powerpc/64: Fix perf profiling asynchronous interrupt handlers

Interrupt entry sets the soft mask to IRQS_ALL_DISABLED to match the
hard irq disabled state. So when should_hard_irq_enable() returns true
because we want PMI interrupts in irq handlers, MSR[EE] is enabled but
PMIs just get soft-masked. Fix this by clearing IRQS_PMI_DISABLED before
enabling MSR[EE].

This also tidies some of the warnings, no need to duplicate them in
both should_hard_irq_enable() and do_hard_irq_enable().
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20230121100156.2824054-1-npiggin@gmail.com

arch/powerpc/include/asm/hw_irq.h

@@ -173,6 +173,15 @@ static inline notrace unsigned long irq_soft_mask_or_return(unsigned long mask)
 	return flags;
 }
 
+static inline notrace unsigned long irq_soft_mask_andc_return(unsigned long mask)
+{
+	unsigned long flags = irq_soft_mask_return();
+
+	irq_soft_mask_set(flags & ~mask);
+
+	return flags;
+}
+
 static inline unsigned long arch_local_save_flags(void)
 {
 	return irq_soft_mask_return();
@@ -331,10 +340,11 @@ bool power_pmu_wants_prompt_pmi(void);
  * is a different soft-masked interrupt pending that requires hard
  * masking.
  */
-static inline bool should_hard_irq_enable(void)
+static inline bool should_hard_irq_enable(struct pt_regs *regs)
 {
 	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
-		WARN_ON(irq_soft_mask_return() == IRQS_ENABLED);
+		WARN_ON(irq_soft_mask_return() != IRQS_ALL_DISABLED);
+		WARN_ON(!(get_paca()->irq_happened & PACA_IRQ_HARD_DIS));
 		WARN_ON(mfmsr() & MSR_EE);
 	}
 
@@ -347,8 +357,17 @@ static inline bool should_hard_irq_enable(void)
 	 *
 	 * TODO: Add test for 64e
 	 */
-	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && !power_pmu_wants_prompt_pmi())
+	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64)) {
+		if (!power_pmu_wants_prompt_pmi())
 			return false;
+		/*
+		 * If PMIs are disabled then IRQs should be disabled as well,
+		 * so we shouldn't see this condition, check for it just in
+		 * case because we are about to enable PMIs.
+		 */
+		if (WARN_ON_ONCE(regs->softe & IRQS_PMI_DISABLED))
+			return false;
+	}
 
 	if (get_paca()->irq_happened & PACA_IRQ_MUST_HARD_MASK)
 		return false;
@@ -358,18 +377,16 @@ static inline bool should_hard_irq_enable(void)
 
 /*
  * Do the hard enabling, only call this if should_hard_irq_enable is true.
+ * This allows PMI interrupts to profile irq handlers.
  */
 static inline void do_hard_irq_enable(void)
 {
-	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
-		WARN_ON(irq_soft_mask_return() == IRQS_ENABLED);
-		WARN_ON(get_paca()->irq_happened & PACA_IRQ_MUST_HARD_MASK);
-		WARN_ON(mfmsr() & MSR_EE);
-	}
 	/*
-	 * This allows PMI interrupts (and watchdog soft-NMIs) through.
-	 * There is no other reason to enable this way.
+	 * Asynch interrupts come in with IRQS_ALL_DISABLED,
+	 * PACA_IRQ_HARD_DIS, and MSR[EE]=0.
 	 */
+	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64))
+		irq_soft_mask_andc_return(IRQS_PMI_DISABLED);
 	get_paca()->irq_happened &= ~PACA_IRQ_HARD_DIS;
 	__hard_irq_enable();
 }
@@ -452,7 +469,7 @@ static inline bool arch_irq_disabled_regs(struct pt_regs *regs)
 	return !(regs->msr & MSR_EE);
 }
 
-static __always_inline bool should_hard_irq_enable(void)
+static __always_inline bool should_hard_irq_enable(struct pt_regs *regs)
 {
 	return false;
 }

arch/powerpc/kernel/dbell.c

@@ -27,7 +27,7 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(doorbell_exception)
 
 	ppc_msgsync();
 
-	if (should_hard_irq_enable())
+	if (should_hard_irq_enable(regs))
 		do_hard_irq_enable();
 
 	kvmppc_clear_host_ipi(smp_processor_id());

arch/powerpc/kernel/irq.c

@@ -238,7 +238,7 @@ static void __do_irq(struct pt_regs *regs, unsigned long oldsp)
 	irq = static_call(ppc_get_irq)();
 
 	/* We can hard enable interrupts now to allow perf interrupts */
-	if (should_hard_irq_enable())
+	if (should_hard_irq_enable(regs))
 		do_hard_irq_enable();
 
 	/* And finally process it */

arch/powerpc/kernel/time.c

@@ -515,7 +515,7 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(timer_interrupt)
 	}
 
 	/* Conditionally hard-enable interrupts. */
-	if (should_hard_irq_enable()) {
+	if (should_hard_irq_enable(regs)) {
 		/*
 		 * Ensure a positive value is written to the decrementer, or
 		 * else some CPUs will continue to take decrementer exceptions.