ath5k module vs. MadWifi

[Scott Robbins]

I am wondering, for those who have an AR5007EG card, if they are finding
that lately (perhaps 2 updates or so ago) that they are better off
reverting to the MadWifi drivers.  

My results are not consistent.  I imagine there are various differences
in the chipsets that I'm missing, but with two Acer laptops, one the
Aspire One, I get different results.  With the 4720z, the latest kernels
and ath5k module works without any issues.  On the AA1, it will
sometimes show nothing if doing iwlist scan, and other times, though it
will show available networks, will not get a DHCP address, nor work with
a statically assigned address.  I don't use NetworkManager, I configure
it manually, so that shouldn't be a factor. 

I note that the latest Ubuntu is disabling the ath5k module by default,
saying, in their release notes, that it will conflict with the MadWifi
driver--however, they seem a bit ambiguous as to which works better.
(They say that with some of the chipsets, the MadWifi won't work with
WPA, or maybe WPA2.)  In that case, one wonders why they are disabling
the ath5k instead of encouraging people to use it--this is why I felt it
was a bit ambiguous, they just suggest that if your MadWifi driver isn't
working, install a backports modules with updated ath5k drivers. 

The 4720z is working well with the ath5k.  So this might be some issue
with the AA1, however, I have a few installations on this machine--those
that are using the MadWifi driver connect without issue, while Rawhide
isn't. 

So, I'm wondering what other people are experiencing.  Is the ath5k
working for you?  

Thanks for any input--please note that at present, this is a fairly
trivial question as it's quite easy to blacklist and put in the MadWifi
driver.  Then, I uninstall it and put back the ath5k with each update.
:)
 

-- 
Scott Robbins
PGP keyID EB3467D6
( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

Spike:'Made with care for Randy.' (looks at Giles angrily) Randy 
Giles? Why not just call me 'Horny Giles,' or 'Desperate for a 
Shag Giles'? I knew there was a reason I hated you!
Giles: Randy's ... a family name, undoubtedly.

[Arthur Pemberton]

On Fri, Oct 31, 2008 at 10:41 PM, Scott Robbins <scottro(a)nyc.rr.com&gt; wrote:
>  I am wondering, for those who have an AR5007EG card, if they are
> finding
>  that lately (perhaps 2 updates or so ago) that they are better off
>  reverting to the MadWifi drivers. 

Was this working for you with Madwifi? it barely was for me, I
couldn't suspend. Updated to the .27 kernel and using the ath5k
driver, and stuff works now, even suspend.


-- 
Fedora 9 : sulphur is good for the skin
( www.pembo13.com )

[Scott Robbins]

On Sat, Nov 01, 2008 at 12:34:55AM -0500, Arthur Pemberton wrote:
>  On Fri, Oct 31, 2008 at 10:41 PM, Scott Robbins
> <scottro(a)nyc.rr.com&gt; wrote:
>  > I am wondering, for those who have an AR5007EG card, if they are finding
>  > that lately (perhaps 2 updates or so ago) that they are better off
>  > reverting to the MadWifi drivers.
>  
>  
>  Was this working for you with Madwifi? it barely was for me, I
>  couldn't suspend. Updated to the .27 kernel and using the ath5k
>  driver, and stuff works now, even suspend. 
Actually, I never used suspend.  In general, I only use the laptop(s) on
occasion, preferring the desktop.  

However, as far as connecting and keeping the connection goes, the
MadWifi always seemed a bit better, especially with the latest (and
possibly last) 20080903 snapshot.

Thanks for the input.  I'm really beginning to think (from talking to
friends as well) that it's an Aspire One thing.  (On the Aspire One
forums, many people mentioned something similar--the MadWif driver
working better than the ath5k.)

Thank you.

-- 
Scott Robbins
PGP keyID EB3467D6
( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

Angel: I've been around death before. A lot. I've lost
people. I've killed people...
Cordelia: And you are dead.

[Arthur Pemberton]

On Sat, Nov 1, 2008 at 3:58 AM, Scott Robbins <scottro(a)nyc.rr.com&gt; wrote:
>  On Sat, Nov 01, 2008 at 12:34:55AM -0500, Arthur Pemberton wrote:
> > On Fri, Oct 31, 2008 at 10:41 PM, Scott Robbins <scottro(a)nyc.rr.com&gt; wrote:
> > > I am wondering, for those who have an AR5007EG card, if they are finding
> > > that lately (perhaps 2 updates or so ago) that they are better off
> > > reverting to the MadWifi drivers.
> >
> >
> > Was this working for you with Madwifi? it barely was for me, I
> > couldn't suspend. Updated to the .27 kernel and using the ath5k
> > driver, and stuff works now, even suspend.
>
>  Actually, I never used suspend.  In general, I only use the laptop(s) on
>  occasion, preferring the desktop.
>
>  However, as far as connecting and keeping the connection goes, the
>  MadWifi always seemed a bit better, especially with the latest (and
>  possibly last) 20080903 snapshot.
>
>  Thanks for the input.  I'm really beginning to think (from talking to
>  friends as well) that it's an Aspire One thing.  (On the Aspire One
>  forums, many people mentioned something similar--the MadWif driver
>  working better than the ath5k.)
>  

Just so you know ath_pci is essentially dead, you're going to want to
switch to ath5k eventually. I believe Atheros is semi-actively
contributing to that project (I may be wrong here)


-- 
Fedora 9 : sulphur is good for the skin
( www.pembo13.com )

[Jason Farrell]

On Sat, Nov 1, 2008 at 12:34 AM, Arthur Pemberton <pemboa(a)gmail.com&gt; wrote:

>  On Fri, Oct 31, 2008 at 10:41 PM, Scott Robbins
> <scottro(a)nyc.rr.com&gt;
>  wrote:
>  > I am wondering, for those who have an AR5007EG card, if they are finding
>  > that lately (perhaps 2 updates or so ago) that they are better off
>  > reverting to the MadWifi drivers.
>
>
>  Was this working for you with Madwifi? it barely was for me, I
>  couldn't suspend. Updated to the .27 kernel and using the ath5k
>  driver, and stuff works now, even suspend. 

This has been my experience as well, on an acer aspire one: f9 + madwifi
worked well, but didn't work after suspend/resume (unless you manually
rmmod/modprobe again); f10 + ath5k in the .27 kernel works just as well, but
also after resume.

The only real downside of the current ath5k is that the wifi status LEDs no
longer work, but that's a small price to pay for everything working after
suspend/resume, and for not needing the madwifi kmod.

--Jason

[Scott Robbins]

On Sun, Nov 02, 2008 at 11:12:35AM -0500, Jason Farrell wrote:
>  On Sat, Nov 1, 2008 at 12:34 AM, Arthur Pemberton
> <pemboa(a)gmail.com&gt; wrote:
>  
>  
>  The only real downside of the current ath5k is that the wifi status LEDs no
>  longer work, but that's a small price to pay for everything working after
>  suspend/resume, and for not needing the madwifi kmod. 
If you use MadWifi and the 2.6.27 kernel, are you able to get the LED
working?  I thought that was kernel specific as the sysctl options don't
seem to be there (the ledpins and softled.)

Thanks.

-- 
Scott Robbins
PGP keyID EB3467D6
( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

Angelus: No weapons... no friends... no hope. Take all 
that away and what's left? 
Buffy: Me.

[Partha Bagchi]

On Sun, Nov 2, 2008 at 12:48 PM, Scott Robbins <scottro(a)nyc.rr.com&gt; wrote:
>  On Sun, Nov 02, 2008 at 11:12:35AM -0500, Jason Farrell wrote:
> > On Sat, Nov 1, 2008 at 12:34 AM, Arthur Pemberton <pemboa(a)gmail.com&gt; wrote:
> >
> >
> > The only real downside of the current ath5k is that the wifi status LEDs no
> > longer work, but that's a small price to pay for everything working after
> > suspend/resume, and for not needing the madwifi kmod.
>
>  If you use MadWifi and the 2.6.27 kernel, are you able to get the LED
>  working?  I thought that was kernel specific as the sysctl options don't
>  seem to be there (the ledpins and softled.)
>
>  Thanks.
>
>  --
>  Scott Robbins
>  PGP keyID EB3467D6
>  ( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
>  gpg --keyserver pgp.mit.edu --recv-keys EB3467D6
>
>  Angelus: No weapons... no friends... no hope. Take all
>  that away and what's left?
>  Buffy: Me.
>
>  --
>  fedora-test-list mailing list
>  fedora-test-list(a)redhat.com
>  To unsubscribe:
>  https://www.redhat.com/mailman/listinfo/fedora-test-list
>  
Does anyone know why the softled options were removed from 2.6.27?

Thanks,
Partha

[Tim Lauridsen]

Scott Robbins wrote:
>  I am wondering, for those who have an AR5007EG card, if they are
> finding
>  that lately (perhaps 2 updates or so ago) that they are better off
>  reverting to the MadWifi drivers.  
>
>  My results are not consistent.  I imagine there are various differences
>  in the chipsets that I'm missing, but with two Acer laptops, one the
>  Aspire One, I get different results.  With the 4720z, the latest kernels
>  and ath5k module works without any issues.  On the AA1, it will
>  sometimes show nothing if doing iwlist scan, and other times, though it
>  will show available networks, will not get a DHCP address, nor work with
>  a statically assigned address.  I don't use NetworkManager, I configure
>  it manually, so that shouldn't be a factor. 
>
>  I note that the latest Ubuntu is disabling the ath5k module by default,
>  saying, in their release notes, that it will conflict with the MadWifi
>  driver--however, they seem a bit ambiguous as to which works better.
>  (They say that with some of the chipsets, the MadWifi won't work with
>  WPA, or maybe WPA2.)  In that case, one wonders why they are disabling
>  the ath5k instead of encouraging people to use it--this is why I felt it
>  was a bit ambiguous, they just suggest that if your MadWifi driver isn't
>  working, install a backports modules with updated ath5k drivers. 
>
>  The 4720z is working well with the ath5k.  So this might be some issue
>  with the AA1, however, I have a few installations on this machine--those
>  that are using the MadWifi driver connect without issue, while Rawhide
>  isn't. 
>
>  So, I'm wondering what other people are experiencing.  Is the ath5k
>  working for you?  
>
>  Thanks for any input--please note that at present, this is a fairly
>  trivial question as it's quite easy to blacklist and put in the MadWifi
>  driver.  Then, I uninstall it and put back the ath5k with each update.
>  :)
>    
The current ath5k in rawhide, works very well for me on my Lenovo T60 
laptop (using WP2 / NetworkManager)

03:00.0 Ethernet controller: Atheros Communications Inc. AR5212 
802.11abg NIC (rev 01)

It has not be working very well for me in the past and i had to fallback 
to using the madwifi driver.

Disabling ath5k in Ubuntu, seems kind of strange, they could just add a 
backlist file, like the livna/rpmfusion madwifi driver does, but again
do a lot of weird stuff :)

Tim

[Jim]

Scott Robbins wrote:
>  I am wondering, for those who have an AR5007EG card, if they are
> finding
>  that lately (perhaps 2 updates or so ago) that they are better off
>  reverting to the MadWifi drivers.  
>
>  My results are not consistent.  I imagine there are various differences
>  in the chipsets that I'm missing, but with two Acer laptops, one the
>  Aspire One, I get different results.  With the 4720z, the latest kernels
>  and ath5k module works without any issues.  On the AA1, it will
>  sometimes show nothing if doing iwlist scan, and other times, though it
>  will show available networks, will not get a DHCP address, nor work with
>  a statically assigned address.  I don't use NetworkManager, I configure
>  it manually, so that shouldn't be a factor. 
>
>  I note that the latest Ubuntu is disabling the ath5k module by default,
>  saying, in their release notes, that it will conflict with the MadWifi
>  driver--however, they seem a bit ambiguous as to which works better.
>  (They say that with some of the chipsets, the MadWifi won't work with
>  WPA, or maybe WPA2.)  In that case, one wonders why they are disabling
>  the ath5k instead of encouraging people to use it--this is why I felt it
>  was a bit ambiguous, they just suggest that if your MadWifi driver isn't
>  working, install a backports modules with updated ath5k drivers. 
>
>  The 4720z is working well with the ath5k.  So this might be some issue
>  with the AA1, however, I have a few installations on this machine--those
>  that are using the MadWifi driver connect without issue, while Rawhide
>  isn't. 
>
>  So, I'm wondering what other people are experiencing.  Is the ath5k
>  working for you?  
>
>  Thanks for any input--please note that at present, this is a fairly
>  trivial question as it's quite easy to blacklist and put in the MadWifi
>  driver.  Then, I uninstall it and put back the ath5k with each update.
>  :)
>   
>
>     
The only answer to that is update to the kernel-2.6.27 and all your 
troubles will go away.

[Scott Robbins]

On Sat, Nov 01, 2008 at 11:01:52AM -0400, Jim wrote:
>  Scott Robbins wrote:
> > 

> > Thanks for any input--please note that at present, this is a
> fairly
> > trivial question as it's quite easy to blacklist and put in the MadWifi
> > driver.  Then, I uninstall it and put back the ath5k with each update.
> > :)
> >  
> >
> >   
>  The only answer to that is update to the kernel-2.6.27 and all your  
>  troubles will go away. 
Sorry, I thought I'd mentioned that and apparently I didn't.  I've been
using the 2.6.27 kernel for awhile now, the problems seemed to begin
after a recent update to it. 


-- 
Scott Robbins
PGP keyID EB3467D6
( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

Ford: I wanna be like you...A vampire. 
Spike: I've known you for two minutes and I can't stand you. I 
don't really feature you living forever. Can I eat him now love?

[potat0]

The ath5k kernel module was working fine for me until the latest update.  I do not use
network manager preferring to manually activate using system-config-network.  Now whenever
I try to activate I get "ipcalc: netmask or prefix expected".  This does not
seem to be kernel related as it now occurs in both 2.6.27-4.68 and 2.6.27-4.58 where it
was all working prior to 30th October update.


-- 
This is an email sent via The Fedora Community Portal https://fcp.surfsite.org
https://fcp.surfsite.org/modules/newbb/viewtopic.php?post_id=301123&t...
If you think, this is spam, please report this to webmaster(a)fcp.surfsite.org and/or blame
spoffley(a)gmail.com.

[Scott Robbins]

On Sun, Nov 02, 2008 at 06:23:24PM +0100, spoffley wrote:
>  The ath5k kernel module was working fine for me until the latest
> update.  I do not use network manager preferring to manually activate using
> system-config-network.  Now whenever I try to activate I get "ipcalc: netmask or
> prefix expected".  This does not seem to be kernel related as it now occurs in both
> 2.6.27-4.68 and 2.6.27-4.58 where it was all working prior to 30th October update.
>  
>   
Perhaps that's a separate, system-config-network issue?  Did you try
that one with the MadWifi as well and get the same error? 
(That's just a guess as I don't use that, I simply have a little script
if I'm using wireless that starts wpa_supplicant then runs dhclient.)


-- 
Scott Robbins
PGP keyID EB3467D6
( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

Willy: What are you gonna' do with him, anyway? 
Spike: I'm thinkin' maybe dinner and a movie. I don't want to 
rush into anything. I've been hurt, you know.