From mboxrd@z Thu Jan  1 00:00:00 1970
From: Reinhard Meyer <u-boot@emk-elektronik.de>
Date: Thu, 09 Jun 2011 16:10:14 +0200
Subject: [U-Boot] [PATCH] AT91 rework: pm9261, pm9263 and pm9g45
In-Reply-To: <4DF0B9C0.2020001@ronetix.at>
References: <4DF08968.4080501@ronetix.at> <4DF0A71E.7020100@emk-elektronik.de>
	<4DF0B9C0.2020001@ronetix.at>
Message-ID: <4DF0D446.3040106@emk-elektronik.de>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

Dear Asen Dimov,
> Hello Reinhard,
> 
> On 06/09/2011 01:57 PM, Reinhard Meyer wrote:
> ...
>> Dear Asen Dimov,
>> The empty reset_timer() function added there can obviously only
>> solve build issues.
>> On any account reset_timer() must not be used anymore.
>> As such this patch must get a NAK.
>>
> The architectures, except AT91 are using reset_timer() to make epochs
> (start from zero). I don't want to break the other architectures
> and I need the CFI driver for pm92613 and pm9261. I can not think of
> another idea, except an empty reset_timer(). Any suggestions, ideas?

1. an empty reset_timer() will allow you to build, but at runtime it
must break.

2._if_ the current CFI driver is based on reset_timer()
that should actually be fixed there.
Whereby the actual impact on other architectures that have a broken
get_timer() implementation and therefore _require_ reset_timer() is
unclear to me. We just had a lengthy discussion about timer API etc.
In essence, this discussion, whatever exact API will be implemented,
resulted in NOT having any reset_timer() and only a monotonous,
millisecond returning get_timer() function.

The only _interim_ solution would be to reintroduce the original reset_timer()
to AT91, which I am NOT fond of.

Wolfgang?

Best Regards,
Reinhard