On Tue, May 27, 2014 at 9:59 AM, Feng Kan <fkan@xxxxxxx> wrote: > On Tue, May 27, 2014 at 1:25 AM, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote: >> On Wed, May 21, 2014 at 1:54 AM, Feng Kan <fkan@xxxxxxx> wrote: >> >>> I see each bank as separate gpio chip. It is a simple way to >>> abstract the banks since they can operate independently. >> >> I think they should also be separate devices, and separate nodes >> in the device tree. >> >>> It also >>> provided me a way to fix the sysfs gpio base number, regardless if >>> a particular bank node is pulled out. >> >> The GPIO sysfs is unpredictable for this and many other >> reasons and should not be relied upon. >> >>> This is also done in similar way >>> in some other gpio drivers such as the dwapb gpio driver. >> >> I think the dwapb has it's registers mingled while your banks >> seem to be separate chunks. > > Thanks for the comments, I will fix these problems. Just some details > in terms of the register layout. I will break out the gpio to separate > device nodes as instructed, I will have to use the reg attribute to offset > bank registers. Please let me know if it is acceptable. > > gfc_gpio_fl0 GFC42'h1701c00c > gfc_gpio_fl0_mux GFC42'h1701c010 > gfc_gpio_fl1 GFC42'h1701c018 > gfc_gpio_fl1_mux GFC42'h1701c01c > gfc_gpio_fl2 GFC42'h1701c024 > gfc_gpio_fl2_mux GFC42'h1701c028 > gfc_gpio_fl0_od GFC42'h1701c030 > gfc_gpio_fl1_od GFC42'h1701c034 > gfc_gpio_fl2_od GFC42'h1701c038 Please kindly give some guidance on this, as you can see the individual gpio bank is interleaved in the address space, it is not likely to break this into multiple device nodes. > > > >> >> Yours, >> Linus Walleij -- 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