> I'd just jump here from the later test in mitigate sinkhole. […]
oops, you're right, I missed that.
Doing a quick test with gnu as, it would still almost work out by accident though: the first two moves would be conflated into one (into edx, the upper 16 bit containing the opcode for the second move), so instead we'd get ?rash, with the first value being whatever happened to be in %al at the time.
Anybody trying to debug that would probably go crazy though.
To view, visit change 37289. To unsubscribe, or for help writing mail filters, visit settings.