Latency problems related to hostap?
Mon Feb 23 05:44:17 PST 2004
I'm not even sure if I really have a problem and whether it is hostap related at all. Oh... What an entrance, but please go on reading... (-;
If I run a vi-Session from an associated station to my hostap-box (WindowsXP - Debian Linux) I get kind of "hiccups" after a random period of time.
Running a sequence of pings verifies this behaviour. After a random number of pings (between 20 and 50) having latency times of 2ms I get 3 to 4 missing answers and a response time of 2400ms for the next working ping.
Debian Linux testing/unstable, kernel-image 2.4.23 with the prebuild package of the hostap modules being available (HostAP vers. 0.1.2) for this kernel image.
233Mhz Pentium II MMX, 128MB RAM, running several services including a pppoe-Link, samba, mysql, apache... When using the WLAN connecting at maximum connection speed (large file transfer), the CPU utilization is at about 30% (SAMBA and kernel).
3Com 3CRDW696 PCI, Prism 2.5 chipset, upgraded to Firmware Versions 1.1.0 and 1.7.4.
Settings for the AP: channel 3, WEP 104bit key in place (restricted mode), Master mode, essid set explicitly, everything else left to standard values.
wlan0 and eth0 are brought up in bridging mode via the bridging tool brctl using the standard bridging modules delivered with the standard kernel mentioned above.
AMD Athlon XP 1600+, Windows patched with current patch releases, LevelOne WPC-0101 running on an Realtek Chipset with the most current driver, running in infrastructure mode, settings left to Windows via automatic mode (doing this by hand doesn't change the situation at all). Association etc. works like a charm, no problems at all, adress is given via DHCP to the interface. The interface is working in CAM (continuous access mode), all power saving settings are disabled.
The "hiccups" can be monitored in the Windows Driver Tool delivered with the card. There you have a status information field which is showing "associated" during normal operation. On occurance of the problem it is switching to "checking status" for the durance of the hiccup, afterwards it is showing the link qualitiy and the signal state as "very good" again.
In the kernel logs I don't find any associated entries like disassociation or changing of the transfer rates. You just don't find anything on the linux side except the hanging ping.
As I'm quite familiar to Linux (especially the Debian distribution), networking under Windows and Linux I am quite sure, that there's no misconfiguration in the "normal" settings there.
What is definitly totally new to me is everything "wireless" related. (I've been working with several internet sources to bring my setup up without having any serious trouble.) I don't know if this behaviour is kind of normal on a wireless media or if this could be caused by the client adapter. I even don't know which other log-possibilities I have to track this down.
So, questions to this list (if somebody would like to help me in any way):
What additional information could I get from which logging facility in order to track this down? How to configure these logging possibilities?
Which additional information could I give to you?
Does anyone know this problem or is it just normal behaviour on a wireless media? How to fix the problem if it is not "normal".
Might it be a problem with the settings of the hostap-modules?
I haven't found any related information whether in this lists archive nor on the internet in general of the manuals shipped with the driver. If there is something, please point me to it!
It is getting on my nervs having those hiccups during using a vi on a ssh session or during gaming.
Thank you for reading this huge amoung of text and excuse if I am completely wrong here...
sven.packebusch at web.de
Nachrichten, Musik und Spiele schnell und einfach per Quickstart im
WEB.DE Screensaver - Gratis downloaden: http://screensaver.web.de/?mc=021110
More information about the Hostap