David CARLIER

Your proposal makes sense, would be nice to come with tests before and after the change, what do you think ?

Anyhow, snmalloc already uses MADV_FREE for FreeBSD.

Out of curiosity, would the `dev` branch still fails ?

weird because the assembly is supposed to be different.

Right ... what happens if you use this change : ` uintptr_t tcb; UNUSED(ofs);UNUSED(slot); __asm__ volatile ("mrs %0, tpidr_el0" : "=r" (tcb)); res = tcb; ` Just theorical, I wish...

Yes it would be similar-ish solution if it works like : ` uintptr_t tcb; UNUSED(ofs);UNUSED(slot) __asm__ volatile ("mrs %0, tpidr_el0" : "=r" (tcb)); tcb = (uintptr_t)value; ` Thanks for your...

Too bad the changes work on Linux arm64 maybe the M1 gear needs a specific code path in this case. Just an idea, what happens if you comment line 298...

Right so maybe M1 needs a specific code path (specific register like x86 case ?). I see on other projects issues popping up about M1 here and there seems to...

I just did a run with an iPhone 6 :   with the actual dev but with these little changes already mentioned 

Was wondering if, only in the case of darwin ARM devices, we would use, tpidrr0_el0 TID register instead ? tpidr_el0 always return 0 otherwise.