-
Notifications
You must be signed in to change notification settings - Fork 13
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Convert the project from using user-patches back to a tranditional patch #24
Comments
Oh yeah, I was thinking the same thing. I would happy to do this in right way. |
Yeah.. We can still do it a "more right" way without to much trouble I believe (without going as far as being "conforming" but "in the spirit of". I believe and was going to try to use this utility (VS Code extension Git Patch, as I believe it will allow something along the lines of
I already prototyped this once just moving the patches (ones I had from previous effort I was working independently) into the normal patch dir and almost had it working but there were some issues in how I did it (general mistake, unrelated). Then go forward, just diff/patch like normal. |
Well, perhaps we have different definition of "more right" way:) Let's discuss this a bit.
Another one task to adapt all these point to the current project structure. I try to rebase my changes over fresh armbian/main and looks like patch locations were significantly changed. And I am against the idea of keeping all changes in single file. It's better to have a few "topic related" files rather than one single blob. BTW are you aware about |
1/2 were where I stopped on my solo attempts. It was getting the timing of the patch insertion when coming from latest. What I had working at one point before deciding to combine my efforts here was/were
Coming from Renegade actually makes the most sense since that is what the HW probes think it is, and that this is really a "subvariant" No problem about the single file, I probably should have said single "cleanup commit" irrespective of number of files. Definitely the kernel-patch command is handy :) |
Actually, we may want to use a newer feature from this armbian#5124 It adds (as one of the newer *-patch commands) uboot-config, This outputs a defconfig at the end, instead of a patch. Since that's in uboot, we would then be able to take that defconfig (having the current userpatches in it), and use that defconfig for future builds instead of the normal instead of the current Renegade ( roc-cc-rk3328_defconfig ) one. |
But we have no issues with u-boot, patches already in right place. Do you mean |
Actually (after relooking last night and starting an out of band from armbian/build tag 23.11.2.) you are right, the defconfig file can remain untouched. Let me see where I get from this out of band attempt. |
Ok, doing the above, I was able with exceptionally minimal changes, take 2 files from your build, put them into the appropriate place in /patch dir, on 6.1 upstream armbian/build tag 23.11.2. I need to debug why 6.6 has a build error but it is obviously from a Realteck wifi patch not going properly which may be as easy as disabling the xtra wifi drivers. I took things as far as communicating over CANBUS via USB patch between MCU and "pi" sides, with, klipper in usb/can bridge mode for the MCU's bin My testing steps were to
6.1 works, 6.6 doesn't build yet (wifi patch breaks things) This obviously is still preliminary testing, but you will note, this is only using one partition (one of the major changes) which is why the removal of the FS line is needed in the conf. More info as I get to test more. By doing this, the only patches I don't have from your work appear to be the a display you added, and the display modes added. Both would be easier to just reapply with non-duplicate indexes. |
Would you mind to open PR to show these changes? And have you consider to use latest armbian master or mater from this repo? Asking because Armbian recently changed patch dir structure. |
The only reason I haven't proposed a PR yet is because the PR should be (techinically) be against armbian/build, since that's all I'm modifying, just adding 2-3 files from this repository. Doing this to prove how close they are, should this repository rebase back to armbian/build, add the few files to bring functionality back to where this repository is, right this minute, then maintain from there forward. That's what I was trying to get across by "quickly" doing. Essentially putting it where it would be IF it was in upstream. I hope that made some sense. Been working too many hours this week :) Also, for clarity I STARTED from latest github.com/armbian/build branch main (not master, that is archive) tag 23.11.2. I then made the couple of changes I mention above. Also, Disabling EXTRAWIFI did exactly as I expected and build of 6.6 is continuing as we speak :) |
Ok, 6.6 compiled with no issue, I am able to talk to both sides of the board via USB->CANBUS as canbus device and printer shuts down as expected since I haven't wired my printer's heater and temp sensors yet since I have it setup with a Stealthburner head and am going to swap the head and x axis to CF before Looks like it is time to wire my printer and take testing to the next level. Once i've proved it all talks to "stock" ish sensors, I'll hook up my EBB42, get CAN connection working there, and swap heads. |
For me it's pretty difficult to catch up what have you done. this is why I asked about PR.
Please pay attention to:
And looks like original Armbian repo restructured kernel patches recently. Seems like Renegade related stuff was moved to archive :) |
What would make the most sense
Normal operation of board works (i.e. run klipper, connect via CANBUS to the "printer" side)
|
Yes, I periodically sync with this repo. my So if you would nee "clean state" I would branch ether from my
See text bellow.
Sure, after proper testing
OK, but keep in mind the original definition has 2 Ethernet channels. You need disable the first one and enable the second (gmac2phy if I remember correctly)
Totally wrong. Mainline code do not have support for st7796. They have something similar, but that driver cannot rotate image, which sometimes is needed. tsc2046 touchscreen is full surprises too. IRQ definition should be adjusted for kernel after 6.4. Please see changes for
Yeah, for some reasons proper HDMI communication gone from 6.x kernels (5.x was able to read EDIT properly, 6.x not). SO we need this headache to keep using yet another great MKS device :)
What do you mean? DO you have non standard pins for UART?
Be careful, I removed couple of patches in scope of |
Which feature would you like to have?
Goal
Funding
The text was updated successfully, but these errors were encountered: