Re: Minor net/core/sock.c security issue?

[David S. Miller]

Alexey Kuznetsov writes:
 > > 1) have standard inline functions with names that suggest the
 > >    signedness, much like Rusty's netfilter macros.
 > 
 > min/max are macros. I do not know how to make a valid inline
 > for it: cast to long has problems with unsigned longs, cast to unsigned long
 > have the same problems with signedness.

For the time being I've just killed that bogus min define
from sock.c and also open-coded the min/max usage in the
rest of sock.c

This solves the original report, but later I'd like to do
something more satisfactory here.

I mean, grep for "define [min|max]" in just the networking
sources right now, yuck!

Later,
David S. Miller
[e-mail]