On 24/01/2025 16:01, gary%duzan.org@localhost wrote:
On 21/01/2025 16:51, Robert Swindells wrote:Martin Husemann <martin%duskware.de@localhost> wrote:On Tue, Jan 21, 2025 at 07:30:03AM +0100, Martin Husemann wrote:FWIW, I see the same on my VF2 - it does not properly probe the SD card even if I loaded u-boot from it:The boot log was from a boot with u-boot from flash, but the same happens if I boot from the SD card (so u-boot is happy talking to it).I see the same errors now on the Star64 with the same SoC. The SD card did work when I first booted the Star64 on 03/12/24.Can you bisect it please? Thanks, NickFWIW, I just got to where I can build from git, and got very similar results with this commit: commit bebcc544b9a8894082f9159392f088a7a6bbae77 (HEAD) Author: skrll <skrll%NetBSD.org@localhost> Date: Tue Nov 12 07:30:12 2024 +0000 Remove the USB stuff that is already in GENERIC.commonAnd for this commit: commit d3f2e7824b7cae18e4901f55eb4081b16f54288a (HEAD) Author: gutteridge <gutteridge%NetBSD.org@localhost> Date: Tue Oct 29 02:54:38 2024 +0000 mkubootimage.c: consistently use endian.h origin functions
Thanks for trying... I did some tests myself and "something weird" is happening - it seems power on vs reboot makes a difference. More testing required.
p.s. I am cross-building from Linux, so there is some fighting with host tools. If I get past that, I'll see about doing a proper bisect.
I've done ALL my development cross building from linux and there shouldn't have been a reason to change tools recently unless I missed something. Nick