Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1532749

[PATCH v2 0/5] KVM: VMX: Fix single stepping with emulated instructions

Show all headers | View raw

KVM does not currently honor the trap flag when emulating instructions that
cause VM exits. This is observable from guest userspace, try stepping on a
CPUID instruction in gdb in a KVM guest. The program will stop two
instructions after CPUID.

To fix this, add kvm_skip_emulated_instruction that calls
kvm_x86_ops->skip_emulated_instruction and then kvm_vcpu_check_singlestep to
check for RFLAGS.TF. Patch 4 does this. To handle both the guest setting TF
and the KVM_GUESTDBG_SINGLESTEP cases we need to be able to indicate to
callees that an exit to userspace is required. Patches 1-3 are largely
plumbing to make this possible.

Traps triggered by task switch instructions require some additional handling
and are not implemented. KVM_GUESTDBG_SINGLESTEP traps can be squashed by
certain instructions which also trigger userspace exits, such as HALT,
MOV CR8, and IO instructions. I believe (although I have not tested) that
KVM will simply generate another trap on the next instruction, so this is
no worse than the current behavior.

Patch 5 fixes an existing bug where kvm_vcpu_check_singlestep clears
RFLAGS.TF when generating a singlestep trap. Bare silicon does not do this.

These patches only fix this issue for VMX, and the handful of SVM cases that
call the shared kvm_complete_insn_gp. I don't have AMD silicon to test on.

Since v1:
Patch 3: KVM: VMX: Move skip_emulated_instruction out of nested_vmx_check_vmcs12
- No longer creates a separate skip_emulated_instruction_no_trap variant.

Patch 4: KVM: x86: Add kvm_skip_emulated_instruction and use it.
- Instead the existing skip_emulated_instruction continues to not check the TF
  flag.
- kvm_skip_emulated_instruction is added to skip_emulated_instruction and then
  deal with any singlestep traps that should be generated.
- The patch 5 from v1 is effectively subsumed by this patch.

Patch 5: KVM: x86: Do not clear RFLAGS.TF when a singlestep trap occurs.
- New in this version, fixes an existing bug visible to the kvm-unit-tests
  test I wrote for this patch series.

Back to linux.kernel | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

[PATCH v2 0/5] KVM: VMX: Fix single stepping with emulated instructions Kyle Huey <me@kylehuey.com> - 2016-11-29 21:50 +0100
  [PATCH v2 1/5] KVM: x86: Add a return value to kvm_emulate_cpuid Kyle Huey <me@kylehuey.com> - 2016-11-29 21:50 +0100
  [PATCH v2 3/5] KVM: VMX: Move skip_emulated_instruction out of nested_vmx_check_vmcs12 Kyle Huey <me@kylehuey.com> - 2016-11-29 21:50 +0100
  [PATCH v2 2/5] KVM: VMX: Reorder some skip_emulated_instruction calls Kyle Huey <me@kylehuey.com> - 2016-11-29 21:50 +0100
  [PATCH v2 4/5] KVM: x86: Add kvm_skip_emulated_instruction and use it. Kyle Huey <me@kylehuey.com> - 2016-11-29 21:50 +0100
  Re: [PATCH v2 0/5] KVM: VMX: Fix single stepping with emulated  instructions Radim Krčmář <rkrcmar@redhat.com> - 2016-11-30 19:50 +0100
    Re: [PATCH v2 0/5] KVM: VMX: Fix single stepping with emulated instructions Kyle Huey <me@kylehuey.com> - 2016-11-30 20:00 +0100

csiph-web