Messages in this thread | | | Subject | double lock on socket | Date | Mon, 4 Mar 1996 14:12:45 -0600 (CST) | From | (Mike Castle) |
| |
Hmmmm... I've gotten double lock on socket at ..... twice in a few minutes with kernel 1.3.70. In function tcp_sendmesg().
It's a mostly ELF system, both a.out and elf compiled in (not tried a.out as module yet), but everything else is a module.
Using kerneld 1.3.57 and dip with cslip to connect to campus dialups. kerneld runs dip to give ondemand dial in. I use IPMASQ on machine with modem (ylum) to allow other machine (thune) to access network.
Sitting at thune's console. Had one vt logged in as root and rshed to ylum. On another vt, had user logged in running screen and multiple sessions to the net (telnets, lynx, ftp). Occasionally I'd switch to the vt with root rshed to ylum, and as soon as I'd hit a key, I'd get the double lock message.
The message is generated on thune.
Hmmm... upon looking in /var/log/messages, I see the following:
Mar 4 12:06:54 thune linux: 5>double lock on socket at 00134a94 Mar 4 12:52:09 thune linux: 5>double lock on socket at 0013f2d7 Mar 4 13:35:39 thune linux: 5>double lock on socket at 0013f2d7
The last two are tcp_sendmesg() and were noticed in the scenario above. I don't remember seeing the other message, though may have been scrolled off. But the conditions where the same. 00134a94 corresponds to skb_recv_datagram().
Machine seems to work fine. I'll try to pound on things when I get 71 compiled and see if I can generate a set of circumstances that will always create it.
mrc -- Mike Castle .-=NEXUS=-. Life is like a clock: You can work constantly mcastle@cs.umr.edu and be right all the time, or not work at all mcastle@umr.edu and be right at least twice a day. -- mrc We are all of us living in the shadow of Manhattan. -- Watchmen
| |