Re: [PATCH 0/1] hw/arm/aspeed: Add Fuji machine type

[Cédric Le Goater]

On 9/2/21 5:46 PM, Patrick Williams wrote:
> On Thu, Sep 02, 2021 at 03:12:12PM +0000, Peter Delevoryas wrote:
>>
>>> On Sep 2, 2021, at 1:12 AM, Cédric Le Goater <clg@kaod.org> wrote:
> 
>>> Are we sure the flash models fit what the machine expects ?
>>
>> Oh, I see, I’m not exactly sure which device is being used
>> but it should be one using the m25p80 command set, and after
>> switching to one of those devices (e.g. mx66l1g45g) it works:
>>
>> [   54.867027] Creating 6 MTD partitions on "spi0.0":
>> [   54.867506] 0x000000000000-0x0000000e0000 : "u-boot"
>> [   54.873929] 0x0000000e0000-0x0000000f0000 : "env"
>> [   54.876546] 0x0000000f0000-0x000000100000 : "meta"
>> [   54.878948] 0x000000100000-0x000007800000 : "fit"
>> [   54.881290] 0x000007800000-0x000008000000 : "data0"
>> [   54.883965] 0x000000000000-0x000008000000 : "flash0"
> 
> It appears that machine uses the 128MB flash layout dtsi file:
>     
> https://github.com/torvalds/linux/blob/master/arch/arm/boot/dts/facebook-bmc-flash-layout-128.dtsi
> 
> I think that means it would be safe to use the same mx66l1g45g that Sonorapass
> is using in the Fuji config.  Likely we're either using that exact chip or
> something 100% compatible with it.
> 
>>>
>>>  [   41.314338] 6 fixed-partitions partitions found on MTD device spi0.0
>>>  [   41.314600] Creating 6 MTD partitions on "spi0.0":
>>>  [   41.314957] 0x000000000000-0x0000000e0000 : "u-boot"
>>>  [   41.318355] 0x0000000e0000-0x0000000f0000 : "env"
>>>  [   41.319747] 0x0000000f0000-0x000000100000 : "meta"
>>>  [   41.321064] 0x000000100000-0x000007800000 : "fit"
>>>  [   41.321247] mtd: partition "fit" extends beyond the end of device 
>>> "spi0.0" -- size truncated to 0x3f00000
>>>  [   41.322684] 0x000007800000-0x000008000000 : "data0"
>>>  [   41.322827] mtd: partition "data0" is out of reach -- disabled
>>>  [   41.324110] 0x000000000000-0x000008000000 : "flash0"
>>>  [   41.324253] mtd: partition "flash0" extends beyond the end of device 
>>> "spi0.0" -- size truncated to 0x4000000
>>>
>>> More info in the log below.
>>>
>>>
>>> Could we also define the list of expected I2C devices of the machine ? 
>>
>> Yeah sure! I’ll follow up with that.
>>
> 
> The fuji DTS is available upstream:
>     
> https://github.com/torvalds/linux/blob/master/arch/arm/boot/dts/aspeed-bmc-facebook-fuji.dts
> 
> I found that when I did Sonora Pass there were a number of unsupported i2c
> devices (ie. devices without QEMU models) but most of them were behind muxes,
> which are similarly non-supported.  You'll see a pretty large fan-out in the 
> i2c
> busses of that DTS.  I don't see anything obvious in the DTS that is directly
> connected except for the muxes.

The pca954x muxes are supported and these devices : 

  ...
  3-0048 (lm75) - no driver bound
  unable to find driver for 3-0048
  3-0049 (lm75) - no driver bound
  unable to find driver for 3-0049
  3-004a (lm75) - no driver bound
  unable to find driver for 3-004a
  3-004c (tmp422) - no driver bound
  Failed to open /sys/bus/i2c/devices/50-0052/eeprom: No such file or directory
  Failed to parse default EEPROM

are also. It would be good to have them defined in the machine. 

Thanks,

C.