| Age | Commit message (Collapse) | Author |
|---|
|
The device_for_each_child() doesn't require the returned value to be checked.
Thus, drop the dummy variable completely and have no warning anymore:
drivers/spi/spi.c: In function ‘spi_unregister_controller’:
drivers/spi/spi.c:2480:6: warning: variable ‘dummy’ set but not used [-Wunused-but-set-variable]
int dummy;
^~~~~
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Compiler is not happy about spi_set_cs_timing() prototype.
drivers/spi/spi.c:3016:6: warning: no previous prototype for ‘spi_set_cs_timing’ [-Wmissing-prototypes]
void spi_set_cs_timing(struct spi_device *spi, u8 setup, u8 hold,
^~~~~~~~~~~~~~~~~
Let's add it to the header.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Before clearing stream statuses, ensure RUN bit update has taken
effect by reading the value back.
Signed-off-by: Zhu Yingjiang <yingjiang.zhu@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
As per the HW recommendation, after setting the RUN bit
(start as 1, stop as 0), software must read the bit back
to make sure the bit is set right, before modifying related
control registers/re-starting the DMA engine.
Signed-off-by: Zhu Yingjiang <yingjiang.zhu@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Stream status and WAKESTS registers need to be cleared by writing
to them with snd_sof_dsp_write(). snd_sof_dsp_update_bits() only
writes if the value is changed and will result in not clearing
the status.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Unify suspend code by using SOF common function
hda_dsp_ctrl_stop_chip() which can handle both HDA
and non-HDA cases.
Signed-off-by: Zhu Yingjiang <yingjiang.zhu@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Add common hda_dsp_ctrl_stop_chip() function to stop controller with
the same function handling both HDA and non-HDA cases. This function
disables IRQs and clears status masks. When CONFIG_SND_SOC_SOF_HDA
is defined, also disables the CORB/RIRB, and stops i/o.
Signed-off-by: Zhu Yingjiang <yingjiang.zhu@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Unify ppcap function setup by using SOF common functions
for both HDA and non-HDA cases.
Signed-off-by: Zhu Yingjiang <yingjiang.zhu@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
We should use irq disabled mode when read/write hda registers from
thread context, as we need to hold the same bus->reg_lock in interrupt
context hda_dsp_stream_interrupt(), otherwise, when we are holding the
lock in hda_dsp_stream_hw_free() and the interrupt arrives, we will get
deadlock in the interrupt handler.
Error logs like this:
[ 5.603606] CPU0
[ 5.603606] ----
[ 5.603607] lock(&(&bus->reg_lock)->rlock);
[ 5.603608] <Interrupt>
[ 5.603609] lock(&(&bus->reg_lock)->rlock);
[ 5.603610]
*** DEADLOCK ***
[ 5.603611] 2 locks held by pulseaudio/2329:
[ 5.603612] #0: 000000005fcf26c6 (&card->mutex/1){+.+.}, at: dpcm_fe_dai_hw_free+0x2b/0x110 [snd_soc_core]
[ 5.603619] #1: 00000000ef369faf (&rtd->pcm_mutex){+.+.}, at: soc_pcm_hw_free+0x2e/0x1c0 [snd_soc_core]
The fix is simple, let's switch to use spin_lock/unlock_irq().
Reported-by: Xun Zhang <xun2.zhang@intel.com>
Signed-off-by: Keyon Jie <yang.jie@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Modify the stream interrupt handler to always wake up the
IRQ thread if the status register is valid. The IRQ thread
performs the check for stream interrupts and RIRB interrupts
in a loop to handle the case of missed interrupts when an
unsolicited response from the codec is received just before the
stream interrupt handler is completed.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Host and link DMA are decoupled during FE hw_params. So,
they must be coupled in hw_free if the link DMA channel
is idle.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Paused streams do not get suspended when the system enters S3.
So, clear and release link DMA channel for such streams in the
hda_dsp_set_hw_params_upon_resume() callback. Also, invalidate
the link DMA channel in the DAI config before restoring the
dai config upon resume. Also, modify the signature for the
set_hw_params_upon_resume() op to return an int.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Due to the HW programming sequence requirement that the host
and link DMA channels need to be coupled/decoupled during pcm
hw_params, the host DMA channel corresponding to the link
DMA channel in use for hostless streams needs to be reserved.
This is achieved by adding a host_reserved flag in the
sof_intel_hda_stream structure which is checked when assigning
a host DMA channel.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
The recommended HDA HW programming sequence for setting
the DMA format requires that the link DMA and host DMA
channels be coupled before setting the format. This
change means that host DMA or link DMA channels be
reserved even if only one is used.
Statically assigned link DMA channels would mean that
all the corresponding host DMA channels will need to be
reserved, leaving only a few channels available at run-time.
So, the suggestion here is to switch to dynamically assigning
both host DMA channels and link DMA channels are run-time.
The host DMA channel is assigned when the pcm
is opened as before. While choosing the link DMA channel,
if the host DMA channel corresponding to the link DMA channel
is already taken, the proposed method checks to make
sure that the BE is connected to the FE that has been assigned
this host DMA channel. Once the link DMA channel is assigned,
an IPC is sent to the DSP to set the link DMA channel.
The link DMA channel is freed during hw_free() and also in the
SUSPEND trigger callback. It will be re-assigned when hw_params
are set upon resume.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Add the cpu_dai_name member to snd_sof_dai and save the
cpu_dai_name while setting the DAI config.
The internal SOF representation will have to change at a later point
as well when we have multiple CPU dais.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Add a new macro to get sof_intel_hda_stream from hdac_ext_stream.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Add a snd_sof_dev member to sof_intel_hda_stream. This will be
used to access the snd_sof_dev during link hw_params callback.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
While the suspend function is already marked __maybe_unused,
the resume function is not, which leads to a warning when
CONFIG_PM is disabled:
sound/soc/codecs/cx2072x.c:1625:12: error: unused function 'cx2072x_runtime_resume' [-Werror,-Wunused-function]
Mark this one like the other one.
Fixes: a497a4363706 ("ASoC: Add support for Conexant CX2072X CODEC")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
Without this, we get a link error:
sound/soc/rockchip/rockchip_pdm.o: In function `rockchip_pdm_hw_params':
rockchip_pdm.c:(.text+0x754): undefined reference to `rational_best_approximation'
Fixes: 624e8e00acaf ("ASoC: rockchip: pdm: fixup pdm fractional div")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
This patch updates the Qualcomm SoC repo to a new location.
Signed-off-by: Andy Gross <agross@kernel.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Olof Johansson <olof@lixom.net>
|
|
With both the direct-addressed and indirect-addressed CCW paths
simplified to this point, the amount of shared code between them is
(hopefully) more easily visible. Move the processing of IDA-specific
bits into the direct-addressed path, and add some useful commentary of
what the individual pieces are doing. This allows us to remove the
entire ccwchain_fetch_idal() routine and maintain a single function
for any non-TIC CCW.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-10-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
This is purely deck furniture, to help understand the merge of the
direct and indirect handlers.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-9-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
Now that both CCW codepaths build this nested array:
ccwchain->pfn_array_table[1]->pfn_array[#idaws/#pages]
We can collapse this into simply:
ccwchain->pfn_array[#idaws/#pages]
Let's do that, so that we don't have to continually navigate two
nested arrays when the first array always has a count of one.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-8-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
Now that pfn_array_table[] is always an array of 1, it seems silly to
check for the very first entry in an array in the middle of two nested
loops, since we know it'll only ever happen once.
Let's move this outside the loops to simplify things, even though
the "k" variable is still necessary.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-7-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
While processing a channel program, we currently have two nested
arrays that carry a slightly different structure. The direct CCW
path creates this:
ccwchain->pfn_array_table[1]->pfn_array[#pages]
while an IDA CCW creates:
ccwchain->pfn_array_table[#idaws]->pfn_array[1]
The distinction appears to state that each pfn_array_table entry
points to an array of contiguous pages, represented by a pfn_array,
um, array. Since the direct-addressed scenario can ONLY represent
contiguous pages, it makes the intermediate array necessary but
difficult to recognize. Meanwhile, since an IDAL can contain
non-contiguous pages and there is no logic in vfio-ccw to detect
adjacent IDAWs, it is the second array that is necessary but appearing
to be superfluous.
I am not aware of any documentation that states the pfn_array[] needs
to be of contiguous pages; it is just what the code does today.
I don't see any reason for this either, let's just flip the IDA
codepath around so that it generates:
ch_pat->pfn_array_table[1]->pfn_array[#idaws]
This will bring it in line with the direct-addressed codepath,
so that we can understand the behavior of this memory regardless
of what type of CCW is being processed. And it means the casual
observer does not need to know/care whether the pfn_array[]
represents contiguous pages or not.
NB: The existing vfio-ccw code only supports 4K-block Format-2 IDAs,
so that "#pages" == "#idaws" in this area. This means that we will
have difficulty with this overlap in terminology if support for
Format-1 or 2K-block Format-2 IDAs is ever added. I don't think that
this patch changes our ability to make that distinction.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-6-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
It is now pretty apparent that ccwchain_handle_ccw()
(nee ccwchain_handle_tic()) does everything that cp_init()
wants to do.
Let's remove that duplicated code from cp_init() and let
ccwchain_handle_ccw() handle it itself.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-5-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
Refactor ccwchain_handle_tic() into a routine that handles a channel
program address (which itself is a CCW pointer), rather than a CCW pointer
that is only a TIC CCW. This will make it easier to reuse this code for
other CCW commands.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-4-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
Extract the "does the target of this TIC already exist?" check from
ccwchain_handle_tic(), so that it's easier to refactor that function
into one that cp_init() is able to use.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-3-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
The routine cp_free() does nothing but call cp_unpin_free(), and while
most places call cp_free() there is one caller of cp_unpin_free() used
when the cp is guaranteed to have not been marked initialized.
This seems like a dubious way to make a distinction, so let's combine
these routines and make cp_free() do all the work.
Signed-off-by: Eric Farman <farman@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190606202831.44135-2-farman@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
|
|
If cmd19 timeout or response crcerr occurs during execute_tuning(),
it need invoke msdc_reset_hw(). Otherwise SDIO IRQ can't be detected.
Signed-off-by: jjian zhou <jjian.zhou@mediatek.com>
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Yong Mao <yong.mao@mediatek.com>
Fixes: 5215b2e952f3 ("mmc: mediatek: Add MMC_CAP_SDIO_IRQ support")
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
SDIO IRQ is triggered by low level. It need disable SDIO IRQ
detected function. Otherwise the interrupt register can't be cleared.
It will process the interrupt more.
Signed-off-by: Jjian Zhou <jjian.zhou@mediatek.com>
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Yong Mao <yong.mao@mediatek.com>
Fixes: 5215b2e952f3 ("mmc: mediatek: Add MMC_CAP_SDIO_IRQ support")
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
We don't have a reproducible error case, yet our BSP team suggested that
the mmc_switch_status() command in mmc_select_hs400() should come after
the callback into the driver completing HS400 setup. It makes sense to
me because we want the status of a fully setup HS400, so it will
increase the reliability of the mmc_switch_status() command.
Reported-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Fixes: ba6c7ac3a2f4 ("mmc: core: more fine-grained hooks for HS400 tuning")
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
If the cache line size is greater than ARCH_DMA_MINALIGN (128),
the warning shows and it's tainted as TAINT_CPU_OUT_OF_SPEC.
However, it's not good because as discussed in the thread [1], the cpu
cache line size will be problem only on non-coherent devices.
Since the coherent flag is already introduced to struct device,
show the warning only if the device is non-coherent device and
ARCH_DMA_MINALIGN is smaller than the cpu cache size.
[1] https://lore.kernel.org/linux-arm-kernel/20180514145703.celnlobzn3uh5tc2@localhost/
Signed-off-by: Masayoshi Mizuma <m.mizuma@jp.fujitsu.com>
Reviewed-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Tested-by: Zhang Lei <zhang.lei@jp.fujitsu.com>
[catalin.marinas@arm.com: removed 'if' block for WARN_TAINT]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
|
Because RISC-V compliant implementations can cache invalid entries
in TLB, an SFENCE.VMA is necessary after changes to the page table.
This patch adds an SFENCE.vma for the vmalloc_fault path.
Signed-off-by: ShihPo Hung <shihpo.hung@sifive.com>
[paul.walmsley@sifive.com: reversed tab->whitespace conversion,
wrapped comment lines]
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@sifive.com>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: linux-riscv@lists.infradead.org
Cc: stable@vger.kernel.org
|
|
Since raw_cpu_xchg() doesn't need to be IRQ-safe, like
this_cpu_xchg(), we can use a simple load-store instead of the cmpxchg
loop.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Nadav reported that code-gen changed because of the this_cpu_*()
constraints, avoid this for select_idle_cpu() because that runs with
preemption (and IRQs) disabled anyway.
Reported-by: Nadav Amit <nadav.amit@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Nadav reported that since the this_cpu_*() ops got asm-volatile
constraints on, code generation suffered for do_IRQ(), but since this
is all with IRQs disabled we can use __this_cpu_*().
smp_x86_platform_ipi 234 222 -12,+0
smp_kvm_posted_intr_ipi 74 66 -8,+0
smp_kvm_posted_intr_wakeup_ipi 86 78 -8,+0
smp_apic_timer_interrupt 292 284 -8,+0
smp_kvm_posted_intr_nested_ipi 74 66 -8,+0
do_IRQ 195 187 -8,+0
Reported-by: Nadav Amit <nadav.amit@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Nadav reported that since this_cpu_read() became asm-volatile, many
smp_processor_id() users generated worse code due to the extra
constraints.
However since smp_processor_id() is reading a stable value, we can use
__this_cpu_read().
While this does reduce text size somewhat, this mostly results in code
movement to .text.unlikely as a result of more/larger .cold.
subfunctions. Less text on the hotpath is good for I$.
$ ./compare.sh defconfig-build1 defconfig-build2 vmlinux.o
setup_APIC_ibs 90 98 -12,+20
force_ibs_eilvt_setup 400 413 -57,+70
pci_serr_error 109 104 -54,+49
pci_serr_error 109 104 -54,+49
unknown_nmi_error 125 120 -76,+71
unknown_nmi_error 125 120 -76,+71
io_check_error 125 132 -97,+104
intel_thermal_interrupt 730 822 +92,+0
intel_init_thermal 951 945 -6,+0
generic_get_mtrr 301 294 -7,+0
generic_get_mtrr 301 294 -7,+0
generic_set_all 749 754 -44,+49
get_fixed_ranges 352 360 -41,+49
x86_acpi_suspend_lowlevel 369 363 -6,+0
check_tsc_sync_source 412 412 -71,+71
irq_migrate_all_off_this_cpu 662 674 -14,+26
clocksource_watchdog 748 748 -113,+113
__perf_event_account_interrupt 204 197 -7,+0
attempt_merge 1748 1741 -7,+0
intel_guc_send_ct 1424 1409 -15,+0
__fini_doorbell 235 231 -4,+0
bdw_set_cdclk 928 923 -5,+0
gen11_dsi_disable 1571 1556 -15,+0
gmbus_wait 493 488 -5,+0
md_make_request 376 369 -7,+0
__split_and_process_bio 543 536 -7,+0
delay_tsc 96 89 -7,+0
hsw_disable_pc8 696 691 -5,+0
tsc_verify_tsc_adjust 215 228 -22,+35
cpuidle_driver_unref 56 49 -7,+0
blk_account_io_completion 159 148 -11,+0
mtrr_wrmsr 95 99 -29,+33
__intel_wait_for_register_fw 401 419 +18,+0
cpuidle_driver_ref 43 36 -7,+0
cpuidle_get_driver 15 8 -7,+0
blk_account_io_done 535 528 -7,+0
irq_migrate_all_off_this_cpu 662 674 -14,+26
check_tsc_sync_source 412 412 -71,+71
irq_wait_for_poll 170 163 -7,+0
generic_end_io_acct 329 322 -7,+0
x86_acpi_suspend_lowlevel 369 363 -6,+0
nohz_balance_enter_idle 198 191 -7,+0
generic_start_io_acct 254 247 -7,+0
blk_account_io_start 341 334 -7,+0
perf_event_task_tick 682 675 -7,+0
intel_init_thermal 951 945 -6,+0
amd_e400_c1e_apic_setup 47 51 -28,+32
setup_APIC_eilvt 350 328 -22,+0
hsw_enable_pc8 1611 1605 -6,+0
total 12985947 12985892 -994,+939
Reported-by: Nadav Amit <nadav.amit@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Nadav Amit reported that commit:
b59167ac7baf ("x86/percpu: Fix this_cpu_read()")
added a bunch of constraints to all sorts of code; and while some of
that was correct and desired, some of that seems superfluous.
The thing is, the this_cpu_*() operations are defined IRQ-safe, this
means the values are subject to change from IRQs, and thus must be
reloaded.
Also, the generic form:
local_irq_save()
__this_cpu_read()
local_irq_restore()
would not allow the re-use of previous values; if by nothing else,
then the barrier()s implied by local_irq_*().
Which raises the point that percpu_from_op() and the others also need
that volatile.
OTOH __this_cpu_*() operations are not IRQ-safe and assume external
preempt/IRQ disabling and could thus be allowed more room for
optimization.
This makes the this_cpu_*() vs __this_cpu_*() behaviour more
consistent with other architectures.
$ ./compare.sh defconfig-build defconfig-build1 vmlinux.o
x86_pmu_cancel_txn 80 71 -9,+0
__text_poke 919 964 +45,+0
do_user_addr_fault 1082 1058 -24,+0
__do_page_fault 1194 1178 -16,+0
do_exit 2995 3027 -43,+75
process_one_work 1008 989 -67,+48
finish_task_switch 524 505 -19,+0
__schedule_bug 103 98 -59,+54
__schedule_bug 103 98 -59,+54
__sched_setscheduler 2015 2030 +15,+0
freeze_processes 203 230 +31,-4
rcu_gp_kthread_wake 106 99 -7,+0
rcu_core 1841 1834 -7,+0
call_timer_fn 298 286 -12,+0
can_stop_idle_tick 146 139 -31,+24
perf_pending_event 253 239 -14,+0
shmem_alloc_page 209 213 +4,+0
__alloc_pages_slowpath 3284 3269 -15,+0
umount_tree 671 694 +23,+0
advance_transaction 803 798 -5,+0
con_put_char 71 51 -20,+0
xhci_urb_enqueue 1302 1295 -7,+0
xhci_urb_enqueue 1302 1295 -7,+0
tcp_sacktag_write_queue 2130 2075 -55,+0
tcp_try_undo_loss 229 208 -21,+0
tcp_v4_inbound_md5_hash 438 411 -31,+4
tcp_v4_inbound_md5_hash 438 411 -31,+4
tcp_v6_inbound_md5_hash 469 411 -33,-25
tcp_v6_inbound_md5_hash 469 411 -33,-25
restricted_pointer 434 420 -14,+0
irq_exit 162 154 -8,+0
get_perf_callchain 638 624 -14,+0
rt_mutex_trylock 169 156 -13,+0
avc_has_extended_perms 1092 1089 -3,+0
avc_has_perm_noaudit 309 306 -3,+0
__perf_sw_event 138 122 -16,+0
perf_swevent_get_recursion_context 116 102 -14,+0
__local_bh_enable_ip 93 72 -21,+0
xfrm_input 4175 4161 -14,+0
avc_has_perm 446 443 -3,+0
vm_events_fold_cpu 57 56 -1,+0
vfree 68 61 -7,+0
freeze_processes 203 230 +31,-4
_local_bh_enable 44 30 -14,+0
ip_do_fragment 1982 1944 -38,+0
do_exit 2995 3027 -43,+75
__do_softirq 742 724 -18,+0
cpu_init 1510 1489 -21,+0
account_system_time 80 79 -1,+0
total 12985281 12984819 -742,+280
Reported-by: Nadav Amit <nadav.amit@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20181206112433.GB13675@hirez.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
My @arm.com address will stop working at the end of August, so update to
my @kernel.org address where you'll still be able to reach me.
When I say "stop working" I really mean "will go to my line manager", so
send patches there at your peril because they may reply with roadmaps
and spreadsheets. You have been warned.
Cc: Will Deacon <will@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: arm-soc <arm@kernel.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
|
The upper bits of the count field is used as reader count. When
sufficient number of active readers are present, the most significant
bit will be set and the count becomes negative. If the number of active
readers keep on piling up, we may eventually overflow the reader counts.
This is not likely to happen unless the number of bits reserved for
reader count is reduced because those bits are need for other purpose.
To prevent this count overflow from happening, the most significant
bit is now treated as a guard bit (RWSEM_FLAG_READFAIL). Read-lock
attempts will now fail for both the fast and slow paths whenever this
bit is set. So all those extra readers will be put to sleep in the wait
list. Wakeup will not happen until the reader count reaches 0.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-17-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Reader optimistic spinning is helpful when the reader critical section
is short and there aren't that many readers around. It makes readers
relatively more preferred than writers. When a writer times out spinning
on a reader-owned lock and set the nospinnable bits, there are two main
reasons for that.
1) The reader critical section is long, perhaps the task sleeps after
acquiring the read lock.
2) There are just too many readers contending the lock causing it to
take a while to service all of them.
In the former case, long reader critical section will impede the progress
of writers which is usually more important for system performance.
In the later case, reader optimistic spinning tends to make the reader
groups that contain readers that acquire the lock together smaller
leading to more of them. That may hurt performance in some cases. In
other words, the setting of nonspinnable bits indicates that reader
optimistic spinning may not be helpful for those workloads that cause it.
Therefore, any writers that have observed the setting of the writer
nonspinnable bit for a given rwsem after they fail to acquire the lock
via optimistic spinning will set the reader nonspinnable bit once they
acquire the write lock. Similarly, readers that observe the setting
of reader nonspinnable bit at slowpath entry will also set the reader
nonspinnable bit when they acquire the read lock via the wakeup path.
Once the reader nonspinnable bit is on, it will only be reset when
a writer is able to acquire the rwsem in the fast path or somehow a
reader or writer in the slowpath doesn't observe the nonspinable bit.
This is to discourage reader optmistic spinning on that particular
rwsem and make writers more preferred. This adaptive disabling of reader
optimistic spinning will alleviate some of the negative side effect of
this feature.
In addition, this patch tries to make readers in the spinning queue
follow the phase-fair principle after quitting optimistic spinning
by checking if another reader has somehow acquired a read lock after
this reader enters the optimistic spinning queue. If so and the rwsem
is still reader-owned, this reader is in the right read-phase and can
attempt to acquire the lock.
On a 2-socket 40-core 80-thread Skylake system, the page_fault1 test of
the will-it-scale benchmark was run with various number of threads. The
number of operations done before reader optimistic spinning patches,
this patch and after this patch were:
Threads Before rspin Before patch After patch %change
------- ------------ ------------ ----------- -------
20 5541068 5345484 5455667 -3.5%/ +2.1%
40 10185150 7292313 9219276 -28.5%/+26.4%
60 8196733 6460517 7181209 -21.2%/+11.2%
80 9508864 6739559 8107025 -29.1%/+20.3%
This patch doesn't recover all the lost performance, but it is more
than half. Given the fact that reader optimistic spinning does benefit
some workloads, this is a good compromise.
Using the rwsem locking microbenchmark with very short critical section,
this patch doesn't have too much impact on locking performance as shown
by the locking rates (kops/s) below with equal numbers of readers and
writers before and after this patch:
# of Threads Pre-patch Post-patch
------------ --------- ----------
2 4,730 4,969
4 4,814 4,786
8 4,866 4,815
16 4,715 4,511
32 3,338 3,500
64 3,212 3,389
80 3,110 3,044
When running the locking microbenchmark with 40 dedicated reader and writer
threads, however, the reader performance is curtailed to favor the writer.
Before patch:
40 readers, Iterations Min/Mean/Max = 204,026/234,309/254,816
40 writers, Iterations Min/Mean/Max = 88,515/95,884/115,644
After patch:
40 readers, Iterations Min/Mean/Max = 33,813/35,260/36,791
40 writers, Iterations Min/Mean/Max = 95,368/96,565/97,798
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-16-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
When the rwsem is owned by reader, writers stop optimistic spinning
simply because there is no easy way to figure out if all the readers
are actively running or not. However, there are scenarios where
the readers are unlikely to sleep and optimistic spinning can help
performance.
This patch provides a simple mechanism for spinning on a reader-owned
rwsem by a writer. It is a time threshold based spinning where the
allowable spinning time can vary from 10us to 25us depending on the
condition of the rwsem.
When the time threshold is exceeded, the nonspinnable bits will be set
in the owner field to indicate that no more optimistic spinning will
be allowed on this rwsem until it becomes writer owned again. Not even
readers is allowed to acquire the reader-locked rwsem by optimistic
spinning for fairness.
We also want a writer to acquire the lock after the readers hold the
lock for a relatively long time. In order to give preference to writers
under such a circumstance, the single RWSEM_NONSPINNABLE bit is now split
into two - one for reader and one for writer. When optimistic spinning
is disabled, both bits will be set. When the reader count drop down
to 0, the writer nonspinnable bit will be cleared to allow writers to
spin on the lock, but not the readers. When a writer acquires the lock,
it will write its own task structure pointer into sem->owner and clear
the reader nonspinnable bit in the process.
The time taken for each iteration of the reader-owned rwsem spinning
loop varies. Below are sample minimum elapsed times for 16 iterations
of the loop.
System Time for 16 Iterations
------ ----------------------
1-socket Skylake ~800ns
4-socket Broadwell ~300ns
2-socket ThunderX2 (arm64) ~250ns
When the lock cacheline is contended, we can see up to almost 10X
increase in elapsed time. So 25us will be at most 500, 1300 and 1600
iterations for each of the above systems.
With a locking microbenchmark running on 5.1 based kernel, the total
locking rates (in kops/s) on a 8-socket IvyBridge-EX system with
equal numbers of readers and writers before and after this patch were
as follows:
# of Threads Pre-patch Post-patch
------------ --------- ----------
2 1,759 6,684
4 1,684 6,738
8 1,074 7,222
16 900 7,163
32 458 7,316
64 208 520
128 168 425
240 143 474
This patch gives a big boost in performance for mixed reader/writer
workloads.
With 32 locking threads, the rwsem lock event data were:
rwsem_opt_fail=79850
rwsem_opt_nospin=5069
rwsem_opt_rlock=597484
rwsem_opt_wlock=957339
rwsem_sleep_reader=57782
rwsem_sleep_writer=55663
With 64 locking threads, the data looked like:
rwsem_opt_fail=346723
rwsem_opt_nospin=6293
rwsem_opt_rlock=1127119
rwsem_opt_wlock=1400628
rwsem_sleep_reader=308201
rwsem_sleep_writer=72281
So a lot more threads acquired the lock in the slowpath and more threads
went to sleep.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-15-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
The rwsem->owner contains not just the task structure pointer, it also
holds some flags for storing the current state of the rwsem. Some of
the flags may have to be atomically updated. To reflect the new reality,
the owner is now changed to an atomic_long_t type.
New helper functions are added to properly separate out the task
structure pointer and the embedded flags.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-14-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
This patch enables readers to optimistically spin on a
rwsem when it is owned by a writer instead of going to sleep
directly. The rwsem_can_spin_on_owner() function is extracted
out of rwsem_optimistic_spin() and is called directly by
rwsem_down_read_slowpath() and rwsem_down_write_slowpath().
With a locking microbenchmark running on 5.1 based kernel, the total
locking rates (in kops/s) on a 8-socket IvyBrige-EX system with equal
numbers of readers and writers before and after the patch were as
follows:
# of Threads Pre-patch Post-patch
------------ --------- ----------
4 1,674 1,684
8 1,062 1,074
16 924 900
32 300 458
64 195 208
128 164 168
240 149 143
The performance change wasn't significant in this case, but this change
is required by a follow-on patch.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-13-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Bit 1 of sem->owner (RWSEM_ANONYMOUSLY_OWNED) is used to designate an
anonymous owner - readers or an anonymous writer. The setting of this
anonymous bit is used as an indicator that optimistic spinning cannot
be done on this rwsem.
With the upcoming reader optimistic spinning patches, a reader-owned
rwsem can be spinned on for a limit period of time. We still need
this bit to indicate a rwsem is nonspinnable, but not setting this
bit loses its meaning that the owner is known. So rename the bit
to RWSEM_NONSPINNABLE to clarify its meaning.
This patch also fixes a DEBUG_RWSEMS_WARN_ON() bug in __up_write().
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-12-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
When the front of the wait queue is a reader, other readers
immediately following the first reader will also be woken up at the
same time. However, if there is a writer in between. Those readers
behind the writer will not be woken up.
Because of optimistic spinning, the lock acquisition order is not FIFO
anyway. The lock handoff mechanism will ensure that lock starvation
will not happen.
Assuming that the lock hold times of the other readers still in the
queue will be about the same as the readers that are being woken up,
there is really not much additional cost other than the additional
latency due to the wakeup of additional tasks by the waker. Therefore
all the readers up to a maximum of 256 in the queue are woken up when
the first waiter is a reader to improve reader throughput. This is
somewhat similar in concept to a phase-fair R/W lock.
With a locking microbenchmark running on 5.1 based kernel, the total
locking rates (in kops/s) on a 8-socket IvyBridge-EX system with
equal numbers of readers and writers before and after this patch were
as follows:
# of Threads Pre-Patch Post-patch
------------ --------- ----------
4 1,641 1,674
8 731 1,062
16 564 924
32 78 300
64 38 195
240 50 149
There is no performance gain at low contention level. At high contention
level, however, this patch gives a pretty decent performance boost.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-11-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
An RT task can do optimistic spinning only if the lock holder is
actually running. If the state of the lock holder isn't known, there
is a possibility that high priority of the RT task may block forward
progress of the lock holder if it happens to reside on the same CPU.
This will lead to deadlock. So we have to make sure that an RT task
will not spin on a reader-owned rwsem.
When the owner is temporarily set to NULL, there are two cases
where we may want to continue spinning:
1) The lock owner is in the process of releasing the lock, sem->owner
is cleared but the lock has not been released yet.
2) The lock was free and owner cleared, but another task just comes
in and acquire the lock before we try to get it. The new owner may
be a spinnable writer.
So an RT task is now made to retry one more time to see if it can
acquire the lock or continue spinning on the new owning writer.
When testing on a 8-socket IvyBridge-EX system, the one additional retry
seems to improve locking performance of RT write locking threads under
heavy contentions. The table below shows the locking rates (in kops/s)
with various write locking threads before and after the patch.
Locking threads Pre-patch Post-patch
--------------- --------- -----------
4 2,753 2,608
8 2,529 2,520
16 1,727 1,918
32 1,263 1,956
64 889 1,343
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-10-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
With the use of wake_q, we can do task wakeups without holding the
wait_lock. There is one exception in the rwsem code, though. It is
when the writer in the slowpath detects that there are waiters ahead
but the rwsem is not held by a writer. This can lead to a long wait_lock
hold time especially when a large number of readers are to be woken up.
Remediate this situation by releasing the wait_lock before waking
up tasks and re-acquiring it afterward. The rwsem_try_write_lock()
function is also modified to read the rwsem count directly to avoid
stale count value.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-9-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Because of writer lock stealing, it is possible that a constant
stream of incoming writers will cause a waiting writer or reader to
wait indefinitely leading to lock starvation.
This patch implements a lock handoff mechanism to disable lock stealing
and force lock handoff to the first waiter or waiters (for readers)
in the queue after at least a 4ms waiting period unless it is a RT
writer task which doesn't need to wait. The waiting period is used to
avoid discouraging lock stealing too much to affect performance.
The setting and clearing of the handoff bit is serialized by the
wait_lock. So racing is not possible.
A rwsem microbenchmark was run for 5 seconds on a 2-socket 40-core
80-thread Skylake system with a v5.1 based kernel and 240 write_lock
threads with 5us sleep critical section.
Before the patch, the min/mean/max numbers of locking operations for
the locking threads were 1/7,792/173,696. After the patch, the figures
became 5,842/6,542/7,458. It can be seen that the rwsem became much
more fair, though there was a drop of about 16% in the mean locking
operations done which was a tradeoff of having better fairness.
Making the waiter set the handoff bit right after the first wakeup can
impact performance especially with a mixed reader/writer workload. With
the same microbenchmark with short critical section and equal number of
reader and writer threads (40/40), the reader/writer locking operation
counts with the current patch were:
40 readers, Iterations Min/Mean/Max = 1,793/1,794/1,796
40 writers, Iterations Min/Mean/Max = 1,793/34,956/86,081
By making waiter set handoff bit immediately after wakeup:
40 readers, Iterations Min/Mean/Max = 43/44/46
40 writers, Iterations Min/Mean/Max = 43/1,263/3,191
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Link: https://lkml.kernel.org/r/20190520205918.22251-8-longman@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
