You can still see the hdmi output (at least) on your R2.
currently i have no output but maybe the dts-node are changed…only updated nodes from your resend patch after adding nodes from my other branch
i try to make a clean branch with these patches based on
- https://patchwork.kernel.org/project/linux-mediatek/list/?series=15283
- https://patchwork.kernel.org/patch/10588951/
- https://patchwork.kernel.org/project/linux-mediatek/list/?series=25989
and your additional patch
https://github.com/frank-w/BPI-R2-4.14/tree/4.19-rc7-hdmiv5
still “unsupported mode”, no mediatek-hdmi-ddc, no other fail-message
if i kill xinit i get endless repeating
mtk-iommu-v1 10205000.mmsys_iommu: fault type=0x5 iova=0x731000 pa=0x0 larb=0 port=1
drm.log (31,4 KB)
but you’ve said, that these patches are incomplete…so i just wait
but you’ve said, that these patches are incomplete…so i just wait
Though we have some patches in the queue, we still could use drm/hdmi normally.
i try to make a clean branch with these patches based on
https://patchwork.kernel.org/project/linux-mediatek/list/?series=15283 1 https://patchwork.kernel.org/patch/10588951/ 1 https://patchwork.kernel.org/project/linux-mediatek/list/?series=25989 2
I tested kernel 4.19 with these patches (and it works), so I assumed you should get a successful result.
i’ve added these patches to above repo (https://github.com/frank-w/BPI-R2-4.14/tree/4.19-rc7-hdmiv5) but it does not work on my TV, imho v3-patches show xserver, but v5 does not work
same with https://patchwork.kernel.org/patch/10609007/ (v3 12/12) applied
log including x-server: 4.19-rc7-hdmiv5.log (60,4 KB)
from logs it looks well, but unsupported mode on TV instead of working X-Display…i guess it’s a hsync/vsync-issue
Have you found a problem in these patches? Hoping these do not go in mainline because then i need to patch them out to include my old patches
I still could not reproduce your problem in my R2 (I can see debian desktop).
Some of these patches have already been merged, and the only problem is that we need to reconnect the cable if we want to boot up with the 720p (1080p is ok). So… I guess you can try to reconnect it.
I have used 1080p and have reconnected cable…have you compiled my branch for testin (to compare with your code)?
I guess it is a hsync/vsync-problem because of the message,but it does not show settings it gets… btw. I have a 10m cable between r2 and tv
i see no merge in 4.19-rc…maybe it’s merged to matthias bruggers or drm-tree, but not yet in 4.19
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/mediatek
drivers/gpu/drm/mediatek/mtk_hdmi_phy.c and drivers/gpu/drm/mediatek/mtk_mt2701_hdmi_phy.c are missing, mst of the other files related to mt7623 are last changed before september 2018
as you see in my logs i try to bootup with 1080p as this setup works with old drivers…all other does not.
Hi @Ryder.Lee!
I’ve compiled and tested Frank’s https://github.com/frank-w/BPI-R2-4.14/tree/4.19-rc7-hdmiv5 branch. It’s doesn’t work for me: no signal on boot, and this after hdmi reconnect.
P.S. my TV is 1080p.
if anyone wants to test, i have created a branch where all (wifi,poweroff,hdmiv1+fbdev) is merged together (don’t want to merge to main-branch, because i have not tested much):
and uploaded to my gdrive https://drive.google.com/open?id=1MY8owoBdyv060u0QcE0GVC1YEtIPgu5E
hdmi, wifi and poweroff seem to work on a quick test 
please give response if it works for you or if you have any issues
Is it barnch named 4.19-main on your github?
the merged is named 4.19-merged, see my link 
4.19-main currently have only build-in-features of 4.19 without wifi, hdmi and poweroff (have not much tested these). will merge to main if no new issue is found
I got xserver running with additional patches from bibby (askes him if i can push these 2 patches before patchwork-release),currently try to get fbdev working with @deadmeat
with the help from bibby we got hdmi working and @DeadMeat helped me with fbdev
https://github.com/frank-w/BPI-R2-4.14/tree/4.19-hdmiv5
in my tests it works so far for
- 1080p hdmi direct on TV
- 1280x1024 (hdmi2dvi), 1024x768 (hdmi2dvi) => in xserver seems ignoring my “video=” cmdline (always 1280x1024 is selected), but i see larger font on fbdev
- 1024x768 (hdmi2vga) now works
- 640x480 did not display anything
*x600 seems still broken
- 800x600 tested…Looks weird
compared with v1-patches 1024x768 is now supported
@DeadMeat please test your resolutions you’ve tested with v1 Searching testing people for hdmi + wifi in Kernel 4.16
Ok. I,ve tested: 1920x1080@60, 1680x1050@60, 1600x900@60, 1280x1024@60;75, 1440x900@60, 1366x768@60, 1280x800@60, 1152x864@75, 1280x720@60, 1024x768@60
Theese modes are ok. next are available in Xorg, but not working for me: 832x624, 800x600, 720x576, 720x480, 640x480, and 720x400.
TV shows no hdmi signal, so i don’t know if it’s a TV issue or BPI-r2. Also I’ve tested 1024x600 on 7" LCD - the screen is the same(or very similar) as on Frank’s 800x600 image.
1 Like
Thank you for your tests
Can you try to prepare fbdev-patch so we can send it to mainline?
Let’s try, are there any requirements/manual? I’ve never sent any patches to mainline
In my opinion the code must meet some mainline requirements.
For now mtk’s fbdev is pretty similar to rockchip’s fbdev, which is in mainline (literally rockchip_drm_fbdev was kinda of manual for porting mtk_drm_fbdev from 4.4 to 4.14), so I’ll really appreciate anyone’s help to find a start point, what to do :).
First check if you have unnecessary code or whitespace-errors. Then commit with -s (sign).
Get patch (navigate to commot on github and add .patch to url)
Then run scripts/checkpatch.pl and if it find no problems run scripts/get_maintainer.pl
Send as textmail to the emails printed by maintainers-script and look what they say
created new branch based on 4.19-main (for easier merging) and with fixed author-information
seems i miss anything…i got purple with stretched (width) console-font and it seems to hang (does not start xserver) and did not get a login-prompt…login from debug-uart is possible
fixed that…i had forgotten to cherry-pick the Patch “drm/mediatek: config component output by device node port”…now it works, sorry my fault 
Tested - works good for me. Both fbdev and Xorg.
Can you check for Copyright/SPDX? Then modify the patch and send it to mainline…
You can also compare your fbdev with your teacher-driver…maybe one combined driver is possible
Bibby wrote that they will send new patchset maybe next week (including the 3 fixes)