On 9/25/19 4:36 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> In xfs_iterate_filemaps, we query the number of bmaps for a given file
> that we're going to iterate, so feed that information to bmap so that
> the kernel won't waste time allocating in-kernel memory unnecessarily.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> ---
> scrub/filemap.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>
> diff --git a/scrub/filemap.c b/scrub/filemap.c
> index bdc6d8f9..aaaa0521 100644
> --- a/scrub/filemap.c
> +++ b/scrub/filemap.c
> @@ -71,7 +71,6 @@ xfs_iterate_filemaps(
> map->bmv_length = ULLONG_MAX;
> else
> map->bmv_length = BTOBB(key->bm_length);
> - map->bmv_count = BMAP_NR;
> map->bmv_iflags = BMV_IF_NO_DMAPI_READ | BMV_IF_PREALLOC |
> BMV_IF_NO_HOLES;
> switch (whichfork) {
> @@ -96,6 +95,7 @@ xfs_iterate_filemaps(
> moveon = false;
> goto out;
> }
> + map->bmv_count = min(fsx.fsx_nextents + 2, BMAP_NR);
Was going to ask you to document the magical +2 here but IRC discussion suggests
that it is in case fsx_nextents is 0, and we need to send in a count of at least
2 (header + 1 structure?)
But if there are no extents, what are we trying to map in the loop below,
anyway?
>
> while ((error = ioctl(fd, XFS_IOC_GETBMAPX, map)) == 0) {
> for (i = 0, p = &map[i + 1]; i < map->bmv_entries; i++, p++) {
>
