(04-06-2022, 03:22 PM)wdt Wrote: likely the problem is video display is wrong
The power LED never turns green, and there is no further output on the serial console. So, would that be caused by a wrong video setting?
The pwr led is controlled by uboot, different uboots have different settings
If there is something wrong with uboot you won't get the various xxxx bytes read statements
After kernel, initrd and dtb is loaded, uboot is done (assuming correct load addresses)
For further serial comms, you have to change console= statements in boot.txt and recompile(->boot.scr)
Maybe the video mis-config are in the dtb, or could be in kernel config
Well, we're still talking about an image that works when you boot it from sd but not when you boot it from emmc.
04-08-2022, 05:06 PM
(This post was last modified: 04-08-2022, 05:08 PM by wdt.)
What uboot is on the emmc? From the boot rom, search order for idbloader
is SPI, emmc, uSD.... IIUR uboot is assumed to be on same media as idbloader
OR
will the sd boot succeed if emmc is disabled/removed?
(04-08-2022, 05:06 PM)wdt Wrote: What uboot is on the emmc? From the boot rom, search order for idbloader
is SPI, emmc, uSD.... IIUR uboot is assumed to be on same media as idbloader
OR
will the sd boot succeed if emmc is disabled/removed?
I dunno what to tell you about the uboot version other than what is in the text i copied from my terminal session.
When i disabled the emmc with the switch on the board, it still didn't boot, but i did have the console cable connected still.
I have since used the simple usd-to-emmc adapter to write a manjaro image to the emmc and it is booting into that normally. But i did note that if the console cable was connected i did not get a video terminal with manjaro. But it did sit there and tell me it was failing to launch xorg.
04-09-2022, 11:50 AM
(This post was last modified: 04-09-2022, 11:53 AM by wdt.)
That, for sure, is a deficiency in uboot, not to tell version -clearly- somewhere
That it WILL boot with emmc enabled and NOT boot with it disabled strongly
suggest that the uboots on emmc and uSD differ
(but not for sure, it could be the OS doesn't drive the display properly, ie booted but no video)
Your last para is not so clear, maybe manjaro boots and can run x11? (but no serial uboot comms)??
But with uSD (armbian) no x11? (maybe try a different dm - display manager. gdm, xdm, sddm, lightdm etc)
(04-09-2022, 11:50 AM)wdt Wrote: That, for sure, is a deficiency in uboot, not to tell version -clearly- somewhere
That it WILL boot with emmc enabled and NOT boot with it disabled strongly
suggest that the uboots on emmc and uSD differ
(but not for sure, it could be the OS doesn't drive the display properly, ie booted but no video)
Your last para is not so clear, maybe manjaro boots and can run x11? (but no serial uboot comms)??
But with uSD (armbian) no x11? (maybe try a different dm - display manager. gdm, xdm, sddm, lightdm etc)
With the current manjaro i get either a text console over serial or x11 on the display.
My PBP was purchased almost exactly 2 years ago and i vaguely recall installing a uboot update to enable some better sleep mode or something.
I don't know if you achieved 'good' sleep -- s3-mem which is ~8%/day?
While tow-boot will get a somewhat reasonable sleep, my tests showed (IIRR) 11-12%/d
BSP uboot has 3 files, all *.img (idbloader, uboot, trust)
Mainline (also tow-boot) has 2 files, idbloader and u-boot.itb
Most people don't seem to care much about sleep, nearly all mainline can't do s3-mem (called suspend)
So the sleep that results is very poor, 10x worse or more
A fresh install of manjaro will have mainline, there is a BSP in repo
Unlike mrfixit, the install is manual, ie you have to do the dd commands
(I am referring to the mrfixit update)
A specific kernel is also required (for s3-mem), but no-one seems to care much to fix this mess
Anyway, if you examine the console statements in manjaro, you can see
what is right for serial comms, to fix armbian to at least have serial and maybe try to trouble shoot video
(04-10-2022, 11:56 AM)wdt Wrote: I don't know if you achieved 'good' sleep -- s3-mem which is ~8%/day?
While tow-boot will get a somewhat reasonable sleep, my tests showed (IIRR) 11-12%/d
BSP uboot has 3 files, all *.img (idbloader, uboot, trust)
Mainline (also tow-boot) has 2 files, idbloader and u-boot.itb
Most people don't seem to care much about sleep, nearly all mainline can't do s3-mem (called suspend)
So the sleep that results is very poor, 10x worse or more
A fresh install of manjaro will have mainline, there is a BSP in repo
Unlike mrfixit, the install is manual, ie you have to do the dd commands
(I am referring to the mrfixit update)
A specific kernel is also required (for s3-mem), but no-one seems to care much to fix this mess
Anyway, if you examine the console statements in manjaro, you can see
what is right for serial comms, to fix armbian to at least have serial and maybe try to trouble shoot video
I dunno about most people but one of the key things i like about laptops is that they are lightweight, compact computers with a screen and a keyboard on a hinge, and that i can configure them so that when i decide it's time to stop using them i can just fold them shut and shove them in a bag and go.
You'd think it wouldn't be much of a hassle to save my work and do a clean shut-down but it turns out that I've had the PBP for almost 2 years now and even for the task of "something over by the 3d printers to access octoprint" it turns out I'd rather use my 2010 era asus 1215n. Because i can fold it up and shove it somewhere and when i open it back up it will still work, and i can't say that about the PBP, even if it stays plugged in.
If the magnet is mis-placed, lid close will not work for sleep
It will sleep when 1 inch from closed, then wake when fully closed
I use ONLY pwr button, disable lid for sleep (all this in systemsettings)
|
