block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead

commit e692cb668fdd5a712c6ed2a2d6f2a36ee83997b4 upstream.

When stacking devices, a request_queue is not always available. This
forced us to have a no_cluster flag in the queue_limits that could be
used as a carrier until the request_queue had been set up for a
metadevice.

There were several problems with that approach. First of all it was up
to the stacking device to remember to set queue flag after stacking had
completed. Also, the queue flag and the queue limits had to be kept in
sync at all times. We got that wrong, which could lead to us issuing
commands that went beyond the max scatterlist limit set by the driver.

The proper fix is to avoid having two flags for tracking the same thing.
We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the
block layer merging functions. The queue_limit 'no_cluster' is turned
into 'cluster' to avoid double negatives and to ease stacking.
Clustering defaults to being enabled as before. The queue flag logic is
removed from the stacking function, and explicitly setting the cluster
flag is no longer necessary in DM and MD.

Reported-by: Ed Lin <ed.lin@promise.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/block/blk-merge.c b/block/blk-merge.c
index 99cb5cf1f447fa8759d73f06bf64fe4b20eec580..1912090f764e495d159674332bd6f4816af003b9 100644 (file)
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -22,7 +22,7 @@ static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
                return 0;
 
        fbio = bio;
-       cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
+       cluster = blk_queue_cluster(q);
        seg_size = 0;
        phys_size = nr_phys_segs = 0;
        for_each_bio(bio) {
@@ -88,7 +88,7 @@ EXPORT_SYMBOL(blk_recount_segments);
 static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
                                   struct bio *nxt)
 {
-       if (!test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags))
+       if (!blk_queue_cluster(q))
                return 0;
 
        if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
@@ -124,7 +124,7 @@ int blk_rq_map_sg(struct request_queue *q, struct request *rq,
        int nsegs, cluster;
 
        nsegs = 0;
-       cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
+       cluster = blk_queue_cluster(q);
 
        /*
         * for each bio in rq

diff --git a/block/blk-settings.c b/block/blk-settings.c
index 7c7b8c1c190f75faadb5318ded7cfa1febc71a8d..112c4f7c7121f6e9602f50f407a89c9082f323a2 100644 (file)
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -103,7 +103,7 @@ void blk_set_default_limits(struct queue_limits *lim)
        lim->alignment_offset = 0;
        lim->io_opt = 0;
        lim->misaligned = 0;
-       lim->no_cluster = 0;
+       lim->cluster = 1;
 }
 EXPORT_SYMBOL(blk_set_default_limits);
 
@@ -477,15 +477,6 @@ EXPORT_SYMBOL(blk_queue_io_opt);
 void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
 {
        blk_stack_limits(&t->limits, &b->limits, 0);
-
-       if (!t->queue_lock)
-               WARN_ON_ONCE(1);
-       else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) {
-               unsigned long flags;
-               spin_lock_irqsave(t->queue_lock, flags);
-               queue_flag_clear(QUEUE_FLAG_CLUSTER, t);
-               spin_unlock_irqrestore(t->queue_lock, flags);
-       }
 }
 EXPORT_SYMBOL(blk_queue_stack_limits);
 
@@ -561,7 +552,7 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
        t->io_min = max(t->io_min, b->io_min);
        t->io_opt = lcm(t->io_opt, b->io_opt);
 
-       t->no_cluster |= b->no_cluster;
+       t->cluster &= b->cluster;
 
        /* Physical block size a multiple of the logical block size? */
        if (t->physical_block_size & (t->logical_block_size - 1)) {
@@ -652,17 +643,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
                printk(KERN_NOTICE "%s: Warning: Device %s is misaligned\n",
                       top, bottom);
        }
-
-       if (!t->queue_lock)
-               WARN_ON_ONCE(1);
-       else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) {
-               unsigned long flags;
-
-               spin_lock_irqsave(t->queue_lock, flags);
-               if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags))
-                       queue_flag_clear(QUEUE_FLAG_CLUSTER, t);
-               spin_unlock_irqrestore(t->queue_lock, flags);
-       }
 }
 EXPORT_SYMBOL(disk_stack_limits);
 

diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index e86912864e0443a348f3a829f64f5ed48c2212fb..57f32f075e349331f600092618d5c6426b34292f 100644 (file)
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -1082,11 +1082,6 @@ void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
         */
        q->limits = *limits;
 
-       if (limits->no_cluster)
-               queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q);
-       else
-               queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, q);
-
        dm_table_set_integrity(t);
 
        /*

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 81bdfd810e3dd691178e3c68766cafbf31a827fb..2c66c7ea32007985fde4d164e1a8c0859fe6ea50 100644 (file)
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -3959,9 +3959,6 @@ static int md_alloc(dev_t dev, char *name)
                goto abort;
        mddev->queue->queuedata = mddev;
 
-       /* Can be unlocked because the queue is new: no concurrency */
-       queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, mddev->queue);
-
        blk_queue_make_request(mddev->queue, md_make_request);
 
        disk = alloc_disk(1 << shift);

diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index b87fc30fad6ba17f1f8158acdcb436e6e589afb8..d78828ff92c1debf9b42e69f24ab853cbe040851 100644 (file)
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1636,9 +1636,8 @@ struct request_queue *__scsi_alloc_queue(struct Scsi_Host *shost,
 
        blk_queue_max_segment_size(q, dma_get_max_seg_size(dev));
 
-       /* New queue, no concurrency on queue_flags */
        if (!shost->use_clustering)
-               queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q);
+               q->limits.cluster = 0;
 
        /*
         * set a reasonable default alignment on word boundaries: the

diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 05f6018b928eb1860761e047bdecdebfecb1bc51..a06bfabec4d096609475c80eedf352b2c80d2e70 100644 (file)
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -318,7 +318,7 @@ struct queue_limits {
        unsigned short          max_phys_segments;
 
        unsigned char           misaligned;
-       unsigned char           no_cluster;
+       unsigned char           cluster;
 };
 
 struct request_queue
@@ -440,7 +440,6 @@ struct request_queue
 #endif
 };
 
-#define QUEUE_FLAG_CLUSTER     0       /* cluster several segments into 1 */
 #define QUEUE_FLAG_QUEUED      1       /* uses generic tag queueing */
 #define QUEUE_FLAG_STOPPED     2       /* queue is stopped */
 #define        QUEUE_FLAG_SYNCFULL     3       /* read queue has been filled */
@@ -461,7 +460,6 @@ struct request_queue
 #define QUEUE_FLAG_DISCARD     17      /* supports DISCARD */
 
 #define QUEUE_FLAG_DEFAULT     ((1 << QUEUE_FLAG_IO_STAT) |            \
-                                (1 << QUEUE_FLAG_CLUSTER) |            \
                                 (1 << QUEUE_FLAG_STACKABLE)    |       \
                                 (1 << QUEUE_FLAG_SAME_COMP))
 
@@ -627,6 +625,11 @@ enum {
 
 #define rq_data_dir(rq)                ((rq)->cmd_flags & 1)
 
+static inline unsigned int blk_queue_cluster(struct request_queue *q)
+{
+       return q->limits.cluster;
+}
+
 /*
  * We regard a request as sync, if either a read or a sync write
  */