Re: [RFC] [PATCH 1/2 v2] x86: introduce int3-based instruction patching

[H. Peter Anvin]

On 07/10/2013 02:48 PM, Borislav Petkov wrote:
> On Wed, Jul 10, 2013 at 02:36:41PM -0700, H. Peter Anvin wrote:
>> I'm wondering if it would be easier/more general to just return to the
>> instruction. The "more general" bit would allow this to be used for
>> other things, like alternatives, and perhaps eventually dynamic call
>> patching.
> 
> Well, the aspect of not using stop_machine in alternatives is a don't
> care because there we do text_poke_early on the BSP anyway. However,
> there we toggle interrupts so it would probably be interesting to see
> whether a non-interrupt-disabling variant would be faster.
> 
>> Returning to the instruction will, in effect, be a busy-wait for
>> the faulted CPU until the patch is complete; more or less what
>> stop_machine would do, but only for a CPU which actually strays into
>> the affected region.
> 
> Oh, something like we patch in a two-byte jump first:
> 
> 1:
> 	jmp 1b
> 
> until we finish patching the rest? Ha, interesting :).
> 

No, the idea is that the affected CPU will simply execute int3 -> iret
ad nauseam until the first byte is repatched, at that point execution
will resume normally.

	-hpa

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/