kernel.git/arch/arm64/kvm/debug.c, branch linux-6.15.y

KVM: arm64: Support trace filtering for guests

2025-01-12T12:50:11Z

For nVHE, switch the filter value in and out if the Coresight driver asks for it. This will support filters for guests when sinks other than TRBE are used. For VHE, just write the filter directly to TRFCR_EL1 where trace can be used even with TRBE sinks. Signed-off-by: James Clark Link: https://lore.kernel.org/r/20250106142446.628923-7-james.clark@linaro.org Signed-off-by: Marc Zyngier

KVM: arm64: coresight: Give TRBE enabled state to KVM

2025-01-12T12:50:11Z

Currently in nVHE, KVM has to check if TRBE is enabled on every guest switch even if it was never used. Because it's a debug feature and is more likely to not be used than used, give KVM the TRBE buffer status to allow a much simpler and faster do-nothing path in the hyp. Protected mode now disables trace regardless of TRBE (because trfcr_while_in_guest is always 0), which was not previously done. However, it continues to flush whenever the buffer is enabled regardless of the filter status. This avoids the hypothetical case of a host that had disabled the filter but not flushed which would arise if only doing the flush when the filter was enabled. Signed-off-by: James Clark Link: https://lore.kernel.org/r/20250106142446.628923-6-james.clark@linaro.org Signed-off-by: Marc Zyngier

KVM: arm64: Avoid reading ID_AA64DFR0_EL1 for debug save/restore

2024-12-20T09:04:14Z

Similar to other per-CPU profiling/debug features we handle, store the number of breakpoints/watchpoints in kvm_host_data to avoid reading the ID register 4 times on every guest entry/exit. And if you're in the nested virt business that's quite a few avoidable exits to the L0 hypervisor. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-18-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Manage software step state at load/put

2024-12-20T09:04:06Z

KVM takes over the guest's software step state machine if the VMM is debugging the guest, but it does the save/restore fiddling for every guest entry. Note that the only constraint on host usage of software step is that the guest's configuration remains visible to userspace via the ONE_REG ioctls. So, we can cut down on the amount of fiddling by doing this at load/put instead. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-16-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Don't hijack guest context MDSCR_EL1

2024-12-20T09:02:51Z

Stealing MDSCR_EL1 in the guest's kvm_cpu_context for external debugging is rather gross. Just add a field for this instead and let the context switch code pick the correct one based on the debug owner. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-15-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Compute MDCR_EL2 at vcpu_load()

2024-12-20T09:01:25Z

KVM has picked up several hacks to cope with vcpu->arch.mdcr_el2 needing to be prepared before vcpu_load(), which is when it gets programmed into hardware on VHE. Now that the flows for reprogramming MDCR_EL2 have been simplified, move that computation to vcpu_load(). Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-14-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Reload vCPU for accesses to OSLAR_EL1

2024-12-20T09:01:25Z

KVM takes ownership of the debug regs if the guest enables the OS lock, as it needs to use MDSCR_EL1 to mask debug exceptions. Just reload the vCPU if the guest toggles the OS lock, relying on kvm_vcpu_load_debug() to update the debug owner and get the right trap configuration in place. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-13-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Use debug_owner to track if debug regs need save/restore

2024-12-20T09:01:25Z

Use the debug owner to determine if the debug regs are in use instead of keeping around the DEBUG_DIRTY flag. Debug registers are now saved/restored after the first trap, regardless of whether it was a read or a write. This also shifts the point at which KVM becomes lazy to vcpu_put() rather than the next exception taken from the guest. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-12-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Remove vestiges of debug_ptr

2024-12-20T09:01:25Z

Delete the remnants of debug_ptr now that debug registers are selected based on the debug owner instead. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-11-oliver.upton@linux.dev Signed-off-by: Marc Zyngier

KVM: arm64: Remove debug tracepoints

2024-12-20T09:01:25Z

The debug tracepoints are a useless firehose of information that track implementation detail rather than well-defined events. These are going to be rather difficult to uphold now that the implementation is getting redone, so throw them out instead of bending over backwards. Tested-by: James Clark Signed-off-by: Oliver Upton Link: https://lore.kernel.org/r/20241219224116.3941496-10-oliver.upton@linux.dev [maz: fixed compilation after trace-ectomy] Signed-off-by: Marc Zyngier