21 Jan, 2010
1 commit
-
Anton reported that perf record kept receiving events even after calling
ioctl(PERF_EVENT_IOC_DISABLE). It turns out that FORK,COMM and MMAP
events didn't respect the disabled state and kept flowing in.Reported-by: Anton Blanchard
Signed-off-by: Peter Zijlstra
Tested-by: Anton Blanchard
LKML-Reference:
CC: stable@kernel.org
Signed-off-by: Ingo Molnar
01 Jan, 2010
1 commit
-
…el/git/tip/linux-2.6-tip
* 'perf-fixes-for-linus-2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
perf: Fix NULL deref in inheritance code
perf: Pass appropriate frame pointer to dump_trace()
31 Dec, 2009
1 commit
-
Liming found a NULL deref when a task has a perf context but no
counters when it forks.This can occur in two cases, a race during construction where
the fork hits after installing the context but before the first
counter gets inserted, or more reproducably, a fork after the
last counter is closed (which leaves the context around).Reported-by: Wang Liming
Signed-off-by: Peter Zijlstra
Cc: Frederic Weisbecker
Cc: Paul Mackerras
CC:
LKML-Reference:
Signed-off-by: Ingo Molnar
23 Dec, 2009
1 commit
-
It seems a couple places such as arch/ia64/kernel/perfmon.c and
drivers/infiniband/core/uverbs_main.c could use anon_inode_getfile()
instead of a private pseudo-fs + alloc_file(), if only there were a way
to get a read-only file. So provide this by having anon_inode_getfile()
create a read-only file if we pass O_RDONLY in flags.Signed-off-by: Roland Dreier
Signed-off-by: Al Viro
20 Dec, 2009
1 commit
-
…/git/tip/linux-2.6-tip
* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
perf session: Make events_stats u64 to avoid overflow on 32-bit arches
hw-breakpoints: Fix hardware breakpoints -> perf events dependency
perf events: Dont report side-band events on each cpu for per-task-per-cpu events
perf events, x86/stacktrace: Fix performance/softlockup by providing a special frame pointer-only stack walker
perf events, x86/stacktrace: Make stack walking optional
perf events: Remove unused perf_counter.h header file
perf probe: Check new event name
kprobe-tracer: Check new event/group name
perf probe: Check whether debugfs path is correct
perf probe: Fix libdwarf include path for Debian
17 Dec, 2009
4 commits
-
Acme noticed that his FORK/MMAP numbers were inflated by about
the same factor as his cpu-count.This led to the discovery of a few more sites that need to
respect the event->cpu filter.Reported-by: Arnaldo Carvalho de Melo
Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar -
Also, we want to check against nr_cpu_ids, not num_possible_cpus().
The latter works, but the correct bounds check is < nr_cpu_ids.Signed-off-by: Rusty Russell
To: Thomas Gleixner -
…/git/tip/linux-2.6-tip
* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (52 commits)
perf record: Use per-task-per-cpu events for inherited events
perf record: Properly synchronize child creation
perf events: Allow per-task-per-cpu counters
perf diff: Percent calcs should use double values
perf diff: Change the default sort order to "dso,symbol"
perf diff: Use perf_session__fprintf_hists just like 'perf record'
perf report: Fix cut'n'paste error recently introduced
perf session: Move perf report specific hits out of perf_session__fprintf_hists
perf tools: Move hist entries printing routines from perf report
perf report: Generalize perf_session__fprintf_hists()
perf symbols: Move symbol filtering to event__preprocess_sample()
perf symbols: Adopt the strlists for dso, comm
perf symbols: Make symbol_conf global
perf probe: Fix to show which probe point is not found
perf probe: Check symbols in symtab/kallsyms
perf probe: Check build-id of vmlinux
perf probe: Reject second attempt of adding same-name event
perf probe: Support event name for --add option
perf probe: Add glob matching support on --del
perf probe: Use strlist__for_each macros in probe-event.c
... -
In order to allow for per-task-per-cpu counters, useful for
scalability when profiling task hierarchies, we allow installing
events with event->cpu != -1 in task contexts.__perf_event_sched_in() already skips events where ->cpu
mis-matches the current cpu, fix up __perf_install_in_context()
and __perf_event_enable() to also respect this filter.This does lead to vary hard to interpret enabled/running times
for such counters, but I don't see a simple solution for that.Signed-off-by: Peter Zijlstra
Cc: Arnaldo Carvalho de Melo
Cc: fweisbec@gmail.com
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar
16 Dec, 2009
1 commit
-
The miss-alignment of bp_addr created a 32bit hole, causing
different structure packings on 32 and 64 bit machines.Fix that by moving __reserve_2 into that hole.
Further, remove the useless struct and redundant __bp_reserve
muck.Signed-off-by: Peter Zijlstra
Cc: Frederic Weisbecker
Cc: Arnaldo Carvalho de Melo
LKML-Reference:
Signed-off-by: Ingo Molnar
15 Dec, 2009
2 commits
-
It is quite legitimate for CPUs to be numbered sparsely, meaning
that it possible for an online CPU to have a number which is
greater than the total count of possible CPUs.Currently find_get_context() has a sanity check on the cpu
number where it checks it against num_possible_cpus(). This
test can fail for a legitimate cpu number if the
cpu_possible_mask is sparsely populated.This fixes the problem by checking the CPU number against
nr_cpumask_bits instead, since that is the appropriate check to
ensure that the cpu number is same to pass to cpu_isset()
subsequently.Reported-by: Michael Neuling
Signed-off-by: Paul Mackerras
Tested-by: Michael Neuling
Acked-by: Peter Zijlstra
Cc:
LKML-Reference:
Signed-off-by: Ingo Molnar -
Convert locks which cannot be sleeping locks in preempt-rt to
raw_spinlocks.Signed-off-by: Thomas Gleixner
Acked-by: Peter Zijlstra
Acked-by: Ingo Molnar
12 Dec, 2009
1 commit
-
…/git/tip/linux-2.6-tip
* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (57 commits)
x86, perf events: Check if we have APIC enabled
perf_event: Fix variable initialization in other codepaths
perf kmem: Fix unused argument build warning
perf symbols: perf_header__read_build_ids() offset'n'size should be u64
perf symbols: dsos__read_build_ids() should read both user and kernel buildids
perf tools: Align long options which have no short forms
perf kmem: Show usage if no option is specified
sched: Mark sched_clock() as notrace
perf sched: Add max delay time snapshot
perf tools: Correct size given to memset
perf_event: Fix perf_swevent_hrtimer() variable initialization
perf sched: Fix for getting task's execution time
tracing/kprobes: Fix field creation's bad error handling
perf_event: Cleanup for cpu_clock_perf_event_update()
perf_event: Allocate children's perf_event_ctxp at the right time
perf_event: Clean up __perf_event_init_context()
hw-breakpoints: Modify breakpoints without unregistering them
perf probe: Update perf-probe document
perf probe: Support --del option
trace-kprobe: Support delete probe syntax
...
11 Dec, 2009
1 commit
-
Signed-off-by: Xiao Guangrong
Cc: Peter Zijlstra
Cc: Frederic Weisbecker
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar
10 Dec, 2009
1 commit
-
fix:
[] ? printk+0x1d/0x24
[] ? perf_prepare_sample+0x269/0x280
[] warn_slowpath_common+0x71/0xd0
[] ? perf_prepare_sample+0x269/0x280
[] warn_slowpath_null+0x1a/0x20
[] perf_prepare_sample+0x269/0x280
[] ? cpu_clock+0x53/0x90
[] __perf_event_overflow+0x2a8/0x300
[] perf_event_overflow+0x1b/0x30
[] perf_swevent_hrtimer+0x7f/0x120This is because 'data.raw' variable not initialize.
Signed-off-by: Xiao Guangrong
Acked-by: Peter Zijlstra
Cc: Frederic Weisbecker
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar
09 Dec, 2009
4 commits
-
Using atomic64_xchg() instead of atomic64_read() and
atomic64_set().Signed-off-by: Xiao Guangrong
Reviewed-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar -
In current code, children task will allocate memory for
'child->perf_event_ctxp' if the parent is counted, we can
do it only if the parent allowed children inherit it.It can save memory and reduce overhead.
Signed-off-by: Xiao Guangrong
Reviewed-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar -
Clean up the code a bit:
- define 'perf_cpu_context' variable with 'static'
- use kzalloc() instead of kmalloc() and memset()Signed-off-by: Xiao Guangrong
Reviewed-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar -
Currently, when ptrace needs to modify a breakpoint, like disabling
it, changing its address, type or len, it calls
modify_user_hw_breakpoint(). This latter will perform the heavy and
racy task of unregistering the old breakpoint and registering a new
one.This is racy as someone else might steal the reserved breakpoint
slot under us, which is undesired as the breakpoint is only
supposed to be modified, sometimes in the middle of a debugging
workflow. We don't want our slot to be stolen in the middle.So instead of unregistering/registering the breakpoint, just
disable it while we modify its breakpoint fields and re-enable it
after if necessary.Signed-off-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Prasad
LKML-Reference:
Signed-off-by: Ingo Molnar
08 Dec, 2009
1 commit
-
Conflicts:
kernel/irq/chip.c
06 Dec, 2009
1 commit
-
struct perf_event::event callback was called when a breakpoint
triggers. But this is a rather opaque callback, pretty
tied-only to the breakpoint API and not really integrated into perf
as it triggers even when we don't overflow.We prefer to use overflow_handler() as it fits into the perf events
rules, being called only when we overflow.Reported-by: Peter Zijlstra
Signed-off-by: Frederic Weisbecker
Cc: Paul Mackerras
Cc: Arnaldo Carvalho de Melo
Cc: "K. Prasad"
04 Dec, 2009
1 commit
-
That is "success", "unknown", "through", "performance", "[re|un]mapping"
, "access", "default", "reasonable", "[con]currently", "temperature"
, "channel", "[un]used", "application", "example","hierarchy", "therefore"
, "[over|under]flow", "contiguous", "threshold", "enough" and others.Signed-off-by: André Goddard Rosa
Signed-off-by: Jiri Kosina
02 Dec, 2009
1 commit
-
In the CONFIG_PERF_USE_VMALLOC case, perf_mmap_data_free() only
schedules the cleanup of the perf_mmap_data struct. In that
case we have to wait until the work has been done before we free
data.Signed-off-by: Kristian Høgsberg
Cc: David S. Miller
Cc: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
Cc:
LKML-Reference:
Signed-off-by: Ingo Molnar
01 Dec, 2009
1 commit
-
In current code in perf_swevent_hrtimer(), data.period is not
initialized, The result is obvious wrong:# ./perf record -f -e cpu-clock make
# ./perf report
# Samples: 1740
#
# Overhead Command ......
# ........ ........ ..........................................
#
1025422183050275328.00% sh libc-2.9.90.so ...
1025422183050275328.00% perl libperl.so ...
1025422168240043264.00% perl [kernel] ...
1025422030011210752.00% perl [kernel] ...Signed-off-by: Xiao Guangrong
Acked-by: Peter Zijlstra
Cc: Frederic Weisbecker
Cc:
LKML-Reference:
Signed-off-by: Ingo Molnar
27 Nov, 2009
1 commit
-
When a pinned group cannot be scheduled it goes into error state.
Normally a group cannot go out of error state without being
explicitly re-enabled or disabled. There was a bug in per-thread
mode, whereby upon termination of the thread, the group would
transition from error to off leading to bogus counts and timing
information returned by read().Fix it by clearing the error state.
Signed-off-by: Stephane Eranian
Acked-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: perfmon2-devel@lists.sourceforge.net
LKML-Reference:
Signed-off-by: Ingo Molnar
26 Nov, 2009
2 commits
-
bp_perf_event_destroy() is unused in its off-case version, let's
remove it to fix the following warning reported by Stephen
Rothwell in linux-next:kernel/perf_event.c:4306: warning: 'bp_perf_event_destroy' defined but not used
Reported-by: Stephen Rothwell
Signed-off-by: Frederic Weisbecker
Cc: Peter Zijlstra
LKML-Reference:
Signed-off-by: Ingo Molnar -
In fail case, perf_event_create_kernel_counter() returns NULL
instead of an error, which doesn't help us to inform the user
about the origin of the problem from the outer most callers.
Often we can just return -EINVAL, which doesn't help anyone when
it's eventually about a memory allocation failure.Then, this patch makes perf_event_create_kernel_counter() always
return a detailed error code.Signed-off-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Arnaldo Carvalho de Melo
Cc: Paul Mackerras
Cc: Prasad
LKML-Reference:
Signed-off-by: Ingo Molnar
25 Nov, 2009
1 commit
-
Commit 4ed7c92d68a5387ba5f7030dc76eab03558e27f5
(perf_events: Undo some recursion damage) has introduced a bad
reference counting of the recursion context. putting the context
behaves like getting it, dropping every software/trace events
after the first one in a context.Signed-off-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Arnaldo Carvalho de Melo
Cc: Paul Mackerras
Cc: Arjan van de Ven
Cc: Li Zefan
Cc: Steven Rostedt
LKML-Reference:
Signed-off-by: Ingo Molnar
24 Nov, 2009
2 commits
-
When using an event group, the value and id for non leaders events
were wrong due to invalid offset into the outgoing buffer.Signed-off-by: Stephane Eranian
Acked-by: Peter Zijlstra
Cc: paulus@samba.org
Cc: perfmon2-devel@lists.sourceforge.net
LKML-Reference:
Signed-off-by: Ingo Molnar -
Add the remaining necessary bits to support breakpoints created
through perf syscall.We don't use the software counter interface as:
- We don't need to check against recursion, this is already done
in hardware breakpoints arch level.- We already know the perf event we are dealing with when the
event is to be committed.Signed-off-by: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Arnaldo Carvalho de Melo
Cc: Paul Mackerras
Cc: Prasad
LKML-Reference:
Signed-off-by: Ingo Molnar
23 Nov, 2009
8 commits
-
It is quite possible to call update_event_times() on a context
that isn't actually running and thereby confuse the thing.perf stat was reporting !100% scale values for software counters
(2e2af50b perf_events: Disable events when we detach them,
solved the worst of that, but there was still some left).The thing that happens is that because we are not self-reaping
(we have a caring parent) there is a time between the last
schedule (out) and having do_exit() called which will detach the
events.This period would be accounted as enabled,!running because the
event->state==INACTIVE, even though !event->ctx->is_active.Similar issues could have been observed by calling read() on a
event while the attached task was not scheduled in.Solve this by teaching update_event_times() about
ctx->is_active.Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Mike Galbraith
Cc: Arnaldo Carvalho de Melo
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar -
Make perf_swevent_get_recursion_context return a context number
and disable preemption.This could be used to remove the IRQ disable from the trace bit
and index the per-cpu buffer with.Signed-off-by: Peter Zijlstra
Cc: Frederic Weisbecker
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar -
Move the update_event_times() call in __perf_event_exit_task()
into list_del_event() because that holds the proper lock
(ctx->lock) and seems a more natural place to do the last time
update.Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar -
It appeared we did call update_event_times() on exit, but we
failed to update the context time, which renders the former
moot.Locking is a bit iffy, we call update_event_times under
ctx->mutex instead of ctx->lock - the next patch fixes this.Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar -
If we leave the event in STATE_INACTIVE, any read of the event
after the detach will increase the running count but not the
enabled count and cause funny scaling artefacts.Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar -
Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar -
We had two almost identical functions, avoid the duplication.
Signed-off-by: Peter Zijlstra
Cc: Paul Mackerras
Cc: Frederic Weisbecker
Cc: Mike Galbraith
Cc: Arnaldo Carvalho de Melo
LKML-Reference:
Signed-off-by: Ingo Molnar -
The structure init creates a bit memcpy, which shows
up big time in perf annotate output:: ffffffff810a859d :
1.68 : ffffffff810a859d: 55 push %rbp
1.69 : ffffffff810a859e: 41 89 fa mov %edi,%r10d
0.01 : ffffffff810a85a1: 49 89 c9 mov %rcx,%r9
0.00 : ffffffff810a85a4: 31 c0 xor %eax,%eax
1.71 : ffffffff810a85a6: b9 16 00 00 00 mov $0x16,%ecx
0.00 : ffffffff810a85ab: 48 89 e5 mov %rsp,%rbp
0.00 : ffffffff810a85ae: 48 83 ec 60 sub $0x60,%rsp
1.52 : ffffffff810a85b2: 48 8d 7d a0 lea -0x60(%rbp),%rdi
85.20 : ffffffff810a85b6: f3 ab rep stos %eax,%es:(%rdi)None of the callees depends on the structure being pre-initialized,
so only initialize ->addr. This gets rid of the memcpy overhead.Cc: Peter Zijlstra
Cc: Mike Galbraith
Cc: Paul Mackerras
Cc: Arnaldo Carvalho de Melo
Cc: Frederic Weisbecker
LKML-Reference:
Signed-off-by: Ingo Molnar
22 Nov, 2009
2 commits
-
Fix:
ERROR: "perf_swevent_put_recursion_context" [fs/ext4/ext4.ko] undefined!
ERROR: "perf_swevent_get_recursion_context" [fs/ext4/ext4.ko] undefined!Cc: Frederic Weisbecker
Cc: Peter Zijlstra
Cc: Arnaldo Carvalho de Melo
Cc: Paul Mackerras
Cc: Steven Rostedt
Cc: Masami Hiramatsu
Cc: Jason Baron
LKML-Reference:
Signed-off-by: Ingo Molnar -
The buffer is first zeroed out by memset(). Then strncpy() is
used to fill the content. The strncpy() function also pads the
string till the end of the specified length, which is redundant.
The strncpy() does not ensures that the string will be properly
closed with 0. Use strlcpy() instead.The semantic match that finds this kind of pattern is as
follows: (http://coccinelle.lip6.fr/)//
@@
expression buffer;
expression size;
expression str;
@@
memset(buffer, 0, size);
...
- strncpy(
+ strlcpy(
buffer, str, sizeof(buffer)
);
@@
expression buffer;
expression size;
expression str;
@@
memset(&buffer, 0, size);
...
- strncpy(
+ strlcpy(
&buffer, str, sizeof(buffer));
@@
expression buffer;
identifier field;
expression size;
expression str;
@@
memset(buffer, 0, size);
...
- strncpy(
+ strlcpy(
buffer->field, str, sizeof(buffer->field)
);
@@
expression buffer;
identifier field;
expression size;
expression str;
@@
memset(&buffer, 0, size);
...
- strncpy(
+ strlcpy(
buffer.field, str, sizeof(buffer.field));
//On strncpy() vs strlcpy() see
http://www.gratisoft.us/todd/papers/strlcpy.html .Signed-off-by: Márton Németh
Cc: Julia Lawall
Cc: cocci@diku.dk
Cc: Peter Zijlstra
Cc: Paul Mackerras
LKML-Reference:
Signed-off-by: Ingo Molnar
