:
        Frederic Weisbecker's idea of using a mask instead of bits,
        thus we can output GFP_KERNEL instead of GPF_WAIT|GFP_IO|GFP_FS.
        Li Zefan's idea of allowing the caller of __print_flags to add their
        own delimiter (or no delimiter) where we can get for file permissions
        rwx instead of r|w|x.
        v3 changes:
         Christoph Hellwig's idea of using an array instead of va_args.
      [ Impact: better displaying of flags in trace output ]
      Signed-off-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Zhaolei's avatar
      ftrace: clean up of using ftrace_event_enable_disable() · 0e907c99
      Zhaolei authored
      Always use ftrace_event_enable_disable() to enable/disable an event
      so that we can factorize out the event toggling code.
      [ Impact: factorize and cleanup event tracing code ]
      Signed-off-by: default avatarZhao Lei <zhaolei@cn.fujitsu.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      Cc: Tom Zanussi <tzanussi@gmail.com>
      LKML-Reference: <4A14FDFE.2080402@cn.fujitsu.com>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Zhaolei's avatar
      ftrace: Add task_comm support for trace_event · b11c53e1
      Zhaolei authored
      If we enable a trace event alone without any tracer running (such as
      function tracer, sched switch tracer, etc...) it can't output enough
      task command information.
      We need to use the tracing_{start/stop}_cmdline_record() helpers
      which are designed to keep track of cmdlines for any tasks that
      were scheduled during the tracing.
      Before this patch:
       # echo 1 > debugfs/tracing/events/sched/sched_switch/enable
       # cat debugfs/tracing/trace
       # tracer: nop
       #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
       #              | |       |          |         |
                  <...>-2289  [000] 526276.724790: sched_switch: task bash:2289 [120] ==> sshd:2287 [120]
                  <...>-2287  [000] 526276.725231: sched_switch: task sshd:2287 [120] ==> bash:2289 [120]
                  <...>-2289  [000] 526276.725452: sched_switch: task bash:2289 [120] ==> sshd:2287 [120]
                  <...>-2287  [000] 526276.727181: sched_switch: task sshd:2287 [120] ==> swapper:0 [140]
                 <idle>-0     [000] 526277.032734: sched_switch: task swapper:0 [140] ==> events/0:5 [115]
                  <...>-5     [000] 526277.032782: sched_switch: task events/0:5 [115] ==> swapper:0 [140]
      After this patch:
       # tracer: nop
       #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
       #              | |       |          |         |
                   bash-2269  [000] 527347.989229: sched_switch: task bash:2269 [120] ==> sshd:2267 [120]
                   sshd-2267  [000] 527347.990960: sched_switch: task sshd:2267 [120] ==> bash:2269 [120]
                   bash-2269  [000] 527347.991143: sched_switch: task bash:2269 [120] ==> sshd:2267 [120]
                   sshd-2267  [000] 527347.992959: sched_switch: task sshd:2267 [120] ==> swapper:0 [140]
                 <idle>-0     [000] 527348.531989: sched_switch: task swapper:0 [140] ==> events/0:5 [115]
               events/0-5     [000] 527348.532115: sched_switch: task events/0:5 [115] ==> swapper:0 [140]
      v1->v2: Update Kconfig to select CONTEXT_SWITCH_TRACER in
      v2->v3: v2 can solve problem that was caused by config EVENT_TRACING
              alone, but when CONFIG_FTRACE is off and CONFIG_TRACING is
              selected by other config, compile fail happened again.
              This version solves it.
      [ Impact: fix incomplete output of event tracing ]
      Signed-off-by: default avatarZhao Lei <zhaolei@cn.fujitsu.com>
      Cc: Tom Zanussi <tzanussi@gmail.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      LKML-Reference: <4A14FDFE.2080402@cn.fujitsu.com>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Pekka Enberg's avatar
      kmemtrace: fix kernel parameter documentation · 29fcefba
      Pekka Enberg authored
      The kmemtrace.enable kernel parameter no longer works. To enable
      kmemtrace at boot-time, you must pass "ftrace=kmemtrace" instead.
      [ Impact: remove obsolete kernel parameter documentation ]
      Cc: Eduard - Gabriel Munteanu <eduard.munteanu@linux360.ro>
      Signed-off-by: default avatarPekka Enberg <penberg@cs.helsinki.fi>
      LKML-Reference: <alpine.DEB.2.00.0905241112190.10296@rocky>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
  4. 25 May, 2009 2 commits
    • Li Zefan's avatar
      tracing/events: change the type of __str_loc_item to unsigned short · b0aae68c
      Li Zefan authored
      When defining a dynamic size string, we add __str_loc_##item to the
      trace entry, and it stores the location of the actual string in
      'unsigned short' should be sufficient to store this information, thus
      we save 2 bytes per dyn-size string in the ring buffer.
      [ Impact: reduce memory occupied by dyn-size strings in ring buffer ]
      Signed-off-by: default avatarLi Zefan <lizf@cn.fujitsu.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      LKML-Reference: <4A14EDB6.2050507@cn.fujitsu.com>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
    • Lai Jiangshan's avatar
      tracing: add trace_event_read_lock() · 4f535968
      Lai Jiangshan authored
      I found that there is nothing to protect event_hash in
      ftrace_find_event(). Rcu protects the event hashlist
      but not the event itself while we use it after its extraction
      through ftrace_find_event().
      This lack of a proper locking in this spot opens a race
      window between any event dereferencing and module removal.
      --Task A--
      print_trace_line(trace) {
        event = find_ftrace_event(trace)
      --Task B--
      trace_module_remove_events(mod) {
        list_trace_events_module(ev, mod) {
          unregister_ftrace_event(ev->event) {
      |--> module removed, the event has been dropped
      --Task A--
        event->print(trace); // Dereferencing freed memory
      If the event retrieved belongs to a module and this module
      is concurrently removed, we may end up dereferencing a data
      from a freed module.
      RCU could solve this, but it would add latency to the kernel and
      forbid tracers output callbacks to call any sleepable code.
      So this fix converts 'trace_event_mutex' to a read/write semaphore,
      and adds trace_event_read_lock() to protect ftrace_find_event().
      [ Impact: fix possible freed memory dereference in ftrace ]
      Signed-off-by: default avatarLai Jiangshan <laijs@cn.fujitsu.com>
      Acked-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      LKML-Reference: <4A114806.7090302@cn.fujitsu.com>
      Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
  5. 20 May, 2009 1 commit
  6. 19 May, 2009 2 commits
    • Stefan Raspl's avatar
      blktrace: remove debugfs entries on bad path · fd51d251
      Stefan Raspl authored
      debugfs directory entries for devices are not removed on some
      of the failure pathes in do_blk_trace_setup().
      One way to reproduce is to start blktrace on multiple devices
      with insufficient Vmalloc space: Devices will fail with
      a message like this:
      	BLKTRACESETUP(2) /dev/sdu failed: 5/Input/output error
      If so, the respective entries in debugfs
      (e.g. /sys/kernel/debug/block/sdu) will remain and subsequent
      attempts to start blktrace on the respective devices will not
      succeed due to existing directories.
      [ Impact: fix /debug/tracing file cleanup corner case ]
      Signed-off-by: default avatarStefan Raspl <stefan.raspl@linux.vnet.ibm.com>
      Acked-by: default avatarLi Zefan <lizf@cn.fujitsu.com>
      Cc: Li Zefan <lizf@cn.fujitsu.com>
      Cc: schwidefsky@de.ibm.com
      Cc: heiko.carstens@de.ibm.com
      LKML-Reference: <4A1266CC.5040801@linux.vnet.ibm.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
    • Li Zefan's avatar
      tracing/events: Documentation updates · 143c145e
      Li Zefan authored
      - fix some typos
      - document the difference between '>' and '>>'
      - document the 'enable' toggle
      - remove section "Defining an event-enabled tracepoint", since it's
        out-dated and sample/trace_events/ already serves this purpose.
      v2: add "Updated by Li Zefan"
      [ Impact: make documentation up-to-date ]
      Signed-off-by: default avatarLi Zefan <lizf@cn.fujitsu.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      Cc: Frederic Weisbecker <fweisbec@gmail.com>
      Cc: "Theodore Ts'o" <tytso@mit.edu>
      LKML-Reference: <4A125503.5060406@cn.fujitsu.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
  7. 18 May, 2009 2 commits
  8. 16 May, 2009 1 commit
  9. 15 May, 2009 19 commits