04:34 <ndufresne> eh, just tried to rebuild mainline u-boot following mainline readme, all steps worked, but it won't boot, traces aml log : SIG CHK : 191 for address 0x01700000 and then reset

04:59 <lvrp16> ndufresne: here's my script for it - https://pastebin.com/ZzMMgJ9g

05:00 <lvrp16> you need to rip src-amlogic-openlinux from one of amlogic's openlinux

05:01 <lvrp16> are you on 2019.01 or 2019.04rc1?

05:05 <ndufresne> I was trying master, mostly for curiosity

05:06 <ndufresne> I've tested now with libretech latest ubuntu, and that one works, so clearly my build (or some step on my side) is incorrect

05:07 <lvrp16> i'm zipping up some stuff

05:08 <ndufresne> it's quite similar to the steps I followed (which I also wrapped into a script)

05:08 <ndufresne> Except I'm guessing you didn't have to build the vendor u-boot

05:09 <lvrp16> vendor u-boot: http://share.loverpi.com/board/libre-computer-project/libre-computer-board-aml-s905x-cc/u-boot/src-amlogic-openlinux.tar.gz

05:10 <lvrp16> my upstream u-boot has too many changes to share

05:10 <lvrp16> i know 2019.04rc1 has a load of bugs for sunxi, i have not tested meson

05:10 <lvrp16> to the point where i can't even compile

05:11 <ndufresne> noted, but I'll give a second try tomorrow

05:12 <ndufresne> at worst I found that the u-boot ripped from the librethech ubuntu image is enough for what I initially needed (just needed a more recent pxe command)

05:13 <ndufresne> thanks btw

05:13 <lvrp16> haha cool

05:15 <lvrp16> my own upstream branch is in tatters right now because i'm trying to add a direct boot mechanism since grub is incredibly slow.

05:15 <lvrp16> while simultaneously supporting normal boot_efi

05:16 <ndufresne> don't it normally look for a boot.scr first ? before doing efi ?

05:16 <lvrp16> and yes the USB/eMMC/SD scanning is incredibly slow also trying to have something to short circuit that

05:16 <lvrp16> my branch has a lot of stuff not in the images yet

05:16 <ndufresne> great, means cools stuff will come ;-P

05:16 <lvrp16> most of it is in the h3/h5 images but not the s905x since

05:17 <lvrp16> you have bugs like these: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=922101

05:18 <lvrp16> i didn't realize how slow grub was until i used it on these arm boards...wow

05:18 <ndufresne> that's the next board on my todo, the H3, it's currently in my krisman hands, who's suppose to try and fix some issues for a friend, and then my goal is to get the codecs running, so I can do some v4l2 tracing to compare against some work my other colleague is doing on the RK3399

05:18 <lvrp16> 10s boot time haha

05:19 <lvrp16> 10s for u-boot, 10s for grub, 10 seconds to boot after grub selection. 30 seconds wasted doing nothing

05:19 <ndufresne> wow, that's seems like slower then usual

05:19 <ndufresne> but if it tries all emmc, sd, usb etc.

05:19 <lvrp16> yeah, with short circuited u-boot to kernel, it takes like 5 seconds to boot an entire headless image....

05:20 <lvrp16> so grub needs a lot of work before it's usable in reality

05:20 <ndufresne> I'll follow up, I'm using pxe boot, but still, whatever can make it boot faster

05:20 <ndufresne> in the default env, pxe is quite far in fact

05:21 <ndufresne> ok, see you around

05:21 <lvrp16> yep, just take care because i think that u-boot bin looks for pxe at u-boot.scr instead of boot.scr

05:21 <lvrp16> night

06:57 <wens> you could try rewriting / patching the default distro_cmd?