On 10/07/2011 03:15 PM, Sascha Hauer wrote: > Hi Gregory, > Hi Sasha, thanks to have taken some time to have a look on my series. > On Thu, Oct 06, 2011 at 10:05:07AM +0200, Gregory CLEMENT wrote: >> This patch set is a RFC about a backlight framework. The purpose of >> this framework is mainly to allow to add easily a support for a >> backlight with the possibility of setting brightness directly from the >> barebox shell using the brightness parameter. >> >> An implementation is provided for i.MX23 by using the PWM. It was >> tested on a custom i.MX23 base board. > > Two things that bother me in this series. > > First thing is that I wonder if it would be better to not > register a seperate device for backlight. How about a call > like this: > > fb_register_backlight(const char *fb, > void (*set_brightness)(int brightness, void *priv)); > > The core would only need a function to find the struct device_d > by the corresponding "fbx" string. This way we could add the > brightness variable to the framebuffer and not a seperate device, > so fb0.brightness=50. > I wasn't entirely convinced by having a driver with a single function. I came to this by several iterations and it didn't lead me in the best direction. That's why I called this series a RFC. I will take in account your idea and propose a new version. > The second thing is that the pwm you use for the mxs backlight > is a generic pwm which not necessarily drives a backlight. We > should have a generic pwm api for this. Otherwise we end up > with different drivers for the same pwm. As I didn't see any pwm framework or API I wasn't sure it was planned or needed to have it in barebox. But I volunteer to work on it. So how do you see it: as an API or as complete driver ? Gregory -- Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
