Brad Smith

> Thanks. Will look into it. If you could have a look at the compiler-rt patch for asan, specifically what could be upstreamed: > > left issues: needed macro _GNU_SOURCE/_DEFAULT_SOURCE...

@korli I pushed the Driver bits.. https://github.com/llvm/llvm-project/commit/910f7f45f27d1f3cfad779669d0e0f15ff5b9686

Trying to extract what you have in your recipe patch I am at.. https://github.com/brad0/llvm-project/commit/fdbb9476c7af38d09dd5fa3ddcee2c683a410312 Some cleaning, formatting and moving the bits that were patching NetBSD code to their own Haiku...

> left issues: needed macro _GNU_SOURCE/_DEFAULT_SOURCE for some pthread functions Checking this again.. _GNU_SOURCE is not necessary, but _DEFAULT_SOURCE is. The build was complaining about the lack of ```struct dl_phdr_info```...

> I pushed the openmp patch with some changes: [8e9336c](https://github.com/haikuports/haikuports/commit/8e9336c3196f2f445a79f1c32743f71f1683391b) Looking that over I see you added the thread id retrieval and removed the signal disabling. Is there anything else...

https://github.com/llvm/llvm-project/commit/9b7a7e4b9e44a73978c374f95f90332620f87ce8 I pushed the OpenMP patch upstream.

If you have time try to look into implementing the ```__kmp_get_load_balance``` function for Haiku.

> > If you have time try to look into implementing the `__kmp_get_load_balance` function for Haiku. > > Haiku is simply not providing this kind of information ATM (getloadavg). I'll...

> I looked into UNIX "load averages" recently for other reasons, and I don't know if they really make sense on Haiku due to how our scheduler works. We could...

I have rebased the patch for asan. https://github.com/brad0/llvm-project/commit/1229f0929b691e052360246c686a62a25391ffc4 ``` /boot/home/llvm-brad/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_haiku.cpp:46:38: error: arithmetic on a pointer to void 46 | segment->end = (uptr)(info.address + info.size); | ~~~~~~~~~~~~ ^ 1 error generated....