Re: [PATCH 1/3] md:Add support for Raid0->Raid5 takeover

[Neil Brown <neilb@suse.de>]

On Fri, 29 Jan 2010 14:54:03 +0000
"Trela, Maciej" <Maciej.Trela@intel.com> wrote:

> Hello Neil, I'm sending patches with new takeover transitions proposal:
> 0->5(4), 5(4)->0, 10->0, 0->10.
> 
> Regards,
> Maciek
> 
> 
> Signed-off-by: Maciej Trela <maciej.trela@intel.com>
> ---
>  drivers/md/raid5.c |   50 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 files changed, 50 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index e84204e..0fe78ea 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -5614,6 +5614,39 @@ static void raid5_quiesce(mddev_t *mddev, int state)
>  }
>  
>  
> +static void *raid5_takeover_raid0(mddev_t *mddev)
> +{
> +	int chunksect;
> +	
> +	if (mddev->degraded > 0) {
> +	    printk(KERN_ERR "ERROR: Raid0 with degraded disks!: %d\n", mddev->degraded);
> +		return ERR_PTR(-EINVAL);
> +	}
> +	
> +	chunksect = 64*2; /* 64K by default */
> +
> +	/* The array must be an exact multiple of chunksize */
> +	while (chunksect && (mddev->array_sectors & (chunksect-1)))
> +		chunksect >>= 1;
> +
> +	if ((chunksect<<9) < STRIPE_SIZE) {
> +		/* array size does not allow a suitable chunk size */
> +		return ERR_PTR(-EINVAL);
> +	}

This is wrong.  The chunksize for RAID5 must be exactly the same as the
chunksize for RAID0.  There is no place for guessing.



> +	
> +	mddev->new_level = 5;
> +	mddev->new_layout = ALGORITHM_PARITY_N;
> +	mddev->new_chunk_sectors = chunksect;
> +	mddev->degraded++;
> +	mddev->raid_disks++;
> +	mddev->delta_disks++;
> +	/* make sure it will be not marked as dirty */
> +	mddev->recovery_cp = MaxSector;
> +
> +	return setup_conf(mddev);
> +}
> +
> +
>  static void *raid5_takeover_raid1(mddev_t *mddev)
>  {
>  	int chunksect;
> @@ -5737,12 +5770,29 @@ static int raid6_check_reshape(mddev_t *mddev)
>  
>  static void *raid5_takeover(mddev_t *mddev)
>  {
> +	mdk_rdev_t *rdev;
> +	sector_t sectors, dev_sectors;
> +
>  	/* raid5 can take over:
>  	 *  raid0 - if all devices are the same - make it a raid4 layout
>  	 *  raid1 - if there are two drives.  We need to know the chunk size
>  	 *  raid4 - trivial - just use a raid4 layout.
>  	 *  raid6 - Providing it is a *_6 layout
>  	 */
> +	if (mddev->level == 0) {
> +		/* make sure all devices are the same (only one zone is supported) */
> +		sectors = mddev->dev_sectors;
> +		sector_div(sectors, mddev->chunk_sectors);
> +		list_for_each_entry(rdev, &mddev->disks, same_set) {
> +			dev_sectors = rdev->sectors;
> +			sector_div(dev_sectors, mddev->chunk_sectors);
> +			if (dev_sectors != sectors) {
> +				printk("error: cannot takeover raid 0 with different dev sizes.\n");
> +				return ERR_PTR(-EINVAL);
> +			}
> +		}
> +		return raid5_takeover_raid0(mddev);
> +	}

I would rather you just look inside the raid0_private_data to see tha
nr_strip_zones is zero.


Thanks,
NeilBrown