This patch fixes the sampling interrupt throttling mechanism.

It was broken in v3.2. Events were not being unthrottled. The
unthrottling mechanism required that events be checked at each
timer tick.

This patch solves this problem and also separates:

  - unthrottling
  - multiplexing
  - frequency-mode period adjustments

Not all of them need to be executed at each timer tick.

This third version of the patch is based on my original patch +
PeterZ proposal (https://lkml.org/lkml/2012/1/7/87

).

At each timer tick, for each context:

  - if the current CPU has throttled events, we unthrottle events

  - if context has frequency-based events, we adjust sampling periods

  - if we have reached the jiffies interval, we multiplex (rotate)

We decoupled rotation (multiplexing) from frequency-mode sampling
period adjustments.  They should not necessarily happen at the same
rate. Multiplexing is subject to jiffies_interval (currently at 1
but could be higher once the tunable is exposed via sysfs).

We have grouped frequency-mode adjustment and unthrottling into the
same routine to minimize code duplication. When throttled while in
frequency mode, we scan the events only once.

We have fixed the threshold enforcement code in __perf_event_overflow().
There was a bug whereby it would allow more than the authorized rate
because an increment of hwc->interrupts was not executed at the right
place.

The patch was tested with low sampling limit (2000) and fixed periods,
frequency mode, overcommitted PMU.

On a 2.1GHz AMD CPU:

 $ cat /proc/sys/kernel/perf_event_max_sample_rate
 2000

We set a rate of 3000 samples/sec (2.1GHz/3000 = 700000):

 $ perf record -e cycles,cycles -c 700000  noploop 10
 $ perf report -D | tail -21

 Aggregated stats:
           TOTAL events:      80086
            MMAP events:         88
            COMM events:          2
            EXIT events:          4
        THROTTLE events:      19996
      UNTHROTTLE events:      19996
          SAMPLE events:      40000

 cycles stats:
           TOTAL events:      40006
            MMAP events:          5
            COMM events:          1
            EXIT events:          4
        THROTTLE events:       9998
      UNTHROTTLE events:       9998
          SAMPLE events:      20000

 cycles stats:
           TOTAL events:      39996
        THROTTLE events:       9998
      UNTHROTTLE events:       9998
          SAMPLE events:      20000

For 10s, the cap is 2x2000x10 = 40000 samples.
We get exactly that: 20000 samples/event.

Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: <stable@kernel.org> # v3.2+
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20120126160319.GA5655@quad


Signed-off-by: Ingo Molnar <mingo@elte.hu>

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] cinergyT2-fe: Fix bandwdith settings
  [media] V4L: atmel-isi: add clk_prepare()/clk_unprepare() functions
  [media] cxd2820r: sleep on DVB-T/T2 delivery system switch
  [media] anysee: fix CI init
  [media] cxd2820r: remove unused parameter from cxd2820r_attach
  [media] cxd2820r: fix dvb_frontend_ops

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc32: forced setting of mode of sun4m per-cpu timers

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/microcode_amd: Add support for CPU family specific container files
  x86/amd: Add missing feature flag for fam15h models 10h-1fh processors
  x86/boot-image: Don't leak phdrs in arch/x86/boot/compressed/misc.c::Parse_elf()
  x86/numachip: Drop unnecessary conflict with EDAC
  x86/uv: Fix uninitialized spinlocks
  x86/uv: Fix uv_gpa_to_soc_phys_ram() shift

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf: Call perf_cgroup_event_time() directly
  perf: Don't call release_callchain_buffers() if allocation fails

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  rcu: Add missing __cpuinit annotation in rcutorture code
  sched: Add "const" to is_idle_task() parameter
  rcu: Make rcutorture bool parameters really bool (core code)
  memblock: Fix alloc failure due to dumb underflow protection in memblock_find_in_range_node()

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: Fix assembler constraint to prevent overeager gcc optimisation
  mac_esp: rename irq
  mac_scsi: dont enable mac_scsi irq before requesting it
  macfb: fix black and white modes
  m68k/irq: Remove obsolete IRQ_FLG_* definitions

Fix up trivial conflict in arch/m68k/kernel/process_mm.c as per Geert.

We've decided to provide CPU family specific container files
(starting with CPU family 15h). E.g. for family 15h we have to
load microcode_amd_fam15h.bin instead of microcode_amd.bin

Rationale is that starting with family 15h patch size is larger
than 2KB which was hard coded as maximum patch size in various
microcode loaders (not just Linux).

Container files which include patches larger than 2KB cause
different kinds of trouble with such old patch loaders. Thus we
have to ensure that the default container file provides only
patches with size less than 2KB.

Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Cc: Borislav Petkov <borislav.petkov@amd.com>
Cc: <stable@kernel.org>
Link: http://lkml.kernel.org/r/20120120164412.GD24508@alberich.amd.com


[ documented the naming convention and tidied the code a bit. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>

That is the last one missing for those CPUs.

Others were recently added with commits

 fb215366
 (KVM: expose latest Intel cpu new features (BMI1/BMI2/FMA/AVX2) to guest)

and

 commit 969df4b8


 (x86: Report cpb and eff_freq_ro flags correctly)

Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Link: http://lkml.kernel.org/r/20120120163823.GC24508@alberich.amd.com


Signed-off-by: Ingo Molnar <mingo@elte.hu>

We allocate memory with malloc(), but neglect to free it before
the variable 'phdrs' goes out of scope --> leak.

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Link: http://lkml.kernel.org/r/alpine.LNX.2.00.1201232332590.8772@swampdragon.chaosbits.net


[ Mostly harmless. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>

EDAC detection no longer crashes multi-node systems, so don't
conflict on it with NumaChip.

Signed-off-by: Daniel J Blueman <daniel@numascale-asia.com>
Cc: Steffen Persvold <sp@numascale.com>
Link: http://lkml.kernel.org/r/1327473349-28395-1-git-send-email-daniel@numascale-asia.com


Signed-off-by: Ingo Molnar <mingo@elte.hu>

Initialize two spinlocks in tlb_uv.c and also properly define/initialize
the uv_irq_lock.

The lack of explicit initialization seems to be functionally
harmless, but it is diagnosed when these are turned on:

        CONFIG_DEBUG_SPINLOCK=y
        CONFIG_DEBUG_MUTEXES=y
        CONFIG_DEBUG_LOCK_ALLOC=y
        CONFIG_LOCKDEP=y

Signed-off-by: Cliff Wickman <cpw@sgi.com>
Cc: <stable@kernel.org>
Cc: Dimitri Sivanich <sivanich@sgi.com>
Link: http://lkml.kernel.org/r/E1RnXd1-0003wU-PM@eag09.americas.sgi.com


[ Added the uv_irq_lock initialization fix by Dimitri Sivanich ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>

uv_gpa_to_soc_phys_ram() was inadvertently ignoring the
shift values.  This fix takes the shift into account.

Signed-off-by: Russ Anderson <rja@sgi.com>
Cc: <stable@kernel.org>
Link: http://lkml.kernel.org/r/20120119020753.GA7228@sgi.com


Signed-off-by: Ingo Molnar <mingo@elte.hu>

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: sha512 - reduce stack usage to safe number
  crypto: sha512 - make it work, undo percpu message schedule

Quoth Ben Myers:
 "Please pull in the following bugfix for xfs.  We forgot to drop a lock on
  error in xfs_readlink.  It hasn't been through -next yet, but there is no
  -next tree tomorrow.  The fix is clear so I'm sending this request today."

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: Fix missing xfs_iunlock() on error recovery path in xfs_readlink()

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/ttm: fix two regressions since move_notify changes
  drm/radeon: avoid deadlock if GPU lockup is detected in ib_pool_get
  drm/radeon: silence out possible lock dependency warning
  drm: Fix authentication kernel crash
  gma500: Fix shmem mapping
  drm/radeon/kms: refine TMDS dual link checks
  drm/radeon/kms: use drm_detect_hdmi_monitor for picking encoder mode
  drm/radeon/kms: rework modeset sequence for DCE41 and DCE5
  drm/radeon/kms: move panel mode setup into encoder mode set
  drm/radeon/kms: move disp eng pll setup to init path
  drm/radeon: finish getting bios earlier
  drm/radeon: fix invalid memory access in radeon_atrm_get_bios()
  drm/radeon/kms: add some missing semaphore init
  drm/radeon/kms: Add an MSI quirk for Dell RS690
  gpu, drm, sis: Don't return uninitialized variable from sis_driver_load()

* 'fix/asoc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ASoC: wm2000: Fix use-after-free - don't release_firmware() twice on error
  ASoC: wm8958: Use correct format string in dev_err() call
  ASoC: wm8996: Call _POST_PMU callback for CPVDD
  ASoC: mxs: Fix mxs-saif timeout
  ASoC: Disable register synchronisation for low frequency WM8996 SYSCLK
  ASoC: Don't go through cache when applying WM5100 rev A updates
  ASoC: Mark WM5100 register map cache only when going into BIAS_OFF
  ASoC: tlv320aic32x4: always enable analouge block
  ASoC: tlv320aic32x4: always enable dividers
  ASoC: sgtl5000: Fix wrong register name in restore

A fairly simple bugfix for a WARN_ON() which was triggered in the cache
reset support as a result of some subsequent work.  There's only one
mainline user for the code path that's updated right now (wm8994) so
should be low risk.

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: Reset cache status when reinitialsing the cache

 The data encryption was moved from ecryptfs_write_end into
ecryptfs_writepage, this patch moves the corresponding function
comments to be consistent with the modification.

Signed-off-by: Li Wang <liwang@nudt.edu.cn>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Says Tyler:
 "Tim's logging message update will be really helpful to users when
  they're trying to locate a problematic file in the lower filesystem
  with filename encryption enabled.

  You'll recognize the fix from Li, as you commented on that.

  You should also be familiar with my setattr/truncate improvements,
  since you were the one that pointed them out to us (thanks again!).
  Andrew noted the /dev/ecryptfs write count sanitization needed to be
  improved, so I've got a fix in there for that along with some other
  less important cleanups of the /dev/ecryptfs read/write code."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs:
  eCryptfs: Fix oops when printing debug info in extent crypto functions
  eCryptfs: Remove unused ecryptfs_read()
  eCryptfs: Check inode changes in setattr
  eCryptfs: Make truncate path killable
  eCryptfs: Infinite loop due to overflow in ecryptfs_write()
  eCryptfs: Replace miscdev read/write magic numbers
  eCryptfs: Report errors in writes to /dev/ecryptfs
  eCryptfs: Sanitize write counts of /dev/ecryptfs
  ecryptfs: Remove unnecessary variable initialization
  ecryptfs: Improve metadata read failure logging
  MAINTAINERS: Update eCryptfs maintainer address

If pages passed to the eCryptfs extent-based crypto functions are not
mapped and the module parameter ecryptfs_verbosity=1 was specified at
loading time, a NULL pointer dereference will occur.

Note that this wouldn't happen on a production system, as you wouldn't
pass ecryptfs_verbosity=1 on a production system. It leaks private
information to the system logs and is for debugging only.

The debugging info printed in these messages is no longer very useful
and rather than doing a kmap() in these debugging paths, it will be
better to simply remove the debugging paths completely.

https://launchpad.net/bugs/913651



Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Daniel DeFreez
Cc: <stable@vger.kernel.org>

ecryptfs_read() has been ifdef'ed out for years now and it was
apparently unused before then. It is time to get rid of it for good.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

Most filesystems call inode_change_ok() very early in ->setattr(), but
eCryptfs didn't call it at all. It allowed the lower filesystem to make
the call in its ->setattr() function. Then, eCryptfs would copy the
appropriate inode attributes from the lower inode to the eCryptfs inode.

This patch changes that and actually calls inode_change_ok() on the
eCryptfs inode, fairly early in ecryptfs_setattr(). Ideally, the call
would happen earlier in ecryptfs_setattr(), but there are some possible
inode initialization steps that must happen first.

Since the call was already being made on the lower inode, the change in
functionality should be minimal, except for the case of a file extending
truncate call. In that case, inode_newsize_ok() was never being
called on the eCryptfs inode. Rather than inode_newsize_ok() catching
maximum file size errors early on, eCryptfs would encrypt zeroed pages
and write them to the lower filesystem until the lower filesystem's
write path caught the error in generic_write_checks(). This patch
introduces a new function, called ecryptfs_inode_newsize_ok(), which
checks if the new lower file size is within the appropriate limits when
the truncate operation will be growing the lower file.

In summary this change prevents eCryptfs truncate operations (and the
resulting page encryptions), which would exceed the lower filesystem
limits or FSIZE rlimits, from ever starting.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reviewed-by: Li Wang <liwang@nudt.edu.cn>
Cc: <stable@vger.kernel.org>

ecryptfs_write() handles the truncation of eCryptfs inodes. It grabs a
page, zeroes out the appropriate portions, and then encrypts the page
before writing it to the lower filesystem. It was unkillable and due to
the lack of sparse file support could result in tying up a large portion
of system resources, while encrypting pages of zeros, with no way for
the truncate operation to be stopped from userspace.

This patch adds the ability for ecryptfs_write() to detect a pending
fatal signal and return as gracefully as possible. The intent is to
leave the lower file in a useable state, while still allowing a user to
break out of the encryption loop. If a pending fatal signal is detected,
the eCryptfs inode size is updated to reflect the modified inode size
and then -EINTR is returned.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Cc: <stable@vger.kernel.org>

ecryptfs_write() can enter an infinite loop when truncating a file to a
size larger than 4G. This only happens on architectures where size_t is
represented by 32 bits.

This was caused by a size_t overflow due to it incorrectly being used to
store the result of a calculation which uses potentially large values of
type loff_t.

[tyhicks@canonical.com: rewrite subject and commit message]
Signed-off-by: Li Wang <liwang@nudt.edu.cn>
Signed-off-by: Yunchuan Wen <wenyunchuan@kylinos.com.cn>
Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

ecryptfs_miscdev_read() and ecryptfs_miscdev_write() contained many
magic numbers for specifying packet header field sizes and offsets. This
patch defines those values and replaces the magic values.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

Errors in writes to /dev/ecryptfs were being incorrectly reported by
returning 0 or the value of the original write count.

This patch clears up the return code assignment in error paths.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

A malicious count value specified when writing to /dev/ecryptfs may
result in a a very large kernel memory allocation.

This patch peeks at the specified packet payload size, adds that to the
size of the packet headers and compares the result with the write count
value. The resulting maximum memory allocation size is approximately 532
bytes.

Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Sasha Levin <levinsasha928@gmail.com>
Cc: <stable@vger.kernel.org>

Removes unneeded variable initialization in ecryptfs_read_metadata(). Also adds
a small comment to help explain metadata reading logic.

[tyhicks@canonical.com: Pulled out of for-stable patch and wrote commit msg]
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

Print inode on metadata read failure. The only real
way of dealing with metadata read failures is to delete
the underlying file system file. Having the inode
allows one to 'find . -inum INODE`.

[tyhicks@canonical.com: Removed some minor not-for-stable parts]
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Cc: stable@vger.kernel.org
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

Update my email address in MAINTAINERS.

Signed-off-by: Dustin Kirkland <dustin.kirkland@gazzang.com>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>

Both changes in dc97b340

 cause serious
regressions in the nouveau driver.

move_notify() was originally able to presume that bo->mem is the old node,
and new_mem is the new node.  The above commit moves the call to
move_notify() to after move() has been done, which means that now, sometimes,
new_mem isn't the new node at all, bo->mem is, and new_mem points at a
stale, possibly-just-been-killed-by-move node.

This is clearly not a good situation.  This patch reverts this change, and
replaces it with a cleanup in the move() failure path instead.

The second issue is that the call to move_notify() from cleanup_memtype_use()
causes the TTM ghost objects to get passed into the driver.  This is clearly
bad as the driver knows nothing about these "fake" TTM BOs, and ends up
accessing uninitialised memory.

I worked around this in nouveau's move_notify() hook by ensuring the BO
destructor was nouveau's.  I don't particularly like this solution, and
would rather TTM never pass the driver these objects.  However, I don't
clearly understand the reason why we're calling move_notify() here anyway
and am happy to work around the problem in nouveau instead of breaking the
behaviour expected by other drivers.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Jerome Glisse <j.glisse@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Commit b52a360b

 forgot to call xfs_iunlock() when it detected corrupted
symplink and bailed out. Fix it by jumping to 'out' instead of doing return.

CC: stable@kernel.org
CC: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Alex Elder <elder@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

If GPU lockup is detected in ib_pool get we are holding the ib_pool
mutex that will be needed by the GPU reset code. As ib_pool code is
safe to be reentrant from GPU reset code we should not block if we
are trying to get the ib pool lock on the behalf of the same userspace
caller, thus use the radeon_mutex_lock helper.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Silence out the lock dependency warning by moving bo allocation out
of ib mutex protected section. Might lead to useless temporary
allocation but it's not harmful as such things only happen at
initialization.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

If the master tries to authenticate a client using drm_authmagic and
that client has already closed its drm file descriptor,
either wilfully or because it was terminated, the
call to drm_authmagic will dereference a stale pointer into kmalloc'ed memory
and corrupt it.

Typically this results in a hard system hang.

This patch fixes that problem by removing any authentication tokens
(struct drm_magic_entry) open for a file descriptor when that file
descriptor is closed.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

Quoth Len:
 "This fixes a merge-window regression due to a conflict
  between error injection and preparation to remove atomicio.c
  Here we fix that regression and complete the removal
  of atomicio.c.

  This also re-orders some idle initialization code to
  complete the merge window series that allows cpuidle
  to cope with bringing processors on-line after boot."

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux:
  Use acpi_os_map_memory() instead of ioremap() in einj driver
  ACPI, APEI, EINJ, cleanup 0 vs NULL confusion
  ACPI, APEI, EINJ Allow empty Trigger Error Action Table
  thermal: Rename generate_netlink_event
  ACPI / PM: Add Sony Vaio VPCCW29FX to nonvs blacklist.
  ACPI: Remove ./drivers/acpi/atomicio.[ch]
  ACPI, APEI: Add RAM mapping support to ACPI
  ACPI, APEI: Add 64-bit read/write support for APEI on i386
  ACPI processor hotplug: Delay acpi_processor_start() call for hotplugged cores
  ACPI processor hotplug: Split up acpi_processor_add

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc: Fix build on some non-freescale platforms
  powerpc/powernv: Fix PCI resource handling
  powerpc/crash: Fix build error without SMP
  powerpc/cpuidle: Make it a bool, not a tristate
  powerpc/85xx: Add dr_mode property in USB nodes
  powerpc/85xx: Enable USB2 controller node for P1020RDB
  powerpc/85xx: Fix cmd12 bug and add the chip compatible for eSDHC
  arch/powerpc/sysdev/fsl_pci.c: add missing iounmap
  powerpc: fix compile error with 85xx/p1022_ds.c

Commit 9deaa53a

 broke build
on platforms that use legacy_serial.c without also having
CONFIG_SERIAL_8250_FSL enabled due to an unconditional code
to a routine in that module.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Recent changes to the handling of PCI resources for host bridges
are breaking the PowerNV code for assigning resources on IODA.

The root of the problem is that the pci_bus attached to a host
bridge no longer has its "legacy" resource pointers populated
but only uses the newer list instead.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>