22 Jul, 2015
3 commits
-
region_is_ram() looks up the iomem_resource table to check if
a target range is in RAM. However, it always returns with -1
due to invalid range checks. It always breaks the loop at the
first entry of the table.Another issue is that it compares p->flags and flags, but it always
fails. flags is declared as int, which makes it as a negative value
with IORESOURCE_BUSY (0x80000000) set while p->flags is unsigned long.Fix the range check and flags so that region_is_ram() works as
advertised.Signed-off-by: Toshi Kani
Reviewed-by: Dan Williams
Cc: Mike Travis
Cc: Luis R. Rodriguez
Cc: Andrew Morton
Cc: Roland Dreier
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1437088996-28511-4-git-send-email-toshi.kani@hp.com
Signed-off-by: Thomas Gleixner -
__ioremap_caller() calls region_is_ram() to walk through the
iomem_resource table to check if a target range is in RAM, which was
added to improve the lookup performance over page_is_ram() (commit
906e36c5c717 "x86: use optimized ioresource lookup in ioremap
function"). page_is_ram() was no longer used when this change was
added, though.__ioremap_caller() then calls walk_system_ram_range(), which had
replaced page_is_ram() to improve the lookup performance (commit
c81c8a1eeede "x86, ioremap: Speed up check for RAM pages").Since both checks walk through the same iomem_resource table for
the same purpose, there is no need to call both functions.Aside of that walk_system_ram_range() is the only useful check at the
moment because region_is_ram() always returns -1 due to an
implementation bug. That bug in region_is_ram() cannot be fixed
without breaking existing ioremap callers, which rely on the subtle
difference of walk_system_ram_range() versus non page aligned ranges.Once these offending callers are fixed we can use region_is_ram() and
remove walk_system_ram_range().[ tglx: Massaged changelog ]
Signed-off-by: Toshi Kani
Reviewed-by: Dan Williams
Cc: Roland Dreier
Cc: Mike Travis
Cc: Luis R. Rodriguez
Cc: H. Peter Anvin
Cc: Ingo Molnar
Cc: Borislav Petkov
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1437088996-28511-3-git-send-email-toshi.kani@hp.com
Signed-off-by: Thomas Gleixner -
__ioremap_check_ram() has a WARN_ONCE() which is emitted when the
given pfn range is not RAM. The warning is bogus in two aspects:- it never triggers since walk_system_ram_range() only calls
__ioremap_check_ram() for RAM ranges.- the warning message is wrong as it says: "ioremap on RAM' after it
established that the pfn range is not RAM.Move the WARN_ONCE() to __ioremap_caller(), and update the message to
include the address range so we get an actual warning when something
tries to ioremap system RAM.[ tglx: Massaged changelog ]
Signed-off-by: Toshi Kani
Reviewed-by: Dan Williams
Cc: Roland Dreier
Cc: Luis R. Rodriguez
Cc: H. Peter Anvin
Cc: Ingo Molnar
Cc: Borislav Petkov
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1437088996-28511-2-git-send-email-toshi.kani@hp.com
Signed-off-by: Thomas Gleixner
21 Jul, 2015
9 commits
-
Pull an EFI fix from Matt Fleming:
- Fix a bug in the Common Platform Error Record (CPER) driver that
caused old UEFI spec (< 2.3) versions of the memory error record
structure to be declared invalid. (Tony Luck)Signed-off-by: Ingo Molnar
-
On built-in kernels this warning will always splat, even if no ivtvfb
hardware is present, as this is part of the module init:if (WARN(pat_enabled(),
"ivtvfb needs PAT disabled, boot with nopat kernel parameter\n")) {Fix that by shifting the PAT requirement check out under the code
that does the "quasi-probe" for the device.This device driver relies on an existing driver to find its own devices,
it looks for that device driver and its own found devices, then uses
driver_for_each_device() to try to see if it can probe each of those
devices as a frambuffer device with ivtvfb_init_card().We tuck the PAT requiremenet check then on the ivtvfb_init_card() call
making the check at least require an ivtv device present before
complaining.Reported-by: Fengguang Wu [0-day test robot]
Signed-off-by: Luis R. Rodriguez
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: andy@silverblocksystems.net
Cc: benh@kernel.crashing.org
Cc: bp@suse.de
Cc: dan.j.williams@intel.com
Cc: dledford@redhat.com
Cc: jkosina@suse.cz
Cc: julia.lawall@lip6.fr
Cc: luto@amacapital.net
Cc: mchehab@osg.samsung.com
Link: http://lkml.kernel.org/r/1437167245-28273-3-git-send-email-mcgrof@do-not-panic.com
Signed-off-by: Ingo Molnar -
WARN() may confuse users, fix that. ipath_init_one() is part the
device's probe so this would only be triggered if a
corresponding device was found.Signed-off-by: Luis R. Rodriguez
Acked-by: Doug Ledford
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: andy@silverblocksystems.net
Cc: benh@kernel.crashing.org
Cc: bp@suse.de
Cc: dan.j.williams@intel.com
Cc: jkosina@suse.cz
Cc: julia.lawall@lip6.fr
Cc: luto@amacapital.net
Cc: mchehab@osg.samsung.com
Link: http://lkml.kernel.org/r/1437167245-28273-2-git-send-email-mcgrof@do-not-panic.com
Signed-off-by: Ingo Molnar -
Make WT really mean WT (rather than UC).
I can't see why commit 9cd25aac1f ("x86/mm/pat: Emulate PAT when
it is disabled") didn't make this to match its changes to
pat_init().Signed-off-by: Jan Beulich
Cc: Andy Lutomirski
Cc: Borislav Petkov
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: Toshi Kani
Link: http://lkml.kernel.org/r/55ACC3660200007800092E62@mail.emea.novell.com
Signed-off-by: Ingo Molnar -
Complete the set of dependent features that need disabling at
once: XSAVEC, AVX-512 and all currently known to the kernel
extensions to it, as well as MPX need to be disabled too.Signed-off-by: Jan Beulich
Cc: Dave Hansen
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Link: http://lkml.kernel.org/r/55ACC40D0200007800092E6C@mail.emea.novell.com
Signed-off-by: Ingo Molnar -
MPX setups private anonymous mapping, but uses vma->vm_ops too.
This can confuse core VM, as it relies on vm->vm_ops to
distinguish file VMAs from anonymous.As result we will get SIGBUS, because handle_pte_fault() thinks
it's file VMA without vm_ops->fault and it doesn't know how to
handle the situation properly.Let's fix that by not setting ->vm_ops.
We don't really need ->vm_ops here: MPX VMA can be detected with
VM_MPX flag. And vma_merge() will not merge MPX VMA with non-MPX
VMA, because ->vm_flags won't match.The only thing left is name of VMA. I'm not sure if it's part of
ABI, or we can just drop it. The patch keep it by providing
arch_vma_name() on x86.Signed-off-by: Kirill A. Shutemov
Signed-off-by: Dave Hansen
Cc: # Fixes: 6b7339f4 (mm: avoid setting up anonymous pages into file mapping)
Cc: Andy Lutomirski
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: dave@sr71.net
Link: http://lkml.kernel.org/r/20150720212958.305CC3E9@viggo.jf.intel.com
Signed-off-by: Ingo Molnar -
flush_tlb_info->flush_start/end are both normal virtual
addresses. When calculating 'nr_pages' (only used for the
tracepoint), I neglected to put parenthesis in.Thanks to David Koufaty for pointing this out.
Signed-off-by: Dave Hansen
Cc: Andrew Morton
Cc: Linus Torvalds
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: dave@sr71.net
Cc:
Link: http://lkml.kernel.org/r/20150720230153.9E834081@viggo.jf.intel.com
Signed-off-by: Ingo Molnar -
Pull s390 fix from Martin Schwidefsky:
"Fast path fix for the thread_struct breakage"* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
s390: adapt entry.S to the move of thread_struct -
Pull AVR32 update from Hans-Christian Egtvedt.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/egtvedt/linux-avr32:
AVR32/time: Migrate to new 'set-state' interface
20 Jul, 2015
7 commits
-
git commit 0c8c0f03e3a292e031596484275c14cf39c0ab7a
"x86/fpu, sched: Dynamically allocate 'struct fpu'"
moved the thread_struct to the end of the task_struct.This causes some of the offsets used in entry.S to overflow their
instruction operand field. To fix this use aghi to create a
dedicated pointer for the thread_struct.Signed-off-by: Martin Schwidefsky
-
Migrate avr32 driver to the new 'set-state' interface provided by
clockevents core, the earlier 'set-mode' interface is marked obsolete
now.This also enables us to implement callbacks for new states of clockevent
devices, for example: ONESHOT_STOPPED.We want to call cpu_idle_poll_ctrl() in shutdown only if we were in
oneshot or resume state earlier. Create another variable to save this
information and check that in shutdown callback.Cc: Haavard Skinnemoen
Cc: Hans-Christian Egtvedt
Signed-off-by: Viresh Kumar
Acked-by: Hans-Christian Egtvedt -
Pull SCSI fixes from James Bottomley:
"Two fairly simple fixes: one is a change that causes us to have a very
low queue depth leading to performance issues and the other is a null
deref occasionally in tapes thanks to use after put"* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: fix host max depth checking for the 'queue_depth' sysfs interface
st: null pointer dereference panic caused by use after kref_put by st_open -
Pull MIPS fixes from Ralf Baechle:
"Another round of MIPS fixes for 4.2.Things are looking quite decent at this stage but the recent work on
the FPU support took its toll:- fix an incorrect overly restrictive ifdef
- select O32 64-bit FP support for O32 binary compatibility
- remove workarounds for Sibyte SB1250 Pass1 parts. There are rare
fixing the workarounds is not worth the effort.- patch up an outdated and now incorrect comment"
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
MIPS: fpu.h: Allow 64-bit FPU on a 64-bit MIPS R6 CPU
MIPS: SB1: Remove support for Pass 1 parts.
MIPS: Require O32 FP64 support for MIPS64 with O32 compat
MIPS: asm-offset.c: Patch up various comments refering to the old filename. -
Pull parisc fix from Helge Deller:
"A memory leak fix from Christophe Jaillet which was introduced with
kernel 4.0 and which leads to kernel crashes on parisc after 1-3 days"* 'parisc-4.2-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
parisc: mm: Fix a memory leak related to pmd not attached to the pgd -
Pull ARM SoC fixes from Olof Johansson:
"By far most of the fixes here are updates to DTS files to deal with
some mostly minor bugs.There's also a fix to deal with non-PM kernel configs on i.MX, a
regression fix for ethernet on PXA platforms and a dependency fix for
OMAP"* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
ARM: keystone: dts: rename pcie nodes to help override status
ARM: keystone: dts: fix dt bindings for PCIe
ARM: pxa: fix dm9000 platform data regression
ARM: dts: Correct audio input route & set mic bias for am335x-pepper
ARM: OMAP2+: Add HAVE_ARM_SCU for AM43XX
MAINTAINERS: digicolor: add dts files
ARM: ux500: fix MMC/SD card regression
ARM: ux500: define serial port aliases
ARM: dts: OMAP5: Add #iommu-cells property to IOMMUs
ARM: dts: OMAP4: Add #iommu-cells property to IOMMUs
ARM: dts: Fix frequency scaling on Gumstix Pepper
ARM: dts: configure regulators for Gumstix Pepper
ARM: dts: omap3: overo: Update LCD panel names
ARM: dts: cros-ec-keyboard: Add support for some Japanese keys
ARM: imx6: gpc: always enable PU domain if CONFIG_PM is not set
ARM: dts: imx53-qsb: fix TVE entry
ARM: dts: mx23: fix iio-hwmon support
ARM: dts: imx27: Adjust the GPT compatible string
ARM: socfpga: dts: Fix entries order
ARM: socfpga: dts: Fix adxl34x formating and compatible string
19 Jul, 2015
10 commits
-
Commit 6134d94923d0 ("MIPS: asm: fpu: Allow 64-bit FPU on MIPS32 R6")
added support for 64-bit FPU on a 32-bit MIPS R6 processor but it missed
the 64-bit CPU case leading to FPU failures when requesting FR=1 mode
(which is always the case for MIPS R6 userland) when running a 32-bit
kernel on a 64-bit CPU. We also fix the MIPS R2 case.Signed-off-by: Markos Chandras
Fixes: 6134d94923d0 ("MIPS: asm: fpu: Allow 64-bit FPU on MIPS32 R6")
Reviewed-by: Paul Burton
Cc: # 4.0+
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/10734/
Signed-off-by: Ralf Baechle -
Commit 0e0da48dee8d ("parisc: mm: don't count preallocated pmds")
introduced a memory leak.After this commit, the 'return' statement in pmd_free is executed in all
cases. Even for pmd that are not attached to the pgd. So 'free_pages'
can never be called anymore, leading to a memory leak.Signed-off-by: Christophe JAILLET
Acked-by: Kirill A. Shutemov
Acked-by: Mikulas Patocka
Acked-by: Helge Deller
Cc: stable@vger.kernel.org # v4.0+
Signed-off-by: Helge Deller -
Merge "pxa fixes for v4.2" from Robert Jarzmik:
ARM: pxa: fixes for v4.2-rc2
This single fix reenables ethernet cards for several pxa boards,
broken by regulator addition to dm9000 driver.* tag 'pxa-fixes-v4.2-rc2' of https://github.com/rjarzmik/linux:
ARM: pxa: fix dm9000 platform data regression -
Pull ARM fixes from Russell King:
"A small set of ARM fixes for -rc3, most of them not far off
one-liners, with the exception of fixing the V7 cache invalidation for
incoming SMP processors which was causing problems for SoCFPGA
devices"* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
ARM: fix __virt_to_idmap build error on !MMU
ARM: invalidate L1 before enabling coherency
ARM: 8404/1: dma-mapping: fix off-by-one error in bitmap size check
ARM: 8402/1: perf: Don't use of_node after putting it
ARM: 8400/1: use virt_to_idmap to get phys_reset address -
Pull x86 fixes from Ingo Molnar:
"Two families of fixes:- Fix an FPU context related boot crash on newer x86 hardware with
larger context sizes than what most people test. To fix this
without ugly kludges or extensive reverts we had to touch core task
allocator, to allow x86 to determine the task size dynamically, at
boot time.I've tested it on a number of x86 platforms, and I cross-built it
to a handful of architectures:(warns) (warns)
testing x86-64: -git: pass ( 0), -tip: pass ( 0)
testing x86-32: -git: pass ( 0), -tip: pass ( 0)
testing arm: -git: pass ( 1359), -tip: pass ( 1359)
testing cris: -git: pass ( 1031), -tip: pass ( 1031)
testing m32r: -git: pass ( 1135), -tip: pass ( 1135)
testing m68k: -git: pass ( 1471), -tip: pass ( 1471)
testing mips: -git: pass ( 1162), -tip: pass ( 1162)
testing mn10300: -git: pass ( 1058), -tip: pass ( 1058)
testing parisc: -git: pass ( 1846), -tip: pass ( 1846)
testing sparc: -git: pass ( 1185), -tip: pass ( 1185)... so I hope the cross-arch impact 'none', as intended.
(by Dave Hansen)
- Fix various NMI handling related bugs unearthed by the big asm code
rewrite and generally make the NMI code more robust and more
maintainable while at it. These changes are a bit late in the
cycle, I hope they are still acceptable.(by Andy Lutomirski)"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/fpu, sched: Introduce CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT and use it on x86
x86/fpu, sched: Dynamically allocate 'struct fpu'
x86/entry/64, x86/nmi/64: Add CONFIG_DEBUG_ENTRY NMI testing code
x86/nmi/64: Make the "NMI executing" variable more consistent
x86/nmi/64: Minor asm simplification
x86/nmi/64: Use DF to avoid userspace RSP confusing nested NMI detection
x86/nmi/64: Reorder nested NMI checks
x86/nmi/64: Improve nested NMI comments
x86/nmi/64: Switch stacks on userspace NMI entry
x86/nmi/64: Remove asm code that saves CR2
x86/nmi: Enable nested do_nmi() handling for 64-bit kernels -
Pull timer fix from Ingo Molnar:
"Fix for a misplaced export that can cause build failures in certain
(rare) Kconfig situations"* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
tick: Move the export of tick_broadcast_oneshot_control to the proper place -
Pull scheduler fix from Ingo Molnar:
"A oneliner rq throttling fix"* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/fair: Test list head instead of list entry in throttle_cfs_rq() -
Pull perf fixes from Ingo Molnar:
"Mostly tooling fixes, plus a static key fix fixing /sys/devices/cpu/rdpmc"* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf tools: Really allow to specify custom CC, AR or LD
perf auxtrace: Fix misplaced check for HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT
perf hists browser: Take the --comm, --dsos, etc filters into account
perf symbols: Store if there is a filter in place
x86, perf: Fix static_key bug in load_mm_cr4()
tools: Copy lib/hweight.c from the kernel sources
perf tools: Fix the detached tarball wrt rbtree copy
perf thread_map: Fix the sizeof() calculation for map entries
tools lib: Improve clean target
perf stat: Fix shadow declaration of close
perf tools: Fix lockup using 32-bit compat vdso -
Pull irq fixes from Ingo Molnar:
"Misc irq fixes:- two driver fixes
- a Xen regression fix
- a nested irq thread crash fix"* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/gicv3-its: Fix mapping of LPIs to collections
genirq: Prevent resend to interrupts marked IRQ_NESTED_THREAD
genirq: Revert sparse irq locking around __cpu_up() and move it to x86 for now
gpio/davinci: Fix race in installing chained irq handler -
Merge fixes from Andrew Morton:
"25 fixes"* emailed patches from Andrew Morton : (25 commits)
lib/decompress: set the compressor name to NULL on error
mm/cma_debug: correct size input to bitmap function
mm/cma_debug: fix debugging alloc/free interface
mm/page_owner: set correct gfp_mask on page_owner
mm/page_owner: fix possible access violation
fsnotify: fix oops in fsnotify_clear_marks_by_group_flags()
/proc/$PID/cmdline: fixup empty ARGV case
dma-debug: skip debug_dma_assert_idle() when disabled
hexdump: fix for non-aligned buffers
checkpatch: fix long line messages about patch context
mm: clean up per architecture MM hook header files
MAINTAINERS: uclinux-h8-devel is moderated for non-subscribers
mailmap: update Sudeep Holla's email id
Update Viresh Kumar's email address
mm, meminit: suppress unused memory variable warning
configfs: fix kernel infoleak through user-controlled format string
include, lib: add __printf attributes to several function prototypes
s390/hugetlb: add hugepages_supported define
mm: hugetlb: allow hugepages_supported to be architecture specific
revert "s390/mm: make hugepages_supported a boot time decision"
...
18 Jul, 2015
11 commits
-
Pull btrfs fixes from Chris Mason:
"These are all from Filipe, and cover a few problems we've had reported
on the list recently (along with ones he found on his own)"* 'for-linus-4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
Btrfs: fix file corruption after cloning inline extents
Btrfs: fix order by which delayed references are run
Btrfs: fix list transaction->pending_ordered corruption
Btrfs: fix memory leak in the extent_same ioctl
Btrfs: fix shrinking truncate when the no_holes feature is enabled -
Pull rtc fixes from Alexandre Belloni:
"A few fixes for the RTC susbsystem for 4.2.The mt6397 driver was introduce in 4.2 so it is worth fixing before
the final release. I though the compilation warning for armada38x was
fixed by akpm in commit f98b733e93e0 ("rtc-armada38x.c: remove unused
local `flags'") but he actually missed some occurrences of the
variables. Since I received 4 patches for that, I think we can
include it now.Summary:
- fix mt6397 wakealarm creation
- remove a compilation warning for armada38x that was forgotten"* tag 'rtc-v4.2-2' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
rtc: armada38x: Remove unused variable from armada38x_rtc_set_time()
rtc: mt6397: enable wakeup before registering rtc device -
Pull device mapper fixes from Mike Snitzer:
- revert a request-based DM core change that caused IO latency to
increase and adversely impact both throughput and system load- fix for a use after free bug in DM core's device cleanup
- a couple DM btree removal fixes (used by dm-thinp)
- a DM thinp fix for order-5 allocation failure
- a DM thinp fix to not degrade to read-only metadata mode when in
out-of-data-space mode for longer than the 'no_space_timeout'- fix a long-standing oversight in both dm-thinp and dm-cache by now
exporting 'needs_check' in status if it was set in metadata- fix an embarrassing dm-cache busy-loop that caused worker threads to
eat cpu even if no IO was actively being issued to the cache device* tag 'dm-4.2-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
dm cache: avoid calls to prealloc_free_structs() if possible
dm cache: avoid preallocation if no work in writeback_some_dirty_blocks()
dm cache: do not wake_worker() in free_migration()
dm cache: display 'needs_check' in status if it is set
dm thin: display 'needs_check' in status if it is set
dm thin: stay in out-of-data-space mode once no_space_timeout expires
dm: fix use after free crash due to incorrect cleanup sequence
Revert "dm: only run the queue on completion if congested or no requests pending"
dm btree: silence lockdep lock inversion in dm_btree_del()
dm thin: allocate the cell_sort_array dynamically
dm btree remove: fix bug in redistribute3 -
Don't burden architectures without dynamic task_struct sizing
with the overhead of dynamic sizing.Also optimize the x86 code a bit by caching task_struct_size.
Acked-and-Tested-by: Dave Hansen
Cc: Andy Lutomirski
Cc: Borislav Petkov
Cc: Brian Gerst
Cc: Dave Hansen
Cc: Denys Vlasenko
Cc: Linus Torvalds
Cc: Oleg Nesterov
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Link: http://lkml.kernel.org/r/1437128892-9831-3-git-send-email-mingo@kernel.org
Signed-off-by: Ingo Molnar -
The FPU rewrite removed the dynamic allocations of 'struct fpu'.
But, this potentially wastes massive amounts of memory (2k per
task on systems that do not have AVX-512 for instance).Instead of having a separate slab, this patch just appends the
space that we need to the 'task_struct' which we dynamically
allocate already. This saves from doing an extra slab
allocation at fork().The only real downside here is that we have to stick everything
and the end of the task_struct. But, I think the
BUILD_BUG_ON()s I stuck in there should keep that from being too
fragile.Signed-off-by: Dave Hansen
Cc: Andy Lutomirski
Cc: Borislav Petkov
Cc: Brian Gerst
Cc: Dave Hansen
Cc: Denys Vlasenko
Cc: H. Peter Anvin
Cc: Linus Torvalds
Cc: Oleg Nesterov
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Link: http://lkml.kernel.org/r/1437128892-9831-2-git-send-email-mingo@kernel.org
Signed-off-by: Ingo Molnar -
Without this we end up using the previous name of the compressor in the
loop in unpack_rootfs. For example we get errors like "compression
method gzip not configured" even when we have CONFIG_DECOMPRESS_GZIP
enabled.Signed-off-by: Aneesh Kumar K.V
Cc: "H. Peter Anvin"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
In CMA, 1 bit in bitmap means 1 << order_per_bits pages so size of
bitmap is cma->count >> order_per_bits rather than just cma->count.
This patch fixes it.Signed-off-by: Joonsoo Kim
Acked-by: Michal Nazarewicz
Cc: Sasha Levin
Cc: Stefan Strogin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
CMA has alloc/free interface for debugging. It is intended that
alloc/free occurs in specific CMA region, but, currently, alloc/free
interface is on root dir due to the bug so we can't select CMA region
where alloc/free happens.This patch fixes this problem by making alloc/free interface per CMA
region.Signed-off-by: Joonsoo Kim
Acked-by: Michal Nazarewicz
Cc: Sasha Levin
Cc: Stefan Strogin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Currently, we set wrong gfp_mask to page_owner info in case of isolated
freepage by compaction and split page. It causes incorrect mixed
pageblock report that we can get from '/proc/pagetypeinfo'. This metric
is really useful to measure fragmentation effect so should be accurate.
This patch fixes it by setting correct information.Without this patch, after kernel build workload is finished, number of
mixed pageblock is 112 among roughly 210 movable pageblocks.But, with this fix, output shows that mixed pageblock is just 57.
Signed-off-by: Joonsoo Kim
Cc: Mel Gorman
Cc: Vlastimil Babka
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When I tested my new patches, I found that page pointer which is used
for setting page_owner information is changed. This is because page
pointer is used to set new migratetype in loop. After this work, page
pointer could be out of bound. If this wrong pointer is used for
page_owner, access violation happens. Below is error message that I
got.BUG: unable to handle kernel paging request at 0000000000b00018
IP: [] save_stack_address+0x30/0x40
PGD 1af2d067 PUD 166e0067 PMD 0
Oops: 0002 [#1] SMP
...snip...
Call Trace:
print_context_stack+0xcf/0x100
dump_trace+0x15f/0x320
save_stack_trace+0x2f/0x50
__set_page_owner+0x46/0x70
__isolate_free_page+0x1f7/0x210
split_free_page+0x21/0xb0
isolate_freepages_block+0x1e2/0x410
compaction_alloc+0x22d/0x2d0
migrate_pages+0x289/0x8b0
compact_zone+0x409/0x880
compact_zone_order+0x6d/0x90
try_to_compact_pages+0x110/0x210
__alloc_pages_direct_compact+0x3d/0xe6
__alloc_pages_nodemask+0x6cd/0x9a0
alloc_pages_current+0x91/0x100
runtest_store+0x296/0xa50
simple_attr_write+0xbd/0xe0
__vfs_write+0x28/0xf0
vfs_write+0xa9/0x1b0
SyS_write+0x46/0xb0
system_call_fastpath+0x16/0x75This patch fixes this error by moving up set_page_owner().
Signed-off-by: Joonsoo Kim
Cc: Mel Gorman
Cc: Vlastimil Babka
Acked-by: Minchan Kim
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
fsnotify_clear_marks_by_group_flags() can race with
fsnotify_destroy_marks() so when fsnotify_destroy_mark_locked() drops
mark_mutex, a mark from the list iterated by
fsnotify_clear_marks_by_group_flags() can be freed and we dereference free
memory in the loop there.Fix the problem by keeping mark_mutex held in
fsnotify_destroy_mark_locked(). The reason why we drop that mutex is that
we need to call a ->freeing_mark() callback which may acquire mark_mutex
again. To avoid this and similar lock inversion issues, we move the call
to ->freeing_mark() callback to the kthread destroying the mark.Signed-off-by: Jan Kara
Reported-by: Ashish Sangwan
Suggested-by: Lino Sanfilippo
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
