Skip to contentSkip to conversation

ceph: Move a variable assignment behind a condition check in __ceph_remove_cap()

PATCH In review Markus Elfring 3 participants 4 messages Jul 13 – Jul 14

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

PATCH 1/1 ceph: Move a variable assignment behind a condition check in __ceph_remove_cap()

From: Markus Elfring <[email protected]>
Date: Mon, 13 Jul 2026 13:21:29 +0200

The address of a data structure member was determined before a corresponding null pointer check in the implementation of the function “__ceph_remove_cap”.

Thus avoid the risk for undefined behaviour by moving the assignment for the variable “inode” behind a condition check.

This issue was detected by using the Coccinelle software.

Fixes 38d46409c4639a1d659ebfa70e27a8bed6b8ee1d ("ceph: print cluster fsid and client global_id in all debug logs")
---
 fs/ceph/caps.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ceph/caps.c +2 -2
index 4b37d9ffdf7f..5b6640707949 100644
@@ -1124,7 +1124,7 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
 	struct ceph_mds_session *session = cap->session;
 	struct ceph_client *cl = session->s_mdsc->fsc->client;
 	struct ceph_inode_info *ci = cap->ci;
-	struct inode *inode = &ci->netfs.inode;
+	struct inode *inode;
 	struct ceph_mds_client *mdsc;
 	int removed = 0;
 
@@ -1135,7 +1135,7 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
 	}
 
 	lockdep_assert_held(&ci->i_ceph_lock);
-
+	inode = &ci->netfs.inode;
 	doutc(cl, "%p from %p %llx.%llx\n", cap, inode, ceph_vinop(inode));
 
 	mdsc = ceph_inode_to_fs_client(&ci->netfs.inode)->mdsc;
Signature · 1 line
2.55.0
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 41 lines · Markus Elfring

From: Markus Elfring <[email protected]>
Date: Mon, 13 Jul 2026 13:21:29 +0200

The address of a data structure member was determined before a corresponding null pointer check in the implementation of the function “__ceph_remove_cap”.

Thus avoid the risk for undefined behaviour by moving the assignment for the variable “inode” behind a condition check.

This issue was detected by using the Coccinelle software.

Fixes 38d46409c4639a1d659ebfa70e27a8bed6b8ee1d ("ceph: print cluster

fsid and client global_id in all debug logs")

---
 fs/ceph/caps.c | 4 ++--

 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ceph/caps.c +1 -1
index 4b37d9ffdf7f..5b6640707949 100644
@@ -1124,7 +1124,7 @@ void __ceph_remove_cap(struct ceph_cap *cap,
 bool queue_release)
  	struct ceph_mds_session *session = cap->session;
  	struct ceph_client *cl = session->s_mdsc->fsc->client;
  	struct ceph_inode_info *ci = cap->ci;
-	struct inode *inode = &ci->netfs.inode;
+	struct inode *inode;
  	struct ceph_mds_client *mdsc;
  	int removed = 0;

fs/ceph/caps.c +0 -0
@@ -1135,7 +1135,7 @@ void __ceph_remove_cap(struct ceph_cap *cap,
 bool queue_release)
  	}


  lockdep_assert_held(&ci->i_ceph_lock);

fs/ceph/caps.c +1 -1
-
+	inode = &ci->netfs.inode;
Quoted 4 lines

  doutc(cl, "%p from %p %llx.%llx\n", cap, inode, ceph_vinop(inode));

  mdsc = ceph_inode_to_fs_client(&ci->netfs.inode)->mdsc;

Makes sense.

Thanks,
Slava.

#2 lore raw
Quoted 11 lines · Viacheslav Dubeyko
Quoted 10 lines · Markus Elfring

From: Markus Elfring <[email protected]>
Date: Mon, 13 Jul 2026 13:21:29 +0200

The address of a data structure member was determined before a corresponding null pointer check in the implementation of the function “__ceph_remove_cap”.

Thus avoid the risk for undefined behaviour by moving the assignment for the variable “inode” behind a condition check.

Quoted 6 lines
Quoted 6 lines

This issue was detected by using the Coccinelle software.

Fixes 38d46409c4639a1d659ebfa70e27a8bed6b8ee1d ("ceph: print cluster

fsid and client global_id in all debug logs")

I have explained to Markus many times that these are not dereferences, they are just pointer math. So the original code works fine and does not need a Fixes tag or a CC to stable.

And then Markus responds, "the C standard says we are not allowed to dereference NULL pointers"... Which is true, but again, it's not a dereference.

regards,
dan carpenter

#3 lore raw
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

The address of a data structure member was determined before a corresponding null pointer check in the implementation of the function “__ceph_remove_cap”.

Thus avoid the risk for undefined behaviour by moving the assignment for the variable “inode” behind a condition check.

This issue was detected by using the Coccinelle software.

Fixes 38d46409c4639a1d659ebfa70e27a8bed6b8ee1d ("ceph: print cluster

fsid and client global_id in all debug logs")

…

Quoted 2 lines

I have explained to Markus many times that these are not dereferences, they are just pointer math.

How does it help to repeat such a questionable development view?

Quoted 6 lines

So the original code works fine and does
not need a Fixes tag or a CC to stable.

And then Markus responds, "the C standard says we are not allowed to dereference NULL pointers"... Which is true, but again, it's not a dereference.

Why did you get special difficulties with adhering to standard specifications also in the discussed case?

Regards,
Markus