lkml.org 
[lkml]   [1999]   [Oct]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subject[2.2.{12,13p15}] PPP uploads fizzle out and die (tcpdump info incl.)
Hi everyone,

I'm stuck. FTP downloads over my PPP link are fine, they pretty much saturate
the connection when the remote host can send that fast. Uploads are a different
matter: regardless of the remote host I'm uploading to, the xfer rate withers
away to sod-all and a huge gob of data remains resolutely stuck in the
send-Q. I tried switching on IRQ unmasking on my IDE drive, but it didn't make
a difference and I wouldn't really expect it to. After poking and proding all
the things I could think of, I eventually took tcpdump samples from the PPP
interface on my machine and the ethernet segment of a remote host I was trying
to upload to.

They're both restricted to showing an FTP data xfer; this first one is from
the machine with the PPP link to my ISP (loth):

09:37:05.942422 loth.1092 > moria-eth.4935: S 4218716301:4218716301(0) win 32648 <mss 1484,sackOK,timestamp 3176272 0,nop,wscale 0> (DF)
09:37:06.091796 moria-eth.4935 > loth.1092: S 4214918012:4214918012(0) ack 4218716302 win 32120 <mss 1460,nop,nop,sackOK,nop,wscale 0> (DF)
09:37:06.101840 loth.1092 > moria-eth.4935: . ack 1 win 32648 (DF)
09:37:06.283140 loth.1092 > moria-eth.4935: P 1:1461(1460) ack 1 win 32648 (DF)
09:37:06.283860 loth.1092 > moria-eth.4935: P 1461:2921(1460) ack 1 win 32648 (DF)
09:37:06.881815 moria-eth.4935 > loth.1092: . ack 1461 win 30660 (DF) [tos 0x10]
09:37:06.891875 loth.1092 > moria-eth.4935: P 2921:4381(1460) ack 1 win 32648 (DF)
09:37:06.892508 loth.1092 > moria-eth.4935: P 4381:5841(1460) ack 1 win 32648 (DF)
09:37:10.851986 loth.1092 > moria-eth.4935: P 1461:2921(1460) ack 1 win 32648 (DF)
09:37:19.992343 loth.1092 > moria-eth.4935: P 1461:2921(1460) ack 1 win 32648 (DF)
09:37:20.922373 moria-eth.4935 > loth.1092: . ack 2921 win 32120 (DF) [tos 0x10]
09:37:20.932449 loth.1092 > moria-eth.4935: P 2921:4381(1460) ack 1 win 32648 (DF)
09:37:20.933112 loth.1092 > moria-eth.4935: P 4381:5841(1460) ack 1 win 32648 (DF)
09:37:39.213121 loth.1092 > moria-eth.4935: P 2921:4381(1460) ack 1 win 32648 (DF)
09:37:39.803120 moria-eth.4935 > loth.1092: . ack 4381 win 32120 (DF) [tos 0x10]
09:37:39.813196 loth.1092 > moria-eth.4935: P 4381:5841(1460) ack 1 win 32648 (DF)
09:37:40.863160 moria-eth.4935 > loth.1092: . ack 5841 win 32120 (DF) [tos 0x10]
09:37:40.873253 loth.1092 > moria-eth.4935: P 5841:7301(1460) ack 1 win 32648 (DF)
09:37:40.873899 loth.1092 > moria-eth.4935: P 7301:8761(1460) ack 1 win 32648 (DF)
09:37:40.874522 loth.1092 > moria-eth.4935: P 8761:10221(1460) ack 1 win 32648 (DF)
09:38:17.434648 loth.1092 > moria-eth.4935: P 5841:7301(1460) ack 1 win 32648 (DF)
09:38:18.514663 moria-eth.4935 > loth.1092: . ack 7301 win 32120 (DF) [tos 0x10]
09:38:18.524746 loth.1092 > moria-eth.4935: P 7301:8761(1460) ack 1 win 32648 (DF)
09:38:18.525410 loth.1092 > moria-eth.4935: P 8761:10221(1460) ack 1 win 32648 (DF)
09:38:19.584706 moria-eth.4935 > loth.1092: . ack 8761 win 32120 (DF) [tos 0x10]
And this one is what actually appeared on the ethernet of the remote host
(moria):

09:37:15.257370 loth.1092 > moria.4935: S 4218716301:4218716301(0) win 32648 <mss 1484,sackOK,timestamp 3176272[|tcp]> (DF)
09:37:15.257630 moria.4935 > loth.1092: S 4214918012:4214918012(0) ack 4218716302 win 32120 <mss 1460,nop,nop,sackOK,nop,wscale 0> (DF)
09:37:15.417112 loth.1092 > moria.4935: . ack 1 win 32648 (DF)
09:37:16.053080 loth.1092 > moria.4935: P 1:1461(1460) ack 1 win 32648 (DF)
09:37:16.053369 moria.4935 > loth.1092: . ack 1461 win 30660 (DF) [tos 0x10]
09:37:16.641292 loth.1092 > moria.4935: P 1461:2921(1460) ack 1 win 32648 (DF)
09:37:29.611927 loth.1092 > moria.4935: P 1461:2921(1460) ack 1 win 32648 (DF)
09:37:30.102445 moria.4935 > loth.1092: . ack 2921 win 32120 (DF) [tos 0x10]
09:37:30.702576 loth.1092 > moria.4935: P 2921:4381(1460) ack 1 win 32648 (DF)
09:37:31.202241 moria.4935 > loth.1092: . ack 4381 win 32120 (DF) [tos 0x10]
09:37:48.982941 loth.1092 > moria.4935: P 2921:4381(1460) ack 1 win 32648 (DF)
09:37:48.983196 moria.4935 > loth.1092: . ack 4381 win 32120 (DF) [tos 0x10]
09:37:49.552405 loth.1092 > moria.4935: P 4381:5841(1460) ack 1 win 32648 (DF)
09:37:50.049171 moria.4935 > loth.1092: . ack 5841 win 32120 (DF) [tos 0x10]
09:38:27.205645 loth.1092 > moria.4935: P 5841:7301(1460) ack 1 win 32648 (DF)
09:38:27.703029 moria.4935 > loth.1092: . ack 7301 win 32120 (DF) [tos 0x10]
09:38:28.274169 loth.1092 > moria.4935: P 7301:8761(1460) ack 1 win 32648 (DF)
09:38:28.772836 moria.4935 > loth.1092: . ack 8761 win 32120 (DF) [tos 0x10]
(it looks like the clocks on the machines were out of step by ~10s, I don't
know why..; also there's no TCP timestamps because I'd turned them off to see
if that made a difference (it didn't), and forgot to turn them on again --
I can redo the dumps with them enabled if that would be useful)

I've tested with (amongst others) 2.2.12 and 2.2.13pre15 on loth; moria was
running 2.2.12.

So I guess either

i) the modem's getting overruns or similar; but wouldn't this result in
errors being counted on the ppp interface?, or

ii) something's throwing packets on the floor between my ISP and the remote
host, or

iii) the PPP driver is screwing up.

Does anyone have any ideas? Where should I go from here?

Thanks,
Steve

--
Plug-and-Play is really nice, unfortunately it only works 50% of the time.
To be specific the "Plug" almost always works. --unknown source

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:54    [from the cache]
©2003-2011 Jasper Spaans