Dell 7486 nami/vayne - Fedora, no touchscreen. Will this trick fix it?

Support Linux Support
1

Thanks to the people behind these chrultrabook/mrchromebox projects to give our chromebooks an extended life.

I am multi-booting RW_LEGACY on hard disk with ChromeOS and Fedora. Got audio working thanks to WeirdThreeThing.

My Dell has emmc onboard. The emmc stats say that wearoff after 5 years is normal [10-20%]. However loading the kernel/initrd takes almost minutes.

The grub console size is also of the WindowsXP era - I could stretch it max to 1024x768. At least it shows the multi-boot options that I had installed [fedora and chromeOS brunch].

Just to compare what works and what not.
On fedora touchscreen and keyboard mappings do not work yet.

Try to find out why not. In ChromeOS elants_i2s [ELAN touchscreen] is reported by dmesg working well. Under fedora:

❯ sudo dmesg | grep elants
[ 7.434959] elants_i2c i2c-ELAN0001:00: supply vcc33 not found, using dummy regulator
[ 7.435035] elants_i2c i2c-ELAN0001:00: supply vccio not found, using dummy regulator
[ 8.504650] elants_i2c i2c-ELAN0001:00: nothing at this address

lsmod does not show specific kernel modules loaded for touchscreen:

❯ lsmod | grep i2s

❯ lsmod | grep hid
i2c_hid_acpi 12288 0
i2c_hid 45056 1 i2c_hid_acpi

And kernel modules related to touchscreen found in fedora:

❯ ls -lR /lib/modules/6.8.11-300.fc40.x86_64/kernel | grep touchscreen
drwxr-xr-x. 2 root root 4096 Jun 5 17:55 touchscreen
/lib/modules/6.8.11-300.fc40.x86_64/kernel/drivers/input/touchscreen:
-rw-r–r–. 1 root root 14288 May 26 07:00 usbtouchscreen.ko.xz

❯ ls -lR /lib/modules/6.8.11-300.fc40.x86_64/kernel | grep elants
-rw-r–r–. 1 root root 15008 May 26 07:00 elants_i2c.ko.xz

Note that on chromeOS the kernel module elants_i2c.ko was not found. I assumed it is included in the kernel, as in lsmod it is also not reported as a specific loaded module.

Then I found this 2020 bugzilla bug with a fix, that work on another laptop.

The trick is to rebuild the kernel and have the elants_i2s run as a loaded module. In this way there is no race condition with other stuff like vcc33/vccio power supply not ready yet.

Any links to a kernel build tutorial is appreciated!

2

Have you tried running all of these scripts. Also make sure you have the correct layout.

I don’t think I know the rest.

3

and there’s the problem. NAMI has 4 touchscreen models as options, and with the stock firmware, Google expects ChromeOS to probe and see which is actually present. Mainline linux rejected this ACPI “probed” extension, and so things go to hell with the kernel trying to init 4 touchscreens at once.

the UEFI Full ROM firmware does not have this problem since coreboot will power up and detect the correct touchscreen and only inform the kernel of the one actually present.

4

Yeah. The keyboard mappings work great. Thanks!

5

i’m impressed by your detailed knowledge. Thanks for the quick reply.

I’ll stick for the time being to RW_LEGACY, cause from time to time still need ChromeOS.

What I notice is that chromebooks have a good battery standby than my windows laptops. And on Linux this device is after bootup much more quiet than in ChromeOS even with Playstore and Linux VM disabled.

6

Try waydroid or waydroid-linux (beta) to see if your required android apps work in it and then you can dump ChromeOS entirely.

7

I wrote the code for runtime detection of i2c devices like touchpads/touchscreens in coreboot, so I’m pretty darn familiar with how it works and what devices make use of it :wink:

8

You like spi or i2c

9

nah, android on 4GB is asking for trouble on chromebooks. I just wanna check how google is updating older machines. Before AUE expires I’ll move to coreboot.

10

Memory is fine. Works great on my 4GB Asus.

doug@C434:~$ sudo pmap 5801
[sudo] password for doug:
5801: /usr/bin/python3 /usr/bin/waydroid show-full-ui
total 640324K

11

Only if you don’t know what you’re doing, or you do something badly