2010-07-06 04:33:16

by Stephen Rothwell

Subject: linux-next: manual merge of the device-mapper tree with the block tree

Hi Alasdair,

Today's linux-next merge of the device-mapper tree got a conflict in
drivers/md/dm.c between commits 9add80db6089272d6bf13ef6b5dc7b3ddda1a887
("dm: stop using q->prepare_flush_fn") and
5e27e27e73b5bff903b3c30ffd5a0e17eb95c087 ("block: remove
q->prepare_flush_fn completely") from the block tree and commit
90c50ea6a71bcb1bdf1482007932cc7fb0902455
("dm-do-not-initialise-full-request-queue-when-bio-based") from the
device-mapper tree.

I fixed it up (see below) and can carry the fix as necessary.
--
Cheers,
Stephen Rothwell [email protected]

diff --cc drivers/md/dm.c
index d505a96,c49818a..0000000
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@@ -2116,6 -2165,73 +2154,72 @@@ int dm_create(int minor, struct mapped_
return 0;
}

+ /*
+ * Functions to manage md->type.
+ * All are required to hold md->type_lock.
+ */
+ void dm_lock_md_type(struct mapped_device *md)
+ {
+ mutex_lock(&md->type_lock);
+ }
+
+ void dm_unlock_md_type(struct mapped_device *md)
+ {
+ mutex_unlock(&md->type_lock);
+ }
+
+ void dm_set_md_type(struct mapped_device *md, unsigned type)
+ {
+ md->type = type;
+ }
+
+ unsigned dm_get_md_type(struct mapped_device *md)
+ {
+ return md->type;
+ }
+
+ /*
+ * Fully initialize a request-based queue (->elevator, ->request_fn, etc).
+ */
+ static int dm_init_request_based_queue(struct mapped_device *md)
+ {
+ struct request_queue *q = NULL;
+
+ if (md->queue->elevator)
+ return 1;
+
+ /* Fully initialize the queue */
+ q = blk_init_allocated_queue(md->queue, dm_request_fn, NULL);
+ if (!q)
+ return 0;
+
+ md->queue = q;
+ md->saved_make_request_fn = md->queue->make_request_fn;
+ dm_init_md_queue(md);
+ blk_queue_softirq_done(md->queue, dm_softirq_done);
+ blk_queue_prep_rq(md->queue, dm_prep_fn);
+ blk_queue_lld_busy(md->queue, dm_lld_busy);
- blk_queue_ordered(md->queue, QUEUE_ORDERED_DRAIN_FLUSH,
- dm_rq_prepare_flush);
++ blk_queue_ordered(md->queue, QUEUE_ORDERED_DRAIN_FLUSH);
+
+ elv_register_queue(md->queue);
+
+ return 1;
+ }
+
+ /*
+ * Setup the DM device's queue based on md's type
+ */
+ int dm_setup_md_queue(struct mapped_device *md)
+ {
+ if ((dm_get_md_type(md) == DM_TYPE_REQUEST_BASED) &&
+ !dm_init_request_based_queue(md)) {
+ DMWARN("Cannot initialize queue for request-based mapped device");
+ return -EINVAL;
+ }
+
+ return 0;
+ }
+
static struct mapped_device *dm_find_md(dev_t dev)
{
struct mapped_device *md;

2010-07-06 14:30:20

by Mike Snitzer

Subject: Re: linux-next: manual merge of the device-mapper tree with the block tree

On Tue, Jul 06 2010 at 12:33am -0400,
Stephen Rothwell <[email protected]> wrote:

> Hi Alasdair,
>
> Today's linux-next merge of the device-mapper tree got a conflict in
> drivers/md/dm.c between commits 9add80db6089272d6bf13ef6b5dc7b3ddda1a887
> ("dm: stop using q->prepare_flush_fn") and
> 5e27e27e73b5bff903b3c30ffd5a0e17eb95c087 ("block: remove
> q->prepare_flush_fn completely") from the block tree and commit
> 90c50ea6a71bcb1bdf1482007932cc7fb0902455
> ("dm-do-not-initialise-full-request-queue-when-bio-based") from the
> device-mapper tree.
>
> I fixed it up (see below) and can carry the fix as necessary.

Hi Stephen,

Your fix is correct, thanks for carrying this until Alasdair gets a
chance to pick it up.

Mike