On Tue, May 25, 2021 at 1:15 PM Christian Brauner <christian.brauner@xxxxxxxxxx> wrote: > > On Mon, May 24, 2021 at 04:53:21PM +0300, Amir Goldstein wrote: > > Reporting event->pid should depend on the privileges of the user that > > initialized the group, not the privileges of the user reading the > > events. > > I think it's in general a good permission model to not have the result > depend as little on the reader as possible post-open/init. So this makes > a lot of sense to me and I'm a bit surprised it wasn't like that right > away. :) > You are right. We got lot there. The fact that pid_vnr() may not report pid depending on the caller's ns contributed to the confusion for making it depend in the caller's creds, but namespace and credentials behave differently. Thanks for the review, Amir.
