Re: When allocating resync pages,judge the queue limit of mddev.

[NeilBrown <neilb@suse.de>]

[-- Attachment #1: Type: text/plain, Size: 4177 bytes --]

On Fri, 30 Mar 2012 13:53:05 +0800 "majianpeng" <majianpeng@gmail.com> wrote:

> >From d5c0ad3ac03c805747f71338d30282f9f8d8d953 Mon Sep 17 00:00:00 2001
> From: majianpeng <majianpeng@gmail.com>
> Date: Fri, 30 Mar 2012 13:37:42 +0800
> Subject: [PATCH] md/raid1:When allocating resync pages,judge the queue limit  of mddev.
>  When max_sectors of mddev is smaller than
>  RESYNC_PAGES(at present 64k),then: 1:alloc enough resync
>  pags. 2:when do bio_add_page can goto bio_fill.
> 
> 
> Signed-off-by: majianpeng <majianpeng@gmail.com>
> ---
>  drivers/md/raid1.c |   18 +++++++++++-------
>  1 files changed, 11 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
> index 4a40a20..cd100e6 100644
> --- a/drivers/md/raid1.c
> +++ b/drivers/md/raid1.c
> @@ -82,16 +82,19 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data)
>  	struct r1bio *r1_bio;
>  	struct bio *bio;
>  	int i, j;
> +	int sync_pages;
>  
>  	r1_bio = r1bio_pool_alloc(gfp_flags, pi);
>  	if (!r1_bio)
>  		return NULL;
>  
> +	sync_pages = queue_max_sectors(pi->mddev->gendisk->queue) << 9;

I don't thing '<< 9' is correct.  Maybe '<< (PAGE_SHIFT - 9)'.

> +	sync_pages = min(RESYNC_PAGES, (sync_pages + PAGE_SIZE - 1) / PAGE_SIZE);

And this is definitely wrong. you don't want to round up the number of page
to be a multiple of the number of bytes in a page....

What is the value of this patch?  It allocates a few fewer pages for the rare
case where the device cannot accept 64K IO requests?

I don't think it is really worth it, and given the obvious bugs, I won't
apply it.

Thanks,
NeilBrown


>  	/*
>  	 * Allocate bios : 1 for reading, n-1 for writing
>  	 */
>  	for (j = pi->raid_disks ; j-- ; ) {
> -		bio = bio_kmalloc(gfp_flags, RESYNC_PAGES);
> +		bio = bio_kmalloc(gfp_flags, sync_pages);
>  		if (!bio)
>  			goto out_free_bio;
>  		r1_bio->bios[j] = bio;
> @@ -108,7 +111,7 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data)
>  		j = 1;
>  	while(j--) {
>  		bio = r1_bio->bios[j];
> -		for (i = 0; i < RESYNC_PAGES; i++) {
> +		for (i = 0; i < sync_pages; i++) {
>  			page = alloc_page(gfp_flags);
>  			if (unlikely(!page))
>  				goto out_free_pages;
> @@ -119,8 +122,8 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data)
>  	}
>  	/* If not user-requests, copy the page pointers to all bios */
>  	if (!test_bit(MD_RECOVERY_REQUESTED, &pi->mddev->recovery)) {
> -		for (i=0; i<RESYNC_PAGES ; i++)
> -			for (j=1; j<pi->raid_disks; j++)
> +		for (i = 0; i < sync_pages ; i++)
> +			for (j = 1; j < pi->raid_disks; j++)
>  				r1_bio->bios[j]->bi_io_vec[i].bv_page =
>  					r1_bio->bios[0]->bi_io_vec[i].bv_page;
>  	}
> @@ -147,7 +150,7 @@ static void r1buf_pool_free(void *__r1_bio, void *data)
>  	int i,j;
>  	struct r1bio *r1bio = __r1_bio;
>  
> -	for (i = 0; i < RESYNC_PAGES; i++)
> +	for (i = 0; i < r1bio->bios[0]->bi_max_vecs; i++)
>  		for (j = pi->raid_disks; j-- ;) {
>  			if (j == 0 ||
>  			    r1bio->bios[j]->bi_io_vec[i].bv_page !=
> @@ -2243,7 +2246,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
>  	int write_targets = 0, read_targets = 0;
>  	sector_t sync_blocks;
>  	int still_degraded = 0;
> -	int good_sectors = RESYNC_SECTORS;
> +	int good_sectors;
>  	int min_bad = 0; /* number of sectors that are bad in all devices */
>  
>  	if (!conf->r1buf_pool)
> @@ -2296,6 +2299,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
>  	r1_bio = mempool_alloc(conf->r1buf_pool, GFP_NOIO);
>  	raise_barrier(conf);
>  
> +	good_sectors = r1_bio->bios[0]->bi_max_vecs << 9;
>  	conf->next_resync = sector_nr;
>  
>  	rcu_read_lock();
> @@ -2477,7 +2481,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
>  		nr_sectors += len>>9;
>  		sector_nr += len>>9;
>  		sync_blocks -= (len>>9);
> -	} while (r1_bio->bios[disk]->bi_vcnt < RESYNC_PAGES);
> +	} while (r1_bio->bios[disk]->bi_vcnt < r1_bio->bios[0]->bi_max_vecs);
>   bio_full:
>  	r1_bio->sectors = nr_sectors;
>  


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