mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] gpio: omap: Propagate runtime PM failures in probe and request
@ 2026-10-29 10:36 Avi Weiss
  0 siblings, 0 replies; only message in thread
From: Avi Weiss @ 2026-10-29 10:36 UTC (permalink / raw)
  To: Bartosz Golaszewski, Grygorii Strashko
  Cc: Linus Walleij, Santosh Shilimkar, Kevin Hilman, linux-gpio,
	linux-omap, linux-kernel, Avi Weiss

pm_runtime_get_sync() can fail when the OMAP PM domain cannot
resume a GPIO controller, but omap_gpio_probe() and
omap_gpio_request() ignore its return value.

Consequently, probe can register a GPIO controller despite a failed
resume, while a GPIO request can report success and modify GPIO
state even though runtime-PM acquisition failed.

Use pm_runtime_resume_and_get() in both paths and propagate
resume errors. On probe failure, disable runtime PM and unprepare
the debounce clock when applicable.

Reproduced both failures on QEMU N800 by injecting -EIO from the
OMAP PM-domain resume callback. The original driver continues
probe and accepts the GPIO request - while the patched driver
rejects both operations with -EIO. Normal GPIO controller
registration continues to work without fault injection.

Fixes: 77640aabd755 ("OMAP: GPIO: Implement GPIO as a platform device")
Fixes: 55b93c32520d ("gpio/omap: use pm-runtime framework")
Signed-off-by: Avi Weiss <thnkslprpt@gmail.com>
---

Related/similar upstream GPIO runtime-PM commits:

  e9438ab5328a (gpio: zynq: fix runtime PM leak on request error path)
  ff82fc3a4a1d (gpio: xilinx: fix runtime PM leak on request error path)

Both use pm_runtime_resume_and_get() to avoid retaining a PM usage
reference when acquisition fails.

 drivers/gpio/gpio-omap.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/drivers/gpio/gpio-omap.c b/drivers/gpio/gpio-omap.c
index a9b9a3d9631e..31525a8eedf5 100644
--- a/drivers/gpio/gpio-omap.c
+++ b/drivers/gpio/gpio-omap.c
@@ -813,8 +813,11 @@ static int omap_gpio_request(struct gpio_chip *chip, unsigned offset)
 {
 	struct gpio_bank *bank = gpiochip_get_data(chip);
 	unsigned long flags;
+	int ret;
 
-	pm_runtime_get_sync(chip->parent);
+	ret = pm_runtime_resume_and_get(chip->parent);
+	if (ret < 0)
+		return ret;
 
 	raw_spin_lock_irqsave(&bank->lock, flags);
 	omap_enable_gpio_module(bank, offset);
@@ -1459,7 +1462,13 @@ static int omap_gpio_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, bank);
 
 	pm_runtime_enable(dev);
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_resume_and_get(dev);
+	if (ret < 0) {
+		pm_runtime_disable(dev);
+		if (bank->dbck_flag)
+			clk_unprepare(bank->dbck);
+		return ret;
+	}
 
 	if (bank->is_mpuio)
 		omap_mpuio_init(bank);
-- 
2.43.0


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2026-10-29 10:36 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-10-29 10:36 [PATCH] gpio: omap: Propagate runtime PM failures in probe and request Avi Weiss

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

all inboxes | Powered by JetHome®