On 23.11.2021 23:17, Florian Fainelli wrote:
On 11/22/21 2:00 PM, Rafał Miłecki wrote:
On 22.11.2021 22:51, Florian Fainelli wrote:
On 11/15/21 1:11 AM, Rafał Miłecki wrote:
From: Rafał Miłecki <rafal@xxxxxxxxxx>
Broadcom used 2 LEDs hardware blocks for their BCM63xx SoCs:
1. Older one (BCM6318, BCM6328, BCM6362, BCM63268, BCM6838)
2. Newer one (BCM6848, BCM6858, BCM63138, BCM63148, BCM63381, BCM68360)
Just so the existing pattern/regexps continue to work, I would be naming
this "bcm63xx" to be consistent with the rest of existing code-base.
The problem I saw with "bcm63xx" is that it seems to match all SoCs:
those with old block and those with new block. So I guess both groups
have the same right to use that "bcm63xx" based binding.
To avoid favouring old or new block I decided to avoid "bcm63xx".
Given above explanation: do you still prefer using "bcm63xx" based
binding for the new block? I'm OK with that, I just want to make sure
you're aware of that minor issue. Please let me know :)
Maybe we use leds-bcm63138.c then since this is the first chip in the
list that featured that block, similar to how leds-bcm6328.c was
created? Then my second choice would be leds-bcm63xx.c just so the
existing patterns match, really and because it's easy to visually not be
able to tell the difference between two x versus three x.
Sounds good to me, thanks for your review!
