Re: Allwinner D1

[Nick Hudson <nick.hudson%gmx.co.uk@localhost>]

On 05/02/2025 03:44, Rui-Xiang Guo wrote:
> On Sat, Feb 01, 2025 at 01:52:17PM +0800, Rui-Xiang Guo wrote:
> > After enabling clint in the firmware instead of attaching sunxitimer,
> > the host controller starts activities but can't detect the card.
> > [...]
> > [   1.6914471] sunximmc0: mmc intr idst=00000000 mint=00000004
> > [   1.6914471] sunximmc0: mmc intr idst=00000000 mint=00000808
> > [   1.6914471] sunximmc0: host controller error, mint=0x00000808
> > [   1.7099755] sdmmc1: 4-bit width, 50.000 MHz
> > [   1.7099755] sunximmc0: i/o error 5
> > [   1.7099755] sunximmc0: update clock
> > [   1.7220301] sdmmc0: sdmem_mem_send_scr: error = 5
> > [   1.7220301] sdmmc0: SD_SEND_SCR send failed.
> > [   1.7309457] sdmmc0: mem init failed
> > [   1.7309457] sdmmc0: init failed
> > [   1.7309457] sunximmc0: opcode 7 flags 0x0 data 0x0 datalen 0 blklen 0
> > [   1.7439451] sunximmc0: cmdval = 80000000
> > [   1.7439451] sunximmc0: mmc intr idst=00000000 mint=00000004
> > [   1.7534039] sunximmc0: width = 1
> > [   1.7534039] sunximmc0: update clock
> > [...]
> > The complete log is in:
> > https://github.com/picohive/netbsd-mangopi-mq-pro/blob/main/boot-debug-mmc.log
> >
> > and the diff:
> > https://github.com/picohive/netbsd-mangopi-mq-pro/blob/main/sunxi_mmc.c.diff
>
> Would it be related to pmap? OpenBSD's pmap contains special handling for the
> T-Head cores.

oh, yeah. mostly likely.

t-head can do uncached mappings via PTE. I was hoping to get to this
soon(ish)

Nick