On Tue, Jan 26, 2016 at 03:14:16PM +0000, Charles Keepax wrote:

> As we now free the constraints in regulator_dev_release we will still
> call free on the constraints pointer even if we went down an error path
> in regulator_register, as such we need to clear to the pointer to ensure
> the kfree in regulator_dev_release is a no-op.

No, this is just making things worse - if we're doing the free in the
release we should be doing the free in the release not sometimes maybey
in the release.

Attachment: signature.asc
Description: PGP signature

Reply via email to