06:40 <milkylainen> hmm. basic, stupid, driver question. I have an efuse nvmem driver. It controls wp with gpio that's picked up by the nvmem core from dt and handled before and after calling the write cb. The struct nvmem_config is mine but nvmem_device is core private afaict. Is there a more proper way to check if core picked up a gpio so that write works?

06:45 <milkylainen> nvmem core allows me to pick up the gpio myself, but I'd rather not over such a trivial thing.

09:04 <robmur01> milkylainen: I guess you could check your DT node for the presence of "wp-gpios", then you know either you're read-only, or that nvmem_register() will fail if the GPIO is specified but unavailable

09:05 <robmur01> looks like it might be nice if the API had a "I'll definitely be read-only if there's no GPIO" flag the driver could pass in the initial config

10:57 <milkylainen> robmur01: Hmm. I guess I could if the register it is guaranteed to fail.

10:57 <milkylainen> robmur01: I was thinking similarly. Like, being helpful instead of cluelessly failing writes.

11:28 <robmur01> of course even if you know you have an external WP pin, and have no GPIO for it, it doesn't necessarily mean that it's not, say, wired to a physical switch that the user can toggle after you've probed :/

17:12 <robher> geertu: Any comments on my series of cleanups and alternative printk locking fix?: https://lore.kernel.org/all/20230801-dt-changeset-fixes-v2-5-c2b701579dee@kernel.org/#r