nix is working pretty reliably for me now. We're past the will it run stage, now for the clean it up stage. As always, contributors welcome. I note that Paul dialed down all the nasty debug prints, so if you use it, you'll find it nicer.
reminder: TODOs are here: https://github.com/rminnich/9front/blob/nix/TODO.md On Mon, Mar 24, 2025 at 2:35 PM ron minnich <rminn...@gmail.com> wrote: > I've taken it a bit further, and it's odd. > > rfork(RFCORE) works fine. up->errlab is 0 when the rfork(RFCORE) exits. > But here's the interesting part. > up->nerrlab is still 1 when we flip the process over to the ac. This is > because of the call to _procctl in syscall() > > Then the process calls pwrite, and, up->nerrlab is 1 BEFORE the first > waserror() in syscall. > > I think before the _procctl() in syscall, there needs to be a poperror(). > this fixes it for me. > if (up->procctl == Procto ...) {poperror(); _procctl(up); } > The reason is, that when procctl is set to Proc_toac, that call to procctl > never returns, and the poperror never happens. > > Why did this ever work, you say? I don't know. > > Paul, see what you think. > > On Mon, Mar 24, 2025 at 1:56 PM Paul Lalonde <paul.a.lalo...@gmail.com> > wrote: > >> No, it does not block, failing if it can't. >> >> This piece is still buggy - the error labeling (waserror(), poperror(), >> error()) is out of balance, even in the path that doesn't fork. >> So I'm tracing carefully to understand where the missing poperror() or >> extra waserror() is. >> >> Paul >> >> On Mon, Mar 24, 2025 at 1:41 PM Stuart Morrow <morrow.stu...@gmail.com> >> wrote: >> >>> On Mon, 24 Mar 2025 at 15:12, ron minnich <rminn...@gmail.com> wrote: >>> > Further, he fixed the rfork support, such that processes can set up >>> and move from a TC to an AC. This restores a long-lost capability. >>> > >>> > This is a pretty major step forward. It is now even possible to do >>> things like this: >>> > main(){ >>> > /* set it all up */ >>> > rfork(RFACORE); /* something like that */ >>> > >>> > and your process is now running on a dedicated, non-preempted, >>> interrupt-free, no kernel code, core all to its own. >>> >>> Does that block until the resource is available? 9front added the >>> ability for rfork to fail. >> *9fans <https://9fans.topicbox.com/latest>* / 9fans / see discussions >> <https://9fans.topicbox.com/groups/9fans> + participants >> <https://9fans.topicbox.com/groups/9fans/members> + delivery options >> <https://9fans.topicbox.com/groups/9fans/subscription> Permalink >> <https://9fans.topicbox.com/groups/9fans/T25c85a39f00802e6-M578da3e508ed84bd7aa48d11> >> ------------------------------------------ 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/T25c85a39f00802e6-Mf774e3ff70947085ef227e27 Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
