* [PATCH 0/2] xfs: online scrub fixes @ 2018-04-02 19:55 Darrick J. Wong 2018-04-02 19:55 ` [PATCH 1/2] xfs: skip scrub xref if corruption already noted Darrick J. Wong 2018-04-02 19:55 ` [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk Darrick J. Wong 0 siblings, 2 replies; 7+ messages in thread From: Darrick J. Wong @ 2018-04-02 19:55 UTC (permalink / raw) To: darrick.wong; +Cc: linux-xfs Hi all, Here's a couple more patches to fix some inefficiencies in the scrub code. The first patch skips cross-referencing once a cross-referencing problem has been found; and the second patch moves the "walk all AGFL blocks" function to libxfs so that (later) the repair code will be able to use it. If you're going to start using this mess, you probably ought to just pull from my git trees. The kernel patches[1] should apply against 4.16. xfsprogs[2] and xfstests[3] can be found in their usual places. The git trees contain all four series' worth of changes. This is an extraordinary way to eat your data. Enjoy! Comments and questions are, as always, welcome. --D [1] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=djwong-devel [2] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=djwong-devel [3] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=djwong-devel ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH 1/2] xfs: skip scrub xref if corruption already noted 2018-04-02 19:55 [PATCH 0/2] xfs: online scrub fixes Darrick J. Wong @ 2018-04-02 19:55 ` Darrick J. Wong 2018-04-03 1:48 ` Dave Chinner 2018-04-03 3:00 ` [PATCH v2 " Darrick J. Wong 2018-04-02 19:55 ` [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk Darrick J. Wong 1 sibling, 2 replies; 7+ messages in thread From: Darrick J. Wong @ 2018-04-02 19:55 UTC (permalink / raw) To: darrick.wong; +Cc: linux-xfs From: Darrick J. Wong <darrick.wong@oracle.com> Don't bother looking for cross-referencing problems if the metadata is already corrupt or we've already found a cross-referencing problem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- fs/xfs/scrub/alloc.c | 4 +++- fs/xfs/scrub/common.c | 5 +++++ fs/xfs/scrub/ialloc.c | 4 +++- fs/xfs/scrub/refcount.c | 8 ++++++-- fs/xfs/scrub/rmap.c | 8 ++++++-- fs/xfs/scrub/rtbitmap.c | 4 ++++ 6 files changed, 27 insertions(+), 6 deletions(-) diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c index 517c079..096bc53 100644 --- a/fs/xfs/scrub/alloc.c +++ b/fs/xfs/scrub/alloc.c @@ -172,7 +172,9 @@ xfs_scrub_xref_is_used_space( bool is_freesp; int error; - if (!sc->sa.bno_cur) + if (!sc->sa.bno_cur || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; error = xfs_alloc_has_record(sc->sa.bno_cur, agbno, len, &is_freesp); diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c index fe3a2b1..ddb0ff6 100644 --- a/fs/xfs/scrub/common.c +++ b/fs/xfs/scrub/common.c @@ -727,6 +727,11 @@ xfs_scrub_should_check_xref( int *error, struct xfs_btree_cur **curpp) { + /* No point in xref if we already know we're corrupt. */ + if (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT)) + return false; + if (*error == 0) return true; diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c index 106ca4b..6e1ce25 100644 --- a/fs/xfs/scrub/ialloc.c +++ b/fs/xfs/scrub/ialloc.c @@ -496,7 +496,9 @@ xfs_scrub_xref_inode_check( bool has_inodes; int error; - if (!(*icur)) + if (!(*icur) || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; error = xfs_ialloc_has_inodes_at_extent(*icur, agbno, len, &has_inodes); diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c index 400f156..1f9787a 100644 --- a/fs/xfs/scrub/refcount.c +++ b/fs/xfs/scrub/refcount.c @@ -460,7 +460,9 @@ xfs_scrub_xref_is_cow_staging( int has_refcount; int error; - if (!sc->sa.refc_cur) + if (!sc->sa.refc_cur || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; /* Find the CoW staging extent. */ @@ -504,7 +506,9 @@ xfs_scrub_xref_is_not_shared( bool shared; int error; - if (!sc->sa.refc_cur) + if (!sc->sa.refc_cur || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; error = xfs_refcount_has_record(sc->sa.refc_cur, agbno, len, &shared); diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c index 8f2a7c3..b7c2ef4 100644 --- a/fs/xfs/scrub/rmap.c +++ b/fs/xfs/scrub/rmap.c @@ -207,7 +207,9 @@ xfs_scrub_xref_check_owner( bool has_rmap; int error; - if (!sc->sa.rmap_cur) + if (!sc->sa.rmap_cur || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; error = xfs_rmap_record_exists(sc->sa.rmap_cur, bno, len, oinfo, @@ -250,7 +252,9 @@ xfs_scrub_xref_has_no_owner( bool has_rmap; int error; - if (!sc->sa.rmap_cur) + if (!sc->sa.rmap_cur || + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT))) return; error = xfs_rmap_has_record(sc->sa.rmap_cur, bno, len, &has_rmap); diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c index 39c41dfe..54255f3 100644 --- a/fs/xfs/scrub/rtbitmap.c +++ b/fs/xfs/scrub/rtbitmap.c @@ -110,6 +110,10 @@ xfs_scrub_xref_is_used_rt_space( bool is_free; int error; + if (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT)) + return; + xfs_ilock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP); error = xfs_rtalloc_extent_is_free(sc->mp, sc->tp, fsbno, len, &is_free); ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 1/2] xfs: skip scrub xref if corruption already noted 2018-04-02 19:55 ` [PATCH 1/2] xfs: skip scrub xref if corruption already noted Darrick J. Wong @ 2018-04-03 1:48 ` Dave Chinner 2018-04-03 3:00 ` [PATCH v2 " Darrick J. Wong 1 sibling, 0 replies; 7+ messages in thread From: Dave Chinner @ 2018-04-03 1:48 UTC (permalink / raw) To: Darrick J. Wong; +Cc: linux-xfs On Mon, Apr 02, 2018 at 12:55:36PM -0700, Darrick J. Wong wrote: > From: Darrick J. Wong <darrick.wong@oracle.com> > > Don't bother looking for cross-referencing problems if the metadata is > already corrupt or we've already found a cross-referencing problem. > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> > --- > fs/xfs/scrub/alloc.c | 4 +++- > fs/xfs/scrub/common.c | 5 +++++ > fs/xfs/scrub/ialloc.c | 4 +++- > fs/xfs/scrub/refcount.c | 8 ++++++-- > fs/xfs/scrub/rmap.c | 8 ++++++-- > fs/xfs/scrub/rtbitmap.c | 4 ++++ > 6 files changed, 27 insertions(+), 6 deletions(-) > > > diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c > index 517c079..096bc53 100644 > --- a/fs/xfs/scrub/alloc.c > +++ b/fs/xfs/scrub/alloc.c > @@ -172,7 +172,9 @@ xfs_scrub_xref_is_used_space( > bool is_freesp; > int error; > > - if (!sc->sa.bno_cur) > + if (!sc->sa.bno_cur || > + (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | > + XFS_SCRUB_OFLAG_XCORRUPT))) Can you put this check in a static inline? if (xfs_scrub_object_is_corrupt(sc)) return; Looks a bunch neater here. Otherwise it looks fine. -Dave. -- Dave Chinner david@fromorbit.com ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v2 1/2] xfs: skip scrub xref if corruption already noted 2018-04-02 19:55 ` [PATCH 1/2] xfs: skip scrub xref if corruption already noted Darrick J. Wong 2018-04-03 1:48 ` Dave Chinner @ 2018-04-03 3:00 ` Darrick J. Wong 1 sibling, 0 replies; 7+ messages in thread From: Darrick J. Wong @ 2018-04-03 3:00 UTC (permalink / raw) To: linux-xfs; +Cc: Dave Chinner From: Darrick J. Wong <darrick.wong@oracle.com> Don't bother looking for cross-referencing problems if the metadata is already corrupt or we've already found a cross-referencing problem. Since we added a helper function for flags testing, convert existing users to use it. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- v2: add helper function instead of open coding --- fs/xfs/scrub/alloc.c | 2 +- fs/xfs/scrub/common.c | 4 ++++ fs/xfs/scrub/common.h | 6 ++++++ fs/xfs/scrub/ialloc.c | 2 +- fs/xfs/scrub/refcount.c | 4 ++-- fs/xfs/scrub/rmap.c | 4 ++-- fs/xfs/scrub/rtbitmap.c | 3 +++ fs/xfs/scrub/scrub.c | 3 +-- 8 files changed, 20 insertions(+), 8 deletions(-) diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c index 517c079..f5c6993 100644 --- a/fs/xfs/scrub/alloc.c +++ b/fs/xfs/scrub/alloc.c @@ -172,7 +172,7 @@ xfs_scrub_xref_is_used_space( bool is_freesp; int error; - if (!sc->sa.bno_cur) + if (!sc->sa.bno_cur || xfs_scrub_found_corruption(sc->sm)) return; error = xfs_alloc_has_record(sc->sa.bno_cur, agbno, len, &is_freesp); diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c index fe3a2b1..51d2a59 100644 --- a/fs/xfs/scrub/common.c +++ b/fs/xfs/scrub/common.c @@ -727,6 +727,10 @@ xfs_scrub_should_check_xref( int *error, struct xfs_btree_cur **curpp) { + /* No point in xref if we already know we're corrupt. */ + if (xfs_scrub_found_corruption(sc->sm)) + return false; + if (*error == 0) return true; diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h index deaf604..30e9039 100644 --- a/fs/xfs/scrub/common.h +++ b/fs/xfs/scrub/common.h @@ -157,4 +157,10 @@ int xfs_scrub_setup_inode_contents(struct xfs_scrub_context *sc, struct xfs_inode *ip, unsigned int resblks); void xfs_scrub_buffer_recheck(struct xfs_scrub_context *sc, struct xfs_buf *bp); +static inline bool xfs_scrub_found_corruption(struct xfs_scrub_metadata *sm) +{ + return sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | + XFS_SCRUB_OFLAG_XCORRUPT); +} + #endif /* __XFS_SCRUB_COMMON_H__ */ diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c index 106ca4b..05e6191 100644 --- a/fs/xfs/scrub/ialloc.c +++ b/fs/xfs/scrub/ialloc.c @@ -496,7 +496,7 @@ xfs_scrub_xref_inode_check( bool has_inodes; int error; - if (!(*icur)) + if (!(*icur) || xfs_scrub_found_corruption(sc->sm)) return; error = xfs_ialloc_has_inodes_at_extent(*icur, agbno, len, &has_inodes); diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c index 400f156..823bda3 100644 --- a/fs/xfs/scrub/refcount.c +++ b/fs/xfs/scrub/refcount.c @@ -460,7 +460,7 @@ xfs_scrub_xref_is_cow_staging( int has_refcount; int error; - if (!sc->sa.refc_cur) + if (!sc->sa.refc_cur || xfs_scrub_found_corruption(sc->sm)) return; /* Find the CoW staging extent. */ @@ -504,7 +504,7 @@ xfs_scrub_xref_is_not_shared( bool shared; int error; - if (!sc->sa.refc_cur) + if (!sc->sa.refc_cur || xfs_scrub_found_corruption(sc->sm)) return; error = xfs_refcount_has_record(sc->sa.refc_cur, agbno, len, &shared); diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c index 8f2a7c3..9ca92e4 100644 --- a/fs/xfs/scrub/rmap.c +++ b/fs/xfs/scrub/rmap.c @@ -207,7 +207,7 @@ xfs_scrub_xref_check_owner( bool has_rmap; int error; - if (!sc->sa.rmap_cur) + if (!sc->sa.rmap_cur || xfs_scrub_found_corruption(sc->sm)) return; error = xfs_rmap_record_exists(sc->sa.rmap_cur, bno, len, oinfo, @@ -250,7 +250,7 @@ xfs_scrub_xref_has_no_owner( bool has_rmap; int error; - if (!sc->sa.rmap_cur) + if (!sc->sa.rmap_cur || xfs_scrub_found_corruption(sc->sm)) return; error = xfs_rmap_has_record(sc->sa.rmap_cur, bno, len, &has_rmap); diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c index 39c41dfe..c8672f7 100644 --- a/fs/xfs/scrub/rtbitmap.c +++ b/fs/xfs/scrub/rtbitmap.c @@ -110,6 +110,9 @@ xfs_scrub_xref_is_used_rt_space( bool is_free; int error; + if (xfs_scrub_found_corruption(sc->sm)) + return; + xfs_ilock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP); error = xfs_rtalloc_extent_is_free(sc->mp, sc->tp, fsbno, len, &is_free); diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c index 26c7596..cb1e727 100644 --- a/fs/xfs/scrub/scrub.c +++ b/fs/xfs/scrub/scrub.c @@ -446,8 +446,7 @@ xfs_scrub_metadata( } else if (error) goto out_teardown; - if (sc.sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT | - XFS_SCRUB_OFLAG_XCORRUPT)) + if (xfs_scrub_found_corruption(sc.sm)) xfs_alert_ratelimited(mp, "Corruption detected during scrub."); out_teardown: ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk 2018-04-02 19:55 [PATCH 0/2] xfs: online scrub fixes Darrick J. Wong 2018-04-02 19:55 ` [PATCH 1/2] xfs: skip scrub xref if corruption already noted Darrick J. Wong @ 2018-04-02 19:55 ` Darrick J. Wong 2018-04-03 1:53 ` Dave Chinner 2018-04-03 3:01 ` [PATCH v2 " Darrick J. Wong 1 sibling, 2 replies; 7+ messages in thread From: Darrick J. Wong @ 2018-04-02 19:55 UTC (permalink / raw) To: darrick.wong; +Cc: linux-xfs From: Darrick J. Wong <darrick.wong@oracle.com> This function is basically a generic AGFL block iterator, so promote it to libxfs ahead of online repair wanting to use it. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- fs/xfs/libxfs/xfs_alloc.c | 37 +++++++++++++++++++++ fs/xfs/libxfs/xfs_alloc.h | 5 +++ fs/xfs/scrub/agheader.c | 78 ++++++++------------------------------------- fs/xfs/scrub/common.h | 4 -- 4 files changed, 55 insertions(+), 69 deletions(-) diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 4c03ec2..d860116 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -3128,3 +3128,40 @@ xfs_alloc_has_record( return xfs_btree_has_record(cur, &low, &high, exists); } + +/* + * Walk all the blocks in the AGFL. The fn function can return any negative + * error code or XFS_BTREE_QUERY_RANGE_ABORT. + */ +int +xfs_agfl_walk( + struct xfs_mount *mp, + struct xfs_agf *agf, + struct xfs_buf *agflbp, + xfs_agfl_walk_fn fn, + void *priv) +{ + __be32 *agfl_bno; + unsigned int i; + int error; + + agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp); + i = be32_to_cpu(agf->agf_flfirst); + + /* Nothing to walk in an empty AGFL. */ + if (agf->agf_flcount == cpu_to_be32(0)) + return 0; + + /* Otherwise, walk from first to last, wrapping as needed. */ + for (;;) { + error = fn(mp, be32_to_cpu(agfl_bno[i]), priv); + if (error) + return error; + if (i == be32_to_cpu(agf->agf_fllast)) + break; + if (++i == xfs_agfl_size(mp)) + i = 0; + } + + return 0; +} diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index a311a24..d4b83d1 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -249,4 +249,9 @@ bool xfs_verify_fsbno(struct xfs_mount *mp, xfs_fsblock_t fsbno); int xfs_alloc_has_record(struct xfs_btree_cur *cur, xfs_agblock_t bno, xfs_extlen_t len, bool *exist); +typedef int (*xfs_agfl_walk_fn)(struct xfs_mount *mp, xfs_agblock_t bno, + void *priv); +int xfs_agfl_walk(struct xfs_mount *mp, struct xfs_agf *agf, + struct xfs_buf *agflbp, xfs_agfl_walk_fn fn, void *priv); + #endif /* __XFS_ALLOC_H__ */ diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c index 018aabbd..9bf1a1a 100644 --- a/fs/xfs/scrub/agheader.c +++ b/fs/xfs/scrub/agheader.c @@ -38,68 +38,6 @@ #include "scrub/common.h" #include "scrub/trace.h" -/* - * Walk all the blocks in the AGFL. The fn function can return any negative - * error code or XFS_BTREE_QUERY_RANGE_ABORT. - */ -int -xfs_scrub_walk_agfl( - struct xfs_scrub_context *sc, - int (*fn)(struct xfs_scrub_context *, - xfs_agblock_t bno, void *), - void *priv) -{ - struct xfs_agf *agf; - __be32 *agfl_bno; - struct xfs_mount *mp = sc->mp; - unsigned int flfirst; - unsigned int fllast; - int i; - int error; - - agf = XFS_BUF_TO_AGF(sc->sa.agf_bp); - agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, sc->sa.agfl_bp); - flfirst = be32_to_cpu(agf->agf_flfirst); - fllast = be32_to_cpu(agf->agf_fllast); - - /* Nothing to walk in an empty AGFL. */ - if (agf->agf_flcount == cpu_to_be32(0)) - return 0; - - /* first to last is a consecutive list. */ - if (fllast >= flfirst) { - for (i = flfirst; i <= fllast; i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - return 0; - } - - /* first to the end */ - for (i = flfirst; i < xfs_agfl_size(mp); i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - /* the start to last. */ - for (i = 0; i <= fllast; i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - return 0; -} - /* Superblock */ /* Cross-reference with the other btrees. */ @@ -680,6 +618,7 @@ struct xfs_scrub_agfl_info { unsigned int sz_entries; unsigned int nr_entries; xfs_agblock_t *entries; + struct xfs_scrub_context *sc; }; /* Cross-reference with the other btrees. */ @@ -701,12 +640,12 @@ xfs_scrub_agfl_block_xref( /* Scrub an AGFL block. */ STATIC int xfs_scrub_agfl_block( - struct xfs_scrub_context *sc, + struct xfs_mount *mp, xfs_agblock_t agbno, void *priv) { - struct xfs_mount *mp = sc->mp; struct xfs_scrub_agfl_info *sai = priv; + struct xfs_scrub_context *sc = sai->sc; xfs_agnumber_t agno = sc->sa.agno; if (xfs_verify_agbno(mp, agno, agbno) && @@ -717,6 +656,9 @@ xfs_scrub_agfl_block( xfs_scrub_agfl_block_xref(sc, agbno, priv); + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return XFS_BTREE_QUERY_RANGE_ABORT; + return 0; } @@ -796,6 +738,7 @@ xfs_scrub_agfl( goto out; } memset(&sai, 0, sizeof(sai)); + sai.sc = sc; sai.sz_entries = agflcount; sai.entries = kmem_zalloc(sizeof(xfs_agblock_t) * agflcount, KM_NOFS); if (!sai.entries) { @@ -805,7 +748,12 @@ xfs_scrub_agfl( /* Check the blocks in the AGFL. */ xfs_rmap_ag_owner(&sai.oinfo, XFS_RMAP_OWN_AG); - error = xfs_scrub_walk_agfl(sc, xfs_scrub_agfl_block, &sai); + error = xfs_agfl_walk(sc->mp, XFS_BUF_TO_AGF(sc->sa.agf_bp), + sc->sa.agfl_bp, xfs_scrub_agfl_block, &sai); + if (error == XFS_BTREE_QUERY_RANGE_ABORT) { + error = 0; + goto out_free; + } if (error) goto out_free; diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h index deaf604..140e90c 100644 --- a/fs/xfs/scrub/common.h +++ b/fs/xfs/scrub/common.h @@ -141,10 +141,6 @@ int xfs_scrub_ag_read_headers(struct xfs_scrub_context *sc, xfs_agnumber_t agno, void xfs_scrub_ag_btcur_free(struct xfs_scrub_ag *sa); int xfs_scrub_ag_btcur_init(struct xfs_scrub_context *sc, struct xfs_scrub_ag *sa); -int xfs_scrub_walk_agfl(struct xfs_scrub_context *sc, - int (*fn)(struct xfs_scrub_context *, xfs_agblock_t bno, - void *), - void *priv); int xfs_scrub_count_rmap_ownedby_ag(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur, struct xfs_owner_info *oinfo, ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk 2018-04-02 19:55 ` [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk Darrick J. Wong @ 2018-04-03 1:53 ` Dave Chinner 2018-04-03 3:01 ` [PATCH v2 " Darrick J. Wong 1 sibling, 0 replies; 7+ messages in thread From: Dave Chinner @ 2018-04-03 1:53 UTC (permalink / raw) To: Darrick J. Wong; +Cc: linux-xfs On Mon, Apr 02, 2018 at 12:55:42PM -0700, Darrick J. Wong wrote: > From: Darrick J. Wong <darrick.wong@oracle.com> > > This function is basically a generic AGFL block iterator, so promote it > to libxfs ahead of online repair wanting to use it. > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> > --- > fs/xfs/libxfs/xfs_alloc.c | 37 +++++++++++++++++++++ > fs/xfs/libxfs/xfs_alloc.h | 5 +++ > fs/xfs/scrub/agheader.c | 78 ++++++++------------------------------------- > fs/xfs/scrub/common.h | 4 -- > 4 files changed, 55 insertions(+), 69 deletions(-) > > > diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c > index 4c03ec2..d860116 100644 > --- a/fs/xfs/libxfs/xfs_alloc.c > +++ b/fs/xfs/libxfs/xfs_alloc.c > @@ -3128,3 +3128,40 @@ xfs_alloc_has_record( > > return xfs_btree_has_record(cur, &low, &high, exists); > } > + > +/* > + * Walk all the blocks in the AGFL. The fn function can return any negative > + * error code or XFS_BTREE_QUERY_RANGE_ABORT. > + */ > +int > +xfs_agfl_walk( > + struct xfs_mount *mp, > + struct xfs_agf *agf, > + struct xfs_buf *agflbp, > + xfs_agfl_walk_fn fn, xfs_agfl_walk_fn walk_fn, I don't really like "fn" as a variable - it's hard to pick out what it does and where it's called in the code. Also, the comment would make more sense as "The @walk_fn can return any negative.... Otherwise looks OK. -Dave. -- Dave Chinner david@fromorbit.com ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v2 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk 2018-04-02 19:55 ` [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk Darrick J. Wong 2018-04-03 1:53 ` Dave Chinner @ 2018-04-03 3:01 ` Darrick J. Wong 1 sibling, 0 replies; 7+ messages in thread From: Darrick J. Wong @ 2018-04-03 3:01 UTC (permalink / raw) To: linux-xfs; +Cc: Dave Chinner From: Darrick J. Wong <darrick.wong@oracle.com> This function is basically a generic AGFL block iterator, so promote it to libxfs ahead of online repair wanting to use it. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> --- v2: rename fn to walk_fn --- fs/xfs/libxfs/xfs_alloc.c | 37 +++++++++++++++++++++ fs/xfs/libxfs/xfs_alloc.h | 5 +++ fs/xfs/scrub/agheader.c | 78 ++++++++------------------------------------- fs/xfs/scrub/common.h | 4 -- 4 files changed, 55 insertions(+), 69 deletions(-) diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 4c03ec2..03da9a9 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -3128,3 +3128,40 @@ xfs_alloc_has_record( return xfs_btree_has_record(cur, &low, &high, exists); } + +/* + * Walk all the blocks in the AGFL. The @walk_fn can return any negative + * error code or XFS_BTREE_QUERY_RANGE_ABORT. + */ +int +xfs_agfl_walk( + struct xfs_mount *mp, + struct xfs_agf *agf, + struct xfs_buf *agflbp, + xfs_agfl_walk_fn walk_fn, + void *priv) +{ + __be32 *agfl_bno; + unsigned int i; + int error; + + agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp); + i = be32_to_cpu(agf->agf_flfirst); + + /* Nothing to walk in an empty AGFL. */ + if (agf->agf_flcount == cpu_to_be32(0)) + return 0; + + /* Otherwise, walk from first to last, wrapping as needed. */ + for (;;) { + error = walk_fn(mp, be32_to_cpu(agfl_bno[i]), priv); + if (error) + return error; + if (i == be32_to_cpu(agf->agf_fllast)) + break; + if (++i == xfs_agfl_size(mp)) + i = 0; + } + + return 0; +} diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index a311a24..56f9ac4 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -249,4 +249,9 @@ bool xfs_verify_fsbno(struct xfs_mount *mp, xfs_fsblock_t fsbno); int xfs_alloc_has_record(struct xfs_btree_cur *cur, xfs_agblock_t bno, xfs_extlen_t len, bool *exist); +typedef int (*xfs_agfl_walk_fn)(struct xfs_mount *mp, xfs_agblock_t bno, + void *priv); +int xfs_agfl_walk(struct xfs_mount *mp, struct xfs_agf *agf, + struct xfs_buf *agflbp, xfs_agfl_walk_fn walk_fn, void *priv); + #endif /* __XFS_ALLOC_H__ */ diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c index 018aabbd..9bf1a1a 100644 --- a/fs/xfs/scrub/agheader.c +++ b/fs/xfs/scrub/agheader.c @@ -38,68 +38,6 @@ #include "scrub/common.h" #include "scrub/trace.h" -/* - * Walk all the blocks in the AGFL. The fn function can return any negative - * error code or XFS_BTREE_QUERY_RANGE_ABORT. - */ -int -xfs_scrub_walk_agfl( - struct xfs_scrub_context *sc, - int (*fn)(struct xfs_scrub_context *, - xfs_agblock_t bno, void *), - void *priv) -{ - struct xfs_agf *agf; - __be32 *agfl_bno; - struct xfs_mount *mp = sc->mp; - unsigned int flfirst; - unsigned int fllast; - int i; - int error; - - agf = XFS_BUF_TO_AGF(sc->sa.agf_bp); - agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, sc->sa.agfl_bp); - flfirst = be32_to_cpu(agf->agf_flfirst); - fllast = be32_to_cpu(agf->agf_fllast); - - /* Nothing to walk in an empty AGFL. */ - if (agf->agf_flcount == cpu_to_be32(0)) - return 0; - - /* first to last is a consecutive list. */ - if (fllast >= flfirst) { - for (i = flfirst; i <= fllast; i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - return 0; - } - - /* first to the end */ - for (i = flfirst; i < xfs_agfl_size(mp); i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - /* the start to last. */ - for (i = 0; i <= fllast; i++) { - error = fn(sc, be32_to_cpu(agfl_bno[i]), priv); - if (error) - return error; - if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) - return error; - } - - return 0; -} - /* Superblock */ /* Cross-reference with the other btrees. */ @@ -680,6 +618,7 @@ struct xfs_scrub_agfl_info { unsigned int sz_entries; unsigned int nr_entries; xfs_agblock_t *entries; + struct xfs_scrub_context *sc; }; /* Cross-reference with the other btrees. */ @@ -701,12 +640,12 @@ xfs_scrub_agfl_block_xref( /* Scrub an AGFL block. */ STATIC int xfs_scrub_agfl_block( - struct xfs_scrub_context *sc, + struct xfs_mount *mp, xfs_agblock_t agbno, void *priv) { - struct xfs_mount *mp = sc->mp; struct xfs_scrub_agfl_info *sai = priv; + struct xfs_scrub_context *sc = sai->sc; xfs_agnumber_t agno = sc->sa.agno; if (xfs_verify_agbno(mp, agno, agbno) && @@ -717,6 +656,9 @@ xfs_scrub_agfl_block( xfs_scrub_agfl_block_xref(sc, agbno, priv); + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return XFS_BTREE_QUERY_RANGE_ABORT; + return 0; } @@ -796,6 +738,7 @@ xfs_scrub_agfl( goto out; } memset(&sai, 0, sizeof(sai)); + sai.sc = sc; sai.sz_entries = agflcount; sai.entries = kmem_zalloc(sizeof(xfs_agblock_t) * agflcount, KM_NOFS); if (!sai.entries) { @@ -805,7 +748,12 @@ xfs_scrub_agfl( /* Check the blocks in the AGFL. */ xfs_rmap_ag_owner(&sai.oinfo, XFS_RMAP_OWN_AG); - error = xfs_scrub_walk_agfl(sc, xfs_scrub_agfl_block, &sai); + error = xfs_agfl_walk(sc->mp, XFS_BUF_TO_AGF(sc->sa.agf_bp), + sc->sa.agfl_bp, xfs_scrub_agfl_block, &sai); + if (error == XFS_BTREE_QUERY_RANGE_ABORT) { + error = 0; + goto out_free; + } if (error) goto out_free; diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h index 30e9039..907e47f 100644 --- a/fs/xfs/scrub/common.h +++ b/fs/xfs/scrub/common.h @@ -141,10 +141,6 @@ int xfs_scrub_ag_read_headers(struct xfs_scrub_context *sc, xfs_agnumber_t agno, void xfs_scrub_ag_btcur_free(struct xfs_scrub_ag *sa); int xfs_scrub_ag_btcur_init(struct xfs_scrub_context *sc, struct xfs_scrub_ag *sa); -int xfs_scrub_walk_agfl(struct xfs_scrub_context *sc, - int (*fn)(struct xfs_scrub_context *, xfs_agblock_t bno, - void *), - void *priv); int xfs_scrub_count_rmap_ownedby_ag(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur, struct xfs_owner_info *oinfo, ^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2018-04-03 3:01 UTC | newest] Thread overview: 7+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2018-04-02 19:55 [PATCH 0/2] xfs: online scrub fixes Darrick J. Wong 2018-04-02 19:55 ` [PATCH 1/2] xfs: skip scrub xref if corruption already noted Darrick J. Wong 2018-04-03 1:48 ` Dave Chinner 2018-04-03 3:00 ` [PATCH v2 " Darrick J. Wong 2018-04-02 19:55 ` [PATCH 2/2] xfs: hoist xfs_scrub_agfl_walk to libxfs as xfs_agfl_walk Darrick J. Wong 2018-04-03 1:53 ` Dave Chinner 2018-04-03 3:01 ` [PATCH v2 " Darrick J. Wong
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).