Currently, atomic_cmpxchg() is used to get the lock. However, this
is not really necessary if there is more than one task in the queue
and the queue head don't need to reset the tail code. For that case,
a simple write to set the lock bit is enough as the queue head will
be the only one eligible to get the lock as long as it checks that
both the lock and pending bits are not set. The current pending bit
waiting code will ensure that the bit will not be set as soon as the
tail code in the lock is set.
With that change, the are some slight improvement in the performance
of the queued spinlock in the 5M loop micro-benchmark run on a 4-socket
Westere-EX machine as shown in the tables below.
[Standalone/Embedded - same node]
# of tasks Before patch After patch %Change
---------- ----------- ---------- -------
3 2324/2321 2248/2265 -3%/-2%
4 2890/2896 2819/2831 -2%/-2%
5 3611/3595 3522/3512 -2%/-2%
6 4281/4276 4173/4160 -3%/-3%
7 5018/5001 4875/4861 -3%/-3%
8 5759/5750 5563/5568 -3%/-3%
[Standalone/Embedded - different nodes]
# of tasks Before patch After patch %Change
---------- ----------- ---------- -------
3 12242/12237 12087/12093 -1%/-1%
4 10688/10696 10507/10521 -2%/-2%
It was also found that this change produced a much bigger performance
improvement in the newer IvyBridge-EX chip and was essentially to close
the performance gap between the ticket spinlock and queued spinlock.
The disk workload of the AIM7 benchmark was run on a 4-socket
Westmere-EX machine with both ext4 and xfs RAM disks at 3000 users
on a 3.14 based kernel. The results of the test runs were:
AIM7 XFS Disk Test
kernel JPM Real Time Sys Time Usr Time
----- --- --------- -------- --------
ticketlock 5678233 3.17 96.61 5.81
qspinlock 5750799 3.13 94.83 5.97
AIM7 EXT4 Disk Test
kernel JPM Real Time Sys Time Usr Time
----- --- --------- -------- --------
ticketlock 1114551 16.15 509.72 7.11
qspinlock 2184466 8.24 232.99 6.01
The ext4 filesystem run had a much higher spinlock contention than
the xfs filesystem run.
The "ebizzy -m" test was also run with the following results:
kernel records/s Real Time Sys Time Usr Time
----- --------- --------- -------- --------
ticketlock 2075 10.00 216.35 3.49
qspinlock 3023 10.00 198.20 4.80
Signed-off-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Daniel J Blueman <daniel@numascale.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Douglas Hatch <doug.hatch@hp.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paolo Bonzini <paolo.bonzini@gmail.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: virtualization@lists.linux-foundation.org
Cc: xen-devel@lists.xenproject.org
Link: http://lkml.kernel.org/r/1429901803-29771-7-git-send-email-Waiman.Long@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch introduces a new generic queued spinlock implementation that
can serve as an alternative to the default ticket spinlock. Compared
with the ticket spinlock, this queued spinlock should be almost as fair
as the ticket spinlock. It has about the same speed in single-thread
and it can be much faster in high contention situations especially when
the spinlock is embedded within the data structure to be protected.
Only in light to moderate contention where the average queue depth
is around 1-3 will this queued spinlock be potentially a bit slower
due to the higher slowpath overhead.
This queued spinlock is especially suit to NUMA machines with a large
number of cores as the chance of spinlock contention is much higher
in those machines. The cost of contention is also higher because of
slower inter-node memory traffic.
Due to the fact that spinlocks are acquired with preemption disabled,
the process will not be migrated to another CPU while it is trying
to get a spinlock. Ignoring interrupt handling, a CPU can only be
contending in one spinlock at any one time. Counting soft IRQ, hard
IRQ and NMI, a CPU can only have a maximum of 4 concurrent lock waiting
activities. By allocating a set of per-cpu queue nodes and used them
to form a waiting queue, we can encode the queue node address into a
much smaller 24-bit size (including CPU number and queue node index)
leaving one byte for the lock.
Please note that the queue node is only needed when waiting for the
lock. Once the lock is acquired, the queue node can be released to
be used later.
Signed-off-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Daniel J Blueman <daniel@numascale.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Douglas Hatch <doug.hatch@hp.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paolo Bonzini <paolo.bonzini@gmail.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: virtualization@lists.linux-foundation.org
Cc: xen-devel@lists.xenproject.org
Link: http://lkml.kernel.org/r/1429901803-29771-2-git-send-email-Waiman.Long@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In up_write()/up_read(), rwsem_wake() will be called whenever it
detects that some writers/readers are waiting. The rwsem_wake()
function will take the wait_lock and call __rwsem_do_wake() to do the
real wakeup. For a heavily contended rwsem, doing a spin_lock() on
wait_lock will cause further contention on the heavily contended rwsem
cacheline resulting in delay in the completion of the up_read/up_write
operations.
This patch makes the wait_lock taking and the call to __rwsem_do_wake()
optional if at least one spinning writer is present. The spinning
writer will be able to take the rwsem and call rwsem_wake() later
when it calls up_write(). With the presence of a spinning writer,
rwsem_wake() will now try to acquire the lock using trylock. If that
fails, it will just quit.
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Acked-by: Jason Low <jason.low2@hp.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Douglas Hatch <doug.hatch@hp.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1430428337-16802-2-git-send-email-Waiman.Long@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull power management and ACPI fixes from Rafael Wysocki:
"These include three regression fixes (PCI resources management,
ACPI/PNP device enumeration, ACPI SBS on MacBook) and two ACPI
documentation fixes related to GPIO.
Specifics:
- Fix for a PCI resources management regression introduced during the
4.0 cycle and related to the handling of ACPI resources'
Producer/Consumer flags that turn out to be useless (Jiang Liu)
- Fix for a MacBook regression related to the Smart Battery Subsystem
(SBS) driver causing various problems (stalls on boot, failure to
detect or report battery) to happen and introduced during the 3.18
cycle (Chris Bainbridge)
- Fix for an ACPI/PNP device enumeration regression introduced during
the 3.16 cycle caused by failing to include two PNP device IDs into
the list of IDs that PNP device objects need to be created for
(Witold Szczeponik)
- Fixes for two minor mistakes in the ACPI GPIO properties
documentation (Antonio Ospite, Rafael J Wysocki)"
* tag 'pm+acpi-4.1-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI / PNP: add two IDs to list for PNPACPI device enumeration
ACPI / documentation: Fix ambiguity in the GPIO properties document
ACPI / documentation: fix a sentence about GPIO resources
ACPI / SBS: Add 5 us delay to fix SBS hangs on MacBook
x86/PCI/ACPI: Make all resources except [io 0xcf8-0xcff] available on PCI bus
* acpi-resources:
x86/PCI/ACPI: Make all resources except [io 0xcf8-0xcff] available on PCI bus
* acpi-battery:
ACPI / SBS: Add 5 us delay to fix SBS hangs on MacBook
* acpi-doc:
ACPI / documentation: Fix ambiguity in the GPIO properties document
ACPI / documentation: fix a sentence about GPIO resources
* acpi-pnp:
ACPI / PNP: add two IDs to list for PNPACPI device enumeration
Pull f2fs fixes from Jaegeuk Kim:
"Fix a performance regression and a bug"
* tag 'for-f2fs-4.1-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs:
f2fs: fix wrong error hanlder in f2fs_follow_link
Revert "f2fs: enhance multi-threads performance"
Pull pin control fixes from Linus Walleij:
"Here is a smallish set of pin control fixes for the v4.1 cycle,
collected the last two weeks:
- fix a real nasty legacy bug that has screwed up the protection of
adding pinctrl maps dynamically. Normally this didn't happen so
much but Dough Anderson ran into it and fixed it, kudos!
- minor driver fixes for Qualcomm spmi, mediatek and Marvell drivers"
* tag 'pinctrl-v4.1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
pinctrl: Don't just pretend to protect pinctrl_maps, do it for real
pinctrl: mediatek: mtk-common: initialize unmask
pinctrl: qcom-spmi-mpp: Fix input value report
pinctrl: qcom-spmi: Fix pin direction configuration
pinctrl: mvebu: Fix mapping of pin 63 (gpo -> gpio)
Pull infiniband updates from Doug Ledford:
"Minor updates for 4.1-rc
Most of the changes are fairly small and well confined. The iWARP
address reporting changes are the only ones that are a medium size. I
had these queued up prior to rc1, but due to the shuffle in
maintainers, they did not get submitted when I expected. My apologies
for that. I feel comfortable with them however due to the testing
they've received, so I left them in this submission"
* tag 'for-linus' of git://github.com/dledford/linux:
MAINTAINERS: Update InfiniBand subsystem maintainer
MAINTAINERS: add include/rdma/ to InfiniBand subsystem
IPoIB/CM: Fix indentation level
iw_cxgb4: Remove negative advice dmesg warnings
IB/core: Fix unaligned accesses
IB/core: change rdma_gid2ip into void function as it always return zero
IB/qib: use arch_phys_wc_add()
IB/qib: add acounting for MTRR
IB/core: dma unmap optimizations
IB/core: dma map/unmap locking optimizations
RDMA/cxgb4: Report the actual address of the remote connecting peer
RDMA/nes: Report the actual address of the remote connecting peer
RDMA/core: Enable the iWarp Port Mapper to provide the actual address of the connecting peer to its clients
iw_cxgb4: enforce qp/cq id requirements
iw_cxgb4: use BAR2 GTS register for T5 kernel mode CQs
iw_cxgb4: 32b platform fixes
iw_cxgb4: Cleanup register defines/MACROS
RDMA/CMA: Canonize IPv4 on IPV6 sockets properly
Pull xen bug fixes from David Vrabel:
- fix blkback regression if using persistent grants
- fix various event channel related suspend/resume bugs
- fix AMD x86 regression with X86_BUG_SYSRET_SS_ATTRS
- SWIOTLB on ARM now uses frames <4 GiB (if available) so device only
capable of 32-bit DMA work.
* tag 'for-linus-4.1b-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: Add __GFP_DMA flag when xen_swiotlb_init gets free pages on ARM
hypervisor/x86/xen: Unset X86_BUG_SYSRET_SS_ATTRS on Xen PV guests
xen/events: Set irq_info->evtchn before binding the channel to CPU in __startup_pirq()
xen/console: Update console event channel on resume
xen/xenbus: Update xenbus event channel on resume
xen/events: Clear cpu_evtchn_mask before resuming
xen-pciback: Add name prefix to global 'permissive' variable
xen: Suspend ticks on all CPUs during suspend
xen/grant: introduce func gnttab_unmap_refs_sync()
xen/blkback: safely unmap purge persistent grants
Pull x86 fixes from Ingo Molnar:
"EFI fixes, and FPU fix, a ticket spinlock boundary condition fix and
two build fixes"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/fpu: Always restore_xinit_state() when use_eager_cpu()
x86: Make cpu_tss available to external modules
efi: Fix error handling in add_sysfs_runtime_map_entry()
x86/spinlocks: Fix regression in spinlock contention detection
x86/mm: Clean up types in xlate_dev_mem_ptr()
x86/efi: Store upper bits of command line buffer address in ext_cmd_line_ptr
efivarfs: Ensure VariableName is NUL-terminated
Pull perf fixes from Ingo Molnar:
"Mostly tooling fixes, but also an uncore PMU driver fix and an uncore
PMU driver hardware-enablement addition"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf probe: Fix segfault if passed with ''.
perf report: Fix -T/--threads option to work again
perf bench numa: Fix immediate meeting of convergence condition
perf bench numa: Fixes of --quiet argument
perf bench futex: Fix hung wakeup tasks after requeueing
perf probe: Fix bug with global variables handling
perf top: Fix a segfault when kernel map is restricted.
tools lib traceevent: Fix build failure on 32-bit arch
perf kmem: Fix compiles on RHEL6/OL6
tools lib api: Undefine _FORTIFY_SOURCE before setting it
perf kmem: Consistently use PRIu64 for printing u64 values
perf trace: Disable events and drain events when forked workload ends
perf trace: Enable events when doing system wide tracing and starting a workload
perf/x86/intel/uncore: Move PCI IDs for IMC to uncore driver
perf/x86/intel/uncore: Add support for Intel Haswell ULT (lower power Mobile Processor) IMC uncore PMUs
perf/x86/intel: Add cpu_(prepare|starting|dying) for core_pmu
Pull RCU fix from Ingo Molnar:
"An RCU Kconfig fix that eliminates an annoying interactive kconfig
question for CONFIG_RCU_TORTURE_TEST_SLOW_INIT"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
rcu: Control grace-period delays directly from value
Way back, when the world was a simpler place and there was no war, no
evil, and no kernel bugs, there was just a single pinctrl lock. That
was how the world was when (57291ce pinctrl: core device tree mapping
table parsing support) was written. In that case, there were
instances where the pinctrl mutex was already held when
pinctrl_register_map() was called, hence a "locked" parameter was
passed to the function to indicate that the mutex was already locked
(so we shouldn't lock it again).
A few years ago in (42fed7b pinctrl: move subsystem mutex to
pinctrl_dev struct), we switched to a separate pinctrl_maps_mutex.
...but (oops) we forgot to re-think about the whole "locked" parameter
for pinctrl_register_map(). Basically the "locked" parameter appears
to still refer to whether the bigger pinctrl_dev mutex is locked, but
we're using it to skip locks of our (now separate) pinctrl_maps_mutex.
That's kind of a bad thing(TM). Probably nobody noticed because most
of the calls to pinctrl_register_map happen at boot time and we've got
synchronous device probing. ...and even cases where we're
asynchronous don't end up actually hitting the race too often. ...but
after banging my head against the wall for a bug that reproduced 1 out
of 1000 reboots and lots of looking through kgdb, I finally noticed
this.
Anyway, we can now safely remove the "locked" parameter and go back to
a war-free, evil-free, and kernel-bug-free world.
Fixes: 42fed7ba44 ("pinctrl: move subsystem mutex to pinctrl_dev struct")
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Make sure that xen_swiotlb_init allocates buffers that are DMA capable
when at least one memblock is available below 4G. Otherwise we assume
that all devices on the SoC can cope with >4G addresses. We do this on
ARM and ARM64, where dom0 is mapped 1:1, so pfn == mfn in this case.
No functional changes on x86.
From: Chen Baozi <baozich@gmail.com>
Signed-off-by: Chen Baozi <baozich@gmail.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Tested-by: Chen Baozi <baozich@gmail.com>
Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
The following commit:
f893959b08 ("x86/fpu: Don't abuse drop_init_fpu() in flush_thread()")
removed drop_init_fpu() usage from flush_thread(). This seems to break
things for me - the Go 1.4 test suite fails all over the place with
floating point comparision errors (offending commit found through
bisection).
The functional change was that flush_thread() after this commit
only calls restore_init_xstate() when both use_eager_fpu() and
!used_math() are true. drop_init_fpu() (now fpu_reset_state()) calls
restore_init_xstate() regardless of whether current used_math() - apply
the same logic here.
Switch used_math() -> tsk_used_math(tsk) to consistently use the grabbed
tsk instead of current, like in the rest of flush_thread().
Tested-by: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Bobby Powers <bobbypowers@gmail.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Pekka Riikonen <priikone@iki.fi>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Suresh Siddha <sbsiddha@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: f893959b ("x86/fpu: Don't abuse drop_init_fpu() in flush_thread()")
Link: http://lkml.kernel.org/r/1430147441-9820-1-git-send-email-bobbypowers@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull EFI fixes from Matt Fleming:
* Avoid garbage names in efivarfs due to buggy firmware by zeroing
EFI variable name. (Ross Lagerwall)
* Stop erroneously dropping upper 32 bits of boot command line pointer
in EFI boot stub and stash them in ext_cmd_line_ptr. (Roy Franz)
* Fix double-free bug in error handling code path of EFI runtime map
code. (Dan Carpenter)
Signed-off-by: Ingo Molnar <mingo@kernel.org>
