Message ID | 20200918094759.2727564-4-chandanrlinux@gmail.com (mailing list archive) |
---|---|
State | Superseded, archived |
Headers | show |
Series | Bail out if transaction can cause extent count to overflow | expand |
On Fri, Sep 18, 2020 at 03:17:52PM +0530, Chandan Babu R wrote: > The extent mapping the file offset at which a hole has to be > inserted will be split into two extents causing extent count to > increase by 1. > > Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com> > --- > fs/xfs/libxfs/xfs_inode_fork.h | 7 +++++++ > fs/xfs/xfs_bmap_item.c | 5 +++++ > fs/xfs/xfs_bmap_util.c | 10 ++++++++++ > 3 files changed, 22 insertions(+) > > diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h > index 7fc2b129a2e7..bcac769a7df6 100644 > --- a/fs/xfs/libxfs/xfs_inode_fork.h > +++ b/fs/xfs/libxfs/xfs_inode_fork.h > @@ -40,6 +40,13 @@ struct xfs_ifork { > */ > #define XFS_IEXT_ADD_NOSPLIT_CNT (1) > > +/* > + * Punching out an extent from the middle of an existing extent can cause the > + * extent count to increase by 1. > + * i.e. | Old extent | Hole | Old extent | > + */ > +#define XFS_IEXT_PUNCH_HOLE_CNT (1) > + > /* > * Fork handling. > */ > diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c > index ec3691372e7c..5c7d08da8ff1 100644 > --- a/fs/xfs/xfs_bmap_item.c > +++ b/fs/xfs/xfs_bmap_item.c > @@ -519,6 +519,11 @@ xfs_bui_item_recover( > } > xfs_trans_ijoin(tp, ip, 0); > > + error = xfs_iext_count_may_overflow(ip, whichfork, > + XFS_IEXT_PUNCH_HOLE_CNT); I think this ought to be XFS_IEXT_ADD_NOSPLIT_CNT if bui_type is XFS_BMAP_MAP and XFS_IEXT_PUNCH_HOLE_CNT if XFS_BMAP_UNMAP. Whoever created the BUI should have called xfs_iext_count_may_overflow before logging the BUI (and hence this should never occur) but it does pay to be careful. :) The rest of the logic in the patch looks ok. --D > + if (error) > + goto err_inode; > + > count = bmap->me_len; > error = xfs_trans_log_finish_bmap_update(tp, budp, type, ip, whichfork, > bmap->me_startoff, bmap->me_startblock, &count, state); > diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c > index dcd6e61df711..0776abd0103c 100644 > --- a/fs/xfs/xfs_bmap_util.c > +++ b/fs/xfs/xfs_bmap_util.c > @@ -891,6 +891,11 @@ xfs_unmap_extent( > > xfs_trans_ijoin(tp, ip, 0); > > + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, > + XFS_IEXT_PUNCH_HOLE_CNT); > + if (error) > + goto out_trans_cancel; > + > error = xfs_bunmapi(tp, ip, startoffset_fsb, len_fsb, 0, 2, done); > if (error) > goto out_trans_cancel; > @@ -1176,6 +1181,11 @@ xfs_insert_file_space( > xfs_ilock(ip, XFS_ILOCK_EXCL); > xfs_trans_ijoin(tp, ip, 0); > > + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, > + XFS_IEXT_PUNCH_HOLE_CNT); > + if (error) > + goto out_trans_cancel; > + > /* > * The extent shifting code works on extent granularity. So, if stop_fsb > * is not the starting block of extent, we need to split the extent at > -- > 2.28.0 >
On Friday 18 September 2020 9:24:52 PM IST Darrick J. Wong wrote: > On Fri, Sep 18, 2020 at 03:17:52PM +0530, Chandan Babu R wrote: > > The extent mapping the file offset at which a hole has to be > > inserted will be split into two extents causing extent count to > > increase by 1. > > > > Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com> > > --- > > fs/xfs/libxfs/xfs_inode_fork.h | 7 +++++++ > > fs/xfs/xfs_bmap_item.c | 5 +++++ > > fs/xfs/xfs_bmap_util.c | 10 ++++++++++ > > 3 files changed, 22 insertions(+) > > > > diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h > > index 7fc2b129a2e7..bcac769a7df6 100644 > > --- a/fs/xfs/libxfs/xfs_inode_fork.h > > +++ b/fs/xfs/libxfs/xfs_inode_fork.h > > @@ -40,6 +40,13 @@ struct xfs_ifork { > > */ > > #define XFS_IEXT_ADD_NOSPLIT_CNT (1) > > > > +/* > > + * Punching out an extent from the middle of an existing extent can cause the > > + * extent count to increase by 1. > > + * i.e. | Old extent | Hole | Old extent | > > + */ > > +#define XFS_IEXT_PUNCH_HOLE_CNT (1) > > + > > /* > > * Fork handling. > > */ > > diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c > > index ec3691372e7c..5c7d08da8ff1 100644 > > --- a/fs/xfs/xfs_bmap_item.c > > +++ b/fs/xfs/xfs_bmap_item.c > > @@ -519,6 +519,11 @@ xfs_bui_item_recover( > > } > > xfs_trans_ijoin(tp, ip, 0); > > > > + error = xfs_iext_count_may_overflow(ip, whichfork, > > + XFS_IEXT_PUNCH_HOLE_CNT); > > I think this ought to be XFS_IEXT_ADD_NOSPLIT_CNT if bui_type is > XFS_BMAP_MAP and XFS_IEXT_PUNCH_HOLE_CNT if XFS_BMAP_UNMAP. You are right. I will include this change in the next version. > > Whoever created the BUI should have called xfs_iext_count_may_overflow > before logging the BUI (and hence this should never occur) but it does > pay to be careful. :) > > The rest of the logic in the patch looks ok. > > --D > > > + if (error) > > + goto err_inode; > > + > > count = bmap->me_len; > > error = xfs_trans_log_finish_bmap_update(tp, budp, type, ip, whichfork, > > bmap->me_startoff, bmap->me_startblock, &count, state); > > diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c > > index dcd6e61df711..0776abd0103c 100644 > > --- a/fs/xfs/xfs_bmap_util.c > > +++ b/fs/xfs/xfs_bmap_util.c > > @@ -891,6 +891,11 @@ xfs_unmap_extent( > > > > xfs_trans_ijoin(tp, ip, 0); > > > > + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, > > + XFS_IEXT_PUNCH_HOLE_CNT); > > + if (error) > > + goto out_trans_cancel; > > + > > error = xfs_bunmapi(tp, ip, startoffset_fsb, len_fsb, 0, 2, done); > > if (error) > > goto out_trans_cancel; > > @@ -1176,6 +1181,11 @@ xfs_insert_file_space( > > xfs_ilock(ip, XFS_ILOCK_EXCL); > > xfs_trans_ijoin(tp, ip, 0); > > > > + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, > > + XFS_IEXT_PUNCH_HOLE_CNT); > > + if (error) > > + goto out_trans_cancel; > > + > > /* > > * The extent shifting code works on extent granularity. So, if stop_fsb > > * is not the starting block of extent, we need to split the extent at >
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h index 7fc2b129a2e7..bcac769a7df6 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.h +++ b/fs/xfs/libxfs/xfs_inode_fork.h @@ -40,6 +40,13 @@ struct xfs_ifork { */ #define XFS_IEXT_ADD_NOSPLIT_CNT (1) +/* + * Punching out an extent from the middle of an existing extent can cause the + * extent count to increase by 1. + * i.e. | Old extent | Hole | Old extent | + */ +#define XFS_IEXT_PUNCH_HOLE_CNT (1) + /* * Fork handling. */ diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c index ec3691372e7c..5c7d08da8ff1 100644 --- a/fs/xfs/xfs_bmap_item.c +++ b/fs/xfs/xfs_bmap_item.c @@ -519,6 +519,11 @@ xfs_bui_item_recover( } xfs_trans_ijoin(tp, ip, 0); + error = xfs_iext_count_may_overflow(ip, whichfork, + XFS_IEXT_PUNCH_HOLE_CNT); + if (error) + goto err_inode; + count = bmap->me_len; error = xfs_trans_log_finish_bmap_update(tp, budp, type, ip, whichfork, bmap->me_startoff, bmap->me_startblock, &count, state); diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index dcd6e61df711..0776abd0103c 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -891,6 +891,11 @@ xfs_unmap_extent( xfs_trans_ijoin(tp, ip, 0); + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, + XFS_IEXT_PUNCH_HOLE_CNT); + if (error) + goto out_trans_cancel; + error = xfs_bunmapi(tp, ip, startoffset_fsb, len_fsb, 0, 2, done); if (error) goto out_trans_cancel; @@ -1176,6 +1181,11 @@ xfs_insert_file_space( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); + error = xfs_iext_count_may_overflow(ip, XFS_DATA_FORK, + XFS_IEXT_PUNCH_HOLE_CNT); + if (error) + goto out_trans_cancel; + /* * The extent shifting code works on extent granularity. So, if stop_fsb * is not the starting block of extent, we need to split the extent at
The extent mapping the file offset at which a hole has to be inserted will be split into two extents causing extent count to increase by 1. Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com> --- fs/xfs/libxfs/xfs_inode_fork.h | 7 +++++++ fs/xfs/xfs_bmap_item.c | 5 +++++ fs/xfs/xfs_bmap_util.c | 10 ++++++++++ 3 files changed, 22 insertions(+)