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

[Jason Baron <jbaron@akamai.com>]

On 07/11/2013 10:35 AM, Steven Rostedt wrote:
> On Wed, 2013-07-10 at 14:36 -0700, H. Peter Anvin wrote:
>> On 07/10/2013 02:31 PM, Jiri Kosina wrote:
>>> If any CPU instruction execution would collide with the patching,
>>> it'd be trapped by the int3 breakpoint and redirected to the provided
>>> "handler" (which would typically mean just skipping over the patched
>>> region, acting as "nop" has been there, in case we are doing nop -> jump
>>> and jump -> nop transitions).
>>>
>> 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.
>>
>> 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.
>>
> Wont work for ftrace, as it patches all functions, it even patches
> functions used to do the changes. Thus, it would cause a deadlock if a
> breakpoint were to spin till the changes were finished.
>
> -- Steve
>
>

I'm not sure this works for jump labels either. Some tracepoints (which
use jump_labels) have interrupts disabled across them. Thus, they will
spin with interrupts disabled, while we are trying to issue an IPI.

Thanks,

-Jason