[Mesa-dev] Android + tegra + mesa
Thierry Reding
thierry.reding at gmail.com
Wed Jan 23 09:36:38 UTC 2019
On Fri, Jan 18, 2019 at 01:26:30PM +0300, Artyom Bambalov wrote:
> Do you have some notes on how to reproduce your setup? I've been meaning
> to try this out but never found enough time to see it through. Having a
> set of simple instructions for getting this built and booted would help
> reproduce the behavior and debug.
>
> My device is xiaomi mipad 1(codename is mocha). It's similar with with
> shieldtablet(tn8), but uses different lcd panel, pmic(tps65913 like in t114
> instead of as3xxx) and some other secondary hardware.
> My device tree(lineage-16.0 branch): https://github.com/arttttt/
> android_device_xiaomi_mocha_mainline
> My kernel(android-4.19_mocha-changes branch): https://github.com/Insei/linux
>
>
>
> That said, it looks like you've got a completely white display, which
> usually means that you're getting page faults from the SMMU, so maybe
> that'd be an interesting place to look at. You should be seeing errors
> from the SMMU in the kernel log in that case.
>
> The display can be white or black. It depends on real picture color.
>
>
>
> There was a brief period where the SMMU wasn't working properly with
> Nouveau, which could be related to this, especially since it was around
> the timeframe of 4.19. That issue was fixed with this commit:
>
> b59fb482b522 ("drm/nouveau: tegra: Detach from ARM DMA/IOMMU
> mapping")
>
> which was merged into v4.19, but maybe it's not in the tree from Google
> that you're using (for whatever reason).
>
> Come to think of it, that wouldn't really explain the white display,
> though, since that usually happens on SMMU faults for reads by the
> display controller. In any case, the kernel log would hopefully contain
> some clues as to what could be wrong.
>
> Thierry
>
>
> I attached dmesg. It contains pretty much nouveau errors.
[...]
> [ 202.341300] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341332] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341359] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341383] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 17e0 data 0000001e
> [ 202.341418] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341453] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341476] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.341499] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 17e0 data 0000001e
> [ 202.415128] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.415162] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.415188] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 0d78 data 00000004
> [ 202.415213] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 6 [0400441000 RenderThread[924]] subc 0 class a297 mthd 17e0 data 0000001e
These errors certainly look like something's wrong with the IOMMU. Can
you post the contents of this:
$ ls -l /sys/class/iommu/70019000.memory-controller/devices
from your system?
Thierry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20190123/e5d44864/attachment.sig>
More information about the mesa-dev
mailing list