Skip to contentSkip to conversation

driver core: drop __must_check from driver_for_each_device

PATCH In review Heiner Kallweit 2 participants 3 messages Aug 20, 2024 Driver core

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 driver core: drop __must_check from driver_for_each_device

There are several users of driver_for_each_device where the worker function doesn't actually have a return value. One example is __igb_notify_dca(). This results in dead code just to make __must_check happy. So drop this annotation.

---
 include/linux/device/driver.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
include/linux/device/driver.h +2 -2
index 1fc8b6878..aed40242f 100644
@@ -153,8 +153,8 @@ void driver_remove_file(const struct device_driver *driver,
 
 int driver_set_override(struct device *dev, const char **override,
 			const char *s, size_t len);
-int __must_check driver_for_each_device(struct device_driver *drv, struct device *start,
-					void *data, int (*fn)(struct device *dev, void *));
+int driver_for_each_device(struct device_driver *drv, struct device *start,
+			   void *data, int (*fn)(struct device *dev, void *));
 struct device *driver_find_device(const struct device_driver *drv,
 				  struct device *start, const void *data,
 				  int (*match)(struct device *dev, const void *data));
Signature · 1 line
2.46.0
#1 lore raw
Quoted 4 lines · Heiner Kallweit

There are several users of driver_for_each_device where the worker function doesn't actually have a return value. One example is __igb_notify_dca(). This results in dead code just to make __must_check happy. So drop this annotation.

No, that code looks correct, the only "odd" thing is that it's a notifier callback and so it requires a different translation of an error message.

Personally, attempting for a driver to iterate over all devices assigned to it just to print out a kernel log message seems like the big abuse here, why is any of that needed at all? Ah, that's a side affect of the dca api there. Ok, still looks correct, I don't see a problem.

Ignoring error values is not a good idea, let's not make it simple to do so.

thanks,

greg k-h

#2 lore raw
Quoted 6 lines · Greg Kroah-Hartman
Quoted 4 lines · Heiner Kallweit

There are several users of driver_for_each_device where the worker function doesn't actually have a return value. One example is __igb_notify_dca(). This results in dead code just to make __must_check happy. So drop this annotation.

Quoted 9 lines

No, that code looks correct, the only "odd" thing is that it's a notifier callback and so it requires a different translation of an error message.

Personally, attempting for a driver to iterate over all devices assigned to it just to print out a kernel log message seems like the big abuse here, why is any of that needed at all? Ah, that's a side affect of the dca api there. Ok, still looks correct, I don't see a problem.

__igb_notify_dca() always returns 0, for other worker functions of driver_for_each_device it's the same. So we have several cases where driver_for_each_device always returns 0. Then I think it doesn't make sense to urge the caller to check the return value.

Quoted 6 lines

Ignoring error values is not a good idea, let's not make it simple to do so.

thanks,

greg k-h

Heiner