Chains of memory accesses are never good for performance.
The req->task->io_uring->in_cancel in io_req_local_work_add() is there
so that when a task is exiting via io_uring_try_cancel_requests() and
starts waiting for completions, it gets woken up by every new task_work
item queued.
Do a little trick by announcing waiting in io_uring_try_cancel_requests(),
making io_req_local_work_add() wake us up. We also need to check for
deferred tw items after prepare_to_wait(TASK_INTERRUPTIBLE);
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/fb11597e9bbcb365901824f8c5c2cf0d6ee100d0.1680782017.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Every task_work will try to wake the task to be executed, which causes
excessive scheduling and additional overhead. For some tw it's
justified, but others won't do much but post a single CQE.
When a task waits for multiple cqes, every such task_work will wake it
up. Instead, the task may give a hint about how many cqes it waits for,
io_req_local_work_add() will compare against it and skip wake ups
if #cqes + #tw is not enough to satisfy the waiting condition. Task_work
that uses the optimisation should be simple enough and never post more
than one CQE. It's also ignored for non DEFER_TASKRUN rings.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d2b77e99d1e86624d8a69f7037d764b739dcd225.1680782017.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Rather than check this in the fast path issue, it makes more sense to
just assign the copy of the data when we're setting it up anyway. This
makes the code a bit cleaner, and removes the need for this check in
the issue path.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We use ->rsrc_ref_lock spinlock to protect ->rsrc_ref_list in
io_rsrc_node_ref_zero(). Now we removed pcpu refcounting, which means
io_rsrc_node_ref_zero() is not executed from the irq context as an RCU
callback anymore, and we also put it under ->uring_lock.
io_rsrc_node_switch(), which queues up nodes into the list, is also
protected by ->uring_lock, so we can safely get rid of ->rsrc_ref_lock.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/6b60af883c263551190b526a55ff2c9d5ae07141.1680576071.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
One problem with the current rsrc infra is that often updates will
generates lots of rsrc nodes, each carry pcpu refs. That takes quite a
lot of memory, especially if there is a stall, and takes lots of CPU
cycles. Only pcpu allocations takes >50 of CPU with a naive benchmark
updating files in a loop.
Replace pcpu refs with normal refcounting. There is already a hot path
avoiding atomics / refs, but following patches will further improve it.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e9ed8a9457b331a26555ff9443afc64cdaab7247.1680576071.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We already do this manually for the !SQPOLL case, do it in general and
we can also dump the ugly min3() in io_submit_sqes().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It has nothing to do with the SQE at this point, it's a request
submission. While in there, get rid of the 'force_nonblock' argument
which is also dead, as we only pass in true.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Before cond_resched()'ing in handle_tw_list() we also drop the current
ring context, and so the next loop iteration will need to pick/pin a new
context and do trylock.
The chunk removed by this patch was intended to be an optimisation
covering exactly this case, i.e. retaking the lock after reschedule, but
in reality it's skipped for the first iteration after resched as
described and will keep hammering the lock if it's contended.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1ecec9483d58696e248d1bfd52cf62b04442df1d.1679931367.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since the move to PF_IO_WORKER, we don't juggle memory context manually
anymore. Remove that outdated part of the comment for __io_worker_idle().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since commit 0654b05e7e65 ("io_uring: One wqe per wq"), we have just a
single io_wqe instance embedded per io_wq. Drop the extra structure in
favor of accessing struct io_wq directly, cleaning up quite a bit of
dereferences and backpointers.
No functional changes intended. Tested with liburing's testsuite
and mmtests performance microbenchmarks. I didn't observe any
performance regressions.
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
Link: https://lore.kernel.org/r/20230322011628.23359-2-krisman@suse.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
On at least parisc, we have strict requirements on how we virtually map
an address that is shared between the application and the kernel. On
these platforms, IOU_PBUF_RING_MMAP should be used when setting up a
shared ring buffer for provided buffers. If the application is mapping
these pages and asking the kernel to pin+map them as well, then we have
no control over what virtual address we get in the kernel.
For that case, do a sanity check if SHM_COLOUR is defined, and disallow
the mapping request. The application must fall back to using
IOU_PBUF_RING_MMAP for this case, and liburing will do that transparently
with the set of helpers that it has.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add support for KASAN in the alloc_caches (apoll and netmsg_cache).
Thus, if something touches the unused caches, it will raise a KASAN
warning/exception.
It poisons the object when the object is put to the cache, and unpoisons
it when the object is gotten or freed.
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Link: https://lore.kernel.org/r/20230223164353.2839177-2-leitao@debian.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Having cache entries linked using the hlist format brings no benefit, and
also requires an unnecessary extra pointer address per cache entry.
Use the internal io_wq_work_node single-linked list for the internal
alloc caches (async_msghdr and async_poll)
This is required to be able to use KASAN on cache entries, since we do
not need to touch unused (and poisoned) cache entries when adding more
entries to the list.
Suggested-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://lore.kernel.org/r/20230223164353.2839177-2-leitao@debian.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Right now io_wq allocates one io_wqe per NUMA node. As io_wq is now
bound to a task, the task basically uses only the NUMA local io_wqe, and
almost never changes NUMA nodes, thus, the other wqes are mostly
unused.
Allocate just one io_wqe embedded into io_wq, and uses all possible cpus
(cpu_possible_mask) in the io_wqe->cpumask.
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://lore.kernel.org/r/20230310201107.4020580-1-leitao@debian.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The ring mapped provided buffer rings rely on the application allocating
the memory for the ring, and then the kernel will map it. This generally
works fine, but runs into issues on some architectures where we need
to be able to ensure that the kernel and application virtual address for
the ring play nicely together. This at least impacts architectures that
set SHM_COLOUR, but potentially also anyone setting SHMLBA.
To use this variant of ring provided buffers, the application need not
allocate any memory for the ring. Instead the kernel will do so, and
the allocation must subsequently call mmap(2) on the ring with the
offset set to:
IORING_OFF_PBUF_RING | (bgid << IORING_OFF_PBUF_SHIFT)
to get a virtual address for the buffer ring. Normally the application
would allocate a suitable piece of memory (and correctly aligned) and
simply pass that in via io_uring_buf_reg.ring_addr and the kernel would
map it.
Outside of the setup differences, the kernel allocate + user mapped
provided buffer ring works exactly the same.
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In preparation for allowing flags to be set for registration, rename
the padding and use it for that.
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Rather than rely on checking buffer_list->buf_pages or ->buf_nr_pages,
add a separate member that tracks if this is a ring mapped provided
buffer list or not.
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In preparation for allowing the kernel to allocate the provided buffer
rings and have the application mmap it instead, abstract out the
current method of pinning and mapping the user allocated ring.
No functional changes intended in this patch.
Acked-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Some architectures have memory cache aliasing requirements (e.g. parisc)
if memory is shared between userspace and kernel. This patch fixes the
kernel to return an aliased address when asked by userspace via mmap().
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_uring hashes writes to a given file/inode so that it can serialize
them. This is useful if the file system needs exclusive access to the
file to perform the write, as otherwise we end up with a ton of io-wq
threads trying to lock the inode at the same time. This can cause
excessive system time.
But if the file system has flagged that it supports parallel O_DIRECT
writes, then there's no need to serialize the writes. Check for that
through FMODE_DIO_PARALLEL_WRITE and don't hash it if we don't need to.
In a basic test of 8 threads writing to a file on XFS on a gen2 Optane,
with each thread writing in 4k chunks, it improves performance from
~1350K IOPS (or ~5290MiB/sec) to ~1410K IOPS (or ~5500MiB/sec).
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Some filesystems support multiple threads writing to the same file with
O_DIRECT without requiring exclusive access to it. io_uring can use this
hint to avoid serializing dio writes to this inode, instead allowing them
to run in parallel.
XFS and ext4 both fall into this category, so set the flag for both of
them.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull btrfs fixes from David Sterba:
- scan block devices in non-exclusive mode to avoid temporary mkfs
failures
- fix race between quota disable and quota assign ioctls
- fix deadlock when aborting transaction during relocation with scrub
- ignore fiemap path cache when there are multiple paths for a node
* tag 'for-6.3-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: ignore fiemap path cache when there are multiple paths for a node
btrfs: fix deadlock when aborting transaction during relocation with scrub
btrfs: scan device in non-exclusive mode
btrfs: fix race between quota disable and quota assign ioctls
This reverts commit a837e5161c, which broke probing of the venus
driver, at least on the SC7180 SoC HP X2 Chromebook:
qcom-venus aa00000.video-codec: Adding to iommu group 11
qcom-venus aa00000.video-codec: non legacy binding
qcom-venus aa00000.video-codec: failed to reset venus core
qcom-venus: probe of aa00000.video-codec failed with error -110
Matthias Kaehlcke also reported that the same change caused a regression
in SC7180 and sc7280, that prevents AOSS from entering sleep mode during
system suspend. So let's revert this commit for now to fix both issues.
Fixes: a837e5161c ("venus: firmware: Correct non-pix start and end addresses")
Reported-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull driver core fixes from Greg KH:
"Here are three small changes for 6.3-rc5 semi-related to driver core
stuff:
- documentation update where we move the security_bugs file to a more
relevant location.
- mdt/spi-nor debugfs memory leak fix that's been floating around for
a long time and acked by the maintainer
- cacheinfo bugfix for a regression in 6.3-rc1
All have been in linux-next with no reported problems"
* tag 'driver-core-6.3-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
cacheinfo: Fix LLC is not exported through sysfs
Documentation/security-bugs: move from admin-guide/ to process/
mtd: spi-nor: fix memory leak when using debugfs_lookup()
Pull powerpc fixes from Michael Ellerman:
- Fix a false positive warning in __pte_needs_flush() (with DEBUG_VM=y)
- Fix oops when a PF_IO_WORKER thread tries to core dump
- Don't try to reconfigure VAS when it's disabled
Thanks to Benjamin Gray, Haren Myneni, Jens Axboe, Nathan Lynch, and
Russell Currey.
* tag 'powerpc-6.3-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/pseries/vas: Ignore VAS update for DLPAR if copy/paste is not enabled
powerpc: Don't try to copy PPR for task with NULL pt_regs
powerpc/64s: Fix __pte_needs_flush() false positive warning
Pull cifs client fixes from Steve French:
"Four cifs/smb3 client (reconnect and DFS related) fixes, including two
for stable:
- DFS oops fix
- DFS reconnect recursion fix
- An SMB1 parallel reconnect fix
- Trivial dead code removal in smb2_reconnect"
* tag '6.3-rc4-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
cifs: get rid of dead check in smb2_reconnect()
cifs: prevent infinite recursion in CIFSGetDFSRefer()
cifs: avoid races in parallel reconnects in smb1
cifs: fix DFS traversal oops without CONFIG_CIFS_DFS_UPCALL
Pull input fixes from Dmitry Torokhov:
- fixes to ALPS and Focaltech PS/2 drivers dealing with the breakage of
switching to -funsigned-char
- quirks to i8042 to better handle Lifebook A574/H and TUXEDO devices
- a quirk to Goodix touchscreen driver to handle Yoga Book X90F
- a fix for incorrectly merged patch to xpad game controller driver
* tag 'input-for-v6.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: i8042 - add TUXEDO devices to i8042 quirk tables for partial fix
Input: alps - fix compatibility with -funsigned-char
Input: focaltech - use explicitly signed char type
Input: xpad - fix incorrectly applied patch for MAP_PROFILE_BUTTON
Input: goodix - add Lenovo Yoga Book X90F to nine_bytes_report DMI table
Input: i8042 - add quirk for Fujitsu Lifebook A574/H
Pull pin control fixes from Linus Walleij:
"Some pin control fixes for the v6.3 series.
The most notable and urgent one is probably the AMD fix which affects
AMD laptops, found by the Chromium people.
Summary:
- Fix up the Kconfig options for MediaTek MT7981
- Fix the irq domain name in the AT91-PIO4 driver
- Fix some alternative muxing modes in the Ocelot driver
- Allocate the GPIO numbers dynamically in the STM32 driver
- Disable and mask interrupts on resume in the AMD driver
- Fix a typo in the Qualcomm SM8550 pin control device tree bindings"
* tag 'pinctrl-v6.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
dt-bindings: pinctrl: qcom,sm8550-lpass-lpi: allow input-enabled and bias-bus-hold
pinctrl: amd: Disable and mask interrupts on resume
pinctrl: stm32: use dynamic allocation of GPIO base
pinctrl: ocelot: Fix alt mode for ocelot
pinctrl: at91-pio4: fix domain name assignment
pinctrl: mediatek: fix naming inconsistency
pinctrl: mediatek: add missing options to PINCTRL_MT7981
Pull Kbuild fixes from Masahiro Yamada:
- Fix linux-headers debian package
- Fix a merge_config.sh error due to a misspelled variable
- Fix modversion for 32-bit build machines
* tag 'kbuild-fixes-v6.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
modpost: Fix processing of CRCs on 32-bit build machines
scripts: merge_config: Fix typo in variable name.
kbuild: deb-pkg: set version for linux-headers paths
