Re: re(4) checksum offload problem

On Tue, Oct 16, 2007 at 12:22:23PM +0200, Jelte Jansen wrote:
 > -----BEGIN PGP SIGNED MESSAGE-----
 > Hash: SHA1
 > 
 > 
 > Hi,
 > 
 > i'm seeing bad IP/UDP checksums on packets on a realtek gigabit network
 > card. Turning offloading off (ifconfig -txcsum) seems to make the
 > problem disappear.
 > 
 > Attached are pcap files showing the bad checksums; the IP header is set
 > to 0x0000 and then TCP/UDP headers are wrong too. The machine in
 > question is on .99. These were taken on the host itself but i also see
 > them on the other machine.
 > 
 > pciconf:
 > re0_at_pci0:1:8:0: class=0x020000 card=0xee221113 chip=0x816910ec rev=0x10
 > hdr=0x00
 >     vendor     = 'Realtek Semiconductor'
 >     device     = 'RTL8110SB Single-Chip Gigabit LOM Ethernet Controller'
 >     class      = network
 >     subclass   = ethernet
 > 
 > Version:
 > $FreeBSD: src/sys/dev/re/if_re.c,v 1.95 2007/08/14 02:00:04 yongari E\
 > xp $
 > 
 > Jelte

Would you give me more information for your network configuration?
Which host use re(4) driver? Which host was used to capture the
traffic?

If you see bad checksum messages on hosts with Tx checksum enabled
driver it's quite normal to see bad checksummed frames. Hardware
will insert computed checksum into the frame after bpf listeners
see the frame.

-- 
Regards,
Pyun YongHyeon