On Tue, 19 Feb 2019 10:47:38 -0800 Guenter Roeck <linux@xxxxxxxxxxxx> wrote: > Hi Cornelia, > > On 2/18/19 2:30 PM, Cornelia Huck wrote: > > >> What is your qemu command line ? > > > > Ignoring any additional devices: > > > > s390x-softmmu/qemu-system-s390x -M s390-ccw-virtio,accel=tcg -cpu max -m 1024 -nographic -device virtio-scsi-ccw,id=scsi0,devno=fe.0.0001 -drive file=/home/cohuck/vm-images/vm1.qcow2,format=qcow2,if=none,id=drive-scsi0-0-0-0 -device scsi-hd,bus=scsi0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0-0-0-0,id=scsi0-0-0-0,bootindex=1 -sandbox on -kernel ~/git/linux/arch/s390/boot/bzImage -append "root=/dev/sda3" > > > > Is "accel=tcg" and "-cpu max" expected to make any difference ? CPU consumption > on my system seems to be identical to the default (no -M or -cpu on command line). > This is with qemu 3.1. If you're not running on a s390x, QEMU will pick tcg even if not specified (as it needs to emulate anyway), so that should not make any difference. The 'max' cpu model was is not yet available with 3.1 (merged in the current development cycle); it will not make much difference in practice (a few more cpu features are currently provided with it over the default 'qemu' cpu model). As I'm testing various combinations (tcg vs. kvm on an s390x host, different cpu models) when I'm queuing patches, my command line is not necessarily exactly the same every time (this was just the one I happened to be testing the kernel patch with). If you simply want to be able to boot a test kernel on a non-s390x host, relying on the default accelerator (tcg) and the default cpu model (qemu) being picked should be completely fine.
