On 2020-02-10 12:27, Mark Rutland wrote:
Across arm64 we use cpus_have_const_cap() to check for a capability
without a runtime check. Prior to capabilities being finalized
cpus_have_const_cap() falls back to a runtime check of the cpu_hwcaps
array. In some cases we know that code is never invoked prior to the
capabilities being finalized, and the fallback code is redundant (and
unsound if ever executed in hyp context).
So that we can avoid the redundant code and detect when the caps are
unexpectedly checked too early, this series adds a new
cpus_have_final_cap() helper, and migrates the KVM hyp code over to it.
I'm hoping to use this as part of the entry.S -> entry-common.c
conversion, and there are other places in arm64 that could potentially
use this today.
Thanks,
Mark.
Mark Rutland (2):
arm64: cpufeature: add cpus_have_final_cap()
arm64: kvm: hyp: use cpus_have_final_cap()
arch/arm64/include/asm/cpufeature.h | 47
+++++++++++++++++++++++++++++++++----
arch/arm64/kvm/hyp/switch.c | 14 +++++------
arch/arm64/kvm/hyp/sysreg-sr.c | 8 +++----
arch/arm64/kvm/hyp/tlb.c | 8 +++----
4 files changed, 57 insertions(+), 20 deletions(-)
Seems like a valuable optimization. For the series:
Reviewed-by: Marc Zyngier <m...@kernel.org>
M.
--
Jazz is not dead. It just smells funny...
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
