Hello, I see this list mostly for Linux kernel patches, but we are trying to get two-processor virtual mac99 machine going in qemu, and sadly it does not work without singlestepping on boot cpu in
smp_core99_kick_cpu() https://elixir.bootlin.com/linux/v6.12.17/source/arch/powerpc/platforms/powermac/smp.c#L796 just single stepping after switching to second thread/vCPU ends with my naive gdb script erroring out ;) Last email in chain on qemu-ppc list: https://lists.gnu.org/archive/html/qemu-ppc/2025-03/msg00136.html latest gdb log attached to https://lists.gnu.org/archive/html/qemu-ppc/2025-03/msg00132.html MacOS X 10.4.11 strangely works out of the box with two CPUs . Users reported even macos 9 (with selected applications like Quicktime) works! I can't test NetBSD/macppc due to some trouble in my crosscompiler? From netbsd/amd64 machine using standart NetBSD build.sh script. So I put NetBSD aside for now. I use 'remote' gdb running in Debian ppc sid/trixie VM on the same NetBSD/amd64 host. Linux kernel 6.12.17 was compiled under emulation with localmodconfig where I changed ext4 and SCSI disk.drivers to be builtin (for faster bootup without initrd). I disabled CPU_HOTPLUG after disabling sleep/hibernation support in hope it will help booting second cpu. But I was wrong about this. Recompiling might take 30 hours or so I prefer to reuse kernel I have ;) I see mac ppc SMP code was not changed much since at least 2.6.24 times of Gentoo 2008.0 so I hope it works on real machines! I might spotted typo in comment https://elixir.bootlin.com/linux/v6.12.17/source/arch/powerpc/kernel/head_book3s_32.S#L239 "/* core99 pmac starts the seconary here by changing the vector, and putting it back to what it was (unknown_async_exception) when done. */ " secondary not seconary ? but apart from this I am mostly clueless how PowerPC works, especially in those dual G4 machines .... If you need link to deb of this kernel, .config, qemu/openbios patches, gdb script - just ask! Not sure if I should crosspost this back to qemu-ppc list?
