Re: CVS commit: src/sys/netinet

[Thor Lancelot Simon <tls%panix.com@localhost>]

On Thu, Mar 24, 2011 at 02:57:28AM +0000, YAMAMOTO Takashi wrote:
> hi,
> 
> > KERNEL_LOCK.  However, there are a few other ways to cause output
> > which require protection:
> > 
> >     1) direct calls to tcp_output() in tcp_input()
> >     2) fast-forwarding code (ip_flow) -- protected elsewise
> >        against itself by the softnet lock.
> >     3) *Possibly* the ARP code.  I have currently persuaded
> >        myself that it is safe because of how it's called.
> >     4) Possibly the ICMP code.
> 
> what's the status of the rest?
> 
> it seems that ip_forward() at least still has the problem.

I still think the ARP code is OK.  I have not seen problems resulting
from the ICMP code but have not given it any more careful study.  We're
using the result of the KERNEL_LOCK change you reference above, pulled
up into our local netbsd-5 tree, in production and haven't seen any
obvious problems.  If you see anything that appears wrong along these
lines, though, by all means, please fix it!

Thor