On 6/17/25 13:55, Jianan Huang wrote:
> When fewer pages are read, nr_pages may be smaller than nr_cpages. Due
> to the nr_vecs limit, the compressed pages will be split into multiple
> bios and then merged at the block level. In this case, nr_cpages should
> be used to pre-allocate bvecs.
>
> Signed-off-by: Jianan Huang <[email protected]>
> Signed-off-by: Sheng Yong <[email protected]>
> ---
> fs/f2fs/data.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index 31e892842625..c7773b09d83f 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -2303,7 +2303,8 @@ int f2fs_read_multi_pages(struct compress_ctx *cc,
> struct bio **bio_ret,
> }
>
> if (!bio) {
> - bio = f2fs_grab_read_bio(inode, blkaddr, nr_pages,
> + bio = f2fs_grab_read_bio(inode, blkaddr,
> + max(nr_pages, cc->nr_cpages) - i,
Hi Jianan,
e.g.
User wants to read page [1, 5],
page #1,2,3,4 locates in compressed block #1000,1001,1003,
page #5 locate in compressed block #1004,1005
It submits first bio w/ block #1000,1001
It allocates second bio w/ size of max(nr_pages=1, nr_cpages=3) - 2 = 1 ?
However block #1003 and block #1004,1005 can be readed in one bio, we
should allocate larger bio for last continuous blocks which cross clusters.
> f2fs_ra_op_flags(rac),
> folio->index, for_write);
> if (IS_ERR(bio)) {
> @@ -2373,7 +2374,6 @@ static int f2fs_mpage_readpages(struct inode *inode,
> pgoff_t index;
> #endif
> unsigned nr_pages = rac ? readahead_count(rac) : 1;
> - unsigned max_nr_pages = nr_pages;
Maybe we can align both start and end of read range w/ cluster_size, and use
start and end for max_nr_pages calculation, then pass it to
f2fs_read_{multi,single}_pages(), something like this?
max_nr_pages = round_up(end_idx, cluster_size) -
round_down(start_idx, cluster_size);
Its size should always cover size of all cpage and/or rpage.
Thanks,
> int ret = 0;
>
> map.m_pblk = 0;
> @@ -2400,7 +2400,7 @@ static int f2fs_mpage_readpages(struct inode *inode,
> /* there are remained compressed pages, submit them */
> if (!f2fs_cluster_can_merge_page(&cc, index)) {
> ret = f2fs_read_multi_pages(&cc, &bio,
> - max_nr_pages,
> + nr_pages,
> &last_block_in_bio,
> rac, false);
> f2fs_destroy_compress_ctx(&cc, false);
> @@ -2432,7 +2432,7 @@ static int f2fs_mpage_readpages(struct inode *inode,
> read_single_page:
> #endif
>
> - ret = f2fs_read_single_page(inode, folio, max_nr_pages, &map,
> + ret = f2fs_read_single_page(inode, folio, nr_pages, &map,
> &bio, &last_block_in_bio, rac);
> if (ret) {
> #ifdef CONFIG_F2FS_FS_COMPRESSION
> @@ -2450,7 +2450,7 @@ static int f2fs_mpage_readpages(struct inode *inode,
> /* last page */
> if (nr_pages == 1 && !f2fs_cluster_is_empty(&cc)) {
> ret = f2fs_read_multi_pages(&cc, &bio,
> - max_nr_pages,
> + nr_pages,
> &last_block_in_bio,
> rac, false);
> f2fs_destroy_compress_ctx(&cc, false);
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
