https://bugs.exim.org/show_bug.cgi?id=2618
--- Comment #6 from Carlo Marcelo Arenas Belón <carenas@???> ---
(In reply to Stuart from comment #5)
> Apple has provided explicit documentation for Porting JITs to arm64
> hardware, including reference to the MAP_JIT flag for the hardened runtime.
>
> -
> https://developer.apple.com/documentation/apple_silicon/porting_just-in-
> time_compilers_to_apple_silicon
and the documentation points in a different direction than the provided patch,
with the only possible missing part being that dance with
pthread_jit_write_protect_np to make the page temporarily writeable before it
can be turned back into RX and jumped into and that will be likely something
new with macOS 11 (or 10.16 if using the official beta) and regardless of CPU
> We are not using the Hardened Runtime
but sljit (embedded within) and PCRE are libraries, and as such have no control
on what runtime the application that linked to them might be operating.
presume though that being able to test in apple silicon as well as intel (with
both hardened and not hardened) runtime will be likely needed to make sure this
is supported appropriately, but the direction shown in the proposed patch
doesn't seem to be the right one.
would "we" be someone with that access and better if able to provide access
remotely?, it might be better also to develop/validate this in sljit first and
look at integrating it into pcre after IMHO
--
You are receiving this mail because:
You are on the CC list for the bug.
