On Wednesday 24 August 2011 09:32 AM, Rajendra Nayak wrote:

On 8/23/2011 8:04 PM, Santosh wrote:

quoted

+ Rajendra and Benoit to comment on optional clock
handling.

On Thursday 04 August 2011 04:34 PM, Tarun Kanti DebBarma wrote:

[....]

quoted

quoted

+ if (IS_ERR_VALUE(pm_runtime_get_sync(bank->dev)< 0))
+ dev_err(bank->dev, "%s: GPIO bank %d "
+ "pm_runtime_get_sync failed\n",
+ __func__, bank->id);
+
for (j = 0; j< hweight_long(bank->dbck_enable_mask); j++)
clk_enable(bank->dbck);

Optional clock handling also should been addressed using runtime hooks,
isn't it ?

Whats handled by runtime hooks is the device;s *main* clocks, which are
needed for the device to become accessible.
Anything else *optional* which the device would need optionally based on
the various modes it can operate etc are still controlled by the driver
by enabling/disabling them as and when needed using the clock framework.

Thanks Rajendra for clarification. In summary, we continue to use clock
framework for the optional clock handling from the drivers.

Regards
Santosh