[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20260123133614.72586-1-dakr@kernel.org>
Date: Fri, 23 Jan 2026 14:31:56 +0100
From: Danilo Krummrich <dakr@...nel.org>
To: gregkh@...uxfoundation.org,
rafael@...nel.org,
broonie@...nel.org,
will@...nel.org,
grygorii.strashko@...com,
ssantosh@...nel.org,
khilman@...nel.org,
linusw@...nel.org,
brgl@...nel.org
Cc: driver-core@...ts.linux.dev,
linux-kernel@...r.kernel.org,
linux-omap@...r.kernel.org,
linux-gpio@...r.kernel.org,
Danilo Krummrich <dakr@...nel.org>
Subject: [PATCH] gpio: omap: do not register driver in probe()
Commit 11a78b794496 ("ARM: OMAP: MPUIO wake updates") registers the
omap_mpuio_driver from omap_mpuio_init(), which is called from
omap_gpio_probe().
However, it neither makes sense to register drivers from probe()
callbacks of other drivers, nor does the driver core allow registering
drivers with a device lock already being held.
The latter was revealed by commit dc23806a7c47 ("driver core: enforce
device_lock for driver_match_device()") leading to a potential deadlock
condition described in [1].
Additionally, the omap_mpuio_driver is never unregistered from the
driver core, even if the module is unloaded.
Hence, register the omap_mpuio_driver from the module initcall and
unregister it in module_exit().
Link: https://lore.kernel.org/lkml/DFU7CEPUSG9A.1KKGVW4HIPMSH@kernel.org/ [1]
Fixes: dc23806a7c47 ("driver core: enforce device_lock for driver_match_device()")
Fixes: 11a78b794496 ("ARM: OMAP: MPUIO wake updates")
Signed-off-by: Danilo Krummrich <dakr@...nel.org>
---
In addition to this fix, the omap_mpuio_device (struct platform_device) should
not be declared as global static, especially when their backing memory can
become invalue due to the module being unloaded. Devices are reference counted
and should be allocated dynamically. This needs a separate fix.
Besides that, the whole construct seems a bit questionable and I'm not exactly
sure what should be achieved by registering the *same* static device everytime
probe() is called for the omap_gpio_driver.
However, for the purpose of avoiding the described potential deadlock in
combination with commit dc23806a7c47 ("driver core: enforce device_lock for
driver_match_device()"), this patch only addresses the driver registration
issue.
---
drivers/gpio/gpio-omap.c | 17 +++++++++++++----
1 file changed, 13 insertions(+), 4 deletions(-)
diff --git a/drivers/gpio/gpio-omap.c b/drivers/gpio/gpio-omap.c
index e136e81794df..8db71a2db9ff 100644
--- a/drivers/gpio/gpio-omap.c
+++ b/drivers/gpio/gpio-omap.c
@@ -800,9 +800,7 @@ static struct platform_device omap_mpuio_device = {
static inline void omap_mpuio_init(struct gpio_bank *bank)
{
platform_set_drvdata(&omap_mpuio_device, bank);
-
- if (platform_driver_register(&omap_mpuio_driver) == 0)
- (void) platform_device_register(&omap_mpuio_device);
+ (void)platform_device_register(&omap_mpuio_device);
}
/*---------------------------------------------------------------------*/
@@ -1575,13 +1573,24 @@ static struct platform_driver omap_gpio_driver = {
*/
static int __init omap_gpio_drv_reg(void)
{
- return platform_driver_register(&omap_gpio_driver);
+ int ret;
+
+ ret = platform_driver_register(&omap_mpuio_driver);
+ if (ret)
+ return ret;
+
+ ret = platform_driver_register(&omap_gpio_driver);
+ if (ret)
+ platform_driver_unregister(&omap_mpuio_driver);
+
+ return ret;
}
postcore_initcall(omap_gpio_drv_reg);
static void __exit omap_gpio_exit(void)
{
platform_driver_unregister(&omap_gpio_driver);
+ platform_driver_unregister(&omap_mpuio_driver);
}
module_exit(omap_gpio_exit);
base-commit: ed1ac3c977dd6b119405fa36dd41f7151bd5b4de
--
2.52.0
Powered by blists - more mailing lists