On Thu, Jun 01, 2023 at 06:52:03PM +0300, Dmitry Antipov wrote:

On 6/1/23 18:34, Dan Carpenter wrote:

quoted

This is a bug in Clang.

Is it confirmed by LLVM/clang developers? The compiler says that
<any unsigned 32-bit> can't be larger than <max unsigned 64-bit> / 8
(assuming LP64). Why this is wrong?

It's a false positive because the test is obviously intended for 32-bit
longs.

regards,
dan carpenter