<kikuchan>
loki666: Hi. you might need to enable CONFIG_SPI_GPIO, CONFIG_DRM_PANEL_NEWVISION_NV3052C, and CONFIG_BACKLIGHT_GPIO.
<tokyovigilante>
loki666: yup as kikuchan suggests that's probably what's missing, the panel won't be initialised without a backlight it seems
<loki666>
ok thanks
warpme has quit []
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
warpme has joined #linux-sunxi
<loki666>
ok i jave a blinking cursor
<tokyovigilante>
winning!
<tokyovigilante>
should have a message about drmfb and an 80x30 console if all working well
apritzel has joined #linux-sunxi
<MoeIcenowy>
mripard: it seems that multiple RISC-V SoCs try to reuse the Allwinner sunxi-ng CCU facility
<MoeIcenowy>
e.g. TH1520 driver by Yangtao Li / Drew Fustini
<MoeIcenowy>
and Spacemit K1's BSP driver
warpme has quit []
apritzel has quit [Ping timeout: 480 seconds]
warpme has joined #linux-sunxi
apritzel has joined #linux-sunxi
hramrach has joined #linux-sunxi
<apritzel>
tokyovigilante: I think the clock patches are independent and look ready to go. That's the LCD clock patches and the whole sun8i-de33.c
<apritzel>
tokyovigilante: can you please split out the latter into separate patches (but please observe jernej's authorship)
<apritzel>
and I guess we would need some input from jernej about the status of the remaining DE33 hacks
<apritzel>
are those hacks part of the current WIP trees, and require non-trivial DE driver refactoring? Or are they just addressing corner cases, and it somewhat works(TM) for the current use case?
<apritzel>
tokyovigilante: there is also this R40 LCD patch, which I find in your and jernej's tree, although jernej's seems older and uses an r40 compatible string, which seems better
<apritzel>
in general this feels like herding cats, so can we keep track of all those in the Wiki, maybe?
<apritzel>
for the pure panel work, we can use one of the Anbernic pages
<tokyovigilante>
thanks apritzel, that is really helpful. I agree the clocks and the DE33 enablement are pretty self-contained, and I think the YUV enablement for DE2/3 is fairly clean and necessary for subsequent HDMI. The main residual "hack" (according to jernej, it looks reasonable enough to my eye) is the VI scaler, which is only necessary for (I think) scaling for HDMI, so not needed for LCD. Happy to
<tokyovigilante>
revert to the R40 patch, its functionally identical for the H700 but I thought it might be queried as a random other SoC in the middle
<tokyovigilante>
Can definitely update the wiki too
<tokyovigilante>
Also, did you have a feeling on a vendor string for the LCD panel patch? A bit orthogonal but I can send a V2 if I get a steer on that
<Jookia>
tokyovigilante: thanks for dumping the init sequence :D
dsimic is now known as Guest7748
dsimic has joined #linux-sunxi
<apritzel>
tokyovigilante: ah, missed that panel patch, thanks for the heads up
Guest7748 has quit [Ping timeout: 480 seconds]
<Jookia>
tokyovigilante: does your panel have a blue tint?
<Jookia>
i'm not sure what's going on with my setup but linux has a blue tint while u-boot does not
warpme has quit []
<Jookia>
(i have finished getting my LCD panel working in u-boot)
<apritzel>
ah, vendor names of those companies, always fun. It's a bit odd, normally you'd expect people to write the company name all over their product
<apritzel>
If Rob concedes, no vendor prefix seems like the best to me
<tokyovigilante>
Jookia: no seems pretty neutral in linux, i havent tried (but intend to) u-boot
<tokyovigilante>
apritzel: thanks, will submit a v2 and see how it goes.
<Jookia>
tokyovigilante: i will submit a patch series soon that should help you when you get around to it. all you should have to do is modify the DE stuff
<Jookia>
kikuchan: you might need to give reviewed-by instead of signed-off on the kernel list. not sure
<Jookia>
signed-off is usually used as a chain of custody for merging in to the kernel, reviewed-by and tested-by (loki666 could perhaps do that?) are from people on the mailing list
apritzel has joined #linux-sunxi
warpme has joined #linux-sunxi
warpme has quit [Ping timeout: 480 seconds]
<kikuchan>
Jookia: Actually, I'm who found and extracted the initialize code of the panel. I should have been signed off on the line `Co-developed-by:`. I'm sorry for confusing you.
<Jookia>
kikuchan: oh cool. how'd you extract it? :D
<kikuchan>
logic analyzer and scope.
<Jookia>
ah yes that makes sense
<Jookia>
it's very good work :) i hope it gets merged
<kikuchan>
thanks :)
<apritzel>
jernej: tokyovigilante: I see that the "4GB" IOMMU patch dumps the kmem_cache in favour of raw allocation, to guarantee allocations for the page table below 4GB, since the whole IOMMU is still 32-bit only
<apritzel>
wouldn't specifying SLAB_CACHE_DMA32 as an additional flag in the kmem_create_cache() call achieve the same?