On Feb 9, 2021, at 10:09 AM, Linus Torvalds [off-list ref] wrote:
On Tue, Feb 9, 2021 at 8:55 AM Andy Lutomirski [off-list ref] wrote:
quoted
Or we hack up #CP to handle this case. I don’t quite know how I feel about this.
I think that's the sane model - if we've replaced the instruction with
'int3', and we end up getting #CP due to that, just do the #BP
handling.
Anything else would just be insanely complicated, I feel.
The other model is “don’t do that then.”
I suppose a nice property of patching ENDBR to INT3 is that, not only is it atomic, but ENDBR is sort of a NOP, so we don’t need to replace the ENDBR with anything.
Linus