From: Jiri Olsa <jolsa@redhat.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: fweisbec@gmail.com, mingo@redhat.com, paulus@samba.org,
acme@ghostprotocols.net, a.p.zijlstra@chello.nl,
linux-kernel@vger.kernel.org, aarapov@redhat.com
Subject: Re: [PATCHv2 02/10] ftrace: Change mcount call replacement logic
Date: Tue, 20 Dec 2011 14:10:19 +0100 [thread overview]
Message-ID: <20111220131019.GA4393@m.brq.redhat.com> (raw)
In-Reply-To: <1324321380.5916.26.camel@gandalf.stny.rr.com>
On Mon, Dec 19, 2011 at 02:03:00PM -0500, Steven Rostedt wrote:
> On Mon, 2011-12-05 at 18:22 +0100, Jiri Olsa wrote:
> > The current logic of updating calls is to do the mcount replacement
> > only when ftrace_ops is being registered. When ftrace_ops is being
> > unregistered then only in case it was the last registered ftrace_ops,
> > all calls are disabled.
> >
> > This is an issue when ftrace_ops without FTRACE_OPS_FL_GLOBAL flag
> > is being unregistered, because all the functions stays enabled
> > and thus inherrited by global_ops, like in following scenario:
> >
> > - set restricting global filter
> > - enable function trace
> > - register/unregister ftrace_ops with flags != FTRACE_OPS_FL_GLOBAL
> > and with no filter
>
> I don't see this problem. I just changed stack_tracer to have its own
> filter (I've been wanting to do that for a long time, so when I saw this
> email, I decided it's a good time to implement it).
>
> Here's what I did:
>
> # echo schedule > set_ftrace_filter
> # cat set_ftrace_filter
> schedule
> # cat enabled_functions
> schedule (1)
> # echo 1 > /proc/sys/kernel/stack_tracer_enabled
> # cat enabled_functions
> do_one_initcall (1)
> match_dev_by_uuid (1)
> name_to_dev_t (1)
> idle_notifier_unregister (1)
> idle_notifier_register (1)
> start_thread_common.constprop.6 (1)
> enter_idle (1)
> exit_idle (1)
> cpu_idle (1)
> __show_regs (1)
> release_thread (1)
> [...]
> _cond_resched (1)
> preempt_schedule_irq (1)
> schedule (2)
> io_schedule (1)
> yield_to (1)
> yield (1)
>
> // note that schedule is (2)
>
> # echo 0 > /proc/sys/kernel/stack_tracer_enabled
> # cat enabled_functions
> schedule (1)
>
>
> >
> > Now the global_ops will get by all the functions regardless the
> > global_ops filter. So we need all functions that where enabled via
> > this ftrace_ops and are not part of global filter to be disabled.
>
> The global functions are not at issue here. What do you see?
>
> Maybe I fixed something as I'm using the latest tip/perf/core. Note, I
> can send you the stack_tracer patch if you want to take a look at this
> example. I need to clean it up too.
that would be great, thanks
>
> >
> > Note, currently if there are only global ftrace_ops registered,
> > there's no filter hash check and the filter is represented only
> > by enabled records.
> >
> > Changing the ftrace_shutdown logic to ensure the replacement
> > is called for each ftrace_ops being unregistered.
> >
> > Also changing FTRACE_ENABLE_CALLS into FTRACE_UPDATE_CALLS
> > calls which seems more suitable now.
>
> I still think this patch is wrong. What's the problem you are seeing?
let me try with an example..
say we have only 2 traceable functions - A and B ;)
1) set global filter for function A with 'echo A > ./set_ftrace_filter'
a - A is put to the global_ops filter
2) enable function trace with 'echo function > current_tracer'
a - register_ftrace_function is called with function trace ftrace_ops (GLOBAL flag)
b - update_ftrace_function is called, setting ftrace_ops callback function
to be called directly from the assembly entry_* code
c - ftrace_hash_rec_enable is called, and dyn_ftrace record
for function A is updated:
A::flags|FL_MASK = 1
d - ftrace_replace_code(1) is called, and function A is
brought in to life
3) enable function trace via perf ftrace_ops
a - register_ftrace_function is called with perf event ftrace_ops (!GLOBAL flag)
b - update_ftrace_function is called, setting ftrace_ops_list_func
function to be called from the assembly entry_* code and
handle the ftrace_ops' dispatch
c - ftrace_hash_rec_enable is called, and A and B dyn_ftrace records
are updated:
A::flags|FL_MASK = 2
B::flags|FL_MASK = 1
d - ftrace_replace_code(1) is called, and function B is
brought in to life
4) disable function trace via perf ftrace_ops
a - unregister_ftrace_function is called with perf event ftrace_ops (same as in step 3)
b - update_ftrace_function is called, setting global ftrace_ops (from step 2)
callback function to be called directly from the assembly entry_* code
c - ftrace_hash_rec_disable is called, and A and B dyn_ftrace
records are updated:
A::flags|FL_MASK = 1
B::flags|FL_MASK = 0
d - ??? see below..
Now, only the global function trace ftrace_ops is enabled (from step 2),
but both A and B are alive and feeding the tracer despite its filter (function A),
because its ftrace_ops is directly linked to the assembly entry_* code (step 4b).
The reason is that even though we updated the B's dyn_ftrace record (step 4c)
to be 'B::flags|FL_MASK == 0', we did not update the B call itself and disable it.
If we'd call ftrace_replace_code(1) at 4d), the B function would be
disabled, and we would get expected behaviour.
So thats the reason I think we should update the calls (mcount call code)
each time we unregister the ftrace_ops, because some records could be
enabled only for the specific ftrace_ops and we need to put them down
when we disable this ftrace_ops.
hopefully it make any sense.. :)
thanks,
jirka
next prev parent reply other threads:[~2011-12-20 13:10 UTC|newest]
Thread overview: 185+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-11-27 18:04 [RFC] ftrace, perf: Adding support to use function trace Jiri Olsa
2011-11-27 18:04 ` [PATCH 1/9] trace: Fix uninitialized variable compiler warning Jiri Olsa
2011-11-28 16:19 ` Steven Rostedt
2011-11-28 16:25 ` Jiri Olsa
2011-11-28 19:34 ` Steven Rostedt
2011-11-27 18:04 ` [PATCH 2/9] ftrace: Fix possible NULL dereferencing in __ftrace_hash_rec_update Jiri Olsa
2011-11-28 16:24 ` Steven Rostedt
2011-11-27 18:04 ` [PATCH 3/9] ftrace: Fix shutdown to disable calls properly Jiri Olsa
2011-11-28 19:18 ` Steven Rostedt
2011-11-29 11:21 ` Jiri Olsa
2011-11-27 18:04 ` [PATCH 4/9] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2011-11-28 19:26 ` Steven Rostedt
2011-11-28 20:02 ` Peter Zijlstra
2011-11-28 20:05 ` Peter Zijlstra
2011-11-28 20:14 ` Steven Rostedt
2011-11-28 20:20 ` Peter Zijlstra
2011-11-28 20:12 ` Steven Rostedt
2011-11-28 20:15 ` Peter Zijlstra
2011-11-28 20:24 ` Steven Rostedt
2011-11-28 20:21 ` Steven Rostedt
2011-11-29 10:07 ` Jiri Olsa
2011-11-27 18:04 ` [PATCH 5/9] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2011-11-27 18:04 ` [PATCH 6/9] ftrace, perf: Add add/del " Jiri Olsa
2011-11-27 18:04 ` [PATCH 7/9] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2011-11-28 19:58 ` Steven Rostedt
2011-11-28 20:03 ` Peter Zijlstra
2011-11-28 20:13 ` Steven Rostedt
2011-11-29 10:10 ` Jiri Olsa
2011-11-28 20:08 ` Peter Zijlstra
2011-11-28 20:10 ` Peter Zijlstra
2011-11-28 20:16 ` Steven Rostedt
2011-11-28 20:18 ` Peter Zijlstra
2011-11-27 18:04 ` [PATCH 8/9] ftrace, perf: Add FILTER_TRACE_FN event field type Jiri Olsa
2011-11-28 20:01 ` Steven Rostedt
2011-11-29 10:14 ` Jiri Olsa
2011-11-29 11:22 ` Jiri Olsa
2011-11-29 11:51 ` Peter Zijlstra
2011-11-29 12:21 ` Jiri Olsa
2011-11-27 18:04 ` [PATCH 9/9] ftrace, perf: Add filter support for function trace event Jiri Olsa
2011-11-28 20:07 ` Steven Rostedt
2011-12-05 17:22 ` [RFCv2] ftrace, perf: Adding support to use function trace Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 01/10] ftrace: Fix possible NULL dereferencing in __ftrace_hash_rec_update Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 02/10] ftrace: Change mcount call replacement logic Jiri Olsa
2011-12-19 19:03 ` Steven Rostedt
2011-12-20 13:10 ` Jiri Olsa [this message]
2011-12-20 16:33 ` Steven Rostedt
2011-12-20 19:39 ` Steven Rostedt
2011-12-21 9:57 ` Jiri Olsa
2011-12-21 11:34 ` Steven Rostedt
2011-12-21 11:35 ` Steven Rostedt
2011-12-21 11:40 ` Jiri Olsa
2012-01-08 9:13 ` [tip:perf/core] ftrace: Fix unregister ftrace_ops accounting tip-bot for Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 03/10] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2011-12-19 19:19 ` Steven Rostedt
2011-12-19 19:35 ` Steven Rostedt
2011-12-20 14:57 ` Jiri Olsa
2011-12-20 15:25 ` Steven Rostedt
2011-12-20 15:35 ` Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 04/10] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 05/10] ftrace, perf: Add add/del " Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 06/10] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 07/10] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2011-12-19 19:22 ` Steven Rostedt
2011-12-05 17:22 ` [PATCHv2 08/10] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 09/10] ftrace, perf: Add filter support for function trace event Jiri Olsa
2011-12-05 17:22 ` [PATCHv2 10/10] ftrace, graph: Add global_ops filter callback for graph tracing Jiri Olsa
2011-12-19 19:27 ` Steven Rostedt
2011-12-19 13:40 ` [RFCv2] ftrace, perf: Adding support to use function trace Jiri Olsa
2011-12-19 16:45 ` Steven Rostedt
2011-12-19 16:58 ` Frederic Weisbecker
2011-12-21 11:48 ` [PATCHv3 0/8] " Jiri Olsa
2011-12-21 11:48 ` [PATCH 1/8] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2011-12-21 11:48 ` [PATCH 2/8] ftrace: Fix possible NULL dereferencing in __ftrace_hash_rec_update Jiri Olsa
2011-12-21 15:23 ` Steven Rostedt
2011-12-21 11:48 ` [PATCH 3/8] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2011-12-21 16:01 ` Steven Rostedt
2011-12-21 16:43 ` Jiri Olsa
2011-12-21 16:55 ` Steven Rostedt
2012-01-24 1:26 ` Frederic Weisbecker
2011-12-21 11:48 ` [PATCH 4/8] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2011-12-21 11:48 ` [PATCH 5/8] ftrace, perf: Add add/del " Jiri Olsa
2011-12-21 11:48 ` [PATCH 6/8] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2011-12-21 11:48 ` [PATCH 7/8] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2011-12-21 11:48 ` [PATCH 8/8] ftrace, perf: Add filter support for function trace event Jiri Olsa
2011-12-21 18:56 ` [PATCHv4 0/8] ftrace, perf: Adding support to use function trace Jiri Olsa
2011-12-21 18:56 ` [PATCH 1/7] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2011-12-22 0:12 ` Steven Rostedt
2011-12-22 8:01 ` [PATCHv5 " Jiri Olsa
2011-12-21 18:56 ` [PATCH 2/7] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2011-12-21 18:56 ` [PATCH 3/7] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2011-12-21 18:56 ` [PATCH 4/7] ftrace, perf: Add add/del " Jiri Olsa
2011-12-21 18:56 ` [PATCH 5/7] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2011-12-21 18:56 ` [PATCH 6/7] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2011-12-21 18:56 ` [PATCH 7/7] ftrace, perf: Add filter support for function trace event Jiri Olsa
2011-12-21 22:07 ` Frederic Weisbecker
2011-12-22 12:55 ` Jiri Olsa
2011-12-22 15:26 ` [PATCHvFIXED " Jiri Olsa
2011-12-24 2:35 ` Frederic Weisbecker
2011-12-21 19:02 ` [PATCHv4 0/7] ftrace, perf: Adding support to use function trace Jiri Olsa
2012-01-02 9:04 ` [PATCHv5 " Jiri Olsa
2012-01-02 9:04 ` [PATCH 1/7] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2012-02-17 13:46 ` [tip:perf/core] ftrace: Change filter/ notrace " tip-bot for Jiri Olsa
2012-01-02 9:04 ` [PATCH 2/7] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2012-01-17 1:42 ` Frederic Weisbecker
2012-01-17 2:07 ` Steven Rostedt
2012-01-17 2:29 ` Frederic Weisbecker
2012-01-18 13:59 ` Jiri Olsa
2012-01-02 9:04 ` [PATCH 3/7] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2012-01-02 9:04 ` [PATCH 4/7] ftrace, perf: Add add/del " Jiri Olsa
2012-01-02 9:04 ` [PATCH 5/7] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2012-01-02 9:04 ` [PATCH 6/7] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2012-01-02 9:04 ` [PATCH 7/7] ftrace, perf: Add filter support for function trace event Jiri Olsa
2012-01-16 23:59 ` Steven Rostedt
2012-01-18 13:45 ` Jiri Olsa
2012-01-16 8:57 ` [PATCHv5 0/7] ftrace, perf: Adding support to use function trace Jiri Olsa
2012-01-16 16:17 ` Steven Rostedt
2012-01-18 18:44 ` [PATCHv6 " Jiri Olsa
2012-01-18 18:44 ` [PATCH 1/7] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2012-01-19 16:31 ` Frederic Weisbecker
2012-01-18 18:44 ` [PATCH 2/7] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2012-01-20 17:02 ` Frederic Weisbecker
2012-01-25 23:13 ` Steven Rostedt
2012-01-26 2:37 ` Frederic Weisbecker
2012-01-27 10:37 ` Jiri Olsa
2012-01-27 10:38 ` Jiri Olsa
2012-01-27 16:40 ` Frederic Weisbecker
2012-01-27 16:54 ` Jiri Olsa
2012-01-27 17:02 ` Frederic Weisbecker
2012-01-27 17:20 ` Jiri Olsa
2012-01-28 16:39 ` Frederic Weisbecker
2012-01-27 17:21 ` Steven Rostedt
2012-01-18 18:44 ` [PATCH 3/7] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2012-01-18 18:44 ` [PATCH 4/7] ftrace, perf: Add add/del " Jiri Olsa
2012-01-18 18:44 ` [PATCH 5/7] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2012-01-18 18:44 ` [PATCH 6/7] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2012-01-18 18:44 ` [PATCH 7/7] ftrace, perf: Add filter support for function trace event Jiri Olsa
2012-01-18 21:43 ` [PATCHv6 0/7] ftrace, perf: Adding support to use function trace Steven Rostedt
2012-01-28 18:43 ` [PATCHv7 " Jiri Olsa
2012-01-28 18:43 ` [PATCH 1/7] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2012-01-30 5:42 ` Frederic Weisbecker
2012-01-28 18:43 ` [PATCH 2/7] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2012-01-30 5:59 ` Frederic Weisbecker
2012-01-30 9:18 ` Jiri Olsa
2012-02-03 13:42 ` Steven Rostedt
2012-02-03 13:50 ` Jiri Olsa
2012-02-03 14:08 ` Steven Rostedt
2012-02-03 14:22 ` [PATCHv8 0/2] first 2 patches passed review Jiri Olsa
2012-02-03 14:22 ` [PATCH 1/2] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2012-02-03 14:22 ` [PATCH 2/2] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2012-02-04 13:24 ` [PATCHv8 0/2] first 2 patches passed review Frederic Weisbecker
2012-02-03 13:40 ` [PATCH 2/7] ftrace: Add enable/disable ftrace_ops control interface Steven Rostedt
2012-01-28 18:43 ` [PATCH 3/7] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2012-02-02 17:35 ` Frederic Weisbecker
2012-02-03 10:23 ` Jiri Olsa
2012-01-28 18:43 ` [PATCH 4/7] ftrace, perf: Add add/del " Jiri Olsa
2012-02-02 17:42 ` Frederic Weisbecker
2012-01-28 18:43 ` [PATCH 5/7] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2012-02-02 18:14 ` Frederic Weisbecker
2012-02-03 12:54 ` Jiri Olsa
2012-02-03 13:00 ` Jiri Olsa
2012-02-03 14:07 ` Steven Rostedt
2012-02-04 13:21 ` Frederic Weisbecker
2012-02-06 19:35 ` Steven Rostedt
2012-02-03 13:53 ` Steven Rostedt
2012-01-28 18:43 ` [PATCH 6/7] ftrace, perf: Distinguish ftrace function event field type Jiri Olsa
2012-02-03 14:16 ` Steven Rostedt
2012-01-28 18:43 ` [PATCH 7/7] ftrace, perf: Add filter support for function trace event Jiri Olsa
2012-02-07 0:20 ` Jiri Olsa
2012-02-07 19:44 ` [PATCHv8 0/8] ftrace, perf: Adding support to use function trace Jiri Olsa
2012-02-07 19:44 ` [PATCH 1/8] ftrace: Change filter/notrace set functions to return exit code Jiri Olsa
2012-02-07 19:44 ` [PATCH 2/8] ftrace: Add enable/disable ftrace_ops control interface Jiri Olsa
2012-02-07 19:44 ` [PATCH 3/8] ftrace, perf: Add open/close tracepoint perf registration actions Jiri Olsa
2012-02-07 19:44 ` [PATCH 4/8] ftrace, perf: Add add/del " Jiri Olsa
2012-02-07 19:44 ` [PATCH 5/8] ftrace: Add FTRACE_ENTRY_REG macro to allow event registration Jiri Olsa
2012-02-07 19:44 ` [PATCH 6/8] ftrace, perf: Add support to use function tracepoint in perf Jiri Olsa
2012-02-07 19:44 ` [PATCH 7/8] ftrace: Allow to specify filter field type for ftrace events Jiri Olsa
2012-02-07 19:44 ` [PATCH 8/8] ftrace, perf: Add filter support for function trace event Jiri Olsa
2012-02-10 13:27 ` [PATCHv8 0/8] ftrace, perf: Adding support to use function trace Steven Rostedt
2012-02-10 14:45 ` Steven Rostedt
2012-02-10 16:07 ` Jiri Olsa
2012-02-10 16:48 ` Frederic Weisbecker
2012-02-10 18:00 ` Steven Rostedt
2012-02-10 18:05 ` Frederic Weisbecker
2012-02-10 18:23 ` David Ahern
2012-02-13 18:02 ` Steven Rostedt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20111220131019.GA4393@m.brq.redhat.com \
--to=jolsa@redhat.com \
--cc=a.p.zijlstra@chello.nl \
--cc=aarapov@redhat.com \
--cc=acme@ghostprotocols.net \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=paulus@samba.org \
--cc=rostedt@goodmis.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).