From mboxrd@z Thu Jan  1 00:00:00 1970
From: Holger Brunck <holger.brunck@keymile.com>
Date: Mon, 21 Nov 2011 15:27:41 +0100
Subject: [U-Boot] [PATCH] arm, post: add missing post_time_ms for arm
In-Reply-To: <4ECA5BA2.50108@denx.de>
References: <1316064873-5500-1-git-send-email-hs@denx.de><4ECA4AD1.70307@keymile.com>
	<4ECA5BA2.50108@denx.de>
Message-ID: <4ECA5FDD.9030207@keymile.com>
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

Hi Heiko,

On 11/21/2011 03:09 PM, Heiko Schocher wrote:
> Holger Brunck wrote:
>> On 09/15/2011 07:34 AM, Heiko Schocher wrote:
>>> Signed-off-by: Heiko Schocher <hs@denx.de>
>>> Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
>>> ---
>>>  post/post.c |    2 +-
>>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/post/post.c b/post/post.c
>>> index 6c59f91..0bfca75 100644
>>> --- a/post/post.c
>>> +++ b/post/post.c
>>> @@ -488,7 +488,7 @@ void post_reloc(void)
>>>   */
>>>  unsigned long post_time_ms(unsigned long base)
>>>  {
>>> -#ifdef CONFIG_PPC
>>> +#if defined(CONFIG_PPC) || defined(CONFIG_ARM)
>>>  	return (unsigned long)(get_ticks() / (get_tbclk() / CONFIG_SYS_HZ))
>>>  		- base;
>>>  #else
>>
>> this commit breaks post test support for our keymile arm boards:
> 
> :-(
> 
> I am sure I tried a "MAKEALL arm" ... Hmm.. how did the post support

I guess that our post support wasn't mainlined at the time you test the patch.

> worked for you before this patch? You should have seen this warning message:
> 
> #warning "Not implemented yet"
> 
> or?
> 

Yes we got these warning in the past but this wasn't a problem for us, because
we don't use these timer informations in our framework.

>>> make km_kirkwood
>> [...]
>> post/libpost.o: In function `post_time_ms':
>> /daten/ws_root/hbu/u-boot-km.git/post/post.c:504: undefined reference to `get_ticks'
>> /daten/ws_root/hbu/u-boot-km.git/post/post.c:504: undefined reference to `get_tbclk'
>>
>> So it seems that the kirkwood code has no implementation for these functions.
>> Any idea how to fix this?
> 
> We must implement it ... Prafulla, can you comment this?
> 

Or in the first step just add && !(CONFIG_MACH_KIRKWOOD) or similar to the
#ifdef until it is supported in the generic kirkwood code.

regards
Holger