Ingo Molnar [Wed, 13 May 2009 10:54:01 +0000 (12:54 +0200)]
perf_counter: x86: Fix throttling
If counters are disabled globally when a perfcounter IRQ/NMI hits,
and if we throttle in that case, we'll promote the '0' value to
the next lapic IRQ and disable all perfcounters at that point,
permanently ...
Fix it.
[ Impact: fix hung perfcounters under load ]
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Allow specifying a pid instead of always fork+exec'ing a command.
Because the PERF_EVENT_COMM and PERF_EVENT_MMAP events happened before
we connected, we must synthesize them so that 'perf report' can get what
it needs.
[ Impact: add new command line option ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Clark Williams <williams@redhat.com> Cc: John Kacur <jkacur@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20090515015046.GA13664@ghostprotocols.net> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Mackerras [Tue, 12 May 2009 11:59:01 +0000 (21:59 +1000)]
perf_counter: call hw_perf_save_disable/restore around group_sched_in
I noticed that when enabling a group via the PERF_COUNTER_IOC_ENABLE
ioctl on the group leader, the counters weren't enabled and counting
immediately on return from the ioctl, but did start counting a little
while later (presumably after a context switch).
The reason was that __perf_counter_enable calls group_sched_in which
calls hw_perf_group_sched_in, which on powerpc assumes that the caller
has called hw_perf_save_disable already. Until commit 46d686c6
("perf_counter: put whole group on when enabling group leader") it was
true that all callers of group_sched_in had called
hw_perf_save_disable first, and the powerpc hw_perf_group_sched_in
relies on that (there isn't an x86 version).
This fixes the problem by putting calls to hw_perf_save_disable /
hw_perf_restore around the calls to group_sched_in and
counter_sched_in in __perf_counter_enable. Having the calls to
hw_perf_save_disable/restore around the counter_sched_in call is
harmless and makes this call consistent with the other call sites
of counter_sched_in, which have all called hw_perf_save_disable first.
[ Impact: more precise counter group disable/enable functionality ]
Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <18953.25733.53359.147452@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Mackerras [Mon, 11 May 2009 05:50:21 +0000 (15:50 +1000)]
perf_counter: call atomic64_set for counter->count
A compile warning triggered because we are calling
atomic_set(&counter->count). But since counter->count
is an atomic64_t, we have to use atomic64_set.
So the count can be set short, resulting in the reset ioctl
only resetting the low word.
[ Impact: clear counter properly during the reset ioctl ]
Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <18951.48285.270311.981806@drongo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Mackerras [Mon, 11 May 2009 05:46:10 +0000 (15:46 +1000)]
perf_counter: don't count scheduler ticks as context switches
The context-switch software counter gives inflated values at present
because each scheduler tick and each process-wide counter
enable/disable prctl gets counted as a context switch.
This happens because perf_counter_task_tick, perf_counter_task_disable
and perf_counter_task_enable all call perf_counter_task_sched_out,
which calls perf_swcounter_event to record a context switch event.
This fixes it by introducing a variant of perf_counter_task_sched_out
with two underscores in front for internal use within the perf_counter
code, and makes perf_counter_task_{tick,disable,enable} call it. This
variant doesn't record a context switch event, and takes a struct
perf_counter_context *. This adds the new variant rather than
changing the behaviour or interface of perf_counter_task_sched_out
because that is called from other code.
Paul Mackerras [Mon, 11 May 2009 02:08:02 +0000 (12:08 +1000)]
perf_counter: Put whole group on when enabling group leader
Currently, if you have a group where the leader is disabled and there
are siblings that are enabled, and then you enable the leader, we only
put the leader on the PMU, and not its enabled siblings. This is
incorrect, since the enabled group members should be all on or all off
at any given point.
This fixes it by adding a call to group_sched_in in
__perf_counter_enable in the case where we're enabling a group leader.
To avoid the need for a forward declaration this also moves
group_sched_in up before __perf_counter_enable. The actual content of
group_sched_in is unchanged by this patch.
[ Impact: fix bug in counter enable code ]
Signed-off-by: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <18951.34946.451546.691693@drongo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Peter Zijlstra [Fri, 8 May 2009 16:52:24 +0000 (18:52 +0200)]
perf_counter: add PERF_RECORD_CPU
Allow recording the CPU number the event was generated on.
RFC: this leaves a u32 as reserved, should we fill in the
node_id() there, or leave this open for future extention,
as userspace can already easily do the cpu->node mapping
if needed.
[ Impact: extend perfcounter output record format ]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090508170029.008627711@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Peter Zijlstra [Fri, 8 May 2009 16:52:22 +0000 (18:52 +0200)]
perf_counter: rework ioctl()s
Corey noticed that ioctl()s on grouped counters didn't work on
the whole group. This extends the ioctl() interface to take a
second argument that is interpreted as a flags field. We then
provide PERF_IOC_FLAG_GROUP to toggle the behaviour.
Having this flag gives the greatest flexibility, allowing you
to individually enable/disable/reset counters in a group, or
all together.
[ Impact: fix group counter enable/disable semantics ]
Reported-by: Corey Ashford <cjashfor@linux.vnet.ibm.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <20090508170028.837558214@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Peter Zijlstra [Fri, 8 May 2009 16:52:21 +0000 (18:52 +0200)]
perf_counter: optimize perf_counter_task_tick()
perf_counter_task_tick() does way too much work to find out
there's nothing to do. Provide an easy short-circuit for the
normal case where there are no counters on the system.
[ Impact: micro-optimization ]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090508170028.750619201@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Wed, 6 May 2009 06:46:27 +0000 (08:46 +0200)]
Merge branch 'core/locking' into perfcounters/core
Merge reason: we moved a mutex.h commit that originated from the
perfcounters tree into core/locking - but now merge
back that branch to solve a merge artifact and to
pick up cleanups of this commit that happened in
core/locking.
Peter Zijlstra [Tue, 5 May 2009 15:50:22 +0000 (17:50 +0200)]
perf_counter: uncouple data_head updates from wakeups
Keep data_head up-to-date irrespective of notifications. This fixes
the case where you disable a counter and don't get a notification for
the last few pending events, and it also allows polling usage.
[ Impact: increase precision of perfcounter mmap-ed fields ]
Suggested-by: Corey Ashford <cjashfor@linux.vnet.ibm.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <20090505155436.925084300@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
ef7b3e0: perf_counter, x86: remove vendor check in fixed_mode_idx()
Which made x86_pmu.num_counters_fixed matter: if it's nonzero, the
fixed-purpose counters are utilized.
But on v2 perfmon this field is not set (despite there being
fixed-purpose PMCs). So add a quirk to set the number of fixed-purpose
counters to at least three.
[ Impact: add quirk for three fixed-purpose counters on certain Intel CPUs ]
Cc: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-28-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Mon, 4 May 2009 17:23:18 +0000 (19:23 +0200)]
perf_counter: convert perf_resource_mutex to a spinlock
Now percpu counters can be initialized very early. But the init
sequence uses mutex_lock(). Fortunately, perf_resource_mutex should
be a spinlock anyway, so convert it.
Ingo Molnar [Mon, 4 May 2009 17:13:30 +0000 (19:13 +0200)]
perf_counter: initialize the per-cpu context earlier
percpu scheduling for perfcounters wants to take the context lock,
but that lock first needs to be initialized. Currently it is an
early_initcall() - but that is too late, the task tick runs much
sooner than that.
Call it explicitly from the scheduler init sequence instead.
Peter Zijlstra [Fri, 1 May 2009 10:23:16 +0000 (12:23 +0200)]
perf_counter: fix race in perf_output_*
When two (or more) contexts output to the same buffer, it is possible
to observe half written output.
Suppose we have CPU0 doing perf_counter_mmap(), CPU1 doing
perf_counter_overflow(). If CPU1 does a wakeup and exposes head to
user-space, then CPU2 can observe the data CPU0 is still writing.
[ Impact: fix occasionally corrupted profiling records ]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090501102533.007821627@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Thomas Gleixner [Sat, 4 Apr 2009 21:01:06 +0000 (21:01 +0000)]
signals: implement sys_rt_tgsigqueueinfo
sys_kill has the per thread counterpart sys_tgkill. sigqueueinfo is
missing a thread directed counterpart. Such an interface is important
for migrating applications from other OSes which have the per thread
delivery implemented.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Oleg Nesterov <oleg@redhat.com> Acked-by: Roland McGrath <roland@redhat.com> Acked-by: Ulrich Drepper <drepper@redhat.com>
perf_counter tools: fix infinite loop in perf-report on zeroed event records
Bail out early if a record has zero size - we have no chance to make
reliable progress in that case. Print out the offset where this happens,
and print the number of bytes we missed out on.
Andrew Morton [Wed, 29 Apr 2009 22:59:58 +0000 (15:59 -0700)]
mutex: add atomic_dec_and_mutex_lock(), fix
include/linux/mutex.h:136: warning: 'mutex_lock' declared inline after being called
include/linux/mutex.h:136: warning: previous declaration of 'mutex_lock' was here
uninline it.
[ Impact: clean up and uninline, address compiler warning ]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <200904292318.n3TNIsi6028340@imap1.linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs/ecryptfs-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs/ecryptfs-2.6:
eCryptfs: Fix min function comparison warning
ecryptfs: fix printk format warning
Luis Henriques [Wed, 29 Apr 2009 20:54:51 +0000 (21:54 +0100)]
locking, rtmutex.c: Documentation cleanup
Two minor updates on functions documentation:
- Updated documentation for function rt_mutex_unlock(), which contained an
incorrect name
- Removed extra '*' from comment in function rt_mutex_destroy()
[ Impact: cleanup ]
Signed-off-by: Luis Henriques <henrix@sapo.pt> Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <20090429205451.GA23154@hades.domain.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
V4L/DVB (11652): au0828: fix kernel oops regression on USB disconnect.
A regression was introduced in hg changeset 33810c734a0d, which resulted in
a kernel panic whenever the device was disconnected from USB. The call to
4l2_device_register() was overwriting the pointer for usb_set_intfdata(), so
when au0828_usb_disconnect() was called, the usb_get_intfdata() returned a
pointer to the v4l2_device instead of the au0828_dev structure.
Jean Delvare [Fri, 17 Apr 2009 13:56:51 +0000 (10:56 -0300)]
V4L/DVB (11568): cx18: Fix the handling of i2c bus registration error
* Return actual error values as returned by the i2c subsystem, rather
than 0 or 1.
* If the registration of the second bus fails, unregister the first one
before exiting, otherwise we are leaking resources.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Andy Walls <awalls@radix.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Currently drivers/media drivers are linked very early - directly after
base, block, misc, and mfd and before ata, scsi, ide, input, firewire,
usb, and i2c. This breaks static build of video4linux drivers, that use
generic CPU i2c adapter drivers and the v4l2-subdev subsystem, because
during video4linux probing the v4l2-subdev core requires a struct
i2c_adapter context, which cannot be satisfied before the i2c subsystem is
initialised. Moving drivers/media after drivers/i2c fixes this problem.
The best way to trigger action is by submitting a patch:-) So, let's see
what comes out of it - on the one hand I don't see any reason why media
has to be linked this early, and nobody was able to give me one yesterday
as this problem has been discussed on linux-media, OTOH, maybe indeed it
would be better to move i2c the whole way up above media, but that'd be
much bigger of a change, I think.
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Andy Walls [Tue, 14 Apr 2009 00:43:31 +0000 (21:43 -0300)]
V4L/DVB (11494): cx18: Send correct input routing value to external audio multiplexers
A late v4l2_subdev framework change accidentally sent the audio input
routing value to the external multiplexer, instead of the muxer input routing
value to the external multiplexer. This change corrects that error.
Signed-off-by: Andy Walls <awalls@radix.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (24 commits)
e100: do not go D3 in shutdown unless system is powering off
netfilter: revised locking for x_tables
Bluetooth: Fix connection establishment with low security requirement
Bluetooth: Add different pairing timeout for Legacy Pairing
Bluetooth: Ensure that HCI sysfs add/del is preempt safe
net: Avoid extra wakeups of threads blocked in wait_for_packet()
net: Fix typo in net_device_ops description.
ipv4: Limit size of route cache hash table
Add reference to CAPI 2.0 standard
Documentation/isdn/INTERFACE.CAPI
update Documentation/isdn/00-INDEX
ixgbe: Fix WoL functionality for 82599 KX4 devices
veth: prevent oops caused by netdev destructor
xfrm: wrong hash value for temporary SA
forcedeth: tx timeout fix
net: Fix LL_MAX_HEADER for CONFIG_TR_MODULE
mlx4_en: Handle page allocation failure during receive
mlx4_en: Fix cleanup flow on cq activation
vlan: update vlan carrier state for admin up/down
netfilter: xt_recent: fix stack overread in compat code
...
Eric Paris [Mon, 23 Mar 2009 17:22:09 +0000 (18:22 +0100)]
mutex: add atomic_dec_and_mutex_lock()
Much like the atomic_dec_and_lock() function in which we take an hold a
spin_lock if we drop the atomic to 0 this function takes and holds the
mutex if we dec the atomic to 0.
Signed-off-by: Eric Paris <eparis@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090323172417.410913479@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Mackerras [Wed, 29 Apr 2009 12:38:51 +0000 (22:38 +1000)]
perf_counter: powerpc: allow use of limited-function counters
POWER5+ and POWER6 have two hardware counters with limited functionality:
PMC5 counts instructions completed in run state and PMC6 counts cycles
in run state. (Run state is the state when a hardware RUN bit is 1;
the idle task clears RUN while waiting for work to do and sets it when
there is work to do.)
These counters can't be written to by the kernel, can't generate
interrupts, and don't obey the freeze conditions. That means we can
only use them for per-task counters (where we know we'll always be in
run state; we can't put a per-task counter on an idle task), and only
if we don't want interrupts and we do want to count in all processor
modes.
Obviously some counters can't go on a limited hardware counter, but there
are also situations where we can only put a counter on a limited hardware
counter - if there are already counters on that exclude some processor
modes and we want to put on a per-task cycle or instruction counter that
doesn't exclude any processor mode, it could go on if it can use a
limited hardware counter.
To keep track of these constraints, this adds a flags argument to the
processor-specific get_alternatives() functions, with three bits defined:
one to say that we can accept alternative event codes that go on limited
counters, one to say we only want alternatives on limited counters, and
one to say that this is a per-task counter and therefore events that are
gated by run state are equivalent to those that aren't (e.g. a "cycles"
event is equivalent to a "cycles in run state" event). These flags
are computed for each counter and stored in the counter->hw.counter_base
field (slightly wonky name for what it does, but it was an existing
unused field).
Since the limited counters don't freeze when we freeze the other counters,
we need some special handling to avoid getting skew between things counted
on the limited counters and those counted on normal counters. To minimize
this skew, if we are using any limited counters, we read PMC5 and PMC6
immediately after setting and clearing the freeze bit. This is done in
a single asm in the new write_mmcr0() function.
The code here is specific to PMC5 and PMC6 being the limited hardware
counters. Being more general (e.g. having a bitmap of limited hardware
counter numbers) would have meant more complex code to read the limited
counters when freezing and unfreezing the normal counters, with
conditional branches, which would have increased the skew. Since it
isn't necessary for the code to be more general at this stage, it isn't.
This also extends the back-ends for POWER5+ and POWER6 to be able to
handle up to 6 counters rather than the 4 they previously handled.
Signed-off-by: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Robert Richter <robert.richter@amd.com>
LKML-Reference: <18936.19035.163066.892208@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:24 +0000 (12:47 +0200)]
perf_counter, x86: remove vendor check in fixed_mode_idx()
The function fixed_mode_idx() is used generically. Now it checks the
num_counters_fixed value instead of the vendor to decide if fixed
counters are present.
[ Impact: generalize code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-28-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:23 +0000 (12:47 +0200)]
perf_counter, x86: introduce max_period variable
In x86 pmus the allowed counter period to programm differs. This
introduces a max_period value and allows the generic implementation
for all models to check the max period.
[ Impact: generalize code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-27-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:22 +0000 (12:47 +0200)]
perf_counter, x86: return raw count with x86_perf_counter_update()
To check on AMD cpus if a counter overflows, the upper bit of the raw
counter value must be checked. This value is already internally
available in x86_perf_counter_update(). Now, the value is returned so
that it can be used directly to check for overflows.
[ Impact: micro-optimization ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-26-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:21 +0000 (12:47 +0200)]
perf_counter, x86: implement the interrupt handler for AMD cpus
This patch implements the interrupt handler for AMD performance
counters. In difference to the Intel pmu, there is no single status
register and also there are no fixed counters. This makes the handler
very different and it is useful to make the handler vendor
specific. To check if a counter is overflowed the upper bit of the
counter is checked. Only counters where the active bit is set are
checked.
With this patch throttling is enabled for AMD performance counters.
This patch also reenables Linux performance counters on AMD cpus.
[ Impact: re-enable perfcounters on AMD CPUs ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-25-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:20 +0000 (12:47 +0200)]
perf_counter, x86: change and remove pmu initialization checks
Some functions are only called if the pmu was proper initialized. That
initalization checks can be removed. The way to check initialization
changed too. Now, the pointer to the interrupt handler is checked. If
it exists the pmu is initialized. This also removes a static variable
and uses struct x86_pmu as only data source for the check.
[ Impact: simplify code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-24-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
As for the enable function, this patch reworks the disable functions
and introduces x86_pmu_disable_counter(). The internal function i/f in
struct x86_pmu changed too.
[ Impact: refactor and generalize code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-23-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
There is vendor specific code in generic x86 code, and there is vendor
specific code that could be generic. This patch introduces
x86_pmu_enable_counter() for x86 generic code. Fixed counter code for
Intel is moved to Intel only functions. In the end, checks and calls
via function pointers were reduced to the necessary. Also, the
internal function i/f changed.
[ Impact: refactor and generalize code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-22-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:16 +0000 (12:47 +0200)]
perf_counter, x86: generic use of cpuc->active
cpuc->active will now be used to indicate an enabled counter which
implies also valid pointers of cpuc->counters[]. In contrast,
cpuc->used only locks the counter, but it can be still uninitialized.
[ Impact: refactor and generalize code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-20-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:11 +0000 (12:47 +0200)]
perf_counter, x86: make x86_pmu data a static struct
Instead of using a pointer to reference to the x86 pmu we now have one
single data structure that is initialized at the beginning. This saves
the pointer access when using this memory.
[ Impact: micro-optimization ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-15-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:04 +0000 (12:47 +0200)]
perf_counter, x86: rename struct pmc_x86_ops into struct x86_pmu
This patch renames struct pmc_x86_ops into struct x86_pmu. It
introduces a structure to describe an x86 model specific pmu
(performance monitoring unit). It may contain ops and data. The new
name of the structure fits better, is shorter, and thus better to
handle. Where it was appropriate, names of function and variable have
been changed too.
[ Impact: cleanup ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-8-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:03 +0000 (12:47 +0200)]
perfcounters: rename struct hw_perf_counter_ops into struct pmu
This patch renames struct hw_perf_counter_ops into struct pmu. It
introduces a structure to describe a cpu specific pmu (performance
monitoring unit). It may contain ops and data. The new name of the
structure fits better, is shorter, and thus better to handle. Where it
was appropriate, names of function and variable have been changed too.
[ Impact: cleanup ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-7-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:47:02 +0000 (12:47 +0200)]
perf_counter, x86: protect per-cpu variables with compile barriers only
Per-cpu variables needn't to be protected with cpu barriers
(smp_wmb()). Protection is only needed for preemption on the same cpu
(rescheduling or the nmi handler). This can be done using a compiler
barrier only.
[ Impact: micro-optimization ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Cc: Paul Mackerras <paulus@samba.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1241002046-8832-6-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Robert Richter [Wed, 29 Apr 2009 10:46:58 +0000 (12:46 +0200)]
perf_counter, x86: remove X86_FEATURE_ARCH_PERFMON flag for AMD cpus
X86_FEATURE_ARCH_PERFMON is an Intel hardware feature that does not
work on AMD CPUs. The flag is now only used in Intel specific code
(especially initialization).
[ Impact: refactor code ]
Signed-off-by: Robert Richter <robert.richter@amd.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1241002046-8832-2-git-send-email-robert.richter@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
e100: do not go D3 in shutdown unless system is powering off
After experimenting with kexec with the last merges after 2.6.29, I've
had some problems when probing e100. It would not read the eeprom. After
some bisects, I realized this has been like that since forever (at least
2.6.18). The problem is that shutdown is doing the same thing that
suspend does and puts the device in D3 state. I couldn't find a way to
get the device back to a sane state in the probe function. So, based on
some similar patches from Rafael J. Wysocki for e1000, e1000e, and ixgbe,
I wrote this one for e100.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com> Acked-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The x_tables are organized with a table structure and a per-cpu copies
of the counters and rules. On older kernels there was a reader/writer
lock per table which was a performance bottleneck. In 2.6.30-rc, this
was converted to use RCU and the counters/rules which solved the performance
problems for do_table but made replacing rules much slower because of
the necessary RCU grace period.
This version uses a per-cpu set of spinlocks and counters to allow to
table processing to proceed without the cache thrashing of a global
reader lock and keeps the same performance for table updates.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / karo-tx-linux.git / log