Closed Liblor closed 4 years ago
Just saw that 4832e7d is removed in #169 as well.
Doesn't really fix the issue. Seems to be a problem with dynamic memory. We probably try to map memory during free
or something... I tried writing to some parts of the to be freed chunk outside of the lock to have it mapped... didn't work
Fix the mess we have with the
malloc
dynamic split. @leopoldsedev threading tests seem to run great :) The hashtable and list use a backed malloc calledstatic_malloc
same goes for paging code. Gone be the days ofenable_static
and deadlocks.~Current issue: Some stuff now use the
static_malloc
is freed usingfree
instead ofstatic_free
.~ I hope this allows us to go back to the threading version of the nameserver. That way spawning can be implemented nicer.