2024-06-13 09:06:18

by Zhang Yi

Subject: [PATCH -next v5 6/8] xfs: reserve blocks for truncating large realtime inode

From: Zhang Yi <[email protected]>

For a large realtime inode, __xfs_bunmapi() could split the tail written
extent and convert the later one that beyond EOF block to unwritten, but
it couldn't work as expected on truncate down now since the reserved
block is zero in xfs_setattr_size(), fix this by reserving
XFS_DIOSTRAT_SPACE_RES blocks for large realtime inode.

Signed-off-by: Zhang Yi <[email protected]>
---
fs/xfs/xfs_iops.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index 8e7e6c435fb3..8af13fd37f1b 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -17,6 +17,8 @@
#include "xfs_da_btree.h"
#include "xfs_attr.h"
#include "xfs_trans.h"
+#include "xfs_trans_space.h"
+#include "xfs_bmap_btree.h"
#include "xfs_trace.h"
#include "xfs_icache.h"
#include "xfs_symlink.h"
@@ -913,6 +915,7 @@ xfs_setattr_size(
struct xfs_trans *tp;
int error;
uint lock_flags = 0;
+ uint resblks = 0;

xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL);
ASSERT(S_ISREG(inode->i_mode));
@@ -967,7 +970,17 @@ xfs_setattr_size(
if (error)
return error;

- error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
+ /*
+ * For realtime inode with more than one block rtextsize, we need the
+ * block reservation for bmap btree block allocations/splits that can
+ * happen since it could split the tail written extent and convert the
+ * right beyond EOF one to unwritten.
+ */
+ if (xfs_inode_has_bigrtalloc(ip))
+ resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
+
+ error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, resblks,
+ 0, 0, &tp);
if (error)
return error;

--
2.39.2

2024-06-14 05:59:02

by Christoph Hellwig

Subject: Re: [PATCH -next v5 6/8] xfs: reserve blocks for truncating large realtime inode

On Thu, Jun 13, 2024 at 05:00:31PM +0800, Zhang Yi wrote:
> From: Zhang Yi <[email protected]>
>
> For a large realtime inode, __xfs_bunmapi() could split the tail written
> extent and convert the later one that beyond EOF block to unwritten, but
> it couldn't work as expected on truncate down now since the reserved
> block is zero in xfs_setattr_size(), fix this by reserving
> XFS_DIOSTRAT_SPACE_RES blocks for large realtime inode.

Looks good:

Reviewed-by: Christoph Hellwig <[email protected]>