Re: [PATCH] ext4: improve ext4lazyinit scalability V2

[Jan Kara <jack@suse.cz>]

On Mon 05-09-16 23:39:54, Ted Tso wrote:
> Dmitry, thanks for the patch, and Jan, thanks for the review.
> 
> This is what I've added to the ext4 tree, which should reflect all of
> the comments which Jan made.

Hum, one comment still:

> +			if (down_read_trylock(&elr->lr_super->s_umount)) {
> +				if (sb_start_write_trylock(elr->lr_super)) {
> +					progress = 1;
> +					/*
> +					 * We hold sb->s_umount, sb can not
> +					 * be removed from the list, it is
> +					 * now safe to drop li_list_mtx
> +					 */
> +					mutex_unlock(&eli->li_list_mtx);
> +					err = ext4_run_li_request(elr);
> +					sb_end_write(elr->lr_super);
> +					mutex_lock(&eli->li_list_mtx);
> +					n = pos->next;
>  				}
> +				up_read((&elr->lr_super->s_umount));
> +			}
> +			/* error, remove the lazy_init job */
> +			if (err) {
> +				ext4_remove_li_request(elr);
> +				continue;
> +			}
> +			if (!progress) {
> +				elr->lr_next_sched = jiffies +
> +					(prandom_u32()
> +					 % (EXT4_DEF_LI_MAX_START_DELAY * HZ));
> +				if (time_before(elr->lr_next_sched,
> +						next_wakeup))
> +					next_wakeup = elr->lr_next_sched;
>  			}
> -
> -			if (time_before(elr->lr_next_sched, next_wakeup))
> -				next_wakeup = elr->lr_next_sched;
>  		}

ext4_run_li_request() can also update elr->lr_next_sched. So we need to
update next_wakeup even in progress == 1 case (i.e., I'd leave the update
of next_wakeup as is in the old code...). Otherwise the patch looks good.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR