Hi,
Here is my attempt at rolling a fix for the deadlock, based on
your suggestions. Does this resemble anything you had in mind ?
And to be honest, while the original deadlock seems to be gone,
this feels disturbingly 'opportunistic programming' for me. I
haven't touched any filesystem code in the kernel before so I
really have no idea if this is anywhere near as it should be.
Someone who understands something about the VFS and LSM hooks
should check this. And also should check if the resulting side-
effects are okay from a security perspective: the original code
path calls security_inode_setxattr while the new code path for
IMA attributes does not call it.
Cheers,
Krisztian
--
To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
