[Linux-cachefs] [PATCH v2] ceph: clamp length of an OSD read to rsize mount option

[Jeff Layton]

There's a hard limit on how large a read we can do from the OSD, defined
as CEPH_MSG_MAX_DATA_LEN (currently, 64M). It's possible to create a
file that is backed by larger objects than that (and indeed, xfstest
ceph/001 does just that).

Ensure we clamp the final length of a read to the rsize, which defaults
to CEPH_MSG_MAX_DATA_LEN, but can be set lower.

Signed-off-by: Jeff Layton <jlayton at kernel.org>
---
 fs/ceph/addr.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

I think this version is more correct. Again, I'll plan to roll this into
the earlier patch that adds the clamp_length op.

diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 9939100f9f9d..c1570fada3d8 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -205,6 +205,7 @@ static void ceph_netfs_expand_readahead(struct netfs_read_request *rreq)
 static bool ceph_netfs_clamp_length(struct netfs_read_subrequest *subreq)
 {
 	struct inode *inode = subreq->rreq->mapping->host;
+	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	u64 objno, objoff;
 	u32 xlen;
@@ -212,7 +213,7 @@ static bool ceph_netfs_clamp_length(struct netfs_read_subrequest *subreq)
 	/* Truncate the extent at the end of the current block */
 	ceph_calc_file_object_mapping(&ci->i_layout, subreq->start, subreq->len,
 				      &objno, &objoff, &xlen);
-	subreq->len = xlen;
+	subreq->len = min(xlen, fsc->mount_options->rsize);
 	return true;
 }
 
-- 
2.31.1