On Tue, Apr 16, 2019 at 05:10:42PM -0700, Eric Biggers wrote: > > Hi Ted, I assumed you resolved the conflict with "fscrypt: use READ_ONCE() to > access ->i_crypt_info"? The code in fscrypt_d_revalidate() should be: > > dir = dget_parent(dentry); > err = fscrypt_get_encryption_info(d_inode(dir)); > valid = !fscrypt_has_encryption_key(d_inode(dir)); > dput(dir); Yes, I did notice the patch conflict. Thanks for confirming the correct resolution! - Ted
