Hello thanks for your insights. Yes I can report that my machine is not the full HD version. It is 1377x768 so that makes sense.
I found three driver files under the directory /lib/firmware/mrvl in chrome:
I am now hung up on the seemingly trivial problem of mounting my usb disk with driver files from Openmarket OS to read copy the driver binaries, as it does not auto-mount.
I need to be super user to use "mount" but when I try and login as root using 147147 it says that is an invalid password. Is there additional default password for root on these images?
Once I can access my binaries from USB where do I put them and what do I need to do on the Linux side? Is it just a matter of moving them to the right place? Apologies I am not familiar with these matters. Thanks to you all for giving me hope this will not be another piece of e-Waste!
Just to update, I discovered that in the same driver directory under postmarket OS it already contains all of these binaries I copied from Chrome OS. There are about 23 files under /lib/firmware/mrvl. So perhaps the question should be why doesn't the WIFI launch at startup?
I tried running modprobe but it seems it is not installed. What is the recommended way to probe the physical hardware and find the WIFI module? Any thoughts much appreciated.
I post here the output of dmesg for further insight. If I understand correct it seems the WLAN fails to initialize. I get the message "WLAN is not the winner! Skip FW dnld". Further on near the end it says "FW failed to be active in time" and then "_mwiflex_fw_dpc: unregister device". I guess this mwiflex refers to the Marvell WIFI:
I might add, even when booting in Chrome OS the operation of the WIFI seems somewhat hit and miss. Sometimes it works sometimes it doesn't. Seems about half of the time.
I wonder also would there be a plugin USB WIFI as an alternative option?
So perhaps there might be a difference in the wifi module of the unit shipped in Japan. I will take a look inside at the module to see if it is a Marvell.
On a related note I wonder is the command LSPCI appropriate to list connected PCI devices? I can get LSUSB working but not LSPCI but I am not clear this is appropriate with this architecture.
Can anyone advise is there a generic USB wifi I could plug in that could be got working with postmarket OS?
On a related note I wonder is the command LSPCI appropriate to list connected PCI devices? I can get LSUSB working but not LSPCI but I am not clear this is appropriate with this architecture.
lspci is part of the pciutils package, so this would need to be installed for testing
I think I will have to study up a bit then. As there is no C++ compiler already on the image install then it looks like I will have to cross compile the pcituils on another laptop to progress this investigation. Unless there is some means of installing a binary package for this application from Arch for Arm linux (*.pkg package). I guess as postmarket OS is primarily for phones then the model is to cross compile so I will have to work out how to compile and install the required tools. Nevertheless I commend this project and efforts to support all these devices to keep them from landfill!
If anyone could confirm if their WIFI already works (or not) on this machine then that would be great as I could know if I have a hardware fault in that case.
Wifi works fine on Nyan-Big, but recently there are a bunch of issues popping up, current build doesn't complete as initramfs is too big, this casued my Wifi to break. Please check all the current chromebook issues being reported.
#2914 (closed) and #3816.
I have an update. I tried all three releases for this machine with XFCE (Not FHD) and they all have the issue of the wireless LAN not working. Although I can see the module loaded for this chipset when running LSMOD it does not appear under when running ifconfig. As before de-registration of this device is mentioned when I run dmesg.
However I tried today MATE release on 13/9/2024 and can report the WIFI is fine. So that's pretty fantastic news for me with this machine. Kudos to this projects aims to support all these hardware variants.
I installed Firefox ESR to test (sudo apk add firefox-esr) and everything appears to work flawlessly so far. Videos, Audio, Audio and brightness hot keys etc. Audio switches to earphones fine. Battery monitor great. This makes a very impressive and nice lightweight machine!
Just to add regarding booting, as these machines do not have the Mr. Chromebox script to set up the GBB flags:
To set the short boot time and remove the beep, this is the 1st bit called FLAG_DEV_SCREEN_SHORT_DELAY.
To force staying in developer mode and remove hitting space causing it to exit, this is the 4th bit called FORCE_DEV_SWITCH_ON.
So to set bit 0 and bit 4 simultaneously we need to set binary 1001, which corresponds to setting a decimal 9:
0x9 shorten boot time and enforce developer mode
Decimal 9 = Binary 1001
Go to this directory as these are not in the path:
/usr/share/vboot/bin
Type shell to get crosh shell
sudo get_gbb_flags.sh -d -e - Interrogate existing flags
sudo set_gbb_flags - Shows what can be set and which bit is which
sudo set_gbb_flags.sh 0x09 - Sets 2 second boot screen, removes beep and forces developer mode
(turns off the danger that hitting space will cause to remove developer mode)
Further observations when booting the MATE non FHD distributions of 25/9/2024 and 13/9/2024:
With both of these I find that the WIFI works fine on first boot but vanishes on 2nd boot! It is no longer then visible under ifconfig. I did watch a movie though so that feels like huge progress!
Upon reading these much older notes on this ACER CB5-311 machine:
It is mentioned that: "Additionally, the Marvel Chipset used for Wifi in this chromebook comes with two additional devices, an accesspoint (AP) and a Peer2Peer device (P2P) which troubles NetworkManager. Therefore, I've also included a patch that deactivates these two. "
Not being an expert in these matters, I wonder if anyone can inform what I have to edit to load some drivers and not others into the kernel at boot time? Perhaps the above comment is somehow linked to this issue?