Re: ppp bug in 2.1 (?)

David Burrows (snadge@gemcorp.com.au)
Sat, 18 Apr 1998 17:34:48 +1000 (EST)


On Fri, 17 Apr 1998, David S. Miller wrote:

> Date: Sat, 18 Apr 1998 15:57:30 +1000 (EST)
> From: David Burrows <snadge@gemcorp.com.au>
>
> However, I am _still_ getting frame errors on receive which I don't get in
> the pre-2.0.34-6 kernel I was previously using.. ifconfig output follows
> (also of note I get 1 receive error immediately after logging in)
>
> ppp0 Link encap:Point-to-Point Protocol
> inet addr:203.108.221.48 P-t-P:203.108.221.15 Mask:255.255.255.255
> UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1500 Metric:1
> RX packets:13179 errors:17 dropped:0 overruns:0 frame:16
> TX packets:20223 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0
> Memory:c04e801c-c04e8be8
>
> Machine is rh5+updates, AMD 386dx40/8mb ram.. 16550 uarts, external modem.
>
> Do you have an updated ifconfig from the latest nettools? The format
> of /proc/net/dev has changed, cat that and see if the situation is the
> same.

Yes, I have patched nettools which fixed the TX errors problems.. So I
_think_ what nettools is reporting is pretty accurate..

ppp0 Link encap:Point-to-Point Protocol
inet addr:203.108.221.48 P-t-P:203.108.221.15 Mask:255.255.255.255
UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1500 Metric:1
RX packets:17454 errors:22 dropped:0 overruns:0 frame:21
TX packets:27470 errors:0 dropped:0 overruns:0 carrier:0
collisions:0
Memory:c04e801c-c04e8be8

That is what it says currently (same connection), the errors seem to be
rather infrequent.. However check this out:

Socket destroy delayed (r=0 w=2128)
Socket destroy delayed (r=0 w=2128)
Socket destroy delayed (r=0 w=1984)
Socket destroy delayed (r=0 w=2224)
Socket destroy delayed (r=0 w=1984)

(thats what i found at the end of my dmesg)

Oh, and FWIW I'm using pppd 2.3.4 (which a bit of butchering to get
working with glibc), however I got the same symptoms with 2.3.3...

Regards,

Dave.

PS.. Do you know anything about the VJ header compression causing tcp
checksum errors? My temporary workaround was to disable it (which seemed
to reduce the frame errors too).

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu