Re: [Qemu-arm] [PATCH 0/5] ast2400: some cleanups and a simple memory controller model

[Cédric Le Goater]

On 07/12/2016 06:20 PM, Cédric Le Goater wrote:
> Hello Peter,
> 
> On 07/12/2016 04:19 PM, Peter Maydell wrote:
>> On 8 July 2016 at 17:06, Cédric Le Goater <address@hidden> wrote:
>>> Hello,
>>>
>>> Here are a couple of cleanups and a very basic memory controller model
>>> providing a configuration register. This is needed to determine the
>>> RAM size of the SOC, only used by U-Boot as of today.
>>>
>>> Ultimately, we would want to add more support for U-Boot and be able
>>> to boot from flash 0 but the device model of the m25p80 object is not
>>> designed to use a memory region. And so, it is difficult to define a
>>> rom device for the flash backend.
>>>
>>> Some initial patches were proposed but they are not being friendly
>>> with the object interface. Work in progress.
>>
>> Hi; the release process is now at the point where I wouldn't really
>> want to put this into 2.7; I can be persuaded if you have a solid
>> justification, though.
> 
> Here is my take to it  :
> 
> * 1/5 hw/misc: fix typo in Aspeed SCU hw-strap2 property name
>   
>   That's a minor fix
> 
> * 2/5 ast2400: replace aspeed_smc_is_implemented()
> 
>   That is not a fix but it removes some ugliness in the code, which 
>   you had noticed, and which prevents us from fixing possible issues 
>   nicely. So this is a preventive cleanup :)
> 
> * 3/5 ast2400: pretend DMAs are done for U-boot
> 
>   This is a work around fix. U-boot[1] uses DMAs to calibrate the 
>   SPI flash module. It will spin until the DMA transfers are done.
> 
> * 4/5 ast2400: externalize revision numbers
> 
>   This is minor fix de-duplicating code and a prereq for the next
>   patch.
> 
> * 5/5  ast2400: add a memory controller device model
> 
>   That's an enhancement. It adds a basic device model for the memory
>   controller. This is needed by some version of the SDK relying on
>   the configuration register to know how much RAM the SOC has. Older
>   versions used a hardcoded value.  

oh yes and that's a v2. first version was sent on 06/29/2016.

Thanks,

C. 

> The first patches should not be too much of a problem regarding the 
> soft freeze. The last can wait 2.8 if you prefer, we need more work 
> to fully support u-boot anyhow. I would be happy to have review 
> though :) The model is very much like the SCU.
> 
> Thanks,
> 
> C.
> 
> [1] not supported yet in qemu because of the lack of a rom device 
>     memory region needed to boot from the flash module.
> 
>