Hi guys, I've had this patch sitting on my disk for a few months, and haven't done anything more than test it. Before I stuff it into git, and send it to the kernel, would you guys have a look over it and see whether it's technically correct. The patch simply causes a kobject "change" event to be emitted on devmapper block devices when they are renamed. Scott -- Scott James Remnant scott@xxxxxxxxxx
diff -ruNp linux-source-2.6.22~/drivers/md/dm-ioctl.c linux-source-2.6.22/drivers/md/dm-ioctl.c --- linux-source-2.6.22~/drivers/md/dm-ioctl.c 2007-07-10 17:54:44.000000000 +0100 +++ linux-source-2.6.22/drivers/md/dm-ioctl.c 2007-07-10 22:01:09.000000000 +0100 @@ -323,6 +323,8 @@ static int dm_hash_rename(const char *ol hc->name = new_name; list_add(&hc->name_list, _name_buckets + hash_str(new_name)); + dm_changed(hc->md); + /* * Wake up any dm event waiters. */ diff -ruNp linux-source-2.6.22~/drivers/md/dm.c linux-source-2.6.22/drivers/md/dm.c --- linux-source-2.6.22~/drivers/md/dm.c 2007-05-13 19:55:50.000000000 +0100 +++ linux-source-2.6.22/drivers/md/dm.c 2007-07-10 22:01:09.000000000 +0100 @@ -1542,6 +1542,11 @@ int dm_noflush_suspending(struct dm_targ } EXPORT_SYMBOL_GPL(dm_noflush_suspending); +void dm_changed(struct mapped_device *md) +{ + kobject_uevent(&md->disk->kobj, KOBJ_CHANGE); +} + static struct block_device_operations dm_blk_dops = { .open = dm_blk_open, .release = dm_blk_close, diff -ruNp linux-source-2.6.22~/drivers/md/dm.h linux-source-2.6.22/drivers/md/dm.h --- linux-source-2.6.22~/drivers/md/dm.h 2007-05-13 19:55:50.000000000 +0100 +++ linux-source-2.6.22/drivers/md/dm.h 2007-07-10 22:01:09.000000000 +0100 @@ -151,4 +151,6 @@ union map_info *dm_get_mapinfo(struct bi int dm_open_count(struct mapped_device *md); int dm_lock_for_deletion(struct mapped_device *md); +void dm_changed(struct mapped_device *md); + #endif
Attachment:
signature.asc
Description: This is a digitally signed message part
-- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel