Add missing function attributes and assert for refcount_acquire_if_not_zero() and refcount_acquire_if_not_last()

[• hselasky]

Summary

Make sure returned value is checked and assert a valid refcount.
While at it fix a print: Unsigned types cannot be negative.

MFC after: 1 week
Sponsored by: Mellanox Technologies

[• hselasky]

Make sure return value is checked.

[mjg]

there is already refcount_acquire_if_not_zero in the file

[• hselasky]

Update as per @mjg 's suggestion

[kib]

sys/sys/refcount.h

85 ↗

(On Diff #49312)

If acquire_if_nz gets the warning on unused result, then release_if_not_last should too ?

[mjg]

sys/sys/refcount.h

85 ↗

(On Diff #49312)

agreed, some for the assertion. Somewhat related, would be good to do a test build with the annotation added to refcount_release. There are probably very few places where it makes sense to do it without checking and those can be covered with an extra variant (or perhaps a (void) cast will take care of it).

[• hselasky]

I think the return value of refcount_release() should also be checked, but I see many users in the kernel which don't check it.

[• hselasky]

Minor refactor.

[kib]

So both you and mjg mentioned that there are a lot of consumers which do not check the result from release_if_not_last. How is it handled ?

sys/sys/refcount.h

107 ↗

(On Diff #49327)

"zero refcount %p"

[• hselasky]

Update assert message, suggested by @kib .

[• hselasky]

@kib : Only refcount_release() can be called w/o checking the returned result, so I didn't add an attribute for that function. The others should be fine from what I can see.

[kib]

sys/sys/refcount.h

107 ↗

(On Diff #49327)

old cannot be negative. refcounts are unsigned.

[• hselasky]

Address comments from @kib .