peterm6881 has joined #Speedsaver
<peterm6881> i can confirm, as expected, no usb on mangopi-r3c
peterm6881 has quit [Remote host closed the connection]
peterm6881 has joined #Speedsaver
peterm6881 has quit [Remote host closed the connection]
peterm6881 has joined #Speedsaver
peter_ has joined #Speedsaver
peter_ has quit [Client Quit]
peterm6881 has quit [Quit: Leaving]
peterm6881 has joined #Speedsaver
<peterm6881> hey Xogium
<peterm6881> I'm curious, do we know why there's no usb?
<Speedsaver> Title: Press corner | European Commission (at ec.europa.eu)
peterm6881 has quit [Read error: Connection reset by peer]
peterm6881 has joined #Speedsaver
<peterm6881> hey Xogium
<peterm6881> is this the thing you checked?
<peterm6881> usbc0:usbc0@0 {
<peterm6881> device_type = "usbc0";
<peterm6881> compatible = "allwinner,sunxi-otg-manager";
<peterm6881> usb_port_type = <2>; /* 0: for device, 1: for host */
<peterm6881> usb_detect_type = <1>;
<peterm6881> usb_detect_mode = <0>;
<peterm6881> //usb_id_gpio = <&pio PD 9 0 1 1 1>;
<peterm6881> usb_id_gpio = <&pio PE 2 0 1 1 1>;
<peterm6881> //usb_det_vbus_gpio = <&pio PD 10 0 1 1 1>;
<peterm6881> usb_det_vbus_gpio = <&pio PE 2 0 1 1 1>;
<peterm6881> //usb_drv_vbus_gpio = <&pio PD 11 1 1 1 1>;
<peterm6881> usb_host_init_state = <0>;
<peterm6881> usb_regulator_io = "nocare";
<peterm6881> usb_wakeup_suspend = <0>;
<peterm6881> usb_luns = <1>;
<peterm6881> usb_serial_unique = <0>;
<peterm6881> usb_serial_number = "20080411";
<peterm6881> rndis_wceis = <1>;
<peterm6881> status = "okay";
<peterm6881> };
<peterm6881> "usb device works. Just configure usb_port_type to 2 in dts"
<peterm6881> bit weird, since it says 0: for device, 1: for host
<Xogium> I checked usb otg in the dts file
<peterm6881> what mode is specified
<Xogium> in board/widora/mangopi/r3/devicetree/linux/devicetree.dts, in the external tree
<peterm6881> whats it set to
<Xogium> &usb_otg { dr_mode = "peripheral"; /* host peripheral */ status = "okay";
<Xogium> };
<Xogium> that's all
<Xogium> sounds like what you quoted earlier is part of the dtsi, not the dts
peterm6881 has quit [Quit: Leaving]
peterm6881 has joined #Speedsaver
<peterm6881> ok, can you check that then
<Xogium> the dtsi just is for the SoC itself, not the board
<Xogium> when you create a board you includes the dtsi then you alter properties from inside the dts
<peterm6881> yeah, but thats where the problem might be
<Xogium> in the dtsi there is only usb_otg and usbphy
<Xogium> so no idea where you saw that usb thing earlier
<peterm6881> if its not enabled in the SoC, its not gonna work anywhere else
<Xogium> also don't see any of the properties you mentioned earlier
<Speedsaver> Title: f1c100s USB otg device 模式可以用吗? / 全志 SOC / WhyCan Forum(哇酷开发者社区) (at whycan-com.translate.goog)
<Xogium> the device tree from allwinner itself shows like this
<Xogium> usb_otg: usb@1c13000 {
<Xogium> compatible = "allwinner,suniv-f1c100s-musb";
<Xogium> resets = <&ccu RST_BUS_OTG>;
<Xogium> clocks = <&ccu CLK_BUS_OTG>;
<Xogium> reg = <0x01c13000 0x0400>;
<Xogium> interrupts = <26>;
<Xogium> interrupt-names = "mc";
<Xogium> phys = <&usbphy 0>;
<Xogium> phy-names = "usb";
<Xogium> extcon = <&usbphy 0>;
<Xogium> allwinner,sram = <&otg_sram 1>;
<Xogium> status = "disabled";
<Xogium> };
<Xogium> usbphy: phy@1c13400 {
<Xogium> compatible = "allwinner,suniv-f1c100s-usb-phy";
<Xogium> reg = <0x01c13400 0x10>;
<Xogium> reg-names = "phy_ctrl";
<Xogium> clocks = <&ccu CLK_USB_PHY0>;
<Xogium> clock-names = "usb0_phy";
<Xogium> resets = <&ccu RST_USB_PHY0>;
<Xogium> reset-names = "usb0_reset";
<Xogium> #phy-cells = <1>;
<Xogium> status = "disabled";
<Xogium> };
<Xogium> the status in the device tree has been set to okay for both, so they are definitely enabled
<Xogium> in the board's dt that is
<peterm6881> can you check if there is anything useful in this:
<peterm6881> hes definitely using usb
<Xogium> can try yeah
<Xogium> they use the board itself as a usb device in this case, looks like
<Xogium> usb mass storage
<Xogium> hmm
<peterm6881> ok erased
<peterm6881> ok burning the new rootfs.ubi
<Xogium> lets hope things don't go fzzzzt
<peterm6881> lol
<peterm6881> state(2) = dfuIDLE, status(0) = No error condition is present
<peterm6881> Done!
<peterm6881> Resetting USB to switch back to runtime mode
<peterm6881> so far so good
<peterm6881> ah crap I burned the standard speedsaver image by mistake, duh
<Xogium> lol
<peterm6881> erase
<peterm6881> oops
<peterm6881> ok re-erased and now downloading the correct rootfs.ubi
<peterm6881> success
<peterm6881> alright, im gonna be completely honest with you, its exactly the same
<peterm6881> and I deleted the file from yesterday
<peterm6881> nothing in lsusb
<Xogium> oh well
<Xogium> did you try to connect a dongle and check dmesg ? I suppose so
<peterm6881> yeah, its exactly the same as yesterday. Let me try the hub
<Xogium> so then if its the same freaking thing, it means its not just the fact the thing isn't an host but a peripheral by default, there's something else going on
<peterm6881> wow the mangopi-r3c is way shitter than we even gave it credit for
<Xogium> I don't know if its software or hardware being shit here
<peterm6881> is that greg guy still online?
<Xogium> but this is already way above my head
<Xogium> yep, he outright told me to not bother with armv5 architecture earlier lol
<Xogium> not because it buggy software or anything, but because its ultra old
<sc6502> hey all, just a quick visit.
<peterm6881> wouldnt hurt to ask if he can think of anything else
<peterm6881> hey sc6502
<peterm6881> anything to report?
<peterm6881> we cant get stupid otg to work on the mangopi-r3c, so no wifi
<Xogium> 16:43 < gregkh> Xogium: ick, good luck ! and odds are, you don't want to support it.
<Xogium> hah
<peterm6881> so even our alternative use case is a no go
<Xogium> ulo sc6502, hope you're going well
<peterm6881> Xogium, can you do a quick sanity check you definitely applied the setting
<sc6502> Mar 27 16:49:43 <Xogium> that is needed because otherwise you cannot build this with gcc >= 10… Multiple definitions of yylock make the build explode
<sc6502> make licheepi_zero_defconfig
<sc6502> make
<sc6502> Just curious, because I did
<sc6502> and it compiled straight away with gcc-10.3
<Xogium> I definitely applied it, yep
<Xogium> sc6502: hmm with my patch applied or without it ?
<sc6502> With our without, it still compiles to the end (I got an sdcard.img)
<Xogium> hmm
<Xogium> apparently br 2022.02 has some detection to apply this at compile time, if it finds that your u-boot is gonna run into that problem… but they said the euristic is fairly fragile, and I might have somehow bypassed it on my host system
<Xogium> it would stand to reason you haven't
<Xogium> there's a commit for this in buildroot… let me find it again
<sc6502> Are you on the latest release branch rather than master?
<sc6502> I was compiling against master
<Xogium> commit 3b3859cc7df8edce1 in buildroot
<Xogium> yes 2022.02
<Xogium> but that logic should also be present
<Xogium> +# Older versions break on gcc 10+ because of redefined symbols
<Xogium> +define UBOOT_DROP_YYLLOC
<Xogium> +endef
<Xogium> + |xargs -0 -r $(SED) '/^YYLTYPE yylloc;$$/d'
<Xogium> + $(Q)grep -Z -l -r -E '^YYLTYPE yylloc;$$' $(@D) \
<Xogium> +UBOOT_POST_PATCH_HOOKS += UBOOT_DROP_YYLLOC
<Xogium> +
<Xogium> I wonder why that spectacularly failed on my system but evidently didn't on yours
<peterm6881> it failed on mine, and I'm also using Ubuntu LTS (effectively)
<Xogium> peterm6881: you had another error
<peterm6881> 20.04
<peterm6881> yeah, but i cant remember what it was, only that you fixed it
<Xogium> but this u-boot isa down right mess on br 2022.02 and should be updated anyway, if for no other reason than buildroot deprecated and removed python 2 support, both host and target. So when you build it now on ubuntu or similar that still have python 2, it tries to use python from the real system, because that u-boot is so old is still requires python 2
<Xogium> that can lead to extremely weird behaviors unless you're lucky
<sc6502> Yeah, I had to back-install python2 on my system because of that.
<Xogium> so yeah… better to update this either way
<Xogium> even if the yylock thing isn't the primary objective
<Xogium> otherwise this'll get messy real quick
<Xogium> before 2022.02 there was a kconfig option, BR2_TARGET_UBOOT_NEEDS_PYTHON2
<Xogium> or similar
<Xogium> now there is only the python 3 variant of this
<peterm6881> Xogium, apart from the ancient u-boot, is there any other issue we should consider in the Lichee Zero stuff?
<Xogium> no, everything else seems sane… At least, so far. Its hard to say what things are really like without the hardware mind you
<Xogium> but so far its the only thing we could do for it
<peterm6881> understood
<peterm6881> I guess we need to bury Mangopi-R3c, pity, the price and that audio amplifier made it interesting. But its just too much pain. Thanks for trying
<Xogium> well fuck. I realized this would also mark the end of me building my router's system
<Xogium> the python 2 thingy
<Xogium> the u-boot the espressobin uses is so old it most certainly relies on python 2, and it was never mainlined properly far as I'm aware
<peterm6881> they have a nerve jacking the price up, since I doubt hardly anyone is buying it even at the previous price, since armV5 is dead
<peterm6881> Xogium, time for an upgrade ;)
<Xogium> oh well, was fun while it lasted, as they say. It was definitely not suitable on anything above 200 ~300 mbps anyway
<Xogium> the throughput was reduced by the dual core cpu, so even if it was claimed to be 1 gbps it couldn't ever reach this
<peterm6881> https://github.com/mangopi-sbc/buildroot-mangopi-r they havent even enabled Issues in github
<Xogium> lol
<peterm6881> the espressobin is destined for the Xogium museum
<Xogium> yep, sure looks like it
<Xogium> shame, it was a nice board, despite all of its shortcomings
<peterm6881> lets be nice and close this
<Xogium> yeah…
<peterm6881> because im gonna raise a new one about OTG
<peterm6881> unless you wanna do it?
<sc6502> Still puzzled here: checked out the 2022.02.x branch, configure for licheepi_zero_defconfig and that builds also.
<sc6502> Curiously, yylloc is never mentioned in any of the log files.
<Xogium> sc6502: its probably due to that logic I pointed to earlier, your host system somehow manages to do this step correctly
<Xogium> whereas mine falls flat on its face
<Xogium> hey, what's gotten into the bot, that it doesn't give us nice url titles
<Xogium> peterm6881: yep exactly what I was talking about when I said these guys didn't give a shit about mainlining
<peterm6881> thats why as a general rule I feel we should take ownership of that part of the process, since we've proved we cant rely on anyone else
<Xogium> that's what I mean when I said every piece of allwinner that got mainlined was done because of the community
<Xogium> it was done by nobodies with no relation to allwinner or any vendor
<Xogium> because vendors don't give a fuck
<peterm6881> its just a distraction to allwinner, their core business is selling cheap processors for cheap generic chinese goods
<peterm6881> where it becomes a problem when idiots try to make dev boards ;)
<peterm6881> Allwinner never intended to focus on that market
<Xogium> indeed
<peterm6881> Think it was about 10 years ago I bought A10 based tablets, with Android
<peterm6881> for 40 GBP
<peterm6881> for the kids
<Xogium> to be fair, next to no android devices are mainlined either
<peterm6881> its a weird world to dip into
<Xogium> android pretends to be linux but it really isn't… They each run their own linux kernel, with their own set of patches on top, and never upstream
<Xogium> even google
<peterm6881> sounds nightmarish
<Xogium> sure is
<Xogium> a lot of the vendors also don't even make the source code of their kernel public
<Xogium> so yeah, you'd be right to call android nightmarish
<sc6502> Xogium, what is your host?
<sc6502> peterm6881, have you tried to build licheepi_zero_defconfig and got the same result as Xogium ?
<Xogium> sc6502: archlinux
<peterm6881> sc6502, I got a different issue, I cant remember exact details, or if the same fix also fixed my issue as Xogium's
<peterm6881> But I think it did
<peterm6881> itll be in the irc chat
<peterm6881> i cant find it
<Xogium> sc6502: he got the python issue because of deprecation of python
<peterm6881> sounds right
<Xogium> I'm the only one that ran into yylock lol
<sc6502> So with the patch, we no longer need python2?
<Xogium> indeed
<Xogium> u-boot 2022.01 is recent enough that it doesn't require python 2 anymore
<sc6502> OK, I'll test that and use that as the reason :)
<peterm6881> sweet
<peterm6881> Xogium, does your new mangopi image have the playlist?
<peterm6881> im wondering if the usb wifi would work on the ttl micro usb socket
<peterm6881> i dont know enough about usb to know if thats even a thing, given the CH341 is already on that bus
<Xogium> probably wouldn't, I don't think
<peterm6881> confirmed. I waited a minute then pressed WPS, nothing connected to it
<peterm6881> this thing is officially useless, since they didnt expose usb anywhere else
<peterm6881> i wouldnt be interested in it as a tinker board without usb, even though its possible we could get WiFi through SDIO
<peterm6881> ive raised an Issue on Aodzips repo
<peterm6881> go ahead and close that DFU one, it might encourage him to respond. Even though he could have closed it himself
<Xogium> yeah I'll still let him know how it got fixed
<Xogium> least I can do
<peterm6881> :)
<peterm6881> Wish these bloody lichee zeros would arrive
<Xogium> hehe
peterm6881 has quit [Quit: Leaving]