kernel/git/stable/linux.git/drivers/block/nbd.c, branch linux-6.10.y

nbd: correct the maximum value for discard sectors

2024-10-04T14:32:32+00:00

[ Upstream commit 296dbc72d29085d5fc34430d0760423071e9e81d ] The version of the NBD protocol implemented by the kernel driver currently has a 32 bit field for length values. As the NBD protocol uses bytes as a unit of length, length values larger than 2^32 bytes cannot be expressed. Update the max_hw_discard_sectors field to match that. Signed-off-by: Wouter Verhelst Fixes: 268283244c0f ("nbd: use the atomic queue limits API in nbd_set_size") Reviewed-by: Damien Le Moal Cc: Eric Blake Link: https://lore.kernel.org/r/20240812133032.115134-8-w@uter.be Signed-off-by: Jens Axboe Signed-off-by: Sasha Levin

nbd: fix race between timeout and normal completion

2024-10-04T14:32:31+00:00

[ Upstream commit c9ea57c91f03bcad415e1a20113bdb2077bcf990 ] If request timetout is handled by nbd_requeue_cmd(), normal completion has to be stopped for avoiding to complete this requeued request, other use-after-free can be triggered. Fix the race by clearing NBD_CMD_INFLIGHT in nbd_requeue_cmd(), meantime make sure that cmd->lock is grabbed for clearing the flag and the requeue. Cc: Josef Bacik Cc: Yu Kuai Fixes: 2895f1831e91 ("nbd: don't clear 'NBD_CMD_INFLIGHT' flag if request is not completed") Signed-off-by: Ming Lei Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20240830034145.1827742-1-ming.lei@redhat.com Signed-off-by: Jens Axboe Signed-off-by: Sasha Levin

nbd: Remove __force casts

2024-06-12T20:47:07+00:00

Make it again possible for sparse to verify that blk_status_t and Unix error codes are used in the proper context by making nbd_send_cmd() return a blk_status_t instead of an integer. No functionality has been changed. Signed-off-by: Christoph Hellwig [ bvanassche: added description and made two small formatting changes ] Signed-off-by: Bart Van Assche Reviewed-by: Chaitanya Kulkarni Link: https://lore.kernel.org/r/20240604221531.327131-1-bvanassche@acm.org Signed-off-by: Jens Axboe

nbd: Fix signal handling

2024-05-14T13:22:35+00:00

Both nbd_send_cmd() and nbd_handle_cmd() return either a negative error number or a positive blk_status_t value. nbd_queue_rq() converts these return values into a blk_status_t value. There is a bug in the conversion code: if nbd_send_cmd() returns BLK_STS_RESOURCE, nbd_queue_rq() should return BLK_STS_RESOURCE instead of BLK_STS_OK. Fix this, move the conversion code into nbd_handle_cmd() and fix the remaining sparse warnings. This patch fixes the following sparse warnings: drivers/block/nbd.c:673:32: warning: incorrect type in return expression (different base types) drivers/block/nbd.c:673:32: expected int drivers/block/nbd.c:673:32: got restricted blk_status_t [usertype] drivers/block/nbd.c:714:48: warning: incorrect type in return expression (different base types) drivers/block/nbd.c:714:48: expected int drivers/block/nbd.c:714:48: got restricted blk_status_t [usertype] drivers/block/nbd.c:1120:21: warning: incorrect type in assignment (different base types) drivers/block/nbd.c:1120:21: expected int [assigned] ret drivers/block/nbd.c:1120:21: got restricted blk_status_t [usertype] drivers/block/nbd.c:1125:16: warning: incorrect type in return expression (different base types) drivers/block/nbd.c:1125:16: expected restricted blk_status_t drivers/block/nbd.c:1125:16: got int [assigned] ret Cc: Christoph Hellwig Cc: Josef Bacik Cc: Yu Kuai Cc: Markus Pargmann Fixes: fc17b6534eb8 ("blk-mq: switch ->queue_rq return value to blk_status_t") Cc: stable@vger.kernel.org Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240510202313.25209-6-bvanassche@acm.org Signed-off-by: Jens Axboe

nbd: Remove a local variable from nbd_send_cmd()

2024-05-14T13:22:35+00:00

blk_rq_bytes() returns an unsigned int while 'size' has type unsigned long. This is confusing. Improve code readability by removing the local variable 'size'. Cc: Christoph Hellwig Cc: Josef Bacik Cc: Yu Kuai Cc: Markus Pargmann Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240510202313.25209-5-bvanassche@acm.org Signed-off-by: Jens Axboe

nbd: Improve the documentation of the locking assumptions

2024-05-14T13:22:35+00:00

Document locking assumptions with lockdep_assert_held() instead of source code comments. The advantage of lockdep_assert_held() is that it is verified at runtime if lockdep is enabled in the kernel config. Cc: Christoph Hellwig Cc: Josef Bacik Cc: Yu Kuai Cc: Markus Pargmann Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240510202313.25209-4-bvanassche@acm.org Signed-off-by: Jens Axboe

nbd: Remove superfluous casts

2024-05-14T13:22:35+00:00

In Linux kernel code it is preferred not to use a cast when converting a void pointer to another pointer type. Cc: Christoph Hellwig Cc: Josef Bacik Cc: Yu Kuai Cc: Markus Pargmann Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240510202313.25209-3-bvanassche@acm.org Signed-off-by: Jens Axboe

nbd: use the atomic queue limits API in nbd_set_size

2024-03-01T16:08:22+00:00

Use queue_limits_start_update / queue_limits_commit_update to update all the limits in one go and with proper sanity checking. Signed-off-by: Christoph Hellwig Link: https://lore.kernel.org/r/20240229143846.1047223-4-hch@lst.de Signed-off-by: Jens Axboe

nbd: freeze the queue for queue limits updates

2024-03-01T16:08:22+00:00

nbd currently updates the logical and physical block sizes as well as the discard_sectors on a live queue. Freeze the queue first to make sure there are not commands in flight that can see torn or inconsistent limits. Signed-off-by: Christoph Hellwig Link: https://lore.kernel.org/r/20240229143846.1047223-3-hch@lst.de Signed-off-by: Jens Axboe

nbd: don't clear discard_sectors in nbd_config_put

2024-03-01T16:08:22+00:00

nbd_config_put currently clears discard_sectors when unusing a device. This is pretty odd behavior and different from the sector size configuration which is simply left in places and then reconfigured when nbd_set_size is as part of configuring the device. Change nbd_set_size to clear discard_sectors if discard is not supported so that all the queue limits changes are handled in one place. Signed-off-by: Christoph Hellwig Link: https://lore.kernel.org/r/20240229143846.1047223-2-hch@lst.de Signed-off-by: Jens Axboe