Minor cleanup to fix the underlying crustiness that led to an uninitialized variable warning reported by Barret. The first two patches are tagged with Fixes:, but I don't know that they're actually worth backporting to stable. Functionally, everthing works, it's just a bit weird and AFAICT not what is intended. It might be preferable to take Barret's patch[*] first and only mark that for stable, as it fixes the immediate issue without revamping __kvm_gfn_to_hva_cache_init(). [*] https://lkml.kernel.org/r/20200109195855.17353-1-brho@xxxxxxxxxx Sean Christopherson (3): KVM: Check for a bad hva before dropping into the ghc slow path KVM: Clean up __kvm_gfn_to_hva_cache_init() and its callers KVM: Return immediately if __kvm_gfn_to_hva_cache_init() fails virt/kvm/kvm_main.c | 45 ++++++++++++++++++++++++++------------------- 1 file changed, 26 insertions(+), 19 deletions(-) -- 2.24.1
