Jens Axboe [Thu, 16 May 2019 14:07:57 +0000 (08:07 -0600)]
Merge branch 'nvme-5.2' of git://git.infradead.org/nvme into for-5.2/block-post
Pull NVMe fixes from Christoph
* 'nvme-5.2' of git://git.infradead.org/nvme:
nvme: validate cntlid during controller initialisation
nvme: change locking for the per-subsystem controller list
nvme: trace all async notice events
nvme: fix typos in nvme status code values
nvme-fabrics: remove unused argument
nvme-multipath: avoid crash on invalid subsystem cntlid enumeration
nvme-fc: use separate work queue to avoid warning
nvme-rdma: remove redundant reference between ib_device and tagset
nvme-pci: mark expected switch fall-through
nvme-pci: add known admin effects to augument admin effects log page
nvme-pci: init shadow doorbell after each reset
nvme: validate cntlid during controller initialisation
The CNTLID value is required to be unique, and we do rely on this
for correct operation. So reject any controller for which a non-unique
CNTLID has been detected.
Based on a patch from Hannes Reinecke.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com>
nvme: change locking for the per-subsystem controller list
Life becomes a lot simpler if we just use the global
nvme_subsystems_lock to protect this list. Given that it is only
accessed during controller probing and removal that isn't a scalability
problem either.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com>
This patch removes the tracing of the NVMe Async events out of the
switch so that it can trace all the events including the ones which
are not handled in the nvme_handle_aen_notice(). The events which
are not handled in the nvme_handle_aen_notice() such as
NVME_AER_NOTICE_DISC_CHANGED corresponding event identifier needs
to be added in the drivers/nvme/host/trace.h so that it can stringify
the AER .
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Hannes Reinecke [Fri, 3 May 2019 13:37:35 +0000 (15:37 +0200)]
nvme-multipath: avoid crash on invalid subsystem cntlid enumeration
A process holding an open reference to a removed disk prevents it
from completing deletion, so its name continues to exist. A subsequent
gendisk creation may have the same cntlid which risks collision when
using that for the name. Use the unique ctrl->instance instead.
Signed-off-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Keith Busch <keith.busch@intel.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Hannes Reinecke [Fri, 3 May 2019 09:43:52 +0000 (11:43 +0200)]
nvme-fc: use separate work queue to avoid warning
When tearing down a controller the following warning is issued:
WARNING: CPU: 0 PID: 30681 at ../kernel/workqueue.c:2418 check_flush_dependency
This happens as the err_work workqueue item is scheduled on the
system workqueue (which has WQ_MEM_RECLAIM not set), but is flushed
from a workqueue which has WQ_MEM_RECLAIM set.
Fix this by providing an FC-NVMe specific workqueue.
Fixes: 4cff280a5fcc ("nvme-fc: resolve io failures during connect") Signed-off-by: Hannes Reinecke <hare@suse.com> Reviewed-by: James Smart <james.smart@broadcom.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Max Gurtovoy [Mon, 6 May 2019 10:47:55 +0000 (13:47 +0300)]
nvme-rdma: remove redundant reference between ib_device and tagset
In the past, before adding f41725bb ("nvme-rdma: Use mr pool") commit,
we needed a reference on the ib_device as long as the tagset
was alive, as the MRs in the request structures needed a valid ib_device.
Now, we allocate/deallocate MR pool per QP and consume on demand.
Also remove nvme_rdma_free_tagset function and use blk_mq_free_tag_set
instead, as it unneeded anymore.
This commit also fixes a memory leakage and possible segmentation fault.
When configuring the system with NIC teaming (aka bonding), we use 1
network interface to create an HA connection to the target side. In case
one connection breaks down, nvme-rdma driver will get notification from
rdma-cm layer that underlying address was change and will start error
recovery process. During this process, we'll reconnect to the target
via the second interface in the bond without destroying the tagset.
This will cause a leakage of the initial rdma device (ndev) and miscount
in the reference count of the new created rdma device (new ndev). In
the final destruction (or in another error flow), we'll get a warning
dump from the ib_dealloc_pd that we still have inflight MR's related to
that pd. This happens becasue of the miscount of the reference tag of
the rdma device and causing access violation to it's elements (some
queues are not destroyed yet).
Signed-off-by: Max Gurtovoy <maxg@mellanox.com> Signed-off-by: Israel Rukshin <israelr@mellanox.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.
This patch fixes the following warning:
drivers/nvme/host/pci.c: In function ‘nvme_timeout’:
drivers/nvme/host/pci.c:1298:12: warning: this statement may fall through [-Wimplicit-fallthrough=]
shutdown = true;
~~~~~~~~~^~~~~~
drivers/nvme/host/pci.c:1299:2: note: here
case NVME_CTRL_CONNECTING:
^~~~
Warning level 3 was used: -Wimplicit-fallthrough=3
This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com> Reviewed-by: Keith Busch <keith.busch@intel.com> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Maxim Levitsky [Thu, 2 May 2019 11:31:34 +0000 (14:31 +0300)]
nvme-pci: add known admin effects to augument admin effects log page
Add known admin effects even if hardware has known admin effects page,
since hardware can't be ever trusted to report sane values.
(on my Intel DC P3700, it reports no side effects for namespace format)
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
ata_host_alloc() can only fail due to memory allocation failures.
Hence there is no need to print a message, as the memory allocation core
code already takes care of that.
Reviewed-by: Simon Horman <horms+renesas@verge.net.au> Reviewed-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:11 +0000 (20:38 +0200)]
lightnvm: pblk: use nvm_rq_to_ppa_list()
This patch replaces few remaining usages of rqd->ppa_list[] with
existing nvm_rq_to_ppa_list() helpers. This is needed for theoretical
devices with ws_min/ws_opt equal to 1.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:10 +0000 (20:38 +0200)]
lightnvm: pblk: simplify partial read path
This patch changes the approach to handling partial read path.
In old approach merging of data from round buffer and drive was fully
made by drive. This had some disadvantages - code was complex and
relies on bio internals, so it was hard to maintain and was strongly
dependent on bio changes.
In new approach most of the handling is done mostly by block layer
functions such as bio_split(), bio_chain() and generic_make request()
and generally is less complex and easier to maintain. Below some more
details of the new approach.
When read bio arrives, it is cloned for pblk internal purposes. All
the L2P mapping, which includes copying data from round buffer to bio
and thus bio_advance() calls is done on the cloned bio, so the original
bio is untouched. If we found that we have partial read case, we
still have original bio untouched, so we can split it and continue to
process only first part of it in current context, when the rest will be
called as separate bio request which is passed to generic_make_request()
for further processing.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Heiner Litz <hlitz@ucsc.edu> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:09 +0000 (20:38 +0200)]
lightnvm: do not remove instance under global lock
Currently all the target instances are removed under global nvm_lock.
This was needed to ensure that nvm_dev struct will not be freed by
hot unplug event during target removal. However, current implementation
has some drawbacks, since the same lock is used when new nvme subsystem
is registered, so we can have a situation, that due to long process of
target removal on drive A, registration (and listing in OS) of the
drive B will take a lot of time, since it will wait for that lock.
Now when we have kref which ensures that nvm_dev will not be freed in
the meantime, we can easily get rid of this lock for a time when we are
removing nvm targets.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:08 +0000 (20:38 +0200)]
lightnvm: track inflight target creations
When creation process is still in progress, target is not yet on
targets list. This causes a chance for removing whole lightnvm
subsystem by calling nvm_unregister() in the meantime and finally by
causing kernel panic inside target init function.
This patch changes the behaviour by adding kref variable which tracks
all the users of nvm_dev structure. When nvm_dev is allocated, kref
value is set to 1. Then before every target creation the value is
increased and decreased after target removal. The extra reference
is decreased when nvm subsystem is unregistered.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:06 +0000 (20:38 +0200)]
lightnvm: pblk: IO path reorganization
This patch is made in order to prepare read path for new approach to
partial read handling, which is simpler in compare with previous one.
The most important change is to move the handling of completed and
failed bio from the pblk_make_rq() to particular read and write
functions. This is needed, since after partial read path changes,
sometimes completed/failed bio will be different from original one, so
we cannot do this any longer in pblk_make_rq().
Other changes are small read path refactor in order to reduce the size
of the following patch with partial read changes.
Generally the goal of this patch is not to change the functionality,
but just to prepare the code for the following changes.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:05 +0000 (20:38 +0200)]
lightnvm: pblk: GC error handling
Currently when there is an IO error (or similar) on GC read path, pblk
still move the line, which was currently under GC process to free state.
Such a behaviour can lead to silent data mismatch issue.
With this patch, the line which was under GC process on which some IO
errors occurred, will be putted back to closed state (instead of free
state as it was without this patch) and the L2P mapping for such a
failed sectors will not be updated.
Then in case of any user IOs to such a failed sectors, pblk would be
able to return at least real IO error instead of stale data as it is
right now.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:04 +0000 (20:38 +0200)]
lightnvm: pblk: remove internal IO timeout
Currently during pblk padding, there is internal IO timeout introduced,
which is smaller than default NVMe timeout. This can lead to various
use-after-free issues. Since in case of any IO timeouts NVMe and block
layer will handle timeout by themselves and report it back to use,
there is no need to keep this internal timeout in pblk.
Igor Konopko [Sat, 4 May 2019 18:38:03 +0000 (20:38 +0200)]
lightnvm: pblk: wait for inflight IOs in recovery
This patch changes the behaviour of recovery padding in order to
support a case, when some IOs were already submitted to the drive and
some next one are not submitted due to error returned.
Currently in case of errors we simply exit the pad function without
waiting for inflight IOs, which leads to panic on inflight IOs
completion.
After the changes we always wait for all the inflight IOs before
exiting the function.
Igor Konopko [Sat, 4 May 2019 18:38:02 +0000 (20:38 +0200)]
lightnvm: pblk: propagate errors when reading meta
Read errors are not correctly propagated. Errors are cleared before
returning control to the io submitter. Change the behaviour such that
all read errors exept high ecc read warning status is returned
appropriately.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:38:01 +0000 (20:38 +0200)]
lightnvm: pblk: fix update line wp in OOB recovery
In case of OOB recovery, we can hit the scenario when all the data in
line were written and some part of emeta was written too. In such
a case pblk_update_line_wp() function will call pblk_alloc_page()
function which will case to set left_msecs to value below zero
(since this field does not track emeta region) and thus will lead to
multiple kernel warnings. This patch fixes that issue.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
When we trigger nvm target remove during device hot unplug, there is
a probability to hit a general protection fault. This is caused by use
of nvm_dev thay may be freed from another (hot unplug) thread
(in the nvm_unregister function).
Introduce lock in nvme_ioctl_dev_remove function to prevent this
situation.
Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
lightnvm: pblk: set propper line as data_line after gc
In current implementation of l2p recovery, when we are after gc and we
have open line, we are not setting current data line properly (we set
last line from the device instead of last line ordered by seq_nr) and
in consequence, kernel panic and data corruption.
Signed-off-by: Marcin Dziegielewski <marcin.dziegielewski@intel.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Chansol Kim [Sat, 4 May 2019 18:37:56 +0000 (20:37 +0200)]
lightnvm: pblk: fix bio leak when bio is split
For large size io where blk_queue_split needs to be called inside
pblk_rw_io, results in bio leak as bio_endio is not called on the
newly allocated. One way to observe this is to mounting ext4
filesystem on the target and issuing 1MB io with dd, e.g., dd bs=1MB
if=/dev/null of=/mount/myvolume. kmemleak reports:
In case there is a need for a split, blk_queue_split returns the newly
allocated bio to the caller by changing the value of pointer passed as
a reference, while the original is passed to generic_make_requests.
Although pblk_rw_io's local variable bio* has changed and passed to
pblk_submit_read and pblk_write_to_cache, work is done on this new
bio*, and pblk_rw_io returns NVM_IO_DONE, pblk_make_rq calls bio_endio
on the old bio* because it passed bio pointer by value to pblk_rw_io.
pblk_rw_io is unfolded into pblk_make_rq so that there is no copying
of bio* and bio_endio is called on the correct bio*.
Signed-off-by: Chansol Kim <chansol.kim@samsung.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:37:55 +0000 (20:37 +0200)]
lightnvm: Inherit mdts from the parent nvme device
Current lightnvm and pblk implementation does not care about NVMe max
data transfer size, which can be smaller than 64*K=256K. There are
existing NVMe controllers which NVMe max data transfer size is lower
that 256K (for example 128K, which happens for existing NVMe
controllers which are NVMe spec compliant). Such a controllers are not
able to handle command which contains 64 PPAs, since the the size of
DMAed buffer will be above the capabilities of such a controller.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:37:54 +0000 (20:37 +0200)]
lightnvm: pblk: set proper read status in bio
Currently in case of read errors, bi_status is not set properly which
leads to returning inproper data to layers above. This patch fix that
by setting proper status in case of read errors.
Also remove unnecessary warn_once(), which does not make sense
in that place, since user bio is not used for interation with drive
and thus bi_status will not be set here.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:37:53 +0000 (20:37 +0200)]
lightnvm: pblk: cleanly fail when there is not enough memory
L2P table can be huge in many cases, since it typically requires 1GB
of DRAM for 1TB of drive. When there is not enough memory available,
OOM killer turns on and kills random processes, which can be very
annoying for users.
This patch changes the flag for L2P table allocation on order to handle
this situation in more user friendly way.
GFP_KERNEL and __GPF_HIGHMEM are default flags used in parameterless
vmalloc() calls, so they are also keeped in that patch. Additionally
__GFP_NOWARN flag is added in order to hide very long dmesg warn in
case of the allocation failures. The most important flag introduced
in that patch is __GFP_RETRY_MAYFAIL, which would cause allocator
to try use free memory and if not available to drop caches, but not
to run OOM killer.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:37:50 +0000 (20:37 +0200)]
lightnvm: pblk: gracefully handle GC vmalloc fail
Currently when we fail on rq data allocation in gc, it skips moving
active data and moves line straigt to its free state. Losing user
data in the process.
Move the data allocation to an earlier phase of GC, where we can still
fail gracefully by moving line back to the closed state.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Javier González <javier@javigon.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Igor Konopko [Sat, 4 May 2019 18:37:48 +0000 (20:37 +0200)]
lightnvm: pblk: reduce L2P memory footprint
Currently L2P map size is calculated based on the total number of
available sectors, which is redundant, since it contains mapping for
overprovisioning as well (11% by default).
Change this size to the real capacity and thus reduce the memory
footprint significantly - with default op value it is approx.
110MB of DRAM less for every 1TB of media.
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com> Reviewed-by: Hans Holmberg <hans.holmberg@cnexlabs.com> Reviewed-by: Javier González <javier@javigon.com> Signed-off-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:29 +0000 (09:52 +0800)]
block: don't drain in-progress dispatch in blk_cleanup_queue()
Now freeing hw queue resource is moved to hctx's release handler,
we don't need to worry about the race between blk_cleanup_queue and
run queue any more.
So don't drain in-progress dispatch in blk_cleanup_queue().
This is basically revert of c2856ae2f315 ("blk-mq: quiesce queue before
freeing queue").
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reviewed-by: Bart Van Assche <bvanassche@acm.org> Reviewed-by: Hannes Reinecke <hare@suse.com> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:28 +0000 (09:52 +0800)]
blk-mq: move cancel of hctx->run_work into blk_mq_hw_sysfs_release
hctx is always released after requeue is freed.
With holding queue's kobject refcount, it is safe for driver to run queue,
so one run queue might be scheduled after blk_sync_queue() is done.
So moving the cancel of hctx->run_work into blk_mq_hw_sysfs_release()
for avoiding run released queue.
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reviewed-by: Bart Van Assche <bvanassche@acm.org> Reviewed-by: Hannes Reinecke <hare@suse.com> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:27 +0000 (09:52 +0800)]
blk-mq: always free hctx after request queue is freed
In normal queue cleanup path, hctx is released after request queue
is freed, see blk_mq_release().
However, in __blk_mq_update_nr_hw_queues(), hctx may be freed because
of hw queues shrinking. This way is easy to cause use-after-free,
because: one implicit rule is that it is safe to call almost all block
layer APIs if the request queue is alive; and one hctx may be retrieved
by one API, then the hctx can be freed by blk_mq_update_nr_hw_queues();
finally use-after-free is triggered.
Fixes this issue by always freeing hctx after releasing request queue.
If some hctxs are removed in blk_mq_update_nr_hw_queues(), introduce
a per-queue list to hold them, then try to resuse these hctxs if numa
node is matched.
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reviewed-by: Hannes Reinecke <hare@suse.com> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:26 +0000 (09:52 +0800)]
blk-mq: split blk_mq_alloc_and_init_hctx into two parts
Split blk_mq_alloc_and_init_hctx into two parts, and one is
blk_mq_alloc_hctx() for allocating all hctx resources, another
is blk_mq_init_hctx() for initializing hctx, which serves as
counter-part of blk_mq_exit_hctx().
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org Cc: Martin K . Petersen <martin.petersen@oracle.com> Cc: Christoph Hellwig <hch@lst.de> Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:25 +0000 (09:52 +0800)]
blk-mq: free hw queue's resource in hctx's release handler
Once blk_cleanup_queue() returns, tags shouldn't be used any more,
because blk_mq_free_tag_set() may be called. Commit 45a9c9d909b2
("blk-mq: Fix a use-after-free") fixes this issue exactly.
However, that commit introduces another issue. Before 45a9c9d909b2,
we are allowed to run queue during cleaning up queue if the queue's
kobj refcount is held. After that commit, queue can't be run during
queue cleaning up, otherwise oops can be triggered easily because
some fields of hctx are freed by blk_mq_free_queue() in blk_cleanup_queue().
We have invented ways for addressing this kind of issue before, such as:
8dc765d438f1 ("SCSI: fix queue cleanup race before queue initialization is done") c2856ae2f315 ("blk-mq: quiesce queue before freeing queue")
But still can't cover all cases, recently James reports another such
kind of issue:
This issue can be quite hard to address by previous way, given
scsi_run_queue() may run requeues for other LUNs.
Fixes the above issue by freeing hctx's resources in its release handler, and this
way is safe becasue tags isn't needed for freeing such hctx resource.
This approach follows typical design pattern wrt. kobject's release handler.
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reported-by: James Smart <james.smart@broadcom.com> Fixes: 45a9c9d909b2 ("blk-mq: Fix a use-after-free") Cc: stable@vger.kernel.org Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:24 +0000 (09:52 +0800)]
blk-mq: move cancel of requeue_work into blk_mq_release
With holding queue's kobject refcount, it is safe for driver
to schedule requeue. However, blk_mq_kick_requeue_list() may
be called after blk_sync_queue() is done because of concurrent
requeue activities, then requeue work may not be completed when
freeing queue, and kernel oops is triggered.
So moving the cancel of requeue_work into blk_mq_release() for
avoiding race between requeue and freeing queue.
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: Bart Van Assche <bart.vanassche@wdc.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reviewed-by: Bart Van Assche <bvanassche@acm.org> Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Ming Lei [Tue, 30 Apr 2019 01:52:23 +0000 (09:52 +0800)]
blk-mq: grab .q_usage_counter when queuing request from plug code path
Just like aio/io_uring, we need to grab 2 refcount for queuing one
request, one is for submission, another is for completion.
If the request isn't queued from plug code path, the refcount grabbed
in generic_make_request() serves for submission. In theroy, this
refcount should have been released after the sumission(async run queue)
is done. blk_freeze_queue() works with blk_sync_queue() together
for avoiding race between cleanup queue and IO submission, given async
run queue activities are canceled because hctx->run_work is scheduled with
the refcount held, so it is fine to not hold the refcount when
running the run queue work function for dispatch IO.
However, if request is staggered into plug list, and finally queued
from plug code path, the refcount in submission side is actually missed.
And we may start to run queue after queue is removed because the queue's
kobject refcount isn't guaranteed to be grabbed in flushing plug list
context, then kernel oops is triggered, see the following race:
blk_mq_flush_plug_list():
blk_mq_sched_insert_requests()
insert requests to sw queue or scheduler queue
blk_mq_run_hw_queue
Because of concurrent run queue, all requests inserted above may be
completed before calling the above blk_mq_run_hw_queue. Then queue can
be freed during the above blk_mq_run_hw_queue().
Fixes the issue by grab .q_usage_counter before calling
blk_mq_sched_insert_requests() in blk_mq_flush_plug_list(). This way is
safe because the queue is absolutely alive before inserting request.
Cc: Dongli Zhang <dongli.zhang@oracle.com> Cc: James Smart <james.smart@broadcom.com> Cc: linux-scsi@vger.kernel.org, Cc: Martin K . Petersen <martin.petersen@oracle.com>, Cc: Christoph Hellwig <hch@lst.de>, Cc: James E . J . Bottomley <jejb@linux.vnet.ibm.com>, Reviewed-by: Bart Van Assche <bvanassche@acm.org> Tested-by: James Smart <james.smart@broadcom.com> Signed-off-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 2 May 2019 22:07:05 +0000 (16:07 -0600)]
Merge branch 'nvme-5.2' of git://git.infradead.org/nvme into for-5.2/block
Pull NVMe updates from Christoph.
* 'nvme-5.2' of git://git.infradead.org/nvme:
nvmet: protect discovery change log event list iteration
nvme: mark nvme_core_init and nvme_core_exit static
nvme: move command size checks to the core
nvme-fabrics: check more command sizes
nvme-pci: check more command sizes
nvme-pci: remove an unneeded variable initialization
nvme-pci: unquiesce admin queue on shutdown
nvme-pci: shutdown on timeout during deletion
nvme-pci: fix psdt field for single segment sgls
nvme-multipath: don't print ANA group state by default
nvme-multipath: split bios with the ns_head bio_set before submitting
nvme-tcp: fix possible null deref on a timed out io queue connect
Minwoo Im [Thu, 11 Apr 2019 15:18:32 +0000 (00:18 +0900)]
nvme-pci: check more command sizes
All the NVMe command has 64bytes fixed size so that it has been assured
with BUILD_BUG_ON(). The remaining command structures in linux/nvme.h
also need to be checked here.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Minwoo Im [Thu, 11 Apr 2019 15:52:39 +0000 (00:52 +0900)]
nvme-pci: remove an unneeded variable initialization
Variable "n" will be assigned once kstrtoint() succeeds, otherwise it
will not be referred because kstrtoint() will return an error which
means go out from this function.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Keith Busch [Tue, 30 Apr 2019 15:33:41 +0000 (09:33 -0600)]
nvme-pci: unquiesce admin queue on shutdown
Just like IO queues, the admin queue also will not be restarted after a
controller shutdown. Unquiesce this queue so that we do not block
request dispatch on a permanently disabled controller.
Reported-by: Yufen Yu <yuyufen@huawei.com> Signed-off-by: Keith Busch <keith.busch@intel.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Keith Busch [Tue, 30 Apr 2019 15:33:40 +0000 (09:33 -0600)]
nvme-pci: shutdown on timeout during deletion
We do not restart a controller in a deleting state for timeout errors.
When in this state, unblock potential request dispatchers with failed
completions by shutting down the controller on timeout detection.
Reported-by: Yufen Yu <yuyufen@huawei.com> Signed-off-by: Keith Busch <keith.busch@intel.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
The shortcut for single segment SGL requests did not set the PSDT field
to mark the request as using SGLs.
Fixes: 297910571f08 ("nvme-pci: optimize mapping single segment requests using SGLs") Signed-off-by: Klaus Birkelund Jensen <klaus.jensen@cnexlabs.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Hannes Reinecke [Tue, 30 Apr 2019 16:57:09 +0000 (18:57 +0200)]
nvme-multipath: split bios with the ns_head bio_set before submitting
If the bio is moved to a different queue via blk_steal_bios() and
the original queue is destroyed in nvme_remove_ns() we'll be ending
with a crash in bio_endio() as the mempool for the split bio bvecs
had already been destroyed.
So split the bio using the original queue (which will remain during the
lifetime of the bio) before sending it down to the underlying device.
Signed-off-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
This file has no copyright notice, but was added as part of a commit
adding another file using the default kernel GPLv2 license. Add
a matching SPDX tag.
The same page optimization is a rather odd corner case, which is not
used outside bio.c and which really should not be used outside of bio.c
either - we have better highlevel helpers like the rq/bio mapping
helpers.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
block: remove the i argument to bio_for_each_segment_all
We only have two callers that need the integer loop iterator, and they
can easily maintain it themselves.
Suggested-by: Matthew Wilcox <willy@infradead.org> Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de> Acked-by: David Sterba <dsterba@suse.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Acked-by: Coly Li <colyli@suse.de> Reviewed-by: Matthew Wilcox <willy@infradead.org> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Use a variable containing the buffer address instead of the to be
removed integer iterator from bio_for_each_segment_all.
Suggested-by: Matthew Wilcox <willy@infradead.org> Reviewed-by: Hannes Reinecke <hare@suse.com> Acked-by: Coly Li <colyli@suse.de> Reviewed-by: Matthew Wilcox <willy@infradead.org> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Tue, 30 Apr 2019 14:02:25 +0000 (22:02 +0800)]
bcache: remove redundant LIST_HEAD(journal) from run_cache_set()
Commit 95f18c9d1310 ("bcache: avoid potential memleak of list of
journal_replay(s) in the CACHE_SYNC branch of run_cache_set") forgets
to remove the original define of LIST_HEAD(journal), which makes
the change no take effect. This patch removes redundant variable
LIST_HEAD(journal) from run_cache_set(), to make Shenghui's fix
working.
Fixes: 95f18c9d1310 ("bcache: avoid potential memleak of list of journal_replay(s) in the CACHE_SYNC branch of run_cache_set") Reported-by: Juha Aatrokoski <juha.aatrokoski@aalto.fi> Cc: Shenghui Wang <shhuiw@foxmail.com> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Merge branch 'nvme-5.2' of git://git.infradead.org/nvme into for-5.2/block
Pull NVMe changes from Christoph.
* 'nvme-5.2' of git://git.infradead.org/nvme:
nvme: set 0 capacity if namespace block size exceeds PAGE_SIZE
nvme-rdma: fix typo in struct comment
nvme-loop: kill timeout handler
nvme-tcp: rename function to have nvme_tcp prefix
nvme-rdma: fix a NULL deref when an admin connect times out
nvme-tcp: fix a NULL deref when an admin connect times out
nvmet-tcp: don't fail maxr2t greater than 1
nvmet-file: clamp-down file namespace lba_shift
nvmet: include <linux/scatterlist.h>
nvmet: return a specified error it subsys_alloc fails
nvmet: rename nvme_completion instances from rsp to cqe
nvmet-rdma: remove p2p_client initialization from fast-path
Ming Lei [Mon, 15 Apr 2019 01:51:46 +0000 (09:51 +0800)]
nvme-loop: kill timeout handler
Firstly it doesn't make sense to handle timeout for loop: 1) for admin
queue, the request is always completed in code path of queuing IO. 2)
for normal IO request, the timeout on these IOs have been handled by
underlying queue already.
Secondly nvme-loop's timeout handler is simply broken, and easy to
cause issue: 1) no any sync/protection between timeout and normal
completion, and now it is driver's responsibility to deal with
that; 2) bad reset implementation, blk_mq_update_nr_hw_queues()
is called after all NSs's queue is stopped(quiesced), and easy
to trigger deadlock.
So kill the timeout handler.
Signed-off-by: Ming Lei <ming.lei@redhat.com> Reviewd-by: Keith Busch <keith.busch@intel.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de>
nvme-rdma: fix a NULL deref when an admin connect times out
If we timeout the admin startup sequence we might not yet have
an I/O tagset allocated which causes the teardown sequence to crash.
Make nvme_tcp_teardown_io_queues safe by not iterating inflight tags
if the tagset wasn't allocated.
nvme-tcp: fix a NULL deref when an admin connect times out
If we timeout the admin startup sequence we might not yet have
an I/O tagset allocated which causes the teardown sequence to crash.
Make nvme_tcp_teardown_io_queues safe by not iterating inflight tags
if the tagset wasn't allocated.
When the backing file is a tempfile for example, the inode i_blkbits
can be 1M in size which causes problems for hosts to support as the
disk block size. Instead, expose the minimum between i_blkbits and
12 (4K sector size).
drivers/nvme/target/core.c: In function 'nvmet_req_alloc_sgl':
drivers/nvme/target/core.c:939:12: error: implicit declaration of \
function 'sgl_alloc'; did you mean 'bio_alloc'? \
[-Werror=implicit-function-declaration]
req->sg = sgl_alloc(req->transfer_len, GFP_KERNEL, &req->sg_cnt);
^~~~~~~~~
bio_alloc
drivers/nvme/target/core.c:939:10: warning: assignment makes pointer \
from integer without a cast [-Wint-conversion]
req->sg = sgl_alloc(req->transfer_len, GFP_KERNEL, &req->sg_cnt);
^
drivers/nvme/target/core.c: In function 'nvmet_req_free_sgl':
drivers/nvme/target/core.c:952:3: error: implicit declaration of \
function 'sgl_free'; did you mean 'ida_free'? [-Werror=implicit-function-declaration]
sgl_free(req->sg);
^~~~~~~~
ida_free
Cause:
1. missing include to <linux/scatterlist.h>
2. SGL_ALLOC needs to be enabled
Therefore adding the missing include, as well as Kconfig dependency.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Reviewed-by: Minwoo Im <minwoo.im@samsung.com> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by: Christoph Hellwig <hch@lst.de>
Minwoo Im [Sun, 7 Apr 2019 06:28:06 +0000 (15:28 +0900)]
nvmet: return a specified error it subsys_alloc fails
nvmet_subsys_alloc() returns its pointer or NULL if it fails. We can
see three different steps in this function:
1. memory allocation
2. argument check
3. memory allocation for string
But now the callers of this function do not seem to handle case 2 by
returning -ENOMEM only even if it fails with an invalid parameter.
This patch specifies error codes so that caller can pass it to its own
caller.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>. Signed-off-by: Christoph Hellwig <hch@lst.de>
The reason is in journal_reclaim(), when discard is enabled, we send
discard command and reclaim those journal buckets whose seq is old
than the last_seq_now, but before we write a journal with last_seq_now,
the machine is restarted, so the journal with the last_seq_now is not
written to the journal bucket, and the last_seq_wrote in the newest
journal is old than last_seq_now which we expect to be, so when we doing
replay, journals from last_seq_wrote to last_seq_now are missing.
It's hard to write a journal immediately after journal_reclaim(),
and it harmless if those missed journal are caused by discarding
since those journals are already wrote to btree node. So, if miss
seqs are started from the beginning journal, we treat it as normal,
and only print a message to show the miss journal, and point out
it maybe caused by discarding.
Patch v2 add a judgement condition to ignore the missed journal
only when discard enabled as Coly suggested.
(Coly Li: rebase the patch with other changes in bch_journal_replay())
Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com> Tested-by: Dennis Schridde <devurandom@gmx.net> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:40 +0000 (00:48 +0800)]
bcache: improve bcache_reboot()
This patch tries to release mutex bch_register_lock early, to give
chance to stop cache set and bcache device early.
This patch also expends time out of stopping all bcache device from
2 seconds to 10 seconds, because stopping writeback rate update worker
may delay for 5 seconds, 2 seconds is not enough.
After this patch applied, stopping bcache devices during system reboot
or shutdown is very hard to be observed any more.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:39 +0000 (00:48 +0800)]
bcache: add comments for closure_fn to be called in closure_queue()
Add code comments to explain which call back function might be called
for the closure_queue(). This is an effort to make code to be more
understandable for readers.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:38 +0000 (00:48 +0800)]
bcache: Add comments for blkdev_put() in registration code path
Add comments to explain why in register_bcache() blkdev_put() won't
be called in two location. Add comments to explain why blkdev_put()
must be called in register_cache() when cache_alloc() failed.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:37 +0000 (00:48 +0800)]
bcache: add error check for calling register_bdev()
This patch adds return value to register_bdev(). Then if failure happens
inside register_bdev(), its caller register_bcache() may detect and
handle the failure more properly.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:36 +0000 (00:48 +0800)]
bcache: return error immediately in bch_journal_replay()
When failure happens inside bch_journal_replay(), calling
cache_set_err_on() and handling the failure in async way is not a good
idea. Because after bch_journal_replay() returns, registering code will
continue to execute following steps, and unregistering code triggered
by cache_set_err_on() is running in same time. First it is unnecessary
to handle failure and unregister cache set in an async way, second there
might be potential race condition to run register and unregister code
for same cache set.
So in this patch, if failure happens in bch_journal_replay(), we don't
call cache_set_err_on(), and just print out the same error message to
kernel message buffer, then return -EIO immediately caller. Then caller
can detect such failure and handle it in synchrnozied way.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:35 +0000 (00:48 +0800)]
bcache: add comments for kobj release callback routine
Bcache has several routines to release resources in implicit way, they
are called when the associated kobj released. This patch adds code
comments to notice when and which release callback will be called,
- When dc->disk.kobj released:
void bch_cached_dev_release(struct kobject *kobj)
- When d->kobj released:
void bch_flash_dev_release(struct kobject *kobj)
- When c->kobj released:
void bch_cache_set_release(struct kobject *kobj)
- When ca->kobj released
void bch_cache_release(struct kobject *kobj)
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:34 +0000 (00:48 +0800)]
bcache: add failure check to run_cache_set() for journal replay
Currently run_cache_set() has no return value, if there is failure in
bch_journal_replay(), the caller of run_cache_set() has no idea about
such failure and just continue to execute following code after
run_cache_set(). The internal failure is triggered inside
bch_journal_replay() and being handled in async way. This behavior is
inefficient, while failure handling inside bch_journal_replay(), cache
register code is still running to start the cache set. Registering and
unregistering code running as same time may introduce some rare race
condition, and make the code to be more hard to be understood.
This patch adds return value to run_cache_set(), and returns -EIO if
bch_journal_rreplay() fails. Then caller of run_cache_set() may detect
such failure and stop registering code flow immedidately inside
register_cache_set().
If journal replay fails, run_cache_set() can report error immediately
to register_cache_set(). This patch makes the failure handling for
bch_journal_replay() be in synchronized way, easier to understand and
debug, and avoid poetential race condition for register-and-unregister
in same time.
Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Coly Li [Wed, 24 Apr 2019 16:48:33 +0000 (00:48 +0800)]
bcache: never set KEY_PTRS of journal key to 0 in journal_reclaim()
In journal_reclaim() ja->cur_idx of each cache will be update to
reclaim available journal buckets. Variable 'int n' is used to count how
many cache is successfully reclaimed, then n is set to c->journal.key
by SET_KEY_PTRS(). Later in journal_write_unlocked(), a for_each_cache()
loop will write the jset data onto each cache.
The problem is, if all jouranl buckets on each cache is full, the
following code in journal_reclaim(),
529 for_each_cache(ca, c, iter) {
530 struct journal_device *ja = &ca->journal;
531 unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets;
532
533 /* No space available on this device */
534 if (next == ja->discard_idx)
535 continue;
536
537 ja->cur_idx = next;
538 k->ptr[n++] = MAKE_PTR(0,
539 bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
540 ca->sb.nr_this_dev);
541 }
542
543 bkey_init(k);
544 SET_KEY_PTRS(k, n);
If there is no available bucket to reclaim, the if() condition at line
534 will always true, and n remains 0. Then at line 544, SET_KEY_PTRS()
will set KEY_PTRS field of c->journal.key to 0.
Setting KEY_PTRS field of c->journal.key to 0 is wrong. Because in
journal_write_unlocked() the journal data is written in following loop,
649 for (i = 0; i < KEY_PTRS(k); i++) {
650-671 submit journal data to cache device
672 }
If KEY_PTRS field is set to 0 in jouranl_reclaim(), the journal data
won't be written to cache device here. If system crahed or rebooted
before bkeys of the lost journal entries written into btree nodes, data
corruption will be reported during bcache reload after rebooting the
system.
Indeed there is only one cache in a cache set, there is no need to set
KEY_PTRS field in journal_reclaim() at all. But in order to keep the
for_each_cache() logic consistent for now, this patch fixes the above
problem by not setting 0 KEY_PTRS of journal key, if there is no bucket
available to reclaim.
Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
projects / linux.git / log