On Tue, June 9, 2009 2:31 pm, Martin K. Petersen wrote:
>
> Switch MD over to the new disk_stack_limits() function which checks for
> aligment and adjusts preferred I/O sizes when stacking.
>
> Also indicate preferred I/O sizes where applicable.
>
> Signed-off-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx>
>
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index ae12cea..dfd18df 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -2044,7 +2044,7 @@ raid10_size(mddev_t *mddev, sector_t sectors, int
> raid_disks)
> static int run(mddev_t *mddev)
> {
> conf_t *conf;
> - int i, disk_idx;
> + int i, disk_idx, chunk_size;
> mirror_info_t *disk;
> mdk_rdev_t *rdev;
> int nc, fc, fo;
> @@ -2130,6 +2130,10 @@ static int run(mddev_t *mddev)
> spin_lock_init(&conf->device_lock);
> mddev->queue->queue_lock = &conf->device_lock;
>
> + chunk_size = mddev->chunk_sectors << 9;
> + blk_queue_io_min(mddev->queue, chunk_size);
> + blk_queue_io_opt(mddev->queue, chunk_size * (mddev->raid_disks >> 1));
">> 1" is wrong. A raid10 may have more than 2 copies for each block.
This calculation needs to be a bit more subtle and take about of the
particular layout (near, offset, or far).
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index 41dc36e..ea8d3fd 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -4451,7 +4451,7 @@ static raid5_conf_t *setup_conf(mddev_t *mddev)
> static int run(mddev_t *mddev)
> {
> raid5_conf_t *conf;
> - int working_disks = 0;
> + int working_disks = 0, chunk_size;
> mdk_rdev_t *rdev;
>
> if (mddev->recovery_cp != MaxSector)
> @@ -4606,6 +4606,19 @@ static int run(mddev_t *mddev)
> md_set_array_sectors(mddev, raid5_size(mddev, 0, 0));
>
> blk_queue_merge_bvec(mddev->queue, raid5_mergeable_bvec);
> + chunk_size = mddev->chunk_sectors << 9;
> + blk_queue_io_min(mddev->queue, chunk_size);
> +
> + if (mddev->level == 5)
> + blk_queue_io_opt(mddev->queue,
> + chunk_size * (mddev->raid_disks - 1));
> + else
> + blk_queue_io_opt(mddev->queue,
> + chunk_size * (mddev->raid_disks - 2));
> +
You seem to be treating raid4 like raid6 :-)
probably what you want is
blk_queue_io_opt(mddev->queue,
chunk_size * (conf->raid_disks - conf->max_degraded));
I'll be a closer look at the patch in the next day or two.
Thanks,
NeilBrown
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
