Bart Oldeman

Rebased and cleaned up history now, so the step are a bit smaller, and register_hardware_ram is never called anymore without _virtual.

oops! I realized I pushed the wrong branch yesterday. So some of your comments are out of date, sorry for that. The clunky interface (anti-pattern) in mapping.c with do_get_hardware_ram I'll...

Not sure if you are using your sourceforge forwarding address to get notifications, but SF has been plagued by DDOS attacks. Github too, but github survived. I'll send patches addressing...

After looking at the DPMI code a bit further I think a choice needs to be made. Either: 1. FPU state is never copied, and left on the signal stack....

> My assumption was that if you pass NULL, then it will just not change the FPU state. So if it does otherwise, I believe this behaviour is inherently unportable....

> Does the kernel do fpu_clear also when entering the signal handler, or should we do that by hands? I am not sure. For very old hardware it's using FNSAVE...

yes. haven't had much time to work on it, but the i386 sse state fix is still pending.

Putting to draft because this needs a proper test combining `--job --robot --try-*`.

Justed wanted to test locally using `eb HPL-2.3-foss-2023a.eb --try-toolchain=foss,2024a --job`. This works fine. Note that a shared temporary directory is still needed with `--job --from-pr`, but you can use a...

> @bartoldeman Seems like test was implemented, so this shouldn't be a draft PR anymore? @boegel undrafted now