03:57 <wens> reminds me of an intermittent sata issue we had on a whole batch of servers at my previous employer

03:58 <wens> turns out it was some capacitor on some sata line was not discharging correctly

03:58 <wens> the sata drive could not be detected after reboot if the server was powered on for some time, but turning it off and letting it sit for a while made it come back to life

09:54 <a1batross> mriesch, hi! I remember you've been working on camera support on rk356x with mainline kernel? What's the current status on it? Thank you for your work!

10:13 <mmind00> diederik: now that I started looking at the sorting a lot more of those turn up :-D

10:14 <diederik> yeah, there are a couple more indeed. Guess I no longer need to send a (private) mail about those anymore ;)

10:18 <sigmaris> mmind00: I should rly send this to the mailing list, but not subscribed atm & I'm too lazy to reply with git-send-email :). I think the usb-c-connector on the Rock 5 ITX dts should not be a PD sink

10:20 <mmind00> sigmaris: I guess you mean because it can't power the board from it

10:20 <sigmaris> the schematic from Radxa shows TYPEC_VIN only connected to the output of regulator U2603, there is no power path I can see for it to power the board (on sheet 21 https://dl.radxa.com/rock5/5itx/radxa_rock_5_itx_X1100_schematic.pdf )

10:20 <sigmaris> yep, and powering via USB-C isn't mentioned in the other docs

10:23 <mmind00> sigmaris: so power-role = "source" I guess

10:26 <sigmaris> yep, and I'd remove try-power-role and sink-pdos (try-power-role seems optional if power-role is not dual)

10:31 <sigmaris> oh and also remove op-sink-microwatt

10:34 <mmind00> sigmaris: thanks :-) ... I've transcribed that into a reply to the patch now, to not forget ;-)

11:17 <qschulz> sigmaris: you can also download an mbox file from lore.kernel.org BTW

11:34 <diederik> yep, works really well. That's how I responded to "[PATCH v11 09/24] drm/rockchip: dw_hdmi: add regulator support"

11:35 <qschulz> Thunderbird has an add-on for this, called ImportExportToolsNG

14:38 <mriesch> a1batross: our solution works on the basis of v6.9 but is still heavily patched and not at all ready to be mainlined.

14:39 <mriesch> significant effort required to push this towards a v1 and no idea when this could possibly happen :-/

15:26 <a1batross> mriesch, hm, we're using 6.6 with custom patches. Is your work published anywhere? For now, I'm trying to figure out how exactly csidphy and ISP must be configured, and should it be ISP in the first place, or the VICAP, as I don't exactly understand the difference. As the hardware designed to have two cameras, DPHY must be set up in split mode. At least I figured out how to set up it this way.

15:27 <a1batross> two of our programmers tried to straight-up port BSP code on top of 6.6, but failed to do so. It just doesn't work for some reason.

16:05 <CounterPillow> Vicap just gives you the image data without any filtering, which is not very useful for camera input as that needs to be filtered to get a normal looking image. The PHY does the physical layer of the protocol. ISP is both a controller for ingesting the image data (like VICAP) as well as doing the required filtering with hardware

16:15 <a1batross> CounterPillow, thank you very much! I'll probably connect the important camera to ISP then

16:17 <CounterPillow> Since you're Russian this project is probably for killing people since we have seen RK3588 used in drones before so actually fuck you

16:17 <a1batross> CounterPillow, no, this is for agriculture :)

16:18 <a1batross> I'm not Russian citizen and do not intend to work for anything connected with the military

16:19 <CounterPillow> Okay, thank you

16:20 <a1batross> thank you for actually believing me

16:21 <a1batross> I understand that nobody wants to work with Russians these days, drones or not. Really, fuck drones, fuck this war.

21:02 <mriesch> a1batross: no, this work is not yet published anywhere. while it may be feasible to throw something over the fence, i feel i should revise the current state thoroughly first

21:03 <mriesch> i'll give this some more thought after my vacation