Kexec Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2] genirq/kexec: Do not take the bus lock of an interrupt chip on the crash path
@ 2026-10-15 18:51 Igor Velkov via B4 Relay
  2026-10-15 18:58 ` sashiko-bot
  0 siblings, 1 reply; 2+ messages in thread
From: Igor Velkov via B4 Relay @ 2026-10-15 18:51 UTC (permalink / raw)
  To: Thomas Gleixner
  Cc: Radu Rendec, Marc Zyngier, Eliav Farber, Catalin Marinas,
	Will Deacon, Mark Rutland, Baoquan He, Heiko Stuebner, kexec,
	linux-arm-kernel, linux-rockchip, linux-kernel, Igor Velkov

From: Igor Velkov <iav@iav.lv>

machine_kexec_mask_interrupts() calls irq_set_irqchip_state() for every
started interrupt. That function takes the bus lock of the chip and
desc->lock, and syncs the chip when it drops them. On the crash path
interrupts are off and the other CPUs are stopped, so a chip whose bus
transfer is completed by an interrupt never returns from the sync.

On RK3399 boards with an RK808 PMIC the RTC alarm interrupt belongs to
the regmap-irq chip of the RK808, which sits on I2C. rk808_irq_chip sets
init_ack_masked, so regmap_irq_sync_unlock() writes the ack register on
every unlock, and the I2C controller needs its own interrupt to finish
the transfer. A panic with a crash kernel loaded prints

  SMP: stopping secondary CPUs

and then nothing until the watchdog resets the board.

machine_kexec_mask_interrupts() takes neither lock for the other
operations on the descriptor. Move the search for the chip that
implements irq_set_irqchip_state() into __irq_set_irqchip_state(), which
takes no locks, and call it from machine_kexec_mask_interrupts().

Fixes: 78fd584cdec0 ("arm64: kdump: implement machine_crash_shutdown()")
Suggested-by: Thomas Gleixner <tglx@kernel.org>
Assisted-by: LLM
Signed-off-by: Igor Velkov <iav@iav.lv>
---
Tested with sysrq-c and kdump-tools on Kobol Helios64 (RK3399, RK808),
7.3-rc7: the dump is saved 3 of 3. Without a fix it hangs 3 of 3 (7.3-rc5).
---
Changes in v2:
- Clear the active state through a lockless __irq_set_irqchip_state()
  instead of skipping chips with irq_bus_lock. The manage.c part is
  Thomas' proposal from the v1 thread.
- Link to v1: https://lore.kernel.org/r/20261005-kexec-irq-bus-lock-v1-1-09f6b5323460@iav.lv
---
 kernel/irq/internals.h |  1 +
 kernel/irq/kexec.c     |  2 +-
 kernel/irq/manage.c    | 43 ++++++++++++++++++++++++-------------------
 3 files changed, 26 insertions(+), 20 deletions(-)

diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h
index 2c30e7537987..703a062be878 100644
--- a/kernel/irq/internals.h
+++ b/kernel/irq/internals.h
@@ -83,6 +83,7 @@ enum {
 #include "settings.h"
 
 extern int __irq_set_trigger(struct irq_desc *desc, unsigned long flags);
+extern int __irq_set_irqchip_state(struct irq_desc *desc, enum irqchip_irq_state which, bool val);
 extern void __disable_irq(struct irq_desc *desc);
 extern void __enable_irq(struct irq_desc *desc);
 
diff --git a/kernel/irq/kexec.c b/kernel/irq/kexec.c
index 1a3deffe6b5b..f245b9876422 100644
--- a/kernel/irq/kexec.c
+++ b/kernel/irq/kexec.c
@@ -25,7 +25,7 @@ void machine_kexec_mask_interrupts(void)
 			 * First try to remove the active state from an interrupt which is forwarded
 			 * to a VM. If the interrupt is not forwarded, try to EOI the interrupt.
 			 */
-			check_eoi = irq_set_irqchip_state(i, IRQCHIP_STATE_ACTIVE, false);
+			check_eoi = __irq_set_irqchip_state(desc, IRQCHIP_STATE_ACTIVE, false);
 		}
 
 		if (check_eoi && chip->irq_eoi && irqd_irq_inprogress(&desc->irq_data))
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 4fbb8f960dbf..94c2e13ddf8d 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -2747,6 +2747,28 @@ int irq_get_irqchip_state(unsigned int irq, enum irqchip_irq_state which, bool *
 }
 EXPORT_SYMBOL_GPL(irq_get_irqchip_state);
 
+int __irq_set_irqchip_state(struct irq_desc *desc, enum irqchip_irq_state which, bool val)
+{
+	struct irq_data *data = irq_desc_get_irq_data(desc);
+	struct irq_chip *chip;
+
+	do {
+		chip = irq_data_get_irq_chip(data);
+
+		if (WARN_ON_ONCE(!chip))
+			return -ENODEV;
+
+		if (chip->irq_set_irqchip_state)
+			break;
+
+		data = irqd_get_parent_data(data);
+	} while (data);
+
+	if (data)
+		return chip->irq_set_irqchip_state(data, which, val);
+	return -EINVAL;
+}
+
 /**
  * irq_set_irqchip_state - set the state of a forwarded interrupt.
  * @irq:	Interrupt line that is forwarded to a VM
@@ -2761,25 +2783,8 @@ EXPORT_SYMBOL_GPL(irq_get_irqchip_state);
  */
 int irq_set_irqchip_state(unsigned int irq, enum irqchip_irq_state which, bool val)
 {
-	scoped_irqdesc_get_and_buslock(irq, 0) {
-		struct irq_data *data = irq_desc_get_irq_data(scoped_irqdesc);
-		struct irq_chip *chip;
-
-		do {
-			chip = irq_data_get_irq_chip(data);
-
-			if (WARN_ON_ONCE(!chip))
-				return -ENODEV;
-
-			if (chip->irq_set_irqchip_state)
-				break;
-
-			data = irqd_get_parent_data(data);
-		} while (data);
-
-		if (data)
-			return chip->irq_set_irqchip_state(data, which, val);
-	}
+	scoped_irqdesc_get_and_buslock(irq, 0)
+		return __irq_set_irqchip_state(scoped_irqdesc, which, val);
 	return -EINVAL;
 }
 EXPORT_SYMBOL_GPL(irq_set_irqchip_state);

---
base-commit: 4351cc74202857f5b85b3573c477c9f798f5cd88
change-id: 20261005-kexec-irq-bus-lock-f48d5966443d

Best regards,
-- 
Igor Velkov <iav@iav.lv>




^ permalink raw reply related	[flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-10-15 18:58 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-15 18:51 [PATCH v2] genirq/kexec: Do not take the bus lock of an interrupt chip on the crash path Igor Velkov via B4 Relay
2026-10-15 18:58 ` sashiko-bot

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox