perf/arm_pmuv3: Add PMUv3.9 per counter EL0 access control

Armv8.9/9.4 PMUv3.9 adds per counter EL0 access controls. Per counter
access is enabled with the UEN bit in PMUSERENR_EL1 register. Individual
counters are enabled/disabled in the PMUACR_EL1 register. When UEN is
set, the CR/ER bits control EL0 write access and must be set to disable
write access.

With the access controls, the clearing of unused counters can be
skipped.

KVM also configures PMUSERENR_EL1 in order to trap to EL2. UEN does not
need to be set for it since only PMUv3.5 is exposed to guests.

Signed-off-by: Rob Herring (Arm) <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Will Deacon <[e-mail]>

arch/arm/include/asm/arm_pmuv3.h

···
 #define ARMV8_PMU_DFR_VER_V3P1      0x4
 #define ARMV8_PMU_DFR_VER_V3P4      0x5
 #define ARMV8_PMU_DFR_VER_V3P5      0x6
+#define ARMV8_PMU_DFR_VER_V3P9      0x9
 #define ARMV8_PMU_DFR_VER_IMP_DEF   0xF
 
 static inline bool pmuv3_implemented(int pmuver)
···
 static inline bool is_pmuv3p5(int pmuver)
 {
 	return pmuver >= ARMV8_PMU_DFR_VER_V3P5;
+}
+
+static inline bool is_pmuv3p9(int pmuver)
+{
+	return pmuver >= ARMV8_PMU_DFR_VER_V3P9;
 }
 
 static inline u64 read_pmceid0(void)

arch/arm64/include/asm/arm_pmuv3.h

···
 	write_sysreg(val, pmuserenr_el0);
 }
 
+static inline void write_pmuacr(u64 val)
+{
+	write_sysreg_s(val, SYS_PMUACR_EL1);
+}
+
 static inline u64 read_pmceid0(void)
 {
 	return read_sysreg(pmceid0_el0);
···
 static inline bool is_pmuv3p5(int pmuver)
 {
 	return pmuver >= ID_AA64DFR0_EL1_PMUVer_V3P5;
+}
+
+static inline bool is_pmuv3p9(int pmuver)
+{
+	return pmuver >= ID_AA64DFR0_EL1_PMUVer_V3P9;
 }
 
 #endif

arch/arm64/tools/sysreg

···
 	0b0110	V3P5
 	0b0111	V3P7
 	0b1000	V3P8
+	0b1001	V3P9
 	0b1111	IMP_DEF
 EndEnum
 UnsignedEnum	7:4	TraceVer
···
 Field	5	F
 Field	4	P
 Field	3:0	ALIGN
+EndSysreg
+
+Sysreg	PMUACR_EL1	3	0	9	14	4
+Res0	63:33
+Field	32	F0
+Field	31	C
+Field	30:0	P
 EndSysreg
 
 Sysreg	PMSELR_EL0	3	3	9	12	5

drivers/perf/arm_pmuv3.c

···
 	int i;
 	struct pmu_hw_events *cpuc = this_cpu_ptr(cpu_pmu->hw_events);
 
-	/* Clear any unused counters to avoid leaking their contents */
-	for_each_andnot_bit(i, cpu_pmu->cntr_mask, cpuc->used_mask,
-			    ARMPMU_MAX_HWEVENTS) {
-		if (i == ARMV8_PMU_CYCLE_IDX)
-			write_pmccntr(0);
-		else if (i == ARMV8_PMU_INSTR_IDX)
-			write_pmicntr(0);
-		else
-			armv8pmu_write_evcntr(i, 0);
+	if (is_pmuv3p9(cpu_pmu->pmuver)) {
+		u64 mask = 0;
+		for_each_set_bit(i, cpuc->used_mask, ARMPMU_MAX_HWEVENTS) {
+			if (armv8pmu_event_has_user_read(cpuc->events[i]))
+				mask |= BIT(i);
+		}
+		write_pmuacr(mask);
+	} else {
+		/* Clear any unused counters to avoid leaking their contents */
+		for_each_andnot_bit(i, cpu_pmu->cntr_mask, cpuc->used_mask,
+				    ARMPMU_MAX_HWEVENTS) {
+			if (i == ARMV8_PMU_CYCLE_IDX)
+				write_pmccntr(0);
+			else if (i == ARMV8_PMU_INSTR_IDX)
+				write_pmicntr(0);
+			else
+				armv8pmu_write_evcntr(i, 0);
+		}
 	}
 
-	update_pmuserenr(ARMV8_PMU_USERENR_ER | ARMV8_PMU_USERENR_CR);
+	update_pmuserenr(ARMV8_PMU_USERENR_ER | ARMV8_PMU_USERENR_CR | ARMV8_PMU_USERENR_UEN);
 }
 
 static void armv8pmu_enable_event(struct perf_event *event)

include/linux/perf/arm_pmuv3.h

···
 #define ARMV8_PMU_USERENR_SW	(1 << 1) /* PMSWINC can be written at EL0 */
 #define ARMV8_PMU_USERENR_CR	(1 << 2) /* Cycle counter can be read at EL0 */
 #define ARMV8_PMU_USERENR_ER	(1 << 3) /* Event counter can be read at EL0 */
+#define ARMV8_PMU_USERENR_UEN	(1 << 4) /* Fine grained per counter access at EL0 */
 /* Mask for writable bits */
 #define ARMV8_PMU_USERENR_MASK	(ARMV8_PMU_USERENR_EN | ARMV8_PMU_USERENR_SW | \
 				 ARMV8_PMU_USERENR_CR | ARMV8_PMU_USERENR_ER)