[U-Boot] [PATCH V5 4/4] edminiv2: add mvsata_ide and cmd_ide support

[Albert ARIBAUD]

Hi Prafulla et al.,

Le 01/08/2010 18:48, Prafulla Wadaskar a écrit :
>
>
>> -----Original Message-----
>> From: Albert ARIBAUD [mailto:albert.aribaud at free.fr]
>> Sent: Friday, July 23, 2010 4:21 PM
>> To: Prafulla Wadaskar
>> Cc: u-boot at lists.denx.de
>> Subject: Re: [PATCH V5 4/4] edminiv2: add mvsata_ide and
>> cmd_ide support
> ...snip...
>>
>>> Secondly I tried to generate similar patches for Kirkwood
>> oneprd_base (attached) on the top of this patch series
>>> But somehow "ide reset" is unable to detect device on port.
>>> If you have openrd_base board, can you pls give a try?
>>   >
>>   >  May be you may need to look at mpp config.
>>
>> I have an OpenRD Client, not base; client has a SATA
>> connector to which
>> I have a disk connected, so I could give it a try but we'll
>> need to make
>> sur how many SATA ports kirkwood has and which ones are used
>> on either
>> openrd client and base (orion5x has two sata ports and edmini
>> uses port
>> 1, not 0, for its sole SATA conector). I am away from home right now,
>> but I will be able to give it a try as soon as August 1st.
>
> Hi Albert
> Ping ...if you are back

I am, and I have done some work on your problem, but for the moment I am 
still at the same stage as you -- a continuous sequence of dots when 
doing an ide reset, right?

I have #define'd DEBUG in cmd_ide.c and compared the debug messages for 
my ED Mini V2 (working) and my OpenRD-Client (not working). I have also 
performed manual SATA and IDE reset sequences. Short story:

1) the board code correctly initializes the SATA link, which correctly 
detects a disk if there is one:

  Marvell>> md.l f1082300 4
f1082300: 00000113 14010000 00000300 010300b0    ................
Marvell>> md.l f1084300 4
f1084300: 00000000 00000000 00000300 010300b0    ................
Marvell>>

(this is consistent with my system which has a 1.5 GBps disk on port 0).

2) The problem is that the drive appears continuously non-RDY (SStatus 
register remains at 0x80 while on the ED Mini V2 it ends up at 0x50, 
which basically means it is ready).

I will now compare the MPP / GPIO settings performed by the 
FLASH-resident U-Boot (its 'ide reset' command works ok) to those in 
effect when the patched u-boot fails. Maybe one GPIO controls power to 
the internal SATA port, although I doubt it -- if the drive was off, the 
controller would not see it at all, and SStatus at F1082300 would not 
report it present.

> Pls include my email address in your white list, the mails to you bounces back to me.

It's not an issue of (not) whitelisting your e-mail address; it's your 
mail server being blacklisted by my email provider. If you have a recent 
non-delivery report please send it to me in private to my gmail account 
(albert.aribaud at gmail.com), I'll try to sort things out and contact your 
mail server's admin with the relevant details.

> Regards..
> Prafulla . .

Amicalement,
-- 
Albert.