Skip to contentSkip to conversation

fs/ocfs2: fix race in ocfs2_dentry_attach_lock

PATCH Applied Wengang Wang 3 participants 4 messages May 29, 2019 – May 30, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 fs/ocfs2: fix race in ocfs2_dentry_attach_lock

ocfs2_dentry_attach_lock() can be executed in parallel threads against the same dentry. Make that race safe.
The race is like this:

thread A thread B

(A1) enter ocfs2_dentry_attach_lock,
seeing dentry->d_fsdata is NULL,
and no alias found by
ocfs2_find_local_alias, so kmalloc
a new ocfs2_dentry_lock structure
to local variable "dl", dl1

.....

(B1) enter ocfs2_dentry_attach_lock, seeing dentry->d_fsdata is NULL, and no alias found by ocfs2_find_local_alias so kmalloc a new ocfs2_dentry_lock structure to local variable "dl", dl2.

......

(A2) set dentry->d_fsdata with dl1,
call ocfs2_dentry_lock() and increase
dl1->dl_lockres.l_ro_holders to 1 on
success.

              ......

                                    (B2) set dentry->d_fsdata with dl2
                                    call ocfs2_dentry_lock() and increase
				    dl2->dl_lockres.l_ro_holders to 1 on
				    success.

                                                  ......

(A3) call ocfs2_dentry_unlock()
and decrease
dl2->dl_lockres.l_ro_holders to 0
on success.

             ....

                                    (B3) call ocfs2_dentry_unlock(),
                                    decreasing
				    dl2->dl_lockres.l_ro_holders, but
				    see it's zero now, panic

---
v4: return in place on race detection.

v3: add Reviewed-by, Reported-by and Tested-by only

v2: 1) removed lock on dentry_attach_lock at the first access of
dentry->d_fsdata since it helps very little.
2) do cleanups before freeing the duplicated dl
3) return after freeing the duplicated dl found.
---

 fs/ocfs2/dcache.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
fs/ocfs2/dcache.c +12 -0
index 290373024d9d..e8ace3b54e9c 100644
@@ -310,6 +310,18 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry,
 
 out_attach:
 	spin_lock(&dentry_attach_lock);
+	if (unlikely(dentry->d_fsdata && !alias)) {
+		/* d_fsdata is set by a racing thread which is doing
+		 * the same thing as this thread is doing. Leave the racing
+		 * thread going ahead and we return here.
+		 */
+		spin_unlock(&dentry_attach_lock);
+		iput(dl->dl_inode);
+		ocfs2_lock_res_free(&dl->dl_lockres);
+		kfree(dl);
+		return 0;
+	}
+
 	dentry->d_fsdata = dl;
 	dl->dl_count++;
 	spin_unlock(&dentry_attach_lock);
Signature · 1 line
2.13.6
author
#1 lore raw

PATCH v4 1/1 fs/ocfs2: fix race in ocfs2_dentry_attach_lock

Hi,

This is a minor change from v3. It returns in place when racing is detected.

Changwei and Daniel,

I added your Reviewed-by and/or Tested-by with the patch, if you don't think it's good, pls let me know.

thanks,
wengang

Quoted 89 lines · Wengang Wang

ocfs2_dentry_attach_lock() can be executed in parallel threads against the same dentry. Make that race safe.
The race is like this:

thread A thread B

(A1) enter ocfs2_dentry_attach_lock,
seeing dentry->d_fsdata is NULL,
and no alias found by
ocfs2_find_local_alias, so kmalloc
a new ocfs2_dentry_lock structure
to local variable "dl", dl1

.....

(B1) enter ocfs2_dentry_attach_lock, seeing dentry->d_fsdata is NULL, and no alias found by ocfs2_find_local_alias so kmalloc a new ocfs2_dentry_lock structure to local variable "dl", dl2.

......

(A2) set dentry->d_fsdata with dl1,
call ocfs2_dentry_lock() and increase
dl1->dl_lockres.l_ro_holders to 1 on
success.

               ......

                                     (B2) set dentry->d_fsdata with dl2
                                     call ocfs2_dentry_lock() and increase
				    dl2->dl_lockres.l_ro_holders to 1 on
				    success.

                                                   ......

(A3) call ocfs2_dentry_unlock()
and decrease
dl2->dl_lockres.l_ro_holders to 0
on success.

              ....

                                     (B3) call ocfs2_dentry_unlock(),
                                     decreasing
				    dl2->dl_lockres.l_ro_holders, but
				    see it's zero now, panic

---
v4: return in place on race detection.

v3: add Reviewed-by, Reported-by and Tested-by only

v2: 1) removed lock on dentry_attach_lock at the first access of
dentry->d_fsdata since it helps very little.
2) do cleanups before freeing the duplicated dl
3) return after freeing the duplicated dl found.
---

  fs/ocfs2/dcache.c | 12 ++++++++++++
  1 file changed, 12 insertions(+)
fs/ocfs2/dcache.c +12 -0
index 290373024d9d..e8ace3b54e9c 100644
@@ -310,6 +310,18 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry,
 
  out_attach:
  	spin_lock(&dentry_attach_lock);
+	if (unlikely(dentry->d_fsdata && !alias)) {
+		/* d_fsdata is set by a racing thread which is doing
+		 * the same thing as this thread is doing. Leave the racing
+		 * thread going ahead and we return here.
+		 */
+		spin_unlock(&dentry_attach_lock);
+		iput(dl->dl_inode);
+		ocfs2_lock_res_free(&dl->dl_lockres);
+		kfree(dl);
+		return 0;
+	}
+
  	dentry->d_fsdata = dl;
  	dl->dl_count++;
  	spin_unlock(&dentry_attach_lock);
maintainer
#2 lore raw

PATCH v4 1/1 fs/ocfs2: fix race in ocfs2_dentry_attach_lock

In reply to Wengang Wang
Quoted 57 lines · Wengang Wang

ocfs2_dentry_attach_lock() can be executed in parallel threads against the same dentry. Make that race safe.
The race is like this:

thread A thread B

(A1) enter ocfs2_dentry_attach_lock,
seeing dentry->d_fsdata is NULL,
and no alias found by
ocfs2_find_local_alias, so kmalloc
a new ocfs2_dentry_lock structure
to local variable "dl", dl1

.....

(B1) enter ocfs2_dentry_attach_lock, seeing dentry->d_fsdata is NULL, and no alias found by ocfs2_find_local_alias so kmalloc a new ocfs2_dentry_lock structure to local variable "dl", dl2.

......

(A2) set dentry->d_fsdata with dl1,
call ocfs2_dentry_lock() and increase
dl1->dl_lockres.l_ro_holders to 1 on
success.

              ......

                                    (B2) set dentry->d_fsdata with dl2
                                    call ocfs2_dentry_lock() and increase
				    dl2->dl_lockres.l_ro_holders to 1 on
				    success.

                                                  ......

(A3) call ocfs2_dentry_unlock()
and decrease
dl2->dl_lockres.l_ro_holders to 0
on success.

             ....

                                    (B3) call ocfs2_dentry_unlock(),
                                    decreasing
				    dl2->dl_lockres.l_ro_holders, but
				    see it's zero now, panic

---
v4: return in place on race detection.

v3: add Reviewed-by, Reported-by and Tested-by only

Quoted 8 lines

v2: 1) removed lock on dentry_attach_lock at the first access of
dentry->d_fsdata since it helps very little.
2) do cleanups before freeing the duplicated dl
3) return after freeing the duplicated dl found.
---

 fs/ocfs2/dcache.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

<formletter>

This is not the correct way to submit patches for inclusion in the stable kernel tree. Please read:
https://urldefense.proofpoint.com/v2/url?u=https-3A__www.kernel.org_doc_html_latest_process_stable-2Dkernel-2Drules.html&d=DwIBAg&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=C7gAd4uDxlAvTdc0vmU6X8CMk6L2iDY8-HD0qT6Fo7Y&m=k2_veS6gfouGEyAl_KYtO3stkXp1T-mThycjG-C3nVQ&s=qEdFSVeGJPPVIB9XpFnnu7fibsroE5YUT5oujKq10b8&e=
for how to do this properly.

</formletter>

#3 lore raw

PATCH v4 1/1 fs/ocfs2: fix race in ocfs2_dentry_attach_lock

In reply to Wengang Wang
Quoted 48 lines · Wengang Wang

ocfs2_dentry_attach_lock() can be executed in parallel threads against the same dentry. Make that race safe.
The race is like this:

thread A thread B

(A1) enter ocfs2_dentry_attach_lock,
seeing dentry->d_fsdata is NULL,
and no alias found by
ocfs2_find_local_alias, so kmalloc
a new ocfs2_dentry_lock structure
to local variable "dl", dl1

.....

(B1) enter ocfs2_dentry_attach_lock, seeing dentry->d_fsdata is NULL, and no alias found by ocfs2_find_local_alias so kmalloc a new ocfs2_dentry_lock structure to local variable "dl", dl2.

......

(A2) set dentry->d_fsdata with dl1,
call ocfs2_dentry_lock() and increase
dl1->dl_lockres.l_ro_holders to 1 on
success.

              ......

                                    (B2) set dentry->d_fsdata with dl2
                                    call ocfs2_dentry_lock() and increase
				    dl2->dl_lockres.l_ro_holders to 1 on
				    success.

                                                  ......

(A3) call ocfs2_dentry_unlock()
and decrease
dl2->dl_lockres.l_ro_holders to 0
on success.

             ....

                                    (B3) call ocfs2_dentry_unlock(),
                                    decreasing
				    dl2->dl_lockres.l_ro_holders, but
				    see it's zero now, panic
Quoted 4 lines

Looks fine.

Quoted 38 lines

---
v4: return in place on race detection.

v3: add Reviewed-by, Reported-by and Tested-by only

v2: 1) removed lock on dentry_attach_lock at the first access of
dentry->d_fsdata since it helps very little.
2) do cleanups before freeing the duplicated dl
3) return after freeing the duplicated dl found.
---

 fs/ocfs2/dcache.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
fs/ocfs2/dcache.c +12 -0
index 290373024d9d..e8ace3b54e9c 100644
@@ -310,6 +310,18 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry,
 
 out_attach:
 	spin_lock(&dentry_attach_lock);
+	if (unlikely(dentry->d_fsdata && !alias)) {
+		/* d_fsdata is set by a racing thread which is doing
+		 * the same thing as this thread is doing. Leave the racing
+		 * thread going ahead and we return here.
+		 */
+		spin_unlock(&dentry_attach_lock);
+		iput(dl->dl_inode);
+		ocfs2_lock_res_free(&dl->dl_lockres);
+		kfree(dl);
+		return 0;
+	}
+
 	dentry->d_fsdata = dl;
 	dl->dl_count++;
 	spin_unlock(&dentry_attach_lock);