arm64: Relax ICC_PMR_EL1 accesses when ICC_CTLR_EL1.PMHE is clear

The GICv3 architecture specification is incredibly misleading when it
comes to PMR and the requirement for a DSB. It turns out that this DSB
is only required if the CPU interface sends an Upstream Control
message to the redistributor in order to update the RD's view of PMR.

This message is only sent when ICC_CTLR_EL1.PMHE is set, which isn't
the case in Linux. It can still be set from EL3, so some special care
is required. But the upshot is that in the (hopefuly large) majority
of the cases, we can drop the DSB altogether.

This relies on a new static key being set if the boot CPU has PMHE
set. The drawback is that this static key has to be exported to
modules.

Cc: Will Deacon <will@kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Julien Thierry <julien.thierry.kdev@gmail.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

arch/arm64/include/asm/barrier.h

@@ -29,6 +29,18 @@
 						 SB_BARRIER_INSN"nop\n",	\
 						 ARM64_HAS_SB))
 
+#ifdef CONFIG_ARM64_PSEUDO_NMI
+#define pmr_sync()						\
+	do {							\
+		extern struct static_key_false gic_pmr_sync;	\
+								\
+		if (static_branch_unlikely(&gic_pmr_sync))	\
+			dsb(sy);				\
+	} while(0)
+#else
+#define pmr_sync()	do {} while (0)
+#endif
+
 #define mb()		dsb(sy)
 #define rmb()		dsb(ld)
 #define wmb()		dsb(st)

arch/arm64/include/asm/daifflags.h

@@ -8,6 +8,7 @@
 #include <linux/irqflags.h>
 
 #include <asm/arch_gicv3.h>
+#include <asm/barrier.h>
 #include <asm/cpufeature.h>
 
 #define DAIF_PROCCTX		0
@@ -65,7 +66,7 @@ static inline void local_daif_restore(unsigned long flags)
 
 		if (system_uses_irq_prio_masking()) {
 			gic_write_pmr(GIC_PRIO_IRQON);
-			dsb(sy);
+			pmr_sync();
 		}
 	} else if (system_uses_irq_prio_masking()) {
 		u64 pmr;

arch/arm64/include/asm/irqflags.h

@@ -6,6 +6,7 @@
 #define __ASM_IRQFLAGS_H
 
 #include <asm/alternative.h>
+#include <asm/barrier.h>
 #include <asm/ptrace.h>
 #include <asm/sysreg.h>
 
@@ -34,14 +35,14 @@ static inline void arch_local_irq_enable(void)
 	}
 
 	asm volatile(ALTERNATIVE(
-		"msr	daifclr, #2		// arch_local_irq_enable\n"
-		"nop",
-		__msr_s(SYS_ICC_PMR_EL1, "%0")
-		"dsb	sy",
+		"msr	daifclr, #2		// arch_local_irq_enable",
+		__msr_s(SYS_ICC_PMR_EL1, "%0"),
 		ARM64_HAS_IRQ_PRIO_MASKING)
 		:
 		: "r" ((unsigned long) GIC_PRIO_IRQON)
 		: "memory");
+
+	pmr_sync();
 }
 
 static inline void arch_local_irq_disable(void)
@@ -116,14 +117,14 @@ static inline unsigned long arch_local_irq_save(void)
 static inline void arch_local_irq_restore(unsigned long flags)
 {
 	asm volatile(ALTERNATIVE(
-			"msr	daif, %0\n"
-			"nop",
-			__msr_s(SYS_ICC_PMR_EL1, "%0")
-			"dsb	sy",
-			ARM64_HAS_IRQ_PRIO_MASKING)
+		"msr	daif, %0",
+		__msr_s(SYS_ICC_PMR_EL1, "%0"),
+		ARM64_HAS_IRQ_PRIO_MASKING)
 		:
 		: "r" (flags)
 		: "memory");
+
+	pmr_sync();
 }
 
 #endif /* __ASM_IRQFLAGS_H */

arch/arm64/include/asm/kvm_host.h

@@ -600,8 +600,7 @@ static inline void kvm_arm_vhe_guest_enter(void)
 	 * local_daif_mask() already sets GIC_PRIO_PSR_I_SET, we just need a
 	 * dsb to ensure the redistributor is forwards EL2 IRQs to the CPU.
 	 */
-	if (system_uses_irq_prio_masking())
-		dsb(sy);
+	pmr_sync();
 }
 
 static inline void kvm_arm_vhe_guest_exit(void)

arch/arm64/kernel/entry.S

@@ -269,8 +269,10 @@ alternative_else_nop_endif
 alternative_if ARM64_HAS_IRQ_PRIO_MASKING
 	ldr	x20, [sp, #S_PMR_SAVE]
 	msr_s	SYS_ICC_PMR_EL1, x20
-	/* Ensure priority change is seen by redistributor */
-	dsb	sy
+	mrs_s	x21, SYS_ICC_CTLR_EL1
+	tbz	x21, #6, .L__skip_pmr_sync\@	// Check for ICC_CTLR_EL1.PMHE
+	dsb	sy				// Ensure priority change is seen by redistributor
+.L__skip_pmr_sync\@:
 alternative_else_nop_endif
 
 	ldp	x21, x22, [sp, #S_PC]		// load ELR, SPSR

arch/arm64/kvm/hyp/switch.c

@@ -12,7 +12,7 @@
 
 #include <kvm/arm_psci.h>
 
-#include <asm/arch_gicv3.h>
+#include <asm/barrier.h>
 #include <asm/cpufeature.h>
 #include <asm/kprobes.h>
 #include <asm/kvm_asm.h>
@@ -592,7 +592,7 @@ int __hyp_text __kvm_vcpu_run_nvhe(struct kvm_vcpu *vcpu)
 	 */
 	if (system_uses_irq_prio_masking()) {
 		gic_write_pmr(GIC_PRIO_IRQON | GIC_PRIO_PSR_I_SET);
-		dsb(sy);
+		pmr_sync();
 	}
 
 	vcpu = kern_hyp_va(vcpu);

drivers/irqchip/irq-gic-v3.c

@@ -87,6 +87,15 @@ static DEFINE_STATIC_KEY_TRUE(supports_deactivate_key);
  */
 static DEFINE_STATIC_KEY_FALSE(supports_pseudo_nmis);
 
+/*
+ * Global static key controlling whether an update to PMR allowing more
+ * interrupts requires to be propagated to the redistributor (DSB SY).
+ * And this needs to be exported for modules to be able to enable
+ * interrupts...
+ */
+DEFINE_STATIC_KEY_FALSE(gic_pmr_sync);
+EXPORT_SYMBOL(gic_pmr_sync);
+
 /* ppi_nmi_refs[n] == number of cpus having ppi[n + 16] set as NMI */
 static refcount_t *ppi_nmi_refs;
 
@@ -1502,6 +1511,17 @@ static void gic_enable_nmi_support(void)
 	for (i = 0; i < gic_data.ppi_nr; i++)
 		refcount_set(&ppi_nmi_refs[i], 0);
 
+	/*
+	 * Linux itself doesn't use 1:N distribution, so has no need to
+	 * set PMHE. The only reason to have it set is if EL3 requires it
+	 * (and we can't change it).
+	 */
+	if (gic_read_ctlr() & ICC_CTLR_EL1_PMHE_MASK)
+		static_branch_enable(&gic_pmr_sync);
+
+	pr_info("%s ICC_PMR_EL1 synchronisation\n",
+		static_branch_unlikely(&gic_pmr_sync) ? "Forcing" : "Relaxing");
+
 	static_branch_enable(&supports_pseudo_nmis);
 
 	if (static_branch_likely(&supports_deactivate_key))

include/linux/irqchip/arm-gic-v3.h

@@ -487,6 +487,8 @@
 #define ICC_CTLR_EL1_EOImode_MASK	(1 << ICC_CTLR_EL1_EOImode_SHIFT)
 #define ICC_CTLR_EL1_CBPR_SHIFT		0
 #define ICC_CTLR_EL1_CBPR_MASK		(1 << ICC_CTLR_EL1_CBPR_SHIFT)
+#define ICC_CTLR_EL1_PMHE_SHIFT		6
+#define ICC_CTLR_EL1_PMHE_MASK		(1 << ICC_CTLR_EL1_PMHE_SHIFT)
 #define ICC_CTLR_EL1_PRI_BITS_SHIFT	8
 #define ICC_CTLR_EL1_PRI_BITS_MASK	(0x7 << ICC_CTLR_EL1_PRI_BITS_SHIFT)
 #define ICC_CTLR_EL1_ID_BITS_SHIFT	11