Messages in this thread | | | From | Eric.Schenk@dna ... | Subject | Re: slow tcp with isdn | Date | Tue, 29 Jul 1997 13:29:52 +0200 |
| |
Gerd Knorr <kraxel@goldbach.isdn.cs.tu-berlin.de> writes: >Syslog says nothing. ifconfig ippp0 says 0 rx/tx errors, so does >/proc/net/snmp for ip. But ip checksum is the ip header only, right?
Yes.
>I grep'ed net/ipv4/tcp*.c for a bad-checksum-printk(), but hav'nt found >one. Does the 2.0.30 kernel log and/or count bad tcp packets? If not, >is there any way to do this?
You'd have to add a printk or a log count to the code that checks the TCP checksum. Look at the routine tcp_rcv in net/ipv4/tcp_input.c to see where you need to add this. You should do the printk/count just before the jump to discard_it in the checksum testing code at the start of this routine.
>/proc/net/snmp has no somethingError entry for tcp. But there is a >RetransSegs entry. This one lists 5. Hmm... >Is this incoming? outgoing? both?
Outgoing. There is no way of knowing the actual incoming count, although I suppose we could count the number of "duplicate" packets we received.
-- Eric Schenk www: http://www.dna.lth.se/~erics Dept. of Comp. Sci., Lund University email: Eric.Schenk@dna.lth.se Box 118, S-221 00 LUND, Sweden fax: +46-46 13 10 21 ph: +46-46 222 96 38
| |