On Mon, Mar 21, 2022 at 09:04:46AM +0000, David Laight wrote: > From: Dan Carpenter > > Sent: 21 March 2022 08:16 > > > > This version is better, thanks! > > > > Reviewed-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> > > > > What the GCC devs should have done is run their checker on real code and > > silence the common false positives. There are a few ways to do this. > > You could make ~ a special case. They already ignore truncated sign > > extension so they could equally well ignore truncation when all it's a > > 32 bits which are all set to 1. Probably the best answer is to do both. > > Or start using -(BIT_VALUE + 1) instead of ~BIT_VALUE :-) > You are a bad person. :P regards, dan carpenter
