The i_mutex lock use of generic _file_llseek hurts. Independent processes
accessing the same file synchronize over a single lock, even though
they have no need for synchronization at all.
Under high utilization this can cause llseek to scale very poorly on larger
systems.
This patch does some rethinking of the llseek locking model:
First the 64bit f_pos is not necessarily atomic without locks
on 32bit systems. This can already cause races with read() today.
This was discussed on linux-kernel in the past and deemed acceptable.
The patch does not change that.
Let's look at the different seek variants:
SEEK_SET: Doesn't really need any locking.
If there's a race one writer wins, the other loses.
For 32bit the non atomic update races against read()
stay the same. Without a lock they can also happen
against write() now. The read() race was deemed
acceptable in past discussions, and I think if it's
ok for read it's ok for write too.
=> Don't need a lock.
SEEK_END: This behaves like SEEK_SET plus it reads
the maximum size too. Reading the maximum size would have the
32bit atomic problem. But luckily we already have a way to read
the maximum size without locking (i_size_read), so we
can just use that instead.
Without i_mutex there is no synchronization with write() anymore,
however since the write() update is atomic on 64bit it just behaves
like another racy SEEK_SET. On non atomic 32bit it's the same
as SEEK_SET.
=> Don't need a lock, but need to use i_size_read()
SEEK_CUR: This has a read-modify-write race window
on the same file. One could argue that any application
doing unsynchronized seeks on the same file is already broken.
But for the sake of not adding a regression here I'm
using the file->f_lock to synchronize this. Using this
lock is much better than the inode mutex because it doesn't
synchronize between processes.
=> So still need a lock, but can use a f_lock.
This patch implements this new scheme in generic_file_llseek.
I dropped generic_file_llseek_unlocked and changed all callers.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This doesn't change anything for the compiler, but hch thought it would
make the code clearer.
I moved the reference counting into its own little inline.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Add inlines to all the submission path functions. While this increases
code size it also gives gcc a lot of optimization opportunities
in this critical hotpath.
In particular -- together with some other changes -- this
allows gcc to get rid of the unnecessary clearing of
sdio at the beginning and optimize the messy parameter passing.
Any non inlining of a function which takes a sdio parameter
would break this optimization because they cannot be done if the
address of a structure is taken.
Note that benefits are only seen with CONFIG_OPTIMIZE_INLINING
and CONFIG_CC_OPTIMIZE_FOR_SIZE both set to off.
This gives about 2.2% improvement on a large database benchmark
with a high IOPS rate.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Only a single b_private field in the map_bh buffer head is needed after
the submission path. Move map_bh separately to avoid storing
this information in the long term slab.
This avoids the weird 104 byte hole in struct dio_submit which also needed
to be memseted early.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
A direct slab call is slightly faster than kmalloc and can be better cached
per CPU. It also avoids rounding to the next kmalloc slab.
In addition this enforces cache line alignment for struct dio to avoid
any false sharing.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Fix most problems reported by pahole.
There is still a weird 104 byte hole after map_bh. I'm not sure what
causes this.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
There's nothing on the stack, even before my changes.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This large, but largely mechanic, patch moves all fields in struct dio
that are only used in the submission path into a separate on stack
data structure. This has the advantage that the memory is very likely
cache hot, which is not guaranteed for memory fresh out of kmalloc.
This also gives gcc more optimization potential because it can easier
determine that there are no external aliases for these variables.
The sdio initialization is a initialization now instead of memset.
This allows gcc to break sdio into individual fields and optimize
away unnecessary zeroing (after all the functions are inlined)
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We need to move the inode to the end of the list to actually make the
spinning prevention explained in the comment above it work. With a
plain list_move it will simply stay in place as we're always reclaiming
from the head of the list.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruen@kernel.org>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruen@kernel.org>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We are going to add more flags and having them in hex format
make it simpler
Acked-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruen@kernel.org>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This was found by inspection while tracking a similar
bug in compat_statfs64, that has been fixed in mainline
since decemeber.
- This fixes a bug where not all of the f_spare fields
were cleared on mips and s390.
- Add the f_flags field to struct compat_statfs
- Copy f_flags to userspace in case someone cares.
- Use __clear_user to copy the f_spare field to userspace
to ensure that all of the elements of f_spare are cleared.
On some architectures f_spare is has 5 ints and on some
architectures f_spare only has 4 ints. Which makes
the previous technique of clearing each int individually
broken.
I don't expect anyone actually uses the old statfs system
call anymore but if they do let them benefit from having
the compat and the native version working the same.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
* 'drm-core-next' of git://people.freedesktop.org/~airlied/linux: (290 commits)
Revert "drm/ttm: add a way to bo_wait for either the last read or last write"
Revert "drm/radeon/kms: add a new gem_wait ioctl with read/write flags"
vmwgfx: Don't pass unused arguments to do_dirty functions
vmwgfx: Emulate depth 32 framebuffers
drm/radeon: Lower the severity of the radeon lockup messages.
drm/i915/dp: Fix eDP on PCH DP on CPT/PPT
drm/i915/dp: Introduce is_cpu_edp()
drm/i915: use correct SPD type value
drm/i915: fix ILK+ infoframe support
drm/i915: add DP test request handling
drm/i915: read full receiver capability field during DP hot plug
drm/i915/dp: Remove eDP special cases from bandwidth checks
drm/i915/dp: Fix the math in intel_dp_link_required
drm/i915/panel: Always record the backlight level again (but cleverly)
i915: Move i915_read/write out of line
drm/i915: remove transcoder PLL mashing from mode_set per specs
drm/i915: if transcoder disable fails, say which
drm/i915: set watermarks for third pipe on IVB
drm/i915: export a CPT mode set verification function
drm/i915: fix transcoder PLL select masking
...
* 'x86-vdso-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86-64, doc: Remove int 0xcc from entry_64.S documentation
x86, vsyscall: Add missing <asm/fixmap.h> to arch/x86/mm/fault.c
Fix up trivial conflicts in arch/x86/mm/fault.c (asm/fixmap.h vs
asm/vsyscall.h: both work, which to use? Whatever..)
* 'x86-spinlocks-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, ticketlock: remove obsolete comment
x86, cmpxchg: Use __compiletime_error() to make usage messages a bit nicer
x86, ticketlock: Make __ticket_spin_trylock common
x86, ticketlock: Convert __ticket_spin_lock to use xadd()
x86, ticketlock: Convert spin loop to C
x86, ticketlock: Clean up types and accessors
x86: Use xadd helper more widely
x86: Add xadd helper macro
x86, cmpxchg: Unify cmpxchg into cmpxchg.h
x86, cmpxchg: Move 64-bit set64_bit() to match 32-bit
x86, cmpxchg: Move 32-bit __cmpxchg_wrong_size to match 64 bit.
x86, cmpxchg: <linux/alternative.h> has LOCK_PREFIX
* 'x86-rdrand-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, random: Verify RDRAND functionality and allow it to be disabled
x86, random: Architectural inlines to get random integers with RDRAND
random: Add support for architectural random hooks
Fix up trivial conflicts in drivers/char/random.c: the architectural
random hooks touched "get_random_int()" that was simplified to use MD5
and not do the keyptr thing any more (see commit 6e5714eaf7: "net:
Compute protocol sequence numbers and fragment IDs using MD5").
* 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, microcode, AMD: Add microcode revision to /proc/cpuinfo
x86, microcode: Correct microcode revision format
coretemp: Get microcode revision from cpu_data
x86, intel: Use c->microcode for Atom errata check
x86, intel: Output microcode revision in /proc/cpuinfo
x86, microcode: Don't request microcode from userspace unnecessarily
Fix up trivial conflicts in arch/x86/kernel/cpu/amd.c (conflict between
moving AMD BSP code to cpu_dev helper function and adding AMD microcode
revision to /proc/cpuinfo code)
* 'x86-hyperv-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86: Hyper-V: Integrate the clocksource with Hyper-V detection code
Fix up conflicts in drivers/staging/hv/Makefile manually (some of the hv
code has moved out of staging to drivers/hv/)
* 'x86-cpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, amd: Include linux/elf.h since we use stuff from asm/elf.h
x86: cache_info: Update calculation of AMD L3 cache indices
x86: cache_info: Kill the atomic allocation in amd_init_l3_cache()
x86: cache_info: Kill the moronic shadow struct
x86: cache_info: Remove bogus free of amd_l3_cache data
x86, amd: Include elf.h explicitly, prepare the code for the module.h split
x86-32, amd: Move va_align definition to unbreak 32-bit build
x86, amd: Move BSP code to cpu_dev helper
x86: Add a BSP cpu_dev helper
x86, amd: Avoid cache aliasing penalties on AMD family 15h
nfsiostat was failing to find mounted filesystems on kernels after
2.6.38 because of changes to show_vfsstat() by commit
c7f404b40a. This patch adds back the
"device" tag before the nfs server entry so scripts can parse the
mountstats file correctly.
Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
CC: stable@kernel.org [>=2.6.39]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Currently, when you call iov_iter_advance, then the pointer to the iovec
array can be incremented, but it does not decrement the nr_segs value in
the iov_iter struct. The result is a iov_iter struct with a nr_segs
value that goes beyond the end of the array.
While I'm not aware of anything that's specifically broken by this, it
seems odd and a bit dangerous not to decrement that value. If someone
were to trust the nr_segs value to be correct, then they could end up
walking off the end of the array.
Changing this might also provide some micro-optimization when dealing
with the last iovec in an array. Many of the other routines that deal
with iov_iter have optimized codepaths when nr_segs == 1.
Cc: Nick Piggin <npiggin@suse.de>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
fs/cifs/transport.c: In function 'wait_for_response':
fs/cifs/transport.c:328: error: implicit declaration of function 'wait_event_freezekillable'
Caused by commit f06ac72e92 ("cifs, freezer: add
wait_event_freezekillable and have cifs use it"). In this config,
CONFIG_FREEZER is not set.
Reviewed-by: Shirish Pargaonkar <shirishp@us.ibm.com>
CC: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Samba supports a setfs info level to negotiate encrypted
shares. This patch adds the defines so we recognize
this info level. Later patches will add the enablement
for it.
Acked-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Steve French <smfrench@gmail.com>
This reverts commit dfadbbdb57.
Further upstream discussion between Marek and Thomas decided this wasn't
fully baked and needed further work, so revert it before it hits mainline.
Signed-off-by: Dave Airlie <airlied@redhat.com>
This reverts commit d3ed74027f.
Further upstream discussion between Thomas and Marek decided this needed
more work and driver specifics. So revert before it goes upstream.
Signed-off-by: Dave Airlie <airlied@redhat.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (68 commits)
Input: adp5589-keys - add support for the ADP5585 derivatives
Input: imx_keypad - add pm suspend and resume support
Input: force feedback - potential integer wrap in input_ff_create()
Input: tsc2007 - make sure that X plate resistance is specified
Input: serio_raw - fix memory leak when closing char device
Input: serio_raw - kick clients when disconnecting port
Input: serio_raw - explicitly mark disconnected ports as dead
Input: serio_raw - fix coding style issues
Input: serio_raw - use dev_*() for messages
Input: serio_raw - use bool for boolean data
Input: serio_raw - perform proper locking when adding clients to list
Input: serio_raw - rename serio_raw_list to serio_raw_client
Input: serio_raw - use kref instead of rolling out its own refcounting
Input: psmouse - switch to using dev_*() for messages
Input: wacom - correct max Y value on medium bamboos
Input: wacom - add ABS_DISTANCE to Bamboo Pen reports
Input: wacom - remove unneeded touch pressure initialization
Input: lm8323 - wrap suspend and resume in CONFIG_PM_SLEEP
Input: ad7879-i2c - wrap suspend and resume in CONFIG_PM_SLEEP
Input: synaptics_i2c - wrap suspend and resume in CONFIG_PM_SLEEP
...
* 'clk' of http://ftp.arm.linux.org.uk/pub/linux/arm/kernel/git-cur/linux-2.6-arm:
ARM: 7131/1: clkdev: Add Common Macro for clk_lookup
clk: spi-pl022: convert to clk_prepare()/clk_unprepare()
clk: timer-sp: convert to clk_prepare()/clk_unprepare()
clk: sa1111: convert to clk_prepare()/clk_unprepare()
clk: mmci: convert to clk_prepare()/clk_unprepare()
clk: amba-pl011: convert to clk_prepare()/clk_unprepare()
clk: amba-pl010: convert to clk_prepare()/clk_unprepare()
clk: amba-clcd: convert to clk_prepare()/clk_unprepare()
clk: amba bus: convert to clk_prepare()/clk_unprepare()
clk: provide prepare/unprepare functions
In my last patch I did a stupid mistake and broke the exofs
compilation completely. Fix it ASAP.
Instead of obj-y I did obj-$(y)
Really Really sorry. Me totally blushing :-{|
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'upstream-linus' of git://github.com/jgarzik/libata-dev: (50 commits)
pata_efar: fix register naming used in efar_set_piomode()
pata_rdc: add Power Management support
ata_piix: unify code for programming PIO and MWDMA timings
ata_piix: SITRE handling fix
libata: make ata_sff_data_xfer_noirq() work with 32-bit PIO
pata_artop: add Power Management support
pata_sl82c105: add Power Management support
pata_pdc2027x: add Power Management support
pata_legacy: use PIO mask defines
pata_legacy: unify QDI ->set_piomode methods
libata: remove no longer needed pata_qdi driver
pata_it8213: fix register naming used in it8213_set_piomode()
pata_it8213: add UDMA100 and UDMA133 support
pata_rdc: parallel scanning needs an extra locking
pata_via: add via_fixup()
libata: ata_timing_compute() fixup
pata_scc: add proper cable detection method
pata_arasan_cf: remove bogus to_platform_device() calls
pata_atiixp: add proper ->prereset method
pata_serverworks: add serverworks_fixup()
...
* 'slab/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux:
tools, slub: Fix off-by-one buffer corruption after readlink() call
slub: Discard slab page when node partial > minimum partial number
slub: correct comments error for per cpu partial
mm: restrict access to slab files under procfs and sysfs
slub: Code optimization in get_partial_node()
slub: doc: update the slabinfo.c file path
slub: explicitly document position of inserting slab to partial list
slub: update slabinfo tools to report per cpu partial list statistics
slub: per cpu cache for partial pages
slub: return object pointer from get_partial() / new_slab().
slub: pass kmem_cache_cpu pointer to get_partial()
slub: Prepare inuse field in new_slab()
slub: Remove useless statements in __slab_alloc
slub: free slabs without holding locks
slub: use print_hex_dump
slab: use print_hex_dump
* 'for-linus' of git://git.open-osd.org/linux-open-osd: (21 commits)
ore: Enable RAID5 mounts
exofs: Support for RAID5 read-4-write interface.
ore: RAID5 Write
ore: RAID5 read
fs/Makefile: Always inspect exofs/
ore: Make ore_calc_stripe_info EXPORT_SYMBOL
ore/exofs: Change ore_check_io API
ore/exofs: Define new ore_verify_layout
ore: Support for partial component table
ore: Support for short read/writes
exofs: Support for short read/writes
ore: Remove check for ios->kern_buff in _prepare_for_striping to later
ore: cleanup: Embed an ore_striping_info inside ore_io_state
ore: Only IO one group at a time (API change)
ore/exofs: Change the type of the devices array (API change)
ore: Make ore_striping_info and ore_calc_stripe_info public
exofs: Remove unused data_map member from exofs_sb_info
exofs: Rename struct ore_components comps => oc
exofs/super.c: local functions should be static
exofs/ore.c: local functions should be static
...
* 'x86-cleanups-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86-64, unistd: Remove bogus __IGNORE_getcpu
x86, mm, trivial: Remove unnecessary get_order() in free_thread_info()
x86, cleanup: Remove unneeded version.h include from arch/x86/
* 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86-64: Fix CFI data for interrupt frames
x86-64: Don't apply destructive erratum workaround on unaffected CPUs
* 'x86-apic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/irq: Standardize on CONFIG_SPARSE_IRQ=y
x86, ioapic: Clean up ioapic/apic_id usage
x86, ioapic: Factor out print_IO_APIC() to only print one io apic
x86, ioapic: Print out irte with right ioapic index
x86, ioapic: Split up setup_ioapic_entry()
x86, ioapic: Pass struct irq_attr * to setup_ioapic_irq()
apic, i386/bigsmp: Fix false warnings regarding logical APIC ID mismatches
Mark Einon points out that the Kconfig option for NETDEV_1000 no longer
exists, and the merge of the staging drivers should have removed that
for the et131x driver.
And while checking for it, I noticed that slicoss had the same stale
dependency. Remove that one too.
Reported-by: Mark Einon <mark.einon@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (23 commits)
time, s390: Get rid of compile warning
dw_apb_timer: constify clocksource name
time: Cleanup old CONFIG_GENERIC_TIME references that snuck in
time: Change jiffies_to_clock_t() argument type to unsigned long
alarmtimers: Fix error handling
clocksource: Make watchdog reset lockless
posix-cpu-timers: Cure SMP accounting oddities
s390: Use direct ktime path for s390 clockevent device
clockevents: Add direct ktime programming function
clockevents: Make minimum delay adjustments configurable
nohz: Remove "Switched to NOHz mode" debugging messages
proc: Consider NO_HZ when printing idle and iowait times
nohz: Make idle/iowait counter update conditional
nohz: Fix update_ts_time_stat idle accounting
cputime: Clean up cputime_to_usecs and usecs_to_cputime macros
alarmtimers: Rework RTC device selection using class interface
alarmtimers: Add try_to_cancel functionality
alarmtimers: Add more refined alarm state tracking
alarmtimers: Remove period from alarm structure
alarmtimers: Remove interval cap limit hack
...
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (46 commits)
llist: Add back llist_add_batch() and llist_del_first() prototypes
sched: Don't use tasklist_lock for debug prints
sched: Warn on rt throttling
sched: Unify the ->cpus_allowed mask copy
sched: Wrap scheduler p->cpus_allowed access
sched: Request for idle balance during nohz idle load balance
sched: Use resched IPI to kick off the nohz idle balance
sched: Fix idle_cpu()
llist: Remove cpu_relax() usage in cmpxchg loops
sched: Convert to struct llist
llist: Add llist_next()
irq_work: Use llist in the struct irq_work logic
llist: Return whether list is empty before adding in llist_add()
llist: Move cpu_relax() to after the cmpxchg()
llist: Remove the platform-dependent NMI checks
llist: Make some llist functions inline
sched, tracing: Show PREEMPT_ACTIVE state in trace_sched_switch
sched: Remove redundant test in check_preempt_tick()
sched: Add documentation for bandwidth control
sched: Return unused runtime on group dequeue
...
