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 eafc94f68d79f2ceb1c0a89fce8e449c3a5ff3f6..c24bf43d51f033863792b714c71e03b9570a80a4 100644 (file)
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -21,7 +21,7 @@ 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;
        nr_phys_segs = 0;
        for_each_bio(bio) {
@@ -87,7 +87,7 @@ 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 >
@@ -123,7 +123,7 @@ 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 450577d86ac49832852a65f7b3e8bf9f4a9d614c..ea9430d3d7d2ccea8fdef37b9ef6c7353a4969c0 100644 (file)
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -125,7 +125,7 @@ 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);
 
@@ -468,15 +468,6 @@ 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);
 
@@ -547,7 +538,7 @@ 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;
        t->discard_zeroes_data &= b->discard_zeroes_data;
 
        /* Physical block size a multiple of the logical block size? */
@@ -643,7 +634,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
                       sector_t offset)
 {
        struct request_queue *t = disk->queue;
-       struct request_queue *b = bdev_get_queue(bdev);
 
        if (bdev_stack_limits(&t->limits, bdev, offset >> 9) < 0) {
                char top[BDEVNAME_SIZE], bottom[BDEVNAME_SIZE];
@@ -654,17 +644,6 @@ 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/block/blk-sysfs.c b/block/blk-sysfs.c
index 0749b89c68852fc824e4afa862d871814c56ddd8..a26c930db346882b2d23494b64acd8d854c3cfd6 100644 (file)
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -114,7 +114,7 @@ static ssize_t queue_max_segments_show(struct request_queue *q, char *page)
 
 static ssize_t queue_max_segment_size_show(struct request_queue *q, char *page)
 {
-       if (test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags))
+       if (blk_queue_cluster(q))
                return queue_var_show(queue_max_segment_size(q), (page));
 
        return queue_var_show(PAGE_CACHE_SIZE, (page));

diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index f9fc07d7a4b91f9db6aeb688cb5346bf604d1ac4..87e4e78790c00cc5cea202ab02ac66df52425193 100644 (file)
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -1136,11 +1136,6 @@ 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);
-
        if (!dm_table_supports_discards(t))
                queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q);
        else

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 58fc0187122d88163c75061b2e5aacb9d7d2ecc0..4d45012f77b350540cfc8fb5da9a947cbf70d3ec 100644 (file)
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -4289,9 +4289,6 @@ 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 71c4cf7b805180739f3965873b04246d9db5b4b9..5bff8a2396e4201be2ef5ee67e5d6844c672e076 100644 (file)
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1632,9 +1632,8 @@ 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 9e443b9c0223a1728c58b3df63f83b983b7a2de7..f142a7f300f8261a4da27ed37c9687d19b47a190 100644 (file)
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -246,7 +246,7 @@ struct queue_limits {
 
        unsigned char           misaligned;
        unsigned char           discard_misaligned;
-       unsigned char           no_cluster;
+       unsigned char           cluster;
        signed char             discard_zeroes_data;
 };
 
@@ -369,7 +369,6 @@ 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 */
@@ -392,7 +391,6 @@ struct request_queue
 #define QUEUE_FLAG_SECDISCARD  19      /* supports SECDISCARD */
 
 #define QUEUE_FLAG_DEFAULT     ((1 << QUEUE_FLAG_IO_STAT) |            \
-                                (1 << QUEUE_FLAG_CLUSTER) |            \
                                 (1 << QUEUE_FLAG_STACKABLE)    |       \
                                 (1 << QUEUE_FLAG_SAME_COMP)    |       \
                                 (1 << QUEUE_FLAG_ADD_RANDOM))
@@ -550,6 +548,11 @@ 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
  */