On Sun, Jul 18, 2010 at 09:53:57AM +0800, chen huacai wrote: > Hmm, I know. But, why qemu only crashes on user-space code, not on kernel > code?
Probably because a lot more code is executed in user-space than in kernel mode. I have actually tried the current git tree, and it is indeed not working as I was expecting. Last time I tried, it was working correctly for me, it seems something has broken MIPS host support in the meanwhile. It was commit 60bf84cf4ce277e1212929f35674796888a681ce in case you want to try it. > On Fri, Jul 16, 2010 at 8:27 PM, Aurelien Jarno <aurel...@aurel32.net> wrote: > > chen huacai a écrit : > >> Thank you for your information. > >> I'am a little confusing, cacheflush syscall you mentioned is in guest > >> or in host? > >> If you means syscall in guest, why x86 host can boot mips guest? > >> If you means syscall in host, why qemu-0.9.x doesn't crash? > >> > > > > On the host. I guess qemu-0.9.x doesn't crash, because it's not using > > TCG, so it doesn't write the code to memory the same way. > > > > -- > > Aurelien Jarno GPG: 1024D/F1BCDB73 > > aurel...@aurel32.net http://www.aurel32.net > > > > > > -- > Huacai Chen > -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net
