nbd: don't requeue the same request twice.
We can race with the snd timeout and the per-request timeout and end up requeuing the same request twice. We can't use the send_complete completion to tell if everything is ok because we hold the tx_lock during send, so the timeout stuff will block waiting to mark the socket dead, and we could be marked complete and still requeue. Instead add a flag to the socket so we know whether we've been requeued yet. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
parent
70dbcc2254
commit
d7d94d48a2
@ -112,12 +112,15 @@ struct nbd_device {
|
|||||||
struct task_struct *task_setup;
|
struct task_struct *task_setup;
|
||||||
};
|
};
|
||||||
|
|
||||||
|
#define NBD_CMD_REQUEUED 1
|
||||||
|
|
||||||
struct nbd_cmd {
|
struct nbd_cmd {
|
||||||
struct nbd_device *nbd;
|
struct nbd_device *nbd;
|
||||||
int index;
|
int index;
|
||||||
int cookie;
|
int cookie;
|
||||||
struct completion send_complete;
|
struct completion send_complete;
|
||||||
blk_status_t status;
|
blk_status_t status;
|
||||||
|
unsigned long flags;
|
||||||
};
|
};
|
||||||
|
|
||||||
#if IS_ENABLED(CONFIG_DEBUG_FS)
|
#if IS_ENABLED(CONFIG_DEBUG_FS)
|
||||||
@ -146,6 +149,14 @@ static inline struct device *nbd_to_dev(struct nbd_device *nbd)
|
|||||||
return disk_to_dev(nbd->disk);
|
return disk_to_dev(nbd->disk);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
static void nbd_requeue_cmd(struct nbd_cmd *cmd)
|
||||||
|
{
|
||||||
|
struct request *req = blk_mq_rq_from_pdu(cmd);
|
||||||
|
|
||||||
|
if (!test_and_set_bit(NBD_CMD_REQUEUED, &cmd->flags))
|
||||||
|
blk_mq_requeue_request(req, true);
|
||||||
|
}
|
||||||
|
|
||||||
static const char *nbdcmd_to_ascii(int cmd)
|
static const char *nbdcmd_to_ascii(int cmd)
|
||||||
{
|
{
|
||||||
switch (cmd) {
|
switch (cmd) {
|
||||||
@ -343,7 +354,7 @@ static enum blk_eh_timer_return nbd_xmit_timeout(struct request *req,
|
|||||||
nbd_mark_nsock_dead(nbd, nsock, 1);
|
nbd_mark_nsock_dead(nbd, nsock, 1);
|
||||||
mutex_unlock(&nsock->tx_lock);
|
mutex_unlock(&nsock->tx_lock);
|
||||||
}
|
}
|
||||||
blk_mq_requeue_request(req, true);
|
nbd_requeue_cmd(cmd);
|
||||||
nbd_config_put(nbd);
|
nbd_config_put(nbd);
|
||||||
return BLK_EH_DONE;
|
return BLK_EH_DONE;
|
||||||
}
|
}
|
||||||
@ -500,6 +511,7 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index)
|
|||||||
nsock->pending = req;
|
nsock->pending = req;
|
||||||
nsock->sent = sent;
|
nsock->sent = sent;
|
||||||
}
|
}
|
||||||
|
set_bit(NBD_CMD_REQUEUED, &cmd->flags);
|
||||||
return BLK_STS_RESOURCE;
|
return BLK_STS_RESOURCE;
|
||||||
}
|
}
|
||||||
dev_err_ratelimited(disk_to_dev(nbd->disk),
|
dev_err_ratelimited(disk_to_dev(nbd->disk),
|
||||||
@ -541,6 +553,7 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index)
|
|||||||
*/
|
*/
|
||||||
nsock->pending = req;
|
nsock->pending = req;
|
||||||
nsock->sent = sent;
|
nsock->sent = sent;
|
||||||
|
set_bit(NBD_CMD_REQUEUED, &cmd->flags);
|
||||||
return BLK_STS_RESOURCE;
|
return BLK_STS_RESOURCE;
|
||||||
}
|
}
|
||||||
dev_err(disk_to_dev(nbd->disk),
|
dev_err(disk_to_dev(nbd->disk),
|
||||||
@ -805,7 +818,7 @@ static int nbd_handle_cmd(struct nbd_cmd *cmd, int index)
|
|||||||
*/
|
*/
|
||||||
blk_mq_start_request(req);
|
blk_mq_start_request(req);
|
||||||
if (unlikely(nsock->pending && nsock->pending != req)) {
|
if (unlikely(nsock->pending && nsock->pending != req)) {
|
||||||
blk_mq_requeue_request(req, true);
|
nbd_requeue_cmd(cmd);
|
||||||
ret = 0;
|
ret = 0;
|
||||||
goto out;
|
goto out;
|
||||||
}
|
}
|
||||||
@ -818,7 +831,7 @@ static int nbd_handle_cmd(struct nbd_cmd *cmd, int index)
|
|||||||
dev_err_ratelimited(disk_to_dev(nbd->disk),
|
dev_err_ratelimited(disk_to_dev(nbd->disk),
|
||||||
"Request send failed, requeueing\n");
|
"Request send failed, requeueing\n");
|
||||||
nbd_mark_nsock_dead(nbd, nsock, 1);
|
nbd_mark_nsock_dead(nbd, nsock, 1);
|
||||||
blk_mq_requeue_request(req, true);
|
nbd_requeue_cmd(cmd);
|
||||||
ret = 0;
|
ret = 0;
|
||||||
}
|
}
|
||||||
out:
|
out:
|
||||||
@ -843,6 +856,7 @@ static blk_status_t nbd_queue_rq(struct blk_mq_hw_ctx *hctx,
|
|||||||
* done sending everything over the wire.
|
* done sending everything over the wire.
|
||||||
*/
|
*/
|
||||||
init_completion(&cmd->send_complete);
|
init_completion(&cmd->send_complete);
|
||||||
|
clear_bit(NBD_CMD_REQUEUED, &cmd->flags);
|
||||||
|
|
||||||
/* We can be called directly from the user space process, which means we
|
/* We can be called directly from the user space process, which means we
|
||||||
* could possibly have signals pending so our sendmsg will fail. In
|
* could possibly have signals pending so our sendmsg will fail. In
|
||||||
@ -1460,6 +1474,7 @@ static int nbd_init_request(struct blk_mq_tag_set *set, struct request *rq,
|
|||||||
{
|
{
|
||||||
struct nbd_cmd *cmd = blk_mq_rq_to_pdu(rq);
|
struct nbd_cmd *cmd = blk_mq_rq_to_pdu(rq);
|
||||||
cmd->nbd = set->driver_data;
|
cmd->nbd = set->driver_data;
|
||||||
|
cmd->flags = 0;
|
||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
Loading…
Reference in New Issue
Block a user