[PATCH 0/6] fixes for cephfs

[PATCH 0/6] fixes for cephfs

[Yan, Zheng]

Hi,

These patchs are fixes for cephfs kernel client bugs when running
2 MDS with thrash_exports enabled.

These patches are also in:
  git://github.com/ukernel/ceph-client.git master

Regards
Yan, Zheng

[PATCH 1/6] ceph: Don't update i_max_size when handling non-auth cap

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

The cap from non-auth mds doesn't have a meaningful max_size value.

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/caps.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 3251e9c..c633d1d 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2388,7 +2388,7 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant,
 			    &atime);
 
 	/* max size increase? */
-	if (max_size != ci->i_max_size) {
+	if (ci->i_auth_cap == cap && max_size != ci->i_max_size) {
 		dout("max_size %lld -> %llu\n", ci->i_max_size, max_size);
 		ci->i_max_size = max_size;
 		if (max_size >= ci->i_wanted_max_size) {
-- 
1.7.11.7

[PATCH 2/6] ceph: Hold caps_list_lock when adjusting caps_{use,total}_count

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/caps.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index c633d1d..8072aef 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -236,8 +236,10 @@ static struct ceph_cap *get_cap(struct ceph_mds_client *mdsc,
 	if (!ctx) {
 		cap = kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS);
 		if (cap) {
+			spin_lock(&mdsc->caps_list_lock);
 			mdsc->caps_use_count++;
 			mdsc->caps_total_count++;
+			spin_unlock(&mdsc->caps_list_lock);
 		}
 		return cap;
 	}
-- 
1.7.11.7

[PATCH 3/6] ceph: Fix infinite loop in __wake_requests

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

__wake_requests() will enter infinite loop if we use it to wake
requests in the session->s_waiting list. __wake_requests() deletes
requests from the list and __do_request() adds requests back to
the list.

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/mds_client.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 1bcf712..0d9864f 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1876,9 +1876,14 @@ finish:
 static void __wake_requests(struct ceph_mds_client *mdsc,
 			    struct list_head *head)
 {
-	struct ceph_mds_request *req, *nreq;
+	struct ceph_mds_request *req;
+	LIST_HEAD(tmp_list);
+
+	list_splice_init(head, &tmp_list);
 
-	list_for_each_entry_safe(req, nreq, head, r_wait) {
+	while (!list_empty(&tmp_list)) {
+		req = list_entry(tmp_list.next,
+				 struct ceph_mds_request, r_wait);
 		list_del_init(&req->r_wait);
 		__do_request(mdsc, req);
 	}
-- 
1.7.11.7

[PATCH 4/6] ceph: Don't add dirty inode to dirty list if caps is in migration

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

Add dirty inode to cap_dirty_migrating list instead, this can avoid
ceph_flush_dirty_caps() entering infinite loop.

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/caps.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 8072aef..5efa3f5 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1351,11 +1351,15 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask)
 		if (!ci->i_head_snapc)
 			ci->i_head_snapc = ceph_get_snap_context(
 				ci->i_snap_realm->cached_context);
-		dout(" inode %p now dirty snapc %p\n", &ci->vfs_inode,
-			ci->i_head_snapc);
+		dout(" inode %p now dirty snapc %p auth cap %p\n",
+		     &ci->vfs_inode, ci->i_head_snapc, ci->i_auth_cap);
 		BUG_ON(!list_empty(&ci->i_dirty_item));
 		spin_lock(&mdsc->cap_dirty_lock);
-		list_add(&ci->i_dirty_item, &mdsc->cap_dirty);
+		if (ci->i_auth_cap)
+			list_add(&ci->i_dirty_item, &mdsc->cap_dirty);
+		else
+			list_add(&ci->i_dirty_item,
+				 &mdsc->cap_dirty_migrating);
 		spin_unlock(&mdsc->cap_dirty_lock);
 		if (ci->i_flushing_caps == 0) {
 			ihold(inode);
-- 
1.7.11.7

[PATCH 5/6] ceph: Fix __ceph_do_pending_vmtruncate

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

we should set i_truncate_pending to 0 after page cache is truncated
to i_truncate_size

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/inode.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 4b5762e..81613bc 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1466,7 +1466,7 @@ void __ceph_do_pending_vmtruncate(struct inode *inode)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	u64 to;
-	int wrbuffer_refs, wake = 0;
+	int wrbuffer_refs, finish = 0;
 
 retry:
 	spin_lock(&ci->i_ceph_lock);
@@ -1498,15 +1498,18 @@ retry:
 	truncate_inode_pages(inode->i_mapping, to);
 
 	spin_lock(&ci->i_ceph_lock);
-	ci->i_truncate_pending--;
-	if (ci->i_truncate_pending == 0)
-		wake = 1;
+	if (to == ci->i_truncate_size) {
+		ci->i_truncate_pending = 0;
+		finish = 1;
+	}
 	spin_unlock(&ci->i_ceph_lock);
+	if (!finish)
+		goto retry;
 
 	if (wrbuffer_refs == 0)
 		ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL);
-	if (wake)
-		wake_up_all(&ci->i_cap_wq);
+
+	wake_up_all(&ci->i_cap_wq);
 }
 
 
-- 
1.7.11.7

[PATCH 6/6] ceph: call handle_cap_grant() for cap import message

[Yan, Zheng]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

If client sends cap message that requests new max size during
exporting caps, the exporting MDS will drop the message quietly.
So the client may wait for the reply that updates the max size
forever. call handle_cap_grant() for cap import message can
avoid this issue.

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 fs/ceph/caps.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 5efa3f5..a1d9bb3 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2751,6 +2751,7 @@ static void handle_cap_import(struct ceph_mds_client *mdsc,
 
 	/* make sure we re-request max_size, if necessary */
 	spin_lock(&ci->i_ceph_lock);
+	ci->i_wanted_max_size = 0;  /* reset */
 	ci->i_requested_max_size = 0;
 	spin_unlock(&ci->i_ceph_lock);
 }
@@ -2846,8 +2847,6 @@ void ceph_handle_caps(struct ceph_mds_session *session,
 	case CEPH_CAP_OP_IMPORT:
 		handle_cap_import(mdsc, inode, h, session,
 				  snaptrace, snaptrace_len);
-		ceph_check_caps(ceph_inode(inode), 0, session);
-		goto done_unlocked;
 	}
 
 	/* the rest require a cap */
@@ -2864,6 +2863,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
 	switch (op) {
 	case CEPH_CAP_OP_REVOKE:
 	case CEPH_CAP_OP_GRANT:
+	case CEPH_CAP_OP_IMPORT:
 		handle_cap_grant(inode, h, session, cap, msg->middle);
 		goto done_unlocked;
 
-- 
1.7.11.7