[SCSI] block: add support for shared tag maps

The current block queue implementation already contains most of the
machinery for shared tag maps.  The only remaining pieces are a way to
allocate and destroy a tag map independently of the queues (so that
the maps can be managed on the life cycle of the overseeing entity)

Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

2 files changed, 90 insertions, 21 deletions

diff --git a/block/ll_rw_blk.c b/block/ll_rw_blk.c
index ddd9253f9d55..556a3d354eab 100644
--- a/block/ll_rw_blk.c
+++ b/block/ll_rw_blk.c
@@ -848,21 +848,18 @@ struct request *blk_queue_find_tag(request_queue_t *q, int tag)
 EXPORT_SYMBOL(blk_queue_find_tag);
 
 /**
- * __blk_queue_free_tags - release tag maintenance info
- * @q:  the request queue for the device
+ * __blk_free_tags - release a given set of tag maintenance info
+ * @bqt:        the tag map to free
  *
- *  Notes:
- *    blk_cleanup_queue() will take care of calling this function, if tagging
- *    has been used. So there's no need to call this directly.
- **/
-static void __blk_queue_free_tags(request_queue_t *q)
+ * Tries to free the specified @bqt@.  Returns true if it was
+ * actually freed and false if there are still references using it
+ */
+static int __blk_free_tags(struct blk_queue_tag *bqt)
 {
-        struct blk_queue_tag *bqt = q->queue_tags;
-
-        if (!bqt)
-                return;
+        int retval;
 
-        if (atomic_dec_and_test(&bqt->refcnt)) {
+        retval = atomic_dec_and_test(&bqt->refcnt);
+        if (retval) {
                 BUG_ON(bqt->busy);
                 BUG_ON(!list_empty(&bqt->busy_list));
 
@@ -873,12 +870,49 @@ static void __blk_queue_free_tags(request_queue_t *q)
                 bqt->tag_map = NULL;
 
                 kfree(bqt);
+
         }
 
+        return retval;
+}
+
+/**
+ * __blk_queue_free_tags - release tag maintenance info
+ * @q:  the request queue for the device
+ *
+ *  Notes:
+ *    blk_cleanup_queue() will take care of calling this function, if tagging
+ *    has been used. So there's no need to call this directly.
+ **/
+static void __blk_queue_free_tags(request_queue_t *q)
+{
+        struct blk_queue_tag *bqt = q->queue_tags;
+
+        if (!bqt)
+                return;
+
+        __blk_free_tags(bqt);
+
         q->queue_tags = NULL;
         q->queue_flags &= ~(1 << QUEUE_FLAG_QUEUED);
 }
 
+
+/**
+ * blk_free_tags - release a given set of tag maintenance info
+ * @bqt:        the tag map to free
+ *
+ * For externally managed @bqt@ frees the map.  Callers of this
+ * function must guarantee to have released all the queues that
+ * might have been using this tag map.
+ */
+void blk_free_tags(struct blk_queue_tag *bqt)
+{
+        if (unlikely(!__blk_free_tags(bqt)))
+                BUG();
+}
+EXPORT_SYMBOL(blk_free_tags);
+
 /**
  * blk_queue_free_tags - release tag maintenance info
  * @q:  the request queue for the device
@@ -901,7 +935,7 @@ init_tag_map(request_queue_t *q, struct blk_queue_tag *tags, int depth)
         unsigned long *tag_map;
         int nr_ulongs;
 
-        if (depth > q->nr_requests * 2) {
+        if (q && depth > q->nr_requests * 2) {
                 depth = q->nr_requests * 2;
                 printk(KERN_ERR "%s: adjusted depth to %d\n",
                                 __FUNCTION__, depth);
@@ -927,6 +961,38 @@ fail:
         return -ENOMEM;
 }
 
+static struct blk_queue_tag *__blk_queue_init_tags(struct request_queue *q,
+                                                   int depth)
+{
+        struct blk_queue_tag *tags;
+
+        tags = kmalloc(sizeof(struct blk_queue_tag), GFP_ATOMIC);
+        if (!tags)
+                goto fail;
+
+        if (init_tag_map(q, tags, depth))
+                goto fail;
+
+        INIT_LIST_HEAD(&tags->busy_list);
+        tags->busy = 0;
+        atomic_set(&tags->refcnt, 1);
+        return tags;
+fail:
+        kfree(tags);
+        return NULL;
+}
+
+/**
+ * blk_init_tags - initialize the tag info for an external tag map
+ * @depth:      the maximum queue depth supported
+ * @tags: the tag to use
+ **/
+struct blk_queue_tag *blk_init_tags(int depth)
+{
+        return __blk_queue_init_tags(NULL, depth);
+}
+EXPORT_SYMBOL(blk_init_tags);
+
 /**
  * blk_queue_init_tags - initialize the queue tag info
  * @q:  the request queue for the device
@@ -941,16 +1007,10 @@ int blk_queue_init_tags(request_queue_t *q, int depth,
         BUG_ON(tags && q->queue_tags && tags != q->queue_tags);
 
         if (!tags && !q->queue_tags) {
-                tags = kmalloc(sizeof(struct blk_queue_tag), GFP_ATOMIC);
-                if (!tags)
-                        goto fail;
+                tags = __blk_queue_init_tags(q, depth);
 
-                if (init_tag_map(q, tags, depth))
+                if (!tags)
                         goto fail;
-
-                INIT_LIST_HEAD(&tags->busy_list);
-                tags->busy = 0;
-                atomic_set(&tags->refcnt, 1);
         } else if (q->queue_tags) {
                 if ((rc = blk_queue_resize_tags(q, depth)))
                         return rc;
@@ -1002,6 +1062,13 @@ int blk_queue_resize_tags(request_queue_t *q, int new_depth)
         }
 
         /*
+         * Currently cannot replace a shared tag map with a new
+         * one, so error out if this is the case
+         */
+        if (atomic_read(&bqt->refcnt) != 1)
+                return -EBUSY;
+
+        /*
          * save the old state info, so we can copy it back
          */
         tag_index = bqt->tag_index;
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index aafe82788b4e..427b0d61be6c 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -746,6 +746,8 @@ extern void blk_queue_free_tags(request_queue_t *);
 extern int blk_queue_resize_tags(request_queue_t *, int);
 extern void blk_queue_invalidate_tags(request_queue_t *);
 extern long blk_congestion_wait(int rw, long timeout);
+extern struct blk_queue_tag *blk_init_tags(int);
+extern void blk_free_tags(struct blk_queue_tag *);
 
 extern void blk_rq_bio_prep(request_queue_t *, struct request *, struct bio *);
 extern int blkdev_issue_flush(struct block_device *, sector_t *);