Re: [PATCH] Linux Kernel Markers 0.5 for Linux 2.6.17 (with probe management)

[Karim Yaghmour]

Mathieu Desnoyers wrote:
> Here is the implementation :-)

Trigger happy :)

> To change it, we can dynamically overwrite the __mark_near_jump_select_##name
> value (a byte) to (__mark_jump_call_##name - __mark_near_jump_##name).

Hmm... I don't know if you won't still need to resort to int3 and
then overwrite the byte. From my understanding it sounds like you
wouldn't but that's where Richard's insight on the errata stuff
might come in handy.

Have you actually tried to run this code by any chance? *If* it did
work, I think this might just mean that you don't need either
kprobes or djprobes.

> So we have one architecture specific optimisation within the architecture
> agnostic marking mechanism.

That would seem reasonable, I think. You might want to test out
your mechanism, get confirmation from Richard and then post an
update to your patches.

Karim

-
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/