diff options
author | Sage Weil <sage@newdream.net> | 2009-12-22 13:45:45 -0500 |
---|---|---|
committer | Sage Weil <sage@newdream.net> | 2009-12-23 11:17:20 -0500 |
commit | 350b1c32ea58d29e25d63fc25e92dd48f9339546 (patch) | |
tree | 4b9e8b06de8fd7e07c2292307e9a67c121f1e079 | |
parent | ec302645f4a9bd9ec757c30d185557e1c0972c1a (diff) |
ceph: control access to page vector for incoming data
When we issue an OSD read, we specify a vector of pages that the data is to
be read into. The request may be sent multiple times, to multiple OSDs, if
the osdmap changes, which means we can get more than one reply.
Only read data into the page vector if the reply is coming from the
OSD we last sent the request to. Keep track of which connection is using
the vector by taking a reference. If another connection was already
using the vector before and a new reply comes in on the right connection,
revoke the pages from the other connection.
Signed-off-by: Sage Weil <sage@newdream.net>
-rw-r--r-- | fs/ceph/messenger.c | 29 | ||||
-rw-r--r-- | fs/ceph/messenger.h | 2 | ||||
-rw-r--r-- | fs/ceph/osd_client.c | 42 | ||||
-rw-r--r-- | fs/ceph/osd_client.h | 4 |
4 files changed, 67 insertions, 10 deletions
diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c index c03b4185c143..506b638a023b 100644 --- a/fs/ceph/messenger.c +++ b/fs/ceph/messenger.c | |||
@@ -1976,6 +1976,35 @@ void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg) | |||
1976 | } | 1976 | } |
1977 | 1977 | ||
1978 | /* | 1978 | /* |
1979 | * Revoke a page vector that we may be reading data into | ||
1980 | */ | ||
1981 | void ceph_con_revoke_pages(struct ceph_connection *con, struct page **pages) | ||
1982 | { | ||
1983 | mutex_lock(&con->mutex); | ||
1984 | if (con->in_msg && con->in_msg->pages == pages) { | ||
1985 | unsigned data_len = le32_to_cpu(con->in_hdr.data_len); | ||
1986 | |||
1987 | /* skip rest of message */ | ||
1988 | dout("con_revoke_pages %p msg %p pages %p revoked\n", con, | ||
1989 | con->in_msg, pages); | ||
1990 | if (con->in_msg_pos.data_pos < data_len) | ||
1991 | con->in_base_pos = con->in_msg_pos.data_pos - data_len; | ||
1992 | else | ||
1993 | con->in_base_pos = con->in_base_pos - | ||
1994 | sizeof(struct ceph_msg_header) - | ||
1995 | sizeof(struct ceph_msg_footer); | ||
1996 | con->in_msg->pages = NULL; | ||
1997 | ceph_msg_put(con->in_msg); | ||
1998 | con->in_msg = NULL; | ||
1999 | con->in_tag = CEPH_MSGR_TAG_READY; | ||
2000 | } else { | ||
2001 | dout("con_revoke_pages %p msg %p pages %p no-op\n", | ||
2002 | con, con->in_msg, pages); | ||
2003 | } | ||
2004 | mutex_unlock(&con->mutex); | ||
2005 | } | ||
2006 | |||
2007 | /* | ||
1979 | * Queue a keepalive byte to ensure the tcp connection is alive. | 2008 | * Queue a keepalive byte to ensure the tcp connection is alive. |
1980 | */ | 2009 | */ |
1981 | void ceph_con_keepalive(struct ceph_connection *con) | 2010 | void ceph_con_keepalive(struct ceph_connection *con) |
diff --git a/fs/ceph/messenger.h b/fs/ceph/messenger.h index 94b55de90331..7e2aab1d3ce2 100644 --- a/fs/ceph/messenger.h +++ b/fs/ceph/messenger.h | |||
@@ -230,6 +230,8 @@ extern void ceph_con_open(struct ceph_connection *con, | |||
230 | extern void ceph_con_close(struct ceph_connection *con); | 230 | extern void ceph_con_close(struct ceph_connection *con); |
231 | extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg); | 231 | extern void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg); |
232 | extern void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg); | 232 | extern void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg); |
233 | extern void ceph_con_revoke_pages(struct ceph_connection *con, | ||
234 | struct page **pages); | ||
233 | extern void ceph_con_keepalive(struct ceph_connection *con); | 235 | extern void ceph_con_keepalive(struct ceph_connection *con); |
234 | extern struct ceph_connection *ceph_con_get(struct ceph_connection *con); | 236 | extern struct ceph_connection *ceph_con_get(struct ceph_connection *con); |
235 | extern void ceph_con_put(struct ceph_connection *con); | 237 | extern void ceph_con_put(struct ceph_connection *con); |
diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c index a1800fb63237..374f0013956c 100644 --- a/fs/ceph/osd_client.c +++ b/fs/ceph/osd_client.c | |||
@@ -87,6 +87,13 @@ void ceph_osdc_release_request(struct kref *kref) | |||
87 | ceph_msg_put(req->r_request); | 87 | ceph_msg_put(req->r_request); |
88 | if (req->r_reply) | 88 | if (req->r_reply) |
89 | ceph_msg_put(req->r_reply); | 89 | ceph_msg_put(req->r_reply); |
90 | if (req->r_con_filling_pages) { | ||
91 | dout("release_request revoking pages %p from con %p\n", | ||
92 | req->r_pages, req->r_con_filling_pages); | ||
93 | ceph_con_revoke_pages(req->r_con_filling_pages, | ||
94 | req->r_pages); | ||
95 | ceph_con_put(req->r_con_filling_pages); | ||
96 | } | ||
90 | if (req->r_own_pages) | 97 | if (req->r_own_pages) |
91 | ceph_release_page_vector(req->r_pages, | 98 | ceph_release_page_vector(req->r_pages, |
92 | req->r_num_pages); | 99 | req->r_num_pages); |
@@ -687,7 +694,8 @@ static void handle_timeout(struct work_struct *work) | |||
687 | * handle osd op reply. either call the callback if it is specified, | 694 | * handle osd op reply. either call the callback if it is specified, |
688 | * or do the completion to wake up the waiting thread. | 695 | * or do the completion to wake up the waiting thread. |
689 | */ | 696 | */ |
690 | static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) | 697 | static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, |
698 | struct ceph_connection *con) | ||
691 | { | 699 | { |
692 | struct ceph_osd_reply_head *rhead = msg->front.iov_base; | 700 | struct ceph_osd_reply_head *rhead = msg->front.iov_base; |
693 | struct ceph_osd_request *req; | 701 | struct ceph_osd_request *req; |
@@ -715,6 +723,16 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) | |||
715 | ceph_osdc_get_request(req); | 723 | ceph_osdc_get_request(req); |
716 | flags = le32_to_cpu(rhead->flags); | 724 | flags = le32_to_cpu(rhead->flags); |
717 | 725 | ||
726 | /* | ||
727 | * if this connection filled our pages, drop our reference now, to | ||
728 | * avoid a (safe but slower) revoke later. | ||
729 | */ | ||
730 | if (req->r_con_filling_pages == con && req->r_pages == msg->pages) { | ||
731 | dout(" got pages, dropping con_filling_pages ref %p\n", con); | ||
732 | req->r_con_filling_pages = NULL; | ||
733 | ceph_con_put(con); | ||
734 | } | ||
735 | |||
718 | if (req->r_reply) { | 736 | if (req->r_reply) { |
719 | /* | 737 | /* |
720 | * once we see the message has been received, we don't | 738 | * once we see the message has been received, we don't |
@@ -1007,14 +1025,20 @@ static int prepare_pages(struct ceph_connection *con, struct ceph_msg *m, | |||
1007 | } | 1025 | } |
1008 | dout("prepare_pages tid %llu has %d pages, want %d\n", | 1026 | dout("prepare_pages tid %llu has %d pages, want %d\n", |
1009 | tid, req->r_num_pages, want); | 1027 | tid, req->r_num_pages, want); |
1010 | if (likely(req->r_num_pages >= want && !req->r_prepared_pages)) { | 1028 | if (unlikely(req->r_num_pages < want)) |
1011 | m->pages = req->r_pages; | 1029 | goto out; |
1012 | m->nr_pages = req->r_num_pages; | 1030 | |
1013 | req->r_reply = m; /* only for duration of read over socket */ | 1031 | if (req->r_con_filling_pages) { |
1014 | ceph_msg_get(m); | 1032 | dout("revoking pages %p from old con %p\n", req->r_pages, |
1015 | req->r_prepared_pages = 1; | 1033 | req->r_con_filling_pages); |
1016 | ret = 0; /* success */ | 1034 | ceph_con_revoke_pages(req->r_con_filling_pages, req->r_pages); |
1035 | ceph_con_put(req->r_con_filling_pages); | ||
1017 | } | 1036 | } |
1037 | req->r_con_filling_pages = ceph_con_get(con); | ||
1038 | req->r_reply = ceph_msg_get(m); /* for duration of read over socket */ | ||
1039 | m->pages = req->r_pages; | ||
1040 | m->nr_pages = req->r_num_pages; | ||
1041 | ret = 0; /* success */ | ||
1018 | out: | 1042 | out: |
1019 | mutex_unlock(&osdc->request_mutex); | 1043 | mutex_unlock(&osdc->request_mutex); |
1020 | return ret; | 1044 | return ret; |
@@ -1269,7 +1293,7 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) | |||
1269 | ceph_osdc_handle_map(osdc, msg); | 1293 | ceph_osdc_handle_map(osdc, msg); |
1270 | break; | 1294 | break; |
1271 | case CEPH_MSG_OSD_OPREPLY: | 1295 | case CEPH_MSG_OSD_OPREPLY: |
1272 | handle_reply(osdc, msg); | 1296 | handle_reply(osdc, msg, con); |
1273 | break; | 1297 | break; |
1274 | 1298 | ||
1275 | default: | 1299 | default: |
diff --git a/fs/ceph/osd_client.h b/fs/ceph/osd_client.h index 2e4cfd1e9f10..8fef71cc4457 100644 --- a/fs/ceph/osd_client.h +++ b/fs/ceph/osd_client.h | |||
@@ -43,11 +43,13 @@ struct ceph_osd_request { | |||
43 | struct list_head r_osd_item; | 43 | struct list_head r_osd_item; |
44 | struct ceph_osd *r_osd; | 44 | struct ceph_osd *r_osd; |
45 | 45 | ||
46 | struct ceph_connection *r_con_filling_pages; | ||
47 | |||
46 | struct ceph_msg *r_request, *r_reply; | 48 | struct ceph_msg *r_request, *r_reply; |
47 | int r_result; | 49 | int r_result; |
48 | int r_flags; /* any additional flags for the osd */ | 50 | int r_flags; /* any additional flags for the osd */ |
49 | u32 r_sent; /* >0 if r_request is sending/sent */ | 51 | u32 r_sent; /* >0 if r_request is sending/sent */ |
50 | int r_prepared_pages, r_got_reply; | 52 | int r_got_reply; |
51 | int r_num_prealloc_reply; | 53 | int r_num_prealloc_reply; |
52 | 54 | ||
53 | struct ceph_osd_client *r_osdc; | 55 | struct ceph_osd_client *r_osdc; |