Hello Kevin,

Thank you for your review.

On Fri, Nov 12, 2010 at 4:16 PM, Kevin Hilman
<[email protected]> wrote:
>
> Ben Gardiner <[email protected]> writes:
>
> > When the RMII PHY on the UI board is enabled with CONFIG_DA850_UI_RMII then 
> > then
> > following will be printed to the console when warnings are also enabled:
> >
> > WARNING: at drivers/gpio/gpiolib.c:1567 __gpio_set_value+0x4c/0x5c()
> > Modules linked in:
> > [<c002c6ac>] (unwind_backtrace+0x0/0xf8) from [<c003b48c>] 
> > (warn_slowpath_common+0x4c/0x64)
> > [<c003b48c>] (warn_slowpath_common+0x4c/0x64) from [<c003b4c0>] 
> > (warn_slowpath_null+0x1c/0x24)
> > [<c003b4c0>] (warn_slowpath_null+0x1c/0x24) from [<c01aed60>] 
> > (__gpio_set_value+0x4c/0x5c)
> > [<c01aed60>] (__gpio_set_value+0x4c/0x5c) from [<c0033bd4>] 
> > (da850_evm_ui_expander_setup+0x1e4/0x2
> > 44)
> > [<c0033bd4>] (da850_evm_ui_expander_setup+0x1e4/0x244) from [<c02e2e1c>] 
> > (pca953x_probe+0x1f8/0x29
> > 0)
> > <snip>
> >
> > Traced the WARN_ON to the gpio_set_value(rmii_sel,0) call in
> > da850_evm_setup_emac_rmii. Replacing the call with the _cansleep variant
> > results in no more warning.
>
> OK, looks like GPIOs on the GPIO expander used for the UI expander
> (pca953x) are of the 'can_sleep' variety, so any users of these should
> be cansleep.
>
> I suspect you'll still see the similar warnings in the 'teardown' path,
> since those on the same expander.

I can't say I've ever executed the teardown path for the UI expander
but I agree.

> Care to update/test that path as well?
>
> Also, can you update the subject to something like:
>
>      da850-evm: UI expander GPIO usage can sleep, use _cansleep

Will do; I expect to have a re-spin posted Monday.

Best Regards,
Ben Gardiner

---
Nanometrics Inc.
http://www.nanometrics.ca
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source

Reply via email to