Not really related to this oroject, but we at qemu trying to get dual-G4 mac99 machine emulation working, and so far only MacOS 9.2/X works, Linux fails to init second cpu UNLESS I single step function smp_core99_kick_cpu on first (boot) processor remotely via gdb. FreeBSD 14.2 just panic/reboots with second CPU on, boots fine on one.
https://lists.gnu.org/archive/html/qemu-ppc/2025-03/msg00257.html
At this point we a bit out of ideas about how exactly very first steps of second cpu init must be performed? Does second cpu need all interrupts disabled initially, not just halted by openbios?
Do we miss something about timebase sync between CPUs in very early stages of second cpu coming out of halted state?
Not really related to this oroject, but we at qemu trying to get dual-G4 mac99 machine emulation working, and so far only MacOS 9.2/X works, Linux fails to init second cpu UNLESS I single step function smp_core99_kick_cpu on first (boot) processor remotely via gdb. FreeBSD 14.2 just panic/reboots with second CPU on, boots fine on one.
https://lists.gnu.org/archive/html/qemu-ppc/2025-03/msg00257.html
At this point we a bit out of ideas about how exactly very first steps of second cpu init must be performed? Does second cpu need all interrupts disabled initially, not just halted by openbios?
Do we miss something about timebase sync between CPUs in very early stages of second cpu coming out of halted state?