From owner-freebsd-hackers Wed Sep 9 04:39:43 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA28170 for freebsd-hackers-outgoing; Wed, 9 Sep 1998 04:39:43 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from cheops.anu.edu.au (cheops.anu.edu.au [150.203.224.24]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id EAA28165 for ; Wed, 9 Sep 1998 04:39:40 -0700 (PDT) (envelope-from avalon@coombs.anu.edu.au) Message-Id: <199809091139.EAA28165@hub.freebsd.org> Received: by cheops.anu.edu.au (1.37.109.16/16.2) id AA258181169; Wed, 9 Sep 1998 21:39:29 +1000 From: Darren Reed Subject: 2.2.6 zp/tcp fuckups. To: hackers@FreeBSD.ORG Date: Wed, 9 Sep 1998 21:39:29 +1000 (EST) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG okay, a few of you will remember by griping about tcp performance on a laptop with the 3c589d card. well, now I get to experience it all again. right now I have: * a box which can't ping, error message: ping:sendto:No buffer space available * can't be pinged * froze during an ftp * has two non-closed tcp connections, one with a Send-Q of 17520 (CLOSING) and the other has a Send-Q of 106 (LAST_ACK) * netstat -m tells me I have 112 mbufs in use, 34k of which is allocated to network (94%), 64 are for data, 45 for headers, 2 for pcb's, 1 for socket names/addresses and 9/10 mbuf clusers in use. It also says there have been no denials or delays and 0 calls to the drain routines. As far as networking is concerned, the box is fucked. The only hope I have (and what I am doing) to make it work is to reboot. After the reboot... Attempting to ftp to it from solaris resulted in this: 200 PORT command successful. 150 Opening BINARY mode data connection for 'p1' (100663296 bytes). netin: Connection reset by peer 421 Service not available, remote server has closed connection Lost connection No control connection for command: Error 0 No control connection for command: Error 0 ftp> (no, I didn't hit ^C or run anything else, just let it "happen"). if I run "tcpdump -p" on a screen, it seems work okay. btw, I'm quite prepared to blame the zp driver - it fails to properly detect the device when I boot up FreeBSD after running Windows 95. Requires a cold boot to make the card detectable (to FreeBSD) again. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message