From mboxrd@z Thu Jan  1 00:00:00 1970
From: Marcin Niestroj <m.niestroj@grinn-global.com>
Date: Fri, 13 Apr 2018 11:42:31 +0200
Subject: [U-Boot] chiliSOM: USB bug
In-Reply-To: <72160287-31d5-781c-4ff9-9d91ad749de7@onet.eu>
References: <9ce549aa-79a9-81e4-b569-b9a5d574850d@onet.eu>
 <4c2b8ff1-0b8f-85ea-f4bc-f5d51105d5c7@grinn-global.com>
 <f7775296-addf-4e19-2181-b415a8bb223c@onet.eu>
 <dfe90074-a365-42e1-8c2b-c7ffa8d1f249@grinn-global.com>
 <5c725650-8d3a-1a0b-28be-89b835f1efeb@onet.eu>
 <b561c889-2e00-71ee-66ac-9c8b88b2e564@grinn-global.com>
 <13e8ed8d-f302-05d8-9bfc-8c43de0e7487@onet.eu>
 <c96f6b09-8662-9ab2-4cf5-9251e692d67f@grinn-global.com>
 <72160287-31d5-781c-4ff9-9d91ad749de7@onet.eu>
Message-ID: <c0990dad-9e35-74b1-7583-b7dd83f60780@grinn-global.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit
To: u-boot@lists.denx.de



On 13.04.2018 11:12, sdrb wrote:
> Marcin Niestroj wrote:
>>
>> On 12.04.2018 13:35, sdrb wrote:
>>> Marcin Niestroj wrote:
>>>>
>>>>
>>>> On 12.04.2018 11:09, sdrb wrote:
>>>>> Marcin Niestrój wrote:
>>>>>>
>>>>  >> <snip>
>>>>>>
>>>>>> Could you describe what is you BOOT[4:0] configuration? And you 
>>>>>> want to
>>>>>> boot from MMC1, right?
>>>>>
>>>>> If I read the BOOT[4:0] properly it is: 0011.
>>>>> I want to boot from SD card.
>>>>
>>>> Please provide these 5 pins (including 4 and 0) configuration.
>>>>
>>>
>>> Sorry for misprint - SYSBOOT[4:0] pins are as follows: 00011.
>>
>> Unfortunately I can't reproduce this issue on my side with chiliboard.
>> It's hard to tell what could be the problem when using you board.
>> What RAM vendor is on chiliSOM you have on your board? Maybe there
>> DDR is not initialized with proper timing? Don't have other ideas for
>> now.
>>
> 
> If I identified properly the chip it is Samsung.
> 

We had an issue with Samsung DDR on chiliSOM with one of our clients
and his board. We have used a workaround patch to sleep between
DDR initialization and copying u-boot.img to RAM. If there was no delay,
then RAM memory was in undefined behavior, resulting in very strange
runtime errors. We used u-boot 2014.07 back then.

However I see that commit 69b918b65d11 ("am33xx,ddr3: fix ddr3 sdram
configuration") already addresses this issue, but in a different way
than we did back then.

So I have no idea now what goes wrong in your case.

-- 
Marcin Niestroj