sanitize_next_sp might do better to allocate a new rust stack when returning to rust?

[graydon]

Unclear fixme in rust_task.h. Might represent stack-switching bug or perf problem.

[bblum]

It'd be a perf problem the other way around - having crust calls always allocate new stack segments is worse than using existing ones. I tried that and it made crust-call-deep.rs regress.

As it is, it might put you into the redzone, but as long as the redzone is as big as it is, there's no problem. Leaving this open 'cause having a big redzone itself seems unpopular.

[nikomatsakis]

Seems related to #2820

[bblum]

Renominating for production-ready; this is a runtime optimization, not a backward-compatibility issue.

This code will hopefully go away soon, but it will have to be reimplemented in the new runtime, so represents a decision we'll still have to make.

[graydon]

just a bug, removing milestone/nomination.

[alexcrichton]

In today's meeting we have decided to jettison segmented stacks. I also don't think that this code exist any more, so closing.