Re: Failed to scan network with 2200BG
"Cahill, Ben M" <[email protected]>
| Newsgroups | gmane.linux.drivers.ipw2100.devel |
|---|---|
| Message-ID | <4220499A1B034C4FA93B547BA01E1FF006E110@orsmsx413.amr.corp.intel.com> |
> > I expect to find a network on the 11th channel. I use another > 2200BG on a PC under Windows that find a network on this Channel. Thanks. > > In these traces, no NMI was transmitted, but sometimes it happens. > > Another question: I use a PowerPC platform with a little > endian bus on the PCI. Do I have a realignment of dma data? I would guess that Linux infrastructure would do the right thing here, but that's not an area I'm really familiar with. Since the commands are issued to the 2200 via DMA, and the scan seems to be working (at least going through the motions successfully, as the driver sees it), I'd guess that DMA is working okay. Could you try using debug=0x01043fff? This adds the DEBUG_RX output to the log. It should show up something if you're receiving any packets at all. If you want to try some hacking to check the sanity of the data, as the driver sees it, try adding an IPW_DEBUG_RX() or two within ipw_rx(). ipw_rx() is the first function that starts figuring out what to do with a frame notification from the 2200 hardware. You should be able to examine the ieee80211 header there. Within ipw_rx(), search for "ieee80211_hdr_4addr" and see how the "header" variable gets set up. Try IPW_DEBUG_RX for a few values within the ieee80211 header, preferably before ipw_rx() gets into parsing stuff (and perhaps getting confused if data is swapped around!). Don't confuse ieee80211 header (part of data payload) with the packet->header (added by 2200 firmware to tell driver what kind of info it's sending to driver, etc.). For "extra credit", take a look at ipw_report_frame() in the ipw3945 driver, and see if you could port that over to 2200. I put ipw_report_frame() in the 3945 driver a while back, to help debug scanning issues there ... never got to put it into 2200. BTW, if the AP signal is reasonably strong, you might receive packets when tuned to channels other than 11. This is normal behavior for B/G band, because of channel overlap (20 MHz channel width, 5 MHz channel separation). -- Ben -- > > Thanks > > Pascal FIFRE ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642