| Message ID | 20241130111132.1359138-1-alexjlzheng@tencent.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [RESEND,v2] xfs: fix the entry condition of exact EOF block allocation optimization | expand |
On Sat, Nov 30, 2024 at 07:11:32PM +0800, Jinliang Zheng wrote: > When we call create(), lseek() and write() sequentially, offset != 0 > cannot be used as a judgment condition for whether the file already > has extents. > > Furthermore, when xfs_bmap_adjacent() has not given a better blkno, > it is not necessary to use exact EOF block allocation. > > Signed-off-by: Jinliang Zheng <alexjlzheng@tencent.com> > --- > Changelog: > - V2: Fix the entry condition > - V1: https://lore.kernel.org/linux-xfs/ZyFJm7xg7Msd6eVr@dread.disaster.area/T/#t > --- > fs/xfs/libxfs/xfs_bmap.c | 12 +++++++----- > 1 file changed, 7 insertions(+), 5 deletions(-) > > diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c > index 36dd08d13293..c1e5372b6b2e 100644 > --- a/fs/xfs/libxfs/xfs_bmap.c > +++ b/fs/xfs/libxfs/xfs_bmap.c > @@ -3531,12 +3531,14 @@ xfs_bmap_btalloc_at_eof( > int error; > > /* > - * If there are already extents in the file, try an exact EOF block > - * allocation to extend the file as a contiguous extent. If that fails, > - * or it's the first allocation in a file, just try for a stripe aligned > - * allocation. > + * If there are already extents in the file, and xfs_bmap_adjacent() has > + * given a better blkno, try an exact EOF block allocation to extend the > + * file as a contiguous extent. If that fails, or it's the first > + * allocation in a file, just try for a stripe aligned allocation. > */ > - if (ap->offset) { > + if (ap->prev.br_startoff != NULLFILEOFF && > + !isnullstartblock(ap->prev.br_startblock) && > + xfs_bmap_adjacent_valid(ap, ap->blkno, ap->prev.br_startblock)) { There's no need for calling xfs_bmap_adjacent_valid() here - we know that ap->blkno is valid because the bounds checking has already been done by xfs_bmap_adjacent(). Actually, for another patch, the bounds checking in xfs_bmap_adjacent_valid() is incorrect. What happens if the last AG is a runt? i.e. it open codes xfs_verify_fsbno() and gets it wrong. -Dave.
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 36dd08d13293..c1e5372b6b2e 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -3531,12 +3531,14 @@ xfs_bmap_btalloc_at_eof( int error; /* - * If there are already extents in the file, try an exact EOF block - * allocation to extend the file as a contiguous extent. If that fails, - * or it's the first allocation in a file, just try for a stripe aligned - * allocation. + * If there are already extents in the file, and xfs_bmap_adjacent() has + * given a better blkno, try an exact EOF block allocation to extend the + * file as a contiguous extent. If that fails, or it's the first + * allocation in a file, just try for a stripe aligned allocation. */ - if (ap->offset) { + if (ap->prev.br_startoff != NULLFILEOFF && + !isnullstartblock(ap->prev.br_startblock) && + xfs_bmap_adjacent_valid(ap, ap->blkno, ap->prev.br_startblock)) { xfs_extlen_t nextminlen = 0; /*
When we call create(), lseek() and write() sequentially, offset != 0 cannot be used as a judgment condition for whether the file already has extents. Furthermore, when xfs_bmap_adjacent() has not given a better blkno, it is not necessary to use exact EOF block allocation. Signed-off-by: Jinliang Zheng <alexjlzheng@tencent.com> --- Changelog: - V2: Fix the entry condition - V1: https://lore.kernel.org/linux-xfs/ZyFJm7xg7Msd6eVr@dread.disaster.area/T/#t --- fs/xfs/libxfs/xfs_bmap.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-)