17:29 <marc|gonzalez> narmstrong: the "ask TEE to load vdec FW" patch is almost ready I think ( https://paste.debian.net/1308665/ )

17:29 <marc|gonzalez> I asked phh to remember how he chose the sleep durations, he's working on remembering :)

17:30 <marc|gonzalez> Since the loading functions are almost identical, I considered some kind of factorization, but so far, this is what I've come up with

17:31 <phh> there might also be the question of the list of codecs to maintain?

17:31 <phh> idk if we need to explain h265 vs h265_mmu (it should be clear to the person who wrote the original media driver though)

17:33 <marc|gonzalez> (looks like pastebin broke the tabs, erf)

17:33 <marc|gonzalez> (or rather, it look like they have a tabstop=4 on display)

20:28 <minute> hmm, this 6.6.15 kernel is spamming me with "panfrost l2 power transition timeout" and "l2 power transition timeout" on a311d

20:39 <xdarklight> minute: I think chewitt was suspicious about commit 22aa1a209018dc2eca78745f7666db63637cd5dc - maybe try reverting it. if it "solves" that issue then it's time to report it upstream

20:39 <minute> xdarklight: ok thanks, i can try that tomorrow

20:49 <minute> now trying to integrate a311d with tlv320aic3100... so far "aic31xx_setup_pll: Sample rate (48000) and format not supported" weird

20:50 <minute> which of the 3 CLKID_MPLLs are used to clock dai-link-7? (the "analog audio")

20:53 <xdarklight> on older SoCs (with AIU audio instead of the newer AUGE) the MPLLs can feed any output (HDMI, SPDIF, analog codec, ...)

20:53 <minute> mhm

20:53 <xdarklight> on those older SoCs the setup is more like: which MPLL can get closest to the desired output rate

20:53 <minute> this is automatic?

20:54 <minute> ok, i guess i have to understand the algorithm in aic31xx_setup_pll

20:57 <minute> > tlv320aic31xx-codec 1-0018: ## aic31xx_hw_params: width 32 rate 48000

20:57 <minute> 32 bit huh

21:38 <minute> ok bitness and rates doesn't matter, none work

21:41 <minute> oh ok aic31xx_rate_divs is a hardcoded table

21:51 <minute> aha https://patchwork.amarulasolutions.com/patch/1742/

22:41 <Consolatis> recently I tried to hack around wlroots to make it use the overlay plane as a cursor plane but the overlay plane only seems to support various NV / YUV formats which don't play nicely with usual RGB based cursor formats. Additionally the overlay plane seems actually to be an "underlay" plane which makes it pretty useless as a cursor plane. Now I am wondering how many planes for example gxbb actually supports via hardware and if it might be easier t

22:41 <Consolatis> o expose an actual cursor plane with support for RGB formats to userspace instead. Are there maybe already patches floating around for that?

22:42 <Consolatis> I know it is listed as as TODO entry at linux-meson.com but that has been the case for a long time already so maybe there was some work done on that in the meantime