Re: Milk-V Duo S?

[Andrew Cagney <andrew.cagney%gmail.com@localhost>]

On Sun, 16 Jun 2024 at 00:20, T K Spindler (moof) <dogcow%babymeat.com@localhost> wrote:
>
> On Sun, Jun 09, 2024 at 10:57:16AM -0400, Andrew Cagney wrote:
> > Has anyone looked at https://milkv.io/duo-s which is slightly cheaper
> > than anything from beagle.  The chip is SG2000 and seems to be based
> > on the T-Head C906 RISC-V (and that core seems common).
> > For JTAG (which came up at BSDCan), there's this open question
> > https://community.milkv.io/t/jtag-pins-on-duo-s/2161 and the
> > discussion https://community.milkv.io/t/jtag-jtag-debug-guide-for-duo/1138
> > The latter thread points out that for DuoS's sibling Duo256M, uboot
> > re-configures the JTAG serial pins as GPIO (oops) so that needs to be
> > avoided.
>
> The thing about the Duo-S is that Milk-V has not released the source
> or toolchains to rebuild fip.bin - which includes the first stage boot
> loader and u-boot; I'm not sure whether the default pin assignments are
> enough to let 2wire/4wire JTAG work or not, or if u-boot would need to
> be recompiled.

Thanks for the warning.  I've gone ahead and ordered one - too cheap
to resist - so I guess I've set myself a challenge.

> (That not a lot of changes have appeared in
> https://source.denx.de/u-boot/custodians/u-boot-riscv is not exactly
> encouraging - but I don't think there are *that* many differences
> between the SG2002 in the 256M and the SG2000 in the -S, from what
> I've skimmed over in the documentation; that half the stuff in various
> trees refers to the SG2xxx stuff as CV181x makes everything more
> confusing, too.)
>