On Tue, Jul 23, 2002 at 12:40:17PM -0700, Jonathan Stone wrote:
> In message <20020723193331.GA25338@rek.tjls.com>Thor Lancelot Simon writes
> >On Tue, Jul 23, 2002 at 12:19:36PM -0700, Jonathan Stone wrote:
> 
> >Every 3Com NIC shipped in about the past three years has had checksum
> >offload, and the Win32 drivers use it, too.  Not to mention the onboard
> >gigabit interfaces on all of the x86 server boxes shipped by major vendors
> >these days -- I think your "until" has already happened.
> 
> I think you should check with Jason and Allen. Or look at the commit
> logs of the other *bsd's and see where they've had to disable checksum
> offload because it doesn't work.

I'm entirely aware that there are several popular gigabit chipsets with
broken checksum offload support.  What does that have to do with your
statement:

| My guess is this wont change much (if at all) until mass-market OSes
| (windows) start to really use IP/TCP/UDP assist.

and my observation that, in fact, Windows will happily use hardware
checksum offload (they're rather fond of TCP segmentation offload, too)
and that huge numbers of Windows machines are happily using it today?

Broken hardware is broken hardware.  Obviously, if you have broken
checksum offload, it's the same as having none at all: don't use it.

Incidentally, I'm not aware of any revision of the 3c90x that has broken
checksum offload (though I know of many _other_ chips that do).

-- 
 Thor Lancelot Simon	                                      tls@rek.tjls.com
   But as he knew no bad language, he had called him all the names of common
 objects that he could think of, and had screamed: "You lamp!  You towel!  You
 plate!" and so on.              --Sigmund Freud