> Perhaps somebody familiar with the used
> algorhitms can interpret the pattern. I am
> willing to test any experimental patches.
A followup:
tcpdump for border write sizes (localhost
trimmed to have sane line width.
Write size 1480, stall:
...
13:53:20.339402 lh.3333 > lh.1284: P 25161:26641(1480) ack 1 win 31072 (DF)
13:53:20.339402 lh.3333 > lh.1284: P 25161:26641(1480) ack 1 win 31072 (DF)
13:53:20.339450 lh.1284 > lh.3333: . ack 26641 win 4432 (DF)
13:53:20.339450 lh.1284 > lh.3333: . ack 26641 win 4432 (DF)
13:53:20.339864 lh.3333 > lh.1284: P 26641:28121(1480) ack 1 win 31072 (DF)
13:53:20.339864 lh.3333 > lh.1284: P 26641:28121(1480) ack 1 win 31072 (DF)
13:53:20.339903 lh.1284 > lh.3333: . ack 28121 win 2952 (DF)
13:53:20.339903 lh.1284 > lh.3333: . ack 28121 win 2952 (DF)
13:53:20.340299 lh.3333 > lh.1284: P 28121:29601(1480) ack 1 win 31072 (DF)
13:53:20.340299 lh.3333 > lh.1284: P 28121:29601(1480) ack 1 win 31072 (DF)
13:53:20.340337 lh.1284 > lh.3333: . ack 29601 win 1472 (DF)
13:53:20.340337 lh.1284 > lh.3333: . ack 29601 win 1472 (DF)
=== stall ===
The same part of the dump, Write size 1479, no stall:
...
13:55:34.546580 lh.3333 > lh.1286: P 28102:29581(1479) ack 1 win 31072 (DF)
13:55:34.546580 lh.3333 > lh.1286: P 28102:29581(1479) ack 1 win 31072 (DF)
13:55:34.546617 lh.1286 > lh.3333: . ack 29581 win 1492 (DF)
13:55:34.546617 lh.1286 > lh.3333: . ack 29581 win 1492 (DF)
13:55:34.547013 lh.3333 > lh.1286: P 29581:31060(1479) ack 1 win 31072 (DF)
13:55:34.547013 lh.3333 > lh.1286: P 29581:31060(1479) ack 1 win 31072 (DF)
13:55:34.547050 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.547050 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.548690 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.548690 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.549678 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.549678 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.550200 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.550200 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.550705 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.550705 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.551212 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.551212 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.551928 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.551928 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.552438 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.552438 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.552940 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.552940 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.553443 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.553443 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.553943 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.553943 lh.1286 > lh.3333: . ack 31060 win 13 (DF)
13:55:34.554479 lh.1286 > lh.3333: . ack 31060 win 3884 (DF)
13:55:34.554479 lh.1286 > lh.3333: . ack 31060 win 3884 (DF)
13:55:34.554520 lh.3333 > lh.1286: P 31060:34944(3884) ack 1 win 31072 (DF)
13:55:34.554520 lh.3333 > lh.1286: P 31060:34944(3884) ack 1 win 31072 (DF)
13:55:34.555265 lh.1286 > lh.3333: . ack 34944 win 0 (DF)
13:55:34.555265 lh.1286 > lh.3333: . ack 34944 win 0 (DF)
13:55:34.560756 lh.1286 > lh.3333: . ack 34944 win 3884 (DF)
13:55:34.560756 lh.1286 > lh.3333: . ack 34944 win 3884 (DF)
13:55:34.560861 lh.3333 > lh.1286: P 34944:38828(3884) ack 1 win 31072 (DF)
13:55:34.560861 lh.3333 > lh.1286: P 34944:38828(3884) ack 1 win 31072 (DF)
13:55:34.561715 lh.1286 > lh.3333: . ack 38828 win 3884 (DF)
13:55:34.561715 lh.1286 > lh.3333: . ack 38828 win 3884 (DF)
...
BTW, the next parameter the stall depends upon is MTU,
so the people suffering from the problem might try
to /sbin/ifconfig lo <something> to shift the problem
out of the way :-)
Regards
-- Stano
- 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/