[U-Boot] [PATCH v2] x86: Clean up the FSP support codes

[Simon Glass]

Hi Bin,

On 16 December 2014 at 20:30, Bin Meng <bmeng.cn at gmail.com> wrote:
> Hi Simon,
>
> On Wed, Dec 17, 2014 at 11:23 AM, Simon Glass <sjg at chromium.org> wrote:
>> Hi Bin,
>>
>> On 16 December 2014 at 20:12, Bin Meng <bmeng.cn at gmail.com> wrote:
>>> Hi Simon,
>>>
>>> On Tue, Dec 16, 2014 at 1:18 PM, Simon Glass <sjg at chromium.org> wrote:
>>>> Hi Bin,
>>>>
>>>> On 15 December 2014 at 08:03, Bin Meng <bmeng.cn at gmail.com> wrote:
>>>>> This is the follow-on patch to clean up the FSP support codes:
>>>>>
>>>>> - Remove the _t suffix on the structures defines
>>>>> - Use U-Boot's assert()
>>>>> - Use standard bool true/false
>>>>> - Remove read_unaligned64()
>>>>> - Use memcmp() in the compare_guid()
>>>>> - Remove the cast in the memset() call
>>>>> - Replace some magic numbers with macros
>>>>> - Use panic() when no valid FSP image header is found
>>>>> - Change some FSP utility routines to use an fsp_ prefix
>>>>> - Add comment blocks for asm_continuation and fsp_init_done
>>>>> - Add comments to mention find_fsp_header() may be called in a
>>>>>   stackless environment
>>>>> - Add comments to mention init(&params) in fsp_init() cannot
>>>>>   be removed
>>>>>
>>>>> Signed-off-by: Bin Meng <bmeng.cn at gmail.com>
>>>>
>>>> This looks pretty good to me now. I've got a few more comments that
>>>> you can hopefully roll into your final version.
>>>>
>>>> Also I've pushed some updates to u-boot-x86.git branch 'working'.
>>>> Patches 2-7 are the v2 series I just send. Can send a series that
>>>> incorporates those, or put them as your base, perhaps dropping or
>>>> squashing the 'Convert microcode format' patch then resend your
>>>> series? Everything up to your microcode patch is applied to x86/master
>>>> now.
>>>
>>> I will rebase my remaining patches on top of u-boot-x86/master and
>>> send the patches soon.
>>
>> Thanks - also I just sent v3 of the microcode script.
>
> Yep, I see that, will do some test on the v3.
>
>>>
>>>> Then I will retest and push.
>>>>
>>>> BTW I wonder if your series would work with Minnowboard Max?
>>>
>>> There are two versions of the Minnow boards. One is Minnow board based
>>> on Atom E6xx (the same Queensbay platform) and with my series it
>>> should be pretty easy to get U-Boot up and running on that board. The
>>> other version is a newer version called Minnow board Max which has an
>>> Intel Atom E38xx SoC (BayTrail platform). Luckily Intel has released
>>> FSP for BayTrail as well. So it should not take too much effort
>>> supporting Minnow board Max with Intel FSP on top of my series.
>>>
>>> Minnow board: http://www.minnowboard.org/technical-features/
>>> Minnow board Max: http://www.minnowboard.org/meet-minnowboard-max/
>>
>> OK thanks for the info. Is the original board obsolete? I do actually
>> have the minnowboard-max so I will see if the FSP works on it. But
>> I'll wait until your next patch version. I suspect we will want to
>> move some things from queensbay to a common directory if I do that.
>
> I am not sure if the original board is obsolete. These two boards are
> based on two different Intel Atom platforms so I would not say the Max
> board replaces the V1 board. Queensbay is a two chipset solution but
> BayTrail is an SoC. Yep, I originally wanted to put fsp related codes
> under arch/x86/lib however since it is now only validated on one
> platform I decided not to put there. But I did download the BayTrail
> FSP and compared the supported codes with the Queensbay FSP, they are
> almost the same except the FSP VPD/UPD structures (those are platform
> related stuff).

OK I see. Well I'll take a look at some time.

I think the way you have done it is good. We can make things common
when we know what bits are common.

Regards,
Simon