[PATCH 1/2] bitops: Guard GENMASK and GENMASK_ULL against prior defines

[Quentin Schulz]

Hi Simon,

Thanks for doing that! Did you run this in our GitLab CI by any chance?

I'm a bit wary of the sandbox implementation since it isn't actually 
interacting with an EC, but at least *some* test being run in CI would 
be nice :)

On 4/30/26 1:38 PM, Simon Glass wrote:
> Some headers - notably the ec_commands.h synced from upstream
> Chrome OS EC - define GENMASK and GENMASK_ULL themselves before
> linux/bitops.h is pulled in. The unguarded definitions here then trigger
> a 'redefined' warning at the second site, even though both spellings
> produce the same mask.
> 
> Wrap the two macros in '#ifndef' guards so a pre-existing definition
> wins. The U-Boot expressions are unchanged.
> 

Considering how often we may sync cros_ec compared to 
include/linux/bitops.h, I'm tempted to patch cros_ec instead and leave 
include/linux/bitops.h alone.

I'll comment in the next patch what I think may be enough.

Cheers,
Quentin