ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at https://oftc.irclog.whitequark.org/linux-sunxi
evgeny_boger has quit [Ping timeout: 480 seconds]
grming has quit [Quit: Konversation terminated!]
vagrantc has quit [Quit: leaving]
cnxsoft has joined #linux-sunxi
<smaeul> apritzel: no, I don't plan to have <asm/arch/cpu.h> on RISC-V at all. MMIO addresses come from the devicetree only
<MoeIcenowy> smaeul: we cannot have DT sometimes
<MoeIcenowy> e.g. in size-constrained SPL
<smaeul> thankfully on RISC-V we don't have size-constrained SPL :)
<MoeIcenowy> smaeul: but keeping SPL the same helps code sharing
<smaeul> true, though I'd say not using any code in the first place is even better
<MoeIcenowy> smaeul: well if using of to platdata is it possible to keep A10 working?
<smaeul> you may underestimate just how little additional code is needed to make the normal drivers work in SPL
<smaeul> I'm not suggesting getting rid of those headers for ARM
<MoeIcenowy> smaeul: okay
<MoeIcenowy> but then we will finally meet the situation that to choose to get A10/A13/A20 not maintained or to just keep non-DM SPL
<smaeul> again, I'm happy with keeping anything before H6 with non-DM SPL -- that includes A64, so that means some maintainership at least
<MoeIcenowy> smaeul: well sounds reasonable
<MoeIcenowy> migrate "NCAT" to DM SPL?
<smaeul> yeah, H6 is where the 32k size limit goes away
<smaeul> even I don't want to deal with OF platdata :)
<MoeIcenowy> smaeul: yes it really sucks ;-)
rajkosto has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
DasChaos has joined #linux-sunxi
DasChaos is now known as Guest3359
rajkosto has quit [Read error: Connection reset by peer]
bauen1 has joined #linux-sunxi
evgeny_boger has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
bauen1 has joined #linux-sunxi
apritzel has joined #linux-sunxi
junari has joined #linux-sunxi
<gamiee> Hello everyone, A64 SoC, how to deal with situation, where I need to use both eMMC and SPI, but SPI_MISO is multiplexed with SDC2_DS ? Is it possible to have both enabled? (I saw somewhere that DS is only used for HS400 ?) Thanks
<MoeIcenowy> gamiee: what board?
<MoeIcenowy> and I don't think we support HS400 now
<MoeIcenowy> BTW in mainline I think DS is a dedicated pinctrl node
<gamiee> MoeIcenowy: PINE A64-LTS V2. The strange thing is, that on PINE A64-LTS V1.2, eMMC (works???) and also SPI_MISO works OK, I can communicate with SPI device without any issues. But on V2, it receives nothing.
<MoeIcenowy> gamiee: please ensure whether a flash is present
<MoeIcenowy> my V2 just lacks a flash
<gamiee> Flash is missing on both of them.
<gamiee> I'm using external SPI device mounted on PI-2 bus.
<MoeIcenowy> then what happened?
<MoeIcenowy> I think we mostly use the same DT on V1 and V2
<MoeIcenowy> maybe some PCB design change?
<MoeIcenowy> ah, not mostly, but just the same
<MoeIcenowy> V2 just changed Ethernet PHY
<gamiee> AFAIK, only difference between V1 and V2 is that it have Motorcomm PHY
<MoeIcenowy> which is automatically probed via MDIO
<gamiee> Yes
<MoeIcenowy> gamiee: BTW is Pi-2 bus SPI0?
<gamiee> MoeIcenowy: correct. PC0 - PC3
<MoeIcenowy> oops it is
<gamiee> Which is also used on SPI Flash
<gamiee> And chip selects are the same
apritzel_ has joined #linux-sunxi
<MoeIcenowy> gamiee: try to switch eMMC modules between two boards?
<MoeIcenowy> I think PC1 is always connected to eMMC, even if not used
apritzel_ has quit []
<gamiee> Anyway, it's funny, because, same OS image: PINE A64-LTS V1.2, (SanDisk eMMC), SPI works OK | PINE A64-LTS V2 (FOREESEE eMMC), SPI doesn't work
<gamiee> Yeah, just going to try that
<MoeIcenowy> if changing eMMC module works, maybe the problem will need to be delivered to tllim
<gamiee> Yes, I will do that, as I am testing PINE64 LoRa gateway module.
<gamiee> Yes, using SanDisk eMMC solved the issue and SPI works
<gamiee> Although, eMMC is not probed. Where Foreesee eMMC was possible to probe
<gamiee> https://paste.debian.net/1257446/ here is dmesg, so it have difficulties with probing
<gamiee> MoeIcenowy: is there something I can do to make eMMC working? Since HS400 is not used, the pin can be re-used by SPI instead, or?
bauen1_ has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
bauen1_ has quit [Ping timeout: 480 seconds]
<MoeIcenowy> gamiee: I think some problems are here
<MoeIcenowy> I will contact TL
<gamiee> MoeIcenowy: software side or hardware?
<MoeIcenowy> HW
<gamiee> Oh, so I guess V2 related. Okay, thank you :)
bauen1 has joined #linux-sunxi
<MoeIcenowy> gamiee: not V2 related
<MoeIcenowy> Although eMMC related
<gamiee> Ah, makes sense. So I guess, future PINE64 eMMC modules will keep DS not connected, since driver doesn't support HS400 anyway?
<freemangordon> ok, USB on a33 is broken on every kernel since 4.11, /me is bisecting
<karlp> that smells more like you're missing a now required config setting?
junari has quit [Ping timeout: 480 seconds]
<freemangordon> yes, but I cannot figure out which one
<freemangordon> seems the missing piece is CONFIG_USB_ETH_RNDIS=y
<montjoie> freemangordon: my A33 seems USB okay on 5.19.x
<freemangordon> could you share your .config?
<montjoie> I just did make sunxi_defconfig
<montjoie> let me upload it
<freemangordon> no
<freemangordon> this does not even boot here
<montjoie> booted on sun8i-a33-olinuxino
<freemangordon> thanks
<MoeIcenowy> freemangordon: could I have your config?
<MoeIcenowy> I wonder what happened
<freemangordon> MoeIcenowy: sure, sec
<freemangordon> MoeIcenowy: minimal (based on defconfig with just minimal changes) or with all the other things (like wifi driver TS, etc)?
<freemangordon> MoeIcenowy: https://pastebin.com/eSNvc5Wy
<freemangordon> this is based on defconfig with just few changes so USB to work
<MoeIcenowy> well looks not problematic
<MoeIcenowy> did you try host only in DT?
<freemangordon> no
<freemangordon> as I said ^^^, when you add CONFIG_USB_ETH_RNDIS=y, OTG starts wotking
<freemangordon> *working
<libv> which otg devices are you testing?
<libv> ah, keyboard
<freemangordon> and a mouse
<libv> so perhaps figure out which gpio the id pin is attached and then find out what state that pin is in with and without rndis built in
<freemangordon> id is ok
<freemangordon> it changes state in /sys/kernel/debug/gpio
JohnDoe_71Rus has quit []
<libv> freemangordon: have you fully isolated the rndis config?
<libv> as in, does that pull in dependencies?
<freemangordon> I think yes
<freemangordon> but, I can try again
<libv> you can diff the resulting .config
<freemangordon> that's what I did
<libv> ok
<freemangordon> I have selected few more options though, but I think they are not related
<freemangordon> lemme build it again
<libv> make absolutely sure of both ends here
<freemangordon> with just OTG and RNDIS only enabled
<libv> make sure that it is rndis, and eth_rndis alone that does it
<libv> and make sure that the id pin is being handled just the same with and without this
<libv> you do not want to go down a rabbit hole to prove later that these facts were not solid
<freemangordon> libv: bulding defconfig ATM
<libv> freemangordon: try to get to a point where you just toggle the rndis config
<freemangordon> yep, this is defconfig with OTG enabled only
<freemangordon> if it does not wirk, I will flip RNDIS switch and retry
<freemangordon> *work
<freemangordon> does not work, trying with rndis
<freemangordon> don;t know what happened, now it does not boot at all, will have to restart from scratch
grming has joined #linux-sunxi
<libv> :(
<freemangordon> broken fs on uSD
<libv> happens
<freemangordon> right
<freemangordon> but now OTG does not work with the changes it worked before. this drives me mad :(
<libv> did it start working with 4.11?
<freemangordon> yes
<libv> so perhaps it was a register change done by 4.11 which was triggered by rndis?
<libv> so back to bisecting then :(
JohnDoe_71Rus has joined #linux-sunxi
evgeny_boger has quit [Quit: evgeny_boger]
<karlp> pretty sure that rndis option is just "should the gadget use rndis for greater windows compat, or eem/ecm" so it's still likely to be some dependent config. what's different vs montjoie's config they shared earlier? from defconfig on a33?
<freemangordon> sorru, phone call
<freemangordon> libv: register change that survives power cycle? I doubt
<freemangordon> karlp: I still cannot find the single option that makes it work
<freemangordon> will announce it as soon as I have more data
<freemangordon> what's worse, what used to work an hour ago, does not now
evgeny_boger has joined #linux-sunxi
<freemangordon> so I am kind of confused
evgeny_boger has quit []
cnxsoft has quit []
ftg has joined #linux-sunxi
evgeny_boger has joined #linux-sunxi
evgeny_boger has quit []
<freemangordon> oh, ok, with the config I put on pastebin OTG works, now I have something to compare with
evgeny_boger has joined #linux-sunxi
evgeny_boger has quit []
Guest3359 has quit []
bauen1 has quit [Ping timeout: 480 seconds]
evgeny_boger has joined #linux-sunxi
<freemangordon> oh, CONFIG_USB_ETH_RNDIS must be 'y', not 'm'
<freemangordon> lemme see if is is about libcomposite
<freemangordon> libv: this https://pastebin.com/5QmncxM9
<freemangordon> maybe selecting different gadget will make it work too, will try with NULL later on
apritzel has quit [Ping timeout: 480 seconds]
<freemangordon> but it seems without selecting built-in gadget, OTG does not work, at least here
indy has quit [Ping timeout: 480 seconds]
indy has joined #linux-sunxi
Luke-Jr has quit [Ping timeout: 480 seconds]
<freemangordon> ok, now I only need sound working :)
<libv> freemangordon: so what was the solution?
<freemangordon> CONFIG_USB_ETH_RNDIS=y
<freemangordon> didn't try other gadgets, want to have sound working first, if possible
indy has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has quit []
indy has joined #linux-sunxi
Daanct12 has joined #linux-sunxi
Danct12 has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #linux-sunxi
Luke-Jr has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #linux-sunxi
hentai has quit [Ping timeout: 480 seconds]
bauen1 has joined #linux-sunxi
* karlp guesses something like needing a nop phy to be able to enable musb gadget or something,
<karlp> and that all gets pulled in by the pre-baked gadgets.
<karlp> configfs and co wont' fix anything, it's too late?
ftg has quit [Read error: Connection reset by peer]
gamiee has quit [Ping timeout: 480 seconds]
evgeny_boger has quit [Ping timeout: 480 seconds]