You are not logged in.
I gave up on the i915, it didn't work a single time (~6 attempts) for me yesterday on 4.7.10. For me, the most important feature of the i915 driver is actually the possibility to change the backlight strength.
So I did some hacking and managed to get backlight settings working without the i915 driver.
Here those patches that I use:
- Asus T100 meta keys: Pretty much the same as the original, except key toggle touchpad is properly mapped to KEY_TOUCHPAD_TOGGLE instead of F24
- Asus T100 touchpad multitouch: Change the device id to work with T100HA, plus the whole touchpad can register left button presses, not just left lower part, except lower left that is still the right button.
- backlight: Brutish patch that adds backlight settings support (/sys/class/backlight/crystal_cove_pwm/) without i915
www.lithops.se/Asus-T100HA-patches.tar.bz2
And for those of you that are interested, here is my .config and DSDT.hex with the fix for sdcard and wlan.
www.lithops.se/Asus-T100HA-config.tar.bz2
On the kernel commandline I use:
video=LVDS-1:d fbcon=rotate:3
All patches are for 4.7.10. I'll move them to 4.8, since without working i915 I see no reason to stay on 4.7.
Offline
I have been doing research on sound and apparently this laptop was cursed with needing the cx2072x codec..
https://bugzilla.kernel.org/show_bug.cgi?id=115531
I get the same error "intel_sst_acpi 808622A8:00: No matching machine driver found" as the E200HA and the HP X2 N109. According to Pierre Bossart who works for Intel it is going to be a while before this is supported.
So if you have gotten sound to work on the T100HA.. a How-To would be nice especially since it could be used to help out at least 2 other devices.
jmiahman, I've gone through my DSDT and I've only found three RealTek audio ICs, ALC3270, ALC5642 and ALC5648. Nothing about cx2072x. And when I boot windows, it says my computer has ALC3270.
What does Windows say on your computer?
If you look here:
http://lxr.free-electrons.com/source/so … cpi.c#L377
ALC5642 is supported, but not ALC3270 nor ALC5648.
A wild guess is that they are quite alike, so if you are daring (no warranties, might blow your speakers, burn down your house, etc..)
add:
{"10EC5648", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
{"10EC3270", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
And have a try.
Offline
All patches are for 4.7.10. I'll move them to 4.8, since without working i915 I see no reason to stay on 4.7.
I'll give this a try on 4.9-RC2 and see where it goes. I already have the touchpad working as a Synaptics, I am used to 2 finger scrolling and 2 finger right click so I had already foind the t100ta patch edited and applied it. i915 has never worked for me, but I started with 4.8.
{"10EC5648", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
{"10EC3270", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
And have a try.
I don't think it will be that easy, but I will give this a try on 4.9 as well.
Last edited by jmiahman (2016-10-27 19:29:35)
Offline
gmelchett wrote:{"10EC5648", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
{"10EC3270", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
And have a try.I don't think it will be that easy, but I will give this a try on 4.9 as well.
Nope did not work...
Offline
gmelchett wrote:All patches are for 4.7.10. I'll move them to 4.8, since without working i915 I see no reason to stay on 4.7.
I'll give this a try on 4.9-RC2 and see where it goes. I already have the touchpad working as a Synaptics, I am used to 2 finger scrolling and 2 finger right click so I had already foind the t100ta patch edited and applied it. i915 has never worked for me, but I started with 4.8.
Thanks for the idea with configuring the touchpad as synaptics. I was about to implement tap in the driver.
Btw, got the touchscreen rotated correctly? I use voidptr's settings, (with 1133 instead of 1137) but the touchscreen is still in portrate mode. And xinput list-props misses rotatation option.
gmelchett wrote:{"10EC5648", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
{"10EC3270", "cht-bsw-rt5645", "intel/fw_sst_22a8.bin", "cht-bsw", NULL, &chv_platform_data },
And have a try.I don't think it will be that easy, but I will give this a try on 4.9 as well.
I admit, I was a very long shot. Too bad there is no I2S plugin in Wireshark for windows. Getting the basics can't be that hard once you have the information.
Edit: Was missing: Driver "evdev" in my xorg.conf in order to get the touchscreen rotated.
Last edited by gmelchett (2016-10-28 10:26:06)
Offline
Your patches work quite well on 4.9rc2. The brightness patch actually allows KDE to control the brightness and the touchpad. I had seen the touchpad patch before for the t100ta but not having the change the hardware ID is nice. Thanks
Offline
I admit, I was a very long shot. Too bad there is no I2S plugin in Wireshark for windows. Getting the basics can't be that hard once you have the information.
Edit: Was missing: Driver "evdev" in my xorg.conf in order to get the touchscreen rotated.
do you mean fbdev?.. yeah it's too bad it uses the frame buffer instead of i915. I have tried a few things and haven't been able to get it to work, however surfing the internet and the basics do okay with fbdev so I can't complain.. unlike sound, at least it works. So far when I need sound I have just been using a USB headset. I am tempted to buy a USB-C set of headphones (or adapter) and see if they work. I have another Laptop that has a USB-C port and my phone does as well, so it wouldn't be a complete waste if it didn't. If it has it's own DSP that's supported it would be a nice way to free up the USB port for a mouse and not have to carry around a hub.
Last edited by jmiahman (2016-10-28 16:23:31)
Offline
Your patches work quite well on 4.9rc2. The brightness patch actually allows KDE to control the brightness and the touchpad. I had seen the touchpad patch before for the t100ta but not having the change the hardware ID is nice. Thanks
Great! Thanks for testing!
Offline
gmelchett wrote:I admit, I was a very long shot. Too bad there is no I2S plugin in Wireshark for windows. Getting the basics can't be that hard once you have the information.
Edit: Was missing: Driver "evdev" in my xorg.conf in order to get the touchscreen rotated.
do you mean fbdev?.. yeah it's too bad it uses the frame buffer instead of i915. I have tried a few things and haven't been able to get it to work, however surfing the internet and the basics do okay with fbdev so I can't complain..
No, I mean the touchscreen, not the screen. The touchscreen gets by default the X "libinput" driver, which doesn't know about rotating/swapping. The X "evdev" driver does, but the touchscreen then becomes a single button mouse, which isn't that useful.
I've tried to use the X "synaptics" driver instead, but I've failed this far. I've changed synatics to allow not just touchpads, but somthing is lacking. I guess the kernel driver is missing somthing..
unlike sound, at least it works. So far when I need sound I have just been using a USB headset. I am tempted to buy a USB-C set of headphones (or adapter) and see if they work. I have another Laptop that has a USB-C port and my phone does as well, so it wouldn't be a complete waste if it didn't. If it has it's own DSP that's supported it would be a nice way to free up the USB port for a mouse and not have to carry around a hub.
There are whiskey cove patches for 4.9, so there is atleast a chance that the USB C port works.
Last edited by gmelchett (2016-10-29 07:03:22)
Offline
In case anyone is wondering for now, until the onboard sound starts working I bought a USB-C to headphone adapter for $7.99 from Amazon. It has a built in DSP, so it's works as it's own Sound Card. Because it uses USB-C and not the normal USB port it leaves my USB port open for a mouse. I have a USB-C phone as well and it works on my phone, so not a complete waste. It does have issues with using headphones that have built in mics. The audio is quiet and some what muted on such head phones, but normal 3.5 plug based headphones without a built in microphone work well. Most likely has to do with the 3rd connection for the mic.
It identifies as 0d8c:0013 C-Media Electronics, Inc.
https://www.amazon.com/Zinsoko-Headphon … B01M06D2Z9
Last edited by jmiahman (2016-10-31 00:05:42)
Offline
I've been trying to get the touchscreen to work as in windows, with tapping, two finger scrolling,..
As far as I've understood, libinput is nowadays the prefered way of handling inputs in X, evdev and synaptics beeing the old ones.
Unfortunally, I've ran into some X issue, from Xorg.0.log:
(II) Using input driver 'libinput' for 'SIS0457:00 0457:1133'
(**) SIS0457:00 0457:1133: always reports core events
(**) Option "Device" "/dev/input/mouse1"
(**) Option "_source" "server/udev"
(II) failed to create input device '/dev/input/mouse1'.
(EE) libinput: SIS0457:00 0457:1133: Failed to create a device for /dev/input/mouse1
(EE) PreInit returned 2 for "SIS0457:00 0457:1133"
(II) UnloadModule: "libinput"
It looks like it might already been fixed: https://bugs.freedesktop.org/show_bug.cgi?id=97928 but I don't feel for using bleeding edge on xserver + dependencies.
I've tested with building the driver hid-multitouch which is the touchscreen driver into the kernel, IE not as module, but that didn't help.
My xorg.conf, with libinput instead of synaptics for touchpad (tapping and multitouch working fine):
Section "Device"
Identifier "Video Device"
Driver "fbdev"
Option "UseFBDev" "true"
Option "Rotate" "CCW"
EndSection
Section "Screen"
Identifier "Default Screen"
Monitor "Main LCD"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "800x1280"
EndSubSection
EndSection
Section "InputClass"
Identifier "touchpad"
Driver "libinput"
MatchIsTouchpad "on"
Option "NaturalScrolling" "1"
Option "ScrollMethod" "twofinger"
Option "HorizontalScrolling" "1"
Option "SendEventsMode" "disabled-on-external-mouse"
Option "Tapping" "1"
Option "TappingButtonMap" "lrm"
Option "TappingDrag" "1"
Option "TappingDragLock" "0"
Option "DisableWhileTyping" "1"
EndSection
Section "InputClass"
Identifier "touchscreen"
Driver "libinput"
##Option "Device" "/dev/input/event11"
MatchIsTouchScreen "True"
Option "NaturalScrolling" "1"
Option "ScrollMethod" "twofinger"
Option "HorizontalScrolling" "1"
Option "RotationAngle" "270.0"
Option "Tapping" "1"
Option "TappingButtonMap" "lrm"
Option "TappingDrag" "1"
Option "TappingDragLock" "0"
Option "DisableWhileTyping" "0"
EndSection
Offline
There should be away to get the i915 driver working on 4.9rc3. See last comment on:
https://bugs.freedesktop.org/show_bug.cgi?id=97529
I tried to remove those lines from 4.8.4, but nothing changed there.
Offline
There should be away to get the i915 driver working on 4.9rc3. See last comment on:
https://bugs.freedesktop.org/show_bug.cgi?id=97529I tried to remove those lines from 4.8.4, but nothing changed there.
Hi,
did you try to off/on the output? Like this (or try DSI-1):
DISPLAY=:0 xrandr --output DSI1 --off
sleep 1
DISPLAY=:0 xrandr --output DSI1 --auto --rotate left
Offline
Does anyone have any information about the ALC3270? I`d like to make it work as well, but all I`ve found are windows drivers, information that it is probably used in some HP tablet and Japanese eshop selling this chip ...
Offline
gmelchett wrote:There should be away to get the i915 driver working on 4.9rc3. See last comment on:
https://bugs.freedesktop.org/show_bug.cgi?id=97529I tried to remove those lines from 4.8.4, but nothing changed there.
Hi,
did you try to off/on the output? Like this (or try DSI-1):DISPLAY=:0 xrandr --output DSI1 --off sleep 1 DISPLAY=:0 xrandr --output DSI1 --auto --rotate left
Yes, but modprobe got stuck (there is a long timeout) as it usually does with 4.8, with the backlight at maximum, so I doubt any commands after
modprobe got executed. But I missed the sleep 1.
I will retry via ssh, little easier to know what is going on.
Also, I'll have a check if these lines could be removed from 4.7, since that seems to be closer to functional than 4.8.
Offline
Does anyone have any information about the ALC3270? I`d like to make it work as well, but all I`ve found are windows drivers, information that it is probably used in some HP tablet and Japanese eshop selling this chip ...
About what I managed to find. I searched at realtek's site, but nothing there. :-(
The situation is about the same for bluetooth. "BCM2E72" suggests that it is broadcom, but the E doesn't follow their naming standard.
What the DSDT says is that it is connected to ttyS0 at 115200 with hardware flowcontrol.
Offline
Yes, but modprobe got stuck (there is a long timeout) as it usually does with 4.8, with the backlight at maximum, so I doubt any commands after
modprobe got executed. But I missed the sleep 1.
I will retry via ssh, little easier to know what is going on.
Strange, I`d expect 4.8 to work with this patch. There are still some timeouts, but it works. I am not sure that the sleep is necessary, I am using it just to be sure.
Also, I'll have a check if these lines could be removed from 4.7, since that seems to be closer to functional than 4.8.
The patch is reverting of two commits: a0a6d4ffd2adc76c192ebe7ef89b81e3f2f4a078, which is 4.6.0-rc3 and ea0000f0d369a59c2086fe9c489e0a2a86e080ba which is 4.7.0-rc2. So I`d expect 4.8 to work.
EDIT: Maybe you don`t have all necessary options in your config. I remember adding some, but I don`t know which ones You can take a look here https://github.com/p-an/T100HA-DSDT/blo … onfig-test I used this one for bisecting. It modified for this purpose.
EDIT: Feeling stupid... I described those options in the bugreport (https://bugs.freedesktop.org/show_bug.cgi?id=97529) and forgot about it
Last edited by pavan (2016-11-01 20:46:40)
Offline
gmelchett wrote:Yes, but modprobe got stuck (there is a long timeout) as it usually does with 4.8, with the backlight at maximum, so I doubt any commands after
modprobe got executed. But I missed the sleep 1.
I will retry via ssh, little easier to know what is going on.Strange, I`d expect 4.8 to work with this patch. There are still some timeouts, but it works. I am not sure that the sleep is necessary, I am using it just to be sure.
gmelchett wrote:Also, I'll have a check if these lines could be removed from 4.7, since that seems to be closer to functional than 4.8.
The patch is reverting of two commits: a0a6d4ffd2adc76c192ebe7ef89b81e3f2f4a078, which is 4.6.0-rc3 and ea0000f0d369a59c2086fe9c489e0a2a86e080ba which is 4.7.0-rc2. So I`d expect 4.8 to work.
EDIT: Maybe you don`t have all necessary options in your config. I remember adding some, but I don`t know which ones You can take a look here https://github.com/p-an/T100HA-DSDT/blo … onfig-test I used this one for bisecting. It modified for this purpose.
Your .config is quite different from mine. I based mine on what Arch Linux uses (=m on everything I realized). I'm doing two build with these two commits reverted, one on 4.8.6 and one on 4.9-rc3,
with my .config. If it doesn't work, I'll try with your .config.
EDIT: Feeling stupid... I described those options in the bugreport (https://bugs.freedesktop.org/show_bug.cgi?id=97529) and forgot about it
Offline
Your .config is quite different from mine. I based mine on what Arch Linux uses (=m on everything I realized). I'm doing two build with these two commits reverted, one on 4.8.6 and one on 4.9-rc3,
with my .config. If it doesn't work, I'll try with your .config.
That config was designed to do a quick build and test of kms, it was never intended for production use. Just check if you have the options mentioned in the bugreport. And one more note - the patch is not meant as a solution (although it works), it more a hint where to look while creating a proper fix.It is just reverting of two commit, which came out bisection as first bad.
Offline
Tested first with 4.9rc3. And it works fine with the xrandr off/auto workaround! Just three different i915 timeouts occurs in the darkness.
All three goes through drm_atomic_helper_check_planes(), so maybe removing these calls might help..
However, the touchpad stopped working as compiled in in rc3, so I will test with 4.8.6 as well.
Is reverting the cherryview gpio patch really needed? It looks a little out of place to me - who knows nothing about the i915 driver.
Offline
Tested first with 4.9rc3. And it works fine with the xrandr off/auto workaround! Just three different i915 timeouts occurs in the darkness.
All three goes through drm_atomic_helper_check_planes(), so maybe removing these calls might help..
Great
Maybe, but this requires deeper understanding how it works, which I don`t have.
However, the touchpad stopped working as compiled in in rc3, so I will test with 4.8.6 as well.
Does it work with default xorg configuration? Some of configs posted here disabled it for me.
It works for me with 4.9-rc2. I am using this xorg config for the rotation:
Section "InputClass"
Identifier "calibration"
MatchProduct "SIS0457:00 0457:113D"
Option "TransformationMatrix" "0 -1 1 1 0 0 0 0 1"
EndSection
My full xorg.conf is here http://www.transformerforums.com/forum/ … post359054
Is reverting the cherryview gpio patch really needed? It looks a little out of place to me - who knows nothing about the i915 driver.
I tried to keep it there and the screen was black. This was the firs bad commit I found. I guess that there is some error in that function, which breaks it on T100HA and it is better without it. But as I mentioned - the patch is not definitely a proper fix.
Offline
gmelchett wrote:Tested first with 4.9rc3. And it works fine with the xrandr off/auto workaround! Just three different i915 timeouts occurs in the darkness.
All three goes through drm_atomic_helper_check_planes(), so maybe removing these calls might help..Great
Maybe, but this requires deeper understanding how it works, which I don`t have.
Neither have I. Did have a look at the three back traces, two are identical, but neither looks like you could just comment out a line or two.
gmelchett wrote:However, the touchpad stopped working as compiled in in rc3, so I will test with 4.8.6 as well.
Does it work with default xorg configuration? Some of configs posted here disabled it for me.
It works for me with 4.9-rc2. I am using this xorg config for the rotation:
Section "InputClass" Identifier "calibration" MatchProduct "SIS0457:00 0457:113D" Option "TransformationMatrix" "0 -1 1 1 0 0 0 0 1" EndSection
Thanks, but it is a kernel issue. It said something about not being able to reset the touchpad at boot. But it only happened once.
My full xorg.conf is here http://www.transformerforums.com/forum/ … post359054
gmelchett wrote:Is reverting the cherryview gpio patch really needed? It looks a little out of place to me - who knows nothing about the i915 driver.
I tried to keep it there and the screen was black. This was the firs bad commit I found. I guess that there is some error in that function, which breaks it on T100HA and it is better without it. But as I mentioned - the patch is not definitely a proper fix.
Did a quick test as well, it is needs to be removed.
I also tested on 4.8.6 - both patches reverted. The workaround trick didn't work. But to my surprise, "shutdown -h now" does reset the console and that reset worked, so the shutdown log was shown on the screen.
EDIT: Second attempt with 4.8.6, manually running the xrandr commands, and it started to work. Don't know why it didn't the first time.
EDIT2: Forget what I wrote about the touchpad. I was all wrong.
Last edited by gmelchett (2016-11-02 12:11:34)
Offline
Did a quick test as well, it is needs to be removed.
I also tested on 4.8.6 - both patches reverted. The workaround trick didn't work. But to my surprise, "shutdown -h now" does reset the console and that reset worked, so the shutdown log was shown on the screen.
EDIT: Second attempt with 4.8.6, manually running the xrandr commands, and it started to work. Don't know why it didn't the first time.
EDIT2: Forget what I wrote about the touchpad. I was all wrong.
I'm building my Kernel 4.9RC3-t100ha kernel now. I'll let you know if I have different results. From reading it seems there is a confusion between the TouchPad/Mouse (which works for me with the Synaptics Driver and synclient) and the TouchScreen that does not work for me at all. Has anyone figured out the TouchScreen at this point?
Also.. for i915 should I black list the i915 driver and then manually load it at boot? Or just have a boot script run..
DISPLAY=:0 xrandr --output DSI1 --off
sleep 1
DISPLAY=:0 xrandr --output DSI1 --auto --rotate left
Last edited by jmiahman (2016-11-02 21:05:59)
Offline
gmelchett wrote:Did a quick test as well, it is needs to be removed.
I also tested on 4.8.6 - both patches reverted. The workaround trick didn't work. But to my surprise, "shutdown -h now" does reset the console and that reset worked, so the shutdown log was shown on the screen.
EDIT: Second attempt with 4.8.6, manually running the xrandr commands, and it started to work. Don't know why it didn't the first time.
EDIT2: Forget what I wrote about the touchpad. I was all wrong.I'm building my Kernel 4.9RC3-t100ha kernel now. I'll let you know if I have different results. From reading it seems there is a confusion between the TouchPad/Mouse (which works for me with the Synaptics Driver and synclient) and the TouchScreen that does not work for me at all. Has anyone figured out the TouchScreen at this point?
If you have hid-multitouch (CONFIG_HID_MULTITOUCH) driver running, the touchscreen in fully supported on kernel level. You can easily check with:
sudo evtest
and:
SIS0457:00 0457:1133
should be present. You can see how it acts differently on single and multitouch. Notice, it looks like not all T100HA touchscreen has device id 1133. Last figure differs.
However, I've not managed to get the touchscreen to work as a touchscreen in X, only as a mouse.
Also.. for i915 should I black list the i915 driver and then manually load it at boot? Or just have a boot script run..
DISPLAY=:0 xrandr --output DSI1 --off
sleep 1
DISPLAY=:0 xrandr --output DSI1 --auto --rotate left
You could probably do it either way, I have i915 blacklisted and I'm starting X like this:
sudo modprobe i915; startx
and I have the xrandr lines in my .xinitrc.
Offline
Asus T100HA status:
Wireless LA/N Yes, with DSDT fix, (brcmfmac)
SD card reader Yes, with DSDT fix, same as with wlan. card detect works
MMC Yes
Touchpad, multitouch Yes, with kernel patch - synaptics or libinput in X
Touchscreen, multitouch No, works on kernel level, but not as touchscreen in X
Framebuffer Yes, add fbcon=rotate:3 to kernel command line
I915, opengl Yes, requires two kernel patches to be reverted and a xrandr trick. Add video=LVDS-1:d to kernel command line.
Suspend/Resume No, i915 freezes (Not working with fb either). rmmod brcfmac required
Hardware buttons (volume up/down) No
Shutdown Yes
Sensors, rotation Yes, with 4.9 or later
Sensors, temperature Yes
Battery status Yes
Special laptop keys Yes, with patch - except toggle external monitor key.
Bluetooth No, BCM2E72 (Broadcom?) is not supported - Might be misnamed is the DSDT.
Audio speakers/3.5" connector No, RealTek ALC3270 is not supported.
USB 1.1/2.0 Yes
USB Type C Yes, at least audio. 4.9 or later required.
Cameras front/back Untested. One camera depends on a partly unpowered I2C bus, and will require some manual work.
Flash led Untested
Backlight settings Yes, with i915. Patch required for framebuffers
EDIT: Added backlight. Anything else missing and/or incorrect?
EDIT2: Rotation works
Last edited by gmelchett (2016-11-08 06:36:00)
Offline