dio: return -ENOMEM when kzalloc() fails
PATCH 1/1 dio: return -ENOMEM when kzalloc() fails
Return -ENOMEM when kzalloc() fails in order to inform the caller of the failure.
Signed-off-by Salah Triki <[email protected]>
--- drivers/dio/dio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/dio/dio.c +1 -1
@@ -219,7 +219,7 @@ static int __init dio_init(void)/* Found a board, allocate it an entry in the list */ dev =kzalloc (sizeof (struct dio_dev ),GFP_KERNEL );if (!dev) -return 0 ; +return -ENOMEM ; dev->bus = &dio_bus; dev->dev.parent = &dio_bus.dev;
Signature · 1 line
2.25.1
Quoted 13 lines · Salah Triki
Return
-ENOMEMwhenkzalloc()fails in order to inform the caller of the failure.Signed-off-by Salah Triki <[email protected]>--- drivers/dio/dio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/dio/dio.c +0
-0
drivers/dio/dio.c static int __init dio_init(void) · line 222
Quoted 5 lines
drivers/dio/dio.c +1
-1@@ -219,7 +219,7 @@ static int __init dio_init(void)/* Found a board, allocate it an entry in the list */ dev =kzalloc (sizeof (struct dio_dev ),GFP_KERNEL );if (!dev) -return 0 ; +return -ENOMEM ;
Do you have this hardware to test with?
While this patch looks correct, it still is leaking lots of resources if this every happens. Can you fix this up "properly" so that all of the resources allocated at this point in time will be correctly freed?
Or, really, this is an impossible error path to hit, given that it is at boot time, so maybe it's just not worth it given that I doubt anyone has this hardware...
thanks,
greg k-h