Re: aarch64 rust problems

[Tobias Nygren <tnn%NetBSD.org@localhost>]

On Thu, 06 Jun 2024 11:53:36 +0200 (CEST)
Havard Eidnes <he%NetBSD.org@localhost> wrote:

> This is reported upstream as
> https://github.com/rust-lang/rust/issues/123551
> but so far it appears nobody else is seeing a similar failure.

Is reproducable on physical hardware as I wrote on the GH issue
last month.

> This makes me wonder if the rust compiler detection of stack
> overflow is faulty, and that this isn't in fact the problem.
> I find it incredible that it actually overflows a 458MB stack.
> Maybe as a last-ditch effort I'll re-try the build from scratch
> instead of trying to just re-start it.

All the code does it hook up a SEGV handler and then checks if the
faulting address is inside the guard page. This could theoretically
fire on things other than a stack overflow but it seems unlikely.
Even if the guard page calculation is wrong there was still an 
underlying SEGV so we are still crashing regardless.

What we need to do is, in the SEGV handler, somehow find the last good
stack frame and hook up the unwinder to that to find out what
the process is doing.

-Tobias