Re: How to cleanly shut down a block device

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Tue, Nov 14, 2006 at 12:33:47PM +0100, Pierre Ossman wrote:
> Jens Axboe wrote:
> > What do you mean by "killing off the queue"? As long as the queue can be
> > gotten at, it needs to remain valid. That is what the references are
> > for.
> >   
> 
> I do:
> 
> del_gendisk();
> (wait for queue to become empty, i.e. elv_next_request() == NULL)
> blk_cleanup_queue();
> 
> and then assume that the request function will no longer be called for
> this queue.
> 
> Suggested patch:
> 
> diff --git a/drivers/mmc/mmc_block.c b/drivers/mmc/mmc_block.c
> index f9027c8..5025abe 100644
> --- a/drivers/mmc/mmc_block.c
> +++ b/drivers/mmc/mmc_block.c
> @@ -83,7 +83,6 @@ static void mmc_blk_put(struct mmc_blk_d
>         md->usage--;
>         if (md->usage == 0) {
>                 put_disk(md->disk);
> -               mmc_cleanup_queue(&md->queue);
>                 kfree(md);
>         }
>         mutex_unlock(&open_lock);
> @@ -553,12 +552,11 @@ static void mmc_blk_remove(struct mmc_ca
>         if (md) {
>                 int devidx;
>  
> +               /* Stop new requests from getting into the queue */
>                 del_gendisk(md->disk);
>  
> -               /*
> -                * I think this is needed.
> -                */
> -               md->disk->queue = NULL;
> +               /* Then flush out any already in there */
> +               mmc_cleanup_queue(&md->queue);
>  
>                 devidx = md->disk->first_minor >> MMC_SHIFT;
>                 __clear_bit(devidx, dev_use);

You now have a disk which may be in use (and a block device) for which
there is no queue.  I couldn't see any locking what so ever in del_gendisk
so it's quite possible that the queue could still be referenced while
the disk is still open.

I also don't think del_gendisk() is sufficient to ensure that no new
requests appear in the queue, which is why I'm setting md->disk->queue
to NULL there.

But shug, I don't know the block layer.  Jens is your best bet.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/