01:02 <apritzel> smaeul: could you please have a brief look at this one: https://github.com/linux-sunxi/sunxi-tools/pull/189 (Oddly enough I cannot assign you as a reviewer)

01:18 <smaeul> apritzel: I don't personally have any plans for it

01:18 <smaeul> the main use I have seen suggested is for projects using the AR100 for real-time GPIO control

01:19 <apritzel> smaeul: alright, thanks, I was just wondering why people care about Linux' hwspinlock driver

01:20 <smaeul> we could assign one lock to each gpio port and use them from the Linux driver, but I don't see much motivation to do that until somebody says that would actually be useful

01:22 <apritzel> yeah, I see that there are problems for the solution ;-), but was curious about actual usage

01:27 <smaeul> that pull request looks reasonable to me

01:27 <apritzel> many thanks!

01:29 <smaeul> some of the SID layouts are documented on the wiki, but it seems to change quite frequently

01:34 <apritzel> yeah, I already converted the two tables from the wiki into code, will post them laster

01:35 <apritzel> most fuses read as 0, though, so it's only mildly interesting

08:03 <LordKalma> Linus Tech Tips releases a video recomending the banana pi 3 lts as the better rpi alternative, since they are unabtanium right now. I think I should thank you all again for supporting the allwinner platform

08:44 <mps> LordKalma: do you have url of it?

09:00 <LordKalma> mps, sure, https://www.youtube.com/watch?v=uJvCVw1yONQ

09:01 <LordKalma> (he talks about a full range of devices, but settles on the bpi 3 lts)

09:02 <LordKalma> And I meant Orange Pi*

09:50 <mps> LordKalma: thanks. also I prefer allwinner, never bought any rpi

09:51 <mps> even bought olimex A64 allwinner based notebook to learn and play with this SoC

14:15 <junari> apritzel: a few days ago you were talking about the 25M clock source for ethernet phy

14:17 <junari> I took up the second ethernet on my devboard and it is without a separate oscillator

14:18 <junari> Does the dwmac-sun8i driver support this feature?

14:18 <apritzel> it's not really a MAC feature. There is a clock output pin, driven by the CCU

14:19 <apritzel> all you would need to do is to enable this gate clock

14:21 <junari> pinctrl in dts?

14:21 <junari> Is there an example? Or a keyword to search for?

14:25 <apritzel> no, ideally this would be a clock property in the PHY DT node

14:25 <apritzel> problem is that the generic PHY binding does not support this yet, IIRC

14:27 <apritzel> which I tried to fix here: https://lore.kernel.org/netdev/20220704004533.17762-1-andre.przywara@arm.com/

14:31 <junari> hmm, interesting. Thanks

14:33 <apritzel> with that patch in, you just add: clocks = <&ccu CLK_EMAC_25M>; to the PHY DT node, and it should work

16:07 <jernej> LordKalma: funny, OPi 3 LTS is not even supported in mainline yet. But I guess many distros have already appropriate patches for it.

16:22 <apritzel> jernej: is it just DT changes? Or are there differences that require code patches? (Ignoring the PHY power up sequence for a moment)

16:25 <jernej> PHY actually works with delay added to regulator, I just used wrong properties

16:26 <jernej> I'm not sure if generic PHY driver works or not (proper one will be in 6.3), but other than that, it's only DT

16:26 <jernej> and maybe those DRAM issues I talked about, but that's not Linux thing

16:39 <apritzel> I see, thanks