00:21 <micken> Hello, I am bringing up ehci in Genode for rk3399. My writes to ehci regs have no effects , nothing changes. At first I thought that was my driver, but then I tried it in uboot with same results. Is there a write protection somwhere? The manual refence srgf .. but sgrf isn't documented.

00:22 <micken> EL2 btw

01:55 <jakllsch> micken: hm

01:56 <jakllsch> i assume you 'usb start' in u-boot?

06:04 <dsimic> naoki: ah, you're referring to the following lines in rk3399-rock-4se.dts:

06:04 <dsimic> 60 &vcc5v0_host {

06:04 <dsimic> 61 enable-active-high;

06:04 <dsimic> 62 gpio = <&gpio4 RK_PD1 GPIO_ACTIVE_HIGH>;

06:04 <dsimic> 63 pinctrl-names = "default";

06:04 <dsimic> 64 pinctrl-0 = <&vcc5v0_host_en>;

06:04 <dsimic> 65 };

06:06 <dsimic> those are already found in rk3399-rock-pi-4.dtsi, so yes, please submit a patch that removes those redundant lines from rk3399-rock-4se.dts

06:06 <dsimic> because rk3399-rock-4se.dts already includes rk3399-rock-pi-4.dtsi

06:24 <micken> jakllsch: yes ..

06:29 <mmind00> dsimic: you're late to the party, I already applied naoki 's patch doing that yesterday ;-)

06:32 <dsimic> mmind00: ah, I'm still a couple of weeks behind when it comes to reading email :(

06:32 <dsimic> still looking for a good temporary fix that should stop me from drowning in the incoming messages

06:34 <micken> jakllsch: I tried with stopping usb in uboot , no difference all registers are as they boot with. The status register have error bit (host) set

06:34 <dsimic> mmind00: while we're here, I'd like to know are you fine with the explanation that I offered in the email linked below?

06:34 <dsimic> https://lore.kernel.org/linux-phy/8bb17c887eb9122d90e0887068b32da2@manjaro.org/

06:46 <mmind00> dsimic: I really would just go with dev_err_probe ... as the doc to that function says, it is just fine to use for standard error handling too

06:46 <mmind00> dsimic: also having that one dev_err in there will just trigger someone else in the near future to convert it separately

06:48 <mmind00> afterall there seem to be a number of companies around that measure their employees "performance" in number of commits only

07:11 <dsimic> hmm, but then we should convert _all_ instances of dev_err() to dev_error_probe()

07:11 <dsimic> I mean, there would still be more than a few left

07:13 <dsimic> either we should replace them all, or use dev_err() where it fits better

07:34 <micken> jakllsch: haha found it .. first byte in 0x0 defines the start address of io regs

07:54 <dsimic> mmind00: however, I'll stop arguing with you on this rather irrelevant change, :) and I'll submit the v3 that drops this change

07:55 <mmind00> dsimic: correct ... things that are in the probe path should probably use dev_err_probe ;-)

07:58 <dsimic> true, and perhaps we should convert the remaining instances of dev_err() at some point

07:59 <dsimic> let's not do that now, to limit the scope of this patch series :)

07:59 <mmind00> exactly :-)

08:35 <dsimic> mmind00: here's the v3, please have a look when you get a chance

08:35 <dsimic> https://lore.kernel.org/linux-phy/cover.1725524802.git.dsimic@manjaro.org/T/#u

09:33 <dsimic> mmind00: thanks!

11:29 <naoki> hmm, I had no time for 5C today... zzz...

14:19 <jaganteki> is there any kodi integrations available with mainline codecs on rk3588?

14:25 <qschulz> jaganteki: maybe libreelec is using them?

14:25 <qschulz> considering Kwiboo is working on the ffmpeg v4l-requests support, I assume this may be related to the project :)

14:25 <jaganteki> Kwiboo: any lead?

15:15 <Kwiboo> jaganteki: kodi should support v4l2 request codecs as long as ffmpeg is patched to also support v4l2-request hwaccel, for libreelec we use https://github.com/LibreELEC/LibreELEC.tv/blob/master/packages/multimedia/ffmpeg/patches/v4l2-request/ffmpeg-001-v4l2-request.patch to patch ffmpeg

15:15 <Kwiboo> however using https://ffmpeg.org/pipermail/ffmpeg-devel/2024-August/332034.html kodi will require https://github.com/xbmc/xbmc/pull/25467 due to changes/fixes that was made for ffmpeg v2 submission

15:16 <Kwiboo> with recent addition of hantro nodes to rk3588 DT I would expect at least basic mpeg2/h264/vp8 to work with kodi, at least that works for rk356x

15:17 <jaganteki> Do we have any ingratiation build script that include these on rk3588 libreelec, already?

15:18 <jaganteki> s/ingratiation/intigrated

15:21 <Kwiboo> not something that is in a working state, only have an old libreelec PR for rk356x and an old local test for rk3588 (without hdmi), so need to be re-tested and rebased to get something that is somewhat running on rk3588

15:22 <diederik> The rk356x one: https://github.com/LibreELEC/LibreELEC.tv/pull/7864

15:25 <Kwiboo> my current focus is to complate a ffmpeg v4l2-request v3 series (with minor changes), need to re-test hevc on rpi4/rpi5 and test av1 on rk3588 before it will be sent

15:26 <Kwiboo> diederik: hehe, yeah, I should probably push what I currently have locally to that branch/PR ;-)

15:28 <diederik> I'm in no hurry :) I had already looked it up when I saw the conversation, so figured might as well share it

21:25 <linkmauve> Kwiboo, if you send me v3 I can test AV1, I have a rk3588 board setup already.