On Sat, Feb 20, 2010 at 5:00 AM, Pavel Roskin <proski@gnu.org> wrote:
On Fri, 2010-02-19 at 15:17 +0530, Vishal Sevani wrote:

>         > I did apply the following patch
>         > https://madwifi-project.org/ticket/2322 , but to no avail.
>         Could you
>         > point out as to what the problem could be?? If you need me
>         to provide
>         > any other debug information please do let me know.
>         I'll appreciate if you post your results to the ticket.  So
>         far there
>         have been no test results for that patch at all.
> Yeah I have posted my experience in the following ticket.
> https://madwifi-project.org/ticket/2359

I actually I meant that you would report a negative result with the
patch in #2322.  But it looks like the problems are different.

Yeah it seems more of a hardware issue with AR71xx borad, (or issue with Openwrt) as pointed out by nenad@panline.net in his comment following yours in the ticket #2359. If Iam correct, it is happening because the value in reg AR_ISR corresponding to flag AR_ISR_HIUERR is being set, which is for unexpected bus error. As mentioned in the following post http://n2.nabble.com/fatal-interrupt-td3225930.html, it is beacuse the following bit, AR_ISR_S2_MCABT  (for master cycle abort) is being set.

Iam clueless as to what could be causing this master cycle abort problem and how should I go about debugging it. Could you please provide any pointers as to how I can rectify this problem.
> Yesterday again I tried with latest release of Openwrt Kamikaze  ie.
> v8.09.2, but still I get the same problem. But the madwifi code works
> properly on IBM latop running Ubuntu with Linux kernel 2.6.24-23. So I
> guess it is the issue with Openwrt or the hardware ie. AR71xx boards.
> Has the latest code been tested on Openwrt and Ar71xx boards??

I'm not aware of it.

Pavel Roskin