Re: [PATCH] gianfar_ptp: use iomem, not ioports resource tree in probe

[Timur Tabi <timur@freescale.com>]

Richard Cochran wrote:

> Getting back to your really ignorant comment, I suggest that you look
> at this review. It was made by Grant Likely. Perhaps you have heard of
> him?
> 
>   https://lkml.org/lkml/2011/2/23/281
> 
> I was the original developer of the PTP code, and my code went through
> fifteen rounds of review. And guess what - I actually listened to the
> reviewer's comments and changed my work accordingly.

I'm sorry for what I said.  It was inappropriate.  I deal with crappy code
from co-workers on a daily basis, and sometimes I forgot that just because
something is done differently than I would have done it, that doesn't mean
it's wrong.

> You can read all about what happened, but you will have to find v15
> yourself. Be sure to pay special attention to the history of
> irq_of_parse_and_map() verses platform_get_irq().

I actually did that research and saw Grant's comments.  I asked him about
it on IRC, and although I understand his reasoning, I'm not sure I agree
with all of it.  In particular, I think
platform_get_resource/request_resource/ioremap is less elegant than just
calling of_iomap.


-- 
Timur Tabi
Linux kernel developer at Freescale