Hello Hans,
I have been wrestling with the u-boot for BananaPi R1 (Lamobo R1) for some time . As you have mentioned and I have also seen , many people are working on this board , but not too many have attempted to submit patches up stream . I am attaching to this email a patch , which combines work of many people such as Igor P - armbian , Zoltan H - openwrt , and others who have scattered their scripts over the internet . I believe that within the patch we have the important information captured for the Lamobo R1 board : SATA power , IR , GMAC . As you can see the dts has evolved from the original bananapi which it uses for basis , and additional bits were put on . I do not know the original revision of the bananapi dts , but I can see that it differs a lot form the current one in terms of structure . I wanted to attempt to re-write it to make it closer to the current one , but was sure that I would brake it , if I did . I leave it to your skill . Since I have a couple of these boards , I am happy to test any mods for you .
Best Regards Milorad
Hi Milorad,
On 08/23/2015 08:40 AM, mo.ucina wrote:
Hello Hans,
I have been wrestling with the u-boot for BananaPi R1 (Lamobo R1) for some time . As you have mentioned and I have also seen , many people are working on this board , but not too many have attempted to submit patches up stream . I am attaching to this email a patch , which combines work of many people such as Igor P - armbian , Zoltan H - openwrt , and others who have scattered their scripts over the internet . I believe that within the patch we have the important information captured for the Lamobo R1 board : SATA power , IR , GMAC . As you can see the dts has evolved from the original bananapi which it uses for basis , and additional bits were put on . I do not know the original revision of the bananapi dts , but I can see that it differs a lot form the current one in terms of structure . I wanted to attempt to re-write it to make it closer to the current one , but was sure that I would brake it , if I did . I leave it to your skill . Since I have a couple of these boards , I am happy to test any mods for you .
We need to have someone officially submitting a patch upstream, to do so you need to create a git commit in a git checkout of the kernel and then apply the changes and do "git commit -as" and then "git send-email HEAD~", this assumes a properly configured git, see: https://revspace.nl/KernelDriverProgrammingCourse2015/Preparations
And then "8) Configure git so that you can send out patches, edit ~/.gitconfig and add the following lines:"
Also for a u-boot patch the dts file needs at a minimum to be submitted to the upstream kernel (preferably it needs to be accepted, but at a minimum it needs to be submitted) through the official submission process.
About the dts, the dts is using direct paths to various dts nodes, now a days the norm is to address existing nodes by reference, see e.g.:
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/arch/arm...
Which is addressing e.g. the ehci0 node as &ehci0, rather then creating the entire path to the ehci0 node.
Also I wonder if you should define 3 uarts in the dts, are the pins of all routed to gpio headers ?
Regards,
Hans
Hello Hans,
Appreciate your help . However I am not getting that far . After I put in the patches to define the lamobo specific parts , and compile the new u-boot . I can not boot on it . My newly created lamobo u-boot-sunxi-with-spl.bin bricks the device . And the only way to rescue it is to install the bananapi version back in with dd comand . So not very close to submitting any code at all . Funny enough all code compiles very nice with no errors . So no idea where I am going wrong . I had a chat with Igor P about his patches , and he is in the process of re-writing the armbian uboot to the 'new format' 2015-07 . But it is not a priority at the moment and is a fair way off.
Best Regards
Milorad
On 24/08/15 03:06, Hans de Goede wrote:
Hi Milorad,
On 08/23/2015 08:40 AM, mo.ucina wrote:
Hello Hans,
I have been wrestling with the u-boot for BananaPi R1 (Lamobo R1) for some time . As you have mentioned and I have also seen , many people are working on this board , but not too many have attempted to submit patches up stream . I am attaching to this email a patch , which combines work of many people such as Igor P - armbian , Zoltan H - openwrt , and others who have scattered their scripts over the internet . I believe that within the patch we have the important information captured for the Lamobo R1 board : SATA power , IR , GMAC . As you can see the dts has evolved from the original bananapi which it uses for basis , and additional bits were put on . I do not know the original revision of the bananapi dts , but I can see that it differs a lot form the current one in terms of structure . I wanted to attempt to re-write it to make it closer to the current one , but was sure that I would brake it , if I did . I leave it to your skill . Since I have a couple of these boards , I am happy to test any mods for you .
We need to have someone officially submitting a patch upstream, to do so you need to create a git commit in a git checkout of the kernel and then apply the changes and do "git commit -as" and then "git send-email HEAD~", this assumes a properly configured git, see: https://revspace.nl/KernelDriverProgrammingCourse2015/Preparations
And then "8) Configure git so that you can send out patches, edit ~/.gitconfig and add the following lines:"
Also for a u-boot patch the dts file needs at a minimum to be submitted to the upstream kernel (preferably it needs to be accepted, but at a minimum it needs to be submitted) through the official submission process.
About the dts, the dts is using direct paths to various dts nodes, now a days the norm is to address existing nodes by reference, see e.g.:
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/arch/arm...
Which is addressing e.g. the ehci0 node as &ehci0, rather then creating the entire path to the ehci0 node.
Also I wonder if you should define 3 uarts in the dts, are the pins of all routed to gpio headers ?
Regards,
Hans
-
Hans de Goede -
mo.ucina