Re: My last word on copy_to_user

Benny Amorsen (amorsen@sscnet.com)
24 Jan 1998 20:13:00 +0100


>>>>> "BH" == Bill Hawes <whawes@star.net> writes:

BH> Another point I'd like to mention again -- copy_xx_user can block,
BH> so it's not correct to call it with interrupts turned off, for
BH> example. So when making the changes to fix the return values, it's
BH> a good idea to check around for race conditions as well.

Can those race conditions be catched by static program analysis, and
if yes, how hard would it be?

Benny