bit-field is not thread-safe nor smp-safe.
struct marker_entry.rcu_pending is not protected by any lock
in rcu-callback free_old_closure().
so we must turn it into a safe type.
detail:
I suppose rcu_pending and ptype are store in struct marker_entry.tmp1
free_old_closure() side: change ptype side:
| load struct marker_entry.tmp1
--------------------------------|--------------------------------
| change ptype bit in tmp1
load struct marker_entry.tmp1 |
change rcu_pending bit in tmp1 |
store tmp1 |
--------------------------------|--------------------------------
| store tmp1
now this result equals that free_old_closure() do not change rcu_pending
bit, bug! This bug will cause redundant rcu_barrier_sched() called.
not too harmful.
----- corresponding:
free_old_closure() side: change ptype side:
load struct marker_entry.tmp1 |
--------------------------------|--------------------------------
| load struct marker_entry.tmp1
change rcu_pending bit in tmp1 |
| change ptype bit in tmp1
| store tmp1
--------------------------------|--------------------------------
store tmp1 |
now this result equals that change ptype side do not change ptype
bit, bug! this bug cause marker_probe_cb() access to invalid memory.
oops!
see also: http://en.wikipedia.org/wiki/Bit_field
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
when the second, third... probe is registered, its format is
not checked, this patch fixes it.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Lai Jiangshan discovered a reentrancy issue with markers and fixed it by
adding synchronize_sched() calls at each registration/unregistraiton.
It works, but it removes the ability to do batch
registration/unregistration and can cause registration of ~100 markers
to take about 30 seconds on a loaded machine (synchronize_sched() is
much slower on such workloads).
This patch implements a version of the fix which won't slow down marker batch
registration/unregistration. It also go back to the original non-synchronized
reg/unreg.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Use the new rcu_read_lock_sched/unlock_sched() in marker code around the call
site instead of preempt_disable/enable(). It helps reviewing the code more
easily.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
unregister bug:
codes using makers are typically calling marker_probe_unregister()
and then destroying the data that marker_probe_func needs(or
unloading this module). This is bug when the corresponding
marker_probe_func is still running(on other cpus),
it is using the destroying/ed data.
we should call synchronize_sched() after marker_update_probes().
reenter bug:
marker_probe_register(), marker_probe_unregister() and
marker_probe_unregister_private_data() are not reentrant safe
functions. these 3 functions release markers_mutex and then
require it again and do "entry->oldptr = old; ...", but entry->oldptr
maybe is using now for these 3 functions may reenter when markers_mutex
is released.
we use synchronize_sched() instead of call_rcu_sched() to fix
this bug. actually we can do:
"
if (entry->rcu_pending)
rcu_barrier_sched();
"
after require markers_mutex again. but synchronize_sched()
is better and simpler. For these 3 functions are not critical path.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul pointed out two incorrect read barriers in the marker handler code in
the path where multiple probes are connected. Those are ordering reads of
"ptype" (single or multi probe marker), "multi" array pointer, and "multi"
array data access.
It should be ordered like this :
read ptype
smp_rmb()
read multi array pointer
smp_read_barrier_depends()
access data referenced by multi array pointer
The code with a single probe connected (optimized case, does not have to
allocate an array) has correct memory ordering.
It applies to kernel 2.6.26.x, 2.6.25.x and linux-next.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
rcu_barrier_sched() and call_rcu_sched() were introduced in 2.6.26 for the
Markers. Change the marker code to use them.
It can be seen as a fix since the marker code was using an ugly,
temporary, #ifdef hack to work around CONFIG_PREEMPT_RCU.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Acked-by: Paul McKenney <paulmck@us.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Denys Vlasenko <vda.linux@googlemail.com> :
> Not in this patch, but I noticed:
>
> #define __trace_mark(name, call_private, format, args...) \
> do { \
> static const char __mstrtab_##name[] \
> __attribute__((section("__markers_strings"))) \
> = #name "\0" format; \
> static struct marker __mark_##name \
> __attribute__((section("__markers"), aligned(8))) = \
> { __mstrtab_##name, &__mstrtab_##name[sizeof(#name)], \
> 0, 0, marker_probe_cb, \
> { __mark_empty_function, NULL}, NULL }; \
> __mark_check_format(format, ## args); \
> if (unlikely(__mark_##name.state)) { \
> (*__mark_##name.call) \
> (&__mark_##name, call_private, \
> format, ## args); \
> } \
> } while (0)
>
> In this call:
>
> (*__mark_##name.call) \
> (&__mark_##name, call_private, \
> format, ## args); \
>
> you make gcc allocate duplicate format string. You can use
> &__mstrtab_##name[sizeof(#name)] instead since it holds the same string,
> or drop ", format," above and "const char *fmt" from here:
>
> void (*call)(const struct marker *mdata, /* Probe wrapper */
> void *call_private, const char *fmt, ...);
>
> since mdata->format is the same and all callees which need it can take it there.
Very good point. I actually thought about dropping it, since it would
remove an unnecessary argument from the stack. And actually, since I now
have the marker_probe_cb sitting between the marker site and the
callbacks, there is no API change required. Thanks :)
Mathieu
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
CC: Denys Vlasenko <vda.linux@googlemail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
With the needlessly global marker_debug being static gcc can optimize the
unused code away.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Following an experimental deletion of the unnecessary directive
#include <linux/slab.h>
from the header file <linux/percpu.h>, these files under kernel/ were exposed
as needing to include one of <linux/slab.h> or <linux/gfp.h>, so explicit
includes were added where necessary.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Markers do not mix well with CONFIG_PREEMPT_RCU because it uses
preempt_disable/enable() and not rcu_read_lock/unlock for minimal
intrusiveness. We would need call_sched and sched_barrier primitives.
Currently, the modification (connection and disconnection) of probes
from markers requires changes to the data structure done in RCU-style :
a new data structure is created, the pointer is changed atomically, a
quiescent state is reached and then the old data structure is freed.
The quiescent state is reached once all the currently running
preempt_disable regions are done running. We use the call_rcu mechanism
to execute kfree() after such quiescent state has been reached.
However, the new CONFIG_PREEMPT_RCU version of call_rcu and rcu_barrier
does not guarantee that all preempt_disable code regions have finished,
hence the race.
The "proper" way to do this is to use rcu_read_lock/unlock, but we don't
want to use it to minimize intrusiveness on the traced system. (we do
not want the marker code to call into much of the OS code, because it
would quickly restrict what can and cannot be instrumented, such as the
scheduler).
The temporary fix, until we get call_rcu_sched and rcu_barrier_sched in
mainline, is to use synchronize_sched before each call_rcu calls, so we
wait for the quiescent state in the system call code path. It will slow
down batch marker enable/disable, but will make sure the race is gone.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
As Paul pointed out, the ACCESS_ONCE are not needed because we already have
the explicit surrounding memory barriers.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Mason <mmlnx@us.ibm.com>
Cc: Dipankar Sarma <dipankar@in.ibm.com>
Cc: David Smith <dsmith@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Adrian Bunk <adrian.bunk@movial.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add comments requested by Andrew.
Updated comments about synchronize_sched(). Since we use call_rcu and
rcu_barrier now, these comments were out of sync with the code.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Mason <mmlnx@us.ibm.com>
Cc: Dipankar Sarma <dipankar@in.ibm.com>
Cc: David Smith <dsmith@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Adrian Bunk <adrian.bunk@movial.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
get_marker() may return NULL, so test for it.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
RCU style multiple probes support for the Linux Kernel Markers. Common case
(one probe) is still fast and does not require dynamic allocation or a
supplementary pointer dereference on the fast path.
- Move preempt disable from the marker site to the callback.
Since we now have an internal callback, move the preempt disable/enable to the
callback instead of the marker site.
Since the callback change is done asynchronously (passing from a handler that
supports arguments to a handler that does not setup the arguments is no
arguments are passed), we can safely update it even if it is outside the
preempt disable section.
- Move probe arm to probe connection. Now, a connected probe is automatically
armed.
Remove MARK_MAX_FORMAT_LEN, unused.
This patch modifies the Linux Kernel Markers API : it removes the probe
"arm/disarm" and changes the probe function prototype : it now expects a
va_list * instead of a "...".
If we want to have more than one probe connected to a marker at a given
time (LTTng, or blktrace, ssytemtap) then we need this patch. Without it,
connecting a second probe handler to a marker will fail.
It allow us, for instance, to do interesting combinations :
Do standard tracing with LTTng and, eventually, to compute statistics
with SystemTAP, or to have a special trigger on an event that would call
a systemtap script which would stop flight recorder tracing.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Mason <mmlnx@us.ibm.com>
Cc: Dipankar Sarma <dipankar@in.ibm.com>
Cc: David Smith <dsmith@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: "Frank Ch. Eigler" <fche@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Upon module load, we must take the markers mutex. It implies that the marker
mutex must be nested inside the module mutex.
It implies changing the nesting order : now the marker mutex nests inside the
module mutex. Make the necessary changes to reverse the order in which the
mutexes are taken.
Includes some cleanup from Dave Hansen <haveblue@us.ibm.com>.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The marker activation functions sits in kernel/marker.c. A hash table is used
to keep track of the registered probes and armed markers, so the markers
within a newly loaded module that should be active can be activated at module
load time.
marker_query has been removed. marker_get_first, marker_get_next and
marker_release should be used as iterators on the markers.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Acked-by: "Frank Ch. Eigler" <fche@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mike Mason <mmlnx@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
