I was reading (and it’s reproducible) a web page with web, containing nearly only textual information: https://www.nanolounge.de/28875/berechnen-geschwindigkeit-satelliten-kreisformigen-umlaufbahn
and noticed that the device gets warmer and warmer. top shows a proc epiphany with 50% CPU usage and a lot of WebKitWebProces processes, perhaps one for any tab in the browser web.
The 2nd time I watched also the device with the app Usage which shows this picture below. Why is the thermal bq25890-charger so high at 85°C. The device was not connected to the charger. And the temperature remains like this even after power-off reboot.
UPDATE: It went down to 37°C when I plugged in the charger, i.e. the value of 85°C is not reset correctly on power-cycle.
I do every update PureOS Store is offering, right now I did so, and I only have 5.15.0-1 and today morning when I faced the problem only 5.13.x and updated yesterday or the day before, IIRC.
OK, thanks. This shows now version 5.15.9pureos1. But, when I unplug the charger cable the temp. jumps up to 85°C and goes down to some 37°C when I plug it in again.
And, coming back to the original problem, why the device runs so hot when rendering a web page?
That’s interesting then, AFAIK this shouldn’t happen anymore. I’ll take a closer look.
No idea, but this particular website also hogs the CPU and makes fans go loud in Epiphany on my laptop, so seems like either the website is buggy, or it triggers some webkit bug that causes this behavior.
Yes, 85°C maps to the reading of exactly zero in the charger’s temperature register. Recently there were some fixes related to how the temperature is getting read by the driver (both by us and by other kernel devs upstream), and as far as I’m aware incorrect temperature readings should be fixed now in 5.15, but maybe something is still missing somewhere.
Yes, there must be something broken. If you look at the graph in the screenshot attached above the temp goes down to the level of the others and jumps up again to 85°C
2 days before battery working well but after when I trying to charge phone I see slow blinking RED, I stop charging and power off phone.
After hour I try to power on phone and it show me that battery not found!
I restart phone and it start fast blinking RED. Press power button and phone going to bootloop 3 times and wont power on.
At next day I charge phone without power on. And now phone always show 85* at bq25890-charger.
Last update nothing change.
…which with up-to-date u-boot can usually be replaced with a much simpler one:
Unplug USB.
Hold the power button for about 20 seconds (or until the LED shines up)
Plug in the provided USB-C PD charger (or another PD charger that’s known to negotiate enough amperage at 5V when used with Librem 5 - the one included in the box does 3A)
That complicated procedure can be useful when the battery is completely depleted, but personally I didn’t have to use it for many months now.
Phone working (charging working too), but bq25890-charger without USB charge cable always show 85*C.
I have a L5 from latest delivery batch (end of 2021) not USA.
Now I want to check does bq25890-charger still working with old kernel bu I cant find a correct way to try boot prevision kernel…
I correct understand:
What I have installed:
ii linux-image-5.13.0-1-librem5 5.13.19pureos5
ii linux-image-5.15.0-1-librem5 5.15.9pureos1
ii linux-image-librem5 5.15.9pureos1
I just need change symlink /boot/stb to dtbs/5.13.0-1-librem5/freescale/imx8mq-librem5-r4.dtb? I see L5 still dont have a tested bootloader menu?
Have I any chance to fix /boot partition if I make mistake or something going wrong? (create backup or maybe boot from SD)
Can anyone give me a link where to find instructions or learn how to L5 booting?
No boot from SD. (I believe it is doable via a chain boot but if you break the content of the eMMC drive then you may not be able to boot from SD either.)
Jumpdrive is what you want. With Jumpdrive you can create a backup (image the entire eMMC drive) so, worst case, if you break it badly then you can restore the image of the entire eMMC drive. If you break it just a little bit then you can telnet in while Jumpdrive is booted and potentially fix whatever the problem is.
(Yes, if you are careful and know what you are doing, you can probably image just the /boot partition. That will be much faster to do the backup and, if required, to do the restore. However I always think that having a full backup is good for any computer.)
I am pretty sure that this is just a bug. Any time you see such a reading, you should assume that it is bogus.
If you are running byzantium then a Purism developer says above that this was literally being fixed as we speak. So maybe an update slipped in to your phone over the last few days.
No. Problem appears after all updates and was no longer any updates. I switch off autoupdate and control it myself.
Thanks for advice. But I dont charge battery out of phone. Excuse me, I expressed wrong, I mean that phone charged in poweroff mode to the status when RED light is off (100%) and bq25890-charger sensor now magical fixed (I dont install any updates).
I started this thread because, as the title of the thread says, the device gets hot by having some page in the browser and, perhaps due to the temp., the red LED starts flashing, reproducible.
Why this thread was high-jacked with other questions, problems and hints.