[ath9k-devel] still same problem

[ath9k-devel] still same problem

[Dmitri Seletski]

Hi guys.
After quite some time has passed since my last message, i have pulled 
wireless drivers from web site several times, still same problem,  
refresh of drivers didn't seem to affect the problem at all.
I am lost, since you don't  seem to take it seriously, but it hangs 
machine! Can someone give me an advice what I can do in order to help 
you to troubleshoot the issue?
I am not a programmer, but looks like a memory leak somewhere. If you 
give clear understandable instructions - I am glad to follow them!
Regards

[ath9k-devel] still same problem

[Barry Green]

Hi Dimitri,

I"m getting a similar problem to you, in that my computer is hanging 
shortly after loading ath9k and trying to set up a wireless connection. 
I've logged a bug report here:

http://bugzilla.kernel.org/show_bug.cgi?id=12110

but unfortunately, there hasn't been any progress or feedback on my 
comments since the 6th December.  I'm willing to test, but I need some 
guidance on exactly how to get some useful debugging information.

Barry.

Dmitri Seletski wrote:
> Hi guys.
> After quite some time has passed since my last message, i have pulled 
> wireless drivers from web site several times, still same problem,  
> refresh of drivers didn't seem to affect the problem at all.
> I am lost, since you don't  seem to take it seriously, but it hangs 
> machine! Can someone give me an advice what I can do in order to help 
> you to troubleshoot the issue?
> I am not a programmer, but looks like a memory leak somewhere. If you 
> give clear understandable instructions - I am glad to follow them!
> Regards
> _______________________________________________
> ath9k-devel mailing list
> ath9k-devel at lists.ath9k.org
> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
> 

[ath9k-devel] still same problem

[Luis R. Rodriguez]

On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
> Hi Dimitri,
>
> I"m getting a similar problem to you, in that my computer is hanging
> shortly after loading ath9k and trying to set up a wireless connection.
> I've logged a bug report here:
>
> http://bugzilla.kernel.org/show_bug.cgi?id=12110

This is for 2.6.27, the issue being reported here is for
wireless-testing. Regardless we don't have any oops messages for
either issues.

> but unfortunately, there hasn't been any progress or feedback on my
> comments since the 6th December.  I'm willing to test, but I need some
> guidance on exactly how to get some useful debugging information.

We need an oops trace of your crashes so far it seems you have not
been able to get one, lets work on that. Until then it is not easy for
us to provide you with patches to fix your issues.

Please test with the latest drivers from
wireless-testing/compat-wireless and restart in single user mode, we
want to ensure X is not running at all and you are in a virtual
terminal. We also do not want Network Manager or wpa_supplicant
running in the background. To increase the resolution upon bootup try
booting with vga=ask (vga=791 works well for me), then try to
associate to your access point without encryption manually using
iwconfig.

If there is an oops it should hopefully come up.

  Luis

[ath9k-devel] still same problem

[Jens Ott - Opteamax]

Hi,

Luis R. Rodriguez schrieb:
> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
> [...]
> We need an oops trace of your crashes so far it seems you have not
> been able to get one, lets work on that. Until then it is not easy for
> us to provide you with patches to fix your issues.

I also have similar problems with an D-LINK DWA-547 and I'd really like 
to provide the oops trace, but unfortunately the machine seems to hang 
that fast, that even no oops trace is shown. Also magic-sysreq don't 
work anymore, neither via serial console nor via keyboard. Only chance I 
have in this moment is to press reset-button, which already caused a 
full xfs-filsystem-corruption with dataloss once :-(

My problem is, that the hang happens not immediately after loading the 
driver but "somewhen" while it is loaded.

I have seen two types of crash:

1.) first investigations on my side seem to show that it happens, as 
soon as there is high dataload transfered, but I don't have detailed 
info yet.
2.) Running hostapd works fine while started first time, but no client 
is connected. When a client connects a system freeze occures 
"sometimes". Reproducably the system freezes immediately when starting 
hostapd, stop it and then restart hostapd. So every change in hostapd 
conf needs a full reboot, unloading the ath9k module and reload it does 
not prevent freeze, but after reboot I can start hostapd once without a 
problem (only the metioned "sometime" freezes on client connect or high 
data-volume).

As this machine is productive, I suspended testing for now. I will 
change the card into a non-productive environment in the next days. As 
soon as I find more details, I'll report.

> 
> Please test with the latest drivers from
> wireless-testing/compat-wireless and restart in single user mode, we
> want to ensure X is not running at all and you are in a virtual
> terminal. We also do not want Network Manager or wpa_supplicant
> running in the background. To increase the resolution upon bootup try
> booting with vga=ask (vga=791 works well for me), then try to
> associate to your access point without encryption manually using
> iwconfig.

Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant 
is running.

Best regards
Jens

> 
> If there is an oops it should hopefully come up.
> 
>   Luis
> _______________________________________________
> ath9k-devel mailing list
> ath9k-devel at lists.ath9k.org
> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
> 
> !DSPAM:2,49602e56281331804284693!
> 

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 04 January 2009, Dmitri Seletski wrote:
> Hi guys.
> After quite some time has passed since my last message, i have pulled
> wireless drivers from web site several times, still same problem,
> refresh of drivers didn't seem to affect the problem at all.
> I am lost, since you don't  seem to take it seriously, but it hangs
> machine! Can someone give me an advice what I can do in order to help
> you to troubleshoot the issue?
> I am not a programmer, but looks like a memory leak somewhere. If you
> give clear understandable instructions - I am glad to follow them!
> Regards

I am facing the same problems. Kernel hangups without any oops or log-trace.
Sometimes it freezes directly after hostapd (0.6.6) is loading. But once 
running hostapd I can connect with one station to the AP (AR5416). A second
station can be added but sometimes it causes a hangup.

For what I can see the problems may also have to do with the total number of
sessions to the AP or the number of sessions setup in a short time.

If I for example start my notebook and connect to the AP all is (mostly) ok.
I can start a browser. But if I start Kmail with 6 mailboxes it hangs within a 
second.

I am desperate looking for a solution.....


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 04 January 2009, Dmitri Seletski wrote:
> Hi guys.
> After quite some time has passed since my last message, i have pulled
> wireless drivers from web site several times, still same problem,
> refresh of drivers didn't seem to affect the problem at all.
> I am lost, since you don't  seem to take it seriously, but it hangs
> machine! Can someone give me an advice what I can do in order to help
> you to troubleshoot the issue?
> I am not a programmer, but looks like a memory leak somewhere. If you
> give clear understandable instructions - I am glad to follow them!
> Regards

I am facing the same problems. Kernel hangups without any oops or
log-trace. Sometimes it freezes directly after hostapd (0.6.6) is loading.
But once running hostapd I can connect with one station to the AP (AR5416).
A second station can be added but sometimes it causes a hangup.

For what I can see the problems may also have to do with the total number
of sessions to the AP or the number of sessions setup in a short time.

If I for example start my notebook and connect to the AP all is (mostly)
ok. I can start a browser. But if I start Kmail with 6 mailboxes it hangs
within a second.

I am desperate looking for a solution.....

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Luis R. Rodriguez]

On Sun, Jan 4, 2009 at 1:58 AM, Jens Ott - Opteamax <jo@opteamax.de> wrote:
> Hi,
>
> Luis R. Rodriguez schrieb:
>>
>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
>> [...]
>> We need an oops trace of your crashes so far it seems you have not
>> been able to get one, lets work on that. Until then it is not easy for
>> us to provide you with patches to fix your issues.
>
> I also have similar problems with an D-LINK DWA-547 and I'd really like to
> provide the oops trace, but unfortunately the machine seems to hang that
> fast, that even no oops trace is shown. Also magic-sysreq don't work
> anymore, neither via serial console nor via keyboard. Only chance I have in
> this moment is to press reset-button, which already caused a full
> xfs-filsystem-corruption with dataloss once :-(

What kernel are you using? Please note that 2.6.27 had some DMA issues
which were fixed since 2.6.27.8. So if you are using an older kenrel
than 2.6.27.8 you need to upgrade.

If you are using compat-wireless drivers then that's the latest and greatest.

> My problem is, that the hang happens not immediately after loading the
> driver but "somewhen" while it is loaded.
>
> I have seen two types of crash:
>
> 1.) first investigations on my side seem to show that it happens, as soon as
> there is high dataload transfered, but I don't have detailed info yet.
> 2.) Running hostapd works fine while started first time, but no client is
> connected. When a client connects a system freeze occures "sometimes".
> Reproducably the system freezes immediately when starting hostapd, stop it
> and then restart hostapd. So every change in hostapd conf needs a full
> reboot, unloading the ath9k module and reload it does not prevent freeze,
> but after reboot I can start hostapd once without a problem (only the
> metioned "sometime" freezes on client connect or high data-volume).
>
> As this machine is productive, I suspended testing for now. I will change
> the card into a non-productive environment in the next days. As soon as I
> find more details, I'll report.

I'll reply to the thread with some more suggestions as to what can be
done to help us debug this issue in another e-mail, please see that.
You may be able to help as you can reproduce these issues easily it
seems.

  Luis

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> Hi,
>
> Luis R. Rodriguez schrieb:
> > On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
> > [...]
> > We need an oops trace of your crashes so far it seems you have not
> > been able to get one, lets work on that. Until then it is not easy for
> > us to provide you with patches to fix your issues.
>
> I also have similar problems with an D-LINK DWA-547 and I'd really like
> to provide the oops trace, but unfortunately the machine seems to hang
> that fast, that even no oops trace is shown. Also magic-sysreq don't
> work anymore, neither via serial console nor via keyboard. Only chance I
> have in this moment is to press reset-button, which already caused a
> full xfs-filsystem-corruption with dataloss once :-(
>
> My problem is, that the hang happens not immediately after loading the
> driver but "somewhen" while it is loaded.
>
> I have seen two types of crash:
>
> 1.) first investigations on my side seem to show that it happens, as
> soon as there is high dataload transfered, but I don't have detailed
> info yet.
> 2.) Running hostapd works fine while started first time, but no client
> is connected. When a client connects a system freeze occures
> "sometimes". Reproducably the system freezes immediately when starting
> hostapd, stop it and then restart hostapd. So every change in hostapd
> conf needs a full reboot, unloading the ath9k module and reload it does
> not prevent freeze, but after reboot I can start hostapd once without a
> problem (only the metioned "sometime" freezes on client connect or high
> data-volume).
>
> As this machine is productive, I suspended testing for now. I will
> change the card into a non-productive environment in the next days. As
> soon as I find more details, I'll report.
>
> > Please test with the latest drivers from
> > wireless-testing/compat-wireless and restart in single user mode, we
> > want to ensure X is not running at all and you are in a virtual
> > terminal. We also do not want Network Manager or wpa_supplicant
> > running in the background. To increase the resolution upon bootup try
> > booting with vga=ask (vga=791 works well for me), then try to
> > associate to your access point without encryption manually using
> > iwconfig.
>
> Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant
> is running.
>

Hi,

What is the status of your problem? I still have problems with it and have no
progress...

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Jens Ott - Opteamax]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I did not find the time to further investigate this issue. Well what I
did today was compile the latest git-update and the latest hostapd. At
least the machine does not crash that often, but with 802.11n I can't
connect to the card ... but this might also be a problem with the card
in my laptop. I hope I'll find the time to go on testing next week.

Best regards
Jens

W. van den Akker schrieb:
> On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
>> Hi,
>>
>> Luis R. Rodriguez schrieb:
>>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
>>> [...]
>>> We need an oops trace of your crashes so far it seems you have not
>>> been able to get one, lets work on that. Until then it is not easy for
>>> us to provide you with patches to fix your issues.
>> I also have similar problems with an D-LINK DWA-547 and I'd really like
>> to provide the oops trace, but unfortunately the machine seems to hang
>> that fast, that even no oops trace is shown. Also magic-sysreq don't
>> work anymore, neither via serial console nor via keyboard. Only chance I
>> have in this moment is to press reset-button, which already caused a
>> full xfs-filsystem-corruption with dataloss once :-(
>>
>> My problem is, that the hang happens not immediately after loading the
>> driver but "somewhen" while it is loaded.
>>
>> I have seen two types of crash:
>>
>> 1.) first investigations on my side seem to show that it happens, as
>> soon as there is high dataload transfered, but I don't have detailed
>> info yet.
>> 2.) Running hostapd works fine while started first time, but no client
>> is connected. When a client connects a system freeze occures
>> "sometimes". Reproducably the system freezes immediately when starting
>> hostapd, stop it and then restart hostapd. So every change in hostapd
>> conf needs a full reboot, unloading the ath9k module and reload it does
>> not prevent freeze, but after reboot I can start hostapd once without a
>> problem (only the metioned "sometime" freezes on client connect or high
>> data-volume).
>>
>> As this machine is productive, I suspended testing for now. I will
>> change the card into a non-productive environment in the next days. As
>> soon as I find more details, I'll report.
>>
>>> Please test with the latest drivers from
>>> wireless-testing/compat-wireless and restart in single user mode, we
>>> want to ensure X is not running at all and you are in a virtual
>>> terminal. We also do not want Network Manager or wpa_supplicant
>>> running in the background. To increase the resolution upon bootup try
>>> booting with vga=ask (vga=791 works well for me), then try to
>>> associate to your access point without encryption manually using
>>> iwconfig.
>> Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant
>> is running.
>>
> 
> Hi,
> 
> What is the status of your problem? I still have problems with it and have no
> progress...
> 
> gr,
> Willem
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAklyiMoACgkQMf0yjMLKfXpcewCfWW65ZXnBL2tuuuBYAGWLlb+g
u7YAn1TrkV80Czyy1GTKqxiLeShnE+KL
=0wWG
-----END PGP SIGNATURE-----

[ath9k-devel] still same problem

[Dmitri Seletski]

W. van den Akker wrote:
> On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
>   
>> Hi,
>>
>> Luis R. Rodriguez schrieb:
>>     
>>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
>>> [...]
>>> We need an oops trace of your crashes so far it seems you have not
>>> been able to get one, lets work on that. Until then it is not easy for
>>> us to provide you with patches to fix your issues.
>>>       
>> I also have similar problems with an D-LINK DWA-547 and I'd really like
>> to provide the oops trace, but unfortunately the machine seems to hang
>> that fast, that even no oops trace is shown. Also magic-sysreq don't
>> work anymore, neither via serial console nor via keyboard. Only chance I
>> have in this moment is to press reset-button, which already caused a
>> full xfs-filsystem-corruption with dataloss once :-(
>>
>> My problem is, that the hang happens not immediately after loading the
>> driver but "somewhen" while it is loaded.
>>
>> I have seen two types of crash:
>>
>> 1.) first investigations on my side seem to show that it happens, as
>> soon as there is high dataload transfered, but I don't have detailed
>> info yet.
>> 2.) Running hostapd works fine while started first time, but no client
>> is connected. When a client connects a system freeze occures
>> "sometimes". Reproducably the system freezes immediately when starting
>> hostapd, stop it and then restart hostapd. So every change in hostapd
>> conf needs a full reboot, unloading the ath9k module and reload it does
>> not prevent freeze, but after reboot I can start hostapd once without a
>> problem (only the metioned "sometime" freezes on client connect or high
>> data-volume).
>>
>> As this machine is productive, I suspended testing for now. I will
>> change the card into a non-productive environment in the next days. As
>> soon as I find more details, I'll report.
>>
>>     
>>> Please test with the latest drivers from
>>> wireless-testing/compat-wireless and restart in single user mode, we
>>> want to ensure X is not running at all and you are in a virtual
>>> terminal. We also do not want Network Manager or wpa_supplicant
>>> running in the background. To increase the resolution upon bootup try
>>> booting with vga=ask (vga=791 works well for me), then try to
>>> associate to your access point without encryption manually using
>>> iwconfig.
>>>       
>> Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant
>> is running.
>>
>>     
>
> Hi,
>
> What is the status of your problem? I still have problems with it and have no
> progress...
>
> gr,
> Willem
>
>   
Hello Willem.

I think I have had similar issue to yours.

Car was working for short period of time and after a while crashed computer.
I was advised to change kernel version, i updated it to 2.6.27.9 and 
problem has gone away since.
It's something to do with DMA.

Perhaps try to upgrade kernel?

Regards
Dmitri

Re: [ath9k-devel] still same problem

[W. van den Akker]

On Sunday 18 January 2009, you wrote:
> Hi,
>
> I did not find the time to further investigate this issue. Well what I
> did today was compile the latest git-update and the latest hostapd. At
> least the machine does not crash that often, but with 802.11n I can't
> connect to the card ... but this might also be a problem with the card
> in my laptop. I hope I'll find the time to go on testing next week.
>

I have the same problem with my laptop. Cant make a connection if 802.11n
is enabled. I have tested with hostapd 0.6.7 and I didnt notice any changes.

Let me know if you have any progress.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 18 January 2009, Dmitri Seletski wrote:
> W. van den Akker wrote:
> > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> >> Hi,
> >>
> >> Luis R. Rodriguez schrieb:
> >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
> >>> [...]
> >>> We need an oops trace of your crashes so far it seems you have not
> >>> been able to get one, lets work on that. Until then it is not easy for
> >>> us to provide you with patches to fix your issues.
> >>
> >> I also have similar problems with an D-LINK DWA-547 and I'd really like
> >> to provide the oops trace, but unfortunately the machine seems to hang
> >> that fast, that even no oops trace is shown. Also magic-sysreq don't
> >> work anymore, neither via serial console nor via keyboard. Only chance I
> >> have in this moment is to press reset-button, which already caused a
> >> full xfs-filsystem-corruption with dataloss once :-(
> >>
> >> My problem is, that the hang happens not immediately after loading the
> >> driver but "somewhen" while it is loaded.
> >>
> >> I have seen two types of crash:
> >>
> >> 1.) first investigations on my side seem to show that it happens, as
> >> soon as there is high dataload transfered, but I don't have detailed
> >> info yet.
> >> 2.) Running hostapd works fine while started first time, but no client
> >> is connected. When a client connects a system freeze occures
> >> "sometimes". Reproducably the system freezes immediately when starting
> >> hostapd, stop it and then restart hostapd. So every change in hostapd
> >> conf needs a full reboot, unloading the ath9k module and reload it does
> >> not prevent freeze, but after reboot I can start hostapd once without a
> >> problem (only the metioned "sometime" freezes on client connect or high
> >> data-volume).
> >>
> >> As this machine is productive, I suspended testing for now. I will
> >> change the card into a non-productive environment in the next days. As
> >> soon as I find more details, I'll report.
> >>
> >>> Please test with the latest drivers from
> >>> wireless-testing/compat-wireless and restart in single user mode, we
> >>> want to ensure X is not running at all and you are in a virtual
> >>> terminal. We also do not want Network Manager or wpa_supplicant
> >>> running in the background. To increase the resolution upon bootup try
> >>> booting with vga=ask (vga=791 works well for me), then try to
> >>> associate to your access point without encryption manually using
> >>> iwconfig.
> >>
> >> Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant
> >> is running.
> >
> > Hi,
> >
> > What is the status of your problem? I still have problems with it and
> > have no progress...
> >
> > gr,
> > Willem
>
> Hello Willem.
>
> I think I have had similar issue to yours.
>
> Car was working for short period of time and after a while crashed
> computer. I was advised to change kernel version, i updated it to 2.6.27.9
> and problem has gone away since.
> It's something to do with DMA.
>
> Perhaps try to upgrade kernel?
>
> Regards
> Dmitri

I have tried all kernel versions till 2.6.28 and still have the same hang-ups.
I think indeed its a DMA-hardware-ath9k related problem. ?The patch in 
2.6.27.8 didnt work for me. I have filled a bugreport on bugzilla.kernel but 
because AP isnt supported untill 2.6.29 the report is closed.

I will test today 2.6.29.rc2 and see if it has any progress.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Luis R. Rodriguez]

On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
> On Sunday 18 January 2009, Dmitri Seletski wrote:
> > W. van den Akker wrote:
> > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> > >> Hi,
> > >>
> > >> Luis R. Rodriguez schrieb:
> > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net> wrote:
> > >>> [...]
> > >>> We need an oops trace of your crashes so far it seems you have not
> > >>> been able to get one, lets work on that. Until then it is not easy for
> > >>> us to provide you with patches to fix your issues.
> > >>
> > >> I also have similar problems with an D-LINK DWA-547 and I'd really like
> > >> to provide the oops trace, but unfortunately the machine seems to hang
> > >> that fast, that even no oops trace is shown. Also magic-sysreq don't
> > >> work anymore, neither via serial console nor via keyboard. Only chance I
> > >> have in this moment is to press reset-button, which already caused a
> > >> full xfs-filsystem-corruption with dataloss once :-(
> > >>
> > >> My problem is, that the hang happens not immediately after loading the
> > >> driver but "somewhen" while it is loaded.
> > >>
> > >> I have seen two types of crash:
> > >>
> > >> 1.) first investigations on my side seem to show that it happens, as
> > >> soon as there is high dataload transfered, but I don't have detailed
> > >> info yet.
> > >> 2.) Running hostapd works fine while started first time, but no client
> > >> is connected. When a client connects a system freeze occures
> > >> "sometimes". Reproducably the system freezes immediately when starting
> > >> hostapd, stop it and then restart hostapd. So every change in hostapd
> > >> conf needs a full reboot, unloading the ath9k module and reload it does
> > >> not prevent freeze, but after reboot I can start hostapd once without a
> > >> problem (only the metioned "sometime" freezes on client connect or high
> > >> data-volume).
> > >>
> > >> As this machine is productive, I suspended testing for now. I will
> > >> change the card into a non-productive environment in the next days. As
> > >> soon as I find more details, I'll report.
> > >>
> > >>> Please test with the latest drivers from
> > >>> wireless-testing/compat-wireless and restart in single user mode, we
> > >>> want to ensure X is not running at all and you are in a virtual
> > >>> terminal. We also do not want Network Manager or wpa_supplicant
> > >>> running in the background. To increase the resolution upon bootup try
> > >>> booting with vga=ask (vga=791 works well for me), then try to
> > >>> associate to your access point without encryption manually using
> > >>> iwconfig.
> > >>
> > >> Ok, that'll be a thing I'd need to test to, as in my case wpa_supplicant
> > >> is running.
> > >
> > > Hi,
> > >
> > > What is the status of your problem? I still have problems with it and
> > > have no progress...
> > >
> > > gr,
> > > Willem
> >
> > Hello Willem.
> >
> > I think I have had similar issue to yours.
> >
> > Car was working for short period of time and after a while crashed
> > computer. I was advised to change kernel version, i updated it to 2.6.27.9
> > and problem has gone away since.
> > It's something to do with DMA.
> >
> > Perhaps try to upgrade kernel?
> >
> > Regards
> > Dmitri
> 
> I have tried all kernel versions till 2.6.28 and still have the same hang-ups.
> I think indeed its a DMA-hardware-ath9k related problem.  The patch in
> 2.6.27.8 didnt work for me. I have filled a bugreport on bugzilla.kernel but
> because AP isnt supported untill 2.6.29 the report is closed.

A better description of this is that AP support does not work (is not allowed)
and therefore not supported on < 2.6.28.

> I will test today 2.6.29.rc2 and see if it has any progress.

You can also try wireless-testing (entire kernel) or compat-wireless (just updates
your wireless stuff) for bleeding edge stuff.

http://wireless.kernel.org/en/developers/Documentation/git-guide
http://wireless.kernel.org/en/users/Download

  Luis

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 18 January 2009, Luis R. Rodriguez wrote:
> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
> > On Sunday 18 January 2009, Dmitri Seletski wrote:
> > > W. van den Akker wrote:
> > > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> > > >> Hi,
> > > >>
> > > >> Luis R. Rodriguez schrieb:
> > > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net>
> > > >>> wrote: [...]
> > > >>> We need an oops trace of your crashes so far it seems you have not
> > > >>> been able to get one, lets work on that. Until then it is not easy
> > > >>> for us to provide you with patches to fix your issues.
> > > >>
> > > >> I also have similar problems with an D-LINK DWA-547 and I'd really
> > > >> like to provide the oops trace, but unfortunately the machine seems
> > > >> to hang that fast, that even no oops trace is shown. Also
> > > >> magic-sysreq don't work anymore, neither via serial console nor via
> > > >> keyboard. Only chance I have in this moment is to press
> > > >> reset-button, which already caused a full xfs-filsystem-corruption
> > > >> with dataloss once :-(
> > > >>
> > > >> My problem is, that the hang happens not immediately after loading
> > > >> the driver but "somewhen" while it is loaded.
> > > >>
> > > >> I have seen two types of crash:
> > > >>
> > > >> 1.) first investigations on my side seem to show that it happens, as
> > > >> soon as there is high dataload transfered, but I don't have detailed
> > > >> info yet.
> > > >> 2.) Running hostapd works fine while started first time, but no
> > > >> client is connected. When a client connects a system freeze occures
> > > >> "sometimes". Reproducably the system freezes immediately when
> > > >> starting hostapd, stop it and then restart hostapd. So every change
> > > >> in hostapd conf needs a full reboot, unloading the ath9k module and
> > > >> reload it does not prevent freeze, but after reboot I can start
> > > >> hostapd once without a problem (only the metioned "sometime" freezes
> > > >> on client connect or high data-volume).
> > > >>
> > > >> As this machine is productive, I suspended testing for now. I will
> > > >> change the card into a non-productive environment in the next days.
> > > >> As soon as I find more details, I'll report.
> > > >>
> > > >>> Please test with the latest drivers from
> > > >>> wireless-testing/compat-wireless and restart in single user mode,
> > > >>> we want to ensure X is not running at all and you are in a virtual
> > > >>> terminal. We also do not want Network Manager or wpa_supplicant
> > > >>> running in the background. To increase the resolution upon bootup
> > > >>> try booting with vga=ask (vga=791 works well for me), then try to
> > > >>> associate to your access point without encryption manually using
> > > >>> iwconfig.
> > > >>
> > > >> Ok, that'll be a thing I'd need to test to, as in my case
> > > >> wpa_supplicant is running.
> > > >
> > > > Hi,
> > > >
> > > > What is the status of your problem? I still have problems with it and
> > > > have no progress...
> > > >
> > > > gr,
> > > > Willem
> > >
> > > Hello Willem.
> > >
> > > I think I have had similar issue to yours.
> > >
> > > Car was working for short period of time and after a while crashed
> > > computer. I was advised to change kernel version, i updated it to
> > > 2.6.27.9 and problem has gone away since.
> > > It's something to do with DMA.
> > >
> > > Perhaps try to upgrade kernel?
> > >
> > > Regards
> > > Dmitri
> >
> > I have tried all kernel versions till 2.6.28 and still have the same
> > hang-ups. I think indeed its a DMA-hardware-ath9k related problem.  The
> > patch in 2.6.27.8 didnt work for me. I have filled a bugreport on
> > bugzilla.kernel but because AP isnt supported untill 2.6.29 the report is
> > closed.
>
> A better description of this is that AP support does not work (is not
> allowed) and therefore not supported on < 2.6.28.
>
> > I will test today 2.6.29.rc2 and see if it has any progress.
>
> You can also try wireless-testing (entire kernel) or compat-wireless (just
> updates your wireless stuff) for bleeding edge stuff.
>
> http://wireless.kernel.org/en/developers/Documentation/git-guide
> http://wireless.kernel.org/en/users/Download
>
>   Luis

I have tried all kernel versions till 2.6.28 and still have the same hang-ups.
I think indeed its a DMA-hardware-ath9k related problem. ?The patch in 
2.6.27.8 didnt work for me. I have filled a bugreport on bugzilla.kernel but 
because AP isnt supported untill 2.6.29 the report is closed.

I will test today 2.6.29.rc2 and see if it has any progress.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[W. van den Akker]

On Sunday 18 January 2009, Luis R. Rodriguez wrote:
> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
> > On Sunday 18 January 2009, Dmitri Seletski wrote:
> > > W. van den Akker wrote:
> > > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> > > >> Hi,
> > > >>
> > > >> Luis R. Rodriguez schrieb:
> > > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net>
> > > >>> wrote: [...]
> > > >>> We need an oops trace of your crashes so far it seems you have not
> > > >>> been able to get one, lets work on that. Until then it is not easy
> > > >>> for us to provide you with patches to fix your issues.
> > > >>
> > > >> I also have similar problems with an D-LINK DWA-547 and I'd really
> > > >> like to provide the oops trace, but unfortunately the machine seems
> > > >> to hang that fast, that even no oops trace is shown. Also
> > > >> magic-sysreq don't work anymore, neither via serial console nor via
> > > >> keyboard. Only chance I have in this moment is to press
> > > >> reset-button, which already caused a full xfs-filsystem-corruption
> > > >> with dataloss once :-(
> > > >>
> > > >> My problem is, that the hang happens not immediately after loading
> > > >> the driver but "somewhen" while it is loaded.
> > > >>
> > > >> I have seen two types of crash:
> > > >>
> > > >> 1.) first investigations on my side seem to show that it happens, as
> > > >> soon as there is high dataload transfered, but I don't have detailed
> > > >> info yet.
> > > >> 2.) Running hostapd works fine while started first time, but no
> > > >> client is connected. When a client connects a system freeze occures
> > > >> "sometimes". Reproducably the system freezes immediately when
> > > >> starting hostapd, stop it and then restart hostapd. So every change
> > > >> in hostapd conf needs a full reboot, unloading the ath9k module and
> > > >> reload it does not prevent freeze, but after reboot I can start
> > > >> hostapd once without a problem (only the metioned "sometime" freezes
> > > >> on client connect or high data-volume).
> > > >>
> > > >> As this machine is productive, I suspended testing for now. I will
> > > >> change the card into a non-productive environment in the next days.
> > > >> As soon as I find more details, I'll report.
> > > >>
> > > >>> Please test with the latest drivers from
> > > >>> wireless-testing/compat-wireless and restart in single user mode,
> > > >>> we want to ensure X is not running at all and you are in a virtual
> > > >>> terminal. We also do not want Network Manager or wpa_supplicant
> > > >>> running in the background. To increase the resolution upon bootup
> > > >>> try booting with vga=ask (vga=791 works well for me), then try to
> > > >>> associate to your access point without encryption manually using
> > > >>> iwconfig.
> > > >>
> > > >> Ok, that'll be a thing I'd need to test to, as in my case
> > > >> wpa_supplicant is running.
> > > >
> > > > Hi,
> > > >
> > > > What is the status of your problem? I still have problems with it and
> > > > have no progress...
> > > >
> > > > gr,
> > > > Willem
> > >
> > > Hello Willem.
> > >
> > > I think I have had similar issue to yours.
> > >
> > > Car was working for short period of time and after a while crashed
> > > computer. I was advised to change kernel version, i updated it to
> > > 2.6.27.9 and problem has gone away since.
> > > It's something to do with DMA.
> > >
> > > Perhaps try to upgrade kernel?
> > >
> > > Regards
> > > Dmitri
> >
> > I have tried all kernel versions till 2.6.28 and still have the same
> > hang-ups. I think indeed its a DMA-hardware-ath9k related problem.  The
> > patch in 2.6.27.8 didnt work for me. I have filled a bugreport on
> > bugzilla.kernel but because AP isnt supported untill 2.6.29 the report is
> > closed.
>
> A better description of this is that AP support does not work (is not
> allowed) and therefore not supported on < 2.6.28.
>
> > I will test today 2.6.29.rc2 and see if it has any progress.
>
> You can also try wireless-testing (entire kernel) or compat-wireless (just
> updates your wireless stuff) for bleeding edge stuff.
>

I tested it today with kernel 2.6.29-rc2. But no change. The kernel hangs 
completly (without oops or what so ever, even nmi-watchdog has no results).
It happend when I started the connection (from a IW4965 card). Then started up
a Kmail connection, start up a browser sessie. Stop Kmail, and start Kmail.

It also happens when I do a hostapd start, hostapd stop, hostapd start. 

Its a IBM 206 eserver, with no X running on it. The card is a WMP300N.

Any hints?

gr,
Willem


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Luis R. Rodriguez]

On Mon, Jan 19, 2009 at 2:31 PM, W. van den Akker <listsrv@wilsoft.nl> wrote:
> On Sunday 18 January 2009, Luis R. Rodriguez wrote:
>> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
>> > On Sunday 18 January 2009, Dmitri Seletski wrote:
>> > > W. van den Akker wrote:
>> > > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
>> > > >> Hi,
>> > > >>
>> > > >> Luis R. Rodriguez schrieb:
>> > > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net>
>> > > >>> wrote: [...]
>> > > >>> We need an oops trace of your crashes so far it seems you have not
>> > > >>> been able to get one, lets work on that. Until then it is not easy
>> > > >>> for us to provide you with patches to fix your issues.
>> > > >>
>> > > >> I also have similar problems with an D-LINK DWA-547 and I'd really
>> > > >> like to provide the oops trace, but unfortunately the machine seems
>> > > >> to hang that fast, that even no oops trace is shown. Also
>> > > >> magic-sysreq don't work anymore, neither via serial console nor via
>> > > >> keyboard. Only chance I have in this moment is to press
>> > > >> reset-button, which already caused a full xfs-filsystem-corruption
>> > > >> with dataloss once :-(
>> > > >>
>> > > >> My problem is, that the hang happens not immediately after loading
>> > > >> the driver but "somewhen" while it is loaded.
>> > > >>
>> > > >> I have seen two types of crash:
>> > > >>
>> > > >> 1.) first investigations on my side seem to show that it happens, as
>> > > >> soon as there is high dataload transfered, but I don't have detailed
>> > > >> info yet.
>> > > >> 2.) Running hostapd works fine while started first time, but no
>> > > >> client is connected. When a client connects a system freeze occures
>> > > >> "sometimes". Reproducably the system freezes immediately when
>> > > >> starting hostapd, stop it and then restart hostapd. So every change
>> > > >> in hostapd conf needs a full reboot, unloading the ath9k module and
>> > > >> reload it does not prevent freeze, but after reboot I can start
>> > > >> hostapd once without a problem (only the metioned "sometime" freezes
>> > > >> on client connect or high data-volume).
>> > > >>
>> > > >> As this machine is productive, I suspended testing for now. I will
>> > > >> change the card into a non-productive environment in the next days.
>> > > >> As soon as I find more details, I'll report.
>> > > >>
>> > > >>> Please test with the latest drivers from
>> > > >>> wireless-testing/compat-wireless and restart in single user mode,
>> > > >>> we want to ensure X is not running at all and you are in a virtual
>> > > >>> terminal. We also do not want Network Manager or wpa_supplicant
>> > > >>> running in the background. To increase the resolution upon bootup
>> > > >>> try booting with vga=ask (vga=791 works well for me), then try to
>> > > >>> associate to your access point without encryption manually using
>> > > >>> iwconfig.
>> > > >>
>> > > >> Ok, that'll be a thing I'd need to test to, as in my case
>> > > >> wpa_supplicant is running.
>> > > >
>> > > > Hi,
>> > > >
>> > > > What is the status of your problem? I still have problems with it and
>> > > > have no progress...
>> > > >
>> > > > gr,
>> > > > Willem
>> > >
>> > > Hello Willem.
>> > >
>> > > I think I have had similar issue to yours.
>> > >
>> > > Car was working for short period of time and after a while crashed
>> > > computer. I was advised to change kernel version, i updated it to
>> > > 2.6.27.9 and problem has gone away since.
>> > > It's something to do with DMA.
>> > >
>> > > Perhaps try to upgrade kernel?
>> > >
>> > > Regards
>> > > Dmitri
>> >
>> > I have tried all kernel versions till 2.6.28 and still have the same
>> > hang-ups. I think indeed its a DMA-hardware-ath9k related problem.  The
>> > patch in 2.6.27.8 didnt work for me. I have filled a bugreport on
>> > bugzilla.kernel but because AP isnt supported untill 2.6.29 the report is
>> > closed.
>>
>> A better description of this is that AP support does not work (is not
>> allowed) and therefore not supported on < 2.6.28.
>>
>> > I will test today 2.6.29.rc2 and see if it has any progress.
>>
>> You can also try wireless-testing (entire kernel) or compat-wireless (just
>> updates your wireless stuff) for bleeding edge stuff.
>>
>
> I tested it today with kernel 2.6.29-rc2. But no change. The kernel hangs
> completly (without oops or what so ever, even nmi-watchdog has no results).
> It happend when I started the connection (from a IW4965 card). Then started up
> a Kmail connection, start up a browser sessie. Stop Kmail, and start Kmail.
>
> It also happens when I do a hostapd start, hostapd stop, hostapd start.
>
> Its a IBM 206 eserver, with no X running on it. The card is a WMP300N.
>
> Any hints?

Since you are not getting any panic can you try wireless-testing git
tree [1], this will get you the latest and greatest wireless drivers
and mac80211/cfg80211 which are now being queued up for 2.6.30. See if
you see the issue there. If so then you can use git bisect.

[1] http://wireless.kernel.org/en/developers/Documentation/git-guide

  Luis

[ath9k-devel] still same problem

[W. van den Akker]

On Tuesday 20 January 2009, Luis R. Rodriguez wrote:
> On Mon, Jan 19, 2009 at 2:31 PM, W. van den Akker <listsrv@wilsoft.nl> 
wrote:
> > On Sunday 18 January 2009, Luis R. Rodriguez wrote:
> >> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
> >> > On Sunday 18 January 2009, Dmitri Seletski wrote:
> >> > > W. van den Akker wrote:
> >> > > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> >> > > >> Hi,
> >> > > >>
> >> > > >> Luis R. Rodriguez schrieb:
> >> > > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net>
> >> > > >>> wrote: [...]
> >> > > >>> We need an oops trace of your crashes so far it seems you have
> >> > > >>> not been able to get one, lets work on that. Until then it is
> >> > > >>> not easy for us to provide you with patches to fix your issues.
> >> > > >>
> >> > > >> I also have similar problems with an D-LINK DWA-547 and I'd
> >> > > >> really like to provide the oops trace, but unfortunately the
> >> > > >> machine seems to hang that fast, that even no oops trace is
> >> > > >> shown. Also magic-sysreq don't work anymore, neither via serial
> >> > > >> console nor via keyboard. Only chance I have in this moment is to
> >> > > >> press
> >> > > >> reset-button, which already caused a full
> >> > > >> xfs-filsystem-corruption with dataloss once :-(
> >> > > >>
> >> > > >> My problem is, that the hang happens not immediately after
> >> > > >> loading the driver but "somewhen" while it is loaded.
> >> > > >>
> >> > > >> I have seen two types of crash:
> >> > > >>
> >> > > >> 1.) first investigations on my side seem to show that it happens,
> >> > > >> as soon as there is high dataload transfered, but I don't have
> >> > > >> detailed info yet.
> >> > > >> 2.) Running hostapd works fine while started first time, but no
> >> > > >> client is connected. When a client connects a system freeze
> >> > > >> occures "sometimes". Reproducably the system freezes immediately
> >> > > >> when starting hostapd, stop it and then restart hostapd. So every
> >> > > >> change in hostapd conf needs a full reboot, unloading the ath9k
> >> > > >> module and reload it does not prevent freeze, but after reboot I
> >> > > >> can start hostapd once without a problem (only the metioned
> >> > > >> "sometime" freezes on client connect or high data-volume).
> >> > > >>
> >> > > >> As this machine is productive, I suspended testing for now. I
> >> > > >> will change the card into a non-productive environment in the
> >> > > >> next days. As soon as I find more details, I'll report.
> >> > > >>
> >> > > >>> Please test with the latest drivers from
> >> > > >>> wireless-testing/compat-wireless and restart in single user
> >> > > >>> mode, we want to ensure X is not running at all and you are in a
> >> > > >>> virtual terminal. We also do not want Network Manager or
> >> > > >>> wpa_supplicant running in the background. To increase the
> >> > > >>> resolution upon bootup try booting with vga=ask (vga=791 works
> >> > > >>> well for me), then try to associate to your access point without
> >> > > >>> encryption manually using iwconfig.
> >> > > >>
> >> > > >> Ok, that'll be a thing I'd need to test to, as in my case
> >> > > >> wpa_supplicant is running.
> >> > > >
> >> > > > Hi,
> >> > > >
> >> > > > What is the status of your problem? I still have problems with it
> >> > > > and have no progress...
> >> > > >
> >> > > > gr,
> >> > > > Willem
> >> > >
> >> > > Hello Willem.
> >> > >
> >> > > I think I have had similar issue to yours.
> >> > >
> >> > > Car was working for short period of time and after a while crashed
> >> > > computer. I was advised to change kernel version, i updated it to
> >> > > 2.6.27.9 and problem has gone away since.
> >> > > It's something to do with DMA.
> >> > >
> >> > > Perhaps try to upgrade kernel?
> >> > >
> >> > > Regards
> >> > > Dmitri
> >> >
> >> > I have tried all kernel versions till 2.6.28 and still have the same
> >> > hang-ups. I think indeed its a DMA-hardware-ath9k related problem. 
> >> > The patch in 2.6.27.8 didnt work for me. I have filled a bugreport on
> >> > bugzilla.kernel but because AP isnt supported untill 2.6.29 the report
> >> > is closed.
> >>
> >> A better description of this is that AP support does not work (is not
> >> allowed) and therefore not supported on < 2.6.28.
> >>
> >> > I will test today 2.6.29.rc2 and see if it has any progress.
> >>
> >> You can also try wireless-testing (entire kernel) or compat-wireless
> >> (just updates your wireless stuff) for bleeding edge stuff.
> >
> > I tested it today with kernel 2.6.29-rc2. But no change. The kernel hangs
> > completly (without oops or what so ever, even nmi-watchdog has no
> > results). It happend when I started the connection (from a IW4965 card).
> > Then started up a Kmail connection, start up a browser sessie. Stop
> > Kmail, and start Kmail.
> >
> > It also happens when I do a hostapd start, hostapd stop, hostapd start.
> >
> > Its a IBM 206 eserver, with no X running on it. The card is a WMP300N.
> >
> > Any hints?
>
> Since you are not getting any panic can you try wireless-testing git
> tree [1], this will get you the latest and greatest wireless drivers
> and mac80211/cfg80211 which are now being queued up for 2.6.30. See if
> you see the issue there. If so then you can use git bisect.
>
> [1] http://wireless.kernel.org/en/developers/Documentation/git-guide

I will use the latest drivers from wl. The last I tried was 18-1 and that 
didnt had any success.

To be continued.

Willem



-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[W. van den Akker]

On Tuesday 20 January 2009, Luis R. Rodriguez wrote:
> On Mon, Jan 19, 2009 at 2:31 PM, W. van den Akker <listsrv@wilsoft.nl> 
wrote:
> > On Sunday 18 January 2009, Luis R. Rodriguez wrote:
> >> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
> >> > On Sunday 18 January 2009, Dmitri Seletski wrote:
> >> > > W. van den Akker wrote:
> >> > > > On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
> >> > > >> Hi,
> >> > > >>
> >> > > >> Luis R. Rodriguez schrieb:
> >> > > >>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green <barry@lemonpi.net>
> >> > > >>> wrote: [...]
> >> > > >>> We need an oops trace of your crashes so far it seems you have
> >> > > >>> not been able to get one, lets work on that. Until then it is
> >> > > >>> not easy for us to provide you with patches to fix your issues.
> >> > > >>
> >> > > >> I also have similar problems with an D-LINK DWA-547 and I'd
> >> > > >> really like to provide the oops trace, but unfortunately the
> >> > > >> machine seems to hang that fast, that even no oops trace is
> >> > > >> shown. Also magic-sysreq don't work anymore, neither via serial
> >> > > >> console nor via keyboard. Only chance I have in this moment is to
> >> > > >> press
> >> > > >> reset-button, which already caused a full
> >> > > >> xfs-filsystem-corruption with dataloss once :-(
> >> > > >>
> >> > > >> My problem is, that the hang happens not immediately after
> >> > > >> loading the driver but "somewhen" while it is loaded.
> >> > > >>
> >> > > >> I have seen two types of crash:
> >> > > >>
> >> > > >> 1.) first investigations on my side seem to show that it happens,
> >> > > >> as soon as there is high dataload transfered, but I don't have
> >> > > >> detailed info yet.
> >> > > >> 2.) Running hostapd works fine while started first time, but no
> >> > > >> client is connected. When a client connects a system freeze
> >> > > >> occures "sometimes". Reproducably the system freezes immediately
> >> > > >> when starting hostapd, stop it and then restart hostapd. So every
> >> > > >> change in hostapd conf needs a full reboot, unloading the ath9k
> >> > > >> module and reload it does not prevent freeze, but after reboot I
> >> > > >> can start hostapd once without a problem (only the metioned
> >> > > >> "sometime" freezes on client connect or high data-volume).
> >> > > >>
> >> > > >> As this machine is productive, I suspended testing for now. I
> >> > > >> will change the card into a non-productive environment in the
> >> > > >> next days. As soon as I find more details, I'll report.
> >> > > >>
> >> > > >>> Please test with the latest drivers from
> >> > > >>> wireless-testing/compat-wireless and restart in single user
> >> > > >>> mode, we want to ensure X is not running at all and you are in a
> >> > > >>> virtual terminal. We also do not want Network Manager or
> >> > > >>> wpa_supplicant running in the background. To increase the
> >> > > >>> resolution upon bootup try booting with vga=ask (vga=791 works
> >> > > >>> well for me), then try to associate to your access point without
> >> > > >>> encryption manually using iwconfig.
> >> > > >>
> >> > > >> Ok, that'll be a thing I'd need to test to, as in my case
> >> > > >> wpa_supplicant is running.
> >> > > >
> >> > > > Hi,
> >> > > >
> >> > > > What is the status of your problem? I still have problems with it
> >> > > > and have no progress...
> >> > > >
> >> > > > gr,
> >> > > > Willem
> >> > >
> >> > > Hello Willem.
> >> > >
> >> > > I think I have had similar issue to yours.
> >> > >
> >> > > Car was working for short period of time and after a while crashed
> >> > > computer. I was advised to change kernel version, i updated it to
> >> > > 2.6.27.9 and problem has gone away since.
> >> > > It's something to do with DMA.
> >> > >
> >> > > Perhaps try to upgrade kernel?
> >> > >
> >> > > Regards
> >> > > Dmitri
> >> >
> >> > I have tried all kernel versions till 2.6.28 and still have the same
> >> > hang-ups. I think indeed its a DMA-hardware-ath9k related problem. 
> >> > The patch in 2.6.27.8 didnt work for me. I have filled a bugreport on
> >> > bugzilla.kernel but because AP isnt supported untill 2.6.29 the report
> >> > is closed.
> >>
> >> A better description of this is that AP support does not work (is not
> >> allowed) and therefore not supported on < 2.6.28.
> >>
> >> > I will test today 2.6.29.rc2 and see if it has any progress.
> >>
> >> You can also try wireless-testing (entire kernel) or compat-wireless
> >> (just updates your wireless stuff) for bleeding edge stuff.
> >
> > I tested it today with kernel 2.6.29-rc2. But no change. The kernel hangs
> > completly (without oops or what so ever, even nmi-watchdog has no
> > results). It happend when I started the connection (from a IW4965 card).
> > Then started up a Kmail connection, start up a browser sessie. Stop
> > Kmail, and start Kmail.
> >
> > It also happens when I do a hostapd start, hostapd stop, hostapd start.
> >
> > Its a IBM 206 eserver, with no X running on it. The card is a WMP300N.
> >
> > Any hints?
>
> Since you are not getting any panic can you try wireless-testing git
> tree [1], this will get you the latest and greatest wireless drivers
> and mac80211/cfg80211 which are now being queued up for 2.6.30. See if
> you see the issue there. If so then you can use git bisect.
>

Also the latest wireless testing git has the same problem.
No Oops, trace or any panic visible. 

Some nasty problem here.....

gr,
Willem


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Chris Kennedy]

On Jan 20, 2009, at 4:22 PM, W. van den Akker wrote:

> On Tuesday 20 January 2009, Luis R. Rodriguez wrote:
>> On Mon, Jan 19, 2009 at 2:31 PM, W. van den Akker  
>> <listsrv@wilsoft.nl>
> wrote:
>>> On Sunday 18 January 2009, Luis R. Rodriguez wrote:
>>>> On Sun, Jan 18, 2009 at 01:34:03AM -0800, W. van den Akker wrote:
>>>>> On Sunday 18 January 2009, Dmitri Seletski wrote:
>>>>>> W. van den Akker wrote:
>>>>>>> On Sunday 04 January 2009 10:58:02 Jens Ott - Opteamax wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Luis R. Rodriguez schrieb:
>>>>>>>>> On Sat, Jan 3, 2009 at 5:53 PM, Barry Green  
>>>>>>>>> <barry@lemonpi.net>
>>>>>>>>> wrote: [...]
>>>>>>>>> We need an oops trace of your crashes so far it seems you have
>>>>>>>>> not been able to get one, lets work on that. Until then it is
>>>>>>>>> not easy for us to provide you with patches to fix your  
>>>>>>>>> issues.
>>>>>>>>
>>>>>>>> I also have similar problems with an D-LINK DWA-547 and I'd
>>>>>>>> really like to provide the oops trace, but unfortunately the
>>>>>>>> machine seems to hang that fast, that even no oops trace is
>>>>>>>> shown. Also magic-sysreq don't work anymore, neither via serial
>>>>>>>> console nor via keyboard. Only chance I have in this moment  
>>>>>>>> is to
>>>>>>>> press
>>>>>>>> reset-button, which already caused a full
>>>>>>>> xfs-filsystem-corruption with dataloss once :-(
>>>>>>>>
>>>>>>>> My problem is, that the hang happens not immediately after
>>>>>>>> loading the driver but "somewhen" while it is loaded.
>>>>>>>>
>>>>>>>> I have seen two types of crash:
>>>>>>>>
>>>>>>>> 1.) first investigations on my side seem to show that it  
>>>>>>>> happens,
>>>>>>>> as soon as there is high dataload transfered, but I don't have
>>>>>>>> detailed info yet.
>>>>>>>> 2.) Running hostapd works fine while started first time, but no
>>>>>>>> client is connected. When a client connects a system freeze
>>>>>>>> occures "sometimes". Reproducably the system freezes  
>>>>>>>> immediately
>>>>>>>> when starting hostapd, stop it and then restart hostapd. So  
>>>>>>>> every
>>>>>>>> change in hostapd conf needs a full reboot, unloading the ath9k
>>>>>>>> module and reload it does not prevent freeze, but after  
>>>>>>>> reboot I
>>>>>>>> can start hostapd once without a problem (only the metioned
>>>>>>>> "sometime" freezes on client connect or high data-volume).
>>>>>>>>
>>>>>>>> As this machine is productive, I suspended testing for now. I
>>>>>>>> will change the card into a non-productive environment in the
>>>>>>>> next days. As soon as I find more details, I'll report.
>>>>>>>>
>>>>>>>>> Please test with the latest drivers from
>>>>>>>>> wireless-testing/compat-wireless and restart in single user
>>>>>>>>> mode, we want to ensure X is not running at all and you are  
>>>>>>>>> in a
>>>>>>>>> virtual terminal. We also do not want Network Manager or
>>>>>>>>> wpa_supplicant running in the background. To increase the
>>>>>>>>> resolution upon bootup try booting with vga=ask (vga=791 works
>>>>>>>>> well for me), then try to associate to your access point  
>>>>>>>>> without
>>>>>>>>> encryption manually using iwconfig.
>>>>>>>>
>>>>>>>> Ok, that'll be a thing I'd need to test to, as in my case
>>>>>>>> wpa_supplicant is running.
>>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> What is the status of your problem? I still have problems  
>>>>>>> with it
>>>>>>> and have no progress...
>>>>>>>
>>>>>>> gr,
>>>>>>> Willem
>>>>>>
>>>>>> Hello Willem.
>>>>>>
>>>>>> I think I have had similar issue to yours.
>>>>>>
>>>>>> Car was working for short period of time and after a while  
>>>>>> crashed
>>>>>> computer. I was advised to change kernel version, i updated it to
>>>>>> 2.6.27.9 and problem has gone away since.
>>>>>> It's something to do with DMA.
>>>>>>
>>>>>> Perhaps try to upgrade kernel?
>>>>>>
>>>>>> Regards
>>>>>> Dmitri
>>>>>
>>>>> I have tried all kernel versions till 2.6.28 and still have the  
>>>>> same
>>>>> hang-ups. I think indeed its a DMA-hardware-ath9k related problem.
>>>>> The patch in 2.6.27.8 didnt work for me. I have filled a  
>>>>> bugreport on
>>>>> bugzilla.kernel but because AP isnt supported untill 2.6.29 the  
>>>>> report
>>>>> is closed.
>>>>
>>>> A better description of this is that AP support does not work  
>>>> (is not
>>>> allowed) and therefore not supported on < 2.6.28.
>>>>
>>>>> I will test today 2.6.29.rc2 and see if it has any progress.
>>>>
>>>> You can also try wireless-testing (entire kernel) or compat- 
>>>> wireless
>>>> (just updates your wireless stuff) for bleeding edge stuff.
>>>
>>> I tested it today with kernel 2.6.29-rc2. But no change. The  
>>> kernel hangs
>>> completly (without oops or what so ever, even nmi-watchdog has no
>>> results). It happend when I started the connection (from a IW4965  
>>> card).
>>> Then started up a Kmail connection, start up a browser sessie. Stop
>>> Kmail, and start Kmail.
>>>
>>> It also happens when I do a hostapd start, hostapd stop, hostapd  
>>> start.
>>>
>>> Its a IBM 206 eserver, with no X running on it. The card is a  
>>> WMP300N.
>>>
>>> Any hints?
>>
>> Since you are not getting any panic can you try wireless-testing git
>> tree [1], this will get you the latest and greatest wireless drivers
>> and mac80211/cfg80211 which are now being queued up for 2.6.30.  
>> See if
>> you see the issue there. If so then you can use git bisect.
>>
>
> Also the latest wireless testing git has the same problem.
> No Oops, trace or any panic visible.
>
> Some nasty problem here.....
>
> gr,
> Willem
>
>

This is interesting since I had the same lockups using the module  
ath9k, including Gentoo's newest 2.6.27-gentoo-r7, the compat- 
wireless newest git compiled with wireless-testing, and using  
wireless-testing itself (newest as of a day ago).  So all of those,  
when used on the console with just hostapd (version 0.6.7) to setup  
an AP, would freeze the system without any messages or information.   
This was on an HP xw8200 workstation using 2 Xeon hyperthreaded  
CPU's, so a very large system and lots of possible locking-bugs  
finding there.  So I built a test system with a Uni Processor system,  
an older compaq P4 workstation that isn't SMP and not multithreaded.   
Now it works very stable on the UP system, compared with the SMP  
Hyperthreaded one.

So I've been really happy with it now, I've been able to get the AP  
going using the newest wireless-testing using WPA-PSK encryption with  
the newest hostapd.  This was doing exactly what your reporting on my  
HP SMP system,and I duplicated the system when I setup the Compaq UP  
system, so nothing is different except hardware.

I have been watching this thread and as of last night since I got it  
working on the UP system, and today tweaked the system to match my  
Apple Airport Extreme AP (have 802.11n enabled, using the setting  
"ht_capab=[HT40+][SHORT-GI-40][DSSS_CCK-40]" which seems to have  
gotten it to be faster than the Apple AP for my Windows systems using  
Linksys WMP300N cards (the American ones with the Broadcom chips).   
Since today from this working now, and your newest email about using  
the exact same wireless-testing kernel I'm using (mine is newest git  
version as-of last night).

So my theory is that I was triggering the same problem using SMP/ 
Hyperthreading, I suspect your system is not a smaller single  
processor system?  That would at least support that theory.  I think  
the problem could be a spinlock locking problem in the interrupt  
handler, or something like that.  I've programmed drivers, wrote a  
good part of the IVTV driver, and this behavior reminds me of what I  
saw there at times during development  (it would be stable on UP  
systems, lock like this on SMP systems).  I have been peeking around  
in the code some to get familiar with it, but of course may take some  
time for me to start grasping what is going on to be able to  
contribute any help myself.  So this is just what I'm thinking, since  
freezes like this often seem to be something wrong in the interrupt  
handler for SMP/preemption configured systems.  If you are using SMP,  
maybe trying it with just one processor enabled would be a good test  
to see if this is really true there too.

Thanks,
Chris

[ath9k-devel] still same problem

[W. van den Akker]

>
> So my theory is that I was triggering the same problem using SMP/
> Hyperthreading, I suspect your system is not a smaller single
> processor system?  That would at least support that theory.  I think
> the problem could be a spinlock locking problem in the interrupt
> handler, or something like that.  I've programmed drivers, wrote a
> good part of the IVTV driver, and this behavior reminds me of what I
> saw there at times during development  (it would be stable on UP
> systems, lock like this on SMP systems).  I have been peeking around
> in the code some to get familiar with it, but of course may take some
> time for me to start grasping what is going on to be able to
> contribute any help myself.  So this is just what I'm thinking, since
> freezes like this often seem to be something wrong in the interrupt
> handler for SMP/preemption configured systems.  If you are using SMP,
> maybe trying it with just one processor enabled would be a good test
> to see if this is really true there too.
>

Hil Chris,

This is correct. Its a SMP system with HT. So I think your assumption is 
right. It must be in that direction.
 
I will try to test it with one CPU disabled.
I will report back.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[W. van den Akker]

On Wednesday 21 January 2009, W. van den Akker wrote:
> > So my theory is that I was triggering the same problem using SMP/
> > Hyperthreading, I suspect your system is not a smaller single
> > processor system?  That would at least support that theory.  I think
> > the problem could be a spinlock locking problem in the interrupt
> > handler, or something like that.  I've programmed drivers, wrote a
> > good part of the IVTV driver, and this behavior reminds me of what I
> > saw there at times during development  (it would be stable on UP
> > systems, lock like this on SMP systems).  I have been peeking around
> > in the code some to get familiar with it, but of course may take some
> > time for me to start grasping what is going on to be able to
> > contribute any help myself.  So this is just what I'm thinking, since
> > freezes like this often seem to be something wrong in the interrupt
> > handler for SMP/preemption configured systems.  If you are using SMP,
> > maybe trying it with just one processor enabled would be a good test
> > to see if this is really true there too.
>
> Hil Chris,
>
> This is correct. Its a SMP system with HT. So I think your assumption is
> right. It must be in that direction.
>
> I will try to test it with one CPU disabled.
> I will report back.
>

I have tested with 1 CPU disabled. Running now for about 2 hours without
hangups (I have several connection drops, but it will reconnect).

I have also tested with 2 CPU's and HT-disabled (noht parameter) but that
had no effect.

So SMP looks like the cause of the problem here.

I will be happy to do further testing.

gr,
Willem





-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090121/e24cd9bf/attachment.pgp 

[ath9k-devel] still same problem

[W. van den Akker]

On Wednesday 21 January 2009, W. van den Akker wrote:
> > So my theory is that I was triggering the same problem using SMP/
> > Hyperthreading, I suspect your system is not a smaller single
> > processor system?  That would at least support that theory.  I think
> > the problem could be a spinlock locking problem in the interrupt
> > handler, or something like that.  I've programmed drivers, wrote a
> > good part of the IVTV driver, and this behavior reminds me of what I
> > saw there at times during development  (it would be stable on UP
> > systems, lock like this on SMP systems).  I have been peeking around
> > in the code some to get familiar with it, but of course may take some
> > time for me to start grasping what is going on to be able to
> > contribute any help myself.  So this is just what I'm thinking, since
> > freezes like this often seem to be something wrong in the interrupt
> > handler for SMP/preemption configured systems.  If you are using SMP,
> > maybe trying it with just one processor enabled would be a good test
> > to see if this is really true there too.
>
> Hil Chris,
>
> This is correct. Its a SMP system with HT. So I think your assumption is
> right. It must be in that direction.
>
> I will try to test it with one CPU disabled.
> I will report back.
>

I have tested with 1 CPU disabled. Running now for about 2 hours without
hangups (I have several connection drops, but it will reconnect).

I have also tested with 2 CPU's and HT-disabled (noht parameter) but that
had no effect.

So SMP looks like the cause of the problem here.

I will be happy to do further testing.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] still same problem

[Jouni Malinen]

On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
> I have tested with 1 CPU disabled. Running now for about 2 hours without
> hangups (I have several connection drops, but it will reconnect).
> 
> I have also tested with 2 CPU's and HT-disabled (noht parameter) but that
> had no effect.
> 
> So SMP looks like the cause of the problem here.

Thanks for testing this! I'm running most of my tests with a dual core
system, so SMP is being used, but with two cores, not two separate
chips. I don't think there should be much difference there, but
certainly our hardware configuration is different.

Could you please describe your hardware with more details so that we can
see whether we could find a similar system to try to reproduce this? Is
this the IBM 206 eserver with WMP300N you mentioned in an earlier
message in the thread? That seems to be (by default) a uni-processor
setup, so I would like to make sure we understand what is the exact
hardware used here since I do not think we have been able to reproduce
this type of issue so far in any dual core systems.

- Jouni

[ath9k-devel] [PATCH] still same problem

[Chris Kennedy]

On Jan 21, 2009, at 6:35 AM, Jouni Malinen wrote:

> On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
>> I have tested with 1 CPU disabled. Running now for about 2 hours  
>> without
>> hangups (I have several connection drops, but it will reconnect).
>>
>> I have also tested with 2 CPU's and HT-disabled (noht parameter)  
>> but that
>> had no effect.
>>
>> So SMP looks like the cause of the problem here.
>
> Thanks for testing this! I'm running most of my tests with a dual core
> system, so SMP is being used, but with two cores, not two separate
> chips. I don't think there should be much difference there, but
> certainly our hardware configuration is different.
>
> Could you please describe your hardware with more details so that  
> we can
> see whether we could find a similar system to try to reproduce  
> this? Is
> this the IBM 206 eserver with WMP300N you mentioned in an earlier
> message in the thread? That seems to be (by default) a uni-processor
> setup, so I would like to make sure we understand what is the exact
> hardware used here since I do not think we have been able to reproduce
> this type of issue so far in any dual core systems.
>
> - Jouni
>
>

This is a patch (against yesterdays current wireless-testing kernel)
  that shows basically what we did in the IVTV driver,
essentially holding a spin_lock() in the interrupt handler.  So it
isn't even used, since compiled out, if a system isn't SMP.  When the
system is SMP it prevents multiple instances of the interrupt handler
from happening.  This seemed to be the magic fix there, I'm not sure if
this is totally correct for this driver, but it's a patch to test  
(hopefully the
spinlock I used makes some sense, seems this also is good since
wouldn't want to reset the card and have interrupts happen while
that's going on?).  It's at least an example showing what I suspect  
could
fix it, and would be interesting if you had the ability to test it on  
your SMP
system to see if it is just the interrupt handler where it's happening.

Thanks,
Chris

-------------- next part --------------
A non-text attachment was scrubbed...
Name: irq_handler_spinlock.diff
Type: application/applefile
Size: 1435 bytes
Desc: not available
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090121/36ecdc7d/attachment.bin 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: irq_handler_spinlock.diff
Type: application/octet-stream
Size: 1148 bytes
Desc: not available
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090121/36ecdc7d/attachment.obj 
-------------- next part --------------

[ath9k-devel] still same problem

[W. van den Akker]

On Wednesday 21 January 2009, Jouni Malinen wrote:
> On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
> > I have tested with 1 CPU disabled. Running now for about 2 hours without
> > hangups (I have several connection drops, but it will reconnect).
> >
> > I have also tested with 2 CPU's and HT-disabled (noht parameter) but that
> > had no effect.
> >
> > So SMP looks like the cause of the problem here.
>
> Thanks for testing this! I'm running most of my tests with a dual core
> system, so SMP is being used, but with two cores, not two separate
> chips. I don't think there should be much difference there, but
> certainly our hardware configuration is different.
>
> Could you please describe your hardware with more details so that we can
> see whether we could find a similar system to try to reproduce this? Is
> this the IBM 206 eserver with WMP300N you mentioned in an earlier
> message in the thread? That seems to be (by default) a uni-processor
> setup, so I would like to make sure we understand what is the exact
> hardware used here since I do not think we have been able to reproduce
> this type of issue so far in any dual core systems.
>

Its indeed an IBM eServer x206 8482-2RX with a WMP300N (Linksys,  
Atheros Communications Inc. AR5416 802.11abgn Wireless PCI Adapter (rev 01).
The IBM is a dual-core system with 2 CPU's on one fysical CPU-chip (its boxed 
so I am not for sure).

/proc/cpuinfo shows (with one CPU disabled)
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 3
model name      : Intel(R) Pentium(R) 4 CPU 3.00GHz
stepping        : 4
cpu MHz         : 3000.473
cache size      : 1024 KB
physical id     : 0
siblings        : 1
core id         : 0
cpu cores       : 1
apicid          : 0
initial apicid  : 0
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 5
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca 
cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe constant_tsc 
up pebs bts pni dtes64 monitor ds_cpl cid xtpr
bogomips        : 6000.94
clflush size    : 64
power management:

lscpi -vvv for the adapter shows:
03:02.0 Network controller: Atheros Communications Inc. AR5416 802.11abgn 
Wireless PCI Adapter (rev 01)
        Subsystem: Linksys Device 0061
        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, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 25
        Region 0: Memory@d0200000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [40] #80 [0000]
        Kernel driver in use: ath9k
        Kernel modules: ath9k

If more information is needed I am happy to provide.

Greetings,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] [PATCH] still same problem

[W. van den Akker]

On Wednesday 21 January 2009, Chris Kennedy wrote:
> On Jan 21, 2009, at 6:35 AM, Jouni Malinen wrote:
> > On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
> >> I have tested with 1 CPU disabled. Running now for about 2 hours
> >> without
> >> hangups (I have several connection drops, but it will reconnect).
> >>
> >> I have also tested with 2 CPU's and HT-disabled (noht parameter)
> >> but that
> >> had no effect.
> >>
> >> So SMP looks like the cause of the problem here.
> >
> > Thanks for testing this! I'm running most of my tests with a dual core
> > system, so SMP is being used, but with two cores, not two separate
> > chips. I don't think there should be much difference there, but
> > certainly our hardware configuration is different.
> >
> > Could you please describe your hardware with more details so that
> > we can
> > see whether we could find a similar system to try to reproduce
> > this? Is
> > this the IBM 206 eserver with WMP300N you mentioned in an earlier
> > message in the thread? That seems to be (by default) a uni-processor
> > setup, so I would like to make sure we understand what is the exact
> > hardware used here since I do not think we have been able to reproduce
> > this type of issue so far in any dual core systems.
> >
> > - Jouni
>
> This is a patch (against yesterdays current wireless-testing kernel)
>   that shows basically what we did in the IVTV driver,
> essentially holding a spin_lock() in the interrupt handler.  So it
> isn't even used, since compiled out, if a system isn't SMP.  When the
> system is SMP it prevents multiple instances of the interrupt handler
> from happening.  This seemed to be the magic fix there, I'm not sure if
> this is totally correct for this driver, but it's a patch to test
> (hopefully the
> spinlock I used makes some sense, seems this also is good since
> wouldn't want to reset the card and have interrupts happen while
> that's going on?).  It's at least an example showing what I suspect
> could
> fix it, and would be interesting if you had the ability to test it on
> your SMP
> system to see if it is just the interrupt handler where it's happening.
>

Thanks Chris,

I applied the patch.... but no luck. Systems still hangs.

gr,
Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] [PATCH] still same problem

[W. van den Akker]

On Wednesday 21 January 2009, W. van den Akker wrote:
> On Wednesday 21 January 2009, Chris Kennedy wrote:
> > On Jan 21, 2009, at 6:35 AM, Jouni Malinen wrote:
> > > On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
> > >> I have tested with 1 CPU disabled. Running now for about 2 hours
> > >> without
> > >> hangups (I have several connection drops, but it will reconnect).
> > >>
> > >> I have also tested with 2 CPU's and HT-disabled (noht parameter)
> > >> but that
> > >> had no effect.
> > >>
> > >> So SMP looks like the cause of the problem here.
> > >
> > > Thanks for testing this! I'm running most of my tests with a dual core
> > > system, so SMP is being used, but with two cores, not two separate
> > > chips. I don't think there should be much difference there, but
> > > certainly our hardware configuration is different.
> > >
> > > Could you please describe your hardware with more details so that
> > > we can
> > > see whether we could find a similar system to try to reproduce
> > > this? Is
> > > this the IBM 206 eserver with WMP300N you mentioned in an earlier
> > > message in the thread? That seems to be (by default) a uni-processor
> > > setup, so I would like to make sure we understand what is the exact
> > > hardware used here since I do not think we have been able to reproduce
> > > this type of issue so far in any dual core systems.
> > >
> > > - Jouni
> >
> > This is a patch (against yesterdays current wireless-testing kernel)
> >   that shows basically what we did in the IVTV driver,
> > essentially holding a spin_lock() in the interrupt handler.  So it
> > isn't even used, since compiled out, if a system isn't SMP.  When the
> > system is SMP it prevents multiple instances of the interrupt handler
> > from happening.  This seemed to be the magic fix there, I'm not sure if
> > this is totally correct for this driver, but it's a patch to test
> > (hopefully the
> > spinlock I used makes some sense, seems this also is good since
> > wouldn't want to reset the card and have interrupts happen while
> > that's going on?).  It's at least an example showing what I suspect
> > could
> > fix it, and would be interesting if you had the ability to test it on
> > your SMP
> > system to see if it is just the interrupt handler where it's happening.
>
> Thanks Chris,
>
> I applied the patch.... but no luck. Systems still hangs.
>

Running for about a week with one CPU. The driver works ok and the 
connection is rock-solid.

Any progress on the dual-CPU lockup?

gr,

Willem

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[ath9k-devel] [PATCH] still same problem

[Chris Kennedy]

On Jan 27, 2009, at 3:30 AM, W. van den Akker wrote:

> On Wednesday 21 January 2009, W. van den Akker wrote:
>> On Wednesday 21 January 2009, Chris Kennedy wrote:
>>> On Jan 21, 2009, at 6:35 AM, Jouni Malinen wrote:
>>>> On Wed, 2009-01-21 at 02:11 -0800, W. van den Akker wrote:
>>>>> I have tested with 1 CPU disabled. Running now for about 2 hours
>>>>> without
>>>>> hangups (I have several connection drops, but it will reconnect).
>>>>>
>>>>> I have also tested with 2 CPU's and HT-disabled (noht parameter)
>>>>> but that
>>>>> had no effect.
>>>>>
>>>>> So SMP looks like the cause of the problem here.
>>>>
>>>> Thanks for testing this! I'm running most of my tests with a  
>>>> dual core
>>>> system, so SMP is being used, but with two cores, not two separate
>>>> chips. I don't think there should be much difference there, but
>>>> certainly our hardware configuration is different.
>>>>
>>>> Could you please describe your hardware with more details so that
>>>> we can
>>>> see whether we could find a similar system to try to reproduce
>>>> this? Is
>>>> this the IBM 206 eserver with WMP300N you mentioned in an earlier
>>>> message in the thread? That seems to be (by default) a uni- 
>>>> processor
>>>> setup, so I would like to make sure we understand what is the exact
>>>> hardware used here since I do not think we have been able to  
>>>> reproduce
>>>> this type of issue so far in any dual core systems.
>>>>
>>>> - Jouni
>>>
>>> This is a patch (against yesterdays current wireless-testing kernel)
>>>   that shows basically what we did in the IVTV driver,
>>> essentially holding a spin_lock() in the interrupt handler.  So it
>>> isn't even used, since compiled out, if a system isn't SMP.  When  
>>> the
>>> system is SMP it prevents multiple instances of the interrupt  
>>> handler
>>> from happening.  This seemed to be the magic fix there, I'm not  
>>> sure if
>>> this is totally correct for this driver, but it's a patch to test
>>> (hopefully the
>>> spinlock I used makes some sense, seems this also is good since
>>> wouldn't want to reset the card and have interrupts happen while
>>> that's going on?).  It's at least an example showing what I suspect
>>> could
>>> fix it, and would be interesting if you had the ability to test  
>>> it on
>>> your SMP
>>> system to see if it is just the interrupt handler where it's  
>>> happening.
>>
>> Thanks Chris,
>>
>> I applied the patch.... but no luck. Systems still hangs.
>>
>
> Running for about a week with one CPU. The driver works ok and the
> connection is rock-solid.
>
> Any progress on the dual-CPU lockup?

Same here, great on my one CPU system, minor kernel panic but that
only happened 2 times when changing settings and restarting things
so not triggered by normal use by me.  Unfortunately my SMP system
is too much 'in use' for me to test more on, and not wanting any risk  
of data loss.
Eventually would like to move the card into it of course since it's  
my main router system and
would be nice to have my wireless network routing directly into my main
router.  These SMP type system freezes seem hard to figure out, I'm not
even sure how to really get detailed data and pinpoint where the locking
problems would be.  It's just frozen and I'm not sure a serial  
console would
even show anything or not.  I think the usual procedure is to review  
the locking
and trial and error on an SMP system, guessing what could be wrong,  
doing it
by feel but basically blind about what your trying to fix.  This is  
probably one of the
hardest parts of kernel development, especially for someone who  
didn't write the code
or knows the code inside and out, and even then I'm suspecting it's  
very elusive and
has to do with both the locking and the way the hardware is being  
accessed.  Also
it's probaby rarer for people to be using a box as an AP that is SMP,  
since really
just need a low grade single processor system to do a dedicated  
routing system like
this.  So I'm not getting my hopes up for my SMP box to be able to  
have the AP
functionality directly on it anytime soon :(, although it would be  
nice I'm getting by
pretty good on my spare UP system.

Thanks,
Chris



>
> gr,
>
> Willem
>
> -- 
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>