ChanServ changed the topic of #msm8937-mainline to: Boot Linux on your MSM8917/37/40 and QM215 mobile! | GitHub: https://github.com/msm89x7-mainline | Logs: https://oftc.irclog.whitequark.org/msm8937-mainline | Bridged to #msm8937-mainline:kde.org on Matrix
<barni2000[m]> but how redmi2 works ootb
<barni2000[m]> maybe i should check the fw package
<NekoCWD[m]> i have no wwan0qmi*
<barni2000[m]> btw redmi 2 also have pop noise
<barni2000[m]> maybe something missing from dts
<barni2000[m]> unfortunately i have zero experience about enabling modem from scratch
<barni2000[m]> maybe worth to check 8916 dts
<barni2000[m]> what makes wwan0qmi interface?
<NekoCWD[m]> Idk, i have no 8916 device
<barni2000[m]> drivers/net/usb/qmi_wwan.c
<barni2000[m]> s//~`/, s//`~/
<barni2000[m]> s//<del>`/, s//`</del>/
<barni2000[m]> oh wait wwan0at0
<barni2000[m]> how about renaming the interface
<barni2000[m]> is it possible?
<NekoCWD[m]> rename to what?
<barni2000[m]> wwan0qmi
<barni2000[m]> * wwan0qmi0
<NekoCWD[m]> i think at/qmi represents protocol which interface use
<barni2000[m]> can you show your wip commit i should try it on my devices
<NekoCWD[m]> btw, mpss_mem size is really important
<NekoCWD[m]> ```qcom,firmware-name = "modem";``` in downstream
<NekoCWD[m]> maybe just wrong firmware?
<NekoCWD[m]> barni2000
<barni2000[m]> is not it same?
<NekoCWD[m]> No, at least mba.mbn is squashed
<NekoCWD[m]> Oh, that's just diffrent firmwares. Driver use both
<NekoCWD[m]> I'm gonna sleep now. Good night 🐱💤
<barni2000[m]> gn
<barni2000[m]> i have feeling modem manager is not supports it
<NekoCWD[m]> I was trying to set operation mode through qmicli, but "Device not ready". Maybe something wrong in my dt
<barni2000[m]> i can make interfaces with qmicli
<barni2000[m]> i still think mm is not supports it
<barni2000[m]> maybe travmurav or minecrell[m] know more about modems, idk if it is working on 8908.
<barni2000[m]> lets make a stable branch for 6.13 i will skip venus changes for a while, audio is cool improvement for next release
<NekoCWD[m]> I think that's main issue
<NekoCWD[m]> Can someone disable modemmanager and run this commands on 8916?
<barni2000[m]> 1 moment
<barni2000[m]> the difference between 8917 and 8916, 8916 makes the interface without manual steps
<barni2000[m]> mm is disabled
<NekoCWD[m]> what module creates that wwans?
<barni2000[m]> yes
<barni2000[m]> xiaomi-wt88047:~$ sudo qmicli -d qrtr://0 --dms-get-operating-mode
<barni2000[m]> error: couldn't create client for the 'dms' service: QMI protocol error (3): 'Internal'
<NekoCWD[m]> Will try to bring up modem on prada, maybe this issue caused by my bootloader unlocking hack. I haven't tested it android on hacky unlocked device
<barni2000[m]> no, i have same issues on riva
<NekoCWD[m]> Any other soc with working modem on mdm9607 pil?
<barni2000[m]> i should try after a reboot
<barni2000[m]> becaue i have made the interfaces
<barni2000[m]> mm still not see it
<NekoCWD[m]> Can you send commands that makes you modem "ready"?
<barni2000[m]> set operation-mode works for me ootb
<barni2000[m]> s/operation/operating/
<NekoCWD[m]> ```qmicli -d qrtr://0 --dpm-open-port="hw-data-ep-type=bam-dmux,hw-data-ep-iface-number=0,hw-data-rx-id=42,hw-data-tx-id=42"```?
<barni2000[m]> it is not needed for set operation mode
<barni2000[m]> s/operation/operating/
<NekoCWD[m]> It's need for modemmanager startup
<NekoCWD[m]> Can you push your riva changes?
<barni2000[m]> same as agassi
<barni2000[m]> NekoCWD[m]: i still belive modemmanager is not supporting this modem yet
<NekoCWD[m]> Yea, but i can at least see modem status in it
<barni2000[m]> ?
<barni2000[m]> i don't see it
<barni2000[m]> wait it is working
<NekoCWD[m]> On agassi
<NekoCWD[m]> But i cant power on it
<barni2000[m]> i have forgot to delete eid :'(
<barni2000[m]> and i cannot remove it from irc bridge
<NekoCWD[m]> And from matrix message detail
<barni2000[m]> wait your issue is now the missing sim
<NekoCWD[m]> I have no sim in it now
<barni2000[m]> put it in and check if you can enable the modem with mm
<barni2000[m]> so it seems modem manager should be patched for initializing the interfaces and enable the modem
<barni2000[m]> issue is mm releasing is painfully slow
<NekoCWD[m]> What mmcli arg used to power on modem? `-e`?
<NekoCWD[m]> neko@Jiji ~> mmcli -m any -e
<NekoCWD[m]> error: couldn't enable the modem: 'GDBus.Error:org.freedesktop.libqmi.Error.Protocol.DeviceNotReady: Couldn't set operating mode: QMI protocol error (52): 'DeviceNotReady''
<barni2000[m]> does mm shows the sim?
<barni2000[m]> maybe you need to restart the uim servcie
<NekoCWD[m]> I've restarted whole device
<NekoCWD[m]> Coom
<NekoCWD[m]> *
<NekoCWD[m]> Try ping something or call
<barni2000[m]> i can try gps maybe
<NekoCWD[m]> Lol, on prada works
<NekoCWD[m]> Oh, or not. It can't see ports
<barni2000[m]> call is working
<NekoCWD[m]> Which command do i need to add ports on 8937?
<NekoCWD[m]> barni2000[m]: Cool
<barni2000[m]> yes
<barni2000[m]> it seems VoiceMMode is not working on riva
<barni2000[m]> so we will need change the dais
<NekoCWD[m]> barni2000[m]: Maybe just worng voicemmode mix?
<barni2000[m]> i don think
<barni2000[m]> data also works
<NekoCWD[m]> I think now we need to make openrc service to automatically open port
<barni2000[m]> we should place the script to common place for openrc and systemd
<barni2000[m]> but first lets make ucm configs
<NekoCWD[m]> I want to bring up modem at least on one of my devies
<NekoCWD[m]> I think errors on agassi because of firmware
<barni2000[m]> xiaomi-land have the oldest fw i can check
<NekoCWD[m]> barni2000 can you dump fw from riva?
<barni2000[m]> it will not work with agassi
<NekoCWD[m]> Maybe with prada
<barni2000[m]> no
<barni2000[m]> modem fws containing sim detection parts etc i have tried change them on msm8953 and they are broken in many way
<barni2000[m]> don't change it
<NekoCWD[m]> Hmm, i was trying to find agassi camera specs, but found qc bsp archive :D
<NekoCWD[m]> I will try to flash oldest prada firmware (idk what firmware it has now, it was bricked when i bought it :D)
<barni2000[m]> i cannot make interface on land
<NekoCWD[m]> barni2000[m]: No `wwan0` in `/sys/class/net` ?
<barni2000[m]> no
<barni2000[m]> Val Packett: if you want please send your PR for nora
<NekoCWD[m]> Now i have 8916 modem in my collection
<ValPackett[m]> <barni2000[m]> "Val Packett: if you want..." <- sure! what about vibration / led drivers btw?
<barni2000[m]> lets enable stuff step by step
<barni2000[m]> we have about 90 commits atop of 8916 tree
<barni2000[m]> if we enabling too much stuff without upstreaming the base i will have too much too rebase next time
<barni2000[m]> but you can open your PRs and i will review it them
<barni2000[m]> also we should test the new features first
<barni2000[m]> i am fine with haptic btw because i have applied it on msm8953 tree
minecrell[m] has joined #msm8937-mainline
<minecrell[m]> <barni2000[m]> "maybe travmurav or minecrell[m..." <- You mean 8909? Modem works fine there, but you need tqftpserv and msm8909-modem-wwan-port.initd in addition to rmtfs etc on 8916. see soc-qcom-msm8909 in pmOS
<ValPackett[m]> is there any page tracking the upstreaming btw?
<barni2000[m]> patchwork.kernel.org
<barni2000[m]> lore.kernel.org
<ValPackett[m]> btw re: modem, when i've tried to enable it there was some three letter thing timing out after the firmware got loaded
<ValPackett[m]> barni2000[m]: i mean not in general lol but like specifically links to lore/etc per qcom tree like from the wiki
<barni2000[m]> it is clear from our git tree atm 8917 is upstreamed in basic way
<barni2000[m]> 8937 is not yet
<barni2000[m]> i have not maintaining a status page about it
<barni2000[m]> you can make one if you want
<barni2000[m]> outdated
<ValPackett[m]> yeah, definitely lol, that's the general problem with wiki stuff i guess
<barni2000[m]> every change against "Add initial support for msm8917" are not upstreamed
<ValPackett[m]> changed https://github.com/msm89x7-mainline/linux-panel-drivers/pull/5 to not do the IC name thing
<barni2000[m]> btw in our three we have charging driver what is managed on other trees so we will not upstream it, some dsi fix, modem patch, small changes in pm8937 what can be upstreamed easily
<barni2000[m]> audio patches what should be refined
<barni2000[m]> there is not too much commit yet
<barni2000[m]> these are atop of 8916 tree
<ValPackett[m]> yea it seems that with qcom there is already a bit of a preference for like 50-200 commit long downstreams to ship something usable in pmos already xD which is why i thought adding a couple more drivers wouldn't hurt
<barni2000[m]> if i remove device commits from this list there will be much less what is matter
<ValPackett[m]> oh btw what was that cpuidle situation again? i remember you've said that it "will be fixed"?
<barni2000[m]> i have still not fixed it but it is better a bit on 6.13
<barni2000[m]> s/``/```/
<barni2000[m]> as i see there is not too much stuff to upstream
<barni2000[m]> i am not good at updating lists like for sdm660
<ValPackett[m]> why do i get a bunch of changes in other panels when regenerating the drivers?
<ValPackett[m]> am i on the wrong branch for panels 0.o
<ValPackett[m]> no i mean like totally weird changes like get_modes being changed from drm_connector_helper_get_modes_fixed to inlined actual 15ish lines of code
<ValPackett[m]> lots of bool prepared
<ValPackett[m]> `tm_otm1901a_supplies` → `ctx->supplies`
<NekoCWD[m]> generator update?
<ValPackett[m]> ohhhhh right it's a submodule lol
<ValPackett[m]> idk why it's so out of date in the actual commit in the panels repo ;)
<barni2000[m]> oh i have forgot maybe motos will need --dumb-dcs but let wait until users are reporting it
<ValPackett[m]> one of the panels has the warning about that
<barni2000[m]> ok if someone will report it is not working or very dimmed then you will need that param
<barni2000[m]> potter and cedric had those issues
<NekoCWD[m]> Agassi has swapped enter/exit sleep mode command :D
<ValPackett[m]> amazing
<ValPackett[m]> <NekoCWD[m]> "https://github.com/NekoCWD/..."; <- oh i guess the regulators are what i was missing for modem
<NekoCWD[m]> Or mpss size
<ValPackett[m]> that i copied from downstream
<NekoCWD[m]> 0x4800000 ?
<ValPackett[m]> 0x5000000 here actually
<ValPackett[m]> btw. will the vibration driver get upstreamed from 8953? i'd like to lower the default voltage to prevent motor damage like what happened to my moto, should i submit a blind PR to that tree?
<barni2000[m]> ValPackett[m]: idk this question should be go for sdm845 guys
<NekoCWD[m]> barni2000modem works fine on riva?
<barni2000[m]> yes
<barni2000[m]> it not works on land
<barni2000[m]> maybe it needs something more
<NekoCWD[m]> On prada also doesn't work. Both on newest and oldest available firmware
<ValPackett[m]> barni2000[m]: hm. the dts max voltage patch only landed in 8953 afaik? not sure if 845 picked that up..
<barni2000[m]> yes 2 commit difference between the 2 driver
<barni2000[m]> but sdm845 variant is the original
<ValPackett[m]> oop, for msm8920 we don't even have a dtsi? would that just reuse msm8917 entirely?
<barni2000[m]> do you have 8920?
<ValPackett[m]> no
<barni2000[m]> we don't have 8920.dtsi yet because i have not imported ipa2-lite from msm8953
<ValPackett[m]> would it make sense to have a 89xx-nora dtsi and a 8917+nora dts using it already?
<barni2000[m]> they are different in more
<barni2000[m]> 8920 nora have different audio architecture
<barni2000[m]> we should not support a device until we don't have one
<barni2000[m]> btw i am thinking about a base based on msm8916-pm8916.dtsi
<barni2000[m]> maybe pm9837-common.dtsi would work or we should do the same for all cpus
<barni2000[m]> msm8917-pm8937 msm8937-pm8937 msm8940-pm8937 etc.
<ValPackett[m]> hmmm with the voltages modem is no longer "PBL boot timed out" but now "dog.c:1522:Watchdog detects stalled initialization"
<barni2000[m]> moto needs tqftpserv
<barni2000[m]> and maybe a service for rfs
<ValPackett[m]> mm that's all when starting rmtfs yeah but i haven't looked into tqftpserv
<barni2000[m]> xiaomis are the simplest devices, every other vendor needs more
<barni2000[m]> motos need rmtfs tqftpserv+this service
<barni2000[m]> sony's needs ta service as an extra
<ValPackett[m]> ayy, thanks! port wwan0at0 attached port wwan0at1 attached
<ValPackett[m]> <barni2000[m]> "https://gitlab.postmarketos.org..." <- hm. that didn't help. that qmi command says "Successfully opened the port" but MM still says "Failed to find a net port in the QMI modem"
<barni2000[m]> you need to restart mm
<barni2000[m]> and wait
<barni2000[m]> check ip a for interfaces
<ValPackett[m]> that was after restarting, i'm running mm from the CLI with --debug
<barni2000[m]> do you have wwan0 interface?
<ValPackett[m]> no
<ValPackett[m]> mm says it's grabbing all the ports and immediately disposing them
<ValPackett[m]> because none of them is a "net" port 0.o
<ValPackett[m]> you'd think the wwan ones would count
<barni2000[m]> it was take while untile mm recognized ports on tiva
<barni2000[m]> s/tiva/riva/
<barni2000[m]> s/untile/until/, s/tiva/riva/
<barni2000[m]> but if you don't have the interface in ip a it will not work
<ValPackett[m]> oh the wwan is MM_PORT_TYPE_AT. looks like we do indeed need a dpm command
<ValPackett[m]> i did get an interesting dmesg message when doing all this: "4080000.remoteproc:bam-dmux: Failed to request RX DMA channel: -EPROBE_DEFER"
<barni2000[m]> <barni2000[m]> "xiaomi-riva:~$ sudo qmicli -d..." <- this
<barni2000[m]> 8909 does not need setting operating mode
<ValPackett[m]> oh - hm - i didn't enable `&bam_dmux_dma` in the dts
<ValPackett[m]> setting online says QMI protocol error (52): 'DeviceNotReady'.. probably due to the above heh
<ValPackett[m]> yes, lol, we need that node okayed, then dpm-open-port and mm sees it
<barni2000[m]> we need to find out how to handle the regulators, i don't like always-on
<ValPackett[m]> pm8937_s1 is mss-supply ("only valid for qcom,msm8974-mss-pil" but turns out also for our mdm9607 too?)
<ValPackett[m]> s3 in downstream literally does just have a qcom,init-voltage
<ValPackett[m]> s4 is actually used by wcd_codec downstream
<barni2000[m]> switches should be placed in HiFi.conf atm but maybe they are not needed in the ucm config
<barni2000[m]> or i will make them conditional in the future and place them in the building blocks
<ValPackett[m]> oh snap i have a carrier phone right i think the vendor even told me
<ValPackett[m]> it's trying to load /readonly/fsg/nora/latam/customer/personal_arg.mbn and tqftpserv doesn't even like the path because the only readonly it likes is /readonly/firmware/image/
<ValPackett[m]> s/vendor/seller/
<barni2000[m]> there is a patch for tqftpserv
<barni2000[m]> i have not remembered correctly
<barni2000[m]> it was just a discussion about it but potter is not needed it
<barni2000[m]> s/is/was/
<barni2000[m]> it was a german variant of Moto G5 Pro but i don't remember if it was carrier locked
<barni2000[m]> remove the panel commit