02:16 <hanetzer> ahh, think I've figured it. now I just need to figure out bias-pullup and bias-pulldown

18:19 <hanetzer> I wonder... could one setup pinctrl-single so as to be able to group 'sparse' address spaces into one group. the soc has 4 non-contiguous blocks of memory, each under 0x100 in size, so something like reg = <addr1 size1>, <addr2 size2>... and so on would be useful I think.

18:53 <abelloni> I'm not sure there is actually adiffenrece versus having 4 pinctrl-single nodes

18:53 <abelloni> since anyway you won't be able to change multiple pins atomically

19:29 <hanetzer> well, it would make referencing the datasheet easier, as the registers are named continuously across all four io spaces, iocfg_regN

19:29 <hanetzer> wrt pinctrl-single,gpio-ranges and whatnot.

19:36 <hanetzer> on that note, I notice that both pinctrl-single.txt and gpio.txt reference a gpio-ranges property. I wonder about their relation, and the ability to use more than one at a given time

19:48 <hanetzer> blerg. now I need to remember the magic to make initramfs booting work :P

19:50 <hanetzer> tftpboot 0x82000000 uImage.hi3516dv300\; tftpboot 0x84000000 rootfs.cpio.uboot\; bootm 0x82000000 0x84000000 seems like it should have worked (I would use a FIT but that currently screws me with invalid dtb/unknown machine id)