Andres reports that buffered IO seems to suck up more cycles than we
would like, and he narrowed it down to the fact that the io-wq workers
will briefly spin for more work on completion of a work item. This was
a win on the networking side, but apparently some other cases take a
hit because of it. Remove the optimization to avoid burning more CPU
than we have to for disk IO.
Reported-by: Andres Freund <andres@anarazel.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After making ext4 support iopoll method:
let ext4_file_operations's iopoll method be iomap_dio_iopoll(),
we found fio can easily hang in fio_ioring_getevents() with below fio
job:
rm -f testfile; sync;
sudo fio -name=fiotest -filename=testfile -iodepth=128 -thread
-rw=write -ioengine=io_uring -hipri=1 -sqthread_poll=1 -direct=1
-bs=4k -size=10G -numjobs=8 -runtime=2000 -group_reporting
with IORING_SETUP_SQPOLL and IORING_SETUP_IOPOLL enabled.
There are two issues that results in this hang, one reason is that
when IORING_SETUP_SQPOLL and IORING_SETUP_IOPOLL are enabled, fio
does not use io_uring_enter to get completed events, it relies on
kernel io_sq_thread to poll for completed events.
Another reason is that there is a race: when io_submit_sqes() in
io_sq_thread() submits a batch of sqes, variable 'inflight' will
record the number of submitted reqs, then io_sq_thread will poll for
reqs which have been added to poll_list. But note, if some previous
reqs have been punted to io worker, these reqs will won't be in
poll_list timely. io_sq_thread() will only poll for a part of previous
submitted reqs, and then find poll_list is empty, reset variable
'inflight' to be zero. If app just waits these deferred reqs and does
not wake up io_sq_thread again, then hang happens.
For app that entirely relies on io_sq_thread to poll completed requests,
let io_iopoll_req_issued() wake up io_sq_thread properly when adding new
element to poll_list, and when io_sq_thread prepares to sleep, check
whether poll_list is empty again, if not empty, continue to poll.
Signed-off-by: Xiaoguang Wang <xiaoguang.wang@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
KASAN is reporting that __blk_add_trace() has a use-after-free issue
when accessing q->blk_trace. Indeed the switching of block tracing (and
thus eventual freeing of q->blk_trace) is completely unsynchronized with
the currently running tracing and thus it can happen that the blk_trace
structure is being freed just while __blk_add_trace() works on it.
Protect accesses to q->blk_trace by RCU during tracing and make sure we
wait for the end of RCU grace period when shutting down tracing. Luckily
that is rare enough event that we can afford that. Note that postponing
the freeing of blk_trace to an RCU callback should better be avoided as
it could have unexpected user visible side-effects as debugfs files
would be still existing for a short while block tracing has been shut
down.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=205711
CC: stable@vger.kernel.org
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reported-by: Tristan Madani <tristmd@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The function was currently used internal by the gpiolib. Since commit
56cc3af4e8 ("pinctrl: da9062: add driver support") it is also used by
drivers so we need to export the symbol.
Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
Link: https://lore.kernel.org/r/20200225093102.10964-1-m.felsch@pengutronix.de
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Now that .eh_frame sections for the files in setup.elf and realmode.elf
are not generated anymore, the linker scripts don't need the special
output section name /DISCARD/ any more.
Remove the one in the main kernel linker script as well, since there are
no .eh_frame sections already, and fix up a comment referencing .eh_frame.
Update the comment in asm/dwarf2.h referring to .eh_frame so it continues
to make sense, as well as being more specific.
[ bp: Touch up commit message. ]
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Tested-by: Nathan Chancellor <natechancellor@gmail.com>
Link: https://lkml.kernel.org/r/20200224232129.597160-3-nivedita@alum.mit.edu
The check on p->sink looks bogus, I believe it should be p->source
since the following code blocks are related to p->source. Fix
this by replacing p->sink with p->source.
Fixes: 24c8d14192 ("ASoC: Intel: mrfld: add DSP core controls")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Addresses-Coverity: ("Copy-paste error")
Link: https://lore.kernel.org/r/20191119113640.166940-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
While discussing a patch to discard .eh_frame from the compressed
vmlinux using the linker script, Fangrui Song pointed out [1] that these
sections shouldn't exist in the first place because arch/x86/Makefile
uses -fno-asynchronous-unwind-tables.
It turns out this is because the Makefiles used to build the compressed
kernel redefine KBUILD_CFLAGS, dropping this flag.
Add the flag to the Makefile for the compressed kernel, as well as the
EFI stub Makefile to fix this.
Also add the flag to boot/Makefile and realmode/rm/Makefile so that the
kernel's boot code (boot/setup.elf) and realmode trampoline
(realmode/rm/realmode.elf) won't be compiled with .eh_frame sections,
since their linker scripts also just discard them.
[1] https://lore.kernel.org/lkml/20200222185806.ywnqhfqmy67akfsa@google.com/
Suggested-by: Fangrui Song <maskray@google.com>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Tested-by: Nathan Chancellor <natechancellor@gmail.com>
Link: https://lkml.kernel.org/r/20200224232129.597160-2-nivedita@alum.mit.edu
Add a new API dw_pcie_ep_init_notify() to let platform drivers
call it when the core is available for initialization.
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
Add support to send notifications to EPF from EPC once the core
registers initialization is complete.
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
Split core initialization code for EP mode into two, one that doesn't
touch core registers and the other that touches core registers. The latter
would be called/skipped based on the EPC feature 'core_init_notifier'.
In platforms where this is skipped, it would be called indirectly
through hooks from the endpoint function driver.
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
Add a new feature core_init_notifier for cores that can notify about
their availability for initialization.
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
Before nftables commit fb9cea50e8b3 ("main: enforce options before
commands"), 'nft list ruleset -a' happened to work, but it's wrong
and won't work anymore. Replace it by 'nft -a list ruleset'.
Reported-by: Chen Yi <yiche@redhat.com>
Fixes: 611973c1e0 ("selftests: netfilter: Introduce tests for sets with range concatenation")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The whole implementation is done in asm-generic/cmpxchg.h file and there is
no reason to duplicate it.
Also do not include asm-generic/cmpxchg-local.h because it is already
included from asm-generic/cmpxchg.h
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Adding SMP requires to have percpu section defined.
Signed-off-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
boot_cpuid is not used on uni processor system that's why can be removed.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Without this header local_save_flags() is not defined.
Signed-off-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Do the real sync by using mbar instruction.
Signed-off-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
There is no reason to keep empty headers if asm-generic one can be used
instead. That's why remove empty file and use asm-generic equivalent.
cputable.h is completely unused that's why remove it.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Early printk has been removed already that's why this setting doesn't make
any sense. Also change printk level from pr_info to pr_err.
Fixes: 96f0e6fcc9 ("microblaze: remove redundant early_printk support")
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
There is no reason to have this empty file if it is just link to
asm-generic/tlb.h.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Covert all headers to SPDX.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
The Fintek F81532A/534A/535/536 is USB-to-2/4/8/12 serial ports device
and the serial ports are default disabled. Each port contains max 3 pins
GPIO and the 3 pins are default pull high with input mode.
When the serial port had activated (running probe()), we'll transform the
3 pins from GPIO function publicly to control Tranceiver privately use.
We'll default set to 0/0/1 for control transceiver to RS232 mode.
Otherwise, If the serial port is not active, the 3 pins is in GPIO mode
and controlled by global GPIO device with VID/PID: 2c42/16f8.
Signed-off-by: Ji-Ze Hong (Peter Hong) <hpeter+linux_kernel@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
The Fintek F81532A/534A/535/536 is USB-to-2/4/8/12 serial ports device
and the serial port is default disabled when plugin computer.
The IC is contains devices as following:
1. HUB (all devices is connected with this hub)
2. GPIO/Control device. (enable serial port and control GPIOs)
3. serial port 1 to x (2/4/8/12)
It's most same with F81232, the UART device is difference as follow:
1. TX/RX bulk size is 128/512bytes
2. RX bulk layout change:
F81232: [LSR(1Byte)+DATA(1Byte)][LSR(1Byte)+DATA(1Byte)]...
F81534A:[LEN][Data.....][LSR]
Signed-off-by: Ji-Ze Hong (Peter Hong) <hpeter+linux_kernel@gmail.com>
[johan: reword an error message]
Signed-off-by: Johan Hovold <johan@kernel.org>
Use devm_kzalloc() to replace kzalloc() in port_probe().
Signed-off-by: Ji-Ze Hong (Peter Hong) <hpeter+linux_kernel@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Add tx_empty() function for F81232. Without this, console redirection will
get garbage data.
Signed-off-by: Ji-Ze Hong (Peter Hong) <hpeter+linux_kernel@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Extract LSR handler to function that can be re-used by
F81532A/534A/535/536.
Signed-off-by: Ji-Ze Hong (Peter Hong) <hpeter+linux_kernel@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
There is a need to disable VM before jump to zero reset vector.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Eventhough the current documentation explains that the reference count
gets incremented by both kref_init() and kref_get(), it is often
misunderstood that only one instance of kref_put() is needed in the
example code. So let's clarify that a bit.
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
This script allows sysctl documentation to be checked against the
kernel source code, to identify missing or obsolete entries. Running
it against 5.5 shows for example that sysctl/kernel.rst has two
obsolete entries and is missing 52 entries.
Signed-off-by: Stephen Kitt <steve@sk2.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Based on the implementation in arch/x86/kernel/acpi/sleep.c, in
particular the acpi_sleep_setup() function.
Signed-off-by: Stephen Kitt <steve@sk2.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
If a SoundWire link is enabled, but there are no Slave devices exposed
in firmware tables for this link, or no Slaves in ATTACHED or ALERT
mode, the CMD_IGNORED/-ENODATA error code on a broadcast write is
perfectly legit.
Filter this case to report errors and let the caller deal with the
CMD_IGNORED case.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-11-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
SoundWire supports two clock stop modes. Add support to handle the
clock stop modes and add pm_runtime calls in the bus.
Credits: this patch is based on an earlier internal contribution by
Vinod Koul, Sanyog Kale, Shreyas Nc and Hardik Shah.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-10-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
There are two types of io errors when processing alert event.
a) the Master detects an ALERT status for e.g. a jack event and
invokes the implementation-defined function in the Slave driver to
check the jack status. At this time the codec is just suspended, so io
registers can't be accessed.
b) when waking up from clock stop mode1 state, where the bus needs a
complete re-enumeration, Slave registers can't be accessed until the
enumeration is complete.
This patch resumes the Slave device and waits for initialization
complete when processing slave alert event, so that registers on the
Slave can be accessed without timeouts or io errors.
Signed-off-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-9-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Before removing the slave device, disable pm_runtime to prevent any
race condition with the resume being executed after the bus and slave
devices are removed.
Since this pm_runtime_disable() is handled in common routines,
implementations of Slave drivers do not need to call it in their
.remove() routine.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-8-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
When resuming with a bus reset, we need to re-enumerate and restart
from UNATTACHED. The helper added in this patch helps implement a more
robust state machine avoiding race conditions on resume.
The unattach request is stored and will be used by Slave drivers, if
needed: Intel validation exposed a corner case where the Slave device
may transition to D3 when streaming stops, but streaming restarts
before the Master transitions to D3. In that case, the Slave status
was not cleared as UNATTACHED by the Master resuming, and the
wait_for_completion will time out.
When the slave resumes, it can check if a Master-initiated
re-enumeration and initialization took place and skip the
wait_for_completion() if there is no reason to wait.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
While handling the Device0, we can safely use sdw_write_no_pm.
This move will also helps us track that all other usages of
sdw_write() happen when the Slave is already enumerated.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add support for pm_runtime with the appropriate error checks for
sdw_write/read functions, e.g. when pm_runtime is not supported.
Also expose internal functions without pm_runtime support, which are
required to perform any sort of suspend/resume operation, as well as
any enumeration tasks.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Waiting for the enumeration to be complete may not be enough for a
Slave driver, there is a possible race condition between resume
operations and initializations handled in an interrupt thread, which
can results in settings not being fully restored after system or
pm_runtime resume.
This patch builds on the changes added for enumeration_complete,
init_completion() is called when the Slave device becomes UNATTACHED,
as done with enumeration_complete.
The difference with the enumeration_complete case is that complete()
is signaled after the Slave device is fully initialized after the
.update_status() callback is called.
A Slave device driver can decide to wait on either of the two
complete() cases, depending on its initialization code and
requirements.
Signed-off-by: Rander Wang <rander.wang@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
This patch adds the signaling needed for Slave drivers to wait until
the enumeration completes so that race conditions when issuing
read/write commands are avoided. The calls for wait_for_completion()
will be added in codec drivers in follow-up patches.
The order between init_completion() and complete() is deterministic,
the Slave is marked as UNATTACHED either during a Master-initiated
HardReset, or when the hardware detects the Slave no longer reports as
ATTACHED.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The driver probe takes care of basic initialization and is invoked
when a Slave becomes attached, after a match between the Slave DevID
registers and ACPI/DT entries.
The update_status callback is invoked when a Slave state changes,
e.g. when it is assigned a non-zero Device Number and it reports with
an ATTACHED/ALERT state.
The state change detection is usually hardware-based and based on the
SoundWire frame rate (e.g. double-digit microseconds) while the probe
is a pure software operation, which may involve a kernel module
load. In corner cases, it's possible that the state changes before the
probe completes.
This patch suggests the use of wait_for_completion to avoid races on
startup, so that the update_status callback does not rely on invalid
pointers/data structures.
Signed-off-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200115000844.14695-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Propagate bus format/flags so that the previous bridge element in the
chain knows which input format the panel bridge expects.
v11:
* Fix a typo in the subject
* Update the commit message so it's readable by itself
v10:
* Add changelog to the commit message
v8 -> v9:
* No changes
v7:
* Set atomic state hooks explicitly
v4 -> v6:
* Not part of the series
v3:
* Adjust things to match the new bus-format negotiation approach
* Use drm_atomic_helper_bridge_propagate_bus_fmt
* Don't implement ->atomic_check() (the core now takes care of bus
flags propagation)
v2:
* Adjust things to match the new bus-format negotiation approach
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-11-boris.brezillon@collabora.com
The drm_bridge_funcs structure is never modified, make it const. Making
it read-only can improve security as the structure contains function
pointers.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200224230056.2157-1-laurent.pinchart+renesas@ideasonboard.com
Factor out the manual container_of() uses to a common to_lvds_codec()
macro to shorten lines (and provide better type safety, although that
won't matter much in this case).
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200224225645.28060-1-laurent.pinchart+renesas@ideasonboard.com
When there was no parallelism (no top-level -j arg and a pre-1.7
sphinx-build), the argument passed would be empty ("") instead of just
being missing, which would (understandably) badly confuse sphinx-build.
Fix this by removing the quotes.
Reported-by: Rafael J. Wysocki <rafael@kernel.org>
Fixes: 51e46c7a40 ("docs, parallelism: Rearrange how jobserver reservations are made")
Cc: stable@vger.kernel.org # v5.5 only
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
MPX was removed in commit 45fc24e89b ("x86/mpx: remove MPX from
arch/x86"), this removes the corresponding entry in the x86 toc.
This was suggested by a Sphinx warning.
Signed-off-by: Stephen Kitt <steve@sk2.org>
Fixes: 45fc24e89b ("x86/mpx: remove MPX from arch/x86")
Acked-by: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
