This call can fail if there are dirty pages. The preceding call to
filemap_write_and_wait_range() will normally remove dirty pages, but
as inode_lock() is not held over calls to ceph_direct_read_write(), it
could race with non-direct writes and pages could be dirtied
immediately after filemap_write_and_wait_range() returns
If there are dirty pages, they will be removed by the subsequent call
to truncate_inode_pages_range(), so having them here is not a problem.
If the 'ret' value is left holding an error, then in the async IO case
(aio_req is not NULL) the loop that would normally call
ceph_osdc_start_request() will see the error in 'ret' and abort all
requests. This doesn't seem like correct behaviour.
So clear 'ret' and ignore the error (other than the dout() message).
Signed-off-by: NeilBrown <neilb@xxxxxxxx>
---
fs/ceph/file.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 0f5375d8e030..1ca6e29edcc9 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -905,8 +905,14 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
ret = invalidate_inode_pages2_range(inode->i_mapping,
pos >> PAGE_SHIFT,
(pos + count) >> PAGE_SHIFT);
- if (ret < 0)
+ if (ret < 0) {
dout("invalidate_inode_pages2_range returned %d\n", ret);
+ /*
+ * Error is not fatal as we truncate_inode_pages_range()
+ * below.
+ */
+ ret = 0;
+ }
flags = CEPH_OSD_FLAG_ORDERSNAP |
CEPH_OSD_FLAG_ONDISK |
--
2.9.3
Attachment:
signature.asc
Description: PGP signature
