ceph: explicitly specify page alignment in network messages

The alignment used for reading data into or out of pages used to be taken
from the data_off field in the message header.  This only worked as long
as the page alignment matched the object offset, breaking direct io to
non-page aligned offsets.

Instead, explicitly specify the page alignment next to the page vector
in the ceph_msg struct, and use that instead of the message header (which
probably shouldn't be trusted).  The alloc_msg callback is responsible for
filling in this field properly when it sets up the page vector.

Signed-off-by: Sage Weil <sage@newdream.net>

1 files changed, 5 insertions, 5 deletions

diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index d379abf873bc..1c7a2ec4f3cc 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -540,8 +540,7 @@ static void prepare_write_message(struct ceph_connection *con)
                 /* initialize page iterator */
                 con->out_msg_pos.page = 0;
                 if (m->pages)
-                        con->out_msg_pos.page_pos =
-                                le16_to_cpu(m->hdr.data_off) & ~PAGE_MASK;
+                        con->out_msg_pos.page_pos = m->page_alignment;
                 else
                         con->out_msg_pos.page_pos = 0;
                 con->out_msg_pos.data_pos = 0;
@@ -1491,7 +1490,7 @@ static int read_partial_message(struct ceph_connection *con)
         struct ceph_msg *m = con->in_msg;
         int ret;
         int to, left;
-        unsigned front_len, middle_len, data_len, data_off;
+        unsigned front_len, middle_len, data_len;
         int datacrc = con->msgr->nocrc;
         int skip;
         u64 seq;
@@ -1527,7 +1526,6 @@ static int read_partial_message(struct ceph_connection *con)
         data_len = le32_to_cpu(con->in_hdr.data_len);
         if (data_len > CEPH_MSG_MAX_DATA_LEN)
                 return -EIO;
-        data_off = le16_to_cpu(con->in_hdr.data_off);
 
         /* verify seq# */
         seq = le64_to_cpu(con->in_hdr.seq);
@@ -1575,7 +1573,7 @@ static int read_partial_message(struct ceph_connection *con)
 
                 con->in_msg_pos.page = 0;
                 if (m->pages)
-                        con->in_msg_pos.page_pos = data_off & ~PAGE_MASK;
+                        con->in_msg_pos.page_pos = m->page_alignment;
                 else
                         con->in_msg_pos.page_pos = 0;
                 con->in_msg_pos.data_pos = 0;
@@ -2300,6 +2298,7 @@ struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags)
 
         /* data */
         m->nr_pages = 0;
+        m->page_alignment = 0;
         m->pages = NULL;
         m->pagelist = NULL;
         m->bio = NULL;
@@ -2369,6 +2368,7 @@ static struct ceph_msg *ceph_alloc_msg(struct ceph_connection *con,
                                type, front_len);
                         return NULL;
                 }
+                msg->page_alignment = le16_to_cpu(hdr->data_off);
         }
         memcpy(&msg->hdr, &con->in_hdr, sizeof(con->in_hdr));