Vince Weaver [Fri, 19 Oct 2012 21:33:38 +0000 (17:33 -0400)]
perf/x86: Remove P6 cpuc->enabled check
Between 2.6.33 and 2.6.34 the PMU code was made modular.
The x86_pmu_enable() call was extended to disable cpuc->enabled
and iterate the counters, enabling one at a time, before calling
enable_all() at the end, followed by re-enabling cpuc->enabled.
Since cpuc->enabled was set to 0, that change effectively caused
the "val |= ARCH_PERFMON_EVENTSEL_ENABLE;" code in p6_pmu_enable_event()
and p6_pmu_disable_event() to be dead code that was never called.
This change removes this code (which was confusing) and adds some
extra commentary to make it more clear what is going on.
Peter Zijlstra [Tue, 23 Oct 2012 15:47:12 +0000 (17:47 +0200)]
sysctl/sched: Fix 'defined but not used' warning
Since commit ("sched/numa: Implement NUMA home-node selection code")
building a kernel with CONFIG_SMP disabled causes the following
warning:
kernel/sysctl.c:259:12: warning: 'min_sched_tunable_scaling' defined but not used [-Wunused-variable]
kernel/sysctl.c:260:12: warning: 'max_sched_tunable_scaling' defined but not used [-Wunused-variable]
Reported-by: Fabio Estevam <fabio.estevam@freescale.com>
[ Ingo preferred extra #ifdef variant over the __maybe_unused ] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/n/tip-p9w5w57ylinrj9zakvhc5zay@git.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
Dave Young [Thu, 18 Oct 2012 06:33:23 +0000 (14:33 +0800)]
Revert "x86/mm: Fix the size calculation of mapping tables"
Commit:
722bc6b16771 x86/mm: Fix the size calculation of mapping tables
Tried to address the issue that the first 2/4M should use 4k pages
if PSE enabled, but extra counts should only be valid for x86_32.
This commit caused a kdump regression: the kdump kernel hangs.
Work is in progress to fundamentally fix the various page table
initialization issues that we have, via the design suggested
by H. Peter Anvin, but it's not ready yet to be merged.
So, to get a working kdump revert to the last known working version,
which is the revert of this commit and of a followup fix (which was
incomplete):
bd2753b2dda7 x86/mm: Only add extra pages count for the first memory range during pre-allocation
Tested kdump on physical and virtual machines.
Signed-off-by: Dave Young <dyoung@redhat.com> Acked-by: Yinghai Lu <yinghai@kernel.org> Acked-by: Cong Wang <xiyou.wangcong@gmail.com> Acked-by: Flavio Leitner <fbl@redhat.com> Tested-by: Flavio Leitner <fbl@redhat.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Cong Wang <xiyou.wangcong@gmail.com> Cc: Flavio Leitner <fbl@redhat.com> Cc: Tejun Heo <tj@kernel.org> Cc: ianfang.cn@gmail.com Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: <stable@kernel.org> Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ingo Molnar [Wed, 24 Oct 2012 07:04:41 +0000 (09:04 +0200)]
numa, sched: Eliminate unused functions
Andrew Morton reported these allnoconfig warnings:
kernel/sched/fair.c:800: warning: 'task_h_load' declared 'static' but never defined
kernel/sched/fair.c:1004: warning: 'account_numa_enqueue' defined but not used
These are only used on CONFIG_SMP - fix it.
We should eventually resolve the Kconfig complexities here by turning
SMP (and NUMA) scheduling either into a separate source code file, or
by creating a single-model scheduler, which happens to build to a small
object file on !CONFIG_SMP or !CONFIG_NUMA kernels not via #ifdefs but
via more clever build time code elimination and zero-size data fields.
That's not a simple patch.
Reported-by: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Link: http://lkml.kernel.org/n/tip-wyctbug9qKulTs0umsxjyixi@git.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
Andre Przywara [Tue, 9 Oct 2012 15:38:35 +0000 (17:38 +0200)]
x86/perf: Fix virtualization sanity check
In check_hw_exists() we try to detect non-emulated MSR accesses
by writing an arbitrary value into one of the PMU registers
and check if it's value after a readout is still the same.
This algorithm silently assumes that the register does not contain
the magic value already, which is wrong in at least one situation.
Fix the algorithm to really do a read-modify-write cycle. This fixes
a warning under Xen under some circumstances on AMD family 10h CPUs.
The reasons in more details actually sound like a story from
Believe It or Not!:
First you need an AMD family 10h/12h CPU. These do not reset the
PERF_CTR registers on a reboot.
Now you boot bare metal Linux, which goes successfully through this
check, but leaves the magic value of 0xabcd in the register. You
don't use the performance counters, but do a reboot (warm reset).
Then you choose to boot Xen. The check will be triggered with a
recent Linux kernel as Dom0 again, trying to write 0xabcd into the
MSR. Xen silently drops the write (expected), but the subsequent read
will return the value in the register, which just happens to be the
expected magic value. Thus the test misleadingly succeeds, leaving
the kernel in the belief that the PMU is available. This will trigger
the following message:
Peter Zijlstra [Mon, 22 Oct 2012 17:21:32 +0000 (19:21 +0200)]
numa, sched, mm: Fix NULL-ptr deref
Dan reported that there's a possible NULL pointer deref in this logic,
fix that. Further fix it to avoid a possible inf. loop (completely
unlikely in the case no vma is migratable). Also don't endlessly loop
on large length, simply truncate at the end of the address-space and
restart on the next go.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Link: http://lkml.kernel.org/n/tip-mnkio02xxtttiepsg9ek6qkw@git.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
Peter Zijlstra [Mon, 22 Oct 2012 18:15:40 +0000 (20:15 +0200)]
numa, sched: Implement slow start for working set sampling
Add a 1 second delay before starting to scan the working set of
a task and starting to balance it amongst nodes.
The theory is that short-run tasks benefit very little from NUMA
placement: they come and go, and they better stick to the node
they were started on. As tasks mature and rebalance to other CPUs
and nodes, so does their NUMA placement have to change and so
does it start to matter more and more.
In practice this change fixes an observable kbuild regression:
# [ a perf stat --null --repeat 10 test of ten bzImage builds to /dev/shm ]
!NUMA:
45.291088843 seconds time elapsed ( +- 0.40% )
45.154231752 seconds time elapsed ( +- 0.36% )
+NUMA, no slow start:
46.172308123 seconds time elapsed ( +- 0.30% )
46.343168745 seconds time elapsed ( +- 0.25% )
The implementation is simple and straightforward, most of the patch
deals with adding the /proc/sys/kernel/sched_numa_scan_delay_ms tunable
knob and with renaming task_period to scan_period.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Link: http://lkml.kernel.org/n/tip-vn7p3ynbwqt3qqewhdlvjltc@git.kernel.org
[ Wrote the changelog, ran measurements, tuned the default. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ingo Molnar [Wed, 24 Oct 2012 05:47:40 +0000 (07:47 +0200)]
Merge tag 'perf-urgent-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent
Pull perf/urgent fixes from Arnaldo Carvalho de Melo:
* Validate syscall id before growing syscall table in 'trace', fixing potential
excessive memory usage.
* Validate perf_sample.raw_data, making 'trace' more robust, avoiding some
potential SEGFAULTs when reading tracepoint fields.
* Fix exclude_guest parse events 'perf test's, from Jiri Olsa.
* Do not flush maps on COMM, that is sent by the kernel when a process is
exec'ed, but also when a process changes its name. Since we were assuming
a COMM always meant an EXEC, we were losing track of a process maps by
flushing its maps. Fix from Luigi Semenzato.
* A recent patch introduced a problem by not initializing what should be
the first kind of pager to use, 'man', instead it was being left as zero
which means no pager. This caused 'perf subcmd --help' to produce no output.
Fix from Namhyung Kim.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Ingo Molnar <mingo@kernel.org>
Linus Torvalds [Wed, 24 Oct 2012 02:17:27 +0000 (05:17 +0300)]
Merge tag 'stable/for-linus-3.7-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen
Pull xen bug-fixes from Konrad Rzeszutek Wilk:
- Fix mysterious SIGSEGV or SIGKILL in applications due to corrupting
of the %eip when returning from a signal handler.
- Fix various ARM compile issues after the merge fallout.
- Continue on making more of the Xen generic code usable by ARM
platform.
- Fix SR-IOV passthrough to mirror multifunction PCI devices.
- Fix various compile warnings.
- Remove hypercalls that don't exist anymore.
* tag 'stable/for-linus-3.7-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen:
xen: dbgp: Fix warning when CONFIG_PCI is not enabled.
xen: arm: comment on why 64-bit xen_pfn_t is safe even on 32 bit
xen: balloon: use correct type for frame_list
xen/x86: don't corrupt %eip when returning from a signal handler
xen: arm: make p2m operations NOPs
xen: balloon: don't include e820.h
xen: grant: use xen_pfn_t type for frame_list.
xen: events: pirq_check_eoi_map is X86 specific
xen: XENMEM_translate_gpfn_list was remove ages ago and is unused.
xen: sysfs: fix build warning.
xen: sysfs: include err.h for PTR_ERR etc
xen: xenbus: quirk uses x86 specific cpuid
xen PV passthru: assign SR-IOV virtual functions to separate virtual slots
xen/xenbus: Fix compile warning.
xen/x86: remove duplicated include from enlighten.c
Linus Torvalds [Wed, 24 Oct 2012 01:07:51 +0000 (04:07 +0300)]
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Ingo Molnar:
"Most of these are uprobes race fixes from Oleg, and their preparatory
cleanups. (It's larger than what I'd normally send for an -rc kernel,
but they looked significant enough to not delay them.)
There's also an oprofile fix and an uncore PMU fix."
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (22 commits)
perf/x86: Disable uncore on virtualized CPUs
oprofile, x86: Fix wrapping bug in op_x86_get_ctrl()
ring-buffer: Check for uninitialized cpu buffer before resizing
uprobes: Fix the racy uprobe->flags manipulation
uprobes: Fix prepare_uprobe() race with itself
uprobes: Introduce prepare_uprobe()
uprobes: Fix handle_swbp() vs unregister() + register() race
uprobes: Do not delete uprobe if uprobe_unregister() fails
uprobes: Don't return success if alloc_uprobe() fails
uprobes/x86: Only rep+nop can be emulated correctly
uprobes: Simplify is_swbp_at_addr(), remove stale comments
uprobes: Kill set_orig_insn()->is_swbp_at_addr()
uprobes: Introduce copy_opcode(), kill read_opcode()
uprobes: Kill set_swbp()->is_swbp_at_addr()
uprobes: Restrict valid_vma(false) to skip VM_SHARED vmas
uprobes: Change valid_vma() to demand VM_MAYEXEC rather than VM_EXEC
uprobes: Change write_opcode() to use FOLL_FORCE
uprobes: Move clear_thread_flag(TIF_UPROBE) to uprobe_notify_resume()
uprobes: Kill UTASK_BP_HIT state
uprobes: Fix UPROBE_SKIP_SSTEP checks in handle_swbp()
...
Linus Torvalds [Wed, 24 Oct 2012 01:07:02 +0000 (04:07 +0300)]
Merge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull core kernel fixes from Ingo Molnar:
"Two small fixes"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
Documentation: Reflect the new location of the NMI watchdog info
nohz: Fix idle ticks in cpu summary line of /proc/stat
Linus Torvalds [Wed, 24 Oct 2012 01:05:56 +0000 (04:05 +0300)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Pull s390 fixes from Martin Schwidefsky:
"Among the usual minor bug fixes the more interesting patches are the
perf counters for the latest machine, the missing select to enable
transparent huge pages and a build fix for the UAPI rework."
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
s390,uapi: do not use uapi/asm-generic/kvm_para.h
s390/cache: fix data/instruction cache output
s390: fix linker script for 31 bit builds
s390/thp: select HAVE_ARCH_TRANSPARENT_HUGEPAGE
s390/kdump: Use 64 bit mode for 0x10000 entry point
perf_cpum_cf: Add support for counters available with IBM zEC12
s390/css: stop stsch loop after cc 3
s390/cio: use generic bitmap functions
s390/chpid: make headers usable (again)
Linus Torvalds [Wed, 24 Oct 2012 01:05:15 +0000 (04:05 +0300)]
Merge branch 'stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile
Pull tile fixes from Chris Metcalf:
"This fixes one issue with compiler flags that can cause modules not to
load, and cleans up some warnings with ELF_R_xxx defines."
* 'stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
arch/tile: avoid build warnings from duplicate ELF_R_xxx #defines
arch/tile: avoid generating .eh_frame information in modules
Chris Metcalf [Fri, 19 Oct 2012 15:43:11 +0000 (11:43 -0400)]
arch/tile: avoid generating .eh_frame information in modules
The tile tool chain uses the .eh_frame information for backtracing.
The vmlinux build drops any .eh_frame sections at link time, but when
present in kernel modules, it causes a module load failure due to the
presence of unsupported pc-relative relocations. When compiling to
use compiler feedback support, the compiler by default omits .eh_frame
information, so we don't see this problem. But when not using feedback,
we need to explicitly suppress the .eh_frame.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com> Cc: stable@vger.kernel.org
Rik van Riel [Thu, 18 Oct 2012 21:20:21 +0000 (17:20 -0400)]
numa, mm: Rename the PROT_NONE fault handling functions to *_numa()
Having the function name indicate what the function is used
for makes the code a little easier to read. Furthermore,
the fault handling code largely consists of do_...._page
functions.
Rename the NUMA working set sampling fault handling functions
to _numa() names, to indicate what they are used for.
This separates the naming from the regular PROT_NONE namings.
Andrea Arcangeli [Tue, 18 Sep 2012 00:14:51 +0000 (02:14 +0200)]
x86, mm: Prevent gcc to re-read the pagetables
GCC is very likely to read the pagetables just once and cache them in
the local stack or in a register, but it is can also decide to re-read
the pagetables. The problem is that the pagetable in those places can
change from under gcc.
In the page fault we only hold the ->mmap_sem for reading and both the
page fault and MADV_DONTNEED only take the ->mmap_sem for reading and we
don't hold any PT lock yet.
In get_user_pages_fast() the TLB shootdown code can clear the pagetables
before firing any TLB flush (the page can't be freed until the TLB
flushing IPI has been delivered but the pagetables will be cleared well
before sending any TLB flushing IPI).
With THP/hugetlbfs the pmd (and pud for hugetlbfs giga pages) can
change as well under gup_fast, it won't just be cleared for the same
reasons described above for the pte in the page fault case.
[ This patch was picked up from the AutoNUMA tree. ]
Originally-by: Andrea Arcangeli <aarcange@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com>
[ Ported to this tree, because we are modifying the page tables
at a high rate here, so this problem is potentially more
likely to show up in practice. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
Mel Gorman [Wed, 12 Oct 2011 19:06:51 +0000 (21:06 +0200)]
mm: Check if PTE is already allocated during page fault
With transparent hugepage support, handle_mm_fault() has to be careful
that a normal PMD has been established before handling a PTE fault. To
achieve this, it used __pte_alloc() directly instead of pte_alloc_map
as pte_alloc_map is unsafe to run against a huge PMD. pte_offset_map()
is called once it is known the PMD is safe.
pte_alloc_map() is smart enough to check if a PTE is already present
before calling __pte_alloc but this check was lost. As a consequence,
PTEs may be allocated unnecessarily and the page table lock taken.
Thi useless PTE does get cleaned up but it's a performance hit which
is visible in page_test from aim9.
This patch simply re-adds the check normally done by pte_alloc_map to
check if the PTE needs to be allocated before taking the page table
lock. The effect is noticable in page_test from aim9.
(While this affects 2.6.38, it is a performance rather than a
functional bug and normally outside the rules -stable. While the big
performance differences are to a microbench, the difference in fork
and exec performance may be significant enough that -stable wants to
consider the patch)
Reported-by: Raz Ben Yehuda <raziebe@gmail.com> Signed-off-by: Mel Gorman <mgorman@suse.de> Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Rik van Riel <riel@redhat.com>
[ Picked this up from the AutoNUMA tree to help
it upstream and to allow apples-to-apples
performance comparisons. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
Linus Torvalds [Tue, 23 Oct 2012 05:51:07 +0000 (08:51 +0300)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Pull drm fixes from Dave Airlie:
"Fixes for intel and nouveau mainly.
- intel: disable HSW by default, sdvo fixes, link train regression
fix
- nouveau: acpi rom loading regression fix, with a few other fixes
from the rework
-core: just other minor fixes and race fixes for ttm."
* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (24 commits)
drm/ttm: Fix a theoretical race in ttm_bo_cleanup_refs()
drm/ttm: Fix a theoretical race
drm: platform: Don't initialize driver-private data
drm/debugfs: remove redundant info from gem_names
drm: fb: cma: Fail gracefully on allocation failure
drm: fb: cma: Fix typo in debug message
drm/nouveau/clock: fix missing pll type/addr when matching default entry
drm/nouveau/fb: fix reporting of memory type on GF8+ IGPs
drm/nv41/vm: don't init hw pciegart on boards with agp bridge
drm/nouveau/bios: fetch full 4KiB block to determine ACPI ROM image size
drm/nouveau: validate vbios size
drm/nouveau: warn when trying to free mm which is still in use
drm/nouveau: fix nouveau_mm/nouveau_mm_node leak
drm/nouveau/bios: improve error handling when reading the vbios from ACPI
drm/nouveau: handle same-fb page flips
drm/i915: Initialize obj->pages before use by i915_gem_object_do_bit17_swizzle()
drm/i915: Add no-lvds quirk for Supermicro X7SPA-H
drm/i915: Insert i915_preliminary_hw_support variable.
drm/i915: shut up spurious WARN in the gtt fault handler
Revert "drm/i915: Try harder to complete DP training pattern 1"
...
Linus Torvalds [Tue, 23 Oct 2012 05:48:26 +0000 (08:48 +0300)]
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 fixes from Ted Ts'o:
"Various bug fixes for ext4. The most serious of them fixes a security
bug (CVE-2012-4508) which leads to stale data exposure when we have
fallocate racing against writes to files undergoing delayed
allocation. We also have two fixes for the metadata checksum feature,
the most serious of which can cause the superblock to have a invalid
checksum after a power failure."
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: Avoid underflow in ext4_trim_fs()
ext4: Checksum the block bitmap properly with bigalloc enabled
ext4: fix undefined bit shift result in ext4_fill_flex_info
ext4: fix metadata checksum calculation for the superblock
ext4: race-condition protection for ext4_convert_unwritten_extents_endio
ext4: serialize fallocate with ext4_convert_unwritten_extents
Linus Torvalds [Tue, 23 Oct 2012 05:47:38 +0000 (08:47 +0300)]
Merge tag 'nfs-for-3.7-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Pull NFS client bugfixes from Trond Myklebust:
- Do not call pnfs_return_layout() from an rpciod context
- nfs4_ds_disconnect can cause Oopses. Kill it...
- Fix the return value for nfs_callback_start_svc
- Fix a number of compile warnings
* tag 'nfs-for-3.7-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
NFSv4: Fix the return value for nfs_callback_start_svc
NFSv4.1: Declare osd_pri_2_pnfs_err(), objio_init_read/write to be static
NFSv4: fs/nfs/nfs4getroot.c needs to include "internal.h"
NFSv4.1: Use kcalloc() to allocate zeroed arrays instead of kzalloc()
NFSv4.1: Do not call pnfs_return_layout() from an rpciod context
NFSv4.1: Kill nfs4_ds_disconnect()
Thomas Hellstrom [Mon, 22 Oct 2012 12:51:26 +0000 (12:51 +0000)]
drm/ttm: Fix a theoretical race in ttm_bo_cleanup_refs()
In theory, that function could release the lru lock between
checking for bo on ddestroy list and a successful reserve if the bo
was already reserved, and the function was called with waiting reserves
allowed.
However, all current reservers of a bo on the ddestroy list would
atomically take the bo off the list after a successful reserve so this
race should not have been hit, so no need to backport for stable.
This patch also fixes a case found by Maarten Lankhorst where
ttm_mem_evict_first called with no_wait_gpu would incorrectly
spin waiting for bo idle if trying to evict a busy buffer that
also sits on the ddestroy list.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Thomas Hellstrom [Mon, 22 Oct 2012 12:51:25 +0000 (12:51 +0000)]
drm/ttm: Fix a theoretical race
The ttm_mem_evict_first function could theoretically drop the
lru lock without retrying if a reservation from off the LRU list
ended up waiting.
However, since currently there are no users that could cause a wait
in that situation so this is not suitable for stable
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Thierry Reding [Mon, 15 Oct 2012 18:03:42 +0000 (18:03 +0000)]
drm: platform: Don't initialize driver-private data
Platform device drivers usually use the driver-private data for their
own purposes. Having it overwritten by drm_platform_init() is confusing
and error-prone.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de> Signed-off-by: Dave Airlie <airlied@redhat.com>
Marcin Slusarz [Tue, 16 Oct 2012 21:47:35 +0000 (21:47 +0000)]
drm/debugfs: remove redundant info from gem_names
It's a relic of "drm: Convert proc files to seq_file and introduce debugfs",
which wrongly converted DRM_INFO + sprintf to 2 seq_printfs.
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com> Cc: Ben Gamari <bgamari@gmail.com> Cc: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
Thierry Reding [Sat, 20 Oct 2012 10:32:47 +0000 (10:32 +0000)]
drm: fb: cma: Fail gracefully on allocation failure
The drm_gem_cma_create() function never returns NULL but rather an error
encoded in the return value using the ERR_PTR() macro. Callers therefore
need to check for errors using the IS_ERR() macro. This change allows
drivers to handle contiguous DMA allocation failures gracefully.
Lukas Czerner [Mon, 22 Oct 2012 22:01:19 +0000 (18:01 -0400)]
ext4: Avoid underflow in ext4_trim_fs()
Currently if len argument in ext4_trim_fs() is smaller than one block,
the 'end' variable underflow. Avoid that by returning EINVAL if len is
smaller than file system block.
which enables 'exclude_guest' modifier any time the 'precise'
modifier is detected.
Fixing related tests and adding special comment.
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Jiri Olsa <jolsa@redhat.com> Tested-by: David Ahern <dsahern@gmail.com> Acked-by: David Ahern <dsahern@gmail.com> Cc: David Ahern <dsahern@gmail.com> Cc: Namhyung Kim <namhyung@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Luigi Semenzato [Tue, 21 Aug 2012 21:52:20 +0000 (14:52 -0700)]
perf tools: do not flush maps on COMM for perf report
This fixes a long-standing bug caused by the lack of separate COMM and EXEC
record types, which makes "perf report" lose track of symbols when a process
renames itself.
With this fix (suggested by Stephane Eranian), a COMM (rename) no longer
flushes the maps, which is the correct behavior. An EXEC also no longer
flushes the maps, but this doesn't matter because as new mappings are created
(for the executable and the libraries) the old mappings are automatically
removed. This is not by accident: the functionality is necessary because DLLs
can be explicitly loaded at any time with dlopen(), possibly on top of existing
text, so "perf report" handles correctly the clobbering of new mappings on top
of old ones.
An alternative patch (which I proposed earlier) would be to introduce a
separate PERF_RECORD_EXEC type, but it is a much larger change (about 300
lines) and is not necessary.
Signed-off-by: Luigi Semenzato <semenzato@chromium.org> Tested-by: Stephane Eranian <eranian@google.com> Acked-by: Stephane Eranian <eranian@google.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Andi Kleen <ak@linux.intel.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Lucas De Marchi <lucas.demarchi@profusion.mobi> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Olof Johansson <olofj@chromium.org> Cc: Paul Gortmaker <paul.gortmaker@windriver.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Robert Richter <robert.richter@amd.com> Cc: Sonny Rao <sonnyrao@chromium.org> Cc: Stephane Eranian <eranian@google.com> Cc: Stephen Wilson <wilsons@start.ca> Cc: Tejun Heo <tj@kernel.org> Cc: Vasiliy Kulikov <segoon@openwall.com> Link: http://lkml.kernel.org/r/1345585940-6497-1-git-send-email-semenzato@chromium.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Namhyung Kim [Mon, 22 Oct 2012 07:12:23 +0000 (16:12 +0900)]
perf help: Fix --help for builtins
It seems that commit cc5848213329 ("perf help: Remove use of die and
handle errors") caused the problem - it changed the initial value of
'help_format' from HELP_FORMAT_MAN to HELP_FORMAT_NONE.
This broke the --help option for all builtins, that would produce no
output, while 'man perf-top' would work it MANPATH is properly setup.
Reported-by: Stephane Eranian <eranian@google.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Tested-by: David Ahern <dsahern@gmail.com> Acked-by: David Ahern <dsahern@gmail.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/87r4orj7zc.fsf@sejong.aot.lge.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Dave Airlie [Mon, 22 Oct 2012 07:50:07 +0000 (17:50 +1000)]
Merge branch 'drm-nouveau-fixes' of git://git.freedesktop.org/git/nouveau/linux-2.6 into drm-fixes
Fixes from Ben, off note:
ACPI ROM regression fix,
some IGP and AGP regressions fixes from rework fallout.
* 'drm-nouveau-fixes' of git://git.freedesktop.org/git/nouveau/linux-2.6:
drm/nouveau/clock: fix missing pll type/addr when matching default entry
drm/nouveau/fb: fix reporting of memory type on GF8+ IGPs
drm/nv41/vm: don't init hw pciegart on boards with agp bridge
drm/nouveau/bios: fetch full 4KiB block to determine ACPI ROM image size
drm/nouveau: validate vbios size
drm/nouveau: warn when trying to free mm which is still in use
drm/nouveau: fix nouveau_mm/nouveau_mm_node leak
drm/nouveau/bios: improve error handling when reading the vbios from ACPI
drm/nouveau: handle same-fb page flips
Dmitry Torokhov [Mon, 22 Oct 2012 00:57:20 +0000 (17:57 -0700)]
Input: fix use-after-free introduced with dynamic minor changes
Commit 7f8d4cad1e4e ("Input: extend the number of event (and other)
devices") made evdev, joydev and mousedev to embed struct cdev into
their respective structures representing input devices.
Unfortunately character device structure may outlive the parent
structure unless we do not set it up as parent of character device so
that it will stay pinned until character device is freed.
Also, now that parent structure is pinned while character device exists
we do not need to pin and unpin it every time user opens or closes it.
Reported-by: Dave Jones <davej@redhat.com> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> Acked-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dmitry Torokhov [Mon, 22 Oct 2012 00:57:19 +0000 (17:57 -0700)]
char_dev: pin parent kobject
In certain cases (for example when a cdev structure is embedded into
another object whose lifetime is controlled by a separate kobject) it is
beneficial to tie lifetime of another object to the lifetime of
character device so that related object is not freed until after
char_dev object is freed.
To achieve this let's pin kobject's parent when doing cdev_add() and
unpin when last reference to cdev structure is being released.
Ben Skeggs [Mon, 22 Oct 2012 04:10:16 +0000 (14:10 +1000)]
drm/nouveau/clock: fix missing pll type/addr when matching default entry
This issue is a regression from 70790f4f819875e8f390871fd15bbbf823f28e1b,
and causes us to miss a special-case for C51 (NV4E) chipsets and return
the wrong reference frequency for the VPLLs.
Tao Ma [Mon, 22 Oct 2012 04:34:32 +0000 (00:34 -0400)]
ext4: Checksum the block bitmap properly with bigalloc enabled
In mke2fs, we only checksum the whole bitmap block and it is right.
While in the kernel, we use EXT4_BLOCKS_PER_GROUP to indicate the
size of the checksumed bitmap which is wrong when we enable bigalloc.
The right size should be EXT4_CLUSTERS_PER_GROUP and this patch fixes
it.
Also as every caller of ext4_block_bitmap_csum_set and
ext4_block_bitmap_csum_verify pass in EXT4_BLOCKS_PER_GROUP(sb)/8,
we'd better removes this parameter and sets it in the function itself.
Signed-off-by: Tao Ma <boyu.mt@taobao.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Lukas Czerner <lczerner@redhat.com> Cc: stable@vger.kernel.org
Marcin Slusarz [Sun, 21 Oct 2012 23:59:20 +0000 (09:59 +1000)]
drm/nouveau: validate vbios size
Without checking, we could detect vbios size as 0, allocate 0-byte array
(kmalloc returns invalid pointer for such allocation) and crash in
nouveau_bios_score while checking for vbios signature.
Reported-by: Heinz Diehl <htd@fritha.org> Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Marcin Slusarz [Fri, 5 Oct 2012 10:26:32 +0000 (12:26 +0200)]
drm/nouveau: handle same-fb page flips
It's questionable use case, but weston/wayland already relies on this
behaviour, and other drivers don't care about it, so it's a matter of
compatibility. Without it, process invoking such page flip hangs in
unkillable state, trying to reserve the same buffer twice.
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Sometimes we're segfaulting because we were expecting that the
perf_sample.raw_data field was set as requested, but in some cases
that needs further investigation, that field can be NULL, leading
to segfaults.
Make the tool more robust by checking that before calling any per event
handlers that may try to use that field.
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-g1fmodl6ys4lq8honbj1igoi@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
perf trace: Validate syscall id before growing syscall table
In some cases the ID for a syscall read thru the raw_syscalls tracepoint
is bogus, still needs to be investigated why, but to make the tool more
robust first try to resolve the ID to a name via libaudit and if it
fails, don't grow the table.
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-0lsokw3xor7c4ijo45u6bauh@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Dave Airlie [Sun, 21 Oct 2012 23:55:29 +0000 (09:55 +1000)]
Merge branch 'drm-intel-fixes' of git://people.freedesktop.org/~danvet/drm-intel into drm-fixes
Daniel writes:
The big thing is the disabling of the hsw support by default, cc: stable.
We've aimed for basic hsw support in 3.6, but due to a few bad
happenstances we've screwed up and only 3.8 will have better modeset
support than vesa. To avoid yet another round of fallout from such a
gaffle on for the next platform we've added a module option to disable
early hw support by default. That should also give us more flexibility in
bring-up.
Otherwise just small fixes:
- 3 fixes from Egbert for sdvo corner cases
- invert-brightness quirk entry from Egbert
- revert a dp link training change, it regresses some setups
- and shut up a spurious WARN in our gem fault handler.
- regression fix for an oops on bit17 swizzling machines, introduce in 3.7
- another no-lvds quirk
* 'drm-intel-fixes' of git://people.freedesktop.org/~danvet/drm-intel:
drm/i915: Initialize obj->pages before use by i915_gem_object_do_bit17_swizzle()
drm/i915: Add no-lvds quirk for Supermicro X7SPA-H
drm/i915: Insert i915_preliminary_hw_support variable.
drm/i915: shut up spurious WARN in the gtt fault handler
Revert "drm/i915: Try harder to complete DP training pattern 1"
DRM/i915: Restore sdvo_flags after dtd->mode->dtd Roundrtrip.
DRM/i915: Don't clone SDVO LVDS with analog.
DRM/i915: Add QUIRK_INVERT_BRIGHTNESS for NCR machines.
DRM/i915: Don't delete DPLL Multiplier during DAC init.
Jean Delvare [Sun, 21 Oct 2012 10:05:51 +0000 (12:05 +0200)]
Documentation: Reflect the new location of the NMI watchdog info
Commit 9919cba7 ("watchdog: Update documentation") moved the
NMI watchdog documentation from nmi_watchdog.txt to
lockup-watchdogs.txt. Update the index file accordingly.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp> Cc: Randy Dunlap <rdunlap@xenotime.net> Cc: Don Zickus <dzickus@redhat.com> Link: http://lkml.kernel.org/r/20121021120551.4656d99b@endymion.delvare Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ingo Molnar [Sun, 21 Oct 2012 16:16:39 +0000 (18:16 +0200)]
Merge tag 'numascale_mce_fix' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras into x86/urgent
Pull a fix from Borislav Petkov:
"Fix for a bug causing an OOPS on confederate systems from Numascale
where northbridge descriptors are not unique, causing a lookup failure
of the northbridge descriptor in the amd_nb.c code.
numa, mm: Fix NUMA hinting page faults from gup/gup_fast
Introduce FOLL_NUMA to tell follow_page to check
pte/pmd_numa. get_user_pages must use FOLL_NUMA, and it's safe to do
so because it always invokes handle_mm_fault and retries the
follow_page later.
KVM secondary MMU page faults will trigger the NUMA hinting page
faults through gup_fast -> get_user_pages -> follow_page ->
handle_mm_fault.
Other follow_page callers like KSM should not use FOLL_NUMA, or they
would fail to get the pages if they use follow_page instead of
get_user_pages.
[ This patch was picked up from the AutoNUMA tree. ]
Originally-by: Andrea Arcangeli <aarcange@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com>
[ ported to this tree. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
Peter Zijlstra [Sun, 14 Oct 2012 14:59:13 +0000 (16:59 +0200)]
sched, numa, mm: Implement constant, per task Working Set Sampling (WSS) rate
Previously, to probe the working set of a task, we'd use
a very simple and crude method: mark all of its address
space PROT_NONE.
That method has various (obvious) disadvantages:
- it samples the working set at dissimilar rates,
giving some tasks a sampling quality advantage
over others.
- creates performance problems for tasks with very
large working sets
- over-samples processes with large address spaces but
which only very rarely execute
Improve that method by keeping a rotating offset into the
address space that marks the current position of the scan,
and advance it by a constant rate (in a CPU cycles execution
proportional manner). If the offset reaches the last mapped
address of the mm then it then it starts over at the first
address.
The per-task nature of the working set sampling functionality
in this tree allows such constant rate, per task,
execution-weight proportional sampling of the working set,
with an adaptive sampling interval/frequency that goes from
once per 100 msecs up to just once per 1.6 seconds.
The current sampling volume is 256 MB per interval.
As tasks mature and converge their working set, so does the
sampling rate slow down to just a trickle, 256 MB per 1.6
seconds of CPU time executed.
This, beyond being adaptive, also rate-limits rarely
executing systems and does not over-sample on overloaded
systems.
[ In AutoNUMA speak, this patch deals with the effective sampling
rate of the 'hinting page fault'. AutoNUMA's scanning is
currently rate-limited, but it is also fundamentally
single-threaded, executing in the knuma_scand kernel thread,
so the limit in AutoNUMA is global and does not scale up with
the number of CPUs, nor does it scan tasks in an execution
proportional manner.
So the idea of rate-limiting the scanning was first implemented
in the AutoNUMA tree via a global rate limit. This patch goes
beyond that by implementing an execution rate proportional
working set sampling rate that is not implemented via a single
global scanning daemon. ]
Based-on-idea-by: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Link: http://lkml.kernel.org/n/tip-wt5b48o2226ec63784i58s3j@git.kernel.org
[ Wrote changelog and fixed bug. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
Linus Torvalds [Sat, 20 Oct 2012 16:48:10 +0000 (09:48 -0700)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64
Pull arm64 fixes from Catalin Marinas:
"Main changes:
- AArch64 Linux compilation fixes following 3.7-rc1 changes
(MODULES_USE_ELF_RELA, update_vsyscall() prototype)
- Unnecessary register setting in start_thread() (thanks to Al Viro)
- ptrace fixes"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64:
arm64: fix alignment padding in assembly code
arm64: ptrace: use HW_BREAKPOINT_EMPTY type for disabled breakpoints
arm64: ptrace: make structure padding explicit for debug registers
arm64: No need to set the x0-x2 registers in start_thread()
arm64: Ignore memory blocks below PHYS_OFFSET
arm64: Fix the update_vsyscall() prototype
arm64: Select MODULES_USE_ELF_RELA
arm64: Remove duplicate inclusion of mmu_context.h in smp.c
Marc Zyngier [Fri, 19 Oct 2012 16:33:27 +0000 (17:33 +0100)]
arm64: fix alignment padding in assembly code
An interesting effect of using the generic version of linkage.h
is that the padding is defined in terms of x86 NOPs, which can have
even more interesting effects when the assembly code looks like this:
ENTRY(func1)
mov x0, xzr
ENDPROC(func1)
// fall through
ENTRY(func2)
mov x0, #1
ret
ENDPROC(func2)
Admittedly, the code is not very nice. But having code from another
architecture doesn't look completely sane either.
The fix is to add arm64's version of linkage.h, which causes the insertion
of proper AArch64 NOPs.
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Yan, Zheng [Tue, 21 Aug 2012 09:08:37 +0000 (17:08 +0800)]
perf/x86: Disable uncore on virtualized CPUs
Initializing uncore PMU on virtualized CPU may hang the kernel.
This is because kvm does not emulate the entire hardware. Thers
are lots of uncore related MSRs, making kvm enumerate them all
is a non-trival task. So just disable uncore on virtualized CPU.
projects / karo-tx-linux.git / log