Re: [PATCH 00/34] fs: idmapped mounts
On Thu, Oct 29, 2020 at 05:19:20PM +0100, Christian Brauner wrote:
This approach (always) seemed conceptually wrong to me. Like Tycho
said
somewhere else this basically would act like a global variable which
isn't great.
Elaborating on this a bit: the way Christian has organized this series
(by adding a parameter to the various inode operations) forces people
to make a choice about what user namespace to pass. They can of course
make the wrong choice, but that will show up in a diff somewhere, and
hopefully someone will notice.
If we pass things via current-> somewhere, people could potentially
avoid making this choice, and then we wind more of the kinds of bugs
Eric is worried about elsewhere in the thread.
While it does result in larger patches initally, requiring people to
reason about it explicitly is the right move, IMO.
Tycho