Red Hat Bugzilla – Bug 440478
184.108.40.206-64.fc8 iwl3945 radio "on" lamp flashes irratically
Last modified: 2008-07-15 11:13:53 EDT
Description of problem:
The radio-on light for the iwl3945 wireless card flashes or stays on steadily
during operation of the latest kernel.
Version-Release number of selected component (if applicable):
Simply use the iwl3945 wireless kernel module to drive the wireless card.
Irritating flashing light.
Light steadily on when the radio is on, and light steadily off when radio is
Previous versions of iwl3945 did not turn the light on at all. Although that
was is not correct behaviour, it is much less annoying than having the light
Same on i686
...this is triggered by a new kernel, yet you report it against the firmware
package? Reassigning to the kernel maintainers. Please confirm if this is still
happening with the latest F-8 updated kernel.
(In reply to comment #2)
> ...this is triggered by a new kernel, yet you report it against the firmware
> package? Reassigning to the kernel maintainers. Please confirm if this is
> happening with the latest F-8 updated kernel.
Yes, it is still happening with 220.127.116.11-64.fc8. I have also tried
18.104.22.168-85.fc8 from Koji, and that is the same.
Flashing lights are used to indicate traffic and other status of the device --
this is perfectly normal.
If it's not a bug, then it's a feature regression. The problem remains.
For those of us who have the light right in front of the LCD it's a pita.
Please create an option for a steady light (e.g. reintroduce the previous bug
:-) ) when the radio is on.
(In reply to comment #4)
> Flashing lights are used to indicate traffic and other status of the
> this is perfectly normal.
Well someone should file a bug against Intel's Windows driver, because in that
OS, the light is on and steady when the radio is on and off otherwise :-)
Seriously, with respect, I don't think that is correct (do you have
documentation to confirm?). Yes, I know that many NICs use a light to
indicate status activity, but not in this case, for several reasons:
a) the flashing bears no relation that I can tell to network activity: it
b) it is not like an activity light on a wired NIC: it is much slower, with
flashes variably at between about 0,5 and 2 Hz, with roughly equal
c) the manufacturer's drivers' behaviour, mentioned above; and
d) no-one would have created something quite so annoying. This is a *bright*
blue flashing light right in the user's line-of-sight.
Oh, and the Intel-developed Linux ipw3945 driver also had the light steady on
Perhaps you are unaware that Intel develops this driver?
I did not know that!
Anyway, I have done some research. Looking at the wireless manual for my
computer (HP nx7400), which you can get at
"The wireless light indicates the overall power state of your wireless devices,
not the status of individual devices. If the wireless light is blue, at least
one wireless device is turned on. If the wireless light is turned off, all
wireless devices are turned off."
I think that is clear. Flashing is wrong: it is a bug.
Thanks for your attention.
You are welcome to your opinion about whether or not it should flash.
However, you will forgive me if I do not cede to HP the decision as to how
everyone's LEDs should work. :-)
I'll leave this open for now as a reminder to bug (no pun intended) Intel
about whether or not the current behavior is annoying. FWIW, I think at least
some flashing to indicate wireless activity is completely appropriate. I will
allow for the possibility that the current flashing parameters are
Dell has it as non-flashing:
*** Bug 449733 has been marked as a duplicate of this bug. ***
Well, I'm sorry this behavior is not to your taste. But I don't think we are
going to change it.
Oh, well. Time for some kernel hacking for me, then :-o