An rmdir pushes css's ref count to zero. However, if the associated
directory is open at the time, the dentry ref count is non-zero. If
the fd for this directory is then passed into perf_event_open, it
does a css_get(). This bounces the ref count back up from zero. This
is a problem by itself. But what makes it turn into a crash is the
fact that we end up doing an extra dput, since we perform a dput
when css_put sees the ref count go down to zero.
css_tryget() does not fall into that trap. So, we use that instead.
Reproduction test-case for the bug:
#include <unistd.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <linux/unistd.h>
#include <linux/perf_event.h>
#include <string.h>
#include <errno.h>
#include <stdio.h>
#define PERF_FLAG_PID_CGROUP (1U << 2)
int perf_event_open(struct perf_event_attr *hw_event_uptr,
pid_t pid, int cpu, int group_fd, unsigned long flags) {
return syscall(__NR_perf_event_open,hw_event_uptr, pid, cpu,
group_fd, flags);
}
/*
* Directly poke at the perf_event bug, since it's proving hard to repro
* depending on where in the kernel tree. what moved?
*/
int main(int argc, char **argv)
{
int fd;
struct perf_event_attr attr;
memset(&attr, 0, sizeof(attr));
attr.exclude_kernel = 1;
attr.size = sizeof(attr);
mkdir("/dev/cgroup/perf_event/blah", 0777);
fd = open("/dev/cgroup/perf_event/blah", O_RDONLY);
perror("open");
rmdir("/dev/cgroup/perf_event/blah");
sleep(2);
perf_event_open(&attr, fd, 0, -1, PERF_FLAG_PID_CGROUP);
perror("perf_event_open");
close(fd);
return 0;
}
Signed-off-by: Salman Qazi <sqazi@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Tejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/r/20120614223108.1025.2503.stgit@dungbeetle.mtv.corp.google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf fixes from Arnaldo Carvalho de Melo:
* Set name of tracepoints when reading the perf.data headers, so that
we don't end up using the local ones, from /sys.
* Fix default output file for perf stat, from Stephane Eranian.
* Fix endian handling of features bitmask in perf.data header, from David Ahern.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
A bunch of bugzillas have complained how noisy the nmi_watchdog
is during boot-up especially with its expected failure cases
(like virt and bios resource contention).
This is my attempt to quiet them down and keep it less confusing
for the end user. What I did is print the message for cpu0 and
save it for future comparisons. If future cpus have an
identical message as cpu0, then don't print the redundant info.
However, if a future cpu has a different message, happily print
that loudly.
Before the change, you would see something like:
..TIMER: vector=0x30 apic1=0 pin1=2 apic2=-1 pin2=-1
CPU0: Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz stepping 0a
Performance Events: PEBS fmt0+, Core2 events, Intel PMU driver.
... version: 2
... bit width: 40
... generic registers: 2
... value mask: 000000ffffffffff
... max period: 000000007fffffff
... fixed-purpose events: 3
... event mask: 0000000700000003
NMI watchdog enabled, takes one hw-pmu counter.
Booting Node 0, Processors #1
NMI watchdog enabled, takes one hw-pmu counter.
#2
NMI watchdog enabled, takes one hw-pmu counter.
#3 Ok.
NMI watchdog enabled, takes one hw-pmu counter.
Brought up 4 CPUs
Total of 4 processors activated (22607.24 BogoMIPS).
After the change, it is simplified to:
..TIMER: vector=0x30 apic1=0 pin1=2 apic2=-1 pin2=-1
CPU0: Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz stepping 0a
Performance Events: PEBS fmt0+, Core2 events, Intel PMU driver.
... version: 2
... bit width: 40
... generic registers: 2
... value mask: 000000ffffffffff
... max period: 000000007fffffff
... fixed-purpose events: 3
... event mask: 0000000700000003
NMI watchdog: enabled on all CPUs, permanently consumes one hw-PMU counter.
Booting Node 0, Processors #1#2#3 Ok.
Brought up 4 CPUs
V2: little changes based on Joe Perches' feedback
V3: printk cleanup based on Ingo's feedback; checkpatch fix
V4: keep printk as one long line
V5: Ingo fix ups
Reported-and-tested-by: Nathan Zimmer <nzimmer@sgi.com>
Signed-off-by: Don Zickus <dzickus@redhat.com>
Cc: nzimmer@sgi.com
Cc: joe@perches.com
Link: http://lkml.kernel.org/r/1339594548-17227-1-git-send-email-dzickus@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
I noticed that the LBR fixups were not working anymore
on programs where they used to. I tracked this down to
a recent change to copy_from_user_nmi():
db0dc75d64 ("perf/x86: Check user address explicitly in copy_from_user_nmi()")
This commit added a call to __range_not_ok() to the
copy_from_user_nmi() routine. The problem is that the logic
of the test must be reversed. __range_not_ok() returns 0 if the
range is VALID. We want to return early from copy_from_user_nmi()
if the range is NOT valid.
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Arun Sharma <asharma@fb.com>
Link: http://lkml.kernel.org/r/20120611134426.GA7542@quad
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We need to use the per event info snapshoted at record time to
synthesize the events name, so do it just after reading the perf.data
headers, when we already processed the /sys events data, otherwise we'll
end up using the local /sys that only by sheer luck will have the same
tracepoint ID -> real event association.
Example:
# uname -a
Linux felicio.ghostprotocols.net 3.4.0-rc5+ #1 SMP Sat May 19 15:27:11 BRT 2012 x86_64 x86_64 x86_64 GNU/Linux
# perf record -e sched:sched_switch usleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.015 MB perf.data (~648 samples) ]
# cat /t/events/sched/sched_switch/id
279
# perf evlist -v
sched:sched_switch: sample_freq=1, type: 2, config: 279, size: 80, sample_type: 1159, read_format: 7, disabled: 1, inherit: 1, mmap: 1, comm: 1, enable_on_exec: 1, sample_id_all: 1, exclude_guest: 1
#
So on the above machine the sched:sched_switch has tracepoint id 279, but on
the machine were we'll analyse it it has a different id:
$ cat /t/events/sched/sched_switch/id
56
$ perf evlist -i /tmp/perf.data
kmem:mm_balancedirty_writeout
$ cat /t/events/kmem/mm_balancedirty_writeout/id
279
With this fix:
$ perf evlist -i /tmp/perf.data
sched:sched_switch
Reported-by: Dmitry Antipov <dmitry.antipov@linaro.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-auwks8fpuhmrdpiefs55o5oz@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The following commit:
commit 56f3bae706
Author: Jim Cromie <jim.cromie@gmail.com>
Date: Wed Sep 7 17:14:00 2011 -0600
perf stat: Add --log-fd <N> option to redirect stderr elsewhere
introduced a bug in the way perf stat outputs the results by default,
i.e., without the --log-fd or --output option. It would default to
writing to file descriptor 0, i.e., stdin. Writing to stdin is allowed
and is equivalent to writing to stdout. However, there is a major
difference for any script that was already capturing the output of perf
stat via redirection:
perf stat >/tmp/log .... or perf stat 2>/tmp/log ....
They would not capture anything anymore. They would have to do:
perf stat 0>/tmp/log ...
This breaks compatibility with existing scripts and does not look very
natural.
This patch fixes the problem by looking at output_fd only when it was
modified by user (> 0). It also checks that the value if positive.
Passing --log-fd 0 is ignored.
I would also argue that defaulting to stderr for the results is not the
right thing to do, though this patch does not address this specific
issue.
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jim Cromie <jim.cromie@gmail.com>
Link: http://lkml.kernel.org/r/20120515111111.GA9870@quad
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Based on Jiri's latest attempt:
https://lkml.org/lkml/2012/5/16/61
Basically, adds_features should be byte swapped assuming unsigned
longs are either 8-bytes (u64) or 4-bytes (u32).
Fixes 32-bit ppc dumping 64-bit x86 feature data:
========
captured on: Sun May 20 19:23:23 2012
hostname : nxos-vdc-dev3
os release : 3.4.0-rc7+
perf version : 3.4.rc4.137.g978da3
arch : x86_64
nrcpus online : 16
nrcpus avail : 16
cpudesc : Intel(R) Xeon(R) CPU E5540 @ 2.53GHz
cpuid : GenuineIntel,6,26,5
total memory : 24680324 kB
...
Verified 64-bit x86 can still dump feature data for 32-bit ppc.
Signed-off-by: David Ahern <dsahern@gmail.com>
Reviewed-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/4FBBB539.5010805@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
A recent update to have tracing_on/off() only affect the ftrace ring
buffers instead of all ring buffers had a cut and paste error.
The tracing_off() did the exact same thing as tracing_on() and
would not actually turn off tracing. Unfortunately, tracing_off()
is more important to be working than tracing_on() as this is a key
development tool, as it lets the developer turn off tracing as soon
as a problem is discovered. It is also used by panic and oops code.
This bug also breaks the 'echo func:traceoff > set_ftrace_filter'
Cc: <stable@vger.kernel.org> # 3.4
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Zheng Yan reported that event group validation can wreck event state
when Intel extra_reg allocation changes event state.
Validation shouldn't change any persistent state. Cloning events in
validate_{event,group}() isn't really pretty either, so add a few
special cases to avoid modifying the event state.
The code is restructured to minimize the special case impact.
Reported-by: Zheng Yan <zheng.z.yan@linux.intel.com>
Acked-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1338903031.28282.175.camel@twins
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fix the x86 instruction decoder to decode bsr/bsf/jmpe with
operand-size prefix (66h). This fixes the test case failure
reported by Linus, attached below.
bsf/bsr/jmpe have a special encoding. Opcode map in
Intel Software Developers Manual vol2 says they have
TZCNT/LZCNT variants if it has F3h prefix. However, there
is no information if it has other 66h or F2h prefixes.
Current instruction decoder supposes that those are
bad instructions, but it actually accepts at least
operand-size prefixes.
H. Peter Anvin further explains:
" TZCNT/LZCNT are F3 + BSF/BSR exactly because the F2 and
F3 prefixes have historically been no-ops with most instructions.
This allows software to unconditionally use the prefixed versions
and get TZCNT/LZCNT on the processors that have them if they don't
care about the difference. "
This fixes errors reported by test_get_len:
Warning: arch/x86/tools/test_get_len found difference at <em_bsf>:ffffffff81036d87
Warning: ffffffff81036de5: 66 0f bc c2 bsf %dx,%ax
Warning: objdump says 4 bytes, but insn_get_length() says 3
Warning: arch/x86/tools/test_get_len found difference at <em_bsr>:ffffffff81036ea6
Warning: ffffffff81036f04: 66 0f bd c2 bsr %dx,%ax
Warning: objdump says 4 bytes, but insn_get_length() says 3
Warning: decoded and checked 13298882 instructions with 2 warnings
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Reported-by: Pekka Enberg <penberg@kernel.org>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <yrl.pp-manager.tt@hitachi.com>
Link: http://lkml.kernel.org/r/20120604150911.22338.43296.stgit@localhost.localdomain
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf fixes from Arnaldo Carvalho de Melo:
* Endianness fixes from Jiri Olsa
* Fixes for make perf tarball
* Fix for DSO name in perf script callchains, from David Ahern
* Segfault fixes for perf top --callchain, from Namhyung Kim
* Minor function result fixes from Srikar Dronamraju
* Add missing 3rd ioctl parameter, from Namhyung Kim
* Fix pager usage in minimal embedded systems, from Avik Sil
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull scheduler fixes from Ingo Molnar.
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched: Remove NULL assignment of dattr_cur
sched: Remove the last NULL entry from sched_feat_names
sched: Make sched_feat_names const
sched/rt: Fix SCHED_RR across cgroups
sched: Move nr_cpus_allowed out of 'struct sched_rt_entity'
sched: Make sure to not re-read variables after validation
sched: Fix SD_OVERLAP
sched: Don't try allocating memory from offline nodes
sched/nohz: Fix rq->cpu_load calculations some more
sched/x86: Use cpu_llc_shared_mask(cpu) for coregroup_mask
Pull signal and vfs compile breakage fixes from Al Viro.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/signal:
fixups for signal breakage
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
nommu: fix compilation of nommu.c
Pull cifs fixes from Steve French.
* 'for-next' of git://git.samba.org/sfrench/cifs-2.6:
CIFS: Move get_next_mid to ops struct
CIFS: Make accessing is_valid_oplock/dump_detail ops struct field safe
CIFS: Improve identation in cifs_unlock_range
CIFS: Fix possible wrong memory allocation
Compiling 3.5-rc1 for nommu targets gives:
CC mm/nommu.o
mm/nommu.c: In function ‘sys_mmap_pgoff’:
mm/nommu.c:1489:2: error: ‘ret’ undeclared (first use in this function)
mm/nommu.c:1489:2: note: each undeclared identifier is reported only once for each function it appears in
It is trivially fixed by replacing 'ret' with the local variable that is
already defined for the return value 'retval'.
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Salman Qazi
Ingo Molnar
Don Zickus
Stephane Eranian
Arnaldo Carvalho de Melo
David Ahern
Steven Rostedt
Arun Sharma
Peter Zijlstra
Masami Hiramatsu
Linus Torvalds
Al Viro
Greg Ungerer