Hello, > I'd like to take a closer look at the ehci(4) changes; will come back > to you on those soon.
ok, I am waiting. > The sxiccmu(4) and sxisyscon(4) diffs are ok kettenis@ Currentry, cvs.openbsd.org seems to be something in trouble. I cannot connet with "Connection resed". So I will try again in next morning. > The sxipio(4) diff is missing a const but is otherwise identical to > mine. Shall I commit mine? Yes, please. > The sximmc(4) diff isn't quite correct. We need to restrict the > maximum segment size for the "allwinner,sun50i-a100-emmc" ones > otherwise large reads and writes will fail. Also, we can actually > remove the handling of the "allwinner,sun20i-d1-emmc" compatible since > the Allwinner D1 device trees include "allwinner,sun50i-a100-emmc" as > well. > > ok? I think this is ok because Linux driver does not have "allwinner,sun20i-d1-emmc" entry (https://elixir.bootlin.com/linux/latest/source/drivers/mmc/host/sunxi-mmc.c). and the segment size will be: allwinner,sun4i-a10-mmc 0x02000 allwinner,sun5i-a13-mmc 0x10000 allwinner,sun7i-a20-mmc 0x10000 allwinner,sun8i-a83t-emmc 0x10000 allwinner,sun9i-a80-mmc 0x10000 allwinner,sun20i-d1-mmc 0x02000 allwinner,sun50i-a64-mmc 0x10000 allwinner,sun50i-a64-emmc 0x02000 allwinner,sun50i-a100-mmc 0x10000 allwinner,sun50i-a100-emmc 0x02000 Regards, -- SASANO Takayoshi (JG1UAA) <[email protected]>
