08:36 <qschulz> phh: if Rockchip does it the same way they have done it recently, engineering samples for China for like a year (think Radxa), then Chinese products slowly start to be shipped and maybe a few months later you can source them outside of mainland China

08:37 <qschulz> end of 2026 is extremely optimistic. I would say end of 2026 for the first engineering samples to Chinese companies?

08:37 <qschulz> end of 2027 global release more probably?

08:37 <phh> qschulz: i don't think the availability outside China is made by rockchip, it's just a matter that rockchip overseas customers take more than a year to take a product

08:38 <qschulz> the big question will be if they repeat the same mistake they made with RK3588 of having nothing upstreamed for like a year before Collabora started working on it

08:38 <qschulz> phh: I can tell you they do not sell before they have ironed out everything with their Chinese companies

08:38 <phh> ok

08:39 <phh> qschulz: no idea about 2026 really. all i know is they started teasing it

08:39 <phh> but yes, it's is blurry enough that it can be years in the future

08:39 <qschulz> phh: yeah with absolutely no spec :/ We'll learn more once there are actual specs :)

08:39 <qschulz> phh: it's years in the future for sure :)

08:40 <qschulz> It'll be interesting to see if the GPU will be supported with Panthor fast enough, this would drive adoption of mainline faster

08:40 <mmind00> qschulz: with so many @arm.com people doing panthor patches these days .... :-)

08:40 <qschulz> hopefully the camera stack won't be too far off the RK3588 or newer chips which may get upstream support "soon" (end of the year, mid next year?)

08:41 <qschulz> mmind00: one never knows until it is done :)

08:42 <qschulz> to be clear, what we've been told by Rockchip is that they start their engineering samples only in China so that it's easier for them to iron out silicon bugs and whatnot by speaking the same language and being able to easily travel to the company's premises

08:42 <qschulz> it's not like nationalism (not saying it isn't :) )

08:43 <mmind00> which make sense in a way ... just that they could start upstreaming in parallel too :-)

08:44 <qschulz> yeah, that's what Intel (and likely AMD?) is doing?

08:44 <qschulz> not sure we're talking about the same amount of employees though :D

08:47 <mmind00> still, right now they're already doing a way better job ... the amount of high quality graphics patches from Andy is just great

08:49 <qschulz> yeah, that's encouraging, but if I remember what i've been told, the chips aren't THAT different from the RK3588 which took the biggest hit in terms of time required to upstream stuff

08:49 <qschulz> so if they do that with the RK3688 again...

08:49 <qschulz> but it's a good sign honestly, hopefully they keep at it!

08:57 <Ermine> have they announced the new chip?

08:58 <qschulz> it "leaked" from a Chinese conference last year I believe

08:58 <qschulz> https://www.cnx-software.com/2024/10/04/rockchip-rk3688-armv9-3-aiot-processor-to-feature-a-16-tops-npu-ufs-4-0-storage/

11:39 <CounterPillow> mmind00: I've wondered whether Andy is doing it in his free time or if he gets some time from RK to work on upstreaming. I do hope they pay him, but either way, I'm glad it's happening

14:35 <wens> wait, wasn't the rk3688 model name already used before?

14:38 <wens> I can imagine them ironing out stuff locally

14:39 <wens> on the flip side, you really don't want to work with engineering samples if possible :(

14:46 <robmur01> wens: nah, 3368. Not quite as bad as having 3288 and 3228 in the same generation :)

14:55 <Dex2x> seems like dietpi might have some decent support ?

14:59 <qschulz> Dex2x: upstream has some device tree for the Orange Pi 5 Ultra, maybe go with that?

14:59 <qschulz> depends on what you want to do with it

14:59 <qschulz> https://elixir.bootlin.com/linux/v6.15-rc7/source/arch/arm64/boot/dts/rockchip/rk3588-orangepi-5-ultra.dts

14:59 <qschulz> only in v6.15 (going to be released this week-end if everything goes well)

15:00 <CounterPillow> just build mainline u-boot yourself, flash it, have it pick up a debian rootfs, ye.

15:00 <Dex2x> ah okay will wait for that

15:00 <Dex2x> Would be nice if the hdmi in worked at some point

15:01 <qschulz> Dex2x: there's support for the controller, so could just be a question of modifying the device tree

15:01 <qschulz> c.f. Rock 5B

15:01 <qschulz> https://elixir.bootlin.com/linux/v6.15-rc7/source/arch/arm64/boot/dts/rockchip/rk3588-rock-5b.dts#L260 (the two nodes under the highlighted line)

15:06 <Dex2x> ah nice

17:51 <macromorgan> anyone get DSI panels working on mainline rk3588 with the latest patch series? I can't confirm it's accepting my DSI init sequence, but I can confirm it's timing out when I try to read the panel ID...

18:23 <CounterPillow> macromorgan: sre says it works for him on the EVB1

18:44 <macromorgan> okay I'll see if they're doing something I'm not

18:55 <macromorgan> sre doesn't have a tree floating around I could take a look at, do they?

19:00 <CounterPillow> Not that I could quickly find, but snooping around https://gitlab.collabora.com/sre and https://gitlab.collabora.com/hardware-enablement/rockchip-3588/linux may bring something up that I didn't spot

19:34 <macromorgan> can't find anything he's done in those branches sadly