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
ungeskriptet_ has quit [Remote host closed the connection]
ungeskriptet has joined #msm8937-mainline
ungeskriptet has quit [Remote host closed the connection]
ungeskriptet has joined #msm8937-mainline
ungeskriptet has quit [Remote host closed the connection]
ungeskriptet has joined #msm8937-mainline
<M0xCAFEBABE[m]> barni2000: where's our device-specific kernel configs for 6.14?
<barni2000[m]> There are now device specific config
<M0xCAFEBABE[m]> where?
<barni2000[m]> s/now/no/
<barni2000[m]> s/now/no/, s/config/configs/
<barni2000[m]> I have put everything in pmos config what are needed for these devices
<M0xCAFEBABE[m]> so... how do you build the kernel for our devices?
<M0xCAFEBABE[m]> barni2000[m]: where's that?
funderscore is now known as f_
<M0xCAFEBABE[m]> thanks
<M0xCAFEBABE[m]> why not using msm8953-mainline's config sturcture?
<barni2000[m]> It can be cleand up a bit ig, it would be great make some fragements in the future
<barni2000[m]> M0xCAFEBABE[m]: Those are also out dated
<M0xCAFEBABE[m]> damn...
<barni2000[m]> At msm8953 mainline i do the same store the config in the package
<barni2000[m]> You can use them as a defconfig
<barni2000[m]> Usually i am copying them to the arch/arm64/configs/<soc group>_defconfig
<barni2000[m]> As defconfig they can be easily updated through kernel versions
<barni2000[m]> s/through/by/
<barni2000[m]> Idk how to auto update the fragements
<M0xCAFEBABE[m]> barni2000[m]: why would you have to update your fragments for upstream changes?
<barni2000[m]> Because upstream changing a lot
<barni2000[m]> Check the kernel update commits
<M0xCAFEBABE[m]> then only just change whatever that's needed to change on your fragments?
<barni2000[m]> msm8953.config like an full config
<barni2000[m]> Not a real fragement
<barni2000[m]> Btw this way is easier to maintain to me
<barni2000[m]> I only need to rerun `make msm89x7_defconfig` for update it
<barni2000[m]> after a rebase
<M0xCAFEBABE[m]> barni2000[m]: after pulling from upstream, your config fragments stops working or what...?
<barni2000[m]> Maybe all could be based on arm_defconfig and fragments
<barni2000[m]> M0xCAFEBABE[m]: They are just not updated automatically
<barni2000[m]> Also i am using make menuconfig for enable new stuff for avoid missing dependencies
<barni2000[m]> And it will not update the fragments
<M0xCAFEBABE[m]> barni2000[m]: savedefconfig, diff, and apply what's changed to the fragments manually?
<barni2000[m]> That is my problem, current workflow have less manual work than fragment workflow
<M0xCAFEBABE[m]> okay...
<M0xCAFEBABE[m]> btw
<M0xCAFEBABE[m]> how's the USB support on xiaomi mi89x7 devices right now?
<M0xCAFEBABE[m]> * USB support status on xiaomi
<M0xCAFEBABE[m]> does USB accessory mode work on all devices?
<barni2000[m]> Otg works on msm8937 and msm8940 devices except ugg
<barni2000[m]> On 8917 only motorola nora have otg
<M0xCAFEBABE[m]> what about USB accessory/gadget mode?
<barni2000[m]> It is working pmos using it
<M0xCAFEBABE[m]> on all the devices?
<barni2000[m]> Yes
<M0xCAFEBABE[m]> cool
<barni2000[m]> Are you working on mainline android?
<M0xCAFEBABE[m]> yep
<M0xCAFEBABE[m]> for unknown reason it doesn't boot on santoni, it just instantly reboots, boots on ugg and ugglite
<M0xCAFEBABE[m]> also, goodix ts doesn't work on my ugg, goodix ts works on my ugglite
<Danct12[m]> <M0xCAFEBABE[m]> "for unknown reason it doesn't..." <- time for ramoops
<barni2000[m]> <M0xCAFEBABE[m]> "also, goodix ts doesn't work..." <- only edt enabled on ugg
<barni2000[m]> I have focaltech variant
<M0xCAFEBABE[m]> uhh... you can blindly add goodix from ugglite
<barni2000[m]> Santoni should boot, maybe firmware issue
<M0xCAFEBABE[m]> barni2000[m]: wut, i'm on latest firmware
<M0xCAFEBABE[m]> it rebooted in less than 1 second
<barni2000[m]> I will check my santonis but they were fine last time
<barni2000[m]> Both of my santonis are booting
<barni2000[m]> MSM8940.LA.2.0-00550-STD.PROD-1.160464.1.184703
<M0xCAFEBABE[m]> barni2000[m]: uhh weird, could you try booting my image?
<barni2000[m]> yes
<barni2000[m]> or it depends i don't want to flash the userdata if it not necessary
<M0xCAFEBABE[m]> it won't touch any partition, just a recovery image
<barni2000[m]> ok
<M0xCAFEBABE[m]> <barni2000[m]> "ok" <- (doesn't include lk2nd, supposed to be booted from lk2nd)
<M0xCAFEBABE[m]> it uses your pmos defconfig as base, and all xiaomi mi89x7 dts are included
<M0xCAFEBABE[m]> * (includes lk2nd)
<barni2000[m]> not crashed yet
<barni2000[m]> ig
<barni2000[m]> does it contains the panel drivers?
<M0xCAFEBABE[m]> btw, goodix ts keys on ugglite have incorrect mapping
<M0xCAFEBABE[m]> barni2000[m]: yes, all are there
<barni2000[m]> M0xCAFEBABE[m]: i have not mapped them
<barni2000[m]> i have fake goodix touch in all my ugglite
<barni2000[m]> none of them works correctly
<M0xCAFEBABE[m]> barni2000[m]: oh... what
<barni2000[m]> replacemente displays
<barni2000[m]> * replacement displays
<M0xCAFEBABE[m]> oh
<barni2000[m]> all ugglite replacement display a sh**
<barni2000[m]> 1st redmi 7a replacement display also was a mess
<barni2000[m]> lets check this image with land
<barni2000[m]> none of them boots correctly
<M0xCAFEBABE[m]> what about on ugg and ugglite?
<M0xCAFEBABE[m]> booted fine here
<barni2000[m]> a minute
<barni2000[m]> yes ugglite is booting
<barni2000[m]> btw 6.14 branch already there
<M0xCAFEBABE[m]> yeah i've used msm89x7/6.14
<barni2000[m]> oh wait, ig ugglite was boot from flash
<M0xCAFEBABE[m]> after it's booted, the UI title should show LineageOS 22.2 mi89x7
<M0xCAFEBABE[m]> * 22.2 mi89x7 20250409
<barni2000[m]> fastboot boot is not working with your image
<barni2000[m]> from lk2nd
<barni2000[m]> do you have custom lk2nd?
<M0xCAFEBABE[m]> nope, downloaded latest prebuilt from msm8916-mainline
<M0xCAFEBABE[m]> btw, are our GPUs supported in upstream Mesa now?
<barni2000[m]> yes
<M0xCAFEBABE[m]> nice, since when tho?
<M0xCAFEBABE[m]> which mesa version
<barni2000[m]> a moment
<M0xCAFEBABE[m]> cool
<barni2000[m]> both of them
<barni2000[m]> how do you use this recovery.img i wonder why it is not working for me
<M0xCAFEBABE[m]> fastboot boot recovery.img seems to work, at least you can flash it to real boot partition and reboot
<barni2000[m]> its freezing on ugg
<barni2000[m]> it seems like like and lk2nd img every fastboot boot returns back to lk2nd
<barni2000[m]> i can check with prebuilt lk2nd
<barni2000[m]> no luck
<M0xCAFEBABE[m]> weird...
<barni2000[m]> i have realized ugg have different accel like others
<M0xCAFEBABE[m]> on ugglite the whole device restarts when the CPU load stays mid~high for like a minute...
<M0xCAFEBABE[m]> even if, I plug in AC charger
<M0xCAFEBABE[m]> * whole device freezes or restarts when
<barni2000[m]> strange
<barni2000[m]> gpu could cause these kind of issues mesa pretty broken on a3xx
<barni2000[m]> check dmesg -w during high load
<M0xCAFEBABE[m]> oh... i see, i'll try vkms display
f_ is now known as Guest13267
Guest13267 is now known as f_
<M0xCAFEBABE[m]> it still occurs even if I don't load msm module...
<M0xCAFEBABE[m]> and, occurs on ugg too
<M0xCAFEBABE[m]> finally booted into UI after making multiple CPU cores offline and setting powersave governor and limiting cpufreq
<jojo_autoboy[m]> hm doesn't go past lk2nd on my land ether
<barni2000[m]> my land still works
<barni2000[m]> <jojo_autoboy[m]> "hm doesn't go past lk2nd on my..." <- have you installed it from pmos repo or build it yourself?
<jojo_autoboy[m]> no i mean that recovery test image
<barni2000[m]> oh ok
<barni2000[m]> yes it was not works for me either
f_ has quit [Killed (NickServ (This nickname is registered and protected))]
f_ has joined #msm8937-mainline
<barni2000[m]> i wonder what is the difference why none of my device have issues with cpufreq
<barni2000[m]> <M0xCAFEBABE[m]> "finally booted into UI after..." <- you could try limit frequencies in a53pll opps and cpu opps
<M0xCAFEBABE[m]> <barni2000[m]> "i wonder what is the difference..." <- did you try stressing up the CPU for like a minute?
<M0xCAFEBABE[m]> ... when android is booting, it's like that
<M0xCAFEBABE[m]> * ... when android is booting, it behaves like that
<barni2000[m]> i have played with STK
<barni2000[m]> that should be enough strees
<barni2000[m]> s/strees/stress/
<barni2000[m]> but i could try unix bench also
<M0xCAFEBABE[m]> what's that, does it stress up the GPU or the CPU...
<barni2000[m]> unixbench is a benchmarking tool for cpu
<barni2000[m]> glmark2 is for gpu
<M0xCAFEBABE[m]> and, for what period of time?
<barni2000[m]> it runs alogrithms for testing theirs speed
<barni2000[m]> with different size of data sets
<M0xCAFEBABE[m]> btw, what are the mesa workaround environment variables?
<M0xCAFEBABE[m]> * environment variables that we need?
<barni2000[m]> non
<barni2000[m]> * none
<M0xCAFEBABE[m]> what about in before?
<barni2000[m]> maybe ddraw
<M0xCAFEBABE[m]> mesa still refuses to render anything even after I tried multiple hacks that I've used in before :(
<M0xCAFEBABE[m]> * I've used on msm8953 in before
<barni2000[m]> i can made a highload with video playing :D
<barni2000[m]> there is no hwdec yet
<barni2000[m]> all cores near 80-100%
<barni2000[m]> there is no issue on moto hannah
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "mesa still refuses to render..." <- OHHHH it turns out I have to take adreno firmwares from downstream lol
<barni2000[m]> it should not cause problem
<barni2000[m]> 4 minutes of high load, no issues
<M0xCAFEBABE[m]> and the default cpu freq governor here is schedutil
<M0xCAFEBABE[m]> same for you too?
<M0xCAFEBABE[m]> <barni2000[m]> "it should not cause problem" <- downstream one at least gives boot animation (with some gpu errors too), mainline one doesn't
<barni2000[m]> <M0xCAFEBABE[m]> "and the default cpu freq..." <- let me check ig no
<barni2000[m]> /sys/kernel/debug/clk # cat /sys/devices/system/cpu/cpufreq/policy0/scaling_governor
<barni2000[m]> schedutil
<barni2000[m]> /sys/kernel/debug/clk # cat /sys/devices/system/cpu/cpufreq/policy4/scaling_governor
<barni2000[m]> schedutil
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "finally booted into UI after..." <- limiting to only 2 online CPUs on msm8917 still doesn't prevent it from restarting, only works on msm8940...
<barni2000[m]> msm8917 like msm8937 only with big cluster
<barni2000[m]> should not be too much difference
<M0xCAFEBABE[m]> imagine, the pain of, CPU rendering, with only 2 online CPU cores at 960MHz...
<M0xCAFEBABE[m]> <M0xCAFEBABE[m]> "imagine, the pain of, CPU..." <- after stressing it up that much, it even randomly crashes while booting lol...