ChanServ changed the topic of #linux-msm to:
lumag_ has quit [Remote host closed the connection]
lumag_ has joined #linux-msm
lumag_ has quit [Ping timeout: 480 seconds]
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]
<aka_[m]> konradybcio: rc5 gives some output atleast
<aka_[m]> ICC related tho
<aka_[m]> WARNING: CPU: 2 PID: 73 at drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c:249 dpu_debugfs_create_regset32+0x14/0xb
<aka_[m]> [drm:dpu_core_perf_crtc_check:197] [dpu error]no bandwidth limits specified
<aka_[m]> [drm:dpu_crtc_atomic_check:1235] [dpu error]crtc45 failed performance check -7
<aka_[m]> seems like lacking dpu data
<aka_[m]> weird as i reuse qcm one
jhovold has joined #linux-msm
<aka_[m]> threshold = kms->catalog->perf.max_bw_high;
<aka_[m]> [ 2.916128] [drm:dpu_core_perf_crtc_check] calculated bandwidth=0k
<aka_[m]> [ 2.916151] [drm:dpu_core_perf_crtc_check] final threshold bw limit = 0
<aka_[m]> [ 2.916172] [drm:dpu_core_perf_crtc_check:197] [dpu error]no bandwidth limits specified
<aka_[m]> that sounds right before min and max is same
<aka_[m]> s/before/because/
<aka_[m]> tho on bengal it should be bit less >_>
<aka_[m]> >_>
<aka_[m]> log if someone can take a look
<aka_[m]> compiler threw my warning when i was far away from pc......
<aka_[m]> s/my/me/
<aka_[m]> not errors just warnings
<konradybcio> you are missing dpu_perf_cfg
<konradybcio> or it's incomplete
<konradybcio> the data for it is in dt
<konradybcio> qcom,sde-max-bw-high-kbps etc
<aka_[m]> /dev/dri # ls
<aka_[m]> card0 renderD128
<aka_[m]> /dev/dri #
<aka_[m]> konradybcio: i had & in structs
<aka_[m]> leftover from other tree
<aka_[m]> i mean i was reffering to &perf_data instead of perf_data
<konradybcio> long live C :P
<aka_[m]> still blank screen and debugfs error
<konradybcio> dmesg?
<konradybcio> >[ 0.451117] l5: unsupportable voltage range: 3300000-3056000uV hmm
<aka_[m]> yes this one was assigned to sdcard
<aka_[m]> but sdcard now works
<konradybcio> #define DSI_ERR_STATE_FIFO 0x0004
<konradybcio> this error name tells me exactly nothing, but I think I had it when my panel setup was not correct..
<aka_[m]> maybe i need to cal to last step if 3056000 fits
<aka_[m]> konradybcio: its simple panel driver
<konradybcio> try flipping the reset gpio?
<aka_[m]> should i switch to drm one?
<konradybcio> hm you could
<aka_[m]> it would be so cool to get it working >_>
<aka_[m]> konradybcio: power for what exactly
<konradybcio> panel
<aka_[m]> i have no regulators defined
<konradybcio> can you check /sys/kernel/debug/clk/dispcc0*/clk_rate
<konradybcio> are your phy addresses correct?
<aka_[m]> so its send by panel driver i assume
<konradybcio> yes
<konradybcio> this message tells you a command send timed out
<konradybcio> yes but bring your own init seq
<aka_[m]> uh it would do same
<aka_[m]> konradybcio: now i don't have this cmd error anymore
<aka_[m]> xiaomi-lemon:/dev/dri# ls
<aka_[m]> by-path card0 renderD128
<konradybcio> kgsl smmu, we can ignore for now i guess
<aka_[m]> yea
<aka_[m]> adreno disabled
<aka_[m]> and it depends on disabled gpucc
<aka_[m]> konradybcio: once iwent for pmos_continue_boot display blinked like 4 or 5 times
<konradybcio> ok, are you on lmdpmpmdpgg .c driver, panel-simple or nt36.. driver?
<aka_[m]> 1
<konradybcio> well yes, size is wrong
<konradybcio> qcom,sde-vbif-size = <0x2008>;
<konradybcio> vs .base = 0, .len = 0x1040,
<konradybcio> and there's no regdma on bengal
<konradybcio> so removethat
<aka_[m]> so new vbif to create?
<konradybcio> y
<aka_[m]> konradybcio: and dma to dump like it is on sc7280?
<konradybcio> yes, just remove it from the cfg_init
<aka_[m]> konradybcio: only base then it appears is different
<aka_[m]> rest is same
<aka_[m]> qos and stuff
<konradybcio> base should be 0?
<konradybcio> did you try all 3 drivers?
<aka_[m]> pll fail to lock
<konradybcio> hmm
<konradybcio> do you have an opp table for dsi?
<aka_[m]> well i have all below NOM xD
<aka_[m]> power-domains = <&rpmpd SM6115_VDDMX>;
<aka_[m]> i kinda not remember where dsi opp table comes from, on older i seen that in gcc
<aka_[m]> right
<aka_[m]> changed all opps to require nom yet i don't see a change
<calebccff> I'm trying to bring up the modem on the Pixel 3 and I'm getting this error:
<calebccff> [ 8.548873] qcom-q6v5-mss 4080000.remoteproc: assigning Q6 access to mpss memory failed: -22
<calebccff> Added some logging: https://p.calebs.dev/035fd4@raw
<calebccff> I'm reasonably sure I have the memory regions defined correctly, google's kernel sources are quite nice
<konradybcio> aka_: try just arbitrarily making the levels higher
<konradybcio> or keep just the highest opp and set its level to NOM
<konradybcio> or TURBO or whatever
<konradybcio> >=NOM
<konradybcio> calebccff: check the region in downstream dmesg to be double sure
<aka_[m]> ok i just give up,too dumb for that.
<calebccff> konradybcio: have done, mpss is definitely at the correct address with the correct size
<calebccff> z3ntu: on fp4 did you have trouble with GPI DMA? on pixel 3 I'm hitting some issues with it, touch works fine with i2c-gpio
<z3ntu> calebccff: yes, even though gpi dma seems to init fine, touchscreen still doesn't work with it, but works fine with i2c-gpio
<calebccff> z3ntu: huh alright, so seems like a bug in GPI_DMA somewhere then
<z3ntu> I guess so
<aka_[m]> konradybcio: dpu error in dmesg fixed
<aka_[m]> and i don't see dsi error anymore
<aka_[m]> however PLL still fails to lock
<aka_[m]> i had wrong sspp assigned in data
jhovold has quit [Ping timeout: 480 seconds]
<konradybcio> Hmm need to check techpack then
<konradybcio> Maybe bengal has some quirk
<konradybcio> Doesnt look like it but maybe i missed sth
<aka_[m]> <aka_[m]> "and i don't see dsi error..." <- i see them again with panel driver >_>
<aka_[m]> seems like i will give up for now and try to get rest working
<aka_[m]> then come back maybe once.
<aka_[m]> <konradybcio> "Doesnt look like it but maybe..." <- with debug logs it says it manage to get pll lock at the end
<konradybcio> Hm now that i think about it, it should have been obvious given it reported a rate
<aka_[m]> yea
<aka_[m]> it was happening on older platforms too with pll clocks
<aka_[m]> still something is wrong
svarbanov has quit [Ping timeout: 480 seconds]
svarbanov has joined #linux-msm
<aka_[m]> konradybcio: ...
svarbanov has quit [Ping timeout: 480 seconds]
svarbanov has joined #linux-msm