Skip to contentSkip to conversation

ceph: correctly release memory from capsnap

PATCH Superseded [email protected] 2 participants 3 messages Aug 17, 2021 – Aug 18, 2021 Ceph client

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 ceph: correctly release memory from capsnap

From: Xiubo Li <[email protected]>

When force umounting, it will try to remove all the session caps. If there has any capsnap is in the flushing list, the remove session caps callback will try to release the capsnap->flush_cap memory to "ceph_cap_flush_cachep" slab cache, while which is allocated from kmalloc-256 slab cache.

At the same time switch to list_del_init() because just in case the force umount has removed it from the lists and the handle_cap_flushsnap_ack() comes then the seconds list_del_init() won't crash the kernel.

URL: https://tracker.ceph.com/issues/52283

---

- Only to resolve the crash issue.
- s/list_del/list_del_init/

 fs/ceph/caps.c       | 18 ++++++++++++++----
 fs/ceph/mds_client.c |  7 ++++---
 2 files changed, 18 insertions(+), 7 deletions(-)
fs/ceph/caps.c +14 -4
index 990258cbd836..4ee0ef87130a 100644
@@ -1667,7 +1667,16 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask,
 
 struct ceph_cap_flush *ceph_alloc_cap_flush(void)
 {
-	return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	struct ceph_cap_flush *cf;
+
+	cf = kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	/*
+	 * caps == 0 always means for the capsnap
+	 * caps > 0 means dirty caps being flushed
+	 * caps == -1 means preallocated, not used yet
+	 */
+	cf->caps = -1;
+	return cf;
 }
 
 void ceph_free_cap_flush(struct ceph_cap_flush *cf)
@@ -1704,14 +1713,14 @@ static bool __finish_cap_flush(struct ceph_mds_client *mdsc,
 			prev->wake = true;
 			wake = false;
 		}
-		list_del(&cf->g_list);
+		list_del_init(&cf->g_list);
 	} else if (ci) {
 		if (wake && cf->i_list.prev != &ci->i_cap_flush_list) {
 			prev = list_prev_entry(cf, i_list);
 			prev->wake = true;
 			wake = false;
 		}
-		list_del(&cf->i_list);
+		list_del_init(&cf->i_list);
 	} else {
 		BUG_ON(1);
 	}
@@ -3398,7 +3407,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
 		cf = list_first_entry(&to_remove,
 				      struct ceph_cap_flush, i_list);
 		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
 	}
 
 	if (wake_ci)
fs/ceph/mds_client.c +4 -3
index 2c3d762c7973..dc30f56115fa 100644
@@ -1226,7 +1226,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		spin_lock(&mdsc->cap_dirty_lock);
 
 		list_for_each_entry(cf, &to_remove, i_list)
-			list_del(&cf->g_list);
+			list_del_init(&cf->g_list);
 
 		if (!list_empty(&ci->i_dirty_item)) {
 			pr_warn_ratelimited(
@@ -1266,8 +1266,9 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		struct ceph_cap_flush *cf;
 		cf = list_first_entry(&to_remove,
 				      struct ceph_cap_flush, i_list);
-		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		list_del_init(&cf->i_list);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
 	}
 
 	wake_up_all(&ci->i_cap_wq);
Signature · 1 line
2.27.0
#1 lore raw
Quoted 89 lines · On Tue, 2021-08-17 at 20:35 +0800, [email protected] wrote:

From: Xiubo Li <[email protected]>

When force umounting, it will try to remove all the session caps. If there has any capsnap is in the flushing list, the remove session caps callback will try to release the capsnap->flush_cap memory to "ceph_cap_flush_cachep" slab cache, while which is allocated from kmalloc-256 slab cache.

At the same time switch to list_del_init() because just in case the force umount has removed it from the lists and the handle_cap_flushsnap_ack() comes then the seconds list_del_init() won't crash the kernel.

URL: https://tracker.ceph.com/issues/52283

---

- Only to resolve the crash issue.
- s/list_del/list_del_init/

 fs/ceph/caps.c       | 18 ++++++++++++++----
 fs/ceph/mds_client.c |  7 ++++---
 2 files changed, 18 insertions(+), 7 deletions(-)
fs/ceph/caps.c +14 -4
index 990258cbd836..4ee0ef87130a 100644
@@ -1667,7 +1667,16 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask,
 
 struct ceph_cap_flush *ceph_alloc_cap_flush(void)
 {
-	return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	struct ceph_cap_flush *cf;
+
+	cf = kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	/*
+	 * caps == 0 always means for the capsnap
+	 * caps > 0 means dirty caps being flushed
+	 * caps == -1 means preallocated, not used yet
+	 */
+	cf->caps = -1;
+	return cf;
 }
 
 void ceph_free_cap_flush(struct ceph_cap_flush *cf)
@@ -1704,14 +1713,14 @@ static bool __finish_cap_flush(struct ceph_mds_client *mdsc,
 			prev->wake = true;
 			wake = false;
 		}
-		list_del(&cf->g_list);
+		list_del_init(&cf->g_list);
 	} else if (ci) {
 		if (wake && cf->i_list.prev != &ci->i_cap_flush_list) {
 			prev = list_prev_entry(cf, i_list);
 			prev->wake = true;
 			wake = false;
 		}
-		list_del(&cf->i_list);
+		list_del_init(&cf->i_list);
 	} else {
 		BUG_ON(1);
 	}
@@ -3398,7 +3407,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
 		cf = list_first_entry(&to_remove,
 				      struct ceph_cap_flush, i_list);
 		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
 	}
 
 	if (wake_ci)
fs/ceph/mds_client.c +1 -1
index 2c3d762c7973..dc30f56115fa 100644
@@ -1226,7 +1226,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		spin_lock(&mdsc->cap_dirty_lock);
 
 		list_for_each_entry(cf, &to_remove, i_list)
-			list_del(&cf->g_list);
+			list_del_init(&cf->g_list);
 
 		if (!list_empty(&ci->i_dirty_item)) {
 			pr_warn_ratelimited(
@@ -1266,8 +1266,9 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		struct ceph_cap_flush *cf;

fs/ceph/mds_client.c static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, · line 1274

Quoted 10 lines
fs/ceph/mds_client.c +3 -2
 		cf = list_first_entry(&to_remove,
 				      struct ceph_cap_flush, i_list);
-		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		list_del_init(&cf->i_list);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
 	}
 
 	wake_up_all(&ci->i_cap_wq);

This patch doesn't seem to apply to the current testing branch. Is it against an older tree?

Signature · 1 line
Jeff Layton <[email protected]>
author
#2 lore raw
Quoted 99 lines · Jeff Layton
Quoted 98 lines · On Tue, 2021-08-17 at 20:35 +0800, [email protected] wrote:

From: Xiubo Li <[email protected]>

When force umounting, it will try to remove all the session caps. If there has any capsnap is in the flushing list, the remove session caps callback will try to release the capsnap->flush_cap memory to "ceph_cap_flush_cachep" slab cache, while which is allocated from kmalloc-256 slab cache.

At the same time switch to list_del_init() because just in case the force umount has removed it from the lists and the handle_cap_flushsnap_ack() comes then the seconds list_del_init() won't crash the kernel.

URL: https://tracker.ceph.com/issues/52283

---

- Only to resolve the crash issue.
- s/list_del/list_del_init/

  fs/ceph/caps.c       | 18 ++++++++++++++----
  fs/ceph/mds_client.c |  7 ++++---
  2 files changed, 18 insertions(+), 7 deletions(-)
fs/ceph/caps.c +14 -4
index 990258cbd836..4ee0ef87130a 100644
@@ -1667,7 +1667,16 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask,
 
  struct ceph_cap_flush *ceph_alloc_cap_flush(void)
  {
-	return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	struct ceph_cap_flush *cf;
+
+	cf = kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL);
+	/*
+	 * caps == 0 always means for the capsnap
+	 * caps > 0 means dirty caps being flushed
+	 * caps == -1 means preallocated, not used yet
+	 */
+	cf->caps = -1;
+	return cf;
  }
 
  void ceph_free_cap_flush(struct ceph_cap_flush *cf)
@@ -1704,14 +1713,14 @@ static bool __finish_cap_flush(struct ceph_mds_client *mdsc,
  			prev->wake = true;
  			wake = false;
  		}
-		list_del(&cf->g_list);
+		list_del_init(&cf->g_list);
  	} else if (ci) {
  		if (wake && cf->i_list.prev != &ci->i_cap_flush_list) {
  			prev = list_prev_entry(cf, i_list);
  			prev->wake = true;
  			wake = false;
  		}
-		list_del(&cf->i_list);
+		list_del_init(&cf->i_list);
  	} else {
  		BUG_ON(1);
  	}
@@ -3398,7 +3407,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
  		cf = list_first_entry(&to_remove,
  				      struct ceph_cap_flush, i_list);
  		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
  	}
 
  	if (wake_ci)
fs/ceph/mds_client.c +4 -3
index 2c3d762c7973..dc30f56115fa 100644
@@ -1226,7 +1226,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
  		spin_lock(&mdsc->cap_dirty_lock);
 
  		list_for_each_entry(cf, &to_remove, i_list)
-			list_del(&cf->g_list);
+			list_del_init(&cf->g_list);
 
  		if (!list_empty(&ci->i_dirty_item)) {
  			pr_warn_ratelimited(
@@ -1266,8 +1266,9 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
  		struct ceph_cap_flush *cf;
  		cf = list_first_entry(&to_remove,
  				      struct ceph_cap_flush, i_list);
-		list_del(&cf->i_list);
-		ceph_free_cap_flush(cf);
+		list_del_init(&cf->i_list);
+		if (cf->caps)
+			ceph_free_cap_flush(cf);
  	}
 
Quoted 3 lines

fs/ceph/mds_client.c static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, · line 1274

Quoted 1 line
fs/ceph/mds_client.c +0 -0
  	wake_up_all(&ci->i_cap_wq);

This patch doesn't seem to apply to the current testing branch. Is it against an older tree?

I was trying to test the old tags and forgot to switch it back.

Have send out the V3 to fix it.

Thanks