[PATCH 1/1] md/raid5: allow resync-rebuild requests to run multi-threaded

[PATCH 1/1] md/raid5: allow resync-rebuild requests to run multi-threaded

[Eivind Sarto]

[-- Attachment #1: raid5.patch --]
[-- Type: application/octet-stream, Size: 886 bytes --]

Author: Eivind Sarto <eivind.sarto@gmail.com>
Date: Mon Feb 10 17:42:09 2014

    md/raid5: allow resync/rebuild requests to run multi-threaded

    Sync operations are currently handled within the sync thread, making the
    execution single-threaded.  This patch will allow the processing to be done
    parallel by raid5d (and a pool of kworkers), significantly increasing
    rebuild/resync throughput.

    Signed-off-by: Eivind Sarto <eivind.sarto@gmail.com>

--- a/drivers/md/raid5.c	2014-02-10 17:18:43.000000000 -0800
+++ b/drivers/md/raid5.c	2014-02-10 17:21:49.000000000 -0800
@@ -5026,8 +5026,8 @@ static inline sector_t sync_request(stru
 	bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, still_degraded);
 
 	set_bit(STRIPE_SYNC_REQUESTED, &sh->state);
+	set_bit(STRIPE_HANDLE, &sh->state);
 
-	handle_stripe(sh);
 	release_stripe(sh);
 
 	return STRIPE_SECTORS;

Re: [PATCH 1/1] md/raid5: allow resync-rebuild requests to run multi-threaded

[NeilBrown]

[-- Attachment #1: Type: text/plain, Size: 673 bytes --]


Hi,
 thanks for the patch.
 When you include a patch as an attachment, I find it easier to work with if
 it is "text/plain" rather than "application/octet-stream".

 As your patch is justified as a performance improvement, please include
 some measurements together with a rough description of the hardware (one or
 two sentences).

 Also, the performance improvement requires that multi-threading is enabled,
 which isn't the default.  Please measure the effect of the patch when
 multi-threading not enabled.
 Maybe we should make the change you suggest be conditional on threading
 being enabled, but we can't know without numbers.

Thanks,
NeilBrown

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]