On 28 January 2015 at 13:41, Tobias Jakobi <liquid.acid@xxxxxxx> wrote: > Hello! > > Jaehoon Chung wrote: >> mmc core supported to hw_reset function. >> So i think we can use it. It's called at only initial time to clear the previous status. >> But i think it can be called at reboot time. (it needs to implement codes.) >> how about? > I don't think that's going the work. The problem here is that depending > on the board configuration, the eMMC might carry the bootloader. If the > eMMC subsystem is not properly reset _during_ reboot, the board won't > even start since no bootloader is found. So we don't even reach the > point where the kernel does anything. I guess it depends on what's being done during the reboot sequence. The most proper thing would be to let the boot loader control the GPIO to trigger the HW reset, but that would mean the boot loader need to know about board specific configurations, such as which GPIO pin. So maybe SOC vendors need to state what GPIO pin to use and don't leave that as a configurable option. >From kernel perspective, the best we can do is to the GPIO, when doing a controlled reset (soft reset, or whatever we call it), but I am not sure where that should be done? Is there a guarantee that the mmc bus' ->shutdown() callback gets called in this sequence? Moreover, adding the reset GPIO as part of the initialization procedure in the mmc core, gives us other benefits and it won't hurt. So no matter, I think it's worth to proceed and discuss Marek's proposal. Kind regards Uffe -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
