Now ublk_abort_queue() is moved to ublk char device release handler,
meantime our request queue is "quiesced" because either ->canceling was
set from uring_cmd cancel function or all IOs are inflight and can't be
completed by ublk server, things becomes easy much:
- all uring_cmd are done, so we needn't to mark io as UBLK_IO_FLAG_ABORTED
for handling completion from uring_cmd
- ublk char device is closed, no one can hold IO request reference any more,
so we can simply complete this request or requeue it for ublk_nosrv_should_reissue_outstanding.
Reviewed-by: Uday Shankar <ushankar@purestorage.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250416035444.99569-8-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove __ublk_quiesce_dev() and open code for updating device state as
QUIESCED.
We needn't to drain inflight requests in __ublk_quiesce_dev() any more,
because all inflight requests are aborted in ublk char device release
handler.
Also we needn't to set ->canceling in __ublk_quiesce_dev() any more
because it is done unconditionally now in ublk_ch_release().
Reviewed-by: Uday Shankar <ushankar@purestorage.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250416035444.99569-7-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There are currently two ways in which ublk server exit is detected by
ublk_drv:
1. uring_cmd cancellation. If there are any outstanding uring_cmds which
have not been completed to the ublk server when it exits, io_uring
calls the uring_cmd callback with a special cancellation flag as the
issuing task is exiting.
2. I/O timeout. This is needed in addition to the above to handle the
"saturated queue" case, when all I/Os for a given queue are in the
ublk server, and therefore there are no outstanding uring_cmds to
cancel when the ublk server exits.
There are a couple of issues with this approach:
- It is complex and inelegant to have two methods to detect the same
condition
- The second method detects ublk server exit only after a long delay
(~30s, the default timeout assigned by the block layer). This delays
the nosrv behavior from kicking in and potential subsequent recovery
of the device.
The second issue is brought to light with the new test_generic_06 which
will be added in following patch. It fails before this fix:
selftests: ublk: test_generic_06.sh
dev id is 0
dd: error writing '/dev/ublkb0': Input/output error
1+0 records in
0+0 records out
0 bytes copied, 30.0611 s, 0.0 kB/s
DEAD
dd took 31 seconds to exit (>= 5s tolerance)!
generic_06 : [FAIL]
Fix this by instead detecting and handling ublk server exit in the
character file release callback. This has several advantages:
- This one place can handle both saturated and unsaturated queues. Thus,
it replaces both preexisting methods of detecting ublk server exit.
- It runs quickly on ublk server exit - there is no 30s delay.
- It starts the process of removing task references in ublk_drv. This is
needed if we want to relax restrictions in the driver like letting
only one thread serve each queue
There is also the disadvantage that the character file release callback
can also be triggered by intentional close of the file, which is a
significant behavior change. Preexisting ublk servers (libublksrv) are
dependent on the ability to open/close the file multiple times. To
address this, only transition to a nosrv state if the file is released
while the ublk device is live. This allows for programs to open/close
the file multiple times during setup. It is still a behavior change if a
ublk server decides to close/reopen the file while the device is LIVE
(i.e. while it is responsible for serving I/O), but that would be highly
unusual. This behavior is in line with what is done by FUSE, which is
very similar to ublk in that a userspace daemon is providing services
traditionally provided by the kernel.
With this change in, the new test (and all other selftests, and all
ublksrv tests) pass:
selftests: ublk: test_generic_06.sh
dev id is 0
dd: error writing '/dev/ublkb0': Input/output error
1+0 records in
0+0 records out
0 bytes copied, 0.0376731 s, 0.0 kB/s
DEAD
generic_04 : [PASS]
Signed-off-by: Uday Shankar <ushankar@purestorage.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250416035444.99569-6-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
ublk_ch_release() is called after ublk char device is closed, when all
uring_cmd are done, so it is perfect fine to move ublk device reset to
ublk_ch_release() from ublk_ctrl_start_recovery().
This way can avoid to grab the exiting daemon task_struct too long.
However, reset of the following ublk IO flags has to be moved until ublk
io_uring queues are ready:
- ubq->canceling
For requeuing IO in case of ublk_nosrv_dev_should_queue_io() before device
is recovered
- ubq->fail_io
For failing IO in case of UBLK_F_USER_RECOVERY_FAIL_IO before device is
recovered
- ublk_io->flags
For preventing using io->cmd
With this way, recovery is simplified a lot.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250416035444.99569-5-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now ublk deals with ublk_nosrv_dev_should_queue_io() by keeping request
queue as quiesced. This way is fragile because queue quiesce crosses syscalls
or process contexts.
Switch to rely on ubq->canceling for dealing with
ublk_nosrv_dev_should_queue_io(), because it has been used for this purpose
during io_uring context exiting, and it can be reused before recovering too.
In ublk_queue_rq(), the request will be added to requeue list without
kicking off requeue in case of ubq->canceling, and finally requests added in
requeue list will be dispatched from either ublk_stop_dev() or
ublk_ctrl_end_recovery().
Meantime we have to move reset of ubq->canceling from ublk_ctrl_start_recovery()
to ublk_ctrl_end_recovery(), when IO handling can be recovered completely.
Then blk_mq_quiesce_queue() and blk_mq_unquiesce_queue() are always used
in same context.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Uday Shankar <ushankar@purestorage.com>
Link: https://lore.kernel.org/r/20250416035444.99569-4-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Most uring_cmds issued against ublk character devices are serialized
because each command affects only one queue, and there is an early check
which only allows a single task (the queue's ubq_daemon) to issue
uring_cmds against that queue. However, this mechanism does not work for
FETCH_REQs, since they are expected before ubq_daemon is set. Since
FETCH_REQs are only used at initialization and not in the fast path,
serialize them using the per-ublk-device mutex. This fixes a number of
data races that were previously possible if a badly behaved ublk server
decided to issue multiple FETCH_REQs against the same qid/tag
concurrently.
Reported-by: Caleb Sander Mateos <csander@purestorage.com>
Signed-off-by: Uday Shankar <ushankar@purestorage.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250416035444.99569-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
test may exit early because of missing program or not having required
feature before calling _prep_test(), then $UBLK_TMP isn't cleaned.
Fix it by moving creating $UBLK_TMP into _prep_test(), any resources
created since _prep_test() will be cleaned by _cleanup_test().
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250412023035.2649275-14-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In NUMA machine, ublk IO performance is very sensitive with queue
pthread's affinity setting.
Retrieve queue's affinity and select the 1st cpu as queue thread's sched
affinity, and it is observed that single cpu task affinity can get
stable & good performance if client application is put on proper cpu.
Dump this info when adding one ublk device. Use shmem to communicate
queue's tid between parent and daemon.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250412023035.2649275-9-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Detach ublk daemon from the starting process completely by double-fork and
clearing its process group, so that `_add_ublk_dev` can return from sub-shell.
Then it is more friendly for writing shell test script for adding/recovering
ublk device.
Prepare for running ublk test in parallel.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250412023035.2649275-5-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add io_uring UAPI header so that ublk can work with latest uapi
definition.
Fix the following build failure:
stripe.c: In function ‘stripe_to_uring_op’:
stripe.c:120:29: error: ‘IORING_OP_READV_FIXED’ undeclared (first use in this function); did you mean ‘IORING_OP_READ_FIXED’?
120 | return zc ? IORING_OP_READV_FIXED : IORING_OP_READV;
| ^~~~~~~~~~~~~~~~~~~~~
| IORING_OP_READ_FIXED
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Fixes: 57ed58c132 ("selftests: ublk: enable zero copy for stripe target")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250412023035.2649275-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Placing multiple protection information buffers inside the same page
can lead to oopses because set_page_dirty_lock() can't be called from
interrupt context.
Since a protection information buffer is not backed by a file there is
no point in setting its page dirty, there is nothing to synchronize.
Drop the call to set_page_dirty_lock() and remove the last argument to
bio_integrity_unpin_bvec().
Cc: stable@vger.kernel.org
Fixes: 492c5d4559 ("block: bio-integrity: directly map user buffers")
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/yq1v7r3ev9g.fsf@ca-mkp.ca.oracle.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
During recovery/check operations, the process_checks function loops
through available disks to find a 'primary' source with successfully
read data.
If no suitable source disk is found after checking all possibilities,
the 'primary' index will reach conf->raid_disks * 2. Add an explicit
check for this condition after the loop. If no source disk was found,
print an error message and return early to prevent further processing
without a valid primary source.
Link: https://lore.kernel.org/linux-raid/20250408143808.1026534-1-meir.elisha@volumez.com
Signed-off-by: Meir Elisha <meir.elisha@volumez.com>
Suggested-and-reviewed-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
vfs_iter_{read,write} always perform direct I/O when the file has the
O_DIRECT flag set, which breaks disabling direct I/O using the
LOOP_SET_STATUS / LOOP_SET_STATUS64 ioctls.
This was recenly reported as a regression, but as far as I can tell
was only uncovered by better checking for block sizes and has been
around since the direct I/O support was added.
Fix this by using the existing aio code that calls the raw read/write
iter methods instead. Note that despite the comments there is no need
for block drivers to ever call flush_dcache_page themselves, and the
call is a left-over from prehistoric times.
Fixes: ab1cb278bc ("block: loop: introduce ioctl command of LOOP_SET_DIRECT_IO")
Reported-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Darrick J. Wong <djwong@kernel.org>
Link: https://lore.kernel.org/r/20250409130940.3685677-1-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The original commit message and the wording "uncork" in the code comment
indicate that it is expected that the suppressed event instances are
automatically sent after unsuppressing.
This is not the case, instead they are discarded.
In effect this means that no "changed" events are emitted on the device
itself by default.
While each discovered partition does trigger a changed event on the
device, devices without partitions don't have any event emitted.
This makes udev miss the device creation and prompted workarounds in
userspace. See the linked util-linux/losetup bug.
Explicitly emit the events and drop the confusingly worded comments.
Link: https://github.com/util-linux/util-linux/issues/2434
Fixes: 498ef5c777 ("loop: suppress uevents while reconfiguring the device")
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
Link: https://lore.kernel.org/r/20250415-loop-uevent-changed-v2-1-0c4e6a923b2a@linutronix.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When registering a queue fails after blk_mq_sysfs_register() is
successful but the function later encounters an error, we need
to clean up the blk_mq_sysfs resources.
Add the missing blk_mq_sysfs_unregister() call in the error path
to properly clean up these resources and prevent a memory leak.
Fixes: 320ae51fee ("blk-mq: new multi-queue block IO queueing mechanism")
Signed-off-by: Zheng Qixing <zhengqixing@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Yu Kuai <yukuai3@huawei.com>
Link: https://lore.kernel.org/r/20250412092554.475218-1-zhengqixing@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
blk_mq_alloc_disk() already zero-initializes the destination buffer,
making strscpy() sufficient for safely copying the disk's name. The
additional NUL-padding performed by strscpy_pad() is unnecessary.
If the destination buffer has a fixed length, strscpy() automatically
determines its size using sizeof() when the argument is omitted. This
makes the explicit size argument unnecessary.
The source string is also NUL-terminated and meets the __must_be_cstr()
requirement of strscpy().
No functional changes intended.
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Reviewed-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20250410154727.883207-1-thorsten.blum@linux.dev
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull NVMe updates from Christoph:
"nvme updates for Linux 6.15
- nvmet fc/fcloop refcounting fixes (Daniel Wagner)
- fix missed namespace/ANA scans (Hannes Reinecke)
- fix a use after free in the new TCP netns support (Kuniyuki Iwashima)
- fix a NULL instead of false review in multipath (Uday Shankar)"
* tag 'nvme-6.15-2025-04-10' of git://git.infradead.org/nvme:
nvmet-fc: put ref when assoc->del_work is already scheduled
nvmet-fc: take tgtport reference only once
nvmet-fc: update tgtport ref per assoc
nvmet-fc: inline nvmet_fc_free_hostport
nvmet-fc: inline nvmet_fc_delete_assoc
nvmet-fcloop: add ref counting to lport
nvmet-fcloop: replace kref with refcount
nvmet-fcloop: swap list_add_tail arguments
nvme-tcp: fix use-after-free of netns by kernel TCP socket.
nvme: multipath: fix return value of nvme_available_path
nvme: re-read ANA log page after ns scan completes
nvme: requeue namespace scan on missed AENs
Commit 8284066946 ("ublk: grab request reference when the request is handled
by userspace") doesn't grab request reference in case of recovery reissue.
Then the request can be requeued & re-dispatch & failed when canceling
uring command.
If it is one zc request, the request can be freed before io_uring
returns the zc buffer back, then cause kernel panic:
[ 126.773061] BUG: kernel NULL pointer dereference, address: 00000000000000c8
[ 126.773657] #PF: supervisor read access in kernel mode
[ 126.774052] #PF: error_code(0x0000) - not-present page
[ 126.774455] PGD 0 P4D 0
[ 126.774698] Oops: Oops: 0000 [#1] SMP NOPTI
[ 126.775034] CPU: 13 UID: 0 PID: 1612 Comm: kworker/u64:55 Not tainted 6.14.0_blk+ #182 PREEMPT(full)
[ 126.775676] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-1.fc39 04/01/2014
[ 126.776275] Workqueue: iou_exit io_ring_exit_work
[ 126.776651] RIP: 0010:ublk_io_release+0x14/0x130 [ublk_drv]
Fixes it by always grabbing request reference for aborting the request.
Reported-by: Caleb Sander Mateos <csander@purestorage.com>
Closes: https://lore.kernel.org/linux-block/CADUfDZodKfOGUeWrnAxcZiLT+puaZX8jDHoj_sfHZCOZwhzz6A@mail.gmail.com/
Fixes: 8284066946 ("ublk: grab request reference when the request is handled by userspace")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250409011444.2142010-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Do not leak the tgtport reference when the work is already scheduled.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The reference counting code can be simplified. Instead taking a tgtport
refrerence at the beginning of nvmet_fc_alloc_hostport and put it back
if not a new hostport object is allocated, only take it when a new
hostport object is allocated.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We need to take for each unique association a reference.
nvmet_fc_alloc_hostport for each newly created association.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
No need for this tiny helper with only one user, let's inline it.
And since the hostport ref counter needs to stay in sync, it's not
optional anymore to give back the reference.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
No need for this tiny helper with only one user, just inline it.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The fcloop_lport objects live time is controlled by the user interface
add_local_port and del_local_port. nport, rport and tport objects are
pointing to the lport objects but here is no clear tracking. Let's
introduce an explicit ref counter for the lport objects and prepare the
stage for restructuring how lports are used.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The kref wrapper is not really adding any value ontop of refcount. Thus
replace the kref API with the refcount API.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The newly element to be added to the list is the first argument of
list_add_tail. This fix is missing dcfad4ab4d ("nvmet-fcloop: swap
the list_add_tail arguments").
Fixes: 437c0b824d ("nvme-fcloop: add target to host LS request support")
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Commit 1be52169c3 ("nvme-tcp: fix selinux denied when calling
sock_sendmsg") converted sock_create() in nvme_tcp_alloc_queue()
to sock_create_kern().
sock_create_kern() creates a kernel socket, which does not hold
a reference to netns. If the code does not manage the netns
lifetime properly, use-after-free could happen.
Also, TCP kernel socket with sk_net_refcnt 0 has a socket leak
problem: it remains FIN_WAIT_1 if it misses FIN after close()
because tcp_close() stops all timers.
To fix such problems, let's hold netns ref by sk_net_refcnt_upgrade().
We had the same issue in CIFS, SMC, etc, and applied the same
solution, see commit ef7134c7fc ("smb: client: Fix use-after-free
of network namespace.") and commit 9744d2bf19 ("smc: Fix
use-after-free in tcp_write_timer_handler().").
Fixes: 1be52169c3 ("nvme-tcp: fix selinux denied when calling sock_sendmsg")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The function returns bool so we should return false, not NULL. No
functional changes are expected.
Signed-off-by: Uday Shankar <ushankar@purestorage.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
When scanning for new namespaces we might have missed an ANA AEN.
The NVMe base spec (NVMe Base Specification v2.1, Figure 151 'Asynchonous
Event Information - Notice': Asymmetric Namespace Access Change) states:
A controller shall not send this even if an Attached Namespace
Attribute Changed asynchronous event [...] is sent for the same event.
so we need to re-read the ANA log page after we rescanned the namespace
list to update the ANA states of the new namespaces.
Signed-off-by: Hannes Reinecke <hare@kernel.org>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Scanning for namespaces can take some time, so if the target is
reconfigured while the scan is running we may miss a Attached Namespace
Attribute Changed AEN.
Check if the NVME_AER_NOTICE_NS_CHANGED bit is set once the scan has
finished, and requeue scanning to pick up any missed change.
Signed-off-by: Hannes Reinecke <hare@kernel.org>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The bitmap_get_stats() function incorrectly returns -ENOENT for external
bitmaps.
Remove the external bitmap check as the statistics should be available
regardless of bitmap storage location.
Return -EINVAL only for invalid bitmap with no storage (neither in
superblock nor in external file).
Note: "bitmap_info.external" here refers to a bitmap stored in a separate
file (bitmap_file), not to external metadata.
Fixes: 8d28d0ddb9 ("md/md-bitmap: Synchronize bitmap_get_stats() with bitmap lifetime")
Signed-off-by: Zheng Qixing <zhengqixing@huawei.com>
Link: https://lore.kernel.org/linux-raid/20250403015322.2873369-1-zhengqixing@huaweicloud.com
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
