)]}'
{
  "commit": "8c1b724ddb218f221612d4c649bc9c7819d8d7a6",
  "tree": "0e226f4156b554eec2690adb8f30ba54b15b68cc",
  "parents": [
    "f14a9532ee30c68a56ff502c382860f674cc180c",
    "514ccc194971d0649e4e7ec8a9b3a6e33561d7bf"
  ],
  "author": {
    "name": "Linus Torvalds",
    "email": "torvalds@linux-foundation.org",
    "time": "Thu Apr 02 15:13:15 2020 -0700"
  },
  "committer": {
    "name": "Linus Torvalds",
    "email": "torvalds@linux-foundation.org",
    "time": "Thu Apr 02 15:13:15 2020 -0700"
  },
  "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm\n\nPull kvm updates from Paolo Bonzini:\n \"ARM:\n   - GICv4.1 support\n\n   - 32bit host removal\n\n  PPC:\n   - secure (encrypted) using under the Protected Execution Framework\n     ultravisor\n\n  s390:\n   - allow disabling GISA (hardware interrupt injection) and protected\n     VMs/ultravisor support.\n\n  x86:\n   - New dirty bitmap flag that sets all bits in the bitmap when dirty\n     page logging is enabled; this is faster because it doesn\u0027t require\n     bulk modification of the page tables.\n\n   - Initial work on making nested SVM event injection more similar to\n     VMX, and less buggy.\n\n   - Various cleanups to MMU code (though the big ones and related\n     optimizations were delayed to 5.8). Instead of using cr3 in\n     function names which occasionally means eptp, KVM too has\n     standardized on \"pgd\".\n\n   - A large refactoring of CPUID features, which now use an array that\n     parallels the core x86_features.\n\n   - Some removal of pointer chasing from kvm_x86_ops, which will also\n     be switched to static calls as soon as they are available.\n\n   - New Tigerlake CPUID features.\n\n   - More bugfixes, optimizations and cleanups.\n\n  Generic:\n   - selftests: cleanups, new MMU notifier stress test, steal-time test\n\n   - CSV output for kvm_stat\"\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm: (277 commits)\n  x86/kvm: fix a missing-prototypes \"vmread_error\"\n  KVM: x86: Fix BUILD_BUG() in __cpuid_entry_get_reg() w/ CONFIG_UBSAN\u003dy\n  KVM: VMX: Add a trampoline to fix VMREAD error handling\n  KVM: SVM: Annotate svm_x86_ops as __initdata\n  KVM: VMX: Annotate vmx_x86_ops as __initdata\n  KVM: x86: Drop __exit from kvm_x86_ops\u0027 hardware_unsetup()\n  KVM: x86: Copy kvm_x86_ops by value to eliminate layer of indirection\n  KVM: x86: Set kvm_x86_ops only after -\u003ehardware_setup() completes\n  KVM: VMX: Configure runtime hooks using vmx_x86_ops\n  KVM: VMX: Move hardware_setup() definition below vmx_x86_ops\n  KVM: x86: Move init-only kvm_x86_ops to separate struct\n  KVM: Pass kvm_init()\u0027s opaque param to additional arch funcs\n  s390/gmap: return proper error code on ksm unsharing\n  KVM: selftests: Fix cosmetic copy-paste error in vm_mem_region_move()\n  KVM: Fix out of range accesses to memslots\n  KVM: X86: Micro-optimize IPI fastpath delay\n  KVM: X86: Delay read msr data iff writes ICR MSR\n  KVM: PPC: Book3S HV: Add a capability for enabling secure guests\n  KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs\n  KVM: arm64: GICv4.1: Allow non-trapping WFI when using HW SGIs\n  ...\n",
  "tree_diff": []
}