lkml.org 
[lkml]   [2020]   [Jul]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH RFC V2 17/17] x86/entry: Preserve PKRS MSR across exceptions
Date
Andy Lutomirski <luto@amacapital.net> writes:

> Suppose some kernel code (a syscall or kernel thread) changes PKRS
> then takes a page fault. The page fault handler needs a fresh
> PKRS. Then the page fault handler (say a VMA’s .fault handler) changes
> PKRS. The we get an interrupt. The interrupt *also* needs a fresh
> PKRS and the page fault value needs to be saved somewhere.
>
> So we have more than one saved value per thread, and thread_struct
> isn’t going to solve this problem.

A stack of 7 entries and an index needs 32bytes total which is a
reasonable amount and solves the problem including scheduling from #PF
nicely. Make it 15 and it's still only 64 bytes.

> But idtentry_state is also not great for a couple reasons. Not all
> entries have idtentry_state, and the unwinder can’t find it for
> debugging. For that matter, the page fault logic probably wants to
> know the previous PKRS, so it should either be stashed somewhere
> findable or it should be explicitly passed around.
>
> My suggestion is to enlarge pt_regs. The save and restore logic can
> probably be in C, but pt_regs is the logical place to put a register
> that is saved and restored across all entries.

Kinda, but that still sucks because schedule from #PF will get it wrong
unless you do extra nasties.

> Whoever does this work will have the delightful job of figuring out
> whether BPF thinks that the layout of pt_regs is ABI and, if so,
> fixing the resulting mess.
>
> The fact the new fields will go at the beginning of pt_regs will make
> this an entertaining prospect.

Good luck with all of that.

Thanks,

tglx

\
 
 \ /
  Last update: 2020-07-23 22:22    [W:0.097 / U:0.268 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site