On 08/02/2012 06:33 PM, Olof Johansson wrote: > On Wed, Aug 1, 2012 at 12:03 PM, Sylwester Nawrocki > <sylvester.nawrocki@xxxxxxxxx> wrote: > >> It wouldn't be clear what specific SoCs the "samsung,exynos5-gsc" compatible >> string applies to, would it ? I believe there are already minor differences >> in GScaler parameters on currently available exynos5 SoC. The variant data >> structures are used to handle this and the compatible string determines which >> variant data structure is selected during driver's initialization. >> If you use a wildcard 'compatible' string this won't be possible any more. >> >> Also it would look odd IMO to have two compatible strings like: >> compatible = "samsung,exynos5-gsc", "samsung,exynos5400-gsc"; > > In this particular case, since you're saying that there are subtle > differences between different part numbers, I'm guessing there's good > reason to go specific, but in general there's no need to avoid > exynos5-gsc. > > Your example is also false, since the strings would be in reverse > order (from specific to generic). That would look perfectly normal. You're right, but my intention was more to say that there would have been two entries in the driver's of_match_table, where "samsung,exynos5-gsc" wouldn't have obvious meaning. Devices within these SoCs tend to differ across part numbers and usually there is one common driver handling them. I can't tell for sure now there are differences, but I would have been surprised if there wouldn't. > So, bottom line: I agree in this particular instance, but I disagree > that it's a hard generic rule. Thanks, sorry if it sounded like I'm advocating it as a general rule. I'm no DT expert whatsoever, but in this particular case it just sounded messy to use only exynos5-gsc. -- Regards, Sylwester -- To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
 |