Describe the return values of tracepoint_probe_register(), including
-ENODEV added by commit:

Author: Steven Rostedt <rostedt@goodmis.org>

    tracing: Warn if a tracepoint is not set via debugfs

Link: http://lkml.kernel.org/r/1394499898-1537-2-git-send-email-mathieu.desnoyers@efficios.com

CC: Ingo Molnar <mingo@kernel.org>
CC: Frederic Weisbecker <fweisbec@gmail.com>
CC: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: NMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Describe the @data argument (probe private data).

Link: http://lkml.kernel.org/r/1394587948-27878-1-git-send-email-mathieu.desnoyers@efficios.com

Fixes: 38516ab5 "tracing: Let tracepoints have data passed to tracepoint callbacks"
CC: Ingo Molnar <mingo@kernel.org>
CC: Frederic Weisbecker <fweisbec@gmail.com>
CC: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: NMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

With CONFIG_DYNAMIC_FTRACE=n, I see a warning:
kernel/trace/ftrace.c:240:13: warning: 'control_ops_free' defined but not used
 static void control_ops_free(struct ftrace_ops *ops)
             ^
Move that function around to an already existing #ifdef
CONFIG_DYNAMIC_FTRACE block as the function is used solely from the
dynamic function tracing functions.

Link: http://lkml.kernel.org/r/1394484131-5107-1-git-send-email-jslaby@suse.czSigned-off-by: NJiri Slaby <jslaby@suse.cz>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Ftrace modifies function calls using Int3 breakpoints on x86.
The breakpoints are handled only when the patching is in progress.
If something goes wrong, there is a recovery code that removes
the breakpoints. If this fails, the system might get silently
rebooted when a remaining break is not handled or an invalid
instruction is proceed.

We should BUG() when the breakpoint could not be removed. Otherwise,
the system silently crashes when the function finishes the Int3
handler is disabled.

Note that we need to modify remove_breakpoint() to return non-zero
value only when there is an error. The return value was ignored before,
so it does not cause any troubles.

Link: http://lkml.kernel.org/r/1393258342-29978-4-git-send-email-pmladek@suse.czSigned-off-by: NPetr Mladek <pmladek@suse.cz>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

We should print some warning and kill ftrace functionality when the ftrace
function is not set correctly. Otherwise, ftrace might do crazy things without
an explanation. The error value has been ignored so far.

Note that an error that happens during updating all the traced calls is handled
in ftrace_replace_code(). We print more details about the particular
failing address via ftrace_bug() there.

Link: http://lkml.kernel.org/r/1393258342-29978-3-git-send-email-pmladek@suse.czSigned-off-by: NPetr Mladek <pmladek@suse.cz>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The 'freelist' member was introduced to 'struct dyn_ftrace' in commit
ee000b7f (tracing: use union for
multi-usages field), but the use of this member was later removed in
32082309 (ftrace: Remove usage of
"freed" records). Remove also the 'freelist' member now.

Link: http://lkml.kernel.org/r/1393268401-24379-5-git-send-email-jslaby@suse.czSigned-off-by: NJiri Slaby <jslaby@suse.cz>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

As the data parameter is not really used by any ftrace_dyn_arch_init,
remove that from ftrace_dyn_arch_init. This also removes the addr
local variable from ftrace_init which is now unused.

Note the documentation was imprecise as it did not suggest to set
(*data) to 0.

Link: http://lkml.kernel.org/r/1393268401-24379-4-git-send-email-jslaby@suse.cz

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: linux-arch@vger.kernel.org
Signed-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

No architecture uses the "data" parameter in ftrace_dyn_arch_init() in any
way, it just sets the value to 0. And this is used as a return value
in the caller -- ftrace_init, which just checks the retval against
zero.

Note there is also "return 0" in every ftrace_dyn_arch_init.  So it is
enough to check the retval and remove all the indirect sets of data on
all archs.

Link: http://lkml.kernel.org/r/1393268401-24379-3-git-send-email-jslaby@suse.cz

Cc: linux-arch@vger.kernel.org
Signed-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The function used to do allocations some time ago. This no longer
happens and it only checks the count and prints some info. This patch
inlines the body to the only caller. There are two reasons:
* the name of the function was misleading
* it's clear what is going on in ftrace_init now

Link: http://lkml.kernel.org/r/1393268401-24379-2-git-send-email-jslaby@suse.czSigned-off-by: NJiri Slaby <jslaby@suse.cz>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Some of them can be local to functions, so make them local and pass
them as parameters where needed:
* __start_mcount_loc+__stop_mcount_loc are local to ftrace_init
* ftrace_new_pgs -> new_pgs/start_pg
* ftrace_update_cnt -> local update_cnt in ftrace_update_code

Link: http://lkml.kernel.org/r/1393268401-24379-1-git-send-email-jslaby@suse.czSigned-off-by: NJiri Slaby <jslaby@suse.cz>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Use a temporary variable to store the expansion of the len expression.
If the evaluation is expensive, this commit will ensure it is evaluated
only once inside ftrace_get_offsets_<call>.

Link: http://lkml.kernel.org/r/1393651938-16418-3-git-send-email-filbranden@google.comSigned-off-by: NFilipe Brandenburger <filbranden@google.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

This fixes expansion of the len argument in __dynamic_array macros.
The previous code from commit 7d536cb3 would not fully evaluate the
expression before multiplying its result by the size of the type.

This went unnoticed because the length stored in the high 16 bits of the
offset (which is the one that was broken here) is only used by
filter_pred_strloc which only acts on strings for which the size of the
type is 1.

Link: http://lkml.kernel.org/r/1393651938-16418-2-git-send-email-filbranden@google.comSigned-off-by: NFilipe Brandenburger <filbranden@google.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

There's nothing in the module.h header that requires tracepoint.h to be
included, and there may be cases that tracepoint.h may need to include
module.h, which will cause recursive header issues.

But module.h requires seeing HAVE_JUMP_LABEL which is set in jump_label.h
which it just coincidentally gets from tracepoint.h.

Link: http://lkml.kernel.org/r/20140307084712.5c68641a@gandalf.local.homeAcked-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The trace event headers are required to include tracepoint.h. The only reason
they worked now is because module.h included tracepoint.h, and that will soon
change.

Link: http://lkml.kernel.org/r/20140226190644.591040764@goodmis.org

Fixes: 7b2a2d4a "mm: migrate: Add a tracepoint for migrate_pages"
Acked-by: NMel Gorman <mgorman@suse.de>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The trace event headers are required to include tracepoint.h. The only reason
they worked now is because module.h included tracepoint.h, and that will soon
change.

Link: http://lkml.kernel.org/r/20140226190644.442886305@goodmis.org

Fixes: 455b2864 "writeback: Initial tracing support"
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Tracepoints were made to allow enabling a tracepoint in a module before that
module was loaded. When a tracepoint is enabled and it does not exist, the
name is stored and will be enabled when the tracepoint is created.

The problem with this approach is that when a tracepoint is enabled when
it expects to be there, it gives no warning that it does not exist.

To add salt to the wound, if a module is added and sets the FORCED flag, which
can happen if it isn't signed properly, the tracepoint code will not enabled
the tracepoints, but they will be created in the debugfs system! When a user
goes to enable the tracepoint, the tracepoint code will not see it existing
and will think it is to be enabled later AND WILL NOT GIVE A WARNING.

The tracing will look like it succeeded but will actually be doing nothing.
This will cause lots of confusion and headaches for developers trying to
figure out why they are not seeing their tracepoints.

Link: http://lkml.kernel.org/r/20140213154507.4040fb06@gandalf.local.homeReported-by: NMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reported-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The functions that assign the contents for the ftrace events are
defined by the TRACE_EVENT() macros. Each event has its own unique
way to assign data to its buffer. When you have over 500 events,
that means there's 500 functions assigning data uniquely for each
event (not really that many, as DECLARE_EVENT_CLASS() and multiple
DEFINE_EVENT()s will only need a single function).

By making helper functions in the core kernel to do some of the work
instead, we can shrink the size of the kernel down a bit.

With a kernel configured with 502 events, the change in size was:

   text    data     bss     dec     hex filename
12987390        1913504 9785344 24686238        178ae9e /tmp/vmlinux
12959102        1913504 9785344 24657950        178401e /tmp/vmlinux.patched

That's a total of 28288 bytes, which comes down to 56 bytes per event.

Link: http://lkml.kernel.org/r/20120810034708.370808175@goodmis.orgSigned-off-by: NSteven Rostedt <rostedt@goodmis.org>

The code that shows array fields for events is defined for all events.
This can add up quite a bit when you have over 500 events.

By making helper functions in the core kernel to do the work
instead, we can shrink the size of the kernel down a bit.

With a kernel configured with 502 events, the change in size was:

   text    data     bss     dec     hex filename
12990946        1913568 9785344 24689858        178bcc2 /tmp/vmlinux
12987390        1913504 9785344 24686238        178ae9e /tmp/vmlinux.patched

That's a total of 3556 bytes, which comes down to 7 bytes per event.
Although it's not much, this code is just called at initialization of
the events.

Link: http://lkml.kernel.org/r/20120810034708.084036335@goodmis.orgSigned-off-by: NSteven Rostedt <rostedt@goodmis.org>

The code for trace events to format the raw recorded event data
into human readable format in the 'trace' file is repeated for every
event in the system. When you have over 500 events, this can add up
quite a bit.

By making helper functions in the core kernel to do the work
instead, we can shrink the size of the kernel down a bit.

With a kernel configured with 502 events, the change in size was:

   text    data     bss     dec     hex filename
12991007        1913568 9785344 24689919        178bcff /tmp/vmlinux.orig
12990946        1913568 9785344 24689858        178bcc2 /tmp/vmlinux.patched

Note, this version does not save as much as the version of this patch
I had a few years ago. That is because in the mean time, commit
f71130de ("tracing: Add a helper function for event print functions")
did a lot of the work my original patch did. But this change helps
slightly, and is part of a larger clean up to reduce the size much further.

Link: http://lkml.kernel.org/r/20120810034707.378538034@goodmis.org

Cc: Li Zefan <lizefan@huawei.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Having ftrace_write() return -EPERM on failure, as that's what the callers
return, then we can clean up the code a bit. That is, instead of:

  if (ftrace_write(...))
     return -EPERM;
  return 0;

or

  if (ftrace_write(...)) {
     ret = -EPERM;
     goto_out;
  }

We can instead have:

  return ftrace_write(...);

or

  ret = ftrace_write(...);
  if (ret)
    goto out;
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

No reason to allocate tp_module structures for modules that have no
tracepoints. This just wastes memory.

Fixes: b75ef8b4 "Tracepoint: Dissociate from module mutex"
Acked-by: NMathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

If a failure occurs while modifying ftrace function, it bails out and will
remove the tracepoints to be back to what the code originally was.

There is missing the final sync run across the CPUs after the fix up is done
and before the ftrace int3 handler flag is reset.

Here's the description of the problem:

	CPU0				CPU1
	----				----
  remove_breakpoint();
  modifying_ftrace_code = 0;

				[still sees breakpoint]
				<takes trap>
				[sees modifying_ftrace_code as zero]
				[no breakpoint handler]
				[goto failed case]
				[trap exception - kernel breakpoint, no
				 handler]
				BUG()

Link: http://lkml.kernel.org/r/1393258342-29978-2-git-send-email-pmladek@suse.cz

Fixes: 8a4d0a68 "ftrace: Use breakpoint method to update ftrace caller"
Acked-by: NFrederic Weisbecker <fweisbec@gmail.com>
Acked-by: NH. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: NPetr Mladek <pmladek@suse.cz>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

If a failure occurs while enabling a trace, it bails out and will remove
the tracepoints to be back to what the code originally was. But the fix
up had some bugs in it. By injecting a failure in the code, the fix up
ran to completion, but shortly afterward the system rebooted.

There was two bugs here.

The first was that there was no final sync run across the CPUs after the
fix up was done, and before the ftrace int3 handler flag was reset. That
means that other CPUs could still see the breakpoint and trigger on it
long after the flag was cleared, and the int3 handler would think it was
a spurious interrupt. Worse yet, the int3 handler could hit other breakpoints
because the ftrace int3 handler flag would have prevented the int3 handler
from going further.

Here's a description of the issue:

	CPU0				CPU1
	----				----
  remove_breakpoint();
  modifying_ftrace_code = 0;

				[still sees breakpoint]
				<takes trap>
				[sees modifying_ftrace_code as zero]
				[no breakpoint handler]
				[goto failed case]
				[trap exception - kernel breakpoint, no
				 handler]
				BUG()

The second bug was that the removal of the breakpoints required the
"within()" logic updates instead of accessing the ip address directly.
As the kernel text is mapped read-only when CONFIG_DEBUG_RODATA is set, and
the removal of the breakpoint is a modification of the kernel text.
The ftrace_write() includes the "within()" logic, where as, the
probe_kernel_write() does not. This prevented the breakpoint from being
removed at all.

Link: http://lkml.kernel.org/r/1392650573-3390-1-git-send-email-pmladek@suse.czReported-by: NPetr Mladek <pmladek@suse.cz>
Tested-by: NPetr Mladek <pmladek@suse.cz>
Acked-by: NH. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The ENABLED flag needs to be cleared when a ftrace_ops is unregistered
otherwise it wont be able to be registered again.

This is only for static tracing and does not affect DYNAMIC_FTRACE at
all.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Being able to change the trace clock at boot can be advantageous if
you need a better source of when things happen across CPUs. The default
trace clock is the fastest, but it uses local clocks which may not be
synced across CPUs and it does not let you know when events took place
with respect to events on other CPUs.

The global trace clock can help in this case, and if you do not care
about timings, the counter "clock" is the best, as that is just a  simple
atomic counter that is incremented for every event.

Usage is to add "trace_clock=counter" on the kernel command line. You
can replace counter with "global" or any of the clocks listed in
/sys/kernel/debug/tracing/trace_clock
Suggested-by: NThomas Gleixner <tglx@linutronix.de>
Tested-by: NThomas Gleixner <tglx@linutronix.de>
Reviewed-by: NThomas Gleixner <tglx@linutronix.de>
Appreciated-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

It seems there's no reason to prevent mixed used of ftrace and perf
for a single uprobe event.  At least the kprobes already support it.

Link: http://lkml.kernel.org/r/1389946120-19610-6-git-send-email-namhyung@kernel.orgReviewed-by: NMasami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Add support for event triggering to uprobes.  This is same as kprobes
support added by Tom (plus cleanup by Steven).

Link: http://lkml.kernel.org/r/1389946120-19610-5-git-send-email-namhyung@kernel.orgReviewed-by: NMasami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
Cc: Tom Zanussi <tom.zanussi@linux.intel.com>
Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Support multi-buffer on uprobe-based dynamic events by
using ftrace_event_file.

This patch is based kprobe-based dynamic events multibuffer
support work initially, commited by Masami(commit 41a7dd42),
but revised as below:

Oleg changed the kprobe-based multibuffer design from
array-pointers of ftrace_event_file into simple list,
so this patch also change to the list design.

rcu_read_lock/unlock added into uprobe_trace_func/uretprobe_trace_func,
to synchronize with ftrace_event_file list add and delete.

Even though we allow multi-uprobes instances now,
but TP_FLAG_PROFILE/TP_FLAG_TRACE are still mutually exclusive
in probe_event_enable currently, this means we cannot allow
one user is using uprobe-tracer, and another user is using
perf-probe on same uprobe concurrently.
(Perhaps this will be fix in future, kprobe don't have this
limitation now)

Link: http://lkml.kernel.org/r/1389946120-19610-4-git-send-email-namhyung@kernel.orgReviewed-by: NMasami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Reviewed-by: NOleg Nesterov <oleg@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Nzhangwei(Jovi) <jovi.zhangwei@huawei.com>
Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

A single uprobe event might serve different users like ftrace and
perf.  And this is especially important for upcoming multi buffer
support.  But in this case it'll fetch (same) data from userspace
multiple times.  So move it to the beginning of the dispatcher
function and reuse it for each users.

Link: http://lkml.kernel.org/r/1389946120-19610-3-git-send-email-namhyung@kernel.orgReviewed-by: NMasami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

The uprobe_{trace,perf}_print functions are misnomers since what they
do is not printing.  There's also a real print function named
print_uprobe_event() so they'll only increase confusion IMHO.

Rename them with double underscores to follow convention of kprobe.

Link: http://lkml.kernel.org/r/1389946120-19610-2-git-send-email-namhyung@kernel.orgReviewed-by: NMasami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Create a "set_ftrace_filter" and "set_ftrace_notrace" files in the instance
directories to let users filter of functions to trace for the given instance.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

In preparation for having the function tracing instances be able to
filter on functions, the generic filter functions must first be
converted to take in the global_ops as a parameter.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Allow instances (sub-buffers) to enable function tracing.
Each instance will have its own function tracing capability.
For now, instances will not have function stack tracing, or will
they be able to pick and choose what functions they can trace.

Picking and choosing their own functions will come later.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

As tracers will soon be used by instances, the tracer enabled field
needs to be converted to a counter instead of a boolean.
This counter is protected by the trace_types_lock mutex.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

When an instance is about to be deleted, make sure the tracer
is set to nop. If it isn't reset the tracer and set it to the nop
tracer, otherwise memory leaks and bad pointers may result.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

If global_ops function is being called directly, instead of the global_ops
list function, set the global_ops private to be the same as the ops private
that's being called directly.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Passing data to the function callback was originally done by adding the
ftrace_ops in another structure, and using the container_of() to get
the field. But this adds a bit more complexity than it is worth, and
adding a simple .private field to ftrace_ops makes things a lot easier.

But be warned, the .private data should not be freed once it is used
unless the ftrace_ops itself has gone through the necessary freeing
routines. A simple synchronize_sched() is not enough as functions
can be traced that are called outside the view of RCU and all its
concoctions.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Currently, only the top level instance can have tracing options.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Currently the tracers (function, function_graph, irqsoff, etc) can only
be used by the top level tracing directory (not for instances).

This sets up the infrastructure to allow instances to be able to
run a separate tracer apart from the what the top level tracing is
doing.

As tracers need to adapt for being used by instances, the tracers
must flag if they can be used by instances or not. Currently only the
'nop' tracer can be used by all instances.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

As options (flags) may affect instances instead of being global
the flag_changed() callbacks need to receive the trace_array descriptor
of the instance they will be modifying.
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>