FAILED: patch "[PATCH] irqchip/stm32mp-exti: Fix the unit of the hwspinlock timeout" failed to apply to 5.15-stable tree

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x d31fbbade43f880b7e59e2b3a72722fe2725d93f
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026090813-gallery-jackpot-25e8@gregkh' --subject-prefix 'PATCH 5.15.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From d31fbbade43f880b7e59e2b3a72722fe2725d93f Mon Sep 17 00:00:00 2001
From: Ju Nan <[email protected]>
Date: Fri, 21 Aug 2026 10:47:57 +0800
Subject: [PATCH] irqchip/stm32mp-exti: Fix the unit of the hwspinlock timeout

HWSPNLCK_TIMEOUT is passed to hwspin_lock_timeout_in_atomic(), whose timeout argument is in milliseconds, not microseconds:

  atomic_delay += HWSPINLOCK_RETRY_DELAY_US;
  if (atomic_delay > to * 1000)
          return -ETIMEDOUT;

So stm32mp_exti_set_type() asks for a 1 second timeout where the comment next to the macro says it wants 1 millisecond. The semaphore is polled with udelay() from a section that holds chip_data->rlock, a raw_spinlock_t, so preemption stays disabled for the whole wait on every configuration, PREEMPT_RT included.

The hwspinlock core documents this explicitly:

If the mode is HWLOCK_IN_ATOMIC (called from an atomic context) the timeout is handled with busy-waiting delays, hence shall not exceed few msecs.

index bf3a2def69ca..a19e91fbd010 100644
@@ -22,7 +22,7 @@
 
 #define IRQS_PER_BANK			32
 
-#define HWSPNLCK_TIMEOUT		1000 /* usec */
+#define HWSPNLCK_TIMEOUT_MS		1
 
 #define EXTI_EnCIDCFGR(n)		(0x180 + (n) * 4)
 #define EXTI_HWCFGR1			0x3f0

@@ -376,7 +376,7 @@ static int stm32mp_exti_set_type(struct irq_data *d, unsigned int type)
 	raw_spin_lock(&chip_data->rlock);
 
 	if (hwlock) {
-		err = hwspin_lock_timeout_in_atomic(hwlock, HWSPNLCK_TIMEOUT);
+		err = hwspin_lock_timeout_in_atomic(hwlock, HWSPNLCK_TIMEOUT_MS);
 		if (err) {
 			pr_err("%s can't get hwspinlock (%d)\n", __func__, err);
 			goto unlock;

From: Antonio Borneo <[email protected]>

[ Upstream commit 350755e2e548ccbe941d045900b57233efa906cb ]

Keep only the code for ARMv7m STM32 MCUs in in stm32-exti.c and split out the code for ARMv7a & ARMv8a STM32MPxxx MPUs into stm32mp-exti.c

Backport dependency adaptation for Linux 6.1:

The stable driver predates STM32MP_EXTI selection, standard device PM, interrupts-extended mappings, and secure/RIF event handling. Importing the upstream split would pull in those unrelated changes and new functions. Limit this dependency to relocating the existing combined MCU/MPU driver to irq-stm32mp-exti.c and updating its existing STM32_EXTI Makefile entry. Keep Kconfig, syscore PM, interrupt mappings, and all existing functions.

Retain the upstream removal of unused hwspinlock handling from the MCU set-type callback: MCU host initialization never assigns a hwspinlock. This leaves the MPU callback as the only timeout user, so the target's macro rename fixes all remaining users without leaving an undefined name.

Match the upstream timeout-definition context, including its two register macros, so d31fbbade43f880b7e59e2b3a72722fe2725d93f applies unchanged. The register macros remain unused; no secure/RIF behavior is introduced.

[ sashal: Reduced backport -- upstream 350755e2e548c touches 4 file(s), this
backport carries 2. Not backported here:
drivers/irqchip/irq-stm32-exti.c
drivers/irqchip/Kconfig
This note is generated from the file lists only; see the resolution record for the reasoning. ]

Stable-dep-of: d31fbbade43f ("irqchip/stm32mp-exti: Fix the unit of the hwspinlock timeout")

---
 drivers/irqchip/Makefile                      |  2 +-
 .../{irq-stm32-exti.c => irq-stm32mp-exti.c}  | 21 ++++++-------------
 2 files changed, 7 insertions(+), 16 deletions(-)
 rename drivers/irqchip/{irq-stm32-exti.c => irq-stm32mp-exti.c} (98%)

index f88cbf36a9d28..16a8e5e152a18 100644
@@ -86,7 +86,7 @@ obj-$(CONFIG_MVEBU_SEI)			+= irq-mvebu-sei.o
 obj-$(CONFIG_LS_EXTIRQ)			+= irq-ls-extirq.o
 obj-$(CONFIG_LS_SCFG_MSI)		+= irq-ls-scfg-msi.o
 obj-$(CONFIG_ARCH_ASPEED)		+= irq-aspeed-vic.o irq-aspeed-i2c-ic.o irq-aspeed-scu-ic.o
-obj-$(CONFIG_STM32_EXTI) 		+= irq-stm32-exti.o
+obj-$(CONFIG_STM32_EXTI) 		+= irq-stm32mp-exti.o
 obj-$(CONFIG_QCOM_IRQ_COMBINER)		+= qcom-irq-combiner.o
 obj-$(CONFIG_IRQ_UNIPHIER_AIDET)	+= irq-uniphier-aidet.o
 obj-$(CONFIG_ARCH_SYNQUACER)		+= irq-sni-exiu.o

similarity index 98%
rename from drivers/irqchip/irq-stm32-exti.c
rename to drivers/irqchip/irq-stm32mp-exti.c
index 9c150c402f0ba..71a46e76d8320 100644
@@ -22,9 +22,12 @@
 
 #include <dt-bindings/interrupt-controller/arm-gic.h>
 
-#define IRQS_PER_BANK 32
+#define IRQS_PER_BANK			32
 
-#define HWSPNLCK_TIMEOUT	1000 /* usec */
+#define HWSPNLCK_TIMEOUT		1000 /* usec */
+
+#define EXTI_EnCIDCFGR(n)		(0x180 + (n) * 4)
+#define EXTI_HWCFGR1			0x3f0
 
 struct stm32_exti_bank {
 	u32 imr_ofst;

@@ -306,33 +309,21 @@ static int stm32_irq_set_type(struct irq_data *d, unsigned int type)
 	struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d);
 	struct stm32_exti_chip_data *chip_data = gc->private;
 	const struct stm32_exti_bank *stm32_bank = chip_data->reg_bank;
-	struct hwspinlock *hwlock = chip_data->host_data->hwlock;
 	u32 rtsr, ftsr;
 	int err;
 
 	irq_gc_lock(gc);
 
-	if (hwlock) {
-		err = hwspin_lock_timeout_in_atomic(hwlock, HWSPNLCK_TIMEOUT);
-		if (err) {
-			pr_err("%s can't get hwspinlock (%d)\n", __func__, err);
-			goto unlock;
-		}
-	}
-
 	rtsr = irq_reg_readl(gc, stm32_bank->rtsr_ofst);
 	ftsr = irq_reg_readl(gc, stm32_bank->ftsr_ofst);
 
 	err = stm32_exti_set_type(d, type, &rtsr, &ftsr);
 	if (err)
-		goto unspinlock;
+		goto unlock;
 
 	irq_reg_writel(gc, rtsr, stm32_bank->rtsr_ofst);
 	irq_reg_writel(gc, ftsr, stm32_bank->ftsr_ofst);
 
-unspinlock:
-	if (hwlock)
-		hwspin_unlock_in_atomic(hwlock);
 unlock:
 	irq_gc_unlock(gc);
 

2.53.0

From: Ju Nan <[email protected]>

[ Upstream commit d31fbbade43f880b7e59e2b3a72722fe2725d93f ]

HWSPNLCK_TIMEOUT is passed to hwspin_lock_timeout_in_atomic(), whose timeout argument is in milliseconds, not microseconds:

  atomic_delay += HWSPINLOCK_RETRY_DELAY_US;
  if (atomic_delay > to * 1000)
          return -ETIMEDOUT;

So stm32mp_exti_set_type() asks for a 1 second timeout where the comment next to the macro says it wants 1 millisecond. The semaphore is polled with udelay() from a section that holds chip_data->rlock, a raw_spinlock_t, so preemption stays disabled for the whole wait on every configuration, PREEMPT_RT included.

The hwspinlock core documents this explicitly:

If the mode is HWLOCK_IN_ATOMIC (called from an atomic context) the timeout is handled with busy-waiting delays, hence shall not exceed few msecs.

---
 drivers/irqchip/irq-stm32mp-exti.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 71a46e76d8320..85ed9fe52e6bc 100644
@@ -24,7 +24,7 @@
 
 #define IRQS_PER_BANK			32
 
-#define HWSPNLCK_TIMEOUT		1000 /* usec */
+#define HWSPNLCK_TIMEOUT_MS		1
 
 #define EXTI_EnCIDCFGR(n)		(0x180 + (n) * 4)
 #define EXTI_HWCFGR1			0x3f0

@@ -508,7 +508,7 @@ static int stm32_exti_h_set_type(struct irq_data *d, unsigned int type)
 	raw_spin_lock(&chip_data->rlock);
 
 	if (hwlock) {
-		err = hwspin_lock_timeout_in_atomic(hwlock, HWSPNLCK_TIMEOUT);
+		err = hwspin_lock_timeout_in_atomic(hwlock, HWSPNLCK_TIMEOUT_MS);
 		if (err) {
 			pr_err("%s can't get hwspinlock (%d)\n", __func__, err);
 			goto unlock;

2.53.0