Fwd: kernel .30 BROKE ATH5K with my AR5212 atheros.

Fwd: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

---------- Forwarded message ----------
From: Jasin Colegrove <j.wholesalesupply@gmail.com>
Date: Tue, Jun 30, 2009 at 6:24 PM
Subject: kernel .30 BROKE ATH5K with my AR5212 atheros.
To: linux-wireless@vger.kernel.org


I AM NOT A SUBSCRIBER. Please reply to me directly at this e-mail
address j.wholesalesupply@gmail.com

lspci -vvvvvvvvvv

02:02.0 Ethernet controller: Atheros Communications Inc. AR5212
802.11abg NIC (rev 01)
    Subsystem: Phillips Components Device 8331
    Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop-
ParErr- Stepping- SERR+ FastB2B- DisINTx-
    Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR- INTx-
    Latency: 168 (2500ns min, 7000ns max), Cache Line Size: 32 bytes
    Interrupt: pin A routed to IRQ 11
    Region 0: Memory at c0200000 (32-bit, non-prefetchable) [size=64K]
    Capabilities: <access denied>
    Kernel driver in use: ath5k
    Kernel modules: ath5k


$ lsmod | grep ath
ath5k                 107192  0
mac80211              158652  1 ath5k
led_class               3416  2 thinkpad_acpi,ath5k
cfg80211               36524  2 ath5k,mac80211

Once I installed and ran the .30 release on my laptop my internet
connection was down. Wicd reported that the ping failed because it did
not recieve a response from the access point.  So I started checking
things out, I have no wpa/wep security. I have no mac filtering. If I
try using a static IP to the router, the manual method of ifconfig
192.168.1.4; route add default gw 192.168.1.1 works, but I can not
ping anything outside of my localhost. Pinging the gateway fails,
pinging other computers on the network fails. If I try dhcp, it
timesout waiting for an ip and same thing, no ping outside of
localhost, as expected.

Then I decided to see if it the card was working, So I tried to iwlist
wlan0 scan, it returned the expected results. It see's the multiple
AP's in my area. The card seems fine.

That got me to thinking, maybe I should downgrade the kernel and see
if it works, as soon as i rebooted into .29 kernel release the
authentication to the AP went through and i could ping the gw, other
networked computers etc.

There is clearly a problem with the .30 release of ath5k module using
the atheros chipset AR5212. I would be glad to offer up as much
information as I can, but I don't have the slightest on how to debug
this issue? Any help on that issue is much appreciated. Where can I
file the appropriate bug and what information do I need to post from a
technical standpoint. I am rather competent with linux and the command
line so if i can be of some use researching this bug please let me
know what I can do to help

I am also using the ath9k module on my desktop pc with an atheros
AR5008 and the .30 kernel release and everything works as expected.
this problem seems specific to the ath5k module.

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

On Tue, Jun 30, 2009 at 6:50 PM, Jasin
Colegrove<j.wholesalesupply@gmail.com> wrote:
> I AM NOT A SUBSCRIBER. Please reply to me directly at this e-mail
> address j.wholesalesupply@gmail.com

No need to shout -- also common etiquette on Linux kernel lists is
to "reply-all" anyway :)

> Then I decided to see if it the card was working, So I tried to iwlist
> wlan0 scan, it returned the expected results.

> I would be glad to offer up as much
> information as I can, but I don't have the slightest on how to debug
> this issue?

So to summarize:

2.6.29 works, 2.6.30 doesn't, ath9k works, ath5k doesn't, userspace
setup is unchanged.  ath5k can get scan results (which means RX is
working at least some of the time) but dhcp does not.

You said you are comfortable with the command line, in that case the
easiest way to debug the change is to just bisect the ath5k changes
using git.  There were only 50 or so patches since 2.6.29 so you will
likely only need to compile the kernel 5 or 6 times.  You can do
something like:

$ git bisect start v2.6.30 v2.6.29 -- drivers/net/wireless/ath5k

> Any help on that issue is much appreciated. Where can I
> file the appropriate bug and what information do I need to post from a
> technical standpoint.

http://bugzilla.kernel.org/ is where you would file it, the information
you've already given here is a good start.

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

Thanks bob for the reply. I did not mean to shout, but someone else
told me that I should make it very clear that I was not subscribed.
But, now I am so it doesn't matter :)

I guess I will just wait to file a bug after I get through the bisect
process and test out the newest compat-wireless drivers. Then I should
have a more clear picture of the issue and maybe enough technical
details to file a proper bug report. Thanks again for your time,
greatly appreciated.

On Thu, Jul 2, 2009 at 11:36 AM, Bob Copeland<me@bobcopeland.com> wrote:
> On Tue, Jun 30, 2009 at 6:50 PM, Jasin
> Colegrove<j.wholesalesupply@gmail.com> wrote:
>> I AM NOT A SUBSCRIBER. Please reply to me directly at this e-mail
>> address j.wholesalesupply@gmail.com
>
> No need to shout -- also common etiquette on Linux kernel lists is
> to "reply-all" anyway :)
>
>> Then I decided to see if it the card was working, So I tried to iwlist
>> wlan0 scan, it returned the expected results.
>
>> I would be glad to offer up as much
>> information as I can, but I don't have the slightest on how to debug
>> this issue?
>
> So to summarize:
>
> 2.6.29 works, 2.6.30 doesn't, ath9k works, ath5k doesn't, userspace
> setup is unchanged.  ath5k can get scan results (which means RX is
> working at least some of the time) but dhcp does not.
>
> You said you are comfortable with the command line, in that case the
> easiest way to debug the change is to just bisect the ath5k changes
> using git.  There were only 50 or so patches since 2.6.29 so you will
> likely only need to compile the kernel 5 or 6 times.  You can do
> something like:
>
> $ git bisect start v2.6.30 v2.6.29 -- drivers/net/wireless/ath5k
>
>> Any help on that issue is much appreciated. Where can I
>> file the appropriate bug and what information do I need to post from a
>> technical standpoint.
>
> http://bugzilla.kernel.org/ is where you would file it, the information
> you've already given here is a good start.
>
> --
> Bob Copeland %% www.bobcopeland.com
>

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

I have found the bad commit.

$ git bisect good
1e3e6e8fe2f422f75619b1110f78bb638ae905b9 is first bad commit
commit 1e3e6e8fe2f422f75619b1110f78bb638ae905b9
Author: Nick Kossifidis <mick@madwifi-project.org>
Date:   Mon Feb 9 06:15:42 2009 +0200

    ath5k: Disable BMISS interrupts

     * Don't enable BMISS interrupts as we don't handle them yet

     Signed-off-by: Nick Kossifidis <mickflemm@gmail.com>

    Signed-off-by: John W. Linville <linville@tuxdriver.com>

:040000 040000 4a52fc89878d63cdb5da78e78d1ee17a6b5694c0
ff6f1d15f33d1dd3547e03fa191a779b50ad9ed4 M        drivers

Do you still think it would be wise to file a bug report for this
issue. I have yet to try the latest compat-wireless.

On Sun, Jul 5, 2009 at 8:13 AM, Jasin
Colegrove<j.wholesalesupply@gmail.com> wrote:
> Thanks bob for the reply. I did not mean to shout, but someone else
> told me that I should make it very clear that I was not subscribed.
> But, now I am so it doesn't matter :)
>
> I guess I will just wait to file a bug after I get through the bisect
> process and test out the newest compat-wireless drivers. Then I should
> have a more clear picture of the issue and maybe enough technical
> details to file a proper bug report. Thanks again for your time,
> greatly appreciated.
>
> On Thu, Jul 2, 2009 at 11:36 AM, Bob Copeland<me@bobcopeland.com> wrote:
>> On Tue, Jun 30, 2009 at 6:50 PM, Jasin
>> Colegrove<j.wholesalesupply@gmail.com> wrote:
>>> I AM NOT A SUBSCRIBER. Please reply to me directly at this e-mail
>>> address j.wholesalesupply@gmail.com
>>
>> No need to shout -- also common etiquette on Linux kernel lists is
>> to "reply-all" anyway :)
>>
>>> Then I decided to see if it the card was working, So I tried to iwlist
>>> wlan0 scan, it returned the expected results.
>>
>>> I would be glad to offer up as much
>>> information as I can, but I don't have the slightest on how to debug
>>> this issue?
>>
>> So to summarize:
>>
>> 2.6.29 works, 2.6.30 doesn't, ath9k works, ath5k doesn't, userspace
>> setup is unchanged.  ath5k can get scan results (which means RX is
>> working at least some of the time) but dhcp does not.
>>
>> You said you are comfortable with the command line, in that case the
>> easiest way to debug the change is to just bisect the ath5k changes
>> using git.  There were only 50 or so patches since 2.6.29 so you will
>> likely only need to compile the kernel 5 or 6 times.  You can do
>> something like:
>>
>> $ git bisect start v2.6.30 v2.6.29 -- drivers/net/wireless/ath5k
>>
>>> Any help on that issue is much appreciated. Where can I
>>> file the appropriate bug and what information do I need to post from a
>>> technical standpoint.
>>
>> http://bugzilla.kernel.org/ is where you would file it, the information
>> you've already given here is a good start.
>>
>> --
>> Bob Copeland %% www.bobcopeland.com
>>
>

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

> Colegrove<j.wholesalesupply@gmail.com> wrote:
> > Thanks bob for the reply. I did not mean to shout, but someone else
> > told me that I should make it very clear that I was not subscribed.
> > But, now I am so it doesn't matter :)

No worries, but whoever said that is wrong as far as vger.kernel.org
mailing lists go, as everyone just does reply-all.  (Also, top-posting
is frowned upon.)

> I have found the bad commit.

> $ git bisect good
> 1e3e6e8fe2f422f75619b1110f78bb638ae905b9 is first bad commit
> commit 1e3e6e8fe2f422f75619b1110f78bb638ae905b9
> Author: Nick Kossifidis <mick@madwifi-project.org>
> Date:   Mon Feb 9 06:15:42 2009 +0200
> 
>     ath5k: Disable BMISS interrupts

Hmm, so that's a pretty weird commit to end up with.  Can you
now try whatever your baseline kernel was, with just this patch
reverted and see if it works for you?

We don't actually do anything with BMISS yet so the only thing it
could do is slow down your computer.

Another possibility is that something happened between that
commit and your last good ath5k commit, elsewhere in the tree,
so a bisect would need to be done over the whole tree instead of
just the ath5k directory.

A third possibility is the issue comes and goes, or is timing
related, in which case bisection really isn't useful.

> Do you still think it would be wise to file a bug report for this
> issue. I have yet to try the latest compat-wireless.

Give 2.6.30.1 a shot as it includes a couple of fixes, otherwise
bugzilla is probably the best bet.

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

On Mon, Jul 6, 2009 at 11:07 PM, Bob Copeland<me@bobcopeland.com> wrote:
>> Colegrove<j.wholesalesupply@gmail.com> wrote:
>> > Thanks bob for the reply. I did not mean to shout, but someone else
>> > told me that I should make it very clear that I was not subscribed.
>> > But, now I am so it doesn't matter :)
>
> No worries, but whoever said that is wrong as far as vger.kernel.org
> mailing lists go, as everyone just does reply-all.  (Also, top-posting
> is frowned upon.)
>
>> I have found the bad commit.
>
>> $ git bisect good
>> 1e3e6e8fe2f422f75619b1110f78bb638ae905b9 is first bad commit
>> commit 1e3e6e8fe2f422f75619b1110f78bb638ae905b9
>> Author: Nick Kossifidis <mick@madwifi-project.org>
>> Date:   Mon Feb 9 06:15:42 2009 +0200
>>
>>     ath5k: Disable BMISS interrupts
>
> Hmm, so that's a pretty weird commit to end up with.  Can you
> now try whatever your baseline kernel was, with just this patch
> reverted and see if it works for you?

I haven't tried this yet, but I started up gitk and realized that the
last good commit was the one previous to the bad commit. They where
made 3 minutes apart in the ath5k dir. So I am assuming that the
"Disable BMISS interrupts was not actually at fault. It just doesn't
seem logical now that i look at it.
>
> We don't actually do anything with BMISS yet so the only thing it
> could do is slow down your computer.
>
> Another possibility is that something happened between that
> commit and your last good ath5k commit, elsewhere in the tree,
> so a bisect would need to be done over the whole tree instead of
> just the ath5k directory.

So I started another bisect of drivers/net/. Is this the main tree you
where talking about or do you think I need to go further into the main
tree? If so, how far?
>
> A third possibility is the issue comes and goes, or is timing
> related, in which case bisection really isn't useful.

I don't think this is the case, Only because I haven't noticed it to
be intermittent. Either wicd works on reboot of a new kernel or it
doesn't. Any attempts to get it working fail until I restart with a
good commit of the kernel.
>
>> Do you still think it would be wise to file a bug report for this
>> issue. I have yet to try the latest compat-wireless.
>
> Give 2.6.30.1 a shot as it includes a couple of fixes, otherwise
> bugzilla is probably the best bet.

I tried this and this is where the mystery deepens. With .30.1 it
seems that I can connect to my neighbors network using dhcp but I
still can not hook to my own using a static IP. I have not actually
tried conneting with .30.1/dhcp to my own network, but that is the
next step for me if the bisect on drivers/net fails to provide any
stellar results. Once again, downgrading to kernel .29 fixes
everything without changing any of my other settings. I am still 110%
positive this is a kernel module/driver issue. It does make sense, no?
>
> --
> Bob Copeland %% www.bobcopeland.com
>
>

If you have any other bright ideas keep them coming.

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

On Wed, Jul 08, 2009 at 02:58:11AM -0400, Jasin Colegrove wrote:
> I haven't tried this yet, but I started up gitk and realized that the
> last good commit was the one previous to the bad commit. They where
> made 3 minutes apart in the ath5k dir. So I am assuming that the
> "Disable BMISS interrupts was not actually at fault. It just doesn't
> seem logical now that i look at it.

Yeah, that commit is near some changes that might cause what you
are seeing though, so maybe you made a wrong turn along the way?
E.g. "ath5k: Update reset code" introduced a bug for RF 5413 chips,
maybe others.  Another possible change is the txpower rework.

BTW, can you (re?)post the "Atheros AR5212 chip found (MAC xxx,
PHY xxxx)?"

> So I started another bisect of drivers/net/. Is this the main tree you
> where talking about or do you think I need to go further into the main
> tree? If so, how far?

Also at least /net (you can add multiple paths to git bisect).  If 
you can narrow it down to a small range then a bisect over the entire
tree would eliminate any doubt, but in that case it's easier to make a
bisect mistake.  As you say, it does sound like a driver issue though.

> stellar results. Once again, downgrading to kernel .29 fixes
> everything without changing any of my other settings. I am still 110%
> positive this is a kernel module/driver issue. It does make sense, no?

Yeah, unfortunately I don't really have any good ideas, but I've seen
a trickle of similar reports so it would be good to nail it down.  Thanks
for testing, it really helps a lot.

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

> BTW, can you (re?)post the "Atheros AR5212 chip found (MAC xxx,
> PHY xxxx)?"
>
What exact information do you want and where can I find it?


> Also at least /net (you can add multiple paths to git bisect).

$git bisect start v2.6.30-rc4 v2.6.29 -- drivers/net net/

Is that what you where hinting at? I wasted a whole day trying to fix
errors in drivers/staging before I was told I could just disable
compilation of staging with one switch. Oh well...

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

On Thu, Jul 9, 2009 at 12:42 PM, Jasin
Colegrove<j.wholesalesupply@gmail.com> wrote:
>> BTW, can you (re?)post the "Atheros AR5212 chip found (MAC xxx,
>> PHY xxxx)?"
>>
> What exact information do you want and where can I find it?

Run 'dmesg'; a string like the above is printed out when the module
loads.  In case there was something broken specifically for your
chipset, the mac/phy versions will help narrow it down.

> $git bisect start v2.6.30-rc4 v2.6.29 -- drivers/net net/
>
> Is that what you where hinting at? I wasted a whole day trying to fix
> errors in drivers/staging before I was told I could just disable
> compilation of staging with one switch. Oh well...

Yes, that's what I meant.  And it does help to have a relatively
slimmed down .config, especially when bisecting. :\

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Luis R. Rodriguez]

On Thu, Jul 9, 2009 at 9:55 AM, Bob Copeland<me@bobcopeland.com> wrote:
> On Thu, Jul 9, 2009 at 12:42 PM, Jasin
> Colegrove<j.wholesalesupply@gmail.com> wrote:
>>> BTW, can you (re?)post the "Atheros AR5212 chip found (MAC xxx,
>>> PHY xxxx)?"
>>>
>> What exact information do you want and where can I find it?
>
> Run 'dmesg'; a string like the above is printed out when the module
> loads.  In case there was something broken specifically for your
> chipset, the mac/phy versions will help narrow it down.
>
>> $git bisect start v2.6.30-rc4 v2.6.29 -- drivers/net net/
>>
>> Is that what you where hinting at? I wasted a whole day trying to fix
>> errors in drivers/staging before I was told I could just disable
>> compilation of staging with one switch. Oh well...
>
> Yes, that's what I meant.  And it does help to have a relatively
> slimmed down .config, especially when bisecting. :\

I'd add net/mac80211 and net/wireless in there as well, just to be safe.

 Luis

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

On Thu, Jul 9, 2009 at 1:02 PM, Luis R. Rodriguez<mcgrof@gmail.com> wrote:

>>> $git bisect start v2.6.30-rc4 v2.6.29 -- drivers/net net/

> I'd add net/mac80211 and net/wireless in there as well, just to be safe.

net/ was already in there :)

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Luis R. Rodriguez]

On Thu, Jul 9, 2009 at 10:48 AM, Bob Copeland<me@bobcopeland.com> wrote:
> On Thu, Jul 9, 2009 at 1:02 PM, Luis R. Rodriguez<mcgrof@gmail.com> wrote:
>
>>>> $git bisect start v2.6.30-rc4 v2.6.29 -- drivers/net net/
>
>> I'd add net/mac80211 and net/wireless in there as well, just to be safe.
>
> net/ was already in there :)

Ah yes :)

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

I got one more question and it's driving me mad. I am really confused...

When running git bisect it keeps building 2.6.29-rc kernels. I thought
v2.6.29 was in the future compared to 2.6.29-rc7 for example.
running the command git bisect start v2.6.29-rc7 v2.6.29 disagrees
though and tells me.

$ git bisect start v2.6.29-rc7 v2.6.29
Some good revs are not ancestor of the bad rev.
git bisect cannot work properly in this case.
Maybe you mistake good and bad revs?

But, when I run the command

$ git bisect start v2.6.30 v2.6.29

It co's a commit somewhere around 2.6.29-rc7 which ultimately was a
bad commit. Am I misunderstanding how this all works? Sorry to be a
pain

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Bob Copeland]

On Thu, Jul 9, 2009 at 2:36 PM, Jasin
Colegrove<j.wholesalesupply@gmail.com> wrote:
> When running git bisect it keeps building 2.6.29-rc kernels. I thought
> v2.6.29 was in the future compared to 2.6.29-rc7 for example.
> running the command git bisect start v2.6.29-rc7 v2.6.29 disagrees
> though and tells me.
>
> It co's a commit somewhere around 2.6.29-rc7 which ultimately was a
> bad commit. Am I misunderstanding how this all works? Sorry to be a
> pain

That confuses everyone the first time.  Basically the way it works
is that a commit that was authored before 2.6.29 will still be checked
since it was only merged in 2.6.29.

         +--- B -- D -+
        /              \
    -- A -- C ---- E -- F -- G

  e.g. in the above graph, imagine E is 2.6.29 and G is 2.6.30.  Commits
B and D were written from a base of 2.6.29-rc6 (say) but merged at time
F between 2.6.29 and 2.6.30.  The bisect will check both parents of F if
it needs to since that's when it actually hit the mainline.

-- 
Bob Copeland %% www.bobcopeland.com

Re: kernel .30 BROKE ATH5K with my AR5212 atheros.

[Jasin Colegrove]

Just a quick update on this issue. I have come to the conclusion that
is indeed NOT a kernel driver issue. After removing the cruft and just
going the manual setup route for the complete bisect I was never once
able to repeat the behavior I had expirienced before. That was not
till i tried using wicd again. I have no reported this problem to the
appropriate bug tracker.

Sorry for wasting your precious time, I really do appreciate all your
responses and hard work!