| 1 |
On Wednesday 17 September 2008 22:32:21 Espen Hustad wrote: |
| 2 |
> On Wednesday 17 September 2008 22:03:52 Alan McKinnon wrote: |
| 3 |
> > Hi all, |
| 4 |
> > |
| 5 |
> > I have a Dell XPS M1530 with an Intel IWL3945 wireless card. On |
| 6 |
> > power-up/reboot it sometimes works, sometimes doesn't. I can find no |
| 7 |
> > pattern at all, success does not depend on the machine being hot/cold |
| 8 |
> > rebooted, powered up from shutdown, hibernate, suspend or any other usual |
| 9 |
> > thing. I don't use the wireless-rf kill switch at all. |
| 10 |
> > |
| 11 |
> > Config: |
| 12 |
> > alan@nazgul ~ $ zgrep IWL /proc/config.gz |
| 13 |
> > CONFIG_IWLWIFI=m |
| 14 |
> > CONFIG_IWLCORE=m |
| 15 |
> > # CONFIG_IWLWIFI_LEDS is not set |
| 16 |
> > CONFIG_IWLWIFI_RFKILL=y |
| 17 |
> > # CONFIG_IWL4965 is not set |
| 18 |
> > CONFIG_IWL3945=m |
| 19 |
> > CONFIG_IWL3945_SPECTRUM_MEASUREMENT=y |
| 20 |
> > CONFIG_IWL3945_LEDS=y |
| 21 |
> > # CONFIG_IWL3945_DEBUG is not set |
| 22 |
> > |
| 23 |
> > alan@nazgul /etc $ grep -i iwl /etc/mod*conf |
| 24 |
> > /etc/modprobe.conf:alias wlan0 iwl3945 |
| 25 |
> |
| 26 |
> Hi! |
| 27 |
> |
| 28 |
> Do you have mac80211 compiled as a module or in kernel? See |
| 29 |
> http://www.gentoo-wiki.com/Ipw3945#iwlwifi for explanation. I'm not sure |
| 30 |
> though if this is still a issue, or if it has been fixed. |
| 31 |
|
| 32 |
mac80211 is a module - I learned long ago to make as modules everything that |
| 33 |
can work as a module :-) Some more info I omitted from the original post: |
| 34 |
|
| 35 |
the driver I use is the in-kernel iwl3945 (part of iwlwifi) |
| 36 |
iwl3945-ucode is installed |
| 37 |
As far as I can tell, everything I should have done according to that page and |
| 38 |
to http://www.gentoo-wiki.com/Iwlwifi has been done. |
| 39 |
|
| 40 |
Working purely from a techie's gut feel here (i.e. no evidence whatsoever), |
| 41 |
this behaviour looks like what I might get from the following circumstances: |
| 42 |
|
| 43 |
- a register in the card is not being properly initialised by the firmware, |
| 44 |
sometimes a bit is 0 sometimes a 1 at random and the results are not |
| 45 |
determined |
| 46 |
- there's a module option that I missed when reading the source comments |
| 47 |
- a weird race condition at early boot, results are pot-luck |
| 48 |
- my hardware is busted |
| 49 |
|
| 50 |
But I've run out of ideas on how to falsify these theories :-( |
| 51 |
|
| 52 |
|
| 53 |
-- |
| 54 |
alan dot mckinnon at gmail dot com |
Archives