Skip to content

KVM: arm64: timer fixes and optimisations

Marc Zyngier kvm kvmarm linux-arm-kernel 1 version Applied

Status from this message: Applied to kvmarm/next, thanks!

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest3 patches 1 replyApplied

Patches in v1 Applied

#Subject and filesSizeRepliesApplied
1/3KVM: arm64: Don't arm a hrtimer for an already pending timerarch/arm64/kvm/arch_timer.c+1 -30 replies
2/3KVM: arm64: Reduce overhead of trapped timer sysreg accessesarch/arm64/kvm/arch_timer.c+48 -250 replies
3/3KVM: arm64: timers: Don't BUG() on unhandled timer traparch/arm64/kvm/sys_regs.c+3 -10 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [PATCH 0/3] KVM: arm64: timer fixes and optimisations

Having been busy on the NV front the past few weeks, I collected a small set of fixes/improvements that make sense even outside of NV.

The first one is an interesting fix (actually a regression introduced by the initial set of NV-related patches) reported by Scott and Ganapatrao, where we fail to recognise that a timer that has fired doesn't need to fire again. And again.

The second patch is a definite performance improvement on nVHE systems when accessing the emulated physical timer. It also makes NV bearable in some conditions (with FEAT_ECV, for example).

The last patch is more of a sanity check. There is no reason to BUG() if we can avoid it at all and kill the guest instead!

These patches are 6.3 candidates, although the first one could be a 6.2 fix.

	M.

Marc Zyngier (3):
KVM: arm64: Don't arm a hrtimer for an already pending timer
KVM: arm64: Reduce overhead of trapped timer sysreg accesses
KVM: arm64: timers: Don't BUG() on unhandled timer trap

 arch/arm64/kvm/arch_timer.c | 77 +++++++++++++++++++++++--------------
 arch/arm64/kvm/sys_regs.c   |  4 +-
 2 files changed, 52 insertions(+), 29 deletions(-)