On Thu, Jun 30, 2016 at 08:38:03AM +0300, Artturi Alm wrote: > On Thu, Jun 30, 2016 at 03:21:18PM +1000, Jonathan Gray wrote: > > On Wed, Jun 29, 2016 at 05:39:47PM +0300, Artturi Alm wrote: > > > On Wed, Jun 29, 2016 at 02:44:29PM +1000, Jonathan Gray wrote: > > > > On Wed, Jun 29, 2016 at 04:21:39AM +0000, Karl Hammerschmidt wrote: > > > > > Is anyone else interested in their boards? > > > > > > > > > > They were very willing to provide extra documentation when asked, > > > > > including a schematic for one of their LCD drivers and an unpublished > > > > > datasheet for the AXP228 PMIC used on some of their boards. They also > > > > > offered free boards to the FreeBSD guys several months ago, but there > > > > > were no replies. > > > > > > > > > > It would be a different SoC, though (Allwinner H3, Samsung 4412, 4418, > > > > > 6818). On a positive note, I haven't found any mention yet of closed > > > > > binaries like the ODRIOD C1. (Doesn't mean there aren't any, and I'm > > > > > still looking...) > > > > > > > > > > > > > The focus is on the cubox i4pro and to a lesser extent the beaglebone > > > > black. Boards that aren't included in the main u-boot repository > > > > are unlikely to support the bootloader which requires efi. > > > > > > > > > > Wondering about how i should interpret that; i got yet another board with > > > SoC > > > by Allwinner/sunxi(Pine64+) waiting at the post office for pickup, that i > > > had > > > already began writing the support for(CortexA53) into armv7. Wasted > > > effort? > > > > > > -Artturi > > > > > > > Why not finish support for the a10 first? It still lacks an mmc driver. > > > > http://marc.info/?l=openbsd-arm&m=146352442215755&w=2 > ^s/arm88k/armv7/ for two #includes, and you'll have a working mmc driver. > I was unaware what the code for looking up the pin to use for CD should > look like w/fdt, so i didn't include it, but the driver does work, as is.
You shouldn't claim all those capabilities without checking a register or fdt for them. Card detect gpio pins can be determined via "cd-gpios" in fdt, bus-width can also be determined via fdt. If you can convert it to fdt, perhaps rename it to sximmc, verify it works on real hardware then we could look at getting it committed.
