Commit graph

150,846 commits

Author SHA1 Message Date
Trent Piepho
a7d878af94 leds: Add openfirmware platform device support
Add bindings to support LEDs defined as of_platform devices in addition to
the existing bindings for platform devices.

New options in Kconfig allow the platform binding code and/or the
of_platform code to be turned on.  The of_platform code is of course only
available on archs that have OF support.

The existing probe and remove methods are refactored to use new functions
create_gpio_led(), to create and register one led, and delete_gpio_led(),
to unregister and free one led.  The new probe and remove methods for the
of_platform driver can then share most of the common probe and remove code
with the platform driver.

The suspend and resume methods aren't shared, but they are very short.  The
actual led driving code is the same for LEDs created by either binding.

The OF bindings are based on patch by Anton Vorontsov
<avorontsov@ru.mvista.com>.  They have been extended to allow multiple LEDs
per device.

Signed-off-by: Trent Piepho <tpiepho@freescale.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Sean MacLennan <smaclennan@pikatech.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
2009-04-06 16:06:25 +01:00
Jens Axboe
1aa2a7cc6f block: switch sync_dirty_buffer() over to WRITE_SYNC
We should now have the logic in place to handle this properly
without regressing on the write performance, so re-enable
the sync writes.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:54 -07:00
Jens Axboe
aeb6fafb8f block: Add flag for telling the IO schedulers NOT to anticipate more IO
By default, CFQ will anticipate more IO from a given io context if the
previously completed IO was sync. This used to be fine, since the only
sync IO was reads and O_DIRECT writes. But with more "normal" sync writes
being used now, we don't want to anticipate for those.

Add a bio/request flag that informs the IO scheduler that this is a sync
request that we should not idle for. Introduce WRITE_ODIRECT specifically
for O_DIRECT writes, and make sure that the other sync writes set this
flag.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:54 -07:00
Jens Axboe
644b2d99b7 block: enabling plugging on SSD devices that don't do queuing
For the older SSD devices that don't do command queuing, we do want to
enable plugging to get better merging.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:54 -07:00
Jens Axboe
4194b1eaf1 jbd2: use WRITE_SYNC_PLUG instead of WRITE_SYNC
When you are going to be submitting several sync writes, we want to
give the IO scheduler a chance to merge some of them. Instead of
using the implicitly unplugging WRITE_SYNC variant, use WRITE_SYNC_PLUG
and rely on sync_buffer() doing the unplug when someone does a
wait_on_buffer()/lock_buffer().

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:54 -07:00
Jens Axboe
6c4bac6b33 jbd: use WRITE_SYNC_PLUG instead of WRITE_SYNC
When you are going to be submitting several sync writes, we want to
give the IO scheduler a chance to merge some of them. Instead of
using the implicitly unplugging WRITE_SYNC variant, use WRITE_SYNC_PLUG
and rely on sync_buffer() doing the unplug when someone does a
wait_on_buffer()/lock_buffer().

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:53 -07:00
Jens Axboe
9cf6b720f8 block: fsync_buffers_list() should use SWRITE_SYNC_PLUG
Then it can submit all the buffers without unplugging for each one.
We will kick off the pending IO if we come across a new address space.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:53 -07:00
Jens Axboe
a1f242524c Add WRITE_SYNC_PLUG and SWRITE_SYNC_PLUG
(S)WRITE_SYNC always unplugs the device right after IO submission.
Sometimes we want to build up a queue before doing so, so add
variants that explicitly DON'T unplug the queue. The caller must
then do that after submitting all the IO.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:53 -07:00
Jens Axboe
1faa16d228 block: change the request allocation/congestion logic to be sync/async based
This makes sure that we never wait on async IO for sync requests, instead
of doing the split on writes vs reads.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-06 08:04:53 -07:00
Kay Sievers
30bbf14067 [MTD] struct device - replace bus_id with dev_name(), dev_set_name()
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:24:02 -07:00
Darius Augulis
504a3e7220 [MTD] [NOR] Fixup for Numonyx M29W128 chips
Prevent NUMONYX M29W128G memories from using write buffer,
because it doesn't work properly.

Signed-off-by: Darius Augulis <augulis.darius@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:23:26 -07:00
Yauhen Kharuzhy
d8877f191e [MTD] mtdpart: Make ecc_stats more realistic.
In the existing implementation, ecc_stats fields are incremented only by
one, regardless of master mtd errors number. For example, if there are N
errors were corrected by ECC, partition ecc_stats.corrected will be
incremented by one.

This commit changes simple increment to sum of old value and parent mtd
error count.

Signed-off-by: Yauhen Kharuzhy <jekhor@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:22:28 -07:00
Wolfgang Grandegger
7995c7e94f powerpc/85xx: TQM8548: Update DTS file for multi-chip support
This patch adds multi-chip support for the Micron MT29F8G08FAB NAND
flash memory on the TQM8548 modules.

Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:19:18 -07:00
Wolfgang Grandegger
21e9d9456f powerpc: NAND: FSL UPM: document new bindings
This patch adds documentation for the new NAND FSL UPM bindings for:

 NAND: FSL-UPM: add multi chip support
 NAND: FSL-UPM: Add wait flags to support board/chip specific delays

It also documents the old binding for "chip-delay".

Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Acked-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:18:39 -07:00
Wolfgang Grandegger
ade92a636f [MTD] [NAND] FSL-UPM: Add wait flags to support board/chip specific delays
The NAND flash on the TQM8548_BE modules requires a short delay after
running the UPM pattern. The TQM8548_BE requires a further short delay
after writing out a buffer. Normally the R/B pin should be checked, but
it's not connected on the TQM8548_BE. The existing driver uses similar
fixed delay points. To manage these extra delays in a more general way,
I introduced the "fsl,ump-wait-flags" property allowing the board-
specific driver to specify various types of extra delay.

Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Acked-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:17:59 -07:00
Wolfgang Grandegger
b6e0e8c077 [MTD] [NAND] FSL-UPM: add multi chip support
This patch adds support for multi-chip NAND devices to the FSL-UPM
driver. This requires support for multiple GPIOs for the RNB pins.
The NAND chips are selected through address lines defined by the
FDT property "fsl,upm-addr-line-cs-offsets".

Signed-off-by: Wolfgang Grandegger <wg@grandegger.com>
Acked-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:17:09 -07:00
David Woodhouse
db99a55231 [MTD] [NOR] Add device parent info to physmap_of
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:15:42 -07:00
Wolfgang Grandegger
1b578193af [MTD] [NAND] Add support for NAND on the Socrates board
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Acked-by: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:15:27 -07:00
Anton Vorontsov
f379188958 fsl-diu-fb: Pass the proper device for dma mapping routines
The driver should pass a device that specifies internal DMA ops, but
currently NULL pointers are passed, and thus following bug pops up:

  Freescale DIU driver
  ------------[ cut here ]------------
  kernel BUG at arch/powerpc/include/asm/dma-mapping.h:237!
  Oops: Exception in kernel mode, sig: 5 [#1]
  ...
  NIP [c01658b4] allocate_buf+0x0/0x8
  LR [c0306554] fsl_diu_probe+0x2b4/0x518
  Call Trace:
  [df02be10] [c030638c] fsl_diu_probe+0xec/0x518 (unreliable)
  [df02be60] [c020cdec] of_platform_device_probe+0x5c/0x84
  [df02be80] [c018f5d0] really_probe+0x78/0x1a0
  [df02bea0] [c018f7c0] __driver_attach+0xa4/0xa8
  [df02bec0] [c018ea00] bus_for_each_dev+0x60/0x9c
  [df02bef0] [c018f414] driver_attach+0x24/0x34
  [df02bf00] [c018f168] bus_add_driver+0x12c/0x1cc
  [df02bf20] [c018fbdc] driver_register+0x6c/0x110
  [df02bf30] [c020ccb4] of_register_driver+0x54/0x70
  [df02bf40] [c03d0a50] fsl_diu_init+0x70/0xa4
  ...

This patch fixes the issue.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
2009-04-06 09:12:38 -05:00
Scott Wood
65cc0fa3bd powerpc/pq2fads: Update device tree for use with device-tree-aware u-boot.
Add aliases, and correct CS0 offset to match how u-boot
programs it (this was not a problem with cuImage because
the wrapper would reprogram the localbus to match the device
tree).

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
2009-04-06 09:09:46 -05:00
Scott Wood
4d8107f474 cpm_uart: Disable CPM udbg when re-initing CPM uart, even if not the console.
Previously, if udbg was using the CPM uart, and the normal CPM uart driver
was enabled, but the console was directed elsewhere, udbg would not be
stopped prior to initialization.  This resulted in udbg hanging forever
waiting for the CPM to process a descriptor.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
2009-04-06 09:09:33 -05:00
Scott Wood
bd86ef3784 cpm_uart: Initialize port.dev before it's used.
Previously, this caused NULL to sometimes be passed as a device
to the DMA code.  With recent DMA changes, that now causes a BUG().

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
2009-04-06 09:09:28 -05:00
Kumar Gala
d91e4ea704 powerpc: Add support for CoreInt delivery of interrupts on MPIC
CoreInt provides a mechansim to deliver the IRQ vector directly
into the core on an interrupt (via the SPR EPR) rather than having
to go IACK on the PIC.  This is suppose to provide an improvment
in interrupt latency by reducing the time to get the IRQ vector.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
2009-04-06 09:09:14 -05:00
Thomas Gleixner
81ec5364a5 [MTD] [NAND] Add support for 4KiB pages.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2009-04-06 07:01:56 -07:00
Daniel Glöckner
80fbe6ac9b ASoC: correct s6000 I2S clock polarity
According to the data sheet data is clocked out on the falling edge
and latched on the rising edge of the bit clock. While the left sample
is transmitted the word clock line is low.

Signed-off-by: Daniel Glöckner <dg@emlix.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
2009-04-06 11:18:39 +01:00
Dan Carpenter
09318c47b6 ASoC: Fix null dereference in ak4535_remove()
ak4535_remove() from sound/soc/codecs/ak4535.c calls
i2c_unregister_device() with a possibly null pointer.

This bug was found by smatch (http://repo.or.cz/w/smatch.git/).

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
2009-04-06 10:53:37 +01:00
Darren Hart
52400ba946 futex: add requeue_pi functionality
PI Futexes and their underlying rt_mutex cannot be left ownerless if
there are pending waiters as this will break the PI boosting logic, so
the standard requeue commands aren't sufficient.  The new commands
properly manage pi futex ownership by ensuring a futex with waiters
has an owner at all times.  This will allow glibc to properly handle
pi mutexes with pthread_condvars.

The approach taken here is to create two new futex op codes:

FUTEX_WAIT_REQUEUE_PI:
Tasks will use this op code to wait on a futex (such as a non-pi waitqueue)
and wake after they have been requeued to a pi futex.  Prior to returning to
userspace, they will acquire this pi futex (and the underlying rt_mutex).

futex_wait_requeue_pi() is the result of a high speed collision between
futex_wait() and futex_lock_pi() (with the first part of futex_lock_pi() being
done by futex_proxy_trylock_atomic() on behalf of the top_waiter).

FUTEX_REQUEUE_PI (and FUTEX_CMP_REQUEUE_PI):
This call must be used to wake tasks waiting with FUTEX_WAIT_REQUEUE_PI,
regardless of how many tasks the caller intends to wake or requeue.
pthread_cond_broadcast() should call this with nr_wake=1 and
nr_requeue=INT_MAX.  pthread_cond_signal() should call this with nr_wake=1 and
nr_requeue=0.  The reason being we need both callers to get the benefit of the
futex_proxy_trylock_atomic() routine.  futex_requeue() also enqueues the
top_waiter on the rt_mutex via rt_mutex_start_proxy_lock().

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:03 +02:00
Darren Hart
f801073f87 futex: split out futex value validation code
Refactor the code to validate the expected futex value in order to
reuse it with the requeue_pi code.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:02 +02:00
Darren Hart
9121e4783c futex: distangle futex_requeue()
futex_requeue() is getting a bit long-winded, and will be getting more
so after the requeue_pi patch. Factor out the actual requeueing into a
nicely contained inline function to reduce function length and improve
legibility.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:02 +02:00
Darren Hart
a72188d8a6 futex: add FUTEX_HAS_TIMEOUT flag to restart.futex.flags
Currently restart is only used if there is a timeout. The requeue_pi
functionality requires restarting to futex_lock_pi() on signal after
wakeup in futex_wait_requeue_pi() regardless of if there was a timeout
or not. Using 0 for the timeout value is confusing as that could
indicate an expired timer. The flag makes this explicit. While the
check is not technically needed in futex_wait_restart(), doing so
makes the code consistent with and will avoid confusion should the
need arise to restart wait without a timeout.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:02 +02:00
Darren Hart
8dac456a68 rt_mutex: add proxy lock routines
This patch is a prerequisite for futex requeue_pi. It basically splits
rt_mutex_slowlock() right down the middle, just before the first call
to schedule(). It further adds helper functions which make use of the
split and provide the rt-mutex preliminaries for futex requeue_pi.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:02 +02:00
Darren Hart
dd9739980b futex: split out fixup owner logic from futex_lock_pi()
Refactor the post lock acquisition logic from futex_lock_pi(). This
code will be reused in futex_wait_requeue_pi().

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:02 +02:00
Darren Hart
1a52084d09 futex: split out atomic logic from futex_lock_pi()
Refactor the atomic portion of futex_lock_pi() into futex_lock_pi_atomic().

This logic will be needed by requeue_pi, so modularize it to reduce
code duplication.  The only significant change is passing of the task
to try and take the lock for.  This simplifies the -EDEADLK test as if
the lock is owned by task t, it's a deadlock, regardless of if we are
doing requeue pi or not.  This patch updates the corresponding comment
accordingly.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:01 +02:00
Darren Hart
4b1c486b35 futex: add helper to find the top prio waiter of a futex
Improve legibility by wrapping finding the top waiter in a function.
This will be used by the follow-on patches for enabling requeue pi.

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:01 +02:00
Darren Hart
ca5f9524d6 futex: separate futex_wait_queue_me() logic from futex_wait()
Refactor futex_wait() in preparation for futex_wait_requeue_pi(). In
order to reuse a good chunk of the futex_wait() code for the upcoming
futex_wait_requeue_pi() function, this patch breaks out the
queue-to-wakeup section of futex_wait() into futex_wait_queue_me().

Signed-off-by: Darren Hart <dvhltc@us.ibm.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2009-04-06 11:14:01 +02:00
Russell King
89de3820c7 Merge branch 'devel' of git://git.kernel.org/pub/scm/linux/kernel/git/ycmiao/pxa-linux-2.6 2009-04-06 09:57:38 +01:00
Russell King
3ef48fac6e [ARM] omap: fix omap1 clock usecount decrement bug
Same fix as per a7f8c59, but for OMAP1 instead.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-04-06 09:57:10 +01:00
Peter Zijlstra
92f22a3865 perf_counter: update mmap() counter read
Paul noted that we don't need SMP barriers for the mmap() counter read
because its always on the same cpu (otherwise you can't access the hw
counter anyway).

So remove the SMP barriers and replace them with regular compiler
barriers.

Further, update the comment to include a race free method of reading
said hardware counter. The primary change is putting the pmc_read
inside the seq-loop, otherwise we can still race and read rubbish.

Noticed-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.577951445@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:47 +02:00
Peter Zijlstra
5872bdb88a perf_counter: add more context information
Put in counts to tell which ips belong to what context.

  -----
   | |  hv
   | --
nr | |  kernel
   | --
   | |  user
  -----

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.493101305@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:46 +02:00
Peter Zijlstra
3df70fd623 perf_counter: kerneltop: update to new ABI
Update to reflect the new record_type ABI changes.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.407283141@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:46 +02:00
Peter Zijlstra
c457810ab4 perf_counter: per event wakeups
By request, provide a way to request a wakeup every 'n' events instead
of every page of output.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.323309784@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:45 +02:00
Peter Zijlstra
8a057d8491 perf_counter: move the event overflow output bits to record_type
Per suggestion from Paul, move the event overflow bits to record_type
and sanitize the enums a bit.

Breaks the ABI -- again ;-)

Suggested-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.151921176@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:45 +02:00
Mike Galbraith
9dd499889b perf_counter tools: kerneltop: add real-time data acquisition thread
Decouple kerneltop display from event acquisition by introducing
a separate data acquisition thread. This fixes annnoying kerneltop
display refresh jitter and missed events.

Also add a -r <prio> option, to switch the data acquisition thread
to real-time priority.

Signed-off-by: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Orig-LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:44 +02:00
Peter Zijlstra
4e935e4717 perf_counter: pmc arbitration
Follow the example set by powerpc and try to play nice with oprofile
and the nmi watchdog.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171024.459968444@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:44 +02:00
Peter Zijlstra
d7d59fb323 perf_counter: x86: callchain support
Provide the x86 perf_callchain() implementation.

Code based on the ftrace/sysprof code from Soeren Sandmann Pedersen.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Soeren Sandmann Pedersen <sandmann@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <srostedt@redhat.com>
Orig-LKML-Reference: <20090330171024.341993293@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:43 +02:00
Peter Zijlstra
394ee07623 perf_counter: provide generic callchain bits
Provide the generic callchain support bits. If hw_event->callchain is
set the arch specific perf_callchain() function is called upon to
provide a perf_callchain_entry structure filled with the current
callchain.

If it does so, it is added to the overflow output event.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171024.254266860@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:43 +02:00
Peter Zijlstra
023c54c422 perf_counter tools: kerneltop: update event_types
Go along with the new perf_event_type ABI.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171024.133985461@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:42 +02:00
Peter Zijlstra
5ed00415e3 perf_counter: re-arrange the perf_event_type
Breaks ABI yet again :-)

Change the event type so that [0, 2^31-1] are regular event types, but
[2^31, 2^32-1] forms a bitmask for overflow events.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171024.047961770@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:42 +02:00
Peter Zijlstra
78d613eb12 perf_counter: small cleanup of the output routines
Move the nmi argument to the _begin() function, so that _end() only needs the
handle. This allows the _begin() function to generate a wakeup on event loss.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171023.959404268@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:41 +02:00
Paul Mackerras
31f004df8d perf_counter tools: optionally scale counter values in perfstat mode
Impact: new functionality

This adds add an option to the perfstat mode of kerneltop to scale the
reported counter values according to the fraction of time that each
counter gets to count.  This is invoked with the -l option (I used 'l'
because s, c, a and e were all taken already.)  This uses the new
PERF_RECORD_TOTAL_TIME_{ENABLED,RUNNING} read format options.

With this, we get output like this:

$ ./perfstat -l -e 0:0,0:1,0:2,0:3,0:4,0:5 ./spin

 Performance counter stats for './spin':

     4016072055  CPU cycles           (events)  (scaled from 66.53%)
     2005887318  instructions         (events)  (scaled from 66.53%)
        1762849  cache references     (events)  (scaled from 66.69%)
         165229  cache misses         (events)  (scaled from 66.85%)
     1001298009  branches             (events)  (scaled from 66.78%)
          41566  branch misses        (events)  (scaled from 66.61%)

 Wall-clock time elapsed:  2438.227446 msecs

This also lets us detect when a counter is zero because the counter
never got to go on the CPU at all.  In that case we print <not counted>
rather than 0.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Orig-LKML-Reference: <20090330171023.871484899@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-04-06 09:30:41 +02:00