On Mon, Jul 02, 2018 at 10:01:25PM -0700, Linus Torvalds wrote:
On Mon, Jul 2, 2018 at 9:43 PM Seung-Woo Kim sw0312.kim@samsung.com wrote:
I think the commit itself is required. Simple, but not reliable, workaround fix is like below:
diff --git a/fs/dcache.c b/fs/dcache.c index a34d401..7c751f2 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1879,6 +1879,8 @@ void d_instantiate_new(struct dentry *entry, struct inode *inode) BUG_ON(!hlist_unhashed(&entry->d_u.d_alias)); BUG_ON(!inode); lockdep_annotate_inode_mutex_key(inode);
/* WORKAROUND for calling security_d_instantiate() */entry->d_inode = inode; security_d_instantiate(entry, inode); spin_lock(&inode->i_lock); __d_instantiate(entry, inode);Ugh. That looks horrible even if it might avoid the oops.
I think a much better solution is to back-port commit b296821a7c42 ("xattr_handler: pass dentry and inode as separate arguments of ->get()") to older kernels. Then the inode is passed down all the way, and you don't have people try to get it from the (not yet initialized) dentry.
But there might be other parts missing too, and I didn't look at how easy/painful that backport would be.
Al - comments? This is all because of commit 1e2e547a93a0 ("do d_instantiate/unlock_new_inode combinations safely") being marked for stable, and various cases of security_d_instantiate() calling down to getxattr. Which used to not get the inode at all, so those older kernels use d_inode(dentry), which doesn't work in this path since dentry->d_inode hasn't been instantiated yet..
You also want b96809173e94 and ce23e6401334 there...