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]> land so sensitive
<barni2000[m]> land is booting
<barni2000[m]> i have made very minimal dt for it
<barni2000[m]> i am going to enable features later
<Saalim[m]> Can you send that dt once
<barni2000[m]> only frambuffer is enabled yet
<barni2000[m]> s/frambuffer/framebuffer/
<barni2000[m]> but most of the features will be copy paste from santoni (mostly)
<barni2000[m]> ramoops is not working :(
<barni2000[m]> i have got garbage
<Saalim[m]> You must have gotten something on framebuffer na?
<Saalim[m]> Like a kernel log if thats readable
<barni2000[m]> yes kernel log is on the framebuffer
<barni2000[m]> btw lk2nd writes out the correct framebuffer parameters
<barni2000[m]> the adress is 8dd01000 and not 8dd00000
<Saalim[m]> We can also find the address by hacking abl too
<Saalim[m]> like grepping
<barni2000[m]> no need
<Saalim[m]> barni2000[m]: I can try this on my device within an hour
<barni2000[m]> [110] MDP continuous splash detected: pipe RGB_0, base: 0x8dd01000, stride: 2160, src: 720x1280 (0,0), img: 720x1280, out: 720x1280 (0,0), format: 0x2243f (bpp: 3)
<Saalim[m]> dang nice
<barni2000[m]> montana also booting
<barni2000[m]> these devices are sensitive for framebuffer memory
<hacker420[m]> barni2000 montana booted with simpledrm/simplefb?
<hacker420[m]> <barni2000[m]> "montana also booting" <- on lk2nd or stock bl
<barni2000[m]> both
<hacker420[m]> hm can you send the dts?
<hacker420[m]> what would trip framebuffer memory even
<hacker420[m]> oh
<hacker420[m]> so that'
<hacker420[m]> *'s from montana
<hacker420[m]> ?
<barni2000[m]> yes, montana almost like cedric
<barni2000[m]> 8937 still has some panic i think
<barni2000[m]> maybe it should be limited in cpufreq
<hacker420[m]> montana works with panel or simplefb?
<barni2000[m]> i have only tested with simple drm but panel also should work
<hacker420[m]> well for me with these regions lk2nd reboots and stock stays on splash
<hacker420[m]> hmm
<barni2000[m]> did you applied cpufreq PR for lk2nd?
<hacker420[m]> oh
<hacker420[m]> ah yea
<hacker420[m]> yes I did
<hacker420[m]> currently used dts
<hacker420[m]> hmm
<hacker420[m]> let me toss out the panel stuff
<hacker420[m]> and add simpledrm back in
<hacker420[m]> do you have the nodes for montana? barni2000 \
<barni2000[m]> i have modfied your dt
<barni2000[m]> i have commented out the panel part
<barni2000[m]> and added simplerdm
<hacker420[m]> even if I delete panel it's still unhappy about something
<hacker420[m]> even without the simpledrm node\
<hacker420[m]> either memory, regulators, some ID stuff or GPIO
<hacker420[m]> can you send the montana dts?
<barni2000[m]> enable simpledrm
<barni2000[m]> i have left as disabled
<hacker420[m]> so that boots?
<barni2000[m]> yes
<barni2000[m]> i also see the partitions
<hacker420[m]> I guess I need my own board IDs?
<barni2000[m]> yes maybe but i can check with out board-id
<hacker420[m]> ok
<hacker420[m]> that seems to work
<hacker420[m]> I
<hacker420[m]> I'd need to reinstall rootfs to add the simpledrm module
<hacker420[m]> but it doesn't instantly reboot now
<hacker420[m]> so what now?
<hacker420[m]> even if I uncomment the panel stuff it still doesn't reboot
<hacker420[m]> but panel does not change
<hacker420[m]> from downstream
<hacker420[m]> generated driver
<barni2000[m]> Lab/ibb is not needed because bootloader powers it on
<barni2000[m]> Wled also not needed yet
<barni2000[m]> Let me generate a driver for montana
<barni2000[m]> btw msm module should be in initfs
<hacker420[m]> it is
<hacker420[m]> does the order matter?
<barni2000[m]> no
<barni2000[m]> but it is also rebooting for me
<barni2000[m]> i think it will be something with cpu freq maybe msm8937 needs some more for higher freqs i am going to disable some freqs in opp
<hacker420[m]> barni2000[m]: wait
<hacker420[m]> after some time or right away
<barni2000[m]> right away
<hacker420[m]> weird
<hacker420[m]> it doesn't for me
<barni2000[m]> if i enable frambuffer i see some logs and reboot
<hacker420[m]> it just stays on lk2nd splash
<hacker420[m]> need to see with simpledrm
<barni2000[m]> maybe firmwares are also needed
<hacker420[m]> you didn't run with rootfs?
<barni2000[m]> yes i am just booting the Image appended with dtb
<barni2000[m]> but i can make an boot.img
<barni2000[m]> s/an/a/
<barni2000[m]> i can get some pstore from twrp but it is related to the bootloader and twrp
<hacker420[m]> hm
<hacker420[m]> `status = "disabled";`
<hacker420[m]> do I need to change this to enabled?
<hacker420[m]> or will it auto enabler
<barni2000[m]> where?
<hacker420[m]> in the framebuffer node
<barni2000[m]> yes set to okay
<barni2000[m]> i have just forgot it
<barni2000[m]> s/forgot/forgotten/
<barni2000[m]> panel cause instant crash
<hacker420[m]> hm
<hacker420[m]> it didn't for me
<barni2000[m]> i will check with initfs
<hacker420[m]> hm still nothing on screen
<barni2000[m]> instant crash
<barni2000[m]> time to add firmwares to initfs
<hacker420[m]> wait
<hacker420[m]> are montana and cedric panels the same
<barni2000[m]> almost
<barni2000[m]> dimensions are different cedric is 5" montana is 5.2"
<hacker420[m]> hmmm
<hacker420[m]> shouldn't I also add simpledrm to init modules?
<barni2000[m]> it is not a module for me
<hacker420[m]> weird
<hacker420[m]> is msm a module for you?
<barni2000[m]> no
<barni2000[m]> that could be an issue
<barni2000[m]> ig
<hacker420[m]> I can't enable it as builtin
<hacker420[m]> wait
<hacker420[m]> yeah I can
<hacker420[m]> hang on
<hacker420[m]> will it cause some issue if I leave them in initfs?
<hacker420[m]> or will it just ignore the lack of modules
<barni2000[m]> should not
<barni2000[m]> there are differences between msm8937 and msm8940 gpu
<barni2000[m]> mostly the frequencies
<hacker420[m]> weird
<hacker420[m]> simpledrm still shows nothing
<hacker420[m]> the address is correct
<barni2000[m]> it is working now
<hacker420[m]> that's with panel?
<barni2000[m]> yes
<barni2000[m]> i have compiled them as a module and put them in initfs
<barni2000[m]> here is some error message for proof... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/VeuQdtbXaozRrtwhJWivaRtA>)
<barni2000[m]> i know how to fix this :D
<barni2000[m]> i can try without firmwares
<barni2000[m]> msmdrmfb should work without it
<barni2000[m]> but there are some SoCs where it is required
<barni2000[m]> do you have this qcom,dsi-phy-regulator-ldo-mode;?
<hacker420[m]> barni2000[m]: in downstream?
<barni2000[m]> in your dt
<hacker420[m]> yes
<barni2000[m]> you only need it if you have this downstream also
<barni2000[m]> so remove it
<hacker420[m]> I wonder why simpledrm won't work
<hacker420[m]> or I could go straight to panel
<barni2000[m]> i am checking it that
<hacker420[m]> so panel works for you?
<barni2000[m]> yes
<hacker420[m]> let's try that then
<barni2000[m]> yes it is works without simpledrm
<barni2000[m]> also panel on/off is working
<hacker420[m]> let me put drm as builtin
<barni2000[m]> no
<barni2000[m]> that caused me crash
<barni2000[m]> both should be module
<hacker420[m]> along with panel?
<barni2000[m]> yes panel and msm
<barni2000[m]> let me check panel without firmwares
<barni2000[m]> works without it
<barni2000[m]> [ 3.524911] msm_mdp 1a01000.display-controller: Direct firmware load for qcom/a530_pm4.fw failed with error -2
<barni2000[m]> [ 3.716693] [drm:mdp5_irq_error_handler [msm]] *ERROR* errors: 04000000
<barni2000[m]> [ 3.524947] msm_mdp 1a01000.display-controller: [drm:adreno_request_fw [msm]] *ERROR* failed to load a530_pm4.fw
<hacker420[m]> what do you have in initfs?
<barni2000[m]> msm panel
<hacker420[m]> just msm and panel?
<barni2000[m]> and the default modules for pmos
<hacker420[m]> hm?
<hacker420[m]> can you send yours?
<barni2000[m]> initfs?
<hacker420[m]> yea
<barni2000[m]> a minute
* barni2000[m] posted a file: initramfs-extra (1719KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/clxWCHhMXtATlaNbTqPQyLDp >
* barni2000[m] posted a file: initramfs (7358KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/GROjiHpXOQzJZKHIBrQMGYmE >
<hacker420[m]> nah
<hacker420[m]> I want modules-initfs
<hacker420[m]> I have initramfs
* barni2000[m] posted a file: 00-device-xiaomi-santoni.modules (0KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/XurxurwZXoxMaUkeJJxKWqxR >
* barni2000[m] posted a file: 00-default.modules (0KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/svCppRfJMERtRWHobtjFxTFR >
<hacker420[m]> that's in device package?
<hacker420[m]> I have just this
<hacker420[m]> or does pmos auto add the default ones in rootfs
<barni2000[m]> yes
<barni2000[m]> i have copied those from chroot
<hacker420[m]> k
<barni2000[m]> i don't have device package yet i have just modified the config inside santoni's chroot
<hacker420[m]> nope still nothing
<hacker420[m]> at this point I wonder if my panel driver is broken
<hacker420[m]> can you send the dts for montana with panel?
<hacker420[m]> how did you generate the panel driver?
<barni2000[m]> generate.sh
<barni2000[m]> * generate.sh
<hacker420[m]> wait what
<hacker420[m]> with the panel driver generator?
<hacker420[m]> because I can't see that generate.sh thing anywhere
<barni2000[m]> put your dtb in dtb dir and make a config in the config dir
<hacker420[m]> which dtb
<hacker420[m]> there were multiple
<hacker420[m]> in downstream?
<barni2000[m]> you need downstream dtb because mainline does not contains the panel definitions
<hacker420[m]> yes
<hacker420[m]> and the dtb gets made when I compile it or what
<barni2000[m]> everything stored in the .c file
<barni2000[m]> downstream contains the dsi commands in the dt mainline contains it in the driver
<hacker420[m]> no what I don't get is where do I get the dtb from
<hacker420[m]> do I extract it from dt.img or what
<barni2000[m]> from twrp adb pull /sys/firmware/fdt
<hacker420[m]> ah
<hacker420[m]> now that is helpful
<hacker420[m]> so I just rename it and plop it in there?
<barni2000[m]> yes
<barni2000[m]> it can be also extracted with binwalk from boot.img
<barni2000[m]> or dd
<hacker420[m]> like this?
<hacker420[m]> mdss_mdp.panel=1:dsi:0:qcom,mdss_dsi_mot_tianma_497_1080p_video_v0
<hacker420[m]> just asking what specifically to put in there
<barni2000[m]> generate.sh will write out its name
<barni2000[m]> * ```generate.sh, * generate.sh``` will
<barni2000[m]> remove mot_
<barni2000[m]> motorola,cedric-panel-tianma
<hacker420[m]> so run it from the mainline dir
<barni2000[m]> yes
<barni2000[m]> it will create a squash commit
<hacker420[m]> wtf
<hacker420[m]> Fatal Python error: none_dealloc: deallocating None: bug likely caused by a refcount error in a C extension
<barni2000[m]> did you clone msm89x7 panel-drivers repo?
<barni2000[m]> s/clone/cloned/
<hacker420[m]> yes?
<hacker420[m]> wait
<hacker420[m]> shit
<hacker420[m]> change the readme
<barni2000[m]> ok i will update it later
<hacker420[m]> same crap
<barni2000[m]> fix your system
<hacker420[m]> go blame suse
<barni2000[m]> you can try pyenv
<hacker420[m]> or maybe it's a python version thing
<barni2000[m]> ❯ python -V
<barni2000[m]> Python 3.12.4
<hacker420[m]> I have 3.11.g
<hacker420[m]> I have 3.11.9
<barni2000[m]> it should work
<hacker420[m]> unless it tries to use python2 or smth
<barni2000[m]> but maybe suse has some custom annoying patch on it
<barni2000[m]> i don't have python2
<barni2000[m]> with pyenv you can install more python version and change them for a session
<hacker420[m]> the better question is what does it try to run py files with
<hacker420[m]> hm
<hacker420[m]> can I do the changes manually
<hacker420[m]> as in run the generator manually
<hacker420[m]> I wil send the dtb to a friend
<barni2000[m]> open an MR and i will generate it for you
<barni2000[m]> but if the generator works manual then generator.sh also should work
<barni2000[m]> * but if the generator works manual than generator.sh also should work
<barni2000[m]> cpu freqs are too high for msm8937
<barni2000[m]> maybe missing saw2 driver cause it
<barni2000[m]> 1.3 GHz should be disabled for msm8937 otherwise it is working fine
<hacker420[m]> still no panel here...
<hacker420[m]> does it initialize instantly?
<hacker420[m]> maybe I could try stock bl
<hacker420[m]> but ehhh
<hacker420[m]> barni2000 mind sending the montana dts?
<hacker420[m]> with panel
<barni2000[m]> this will needed for 8937
<barni2000[m]> it should be tested, what happens if both cluster uses the highest clock
<barni2000[m]> * the highest enabled clock
<hacker420[m]> <barni2000[m]> "msm8937-motorola-montana.dts" <- yeah still no panel
<hacker420[m]> what is wrong here
<barni2000[m]> apply the opp changes
<hacker420[m]> what will that change
<barni2000[m]> it will fix the crash
<hacker420[m]> what crash
<barni2000[m]> with 1.3 GHz it is immediately crashing with 1.24 GHz it will reboot later
<barni2000[m]> both are working fine on msm8940
<hacker420[m]> it doesn't even get to the point of crashing for me
<barni2000[m]> if it freezing somewhere it means crashing
<hacker420[m]> just reappend the dtb afterwards?
<barni2000[m]> yep
<barni2000[m]> so top 3 freq will missing 1248000000 1344000000 1401000000, last is because pll driver does not support fractional mode
<hacker420[m]> nope still nothing
<barni2000[m]> these are corner0 freqs with cpr we can reach higher freqs
<hacker420[m]> here maybe you can spot the issue
<barni2000[m]> tianma,497-v0?
<hacker420[m]> yeah I didn't change that compatible
<hacker420[m]> does it really matter if both are the same
<barni2000[m]> two driver cannot have same compatible
<hacker420[m]> wdym by 2 drivers
<barni2000[m]> did you regenerated your panel driver?
<hacker420[m]> yes
<hacker420[m]> it's a newly made one from that dtb I pulled from the phone
<barni2000[m]> let me generate you a driver
<barni2000[m]> send me your dt
<barni2000[m]> s/commit/commits/
<barni2000[m]> "motorola,cedric-497-tianma"
<hacker420[m]> tomorrow
<barni2000[m]> ok