Re: [PATCH] net: phy: marvell10g: add firmware load support

[Baruch Siach <baruch@tkos.co.il>]

Hi Heiner,

On Tue, Mar 31 2020, Heiner Kallweit wrote:
> On 31.03.2020 19:47, Baruch Siach wrote:
>> When Marvell 88X3310 and 88E2110 hardware configuration SPI_CONFIG strap
>> bit is pulled up, the host must load firmware to the PHY after reset.
>> Add support for loading firmware.
>>
>> Firmware files are available from Marvell under NDA.
>>
>
> Loading firmware files that are available under NDA only in GPL-licensed
> code may be problematic. I'd expect firmware files to be available in
> linux-firmware at least.
> I'd be interested in how the other phylib maintainers see this.

The inside-secure crypto acceleration driver
(drivers/crypto/inside-secure/) original had only NDA firmware.

> Two more remarks inline.
>
> Last but not least:
> The patch should have been annotated "net-next", and net-next is closed currently.
>
>> Signed-off-by: Baruch Siach <baruch@tkos.co.il>

[...]

>> +	if ((ret & MV_PMA_BOOT_PROGRESS_MASK) == MV_PMA_BOOT_WAITING) {
>> +		ret = mv3310_load_firmware(phydev);
>> +		if (ret < 0)
>> +			return ret;
>
> You bail out from probe if a firmware file can't be loaded that is
> available under NDA only. That doesn't seem to be too nice.

The code verifies that the PHY is in MV_PMA_BOOT_WAITING state. The PHY
is not usable in this state unless the firmware is loaded. This is just
like the MV_PMA_BOOT_FATAL error in the code above.

In the common case of firmware loaded from SPI flash, the code will not
try to load the firmware.

baruch

>
>> +	}
>> +
>>  	priv = devm_kzalloc(&phydev->mdio.dev, sizeof(*priv), GFP_KERNEL);
>>  	if (!priv)
>>  		return -ENOMEM;
>>

--
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch@tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -