Re: [PATCH 03/16] lib/test_bitmap: don't test bitmap_set if nbits == 0

From: Rasmus Villemoes
Date: Tue Aug 09 2022 - 08:37:36 EST


On 18/07/2022 21.28, Yury Norov wrote:
> Don't test bitmap_set(bitmap, start, 0) as it's useless, most probably
> a sign of error in real code,

No it's not. The nbits can easily be the result of some computation that
ended up resulting in 0 being the right number to copy (or set, or
whatnot), and it's not unreasonable to _not_ check in the caller for
that special case, but rather rely on bitmap_set() to behave sanely - it
has perfectly well-defined semantics to "set 0 bits starting at @start".

The same way that memset() and memcpy() and memcmp() and countless other
functions have perfectly well-defined semantics with a length of 0, and
we don't add caller-side checks for those either.

NAK on this series.

Rasmus