3.8 kernel: support legacy pin mux

GoogleGroups
1

Jason

Here is a working attempt on 3.8 kernel to get the same kind of legacy
userspace mux debugfs entry (/sys/kernel/debug/omap_mux/...)

This looks much easier to me right now to get going than
https://github.com/jadonk/validation-scripts/tree/master/test-capemgr

Hope its of some use.

regards

0001-omap_mux-debugfs-legacy-support.patch (2.78 KB)

2

Jason

Here is a working attempt on 3.8 kernel to get the same kind of legacy
userspace mux debugfs entry (/sys/kernel/debug/omap_mux/...)

This looks much easier to me right now to get going than
https://github.com/jadonk/validation-scripts/tree/master/test-capemgr

Hope its of some use.

Couldn't you have at least used the pinctrl API internally instead of
skipping over the management of pinmux in the kernel entirely? I like
the compatibility and simplicity aspects, but if I'm going to get
Pantelis and Koen to accept putting this in the "vendor" tree, I'm not
sure this will do it.

3

Jason

Here is a working attempt on 3.8 kernel to get the same kind of legacy
userspace mux debugfs entry (/sys/kernel/debug/omap_mux/...)

This looks much easier to me right now to get going than
https://github.com/jadonk/validation-scripts/tree/master/test-capemgr

Hope its of some use.

Couldn't you have at least used the pinctrl API internally instead of
skipping over the management of pinmux in the kernel entirely? I like
the compatibility and simplicity aspects, but if I'm going to get
Pantelis and Koen to accept putting this in the "vendor" tree, I'm not
sure this will do it.

A pointer to the documentation to provide a bit of encouragement:
https://github.com/torvalds/linux/blob/master/Documentation/pinctrl.txt

4

Not posting it as a github pull request also goes a long way to being ignored for inclusion.

5

> Jason
>
> Here is a working attempt on 3.8 kernel to get the same kind of legacy
> userspace mux debugfs entry (/sys/kernel/debug/omap_mux/..**.)
>
> This looks much easier to me right now to get going than
> https://github.com/jadonk/validation-scripts/tree/
master/test-capemgr<https://github.com/jadonk/validation-scripts/tree/master/test-capemgr>
>
> Hope its of some use.

Couldn't you have at least used the pinctrl API internally instead of
skipping over the management of pinmux in the kernel entirely? I like
the compatibility and simplicity aspects, but if I'm going to get
Pantelis and Koen to accept putting this in the "vendor" tree, I'm not
sure this will do it.

Not posting it as a github pull request also goes a long way to being
ignored for inclusion.

Randy,

Is there a chance you can rework the patch to include usage of pinctrl and
provide a github pull request?

6

Jason

7

Does it really make sense to ignore the work upstream for this feature?

http://thread.gmane.org/gmane.linux.kernel/1469197

Contribute to that effort if it’s insufficient for all use cases or pull it in as is.

-Matt

8

I've pointed to that patch multiple times already. It only needs someone to sit down and backport the new pinctrl API additions that patch needs to our 3.8 tree.

9

cd kernel
patch -p1 < abc.patch

10

HI,
your patch ask’s for a file to patch. But I dont’t know what to type in…
Before it say’s : can’t find to patch at input line 19
Does any body know what to do?

Thank you