Re: [PATCH] nvmem: layouts: u-boot-env: remove ethaddr check

[Miquel Raynal <miquel.raynal@bootlin.com>]

On 29/04/2026 at 14:41:36 -07, Rosen Penev <rosenp@gmail.com> wrote:

> On Wed, Apr 29, 2026 at 12:44 AM Miquel Raynal
> <miquel.raynal@bootlin.com> wrote:
>>
>> On 29/04/2026 at 09:39:53 +02, Miquel Raynal <miquel.raynal@bootlin.com> wrote:
>>
>> > Hi Rosen,
>> >
>> >> --- a/drivers/nvmem/layouts/u-boot-env.c
>> >> +++ b/drivers/nvmem/layouts/u-boot-env.c
>> >> @@ -75,11 +75,9 @@ static int u_boot_env_parse_cells(struct device *dev, struct nvmem_device *nvmem
>> >>              info.offset = data_offset + value - data;
>> >>              info.bytes = strlen(value);
>> >>              info.np = of_get_child_by_name(dev->of_node, info.name);
>> >> -            if (!strcmp(var, "ethaddr")) {
>> >> -                    info.raw_len = strlen(value);
>> >> -                    info.bytes = ETH_ALEN;
>> >> -                    info.read_post_process = u_boot_env_read_post_process_ethaddr;
>> >> -            }
>> >> +            info.raw_len = strlen(value);
>> >> +            info.bytes = ETH_ALEN;
>> >
>> > How do you know this is a MAC address?
>>
>> You are overwriting info.bytes above. I am sorry but this change has
>> only been tested with your very own settings and does not look like
>> something we could accept.
>>
>> You cannot attach the "process ethaddr" subprocess to any node, that
>> would not make any sense.
>>
>> I guess we can extend a bit the condition with thinks like "eth%daddr",
>> but that must remain limited.
> how about a static array of the names and strcmp over that?

You can check for eth on the first offsets, and then addr starting
from offset 4, allowing up to 10 ethernet interfaces to by dynamically
assigned, that should be enough. If not enough, you can also try at
offset 5, which would bring that up to 100 possible interfaces. I belive
you can directly give the strings to strcmp, no need for a global
declaration.

Thanks,
Miquèl