05:19 <digetx> tagr: any updates on the memory bw patchset? the display controller doesn't work well without it. I saw that you sent out the 5.14 drm PR, making me wonder what happened to v18

14:45 <digetx> m4t: pgwipeout was trying to use CC2 gpio for the nct1008 interrupt, but the CC2 pin is used by I2C, so it's wrong gpio

14:46 <m4t> hmm, do we know the correct one?

14:48 <digetx> interestingly, nexus7 schematics also has wrong CC2 gpio connected to TEMP_ALERT on one page and S3/S6 on other page, the S3 works

14:56 <digetx> that interrupt is very optional, kernel doesn't make use of it other than to print a warning message when temperature goes out of range, I wanted to let you know about it

14:56 <m4t> there was something where that message kept being repeated over and over again

14:56 <digetx> it's either connected to a different gpio on ouya or not connected at all

14:58 <digetx> the non-stopping messages are the separate driver issue, I'll send patch later today to switch lm90 to use edge trigger instead of level

14:58 <m4t> https://github.com/stephank/linux/commit/4db77995445b00f0f9c05ccf7af3cfbe5937c28d

14:58 <m4t> maybe unrelated but its like the interrupt was never cleared

14:59 <m4t> oh

14:59 <digetx> the CC2 is jumping 1/0 non-stop, I think

14:59 <m4t> i2c clock or something?

15:00 <digetx> ah, it's not i2c, but i2s

15:00 <digetx> nvidia,function = "i2s4" for pcc2

15:02 <digetx> although, i2s4 not used on ouya

15:04 <m4t> https://github.com/stephank/linux/blob/dbb5b36873733f0aa590788b72a1cc64e4215d67/arch/arm/boot/dts/tegra30-cardhu.dtsi#L331-L338

15:05 <m4t> originated there?

15:05 <digetx> maybe

15:06 <digetx> the cc2 also doesn't stop to trigger as interrupt for both low/high trigger levels on nexus7

15:08 <digetx> the lm90 doesn't support clearing interrupt, it just checks the temperature status on interrupt

15:09 <digetx> technically all those device-trees that use IRQ_TYPE_LEVEL_LOW are incorrect, it should be IRQ_TYPE_EDGE_FALLING

15:10 <m4t> makes sense

21:06 <digetx> m4t: have you updated alpine?

21:06 <m4t> no, was there a release?

21:06 <digetx> yes

21:07 <m4t> okay, i will try to tonight

21:07 <digetx> I tried to enable tegra-hda in nexus7 DT and see that pulseaudio autoloads the new ucm

21:08 <digetx> but that's on gentoo

21:08 <digetx> i.e. the most recent ucm + alsalib

21:27 <digetx> seems alpine 3.14 has a fresh alsa-lib, but not fresh alsa-ucm-conf, so you will need to pull recent ucm manually

21:33 <m4t> just using alsa directly on the ouya

22:10 <m4t> alpine is really eager to remove all system packages, hah https://paste.ee/p/a0aGj#rm7WemuLgSb1Qs1hDtLEXGy3pJBgLJU8

22:11 <m4t> (it's because i didn't do an apk update beforehand)

23:10 <m4t> still alsaucm is syntax 3. i can try downgrading the ver of the ouya stuff again...

23:15 <m4t> it seems pretty broken/out of sync anyways, it only looks at virt.conf.d (hard-coded in alsa, i guess):

23:15 <m4t> open("/usr/share/alsa/ucm2/conf.virt.d", O_RDONLY|O_LARGEFILE|O_CLOEXEC|O_DIRECTORY) = -1 ENOENT (No such file or directory)

23:15 <m4t> but that dir doesn't exist in the alsa-ucm-conf package....

23:21 <m4t> okay, so i cloned the repo, mv alsa-ucm-conf/ucm2 /usr/share/alsa; mv virt.conf.d virt.conf.d.bak; mv conf.d virt.conf.d; ln -s virt.conf.d/tegra; ln -s virt.conf.d/tegra-hda;

23:21 <m4t> ouya:/usr/share/alsa/ucm2# alsaucm listcards

23:21 <m4t> 0: tegra-hda

23:21 <m4t> what a PITA