On Sun, May 27, 2012 at 09:39:19AM -0600, Stephen Warren wrote:

On 05/26/2012 10:52 AM, Dong Aisheng wrote:

quoted

On Fri, May 25, 2012 at 10:03 AM, Stephen Warren [off-list ref] wrote:

quoted

On 05/25/2012 07:36 AM, Dong Aisheng wrote:

...

quoted

quoted

If we don't do that, [lock ranges[i].gc] I would argue that we
shouldn't store ranges[i].gc, since it might become invalid - I
believe the only use of it is withinthis function?

In my option, i think it's ok to store it since they're just some data
to describe
hw properties. The gpio function may become invalid but not data.
Is it reasonable to you?

The problem is that if someone tries to dereference the gc field, and
it's no longer valid, which could cause an OOPS. Perhaps we can get away
just with a comment in the struct definition indicating that this field
should only be used by drivers that provided the gc field directly
rather than having it set up by DT, but then why even store it when
creating the ranges from DT in that case?

Yes, you're right.
Maybe we could both not store the gc filed for DT (currently we did not see
the need to store it for dt, right?) and add a comment in the struct definition
as you said. For non-dt users the driver owner should manage that field
correctly with lock since it's provided directly by driver.
Is that ok?

Regards
Dong Aisheng