genirq: Add protection against unsafe usage of generic_handle_irq()

task #29600094

commit c16816acd08697b02a53f56f8936497a9f6f6e7a upstream.
Backport summary: for 4.19 kernel ICX PCIe Gen4 support.

In general calling generic_handle_irq() with interrupts disabled from non
interrupt context is harmless. For some interrupt controllers like the x86
trainwrecks this is outright dangerous as it might corrupt state if an
interrupt affinity change is pending.

Add infrastructure which allows to mark interrupts as unsafe and catch such
usage in generic_handle_irq().

Reported-by: sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NMarc Zyngier <maz@kernel.org>
Link: https://lkml.kernel.org/r/20200306130623.590923677@linutronix.de

(cherry picked from commit c16816acd08697b02a53f56f8936497a9f6f6e7a)
Signed-off-by: NEthan Zhao <haifeng.zhao@intel.com>

Conflicts:
	include/linux/irq.h
Signed-off-by: NArtie Ding <artie.ding@linux.alibaba.com>
Acked-by: NCaspar Zhang <caspar@linux.alibaba.com>

include/linux/irq.h

@@ -208,6 +208,10 @@ struct irq_data {
  * IRQD_SINGLE_TARGET		- IRQ allows only a single affinity target
  * IRQD_DEFAULT_TRIGGER_SET	- Expected trigger already been set
  * IRQD_CAN_RESERVE		- Can use reservation mode
+ * IRQD_MSI_NOMASK_QUIRK	- Non-maskable MSI quirk for affinity change
+ *				  required
+ * IRQD_HANDLE_ENFORCE_IRQCTX	- Enforce that handle_irq_*() is only invoked
+ *				  from actual interrupt context.
  */
 enum {
 	IRQD_TRIGGER_MASK		= 0xf,
@@ -230,6 +234,8 @@ enum {
 	IRQD_SINGLE_TARGET		= (1 << 24),
 	IRQD_DEFAULT_TRIGGER_SET	= (1 << 25),
 	IRQD_CAN_RESERVE		= (1 << 26),
+	IRQD_MSI_NOMASK_QUIRK		= (1 << 27),
+	IRQD_HANDLE_ENFORCE_IRQCTX	= (1 << 28),
 };
 
 #define __irqd_to_state(d) ACCESS_PRIVATE((d)->common, state_use_accessors)
@@ -299,6 +305,16 @@ static inline bool irqd_is_single_target(struct irq_data *d)
 	return __irqd_to_state(d) & IRQD_SINGLE_TARGET;
 }
 
+static inline void irqd_set_handle_enforce_irqctx(struct irq_data *d)
+{
+	__irqd_to_state(d) |= IRQD_HANDLE_ENFORCE_IRQCTX;
+}
+
+static inline bool irqd_is_handle_enforce_irqctx(struct irq_data *d)
+{
+	return __irqd_to_state(d) & IRQD_HANDLE_ENFORCE_IRQCTX;
+}
+
 static inline bool irqd_is_wakeup_set(struct irq_data *d)
 {
 	return __irqd_to_state(d) & IRQD_WAKEUP_STATE;

kernel/irq/internals.h

@@ -424,6 +424,10 @@ static inline struct cpumask *irq_desc_get_pending_mask(struct irq_desc *desc)
 {
 	return desc->pending_mask;
 }
+static inline bool handle_enforce_irqctx(struct irq_data *data)
+{
+	return irqd_is_handle_enforce_irqctx(data);
+}
 bool irq_fixup_move_pending(struct irq_desc *desc, bool force_clear);
 #else /* CONFIG_GENERIC_PENDING_IRQ */
 static inline bool irq_can_move_pcntxt(struct irq_data *data)
@@ -450,6 +454,10 @@ static inline bool irq_fixup_move_pending(struct irq_desc *desc, bool fclear)
 {
 	return false;
 }
+static inline bool handle_enforce_irqctx(struct irq_data *data)
+{
+	return false;
+}
 #endif /* !CONFIG_GENERIC_PENDING_IRQ */
 
 #if !defined(CONFIG_IRQ_DOMAIN) || !defined(CONFIG_IRQ_DOMAIN_HIERARCHY)

kernel/irq/irqdesc.c

@@ -633,9 +633,15 @@ void irq_init_desc(unsigned int irq)
 int generic_handle_irq(unsigned int irq)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
+	struct irq_data *data;
 
 	if (!desc)
 		return -EINVAL;
+
+	data = irq_desc_get_irq_data(desc);
+	if (WARN_ON_ONCE(!in_irq() && handle_enforce_irqctx(data)))
+		return -EPERM;
+
 	generic_handle_irq_desc(desc);
 	return 0;
 }

kernel/irq/resend.c

@@ -72,8 +72,9 @@ void check_irq_resend(struct irq_desc *desc)
 		desc->istate &= ~IRQS_PENDING;
 		desc->istate |= IRQS_REPLAY;
 
-		if (!desc->irq_data.chip->irq_retrigger ||
-		    !desc->irq_data.chip->irq_retrigger(&desc->irq_data)) {
+		if ((!desc->irq_data.chip->irq_retrigger ||
+		    !desc->irq_data.chip->irq_retrigger(&desc->irq_data)) &&
+		    !handle_enforce_irqctx(&desc->irq_data)) {
 #ifdef CONFIG_HARDIRQS_SW_RESEND
 			unsigned int irq = irq_desc_get_irq(desc);