Hi Dmitry,
On Sat, Feb 16, 2013 at 12:49 PM, Dmitry Torokhov
[off-list ref] wrote:
Hi Simon,
On Fri, Feb 15, 2013 at 08:16:12PM -0800, Simon Glass wrote:
quoted
+ for (row = 0; row < ckdev->rows; row++) {
+ if (cros_ec_keyb_row_has_ghosting(ckdev, buf, row))
+ return true;
+ }
No need for curly braces here. I would not care if not for below.
OK I dont't think I even knew about that rule. Actually, what is that rule?
quoted
+
+ return 0;
+
+fail_register:
+ kfree(idev->keycode);
Sorry I did not notice this before, but idev->keycode is devm-managed,
so you either need to use devm_kfree() or just remove call to kfree()
and let it clean up automatically (which will happen if binding fails or
upon removal).
BTW, maybe you should move the whole driver to devm_*? We have
devm_kzalloc() for ckdev and you can use devm_input_allocate_device().
Then you can get rid of entire erro handling path and completely remove
the remove() method as well.
Yes I was thinking about that - might as well do it now.
quoted
+fail_matrix:
+ input_free_device(idev);
+fail_alloc_dev:
+ kfree(ckdev);
+ return err;
+}
+
Thanks.
--
Dmitry
Regards,
Simon