| Message ID | 20230731221458.437440-2-bvanassche@acm.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Improve the performance for zoned UFS devices | expand |
On 8/1/23 07:14, Bart Van Assche wrote: > Writes in sequential write required zones must happen at the write > pointer. Even if the submitter of the write commands (e.g. a filesystem) > submits writes for sequential write required zones in order, the block > layer or the storage controller may reorder these write commands. > > The zone locking mechanism in the mq-deadline I/O scheduler serializes > write commands for sequential zones. Some but not all storage controllers > require this serialization. Introduce a new request queue flag to allow > block drivers to indicate that they preserve the order of write commands > and thus do not require serialization of writes per zone. > > Cc: Christoph Hellwig <hch@lst.de> > Cc: Damien Le Moal <dlemoal@kernel.org> > Cc: Ming Lei <ming.lei@redhat.com> > Signed-off-by: Bart Van Assche <bvanassche@acm.org> Looks OK. Very minor nit below. Reviewed-by: Damien Le Moal <dlemoal@kernel.org> > --- > include/linux/blkdev.h | 10 ++++++++++ > 1 file changed, 10 insertions(+) > > diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h > index 2f5371b8482c..de5e05cc34fa 100644 > --- a/include/linux/blkdev.h > +++ b/include/linux/blkdev.h > @@ -534,6 +534,11 @@ struct request_queue { > #define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */ > #define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */ > #define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */ > +/* > + * Do not serialize sequential writes (REQ_OP_WRITE, REQ_OP_WRITE_ZEROES) sent > + * to a sequential write required zone (BLK_ZONE_TYPE_SEQWRITE_REQ). > + */ I would be very explicit here, for this to be clear to people who are not familiar with zone device write operation handling. Something like: /* * The device supports not using the zone write locking mechanism to serialize * write operations (REQ_OP_WRITE, REQ_OP_WRITE_ZEROES) issued to a sequential * write required zone (BLK_ZONE_TYPE_SEQWRITE_REQ). */ > +#define QUEUE_FLAG_NO_ZONE_WRITE_LOCK 8 > #define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */ > #define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */ > #define QUEUE_FLAG_SYNCHRONOUS 11 /* always completes in submit context */ > @@ -597,6 +602,11 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q); > #define blk_queue_skip_tagset_quiesce(q) \ > test_bit(QUEUE_FLAG_SKIP_TAGSET_QUIESCE, &(q)->queue_flags) > > +static inline bool blk_queue_no_zone_write_lock(struct request_queue *q) > +{ > + return test_bit(QUEUE_FLAG_NO_ZONE_WRITE_LOCK, &q->queue_flags); > +} > + > extern void blk_set_pm_only(struct request_queue *q); > extern void blk_clear_pm_only(struct request_queue *q); >
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 2f5371b8482c..de5e05cc34fa 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -534,6 +534,11 @@ struct request_queue { #define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */ #define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */ #define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */ +/* + * Do not serialize sequential writes (REQ_OP_WRITE, REQ_OP_WRITE_ZEROES) sent + * to a sequential write required zone (BLK_ZONE_TYPE_SEQWRITE_REQ). + */ +#define QUEUE_FLAG_NO_ZONE_WRITE_LOCK 8 #define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */ #define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */ #define QUEUE_FLAG_SYNCHRONOUS 11 /* always completes in submit context */ @@ -597,6 +602,11 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q); #define blk_queue_skip_tagset_quiesce(q) \ test_bit(QUEUE_FLAG_SKIP_TAGSET_QUIESCE, &(q)->queue_flags) +static inline bool blk_queue_no_zone_write_lock(struct request_queue *q) +{ + return test_bit(QUEUE_FLAG_NO_ZONE_WRITE_LOCK, &q->queue_flags); +} + extern void blk_set_pm_only(struct request_queue *q); extern void blk_clear_pm_only(struct request_queue *q);
Writes in sequential write required zones must happen at the write pointer. Even if the submitter of the write commands (e.g. a filesystem) submits writes for sequential write required zones in order, the block layer or the storage controller may reorder these write commands. The zone locking mechanism in the mq-deadline I/O scheduler serializes write commands for sequential zones. Some but not all storage controllers require this serialization. Introduce a new request queue flag to allow block drivers to indicate that they preserve the order of write commands and thus do not require serialization of writes per zone. Cc: Christoph Hellwig <hch@lst.de> Cc: Damien Le Moal <dlemoal@kernel.org> Cc: Ming Lei <ming.lei@redhat.com> Signed-off-by: Bart Van Assche <bvanassche@acm.org> --- include/linux/blkdev.h | 10 ++++++++++ 1 file changed, 10 insertions(+)