On 22/08/2025 17:16, Mike Looijmans wrote:
>>> +
>>> +static struct i2c_driver tmds181_driver = {
>>> + .driver = {
>>> + .owner = THIS_MODULE,
>> Nice coincidence - this stars in one of my talks on OSSE
>> (https://sched.co/25VoV) as my litmus test for crazy old, vendor code.
>> Please come to the session if you are around or just check the slides
>> afterwards.
>
> Oh, Amsterdam this year. Hadn't spotted that in time, or I would have
> been there!
>
> I added coccinelle to my toolkit (would make a good addition to patman).
> Indeed would have prevented the old-skool I2C stuff that I've been
> lugging along in drivers for a decade...
>
> I'll fix those in v4.
>
> And yeah, the real issue is that I didn't submit this driver seven years
> ago. In my defense, the other guys also didn't.
>
> For your talk, maybe add a tip on how to avoid the verbose output.
> make C=1 CHECK=scripts/coccicheck drivers/gpu/drm/bridge/ti-tmds181.o
> I get a gazillion lines of "/usr/bin/spatch -D report ..." and it's
> difficult to spot the warnings. Tried adding "V=0" but that didn't make
> a difference, piping through "grep :" helped a bundle.
make coccicheck M=drivers/gpu/drm/bridge/ DEBUG_FILE=cocci.err
Best regards,
Krzysztof
