Thanks, Dan. I'll send a fix.
On Fri, Aug 15, 2025 at 10:50 AM Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote:
>
> Hello Alex Markuze,
>
> Commit a69ac54928a4 ("ceph: fix race condition where r_parent becomes
> stale before sending message") from Aug 12, 2025 (linux-next), leads
> to the following Smatch static checker warning:
>
> fs/ceph/inode.c:1591 ceph_fill_trace()
> error: 'dir' dereferencing possible ERR_PTR()
>
> fs/ceph/inode.c
> 78 static struct inode *ceph_get_reply_dir(struct super_block *sb,
> 79 struct inode *parent,
> 80 struct ceph_mds_reply_info_parsed *rinfo)
> 81 {
> 82 struct ceph_vino vino;
> 83
> 84 if (unlikely(!rinfo->diri.in))
> 85 return parent; /* nothing to compare against */
> 86
> 87 /* If we didn't have a cached parent inode to begin with, just bail out. */
> 88 if (!parent)
> 89 return NULL;
> ^^^^^^^^^^^^
> This returns NULL
>
> 90
> 91 vino.ino = le64_to_cpu(rinfo->diri.in->ino);
> 92 vino.snap = le64_to_cpu(rinfo->diri.in->snapid);
> 93
> 94 if (likely(ceph_vino_matches_parent(parent, vino)))
> 95 return parent; /* matches – use the original reference */
> 96
> 97 /* Mismatch – this should be rare. Emit a WARN and obtain the correct inode. */
> 98 WARN_ONCE(1, "ceph: reply dir mismatch (parent valid %llx.%llx reply %llx.%llx)\n",
> 99 ceph_ino(parent), ceph_snap(parent), vino.ino, vino.snap);
> 100
> 101 return ceph_get_inode(sb, vino, NULL);
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> This returns error pointers, but the caller only checks for NULL.
>
> 102 }
>
> regards,
> dan carpenter
>
