tokio-techbook: new device

A 7" VIA/Wondermedia 8850 laptop. Got it working with mainline 4.3.0 without patches, but could not get it working with the postmarketos-stable package for some reason.

Depends on !578 (merged) as it needs a custom loading address for its initfs file.

added 1 commit

• 449db0dd - tokio-techbook: Update screen size and "Kernel config based on"

Compare with previous version

By Magnus Gulin on 2019-08-25T20:58:54

added category::device + 1 deleted label

added 1 commit

• 41ffd5a7 - set pkgrel to 0, remove duplicate "deviceinfo_dtb"

Compare with previous version

By Magnus Gulin on 2019-08-27T18:31:59

Update: linux-postmarketos-stable has earlier frozen at bootloader stage, but now I managed to get it booting into a kernel panic. I removed support for CONFIG_ARM_THUMB, CONFIG_ARM_THUMBEE and CONFIG_DEBUG_LL. The last option was probably the problem, as it specifically says "Note that selecting this option will limit the kernel to a single UART definition, as specified below. Attempting to boot the kernel image on a different platform will not work, so this option should not be enabled for kernels that are intended to be portable."

I'll play around with the config settings some more to check which ones are needed, and make a new commit if I get it booting.

By Magnus Gulin on 2019-08-27T20:39:07

resolved all threads

By Oliver Smith on 2019-08-28T21:23:59

Apparently, this may also work on other netbooks that has the same layout/case, since all of them (including this one) is in fact a rebranded system.

By Danct12 on 2019-08-31T10:17:07

@danct12: feel free to suggest a more generic name. We have a generic tablet-x64uefi device, for example.

By Oliver Smith on 2019-08-31T19:30:32

@danct12 Yep, I mentioned some of this on the device wiki page. Still, I figured that it's best to not put a too generic name until I got it verified that this device doesn't have any specific quirks.

One problem is that these clones comes with different ram sizes and the size should be included in the u-boot script. Seems unnecessary to have a separate package for each ram size. Maybe pmbootstrap init could ask for ram size and adjust u-boot script accordingly?

By Magnus Gulin on 2019-08-31T22:33:01

Still, I figured that it's best to not put a too generic name until I got it verified that this device doesn't have any specific quirks.

This is reasonable.

One problem is that these clones comes with different ram sizes and the size should be included in the u-boot script. Seems unnecessary to have a separate package for each ram size. Maybe pmbootstrap init could ask for ram size and adjust u-boot script accordingly?

Sounds like a good idea, but I would wait with implementing this until somebody with a similar netbook shows up, so we can test this feature.

By Oliver Smith on 2019-09-03T20:45:49

@magnusgulin any updates?

By Bart Ribbers on 2019-10-08T11:39:30

@PureTryOut The linux-postmarketos-stable is booting with display output, but no usb/keyboard/touchpad is working. By comparing the dmesg log from a working kernel I think this is possibly connected to PCI drivers, although I've tried to include all related drivers...

Also I haven't been able to run it with an initrd. In the u-boot bootloader I can choose an address to load the initrd into. Depending on which address (and kernel commandline) I choose I either get

1. device hang in bootloader
2. this message in the dmesg log ([ 0.000000] INITRD: 0x00500000+0x00600000 overlaps in-use memory region - disabling initrd)
3. this kernel panic on boot

No filesystem could mount root, tried:
ext3
ext4
ext2
cramfs
squashfs

Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1,0)

The command lines I've tried with is variations like this, with initrd address changing to same address the initrd is loaded into

setenv bootargs mem=${memtotal} root=/dev/mmcblk0p2 rootwait init=/sbin/init initrd=0x01600000,8M console=tty1 logo.nologo PMOS_NO_OUTPUT_REDIRECT
setenv bootargs mem=${memtotal} root=/dev/ram0 rw initrd=0x04000000,6M console=tty1 logo.nologo PMOS_NO_OUTPUT_REDIRECT

I've played around a lot with both kernel configs and command lines, but maybe it would be best to just merge it with the mainline kernel I already have running for now? I can always make a new merge request if I get the linux-postmarketos-stable working.

By Magnus Gulin on 2019-10-08T19:51:00

I've given up on getting mainline to work on this device for now. @ollieparanoid @PureTryOut Would it be ok to merge in this state? It's a working port, and the mainlining process can be continued at a later state if someone is interested.

By Magnus Gulin on 2020-03-19T22:33:13

The source branch is 1011 commits behind the target branch

Cool, but now this needs a rebase on master (devices were moved to subfolder testing/), fixing numerous linter warnings that will appear, using new downstreamkernel_package() in linux- package() function, ... and testing that it works.

Then resolve WIP status and then it probably "it will be ok to merge"...

By Alexey Min on 2020-03-19T23:51:28

OK, I will fix the mentioned issues ASAP

By Magnus Gulin on 2020-03-20T00:03:41

added 1013 commits

• 41ffd5a7...43364cd0 - 1011 commits from branch postmarketOS:master
• 40e5d869 - Merge remote-tracking branch 'upstream/master' into device/tokio-techbook
• d0d2d6cf - tokio-techbook: move devices packages to /testing folder

Compare with previous version

By Magnus Gulin on 2020-03-20T16:56:11

added 2 commits

• 8028b738 - tokio-techbook: fix linting errors
• 76f089ef - tokio-techbook: enable CONFIG_CGROUPS

Compare with previous version

By Magnus Gulin on 2020-03-20T17:42:49