Re: [RESEND][PATCH] ima: Set and clear FMODE_CAN_READ in ima_calc_file_hash()

[Linus Torvalds]

On Tue, Nov 17, 2020 at 3:29 PM Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> On Tue, Nov 17, 2020 at 3:24 PM Mimi Zohar <zohar@xxxxxxxxxxxxx> wrote:
> >
> > I really wish it wasn't needed.
>
> Seriously, I get the feeling that IMA is completely mis-designed, and
> is doing actively bad things.
>
> Who uses this "feature", and who cares? Because I would suggest you
> just change the policy and be done with it.

Another alternative is to change the policy and say "any write-only
open gets turned into a read-write open".

But it needs to be done at *OPEN* time, not randomly afterwards by
just lying to the 'struct file'.

Why? Because the open has told the filesystem that it's only for
writing, and a filesystem could validly do things that make reading
invalid. The simplest example of this is a network filesystem, where
the server might simply not *allow* reads, because the open was for
writing only.

See? IMA really does something fundamentally quite wrong when it tries
to read from a non-readable file. It might "work" by accident, but I
really do think that commit a1f9b1c0439db didn't "break" IMA - it
showed that IMA was doing something fundamentally wrong.

           Linus