Re: [PATCH v2] squashfs: cache partial compressed blocks

[Christoph Hellwig <hch@lst.de>]

On Fri, May 12, 2023 at 03:18:05PM +0200, Vincent Whitchurch wrote:
> +static int squashfs_bio_read_cached(struct bio *fullbio, struct address_space *cache_mapping,
> +				    u64 index, int length, u64 read_start, u64 read_end,
> +				    int page_count)

Please avoid the unreadable formatting and do something like:

static int squashfs_bio_read_cached(struct bio *fullbio,
		struct address_space *cache_mapping, u64 index, int length,
		u64 read_start, u64 read_end, int page_count)


> +		if (!bio || idx != prev_io_idx + 1) {
> +			unsigned int remaining_pages;
> +			unsigned int this_nr_pages;
> +
> +submit_and_retry:
> +			remaining_pages = page_count - idx;
> +			this_nr_pages = min(remaining_pages, BIO_MAX_VECS);
> +			bio = blk_next_bio(bio, bdev, this_nr_pages, REQ_OP_READ,
> +					   GFP_NOIO);
> +			bio->bi_iter.bi_sector = fullbio->bi_iter.bi_sector +
> +						 idx * (PAGE_SIZE / SECTOR_SIZE);
> +		}
> +
> +		retlen = bio_add_page(bio, bv->bv_page, bv->bv_len, bv->bv_offset);
> +		if (retlen != bv->bv_len)
> +			goto submit_and_retry;

Adding data payload to a bio while passing full bio_vec is usually
a sign that instead you should do a (partial) clone of the bio instead.
I think this is such a case.

>  static int squashfs_bio_read(struct super_block *sb, u64 index, int length,
>  			     struct bio **biop, int *block_offset)
>  {
>  	struct squashfs_sb_info *msblk = sb->s_fs_info;
> +	struct inode *cache_inode = msblk->cache_inode;
> +	struct address_space *cache_mapping = cache_inode ? cache_inode->i_mapping : NULL;

Unless I'm badly misreading the squashfs_fill_super changes,
cache_inode can't ever be NULL here or anywhere else in the I/O
code.