On 7/17/2026 9:14 PM, Michael Tokarev wrote: > On 7/17/26 22:40, Pierrick Bouvier wrote: > >> https://gitlab.com/qemu-project/qemu/-/blob/master/hw/vfio/Kconfig? >> ref_type=heads#L5 >> >> ``` >> config VFIO >> bool >> depends on LINUX >> ``` >> >> Incorrect, and it should exclude Linux hosts not supporting VFIO. >> At least sparc64 to fix the current issue. > > Shouldn't there be "depends on LINUX & KVM" ? At least, this > way we can do --disable-kvm and it will work. >
I think it's a safe choice, but not sure if we are supposed to be able to use the exotic combination of vfio with tcg also. Maybe it doesn't make any sense technically though, I don't know the arcane of vfio. @Cédric, Any idea of which matrix of host/target is supposed to be supported? > Is sparc64 the only architecture where kvm is not available? > Somehow I doubt it's the case. Yet, on other architectures > without kvm, the thing builds. > > To me, this smells like a bug in debian, - missing asm/kvm.h > include in linux-headers - so that some headers are missing > while some others are present. Or a bug in kernel itself, with > sparc port lagging in this area so that even dummy asm/kvm.h is > not provided. > In upstream kernel, that's the list of architectures providing this: ./arch/arm64/include/uapi/asm/kvm.h ./arch/loongarch/include/uapi/asm/kvm.h ./arch/mips/include/uapi/asm/kvm.h ./arch/powerpc/include/uapi/asm/kvm.h ./arch/riscv/include/uapi/asm/kvm.h ./arch/s390/include/uapi/asm/kvm.h ./arch/x86/include/uapi/asm/kvm.h It seems like there is no concept of "asm/kvm.h" stub in the kernel. Which seems reasonable considering they have Kconfig for this. > Thanks, > > /mjt Regards, Pierrick
