Comment 16 for bug 291760

Im running Ubuntu 8.10 (Intrepid) on as Acer Aspire 5315-2122. The wireless networking has been an uphill battle. At first it didn't work at all. Googling around led me to forcing it to load ath5k. That got the wireless to sort of work, but whenever the load was heavy on the network, it would stumble and gasp. The broswer seemed to spend an awful lot of time resolving domain names, so I added the dnsmasq package to the system in hopes that a modest cache would help. I think it helped a little, but network performance was still spotty. So I started to watch /var/log/syslog for clues. I disabled IPv6. I tried various network configuration tweeks in sysctl.conf that the web suggested would improve my TCP throughput. I saw lots of messages like this:

Dec 29 12:38:39 drew-laptop kernel: [ 138.432010] APIC error on CPU0: 40(40)
Dec 29 12:38:42 drew-laptop kernel: [ 141.020029] APIC error on CPU0: 40(40)
Dec 29 12:39:08 drew-laptop kernel: [ 167.312006] APIC error on CPU0: 40(40)
Dec 29 12:39:17 drew-laptop kernel: [ 176.085384] APIC error on CPU0: 40(40)
Dec 29 12:39:17 drew-laptop kernel: [ 176.752011] APIC error on CPU0: 40(40)

Upgrading the BIOS from 1.33 to 1.34 got rid of those messages, but the playback still would stall every few seconds when trying to play a youtube video, I know the problem was the network, because if I download the video to the local hard drive, it plays back just fine on this PC.

Taking riskier steps, I found that adding intrepid-proposed to my /etc/apt/sources.list, I could see many more updates. Updating the kernel and whatnot from 2.6.27-9 to 2.6.27-11 made most of the problems go away. youtube now plays OK. But my monitoring of mentions of wlan0 in syslog was still in place and turned up this:

Jan 25 05:25:30 drew-laptop dhclient: DHCPREQUEST of 192.168.1.102 on wlan0 to 192.168.1.1 port 67
Jan 25 05:25:30 drew-laptop NetworkManager: <info> DHCP: device wlan0 state changed bound -> renew
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: Withdrawing address record for 192.168.1.102 on wlan0.
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: Leaving mDNS multicast group on interface wlan0.IPv4 with address 192.168.1.102.
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: Interface wlan0.IPv4 no longer relevant for mDNS.
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: Joining mDNS multicast group on interface wlan0.IPv4 with address 192.168.1.102.
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: New relevant interface wlan0.IPv4 for mDNS.
Jan 25 05:25:30 drew-laptop avahi-daemon[4911]: Registering new address record for 192.168.1.102 on wlan0.IPv4.
Jan 25 05:25:31 drew-laptop NetworkManager: <info> Policy set 'Newell' (wlan0) as default for routing and DNS.

That inspired me to take an unfiltered look at syslog. Before the update to -11, the stalls were associated with messages like this:

Jan 9 07:51:10 drew-laptop kernel: [ 224.900568] wlan0: No ProbeResp from current AP 00:21:29:70:26:31 - assume out of range
Jan 9 07:51:11 drew-laptop kernel: [ 226.120335] wlan0: RX ReassocResp from 00:21:29:70:26:31 (capab=0x11 status=0 aid=1)
Jan 9 07:51:11 drew-laptop kernel: [ 226.146448] wlan0: RX ReassocResp from 00:21:29:70:26:31 (capab=0x11 status=0 aid=1)
Jan 9 07:51:11 drew-laptop kernel: [ 226.164320] wlan0: RX ReassocResp from 00:21:29:70:26:31 (capab=0x11 status=0 aid=1)
Jan 9 07:53:18 drew-laptop kernel: [ 352.920064] wlan0: No ProbeResp from current AP 00:21:29:70:26:31 - assume out of range
Jan 9 07:53:19 drew-laptop kernel: [ 354.075891] wlan0: RX ReassocResp from 00:21:29:70:26:31 (capab=0x11 status=0 aid=1)
Jan 9 07:53:43 drew-laptop kernel: [ 378.076152] wlan0: No ProbeResp from current AP 00:21:29:70:26:31 - assume out of range
Jan 9 07:53:44 drew-laptop kernel: [ 379.206468] wlan0: RX ReassocResp from 00:21:29:70:26:31 (capab=0x11 status=0 aid=1)

Those are gone but I'm seeing this:

Jan 25 05:35:55 drew-laptop NetworkManager: <debug> [1232879755.535897] periodic_update(): Roamed from BSSID (none) ((none)) to 00:21:29:70:26:31 (Newell)
Jan 25 05:37:49 drew-laptop NetworkManager: <debug> [1232879869.625696] periodic_update(): Roamed from BSSID 00:21:29:70:26:31 (Newell) to (none) ((none))
Jan 25 05:37:55 drew-laptop NetworkManager: <debug> [1232879875.627883] periodic_update(): Roamed from BSSID (none) ((none)) to 00:21:29:70:26:31 (Newell)
Jan 25 05:38:57 drew-laptop tcpspy[5438]: disconnect: user drew, local 192.168.1.102:40814, remote 205.188.13.24:aol
Jan 25 05:39:49 drew-laptop NetworkManager: <debug> [1232879989.721521] periodic_update(): Roamed from BSSID 00:21:29:70:26:31 (Newell) to (none) ((none))
Jan 25 05:39:55 drew-laptop NetworkManager: <debug> [1232879995.723907] periodic_update(): Roamed from BSSID (none) ((none)) to 00:21:29:70:26:31 (Newell)
Jan 25 05:40:01 drew-laptop /USR/SBIN/CRON[13592]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Jan 25 05:41:49 drew-laptop NetworkManager: <debug> [1232880109.809445] periodic_update(): Roamed from BSSID 00:21:29:70:26:31 (Newell) to (none) ((none))
Jan 25 05:41:55 drew-laptop NetworkManager: <debug> [1232880115.815902] periodic_update(): Roamed from BSSID (none) ((none)) to 00:21:29:70:26:31 (Newell)
Jan 25 05:43:49 drew-laptop NetworkManager: <debug> [1232880229.900669] periodic_update(): Roamed from BSSID 00:21:29:70:26:31 (Newell) to (none) ((none))
Jan 25 05:43:55 drew-laptop NetworkManager: <debug> [1232880235.903905] periodic_update(): Roamed from BSSID (none) ((none)) to 00:21:29:70:26:31 (Newell)

so I'm guessing my network still isn't living up to its full potential.

I'm thinking of maybe trying wicd in place of network manager. Comments on that?

But first I think I'll let update manager install the rest of the intrepid-proposed updates it sees and I'll see if that matters.

If you have other suggestions or want more info, please let me know.

Drew