01.01.2021 18:43, Sergey Aleynikov wrote:
I've re-checked that with "-vga none" seabios@rel-1.14.0 the specific VM doesn't boot and still doesn't on master. And it boots with an old seabios@e38be2d (though compiling it now requires some effort).
Thank you for confirming this. I'm still unable to reproduce the problem. How do you know qemu is stuck if there's no output? In your case there's no output defined for the guest. If create a serial port (output to a file) and use kernel/initrd from my host, with -append="console=ttyS0", I can see my kernel is panicked being unable to mount root filesystem (which is rather expected), and at this stage it eats about 100% CPU sitting in a tight loop. I tried a buster image, - it works just fine with the parameters you gave in message from 3 Feb 2018, - it boots and I can log in to it over network. So it looks like I'm doing something wrong here.
PS: both host & guest are now debian 10, qemu version is 4.2.
Um.. I'm using qemu 5.2. Lemme retry with qemu from buster... nope, it is the same with qemu-system-x86_3.1+dfsg-8+deb10u8_amd64.deb. What I'm doing wrong? :) Thanks, /mjt
