02:31 <rembo10> I think I'm facing this same issue: https://patchwork.kernel.org/project/linux-rockchip/patch/20190901212413.t561j%25alex14fr@gmail.com/ ... Is it a quirk with the psci? I'm dual booting an ereader with Android and Linux, and I think psci is the only way I can bring up the secondary cpus

10:30 <rembo10> I wonder if someone might be able to help me debug my dts. I'm having trouble with the usb otg and suspend... I think it has something to do with PMIC, but it's the only way I can bring up the secondary cpus

10:30 <rembo10> I mean PSCI

10:30 <rembo10> So many acronymns!

10:31 <phh> you mean usb otg and suspend simultaneously, or is that two independant issues? usb otg should be independant from psci (except during suspend ok)

10:31 <rembo10> Two independent issues

10:32 <rembo10> And also like, on some boots the wifi works fine, on others the interface doesn't show up

10:32 <rembo10> So I'm guessing it's something to do with the pinctrl? But I'm not really sure

10:32 <rembo10> I'm still learning about this stuff

10:32 <phh> pretty hard to say

10:33 <phh> first, when you say usb otg, do you mean that neither peripheral nor host mode works, or that only one works?

10:33 <rembo10> I saw some messages about failing to come out of L1 sleep state in 200us (for the usb), so I thought it was again something to do with power management

10:33 <rembo10> I have it set to peripheral mode in my dts

10:34 <phh> ok, did you setup configfs/functionfs/gadgets ?

10:34 <rembo10> On the board I have... It's a type c connector, so I don't know if that might have some effect? Might need some changes?

10:34 <rembo10> I set up gadgets

10:35 <rembo10> I'm trying to use g_ether

10:35 <rembo10> It has worked every once in a while

10:35 <rembo10> Other times I can ping the device but not ssh

10:35 <phh> I don't know much about type c, but I'm under the impression it requires specific setup yes

10:35 <rembo10> Other times the device freezes up

10:36 <phh> ah, so it's not totally not working

10:36 <rembo10> Yeah exactly, which makes it almost more frustrating!

10:37 <rembo10> And in the uboot.dtb that I managed to extract, there's nothing really fancy going on with the usb

10:37 <phh> so, it doesn't sound like a pinctrl issue. I don't know what's the L1 sleep state error you're speaking about, so yeah sounds like it could be related to psci. I could tell you to disable suspend modes temporarily, but it doesn't sound good for an ereader

10:38 <rembo10> I wonder if it's possible to bring up the secondary cpus without psci?

10:38 <rembo10> Or is it sort of required by the firmware?

10:39 <rembo10> I'm not sure if I posted this link: https://patchwork.kernel.org/project/linux-rockchip/patch/20190901212413.t561j%25alex14fr@gmail.com/

10:39 <rembo10> It's a simlar issue I think, for example I couldn't use timer 6-7 either

10:48 <phh> to bring up secondary CPUs, you need to use whatever bootloader's giving you. Looking at the change you're giving, rk3288 was historically NOT using psci (psci is a rather new thing, kinda mandated on armv8, but that's an armv7 platform), and was letting the OS control the CPUs itself, but newer rk3288 bootloaders implement PSCI and so you must go through PSCI to enable secondary cores

10:49 <rembo10> Interesting

10:49 <rembo10> That sort of fits in with what I was researching

10:50 <rembo10> I think I mentioned before, I'm dual booting with the original android system, so I don't really want to mess with the uboot to render that unusable

10:50 <rembo10> Since I'm not that confident I get the display working aha

10:53 <rembo10> I'm just not sure what to do

12:28 <phh> rembo10: fwiw, there are ways to keep the internal storage intact and boot from sdcard, but i don't think it's worth it in your case

12:29 <phh> (or even from usb)

12:32 <phh> rembo10: you have psci working for secondary-cpu-bringup?

12:33 <phh> rembo10: you can try looking at maybe a rk3328 device-tree that will be using PSCI both for secondary-cpu-bringup and suspend

12:35 <phh> rembo10: wait, i just realized you didn't say what's your soc

13:05 <rembo10> It's an rk3288w

13:05 <rembo10> I can bring up the secondary cpus with psci

13:05 <rembo10> I'll take a look at those device trees

13:07 <rembo10> The only rockchip one I found in mainline was the rk322x.dtsi

13:08 <rembo10> Oh that's interesting, they have opp-suspend in the operating points

13:10 <rembo10> The device I'm using doesn't have an sd card... I was thinking to solder one on but.. Well I think it would be the same issue right? Same bootloader...

13:19 <rembo10> Oh wait sorry I was looking in the arm directory not arm64... so dumb

13:22 <rembo10> I think that might be it... I need to have an idle-states node

13:22 <rembo10> It's weird that it's not in the dtb I extracted from the original system image

13:32 <rembo10> Hm, no dice... Maybe I'm testing it wrong? I have the gpio power key set as a wakeup source, I've also tried by setting the rtc0 wakealarm

13:33 <phh> well it could also be related to pmic

13:35 <rembo10> Wait a minute

13:35 <rembo10> I just realized I have CONFIG_TEE=n

13:37 <rembo10> I'm assuming I need that?

13:37 <phh> you need for it to match your bootloader's configuration

13:37 <phh> if your bootloader has tee then you need tee on

13:37 <rembo10> So I pulled the config from the android system

13:38 <rembo10> They have # CONFIG_TEE is not set

13:38 <rembo10> CONFIG_TEE_SUPPORT=y

13:38 <rembo10> Which seems contradictory aha

13:38 <rembo10> And I don't see CONFIG_TEE_SUPPORT in 5.15

13:41 <rembo10> But on the android side I see some messaages like [ 101.415284] Suspending console(s) (use no_console_suspend to debug)

13:41 <rembo10> INF TEE-CORE:rockchip_cpu_suspend:

13:50 <rembo10> There's also something called the ARM_PSCI_CHECKER

13:50 <rembo10> Run the PSCI checker during startup. This checks that hotplug and suspend operations work correctly when using PSCI.

13:50 <rembo10> So maybe it's all related!

13:54 <rembo10> Thank you so much for all your help btw, you've been enormously helpful

14:16 <rembo10> Ah man it froze my whole computer when I plugged the usb in aha

16:29 <Esmil> mmind00: Oh you have a starlight board too? (i just read the riscv list)

16:30 <mmind00> Esmil: yep ... and since last week a D1-Nezha too :-)

16:32 <Esmil> mmind00: Cool. If you're interested I can have Michael invite you to the StarFive slack channel where I can hazzle them for all the documentation they still haven't released

16:34 <Esmil> all the clock and reset info is basically from some internal excel sheets they sent me :/

16:34 <mmind00> Esmil: probably the only documentation for that at all ;-)

16:35 <Esmil> yeah

16:36 <mmind00> Esmil: so far I've always gotten around having anything todo with slack, so not sure if that is helpful ... I've made the board netboot and am more or less just following your branch for fiddling around with the board ;-)

16:37 <Esmil> mmind00: yes, me too, but both $work, the beagle foundation and riscv are using slack now :(

16:37 <Esmil> anyway the invitation is open. let me know if you want to have a look some day

16:46 <mmind00> hehe ok :-)