Re: [kvm-unit-tests PATCH 4/6] arm: pmu: Fix chain counter enable/disable sequences

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2023-04-21 11:52, Alexandru Elisei wrote:
Hi,

On Wed, Mar 15, 2023 at 12:07:23PM +0100, Eric Auger wrote:
In some ARM ARM ddi0487 revisions it is said that
disabling/enabling a pair of counters that are paired
by a CHAIN event should follow a given sequence:

Disable the low counter first, isb, disable the high counter
Enable the high counter first, isb, enable low counter

This was the case in Fc. However this is not written anymore
in Ia revision.

Introduce 2 helpers to execute those sequences and replace
the existing PMCNTENCLR/ENSET calls.

Also fix 2 write_sysreg_s(0x0, PMCNTENSET_EL0) in subtest 5 & 6
and replace them by PMCNTENCLR writes since writing 0 in
PMCNTENSET_EL0 has no effect.

Signed-off-by: Eric Auger <eric.auger@xxxxxxxxxx>
---
 arm/pmu.c | 37 ++++++++++++++++++++++++++++---------
 1 file changed, 28 insertions(+), 9 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index dde399e2..af679667 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -730,6 +730,22 @@ static void test_chained_sw_incr(bool unused)
 		    read_regn_el0(pmevcntr, 0), \
 		    read_sysreg(pmovsclr_el0))

+static void enable_chain_counter(int even)
+{
+ write_sysreg_s(BIT(even), PMCNTENSET_EL0); /* Enable the high counter first */
+	isb();
+ write_sysreg_s(BIT(even + 1), PMCNTENSET_EL0); /* Enable the low counter */
+	isb();
+}

In ARM DDI 0487F.b, at the bottom of page D7-2727:

"When enabling a pair of counters that are paired by a CHAIN event,
software must:

1. Enable the high counter, by setting PMCNTENCLR_EL0[n+1] to 0 and, if
necessary, setting PMCR_EL0.E to 1.
2. Execute an ISB instruction, or perform another Context synchronization
event.
3. Enable the low counter by setting PMCNTENCLR_EL0[n] to 0."

This particular text seems to have been removed from the H.a and I.a
revisions of the ARM ARM, and I cannot spot any equivalent requirement.

        M.
--
Jazz is not dead. It just smells funny...



[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux