[v6,0/4] arm64: Support perf event modifiers :G and :H

Message

Andrew Murray Dec. 10, 2018, 9:45 a.m. UTC

This patchset provides support for perf event modifiers :G and :H which
allows for filtering of PMU events between host and guests when used
with KVM.

As the underlying hardware cannot distinguish between guest and host
context, the performance counters must be stopped and started upon
entry/exit to the guest. This is performed at EL2 in a way that
minimizes overhead and improves accuracy of recording events that only
occur in the requested context.

This has been tested with VHE and non-VHE kernels with a KVM guest.

Changes from v5:

 - Tweak logic in use of kvm_set_pmu_events

Changes from v4:

 - Prevent unnecessary write_sysreg calls by improving
   __pmu_switch_to_xxx logic.

Changes from v3:

 - Remove confusing _only suffix from bitfields in kvm_cpu_context
 - Remove unnecessary condition when clearing event bits in disable
 - Simplify API of KVM accessors
 - Prevent unnecessary setting of pmcnten when guest/host events are
   the same.

Changes from v2:

 - Ensured that exclude_kernel works for guest
 - Removed unnecessary exclusion of EL2 with exclude_host on !VHE
 - Renamed kvm_clr_set_host_pmu_events to reflect args order
 - Added additional information to isb patch

Changes from v1:

 - Removed unnecessary exclusion of EL1 with exclude_guest on VHE
 - Removed unnecessary isb from existing perf_event.c driver
 - Folded perf_event.c patches together
 - Added additional information to last patch commit message

Andrew Murray (4):
  arm64: arm_pmu: remove unnecessary isb instruction
  arm64: KVM: add accessors to track guest/host only counters
  arm64: arm_pmu: Add support for exclude_host/exclude_guest attributes
  arm64: KVM: Enable support for :G/:H perf event modifiers

 arch/arm64/include/asm/kvm_host.h | 24 ++++++++++++++++++
 arch/arm64/kernel/perf_event.c    | 52 +++++++++++++++++++++++++++++++++------
 arch/arm64/kvm/hyp/switch.c       | 38 ++++++++++++++++++++++++++++
 3 files changed, 106 insertions(+), 8 deletions(-)

Comments

Will Deacon Dec. 10, 2018, 6:19 p.m. UTC | #1

Hi Andrew,

On Mon, Dec 10, 2018 at 09:45:55AM +0000, Andrew Murray wrote:
> This patchset provides support for perf event modifiers :G and :H which
> allows for filtering of PMU events between host and guests when used
> with KVM.
> 
> As the underlying hardware cannot distinguish between guest and host
> context, the performance counters must be stopped and started upon
> entry/exit to the guest. This is performed at EL2 in a way that
> minimizes overhead and improves accuracy of recording events that only
> occur in the requested context.
> 
> This has been tested with VHE and non-VHE kernels with a KVM guest.
> 
> Changes from v5:
> 
>  - Tweak logic in use of kvm_set_pmu_events

This looks good to me, but I'll need Acks from the KVM folks on patches 2
and 4 before I can take it (and Christoffer has some questions on patch 2
as well).

Also, please be aware that I'm planning to close the arm64 tree later this
week so we don't have to deal with late issues going into the holidays. Last
Christmas was ruined by specdown and meltre, so I think many of us are due a
proper break this year :)

Cheers,

Will

Andrew Murray Dec. 10, 2018, 11:48 p.m. UTC | #2

On Mon, Dec 10, 2018 at 06:19:33PM +0000, Will Deacon wrote:
> Hi Andrew,
> 
> On Mon, Dec 10, 2018 at 09:45:55AM +0000, Andrew Murray wrote:
> > This patchset provides support for perf event modifiers :G and :H which
> > allows for filtering of PMU events between host and guests when used
> > with KVM.
> > 
> > As the underlying hardware cannot distinguish between guest and host
> > context, the performance counters must be stopped and started upon
> > entry/exit to the guest. This is performed at EL2 in a way that
> > minimizes overhead and improves accuracy of recording events that only
> > occur in the requested context.
> > 
> > This has been tested with VHE and non-VHE kernels with a KVM guest.
> > 
> > Changes from v5:
> > 
> >  - Tweak logic in use of kvm_set_pmu_events
> 
> This looks good to me, but I'll need Acks from the KVM folks on patches 2
> and 4 before I can take it (and Christoffer has some questions on patch 2
> as well).
> 
> Also, please be aware that I'm planning to close the arm64 tree later this
> week so we don't have to deal with late issues going into the holidays. Last
> Christmas was ruined by specdown and meltre, so I think many of us are due a
> proper break this year :)

Thanks for the heads up and sounding completely reasonable to me.

Thanks,

Andrew Murray

> 
> Cheers,
> 
> Will