> From: "Theo de Raadt" <dera...@openbsd.org> > Date: Thu, 10 Jan 2019 10:23:27 -0700 > > Ted Unangst <t...@tedunangst.com> wrote: > > > Ted Unangst wrote: > > > > > > Does 0xffff come from ACPI? Can we give that a name? > > > > > > I thought sleeping for one tick is kinda weird, but I see what > > > it's doing with the acpi_dotask loop. This feels precarious, but > > > whatever. > > > > So upon further thought, this is pretty bad. If the new task also > > calls sem wait, we'll end up back here, recursing on the event > > loop. And of course, the first sem can't wakeup until after the > > second one does. If you want to say it was like this when you got > > here, ok, but pretending to pause like this is troublesome. There > > was a bug in libc rpc code where it tried to do something similar > > and blew up the stack. > > Not sure I understand. There is only one acpi task.
No there can be many acpi tasks, so there is indeed a risk. But we have to allow other acpi tasks to run while we're blocked on the semaphore otherwise we'll never wake up!
