Re: [PATCH 3/5] SSB extra 1

[Thomas Gleixner <tglx@linutronix.de>]

On Sun, 6 May 2018, speck for Dave Hansen wrote:
> On 05/06/2018 02:48 PM, speck for Thomas Gleixner wrote:
> >   1) I read about trusted and untrusted BPF programs. How is that
> >      determined/defined?
> 
> There's a rather steep learning curve, I fear.
> 
> There's bpf_map->unpriv_array which tells whether or not you have to
> deploy Spectre V1 mitigations.  That's just derived from CAP_SYS_ADMIN
> when the map is allocated.
> 
> Separately, I've seen some hard-coded checks against BPF_PROG_TYPE_*,
> like may_access_direct_pkt_data().
> 
> So I think the trust ends up being anchored in the interfaces that
> allocated the program (or map?) to begin with, but it's somewhat hard to
> follow.
> 
> Is that what you meant?

Kinda. So the question is what is the definition of trusted vs. untrusted,
if there is any.

I digged deeper into the code and the user supplied socket filters run in
SOFTIRQ_RX context, so the question whether this can be restricted to the
context of the program which installed the filter is answered. We
can't.....

Thanks,

	tglx