iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

[Pavel Machek]

Hi!

I tried using ad-hoc mode (against t-mobile G1) on -rc0, and was
unsuccessful -- never got dhcp. I had to reboot back to
2.6.30... Should I try it few more times or is something broken?

(Also wifi LED stopped working on x60, but that's older issue).

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

[reinette chatre]

Hi Pavel,

On Fri, 2009-06-19 at 15:24 -0700, Pavel Machek wrote:
> Hi!
> 
> I tried using ad-hoc mode (against t-mobile G1) on -rc0, and was
> unsuccessful -- never got dhcp. I had to reboot back to
> 2.6.30... Should I try it few more times or is something broken?
> 
> (Also wifi LED stopped working on x60, but that's older issue).
> 

Are you testing Linus's repo here? Should we not maybe wait with testing
until -rc1 is released?

Reinette

Re: iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

[Pavel Machek]

On Mon 2009-06-22 10:39:40, reinette chatre wrote:
> Hi Pavel,
> 
> On Fri, 2009-06-19 at 15:24 -0700, Pavel Machek wrote:
> > Hi!
> > 
> > I tried using ad-hoc mode (against t-mobile G1) on -rc0, and was
> > unsuccessful -- never got dhcp. I had to reboot back to
> > 2.6.30... Should I try it few more times or is something broken?
> > 
> > (Also wifi LED stopped working on x60, but that's older issue).
> > 
> 
> Are you testing Linus's repo here? Should we not maybe wait with testing
> until -rc1 is released?

WTF? Yes, I am testing Linus' tree (hence 2.6.31-rc0). I thought
2.6. stable series is expected to work at all times, and it is bug if
it does not (otherwise git bisect is useless). Did you intentionally
break the tree or what? Or do you have fixes pending?
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

[reinette chatre]

On Mon, 2009-06-22 at 10:40 -0700, Pavel Machek wrote:
> On Mon 2009-06-22 10:39:40, reinette chatre wrote:
> > Hi Pavel,
> > 
> > On Fri, 2009-06-19 at 15:24 -0700, Pavel Machek wrote:
> > > Hi!
> > > 
> > > I tried using ad-hoc mode (against t-mobile G1) on -rc0, and was
> > > unsuccessful -- never got dhcp. I had to reboot back to
> > > 2.6.30... Should I try it few more times or is something broken?
> > > 
> > > (Also wifi LED stopped working on x60, but that's older issue).
> > > 
> > 
> > Are you testing Linus's repo here? Should we not maybe wait with testing
> > until -rc1 is released?
> 
> WTF? Yes, I am testing Linus' tree (hence 2.6.31-rc0). I thought
> 2.6. stable series is expected to work at all times, and it is bug if
> it does not (otherwise git bisect is useless). Did you intentionally
> break the tree or what?

No. We include adhoc testing in our weekly tests on different hardware
(3945) and have not seen any issues. We have not yet done any official
testing on 2.6.31 as we usually do so only after -rc1 is released.

>  Or do you have fixes pending?

All our fixes targeting 2.6.31 have been pushed. We do not have anything
else pending.

This is our first report of an issue in 2.6.31 and we will take a look.

Reinette



						

Re: iwl3945 in 2.6.31-rc0: ad-hoc mode broken?

[reinette chatre]

Hi Pavel,

On Fri, 2009-06-19 at 15:24 -0700, Pavel Machek wrote:
> I tried using ad-hoc mode (against t-mobile G1) on -rc0, and was
> unsuccessful -- never got dhcp. I had to reboot back to
> 2.6.30... Should I try it few more times or is something broken?

I have tried ad-hoc mode in 2.6.31-rc1
(52989765629e7d182b4f146050ebba0abf2cb0b7) and did not have issues. I
tested ad-hoc between 3945 (iwl3945) and 5100 (iwlagn), testing that
both are able to create the IBSS network and both are able to join
another IBSS network.

Do you get any errors? Is it just the dhcp that fails (does association
succeed?) mac80211 does have extra debugging for ad-hoc that can be
enabled with CONFIG_MAC80211_IBSS_DEBUG=y. Could you please try that?

> 
> (Also wifi LED stopped working on x60, but that's older issue).

Is there a bug report for this?

Reinette

iwl3945: fix LED config option

[Pavel Machek]

On Thu 2009-07-09 08:39:46, reinette chatre wrote:
> On Thu, 2009-07-09 at 02:38 -0700, Pavel Machek wrote:
> > > > 
> > > > (Also wifi LED stopped working on x60, but that's older issue).
> > > 
> > > Is there a bug report for this?
> > 
> > Well... I thought that was known issue?
> > 
> > Anyway, wifi LED used to work on x60 before driver rewrite. Now it is
> > broken. I thought it is known/missing functionality? ...but I'd like
> > to get it back :-).
> 
> Sorry ... but I have to ask ... did you compile with
> CONFIG_IWLWIFI_LEDS ? If you still have issues - please do submit a bug
> report.

Aha, this explains it:

  Enable LED support in iwlagn and iwl3945 drivers (IWLWIFI_LEDS) [N/y] (NEW) ?

Sorry, no help available for this option yet.

...maybe LEDS should be default Y?
									Pavel

---

IWLWIFI_LEDS option should certainly have help comment, and should
default to y. Actually I wonder if it should be user configurable at
all, is there good reason?

Signed-off-by: Pavel Machek <pavel@ucw.cz>

diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
index e092af0..95d9aa5 100644
--- a/drivers/net/wireless/iwlwifi/Kconfig
+++ b/drivers/net/wireless/iwlwifi/Kconfig
@@ -9,6 +9,10 @@ config IWLWIFI
 config IWLWIFI_LEDS
 	bool "Enable LED support in iwlagn and iwl3945 drivers"
 	depends on IWLWIFI
+	default y
+	---help---
+	  Useful to get Wifi LED working at least on Thinkpad notebooks.
+
 
 config IWLWIFI_SPECTRUM_MEASUREMENT
 	bool "Enable Spectrum Measurement in iwlagn driver"


-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: iwl3945: fix LED config option

[reinette chatre]

Hi Pavel,

On Sat, 2009-07-18 at 03:20 -0700, Pavel Machek wrote:
> On Thu 2009-07-09 08:39:46, reinette chatre wrote:
> > On Thu, 2009-07-09 at 02:38 -0700, Pavel Machek wrote:
> > > > > 
> > > > > (Also wifi LED stopped working on x60, but that's older issue).
> > > > 
> > > > Is there a bug report for this?
> > > 
> > > Well... I thought that was known issue?
> > > 
> > > Anyway, wifi LED used to work on x60 before driver rewrite. Now it is
> > > broken. I thought it is known/missing functionality? ...but I'd like
> > > to get it back :-).
> > 
> > Sorry ... but I have to ask ... did you compile with
> > CONFIG_IWLWIFI_LEDS ? If you still have issues - please do submit a bug
> > report.
> 
> Aha, this explains it:
> 
>   Enable LED support in iwlagn and iwl3945 drivers (IWLWIFI_LEDS) [N/y] (NEW) ?
> 
> Sorry, no help available for this option yet.
> 
> ...maybe LEDS should be default Y?
> 									Pavel
> 
> ---
> 
> IWLWIFI_LEDS option should certainly have help comment, 

Currently is reads (as you also have below) "Enable LED support in
iwlagn and iwl3945 drivers". This seems clear to me. 

> and should
> default to y. 

This makes sense.

> Actually I wonder if it should be user configurable at
> all, is there good reason?

Some users find the LED annoying and like to have a way to disable it.

> 
> Signed-off-by: Pavel Machek <pavel@ucw.cz>
> 
> diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
> index e092af0..95d9aa5 100644
> --- a/drivers/net/wireless/iwlwifi/Kconfig
> +++ b/drivers/net/wireless/iwlwifi/Kconfig
> @@ -9,6 +9,10 @@ config IWLWIFI
>  config IWLWIFI_LEDS
>  	bool "Enable LED support in iwlagn and iwl3945 drivers"
>  	depends on IWLWIFI
> +	default y

This is reasonable.

> +	---help---
> +	  Useful to get Wifi LED working at least on Thinkpad notebooks.

I do not think this says more than what we already have. I actually
think noting something about a particular platform will increase
confusion (as in "so where is the option if I have an acer laptop?")
because this option is needed to get LED support on all platforms. I do
think what is currently written is sufficient, but if you insist on
having some text there I think it should be simpler. Something like
"Select this if you want LED support."


Reinette

Re: iwl3945: fix LED config option

[Pavel Machek]

> > diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
> > index e092af0..95d9aa5 100644
> > --- a/drivers/net/wireless/iwlwifi/Kconfig
> > +++ b/drivers/net/wireless/iwlwifi/Kconfig
> > @@ -9,6 +9,10 @@ config IWLWIFI
> >  config IWLWIFI_LEDS
> >  	bool "Enable LED support in iwlagn and iwl3945 drivers"
> >  	depends on IWLWIFI
> > +	default y
> 
> This is reasonable.
> 
> > +	---help---
> > +	  Useful to get Wifi LED working at least on Thinkpad notebooks.
> 
> I do not think this says more than what we already have. I actually
> think noting something about a particular platform will increase
> confusion (as in "so where is the option if I have an acer laptop?")
> because this option is needed to get LED support on all platforms. I do
> think what is currently written is sufficient, but if you insist on
> having some text there I think it should be simpler. Something like
> "Select this if you want LED support."

Sounds good. (I really believe it should have _some_ text). Will you
just modify/apply the patch, or should I resent?
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: iwl3945: fix LED config option

[reinette chatre]

Hi Pavel,

On Mon, 2009-07-20 at 15:19 -0700, Pavel Machek wrote:
> > > diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
> > > index e092af0..95d9aa5 100644
> > > --- a/drivers/net/wireless/iwlwifi/Kconfig
> > > +++ b/drivers/net/wireless/iwlwifi/Kconfig
> > > @@ -9,6 +9,10 @@ config IWLWIFI
> > >  config IWLWIFI_LEDS
> > >  	bool "Enable LED support in iwlagn and iwl3945 drivers"
> > >  	depends on IWLWIFI
> > > +	default y
> > 
> > This is reasonable.
> > 
> > > +	---help---
> > > +	  Useful to get Wifi LED working at least on Thinkpad notebooks.
> > 
> > I do not think this says more than what we already have. I actually
> > think noting something about a particular platform will increase
> > confusion (as in "so where is the option if I have an acer laptop?")
> > because this option is needed to get LED support on all platforms. I do
> > think what is currently written is sufficient, but if you insist on
> > having some text there I think it should be simpler. Something like
> > "Select this if you want LED support."
> 
> Sounds good. (I really believe it should have _some_ text). Will you
> just modify/apply the patch, or should I resent?

Sure, I can take this patch and will send it through our usual route
(via linux-wireless).

Reinette

Re: iwl3945: fix LED config option

[Pavel Machek]

On Mon 2009-07-20 20:02:06, reinette chatre wrote:
> Hi Pavel,
> 
> On Mon, 2009-07-20 at 15:19 -0700, Pavel Machek wrote:
> > > > diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
> > > > index e092af0..95d9aa5 100644
> > > > --- a/drivers/net/wireless/iwlwifi/Kconfig
> > > > +++ b/drivers/net/wireless/iwlwifi/Kconfig
> > > > @@ -9,6 +9,10 @@ config IWLWIFI
> > > >  config IWLWIFI_LEDS
> > > >  	bool "Enable LED support in iwlagn and iwl3945 drivers"
> > > >  	depends on IWLWIFI
> > > > +	default y
> > > 
> > > This is reasonable.
> > > 
> > > > +	---help---
> > > > +	  Useful to get Wifi LED working at least on Thinkpad notebooks.
> > > 
> > > I do not think this says more than what we already have. I actually
> > > think noting something about a particular platform will increase
> > > confusion (as in "so where is the option if I have an acer laptop?")
> > > because this option is needed to get LED support on all platforms. I do
> > > think what is currently written is sufficient, but if you insist on
> > > having some text there I think it should be simpler. Something like
> > > "Select this if you want LED support."
> > 
> > Sounds good. (I really believe it should have _some_ text). Will you
> > just modify/apply the patch, or should I resent?
> 
> Sure, I can take this patch and will send it through our usual route
> (via linux-wireless).

thanks!


-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html