diff options
author | Dave Chinner <dchinner@redhat.com> | 2023-09-06 11:21:47 +0200 |
---|---|---|
committer | Carlos Maiolino <cem@kernel.org> | 2023-09-07 11:55:50 +0200 |
commit | ef16737e44b9274fddcc736ec2da5c07d5c4f703 (patch) | |
tree | c8bd0e35b4c7d4c2609e18d2aeae6d4e513115ec | |
parent | d03a3c4b235c6ecdda9b27d5f31115ad710b5912 (diff) | |
download | xfsprogs-dev-ef16737e44b9274fddcc736ec2da5c07d5c4f703.tar.gz |
xfs: use deferred frees for btree block freeing
Source kernel commit: b742d7b4f0e03df25c2a772adcded35044b625ca
Btrees that aren't freespace management trees use the normal extent
allocation and freeing routines for their blocks. Hence when a btree
block is freed, a direct call to xfs_free_extent() is made and the
extent is immediately freed. This puts the entire free space
management btrees under this path, so we are stacking btrees on
btrees in the call stack. The inobt, finobt and refcount btrees
all do this.
However, the bmap btree does not do this - it calls
xfs_free_extent_later() to defer the extent free operation via an
XEFI and hence it gets processed in deferred operation processing
during the commit of the primary transaction (i.e. via intent
chaining).
We need to change xfs_free_extent() to behave in a non-blocking
manner so that we can avoid deadlocks with busy extents near ENOSPC
in transactions that free multiple extents. Inserting or removing a
record from a btree can cause a multi-level tree merge operation and
that will free multiple blocks from the btree in a single
transaction. i.e. we can call xfs_free_extent() multiple times, and
hence the btree manipulation transaction is vulnerable to this busy
extent deadlock vector.
To fix this, convert all the remaining callers of xfs_free_extent()
to use xfs_free_extent_later() to queue XEFIs and hence defer
processing of the extent frees to a context that can be safely
restarted if a deadlock condition is detected.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandan.babu@oracle.com>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
-rw-r--r-- | libxfs/xfs_ag.c | 2 | ||||
-rw-r--r-- | libxfs/xfs_alloc.c | 4 | ||||
-rw-r--r-- | libxfs/xfs_alloc.h | 8 | ||||
-rw-r--r-- | libxfs/xfs_bmap.c | 8 | ||||
-rw-r--r-- | libxfs/xfs_bmap_btree.c | 3 | ||||
-rw-r--r-- | libxfs/xfs_ialloc.c | 8 | ||||
-rw-r--r-- | libxfs/xfs_ialloc_btree.c | 3 | ||||
-rw-r--r-- | libxfs/xfs_refcount.c | 9 | ||||
-rw-r--r-- | libxfs/xfs_refcount_btree.c | 8 |
9 files changed, 29 insertions, 24 deletions
diff --git a/libxfs/xfs_ag.c b/libxfs/xfs_ag.c index c82afbb82a..75711cbaef 100644 --- a/libxfs/xfs_ag.c +++ b/libxfs/xfs_ag.c @@ -983,7 +983,7 @@ xfs_ag_shrink_space( goto resv_err; err2 = __xfs_free_extent_later(*tpp, args.fsbno, delta, NULL, - true); + XFS_AG_RESV_NONE, true); if (err2) goto resv_err; diff --git a/libxfs/xfs_alloc.c b/libxfs/xfs_alloc.c index 107a849207..e5fd360cff 100644 --- a/libxfs/xfs_alloc.c +++ b/libxfs/xfs_alloc.c @@ -2445,6 +2445,7 @@ xfs_defer_agfl_block( xefi->xefi_startblock = XFS_AGB_TO_FSB(mp, agno, agbno); xefi->xefi_blockcount = 1; xefi->xefi_owner = oinfo->oi_owner; + xefi->xefi_agresv = XFS_AG_RESV_AGFL; if (XFS_IS_CORRUPT(mp, !xfs_verify_fsbno(mp, xefi->xefi_startblock))) return -EFSCORRUPTED; @@ -2466,6 +2467,7 @@ __xfs_free_extent_later( xfs_fsblock_t bno, xfs_filblks_t len, const struct xfs_owner_info *oinfo, + enum xfs_ag_resv_type type, bool skip_discard) { struct xfs_extent_free_item *xefi; @@ -2486,6 +2488,7 @@ __xfs_free_extent_later( ASSERT(agbno + len <= mp->m_sb.sb_agblocks); #endif ASSERT(xfs_extfree_item_cache != NULL); + ASSERT(type != XFS_AG_RESV_AGFL); if (XFS_IS_CORRUPT(mp, !xfs_verify_fsbext(mp, bno, len))) return -EFSCORRUPTED; @@ -2494,6 +2497,7 @@ __xfs_free_extent_later( GFP_KERNEL | __GFP_NOFAIL); xefi->xefi_startblock = bno; xefi->xefi_blockcount = (xfs_extlen_t)len; + xefi->xefi_agresv = type; if (skip_discard) xefi->xefi_flags |= XFS_EFI_SKIP_DISCARD; if (oinfo) { diff --git a/libxfs/xfs_alloc.h b/libxfs/xfs_alloc.h index 85ac470be0..a3e519577e 100644 --- a/libxfs/xfs_alloc.h +++ b/libxfs/xfs_alloc.h @@ -232,7 +232,7 @@ xfs_buf_to_agfl_bno( int __xfs_free_extent_later(struct xfs_trans *tp, xfs_fsblock_t bno, xfs_filblks_t len, const struct xfs_owner_info *oinfo, - bool skip_discard); + enum xfs_ag_resv_type type, bool skip_discard); /* * List of extents to be free "later". @@ -245,6 +245,7 @@ struct xfs_extent_free_item { xfs_extlen_t xefi_blockcount;/* number of blocks in extent */ struct xfs_perag *xefi_pag; unsigned int xefi_flags; + enum xfs_ag_resv_type xefi_agresv; }; void xfs_extent_free_get_group(struct xfs_mount *mp, @@ -259,9 +260,10 @@ xfs_free_extent_later( struct xfs_trans *tp, xfs_fsblock_t bno, xfs_filblks_t len, - const struct xfs_owner_info *oinfo) + const struct xfs_owner_info *oinfo, + enum xfs_ag_resv_type type) { - return __xfs_free_extent_later(tp, bno, len, oinfo, false); + return __xfs_free_extent_later(tp, bno, len, oinfo, type, false); } diff --git a/libxfs/xfs_bmap.c b/libxfs/xfs_bmap.c index 5deeb47271..2bd23d40e7 100644 --- a/libxfs/xfs_bmap.c +++ b/libxfs/xfs_bmap.c @@ -567,7 +567,8 @@ xfs_bmap_btree_to_extents( return error; xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, whichfork); - error = xfs_free_extent_later(cur->bc_tp, cbno, 1, &oinfo); + error = xfs_free_extent_later(cur->bc_tp, cbno, 1, &oinfo, + XFS_AG_RESV_NONE); if (error) return error; @@ -5229,8 +5230,9 @@ xfs_bmap_del_extent_real( } else { error = __xfs_free_extent_later(tp, del->br_startblock, del->br_blockcount, NULL, - (bflags & XFS_BMAPI_NODISCARD) || - del->br_state == XFS_EXT_UNWRITTEN); + XFS_AG_RESV_NONE, + ((bflags & XFS_BMAPI_NODISCARD) || + del->br_state == XFS_EXT_UNWRITTEN)); if (error) goto done; } diff --git a/libxfs/xfs_bmap_btree.c b/libxfs/xfs_bmap_btree.c index 751e8165b0..925cc153b2 100644 --- a/libxfs/xfs_bmap_btree.c +++ b/libxfs/xfs_bmap_btree.c @@ -269,7 +269,8 @@ xfs_bmbt_free_block( int error; xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_ino.whichfork); - error = xfs_free_extent_later(cur->bc_tp, fsbno, 1, &oinfo); + error = xfs_free_extent_later(cur->bc_tp, fsbno, 1, &oinfo, + XFS_AG_RESV_NONE); if (error) return error; diff --git a/libxfs/xfs_ialloc.c b/libxfs/xfs_ialloc.c index 42d8863e08..1cb680d4f4 100644 --- a/libxfs/xfs_ialloc.c +++ b/libxfs/xfs_ialloc.c @@ -1848,8 +1848,8 @@ xfs_difree_inode_chunk( /* not sparse, calculate extent info directly */ return xfs_free_extent_later(tp, XFS_AGB_TO_FSB(mp, agno, sagbno), - M_IGEO(mp)->ialloc_blks, - &XFS_RMAP_OINFO_INODES); + M_IGEO(mp)->ialloc_blks, &XFS_RMAP_OINFO_INODES, + XFS_AG_RESV_NONE); } /* holemask is only 16-bits (fits in an unsigned long) */ @@ -1894,8 +1894,8 @@ xfs_difree_inode_chunk( ASSERT(agbno % mp->m_sb.sb_spino_align == 0); ASSERT(contigblk % mp->m_sb.sb_spino_align == 0); error = xfs_free_extent_later(tp, - XFS_AGB_TO_FSB(mp, agno, agbno), - contigblk, &XFS_RMAP_OINFO_INODES); + XFS_AGB_TO_FSB(mp, agno, agbno), contigblk, + &XFS_RMAP_OINFO_INODES, XFS_AG_RESV_NONE); if (error) return error; diff --git a/libxfs/xfs_ialloc_btree.c b/libxfs/xfs_ialloc_btree.c index 560393b05d..0c5d7ba1ce 100644 --- a/libxfs/xfs_ialloc_btree.c +++ b/libxfs/xfs_ialloc_btree.c @@ -159,8 +159,7 @@ __xfs_inobt_free_block( xfs_inobt_mod_blockcount(cur, -1); fsbno = XFS_DADDR_TO_FSB(cur->bc_mp, xfs_buf_daddr(bp)); - return xfs_free_extent(cur->bc_tp, cur->bc_ag.pag, - XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1, + return xfs_free_extent_later(cur->bc_tp, fsbno, 1, &XFS_RMAP_OINFO_INOBT, resv); } diff --git a/libxfs/xfs_refcount.c b/libxfs/xfs_refcount.c index 0006ad7c93..8f30dc196e 100644 --- a/libxfs/xfs_refcount.c +++ b/libxfs/xfs_refcount.c @@ -1151,7 +1151,8 @@ xfs_refcount_adjust_extents( cur->bc_ag.pag->pag_agno, tmp.rc_startblock); error = xfs_free_extent_later(cur->bc_tp, fsbno, - tmp.rc_blockcount, NULL); + tmp.rc_blockcount, NULL, + XFS_AG_RESV_NONE); if (error) goto out_error; } @@ -1212,7 +1213,8 @@ xfs_refcount_adjust_extents( cur->bc_ag.pag->pag_agno, ext.rc_startblock); error = xfs_free_extent_later(cur->bc_tp, fsbno, - ext.rc_blockcount, NULL); + ext.rc_blockcount, NULL, + XFS_AG_RESV_NONE); if (error) goto out_error; } @@ -1980,7 +1982,8 @@ xfs_refcount_recover_cow_leftovers( /* Free the block. */ error = xfs_free_extent_later(tp, fsb, - rr->rr_rrec.rc_blockcount, NULL); + rr->rr_rrec.rc_blockcount, NULL, + XFS_AG_RESV_NONE); if (error) goto out_trans; diff --git a/libxfs/xfs_refcount_btree.c b/libxfs/xfs_refcount_btree.c index 1205e29191..67cb59e335 100644 --- a/libxfs/xfs_refcount_btree.c +++ b/libxfs/xfs_refcount_btree.c @@ -105,19 +105,13 @@ xfs_refcountbt_free_block( struct xfs_buf *agbp = cur->bc_ag.agbp; struct xfs_agf *agf = agbp->b_addr; xfs_fsblock_t fsbno = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp)); - int error; trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_ag.pag->pag_agno, XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1); be32_add_cpu(&agf->agf_refcount_blocks, -1); xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS); - error = xfs_free_extent(cur->bc_tp, cur->bc_ag.pag, - XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1, + return xfs_free_extent_later(cur->bc_tp, fsbno, 1, &XFS_RMAP_OINFO_REFC, XFS_AG_RESV_METADATA); - if (error) - return error; - - return error; } STATIC int |