linux/arch/arm64
Alexandre Chartre ca4f202d08 KVM: arm64: Disabling disabled PMU counters wastes a lot of time
In a KVM guest on arm64, performance counters interrupts have an
unnecessary overhead which slows down execution when using the "perf
record" command and limits the "perf record" sampling period.

The problem is that when a guest VM disables counters by clearing the
PMCR_EL0.E bit (bit 0), KVM will disable all counters defined in
PMCR_EL0 even if they are not enabled in PMCNTENSET_EL0.

KVM disables a counter by calling into the perf framework, in particular
by calling perf_event_create_kernel_counter() which is a time consuming
operation. So, for example, with a Neoverse N1 CPU core which has 6 event
counters and one cycle counter, KVM will always disable all 7 counters
even if only one is enabled.

This typically happens when using the "perf record" command in a guest
VM: perf will disable all event counters with PMCNTENTSET_EL0 and only
uses the cycle counter. And when using the "perf record" -F option with
a high profiling frequency, the overhead of KVM disabling all counters
instead of one on every counter interrupt becomes very noticeable.

The problem is fixed by having KVM disable only counters which are
enabled in PMCNTENSET_EL0. If a counter is not enabled in PMCNTENSET_EL0
then KVM will not enable it when setting PMCR_EL0.E and it will remain
disabled as long as it is not enabled in PMCNTENSET_EL0. So there is
effectively no need to disable a counter when clearing PMCR_EL0.E if it
is not enabled PMCNTENSET_EL0.

Acked-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Alexandre Chartre <alexandre.chartre@oracle.com>
[maz: moved 'mask' close to the actual user, simplifying the patch]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210712170345.660272-1-alexandre.chartre@oracle.com
Link: https://lore.kernel.org/r/20210719123902.1493805-4-maz@kernel.org
2021-08-02 14:26:34 +01:00
..
boot Networking fixes for 5.14-rc4, including fixes from bpf, can, WiFi (mac80211) 2021-07-30 16:01:36 -07:00
configs Additional Qualcomm ARM64 defconfig udpate for v5.14 2021-06-23 19:01:58 -07:00
crypto crypto: arm64 - use a pattern rule for generating *.S files 2021-05-14 19:07:54 +08:00
include Revert "arm64: cache: Lower ARCH_DMA_MINALIGN to 64 (L1_CACHE_BYTES)" 2021-07-12 13:37:34 +01:00
kernel arm64: smccc: Save lr before calling __arm_smccc_sve_check() 2021-07-21 11:23:25 +01:00
kvm KVM: arm64: Disabling disabled PMU counters wastes a lot of time 2021-08-02 14:26:34 +01:00
lib arm64: Avoid premature usercopy failure 2021-07-15 17:29:14 +01:00
mm Revert "mm/pgtable: add stubs for {pmd/pub}_{set/clear}_huge" 2021-07-21 11:28:09 +01:00
net bpf: Introduce BPF nospec instruction for mitigating Spectre v4 2021-07-29 00:20:56 +02:00
tools arm64: Allow mismatched 32-bit EL0 support 2021-06-11 13:25:40 +01:00
xen
Kbuild bpf: Fix BPF_JIT kconfig symbol dependency 2021-05-20 23:48:37 +02:00
Kconfig arm64: Restrict ARM64_BTI_KERNEL to clang 12.0.0 and newer 2021-07-13 16:31:31 +01:00
Kconfig.debug
Kconfig.platforms Merge branch 'arm/fixes' into arm/soc 2021-07-09 10:46:02 -07:00
Makefile Kbuild updates for v5.14 2021-07-10 11:01:38 -07:00