From cfc0bf6640dfd0f43bf8bfec5a475284809baa4d Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Tue, 14 Sep 2010 15:50:59 -0700 Subject: ceph: stop sending FLUSHSNAPs when we hit a dirty capsnap Stop sending FLUSHSNAP messages when we hit a capsnap that has dirty_pages or is still writing. We'll send the newer capsnaps only after the older ones complete. Signed-off-by: Sage Weil --- fs/ceph/caps.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/ceph/caps.c') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index a2069b6680ae..9fbe9019155c 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1227,7 +1227,7 @@ retry: * pages to be written out. */ if (capsnap->dirty_pages || capsnap->writing) - continue; + break; /* * if cap writeback already occurred, we should have dropped @@ -1276,8 +1276,8 @@ retry: &session->s_cap_snaps_flushing); spin_unlock(&inode->i_lock); - dout("flush_snaps %p cap_snap %p follows %lld size %llu\n", - inode, capsnap, next_follows, capsnap->size); + dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n", + inode, capsnap, capsnap->follows, capsnap->flush_tid); send_cap_msg(session, ceph_vino(inode).ino, 0, CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0, capsnap->dirty, 0, capsnap->flush_tid, 0, mseq, -- cgit v1.2.2 From e835124c2be289515b918f2688ced4249e2de566 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Fri, 17 Sep 2010 08:03:08 -0700 Subject: ceph: only send one flushsnap per cap_snap per mds session Sending multiple flushsnap messages is problematic because we ignore the response if the tid doesn't match, and the server may only respond to each one once. It's also a waste. So, skip cap_snaps that are already on the flushing list, unless the caller tells us to resend (because we are reconnecting). Signed-off-by: Sage Weil --- fs/ceph/caps.c | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) (limited to 'fs/ceph/caps.c') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 9fbe9019155c..b01c316a8148 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1195,10 +1195,14 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, * asynchronously back to the MDS once sync writes complete and dirty * data is written out. * + * Unless @again is true, skip cap_snaps that were already sent to + * the MDS (i.e., during this session). + * * Called under i_lock. Takes s_mutex as needed. */ void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession) + struct ceph_mds_session **psession, + int again) __releases(ci->vfs_inode->i_lock) __acquires(ci->vfs_inode->i_lock) { @@ -1240,6 +1244,13 @@ retry: dout("no auth cap (migrating?), doing nothing\n"); goto out; } + + /* only flush each capsnap once */ + if (!again && !list_empty(&capsnap->flushing_item)) { + dout("already flushed %p, skipping\n", capsnap); + continue; + } + mds = ci->i_auth_cap->session->s_mds; mseq = ci->i_auth_cap->mseq; @@ -1314,7 +1325,7 @@ static void ceph_flush_snaps(struct ceph_inode_info *ci) struct inode *inode = &ci->vfs_inode; spin_lock(&inode->i_lock); - __ceph_flush_snaps(ci, NULL); + __ceph_flush_snaps(ci, NULL, 0); spin_unlock(&inode->i_lock); } @@ -1477,7 +1488,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, /* flush snaps first time around only */ if (!list_empty(&ci->i_cap_snaps)) - __ceph_flush_snaps(ci, &session); + __ceph_flush_snaps(ci, &session, 0); goto retry_locked; retry: spin_lock(&inode->i_lock); @@ -1894,7 +1905,7 @@ static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc, if (cap && cap->session == session) { dout("kick_flushing_caps %p cap %p capsnap %p\n", inode, cap, capsnap); - __ceph_flush_snaps(ci, &session); + __ceph_flush_snaps(ci, &session, 1); } else { pr_err("%p auth cap %p not mds%d ???\n", inode, cap, session->s_mds); -- cgit v1.2.2 From a43fb73101eaf6db0b33d22c152b338ab8b3edbb Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Fri, 17 Sep 2010 09:54:08 -0700 Subject: ceph: check mapping to determine if FILE_CACHE cap is used See if the i_data mapping has any pages to determine if the FILE_CACHE capability is currently in use, instead of assuming it is any time the rdcache_gen value is set (i.e., issued -> used). This allows the MDS RECALL_STATE process work for inodes that have cached pages. Signed-off-by: Sage Weil --- fs/ceph/caps.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/ceph/caps.c') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index b01c316a8148..73c153092f72 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -814,7 +814,7 @@ int __ceph_caps_used(struct ceph_inode_info *ci) used |= CEPH_CAP_PIN; if (ci->i_rd_ref) used |= CEPH_CAP_FILE_RD; - if (ci->i_rdcache_ref || ci->i_rdcache_gen) + if (ci->i_rdcache_ref || ci->vfs_inode.i_data.nrpages) used |= CEPH_CAP_FILE_CACHE; if (ci->i_wr_ref) used |= CEPH_CAP_FILE_WR; -- cgit v1.2.2