On 5/17/21 10:48 AM, Jan Kara wrote:
Hi!
Thanks for the patch but I think you missed Amir's comments to your v2 [1]?
Also ...
I think they came in after I sent v3, but I got mailinglist signup spammed last
week, so needless to say my ability to deal with emails for the last week has
been hampered a bit. I'm going through his review and will send out v4
today/tomorrow.
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index 42e5a766d33c..714e6f9b74f5 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -617,12 +617,17 @@ void *devpts_get_priv(struct dentry *dentry)
*/
void devpts_pty_kill(struct dentry *dentry)
{
+ struct inode *dir = d_inode(dentry->d_parent);
+ struct inode *inode = d_inode(dentry);
+
WARN_ON_ONCE(dentry->d_sb->s_magic != DEVPTS_SUPER_MAGIC);
+ ihold(inode);
dentry->d_fsdata = NULL;
drop_nlink(dentry->d_inode);
- fsnotify_unlink(d_inode(dentry->d_parent), dentry);
d_drop(dentry);
+ fsnotify_delete(dir, dentry, inode);
+ iput(inode);
dput(dentry); /* d_alloc_name() in devpts_pty_new() */
}
AFAICT d_drop() actually doesn't make the dentry negative so there's no
need for this inode reference game? And similarly for d_invalidate() below?
Or am I missing something?
Nope you're right, I had it in my head that ___d_drop would do the
dentry_unlink_inode, but it doesn't, I'll fix this, thanks,
Josef
