xen/PMU: Intercept PMU-related MSR and APIC accesses

Provide interfaces for recognizing accesses to PMU-related MSRs and
LVTPC APIC and process these accesses in Xen PMU code.

(The interrupt handler performs XENPMU_flush right away in the beginning
since no PMU emulation is available. It will be added with a later patch).

Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>

arch/x86/xen/apic.c

@@ -7,6 +7,7 @@
 #include <xen/xen.h>
 #include <xen/interface/physdev.h>
 #include "xen-ops.h"
+#include "pmu.h"
 #include "smp.h"
 
 static unsigned int xen_io_apic_read(unsigned apic, unsigned reg)
@@ -72,8 +73,10 @@ static u32 xen_apic_read(u32 reg)
 
 static void xen_apic_write(u32 reg, u32 val)
 {
-	if (reg == APIC_LVTPC)
+	if (reg == APIC_LVTPC) {
+		(void)pmu_apic_update(reg);
 		return;
+	}
 
 	/* Warn to see if there's any stray references */
 	WARN(1,"register: %x, value: %x\n", reg, val);

arch/x86/xen/enlighten.c

@@ -1031,6 +1031,9 @@ static u64 xen_read_msr_safe(unsigned int msr, int *err)
 {
 	u64 val;
 
+	if (pmu_msr_read(msr, &val, err))
+		return val;
+
 	val = native_read_msr_safe(msr, err);
 	switch (msr) {
 	case MSR_IA32_APICBASE:
@@ -1077,17 +1080,13 @@ static int xen_write_msr_safe(unsigned int msr, unsigned low, unsigned high)
 		   Xen console noise. */
 
 	default:
-		ret = native_write_msr_safe(msr, low, high);
+		if (!pmu_msr_write(msr, low, high, &ret))
+			ret = native_write_msr_safe(msr, low, high);
 	}
 
 	return ret;
 }
 
-unsigned long long xen_read_pmc(int counter)
-{
-	return 0;
-}
-
 void xen_setup_shared_info(void)
 {
 	if (!xen_feature(XENFEAT_auto_translated_physmap)) {

arch/x86/xen/pmu.c

@@ -51,6 +51,8 @@ static __read_mostly int amd_num_counters;
 /* Alias registers (0x4c1) for full-width writes to PMCs */
 #define MSR_PMC_ALIAS_MASK          (~(MSR_IA32_PERFCTR0 ^ MSR_IA32_PMC0))
 
+#define INTEL_PMC_TYPE_SHIFT        30
+
 static __read_mostly int intel_num_arch_counters, intel_num_fixed_counters;
 
 
@@ -167,6 +169,91 @@ static int is_intel_pmu_msr(u32 msr_index, int *type, int *index)
 	}
 }
 
+bool pmu_msr_read(unsigned int msr, uint64_t *val, int *err)
+{
+
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
+		if (is_amd_pmu_msr(msr)) {
+			*val = native_read_msr_safe(msr, err);
+			return true;
+		}
+	} else {
+		int type, index;
+
+		if (is_intel_pmu_msr(msr, &type, &index)) {
+			*val = native_read_msr_safe(msr, err);
+			return true;
+		}
+	}
+
+	return false;
+}
+
+bool pmu_msr_write(unsigned int msr, uint32_t low, uint32_t high, int *err)
+{
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
+		if (is_amd_pmu_msr(msr)) {
+			*err = native_write_msr_safe(msr, low, high);
+			return true;
+		}
+	} else {
+		int type, index;
+
+		if (is_intel_pmu_msr(msr, &type, &index)) {
+			*err = native_write_msr_safe(msr, low, high);
+			return true;
+		}
+	}
+
+	return false;
+}
+
+static unsigned long long xen_amd_read_pmc(int counter)
+{
+	uint32_t msr;
+	int err;
+
+	msr = amd_counters_base + (counter * amd_msr_step);
+	return native_read_msr_safe(msr, &err);
+}
+
+static unsigned long long xen_intel_read_pmc(int counter)
+{
+	int err;
+	uint32_t msr;
+
+	if (counter & (1<<INTEL_PMC_TYPE_SHIFT))
+		msr = MSR_CORE_PERF_FIXED_CTR0 + (counter & 0xffff);
+	else
+		msr = MSR_IA32_PERFCTR0 + counter;
+
+	return native_read_msr_safe(msr, &err);
+}
+
+unsigned long long xen_read_pmc(int counter)
+{
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD)
+		return xen_amd_read_pmc(counter);
+	else
+		return xen_intel_read_pmc(counter);
+}
+
+int pmu_apic_update(uint32_t val)
+{
+	int ret;
+	struct xen_pmu_data *xenpmu_data = get_xenpmu_data();
+
+	if (!xenpmu_data) {
+		pr_warn_once("%s: pmudata not initialized\n", __func__);
+		return -EINVAL;
+	}
+
+	xenpmu_data->pmu.l.lapic_lvtpc = val;
+	ret = HYPERVISOR_xenpmu_op(XENPMU_lvtpc_set, NULL);
+
+	return ret;
+}
+
 /* perf callbacks */
 static int xen_is_in_guest(void)
 {
@@ -239,7 +326,7 @@ static void xen_convert_regs(const struct xen_pmu_regs *xen_regs,
 
 irqreturn_t xen_pmu_irq_handler(int irq, void *dev_id)
 {
-	int ret = IRQ_NONE;
+	int err, ret = IRQ_NONE;
 	struct pt_regs regs;
 	const struct xen_pmu_data *xenpmu_data = get_xenpmu_data();
 
@@ -248,6 +335,12 @@ irqreturn_t xen_pmu_irq_handler(int irq, void *dev_id)
 		return ret;
 	}
 
+	err = HYPERVISOR_xenpmu_op(XENPMU_flush, NULL);
+	if (err) {
+		pr_warn_once("%s: failed hypercall, err: %d\n", __func__, err);
+		return ret;
+	}
+
 	xen_convert_regs(&xenpmu_data->pmu.r.regs, &regs,
 			 xenpmu_data->pmu.pmu_flags);
 	if (x86_pmu.handle_irq(&regs))

arch/x86/xen/pmu.h

@@ -7,5 +7,9 @@ irqreturn_t xen_pmu_irq_handler(int irq, void *dev_id);
 void xen_pmu_init(int cpu);
 void xen_pmu_finish(int cpu);
 bool is_xen_pmu(int cpu);
+bool pmu_msr_read(unsigned int msr, uint64_t *val, int *err);
+bool pmu_msr_write(unsigned int msr, uint32_t low, uint32_t high, int *err);
+int pmu_apic_update(uint32_t reg);
+unsigned long long xen_read_pmc(int counter);
 
 #endif /* __XEN_PMU_H */

include/xen/interface/xenpmu.h

@@ -20,6 +20,8 @@
 #define XENPMU_feature_set     3
 #define XENPMU_init            4
 #define XENPMU_finish          5
+#define XENPMU_lvtpc_set       6
+#define XENPMU_flush           7
 
 /* ` } */