[Bug 294058] kernel panic when boot on powerpc64/powerpc64le qemu with more than 8 PCI devices

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D294058

--- Comment #2 from Luke Yasuda <[email protected]> ---
(In reply to Adrian Chadd from comment #1)

Hi, sorry that I forgot to paste my qemu config, then also forgot to reply =
;D

I'm editing the title because I found it inaccurate. After lots of trials, =
the
issue occurs when a PCI bridge is attached to PCI bus 0 with an address gre=
ater
than 0x8. Here is the simplified qemu cmdline I used to reproduce the issue:

/usr/bin/qemu-system-ppc64le \
-name 'random' \
-bios /usr/share/qemu/slof.bin \
-smp '4' \
-nographic -cpu host -m 4096 \
-device 'virtio-scsi-pci,id=3Dvirtioscsi0,bus=3Dpci.0,addr=3D0x3' \
-drive
'if=3Dnone,file=3DFreeBSD-16.0-CURRENT-powerpc-powerpc64le-20260330-e6eba50=
76929-284768-bootonly.iso,id=3Ddrive-scsi0,media=3Dcdrom,aio=3Dio_uring'
\
-device
'scsi-cd,bus=3Dvirtioscsi0.0,channel=3D0,scsi-id=3D0,lun=3D0,drive=3Ddrive-=
scsi0,id=3Dscsi0'
\
-machine 'type=3Dpseries' \
-accel kvm \
-device 'pci-bridge,id=3Dpci.1,chassis_nr=3D1,bus=3Dpci.0,addr=3D0x9' \
-vga none

Then the kernel crashes, both powerpc64 and powerpc64le.

The kernel does not crash and boots normally when I change the pci-bridge a=
ddr
to 0x8 or smaller.

The host is Debian unstable, but this issue can be reproduced on trixie or
bookworm.

--=20
You are receiving this mail because:
You are the assignee for the bug.=