On Wed, Jul 10, 2024 at 5:01 PM Roberto Sassu <roberto.sassu@xxxxxxxxxxxxxxx> wrote: > > It looks ok for me. The only thing we are not doing is to set the > pointer to NULL, but I guess does not matter, since the inode security > blob is being freed. Yes, that shouldn't be an issue, and depending on how things get scheduled, it's entirely possible that the associated inode is completely gone by the time ima_inode_free_rcu() is called. > I also ran some UML kernel tests in the CI, and everything looks good: > > https://github.com/robertosassu/ima-evm-utils/actions/runs/9880817007/job/27291259487 > > Will think a bit, if I'm missing something. Great, thank you. -- paul-moore.com
