I took the 2.6.33-at91 patch from http://maxim.org.za/at91_26.html and forward-ported it to 2.6.34 and 2.6.35. I hope these patches help somebody.

They are attached to this message.

Thanks,
-Chris

Adding at91 patch for 2.6.36.

Thanks-- I've now got 2.6.36 running. I appreciate the time you put in both porting and posting.

Hello,

Thanks for the AT91 Linux patches. I am having an issue however. I don't know if it is related to the kernel patches or not. It has to do with the USB V4L GSPCA driver.

On version 2.6.32 everything works fine.

On version 2.6.34 and 2.6.35 when plugging in a USB camera the kernel crashes with this error.

gspca camera detected
kernel BUG at arch/arm/mm/dma-mapping.c:409!
Unable to handle kernel NULL pointer dereference at virtual address 00000000
etc. etc.

I have googled around for "kernel BUG at arch/arm/mm/dma-mapping.c:409!" and found a few references to it but no apparent fix or reason why. I have searched the Linux TV.org forum for clues, but found nothing.

Has anyone encountered this error before? Can someone that is using the above patches try enabling the GSPCA drivers and then plugging in a USB web camera or other USB device for a test?

Thanks

You have the sources, surely you can dig into it, or diff the respective source trees to understand what changed between releases.

__dma_free_buffer(dma_to_page(dev, handle), size);

My guess would be the dev, or handle value are wrong (ie NULL). Some upstream driver code looks to be initializing a structure to zero, but not initializing a pointer.

Sanity check the values of dev, handle, size and dma_to_page(dev, handle)

Adding at91 patch for 2.6.37.

Adding at91 patch for 2.6.38.

Adding at91 patch for 2.6.39.

Some 3.0 Patches?

Adding patch for v3.0.0.

Do these help anybody? I hope so. I cannot test any other platform than my own, so there may be a few issues with some other boards.

I'm wondering if it might be a good idea to split the board-specific stuff apart from the more core components in this patch... or at least prune off the dead/unmaintained boards. I can test the core stuff, but I can't test the board-specific stuff. As we move to kernel v3.1.0, there is a lot of changes going into the board-*.c files..

-Chris

Chris,

I thank you for your continuous effort to port this patch to an up-to-date kernel.

I am sure that it is very useful to many people!

If you plan to submit little chunks of this patch to the upstream Linux kernel, I will be happy to integrate them in our effort to port the most of AT91 support to Mainline Linux. I know that the difficult part is to cut this big patch in single and self contained enhancements.

Best regards,
--
Nicolas Ferre

Adding at91 patch for 3.1.0.

File downloads appear to be broken here at at91.com. So, I have hosted the AT91 3.1 patch here:

https://coledd.com/software/linux/patch ... .patch.bz2

Thanks,
-Chris Cole
cole /at/ coledd.com

Hi, I wanted to know what's the difference between these patches and the ones that have been included in the upstream kernel since v2.6.30 (in kernel's gitweb you can see that a lot of the original patches have already been merged upstream).

These patches are the stuff that never yet got put into the mainline kernel. it is mostly extra boards support. Also contains some spi driver stuff and some macb driver mods, and a few pm mods. This patch does need broken out and some parts fed upstream, and perhaps some parts dropped. Feel free to help out if you have some spare time