On Wed, Jan 11, 2012 at 4:25 PM, Shubhrajyoti [off-list ref] wrote:

On Wednesday 11 January 2012 07:29 PM, Grazvydas Ignotas wrote:

quoted

On Wed, Jan 11, 2012 at 3:21 PM, Shubhrajyoti D [off-list ref] wrote:

quoted

Currently the i2c driver calls the pm_runtime_enable and never
the disable. This may cause a warning when pm_runtime_enable
checks for the count match.Attempting to fix the same by calling
pm_runtime_disable in the error and the remove path.

I remember seeing Felipe doing the reverse to musb here:
http://marc.info/?l=linux-omap&m=132432610700952&w=2
so I'm confused here.

Strange however ?I see many drivers doing the same in the kernel
on greping . Besides I expect a warn to come up.

Felipe could you explain the issue?

When do you see the warning? kfree(dev) should disable runtime pm, but
without waking up the device.

-- 
Felipe Contreras