[PATCH] acpi/ghes: Prevent sleeping with spinlock held

[PATCH] acpi/ghes: Prevent sleeping with spinlock held

[Ira Weiny]

Smatch caught that cxl_cper_post_event() is called with a spinlock held
or preemption disabled.[1]  There is no need for the callback to sleep.

Replace the RW semaphore with a RW lock.

A static call was considered but ARM does not select HAVE_STATIC_CALL
and in that case setting the function pointer uses a RW semaphore.

[1] https://lore.kernel.org/all/b963c490-2c13-4b79-bbe7-34c6568423c7@moroto.mountain/

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
---
 drivers/acpi/apei/ghes.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
index 7b7c605166e0..bdc0ec2813a3 100644
--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -39,6 +39,7 @@
 #include <linux/aer.h>
 #include <linux/nmi.h>
 #include <linux/sched/clock.h>
+#include <linux/spinlock.h>
 #include <linux/uuid.h>
 #include <linux/ras.h>
 #include <linux/task_work.h>
@@ -677,7 +678,7 @@ static void ghes_defer_non_standard_event(struct acpi_hest_generic_data *gdata,
 /*
  * Only a single callback can be registered for CXL CPER events.
  */
-static DECLARE_RWSEM(cxl_cper_rw_sem);
+static DEFINE_RWLOCK(cxl_callback_lock);
 static cxl_cper_callback cper_callback;
 
 /* CXL Event record UUIDs are formatted as GUIDs and reported in section type */
@@ -721,14 +722,14 @@ static void cxl_cper_post_event(enum cxl_event_type event_type,
 		return;
 	}
 
-	guard(rwsem_read)(&cxl_cper_rw_sem);
+	guard(read_lock_irqsave)(&cxl_callback_lock);
 	if (cper_callback)
 		cper_callback(event_type, rec);
 }
 
 int cxl_cper_register_callback(cxl_cper_callback callback)
 {
-	guard(rwsem_write)(&cxl_cper_rw_sem);
+	guard(write_lock_irq)(&cxl_callback_lock);
 	if (cper_callback)
 		return -EINVAL;
 	cper_callback = callback;
@@ -738,7 +739,7 @@ EXPORT_SYMBOL_NS_GPL(cxl_cper_register_callback, CXL);
 
 int cxl_cper_unregister_callback(cxl_cper_callback callback)
 {
-	guard(rwsem_write)(&cxl_cper_rw_sem);
+	guard(write_lock_irq)(&cxl_callback_lock);
 	if (callback != cper_callback)
 		return -EINVAL;
 	cper_callback = NULL;

---
base-commit: 861c0981648f5b64c86fd028ee622096eb7af05a
change-id: 20240201-cxl-cper-smatch-82b129498498

Best regards,
-- 
Ira Weiny <ira.weiny@intel.com>

Re: [PATCH] acpi/ghes: Prevent sleeping with spinlock held

[Dan Williams]

Ira Weiny wrote:
> Smatch caught that cxl_cper_post_event() is called with a spinlock held
> or preemption disabled.[1]  There is no need for the callback to sleep.
> 
> Replace the RW semaphore with a RW lock.
> 
> A static call was considered but ARM does not select HAVE_STATIC_CALL
> and in that case setting the function pointer uses a RW semaphore.
> 
> [1] https://lore.kernel.org/all/b963c490-2c13-4b79-bbe7-34c6568423c7@moroto.mountain/
> 
> Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
> Signed-off-by: Ira Weiny <ira.weiny@intel.com>
> ---
>  drivers/acpi/apei/ghes.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
> index 7b7c605166e0..bdc0ec2813a3 100644
> --- a/drivers/acpi/apei/ghes.c
> +++ b/drivers/acpi/apei/ghes.c
> @@ -39,6 +39,7 @@
>  #include <linux/aer.h>
>  #include <linux/nmi.h>
>  #include <linux/sched/clock.h>
> +#include <linux/spinlock.h>
>  #include <linux/uuid.h>
>  #include <linux/ras.h>
>  #include <linux/task_work.h>
> @@ -677,7 +678,7 @@ static void ghes_defer_non_standard_event(struct acpi_hest_generic_data *gdata,
>  /*
>   * Only a single callback can be registered for CXL CPER events.
>   */
> -static DECLARE_RWSEM(cxl_cper_rw_sem);
> +static DEFINE_RWLOCK(cxl_callback_lock);
>  static cxl_cper_callback cper_callback;
>  
>  /* CXL Event record UUIDs are formatted as GUIDs and reported in section type */
> @@ -721,14 +722,14 @@ static void cxl_cper_post_event(enum cxl_event_type event_type,
>  		return;
>  	}
>  
> -	guard(rwsem_read)(&cxl_cper_rw_sem);
> +	guard(read_lock_irqsave)(&cxl_callback_lock);
>  	if (cper_callback)
>  		cper_callback(event_type, rec);

This does not help because cxl_cper_event_call() takes a sleeping
device_lock().