Skip to content
Snippets Groups Projects

xiaomi-santoni: Added support for osk-sdl and improvements

Merged Imported Administrator requested to merge xiaomi-santoni into master

I've added support for osk-sdl to the xiaomi-santoni, it works fine, charging-sdl works fine, but full disk encryption doesn't work yet. Here is charging-sdl running: image

Edited by Administrator

Merge request reports

Approved by

Merged by AdministratorAdministrator 5 years ago (May 9, 2019 9:09pm UTC)

Loading

Pipeline #191877 passed

Pipeline passed for eb062251 on master

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Administrator marked as a Work In Progress · Imported

    marked as a Work In Progress

    By Danct12 on 2019-04-28T09:37:12

  • Administrator changed the description · Imported

    changed the description

    By Danct12 on 2019-04-28T09:37:12

  • Author Owner

    Full Disk Encryption barely working on this device. After unlocking, it goes on black screen forever, this is what shown in dmesg:

    [   22.148732] device-mapper: table: 252:1: linear: dm-linear: Device lookup famled
    Y   22.148767]$device-mapper: ioctl: error adding target to table
    [   22.15▒202] device-mapper: table: 252:1▒ linear: dm-linear: Device lookup failed
    [   22.558237] device-mapper: ioctl: error adding target to table

    pmOS is installed in a SD Card.

    By Danct12 on 2019-04-28T17:43:56

    • Author Owner
      Resolved by Administrator

      The patch looks good as is, and it is already an improvement. If you want, we can merge it already, and you can fix it up further from there on.


      Regarding the colors, you would probably need another format than rgba (maybe you can also play with the settings in the kernel and change the mode there?).

      Regarding the dm-crypt error: that is a weird one. I found this random patch, which suggests that maybe your kernel might be reporting a "device lookup failed" error instead of a "device busy" error, which it should display instead. Maybe this is helpful - you could check in your kernel's source if that is the case. And regarding debuggging: try to open it manually: https://wiki.postmarketos.org/wiki/Osk-sdl#Unlocking_via_telnet

      As always, thanks for your great work @danct12 :tada:

      By Oliver Smith on 2019-04-29T12:27:46

      Edited by Ghost User
  • Administrator added 3 commits · Imported

    added 3 commits

    • badaed71 - 1 commit from branch postmarketOS:master
    • 69f399d5 - device-xiaomi-santoni: Added support for osk-sdl
    • f554c858 - linux-xiaomi-santoni: rebased the latest commit from bitrvmpd and patched...

    Compare with previous version

    By Danct12 on 2019-04-30T06:03:48

  • Administrator added 5 commits · Imported

    added 5 commits

    • f554c858...4e9d08d3 - 2 commits from branch postmarketOS:master
    • 531a9f7f - device-xiaomi-santoni: Added support for osk-sdl
    • ba08691e - linux-xiaomi-santoni: rebased the latest commit from bitrvmpd and patched...
    • f86f59b8 - linux-xiaomi-santoni: libudev support

    Compare with previous version

    By Danct12 on 2019-04-30T19:50:30

  • Administrator changed title from WIP: device-xiaomi-santoni: Added support for osk-sdl (with color problems) to WIP: xiaomi-santoni: Added support for osk-sdl and fixed libudev support · Imported

    changed title from WIP: device-xiaomi-santoni: Added support for osk-sdl (with color problems) to WIP: xiaomi-santoni: Added support for osk-sdl and fixed libudev support

    By Danct12 on 2019-04-30T19:51:33

  • Author Owner

    Not sure if this is intended or not, but charging-sdl still kept the backlight turned on after the battery screen disappears. I suppose that the backlight are supposed to turn off, right?

    By Danct12 on 2019-05-01T14:04:12

  • Author Owner

    @danct12 it's not implemented yet, see https://gitlab.com/postmarketOS/charging-sdl/issues/18

    By Daniele Debernardi on 2019-05-01T17:19:48

  • Administrator added 1 commit · Imported

    added 1 commit

    • 3d991cc0 - device-xiaomi-santoni: Added udev rules and scripts for charging led

    Compare with previous version

    By Danct12 on 2019-05-01T17:39:45

  • Administrator changed the description · Imported

    changed the description

    By Danct12 on 2019-05-01T18:02:08

  • Administrator added 1 commit · Imported

    added 1 commit

    • 540c18a8 - device-xiaomi-santoni: Added udev rules and scripts for charging led

    Compare with previous version

    By Danct12 on 2019-05-01T18:21:14

    • Author Owner
      Resolved by Administrator

      I've seen you pushing my patch, awesome. Anyway, I'm currently looking into jack detection:

      • When headphones are plugged, /sys/devices/virtual/switch/h2w/state changes to 1 and 3, or sometimes 4.
      • When nothing is plugged, it changes to 0.

      I'll push a patch to @danct12 when I figured out how the jack detection works.

      By Asriel Dreemurr on 2019-05-01T20:00:35

  • Administrator added 1 commit · Imported

    added 1 commit

    • a56b1121 - device-xiaomi-santoni: Added udev rule for jack detection.

    Compare with previous version

    By Danct12 on 2019-05-02T02:12:52

  • Administrator added 1 commit · Imported

    added 1 commit

    • 496f2789 - device-xiaomi-santoni: Added udev rule for jack detection.

    Compare with previous version

    By Danct12 on 2019-05-02T04:00:31

  • Administrator added 1 commit · Imported

    added 1 commit

    • ec64c2b2 - device-xiaomi-santoni: Added (hacky) udev rule for jack detection.

    Compare with previous version

    By Asriel Dreemurr on 2019-05-02T16:26:12

  • Administrator added 1 commit · Imported

    added 1 commit

    • 195114e4 - xiaomi-santoni: Added pulseaudio files and silenced substream kernel messages

    Compare with previous version

    By Danct12 on 2019-05-02T19:39:17

  • Author Owner

    This should be alright to be merged now, I'll try to think how to get PulseAudio runs as user on startup in the future.

    Changes that I've did with the latest commit:

    • Make all applications to use pulseaudio by default, if pulseaudio doesn't exist (not started) then it'll use the sysdefault fallback.
    • Disabled tsched due to audio being choppy and glitching (had something to do with the CPU or qcom's DSP)
    • Commented kernel spam due to substream runtime not found: [ 3753.419598] msm_pcm_volume_ctl_get substream runtime not found

    By Danct12 on 2019-05-02T19:48:14

    Edited by Administrator
  • Administrator unmarked as a Work In Progress · Imported

    unmarked as a Work In Progress

    By Danct12 on 2019-05-02T19:42:02

  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
Please register or sign in to reply
Loading