Brian | 1 Dec 04:02 2008

Re: Possible Bug Caps lock flashing laptop frozen

All

Brian wrote:

> 
> However, I have only had this problem since using the wireless-testing
> kernel. So just a quick note, in case someone else has the same problem.
> 
While the above statement is true, there was one other change I made at
the same time and that was to put the wireless N adapter in.
I think this must use a bit more power and my laptop must have been
marginal.
Anyway, read on the web that a possible solution was to remove the
battery. Did this and I have been running for a couple of hours with no
problems.
I even have my NX session running to control another machine which is
also wireless. So the wireless is working great and the response time is
instantaneous.

The other thing that has changed is that the adapter LEDs are now
flashing, they were on solid before.

Brian
Pat O'Neil | 1 Dec 18:13 2008
Picon

ar9281

Greetings all,

I've got a recent toshiba qosmio laptop, 64bit that has an ar9281
chipset that was "almost" working under F9, as after downloading the
wireless git and compiling I could get it to come up and associate, but
the performance was extremely slow, with ping times of 10's of seconds,
etc.  I also sometimes got the aforementioned "No ProbeResp from current
AP ... - assume out of range" error.

Since then I've loaded Fedora 10, and have hit a new snag with its
bundled version of ath9k, as in the dmesg log I see the following:

...snip...

ath9k: 0.1
ath_pci: software IOTLB in use, aborting.

...snip...

and no wireless interface shows up at all.  Any ideas?  

Thanks for all the work on atheros drivers!

Pat O'Neil
Luis R. Rodriguez | 1 Dec 18:25 2008
Picon

Re: ar9281

On Mon, Dec 1, 2008 at 9:13 AM, Pat O'Neil <Pat.ONeil <at> fedex.com> wrote:
> Greetings all,
>
> I've got a recent toshiba qosmio laptop, 64bit that has an ar9281
> chipset that was "almost" working under F9, as after downloading the
> wireless git and compiling I could get it to come up and associate, but
> the performance was extremely slow, with ping times of 10's of seconds,
> etc.  I also sometimes got the aforementioned "No ProbeResp from current
> AP ... - assume out of range" error.
>
> Since then I've loaded Fedora 10, and have hit a new snag with its
> bundled version of ath9k, as in the dmesg log I see the following:
>
> ...snip...
>
> ath9k: 0.1
> ath_pci: software IOTLB in use, aborting.
>
> ...snip...

Fedora 10 kernel is broken as a workaround to a kernel bug in ath9k
they decided to disable ath9k when the swiotbl is used due to a bug
that *we actually did fix*. They should fix their kernel. They may
have an updated kernel you may want to check.

http://bugzilla.kernel.org/show_bug.cgi?id=11811

  Luis
Luis R. Rodriguez | 2 Dec 05:49 2008

Re: [ath5k-devel] License of changes to ath5k/ath9k

2008/11/26 Bob Copeland <me@...>:
> On Tue, Nov 25, 2008 at 5:28 PM, Luis R. Rodriguez
> <lrodriguez@...> wrote:
>> Sure, BTW I had created such a script a long time ago, the files just
>> need to be updated (so update them if you download this thingy):
>>
>> http://www.kernel.org/pub/linux/kernel/people/mcgrof/scripts/ath5k-license.pl.txt
>>
>
> Cool, that saved me some work.  I updated this so that the files are
> now up-to-date, and it can be called directly from the commit-msg hook
> (also handles existing Signed-off-by etc).  Feel free to take whatever
> you like from my version.

I've updated the file with your stuff, thanks!

  Luis
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html

advan_xiang | 2 Dec 08:21 2008

Re: ath9k-devel Digest, Vol 5, Issue 14

hi,all
  I would like to switch RX filter bandwidth to 5/10MHz in AR5413 chipset, I fixed 0x9944 register , but I'm not sure whether is right , someone helps me ? thank you in advance .
 
advan
 
advan_xiang
2008-12-02
发件人: ath9k-devel-request <at> lists.ath9k.org
发送时间: 2008-11-24 09:20:07
收件人: ath9k-devel <at> lists.ath9k.org
抄送:
主题: ath9k-devel Digest, Vol 5, Issue 14
 
Send ath9k-devel mailing list submissions to
ath9k-devel <at> lists.ath9k.org
 
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.ath9k.org/mailman/listinfo/ath9k-devel
or, via email, send a message with subject or body 'help' to
ath9k-devel-request <at> lists.ath9k.org
 
You can reach the person managing the list at
ath9k-devel-owner <at> lists.ath9k.org
 
When replying, please edit your Subject line so it is more specific
than "Re: Contents of ath9k-devel digest..."
 
 
Today's Topics:
 
   1. Re: Dirty Laundry List (Luis R. Rodriguez)
   2. Re: Dirty Laundry List (Alistair John Strachan)
   3. Re: Dirty Laundry List (Luis R. Rodriguez)
   4. Re: It Works! (Brian)
   5. Re: Dirty Laundry List (Alistair John Strachan)
   6. Re: It Works! (Sujith)
   7. Re: Issues (Brian)
   8. Re: Issues (Sujith)
 
 
----------------------------------------------------------------------
 
Message: 1
Date: Thu, 20 Nov 2008 11:32:47 -0800
From: "Luis R. Rodriguez"  <lrodriguez <at> atheros.com >
Subject: Re: [ath9k-devel] Dirty Laundry List
To: Sujith Manoharan  <Sujith.Manoharan <at> Atheros.com >
Cc: ath9k-devel <at> lists.ath9k.org
Message-ID:  <20081120193247.GC5899 <at> tesla >
Content-Type: text/plain; charset="us-ascii"
 
Moving this discussion out to the public list. Guys this is our
latest TODO list. We welcome people from the community to take up items
here, but please let us know if you are going to work on something.
 
I'll be tackling the MacBook pro swiotlb issue, I think this should be
fixed now, but not sure, needs a MacBook Pro 3.1 tester. I also already
started working on the ath9k regulatory scrub. Essentially we'll be
relying on CRDA for 11d instead of the EEPROM so this is a huge major
cleanup. Also there is a lot of potential to share this work with ath5k
code. One option which Jouni suggested was a module we can share between
ath9k/ath5k, maybe some sort of ath_common.ko or the like and first
stash in there reg stuff. Maybe later add more shared stuff as we go.
 
On Thu, Nov 20, 2008 at 08:22:21AM -0800, Sujith Manoharan wrote:
> Hi,
> So more dirty linen to be washed, and here it is.
> Issues/Bugs:
> * PCI/DMA suckage ( SWIOTLB issue on macbooks )
 
Taken, need a MacBook pro 3.1 tester.
 
> * Fix xmit ( there is still a difference between xmit and recv )
>   (I don't see the difference in UDP, only in TCP)
> * TX/RX Hang
>   (I run into TX/RX hang after running Iperf for some time, no idea as to what's happening)
> * Fix convoluted aggregation logic
>   (bf_lastbf, bf_lastfrm, bf_next, bf_rifslast, buf_q, tid_q, axq_acq ...)
> * Fix Channel setting mess
>   (Racy, calculation of HT parameters is buggy)
> * RC scaling
>   (More analysis/optimization needs to be done here)
> * Fix probe/auth timeout issues
> * Fix HT handling mess
>   (reassociation with HT40 is a problem, switching between legacy/HT APs fails occasionally too)
> * Fix locking
>   (Yeah, right, like we have locking to begin with)
> * Throughput with legacy APs suck, analyze this
>   (Compare with Minstrel for legacy mode ?)
> * AR5416 based cards show lower throughput numbers than AR9280 and above. Fix this.
> * Handle PCI suspend/resume callbacks properly
> * Check ANI
>   (Does it work ?)
> * WTF is Slow Antenna Diversity and why have we not done anything to enable it so far ?
> * Lots of code for chainmask handling is present ( timers etc.), what are we doing with all that ?
> * Lockdep warning on ath_tid_drain() when association is lost and we try to clean up
> Cleanup love:
> Major Items ( I know, we keep listing these again and again ):
> * 11d (Regulatory Scrub)
 
Taken.
 
> * 11h
> * SoftAP
> * Power Save (legacy)
> * BT Coex
> Sujith
 
 
------------------------------
 
Message: 2
Date: Thu, 20 Nov 2008 23:23:19 +0000
From: Alistair John Strachan  <alistair <at> devzero.co.uk >
Subject: Re: [ath9k-devel] Dirty Laundry List
To: ath9k-devel <at> lists.ath9k.org
Message-ID:  <200811202323.19327.alistair <at> devzero.co.uk >
Content-Type: text/plain;  charset="us-ascii"
 
On Thursday 20 November 2008 19:32:47 Luis R. Rodriguez wrote:
[snip]
>  > * PCI/DMA suckage ( SWIOTLB issue on macbooks )
>
> Taken, need a MacBook pro 3.1 tester.
 
Why does this have to be a MacBook Pro 3.1? If the ath9k driver's got swiotlb 
issues it should be possible to use any supported hardware in combination with 
the sw-iommu right? The IOMMU can be forced on with machines less than ~4GB, 
so just about anybody could test it.
 
If any of the above is correct, I've got a machine I could test the fix on.
 
-- 
Cheers,
Alistair.
 
 
------------------------------
 
Message: 3
Date: Thu, 20 Nov 2008 15:28:47 -0800
From: "Luis R. Rodriguez"  <lrodriguez <at> atheros.com >
Subject: Re: [ath9k-devel] Dirty Laundry List
To: Alistair John Strachan  <alistair <at> devzero.co.uk >
Cc: Luis, "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >,
Sujith Manoharan  <Sujith.Manoharan <at> Atheros.com >, Rodriguez
<Luis.Rodriguez <at> Atheros.com >
Message-ID:  <20081120232847.GE5899 <at> tesla >
Content-Type: text/plain; charset="us-ascii"
 
On Thu, Nov 20, 2008 at 03:23:19PM -0800, Alistair John Strachan wrote:
> On Thursday 20 November 2008 19:32:47 Luis R. Rodriguez wrote:
> [snip]
>  >  > * PCI/DMA suckage ( SWIOTLB issue on macbooks )
>  >
>  > Taken, need a MacBook pro 3.1 tester.
> Why does this have to be a MacBook Pro 3.1? If the ath9k driver's got swiotlb
> issues it should be possible to use any supported hardware in combination with
> the sw-iommu right? The IOMMU can be forced on with machines less than ~4GB,
> so just about anybody could test it.
 
You would think right!? Have you seen the bug report?
 
 
> If any of the above is correct, I've got a machine I could test the fix on.
 
I've been testing on iommu=soft swiotlb=force mem=4G while doing all
sorts of other memory intensive DMA things but no luck. So something is
peculiar about the MacBook pro 3.1. So we have patches now, so what we
need is to see if it still occurs with such patches. If anyone has a
MacBook pro 3.1 and is willing to test please let me know.
 
  Luis
 
 
------------------------------
 
Message: 4
Date: Fri, 21 Nov 2008 10:04:00 +1000
From: Brian  <bnc <at> astronomicalresearchaustralia.org >
Subject: Re: [ath9k-devel] It Works!
Cc: "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >
Message-ID:  <4925FAF0.3040202 <at> astronomicalresearchaustralia.org >
Content-Type: text/plain; charset=ISO-8859-1
 
 
 
Luis,
 
> So we had over 40 patches pending which John just merged today. You
> can do a git pull or wait for compat-wireless to get updated.
Took your suggestion and did a git-pull.
 
Did a recompile after carefully checking menuconfig again.
The only major thing I have done is to remove any reference to IPV6.
 
Rebooted and everything came up and WORKED!:)
Not sure what you changed but it has done the trick.
 
I have been up for about 30mins with no issues.
2.6.28-rc5-wl-W #11 SMP
 
With this result
Mode Rate (Mbps) Signal (%)
802.11n (2.4GHz) 78 100
 
 
So while I am happy the 11n is now working the rate does not seem to be
as expected.
 
ISSUES so far
 
1) When I do a ping there is about a 10-15sec delay before the responses
start coming back.
 
2) I cannot start WPA_GUI or WPA_CLI it comes back with
Selected interface 'wlan0.pid'
Trying to connect to '/var/run/wpa_supplicant/wlan0.pid'
Failed to open control connection to wpa_supplicant.
PING failed - trying to reconnect
 
The pid is correct.
 
 
So many thanks to the atheros team for getting ath9k going.
 
Brian
 
 
 
------------------------------
 
Message: 5
Date: Fri, 21 Nov 2008 00:33:38 +0000
From: Alistair John Strachan  <alistair <at> devzero.co.uk >
Subject: Re: [ath9k-devel] Dirty Laundry List
To: "Luis R. Rodriguez"  <lrodriguez <at> atheros.com >
Cc: "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >, Luis
Rodriguez  <Luis.Rodriguez <at> atheros.com >
Message-ID:  <200811210033.38232.alistair <at> devzero.co.uk >
Content-Type: text/plain;  charset="us-ascii"
 
On Thursday 20 November 2008 23:28:47 Luis R. Rodriguez wrote:
> On Thu, Nov 20, 2008 at 03:23:19PM -0800, Alistair John Strachan wrote:
[snip]
>  > Why does this have to be a MacBook Pro 3.1? If the ath9k driver's got
>  > swiotlb issues it should be possible to use any supported hardware in
>  > combination with the sw-iommu right? The IOMMU can be forced on with
>  > machines less than ~4GB, so just about anybody could test it.
>
> You would think right!? Have you seen the bug report?
>
 
No, very interesting! Fortunately I was able to reproduce the problem on my 
2GB RAM Macbook v2, despite the experiences you guys have had.
 
I built a kernel (2.6.28-rc5) with SWIOTLB set, and then booted with 
"iommu=soft swiotlb=force mem=512M" and then copied a TV show onto the laptop 
from another computer. It broke almost instantly (and took some of my data 
with it :-)).
 
[snip]
> So we have patches now, so what we
> need is to see if it still occurs with such patches. If anyone has a
> MacBook pro 3.1 and is willing to test please let me know.
 
I applied the patches, re-ran my test case, and half a gig later the problem 
seems to be gone. So if this is any use to you at all I'd say it's fixed by 
those patches, which do make a lot of sense.
 
The driver tested was the one in 2.6.28-rc5. I had to apply the patch manually 
because the driver in 2.6.28-rc5 is formatted in a substantially different 
way. No functional differences afaict.
 
-- 
Cheers,
Alistair.
 
 
------------------------------
 
Message: 6
Date: Fri, 21 Nov 2008 07:09:18 +0530
From: Sujith  <m.sujith <at> gmail.com >
Subject: Re: [ath9k-devel] It Works!
To: bnc <at> astronomicalresearchaustralia.org
Cc: "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >
Message-ID:  <18726.4422.322793.34364 <at> localhost.localdomain >
Content-Type: text/plain; charset=us-ascii
 
Brian wrote:
> Did a recompile after carefully checking menuconfig again.
> The only major thing I have done is to remove any reference to IPV6.
> Rebooted and everything came up and WORKED!:)
> Not sure what you changed but it has done the trick.
> I have been up for about 30mins with no issues.
> 2.6.28-rc5-wl-W #11 SMP
> With this result
> Mode Rate (Mbps) Signal (%)
> 802.11n (2.4GHz) 78 100
> So while I am happy the 11n is now working the rate does not seem to be
> as expected.
 
I am assuming you are using HT20 ?
 
> ISSUES so far
> 1) When I do a ping there is about a 10-15sec delay before the responses
> start coming back.
 
This is an issue, and we are looking into this.
 
> 2) I cannot start WPA_GUI or WPA_CLI it comes back with
> Selected interface 'wlan0.pid'
> Trying to connect to '/var/run/wpa_supplicant/wlan0.pid'
> Failed to open control connection to wpa_supplicant.
> PING failed - trying to reconnect
 
Can you show your wpa_supplicant config ?
 
Sujith
-- 
 
 
------------------------------
 
Message: 7
Date: Fri, 21 Nov 2008 14:52:44 +1000
From: Brian  <bnc <at> astronomicalresearchaustralia.org >
Subject: Re: [ath9k-devel] Issues
Cc: "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >
Message-ID:  <49263E9C.3050401 <at> astronomicalresearchaustralia.org >
Content-Type: text/plain; charset=ISO-8859-1
 
 
 
Sujith,
> I am assuming you are using HT20 ?
Sorry, what is that? I did a git-pull about 18hours ago.
 
> > ISSUES so far
> >
> > 1) When I do a ping there is about a 10-15sec delay before the responses
> > start coming back.
> >
> This is an issue, and we are looking into this.
Ok, This could be generic. By that I mean I get a similar delay when I
log into my router. Subjective of course.
 
 
 
> > 2) I cannot start WPA_GUI or WPA_CLI it comes back with
> > Selected interface 'wlan0.pid'
> > Trying to connect to '/var/run/wpa_supplicant/wlan0.pid'
> > Failed to open control connection to wpa_supplicant.
> > PING failed - trying to reconnect
> Can you show your wpa_supplicant config ?
Is their a show command?
In /var/run/wpa_supplicant-wlan0.conf
 
I have
ctrl_interface=/var/run/wpa_supplicant
network={
  scan_ssid=1
  ssid="dlink"
  key_mgmt=WPA-PSK
  psk="secret"
}
 
Brian
 
 
 
------------------------------
 
Message: 8
Date: Fri, 21 Nov 2008 11:46:22 +0530
From: Sujith  <Sujith.Manoharan <at> atheros.com >
Subject: Re: [ath9k-devel] Issues
To: "bnc <at> astronomicalresearchaustralia.org"
<bnc <at> astronomicalresearchaustralia.org >
Cc: "ath9k-devel <at> lists.ath9k.org"  <ath9k-devel <at> lists.ath9k.org >
Message-ID:  <18726.21046.859439.185937 <at> gargle.gargle.HOWL >
Content-Type: text/plain; charset="us-ascii"
 
Brian wrote:
> Sorry, what is that? I did a git-pull about 18hours ago.
 
I meant the mode in which you have connected to the AP ( HT20/HT40 ).
Since you are operating in 2.4 Ghz, it probably is HT20, in which case
the numbers are not too bad.
 
But if your AP is dual-band, you can configure it to use HT40 in 5Ghz
and get improved throughput.
 
>  > > 2) I cannot start WPA_GUI or WPA_CLI it comes back with
>  > > Selected interface 'wlan0.pid'
>  > > Trying to connect to '/var/run/wpa_supplicant/wlan0.pid'
>  > > Failed to open control connection to wpa_supplicant.
>  > > PING failed - trying to reconnect
>  >
>  > Can you show your wpa_supplicant config ?
> Is their a show command?
> In /var/run/wpa_supplicant-wlan0.conf
> I have
> ctrl_interface=/var/run/wpa_supplicant
> network={
>   scan_ssid=1
>   ssid="dlink"
>   key_mgmt=WPA-PSK
>   psk="secret"
> }
 
Does running wpa_supplicant manually connect to the network ?
 
Sujith
 
 
------------------------------
 
_______________________________________________
ath9k-devel mailing list
ath9k-devel <at> lists.ath9k.org
https://lists.ath9k.org/mailman/listinfo/ath9k-devel
 
 
End of ath9k-devel Digest, Vol 5, Issue 14
******************************************
 
_______________________________________________
ath9k-devel mailing list
ath9k-devel <at> lists.ath9k.org
https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Mark Ryden | 2 Dec 10:34 2008
Picon

power save mode in ath9k driver

Hello,

I am performing some tests with various wireless nics and an access point.

I have a question about the driver: As far as I know, this driver
does support power save mode. When the device enters power saving
mode, it should send a control frame (beacon) with power management
bit set to 1, to the access point it is associated with (I am talking
about infrastructure network).

When it exits power save mode, it should send PSPOLL frame to
retrieve the frames that the AP had buffered for it.

I tried to delve into the driver code and could not find the code
which handles entering power save mode. Grepping for
"IEEE80211_FCTL_PM" in the driver code did not find anything
(IEEE80211_FCTL_PM is the power save mode bit in a control frame).
(I am looking at the code of the latest wireless-testing git tree).

Where is the code that handles entering power save mode in this driver?

Rgs,
Mark
Dani Camps | 2 Dec 11:46 2008
Picon

Interaction between retransmission and aggregation mechanisms

Dear all,

I am developing a simulator for 802.11n. I have a question regarding the algorithms that you use for
retransmissions of aggregated MPDUs in the ath9k driver.

The question is the following:

-Consider that the driver builds an aggregated MPDU with N aggregated MPDUs inside. This A-MPDU accesses
the channel through AC_VO, and say that you reach the maximum number of allowed retransmissions and
therefore as stated by the standard you have to drop the frame.

My question is: What do you do in this case? Do you drop the whole A-MPDU and therefore the N aggregated MPDUs
inside ? Or do you somehow de-aggregate and only drop the first of the MPDUs and return the rest to the
transmission buffer ?

In the simulator I could either thing, being of course de-aggregation much more effective, but I would like
to know you opinion in the matter because I am sure there are some constraints in a real device that I am not
aware of.

Best Regards

Daniel
John W. Linville | 2 Dec 21:22 2008

Re: ar9281

On Mon, Dec 01, 2008 at 09:25:33AM -0800, Luis R. Rodriguez wrote:
> On Mon, Dec 1, 2008 at 9:13 AM, Pat O'Neil <Pat.ONeil <at> fedex.com> wrote:
> > Greetings all,
> >
> > I've got a recent toshiba qosmio laptop, 64bit that has an ar9281
> > chipset that was "almost" working under F9, as after downloading the
> > wireless git and compiling I could get it to come up and associate, but
> > the performance was extremely slow, with ping times of 10's of seconds,
> > etc.  I also sometimes got the aforementioned "No ProbeResp from current
> > AP ... - assume out of range" error.
> >
> > Since then I've loaded Fedora 10, and have hit a new snag with its
> > bundled version of ath9k, as in the dmesg log I see the following:
> >
> > ...snip...
> >
> > ath9k: 0.1
> > ath_pci: software IOTLB in use, aborting.
> >
> > ...snip...
> 
> Fedora 10 kernel is broken as a workaround to a kernel bug in ath9k
> they decided to disable ath9k when the swiotbl is used due to a bug
> that *we actually did fix*. They should fix their kernel. They may
> have an updated kernel you may want to check.
> 
> http://bugzilla.kernel.org/show_bug.cgi?id=11811

Not quite that simple.  Those patches as written don't apply to 2.6.27,
and F10 won't be on 2.6.28 for quite some time (since 2.6.28 has yet
to be released).

Now, if you know of 2.6.27 versions of those patches... :-)

John
--

-- 
John W. Linville		Linux should be at the core
linville <at> tuxdriver.com			of your literate lifestyle.
Luis R. Rodriguez | 2 Dec 21:36 2008
Picon

Re: ar9281

On Tue, Dec 2, 2008 at 12:22 PM, John W. Linville
<linville <at> tuxdriver.com> wrote:
> On Mon, Dec 01, 2008 at 09:25:33AM -0800, Luis R. Rodriguez wrote:
>> On Mon, Dec 1, 2008 at 9:13 AM, Pat O'Neil <Pat.ONeil <at> fedex.com> wrote:
>> > Greetings all,
>> >
>> > I've got a recent toshiba qosmio laptop, 64bit that has an ar9281
>> > chipset that was "almost" working under F9, as after downloading the
>> > wireless git and compiling I could get it to come up and associate, but
>> > the performance was extremely slow, with ping times of 10's of seconds,
>> > etc.  I also sometimes got the aforementioned "No ProbeResp from current
>> > AP ... - assume out of range" error.
>> >
>> > Since then I've loaded Fedora 10, and have hit a new snag with its
>> > bundled version of ath9k, as in the dmesg log I see the following:
>> >
>> > ...snip...
>> >
>> > ath9k: 0.1
>> > ath_pci: software IOTLB in use, aborting.
>> >
>> > ...snip...
>>
>> Fedora 10 kernel is broken as a workaround to a kernel bug in ath9k
>> they decided to disable ath9k when the swiotbl is used due to a bug
>> that *we actually did fix*. They should fix their kernel. They may
>> have an updated kernel you may want to check.
>>
>> http://bugzilla.kernel.org/show_bug.cgi?id=11811
>
> Not quite that simple.  Those patches as written don't apply to 2.6.27,
> and F10 won't be on 2.6.28 for quite some time (since 2.6.28 has yet
> to be released).
>
> Now, if you know of 2.6.27 versions of those patches... :-)

I provided a port of the required 3 patches to 2.6.27 in the bug
report, here's the linky:

http://www.kernel.org/pub/linux/kernel/people/mcgrof/patches/ath9k/2008-11-22/27-IOMMU-01/

Let me know if there is anything else I can do to help.

  Luis
Christoph Thiel | 2 Dec 21:37 2008

Re: ar9281

On Tue, Dec 02, 2008 at 03:22:22PM -0500, John W. Linville wrote:
> On Mon, Dec 01, 2008 at 09:25:33AM -0800, Luis R. Rodriguez wrote:
> > On Mon, Dec 1, 2008 at 9:13 AM, Pat O'Neil <Pat.ONeil <at> fedex.com> wrote:
> > > Greetings all,
> > >
> > > I've got a recent toshiba qosmio laptop, 64bit that has an ar9281
> > > chipset that was "almost" working under F9, as after downloading the
> > > wireless git and compiling I could get it to come up and associate, but
> > > the performance was extremely slow, with ping times of 10's of seconds,
> > > etc.  I also sometimes got the aforementioned "No ProbeResp from current
> > > AP ... - assume out of range" error.
> > >
> > > Since then I've loaded Fedora 10, and have hit a new snag with its
> > > bundled version of ath9k, as in the dmesg log I see the following:
> > >
> > > ...snip...
> > >
> > > ath9k: 0.1
> > > ath_pci: software IOTLB in use, aborting.
> > >
> > > ...snip...
> > 
> > Fedora 10 kernel is broken as a workaround to a kernel bug in ath9k
> > they decided to disable ath9k when the swiotbl is used due to a bug
> > that *we actually did fix*. They should fix their kernel. They may
> > have an updated kernel you may want to check.
> > 
> > http://bugzilla.kernel.org/show_bug.cgi?id=11811
> 
> Not quite that simple.  Those patches as written don't apply to 2.6.27,
> and F10 won't be on 2.6.28 for quite some time (since 2.6.28 has yet
> to be released).
> 
> Now, if you know of 2.6.27 versions of those patches... :-)

Check out
http://www.kernel.org/pub/linux/kernel/people/mcgrof/patches/ath9k/2008-11-22/27-IOMMU-01/

I hope we'll get them included in one of the next -stable kernels.

Best,
Christoph

Gmane