KVM: x86: Fix more KVM_X86 Kconfig bugs
Sean Christopherson
maintainercover KVM: x86: Fix more KVM_X86 Kconfig bugs
Fix two more bugs in the KVM_X86 Kconfig. The fix from Arnd resolves a build failure, but practically speaking is a non-issue because the failure only affects 32-bit !SMP builds that effectively disable the local APIC.
The second fix is far more urgent. It resolves a bug where KVM's Kconfig doesn't correctly define KVM_X86, and will build kvm.ko as a module when KVM=y if neither KVM_INTEL nor KVM_AMD is 'y', e.g. if KVM=y, KVM_INTEL,=m, and KVM_AMD=m. I don't know if any distros use that combo, but it broke our production kernel, so it's certainly possible that it broke other setups too.
Arnd's fix was already posted, but I bundled it here to avoid a trivial conflict and to ensure it doesn't get left behind.
Arnd Bergmann (1):
KVM: x86: add back X86_LOCAL_APIC dependency
Sean Christopherson (1):
KVM: x86: Break CONFIG_KVM_X86's direct dependency on KVM_INTEL ||
KVM_AMD
arch/x86/kvm/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
base-commit: adc218676eef25575469234709c2d87185ca223a
Signature · 1 line
2.47.0.338.g60cca15819-goog
Sean Christopherson
maintainerPATCH 1/2 KVM: x86: add back X86_LOCAL_APIC dependency
From: Arnd Bergmann <[email protected]>
Enabling KVM now causes a build failure on x86-32 if X86_LOCAL_APIC is disabled:
arch/x86/kvm/svm/svm.c: In function 'svm_emergency_disable_virtualization_cpu':arch/x86/kvm/svm/svm.c:597:9: error: 'kvm_rebooting' undeclared (first use in this function); did you mean 'kvm_irq_routing'?
597 | kvm_rebooting = true;
| ^~~~~~~~~~~~~
| kvm_irq_routing
arch/x86/kvm/svm/svm.c:597:9: note: each undeclared identifier is reported only once for each function it appears in
make[6]: *** [scripts/Makefile.build:221: arch/x86/kvm/svm/svm.o] Error 1
In file included from include/linux/rculist.h:11,
from include/linux/hashtable.h:14,
from arch/x86/kvm/svm/avic.c:18:arch/x86/kvm/svm/avic.c: In function 'avic_pi_update_irte':arch/x86/kvm/svm/avic.c:909:38: error: 'struct kvm' has no member named 'irq_routing'
909 | irq_rt = srcu_dereference(kvm->irq_routing, &kvm->irq_srcu);
| ^~include/linux/rcupdate.h:538:17: note: in definition of macro '__rcu_dereference_check'
538 | typeof(*p) *local = (typeof(*p) *__force)READ_ONCE(p); \
Move the dependency to the same place as before.
--- arch/x86/kvm/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/Kconfig +1 -1
@@ -19,7 +19,6 @@ if VIRTUALIZATION config KVM_X86 def_tristate KVM if KVM_INTEL || KVM_AMD - depends on X86_LOCAL_APIC select KVM_COMMON select KVM_GENERIC_MMU_NOTIFIER select HAVE_KVM_IRQCHIP
@@ -49,6 +48,7 @@ config KVM_X86 config KVM tristate "Kernel-based Virtual Machine (KVM) support" + depends on X86_LOCAL_APIC help Support hosting fully virtualized guest machines using hardware virtualization extensions. You will need a fairly recent
Signature · 1 line
2.47.0.338.g60cca15819-goog
Sean Christopherson
maintainerPATCH 2/2 KVM: x86: Break CONFIG_KVM_X86's direct dependency on KVM_INTEL || KVM_AMD
In reply to Sean ChristophersonRework CONFIG_KVM_X86's dependency to only check if KVM_INTEL or KVM_AMD is selected, i.e. not 'n'. Having KVM_X86 depend directly on the vendor modules results in KVM_X86 being set to 'm' if at least one of KVM_INTEL or KVM_AMD is enabled, but neither is 'y', regardless of the value of KVM itself.
The documentation for def_tristate doesn't explicitly state that this is the intended behavior, but it does clearly state that the "if" section is parsed as a dependency, i.e. the behavior is consistent with how tristate dependencies are handled in general.
Optionally dependencies for this default value can be added with "if".
--- arch/x86/kvm/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/Kconfig +1 -1
@@ -18,7 +18,7 @@ menuconfig VIRTUALIZATION if VIRTUALIZATION config KVM_X86 - def_tristate KVM if KVM_INTEL || KVM_AMD + def_tristate KVM if (KVM_INTEL != n || KVM_AMD != n) select KVM_COMMON select KVM_GENERIC_MMU_NOTIFIER select HAVE_KVM_IRQCHIP
Signature · 1 line
2.47.0.338.g60cca15819-goog
Queued, thanks.
Paolo