On 06/19/2017 12:15 PM, Florian Weimer wrote: > On 06/19/2017 08:02 PM, Richard Biener wrote: >> Oh, and using push intelligently with first bumping to SP & 4096-1 + 4095 >> would solve the signal atomicity as well. Might be larger and somewhat >> interfere with CPUs stack engine. Who knows... > > On x86-64, PUSH REG is just a single byte, so for sequences that have to > move SP and probe, it's the shortest possible sequence AFAIK. NEG/NOT > can take an offsettable memory operand, but it's three bytes. Right. I think we want guidance from Honza & Uros on what the most runtime efficient mechanisms are (or are likely to be, there's a certain amount of guesswork that has to happen here), then we look at which are the most code space efficient. I'm personally willing to trade off some unwinder table space if it gives us more compact code.
Jeff
