On Nov 20, 2018, at 4:42 AM, Peter Zijlstra [off-list ref] wrote:

On Tue, Nov 13, 2018 at 05:07:20AM -0800, Nadav Amit wrote:

quoted

v4->v5:
- Fix Xen breakage [Damian Tometzki]
- BUG_ON() when poking_mm initialization fails [PeterZ]
- Better comments on "x86/mm: temporary mm struct"
- Cleaner removal of the custom poker

I'll re-iterate my position: it is impossible for the text not to match,
and if it somehow does not match, something went sideways in an
unrecoverably fashion.

text_poke() must not fail, ever. If it does, our text is inconsistent
and we must abort/panic/bug.

The only way I will accept anything else is if someone can come up with
a sensible scenario of text_poke() failing and recovering from it.
AFAICT there is no possible way to gracefully recover.

Consider a jump label with multiple patch sites; we patch the first,
then fail. In order to restore to a sane state, we must undo the
patching of the first, but undoing text_poke() fails again. Then
what?

Allowing text_poke() to fail only creates an unfixable mess. Esp. since
there is no sane scenario under which is can fail.

Ok, ok... I tried to stand my ground, but I guess I failed. I don’t feel
that strongly about this assertion to argue with you. I’m just the “chicken”
kind of guy.

Yet, take into consideration that I will need to use you as my “vest” once I
get being “shot” for adding BUG_ON(). ;-)

I will send another version tonight, assuming no new issues are raised.

Regards,
NAdav