ChanServ changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630)
<bamse> steev: computer says no? or there's no sign of why it doesn't work?
derzahl has quit [Ping timeout: 480 seconds]
<steev> no sign of it at all
<steev> msm is loaded, but nothing in dmesg
<steev> and glxinfo says llvmpipe so definitely not using freedreno
<steev> i wonder if i missed a config option
<steev> no, doesn't look like it
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
<robclark> steev: try something simple like kmscube (ie. without DE running) to sanity check that gpu works... if that doesn't work or you end up w/ llvmpipe make sure kernel is able to load firmware
<steev> robclark: drmGetDevices2 failed: No such file or directory
<steev> firmware and such should be loadable, since it's the same kernel i use on the c630
<steev> but let me double check that i didn't break that
<bamse> Do not try and bend the gpu, that's impossible. Instead, only try to realize the truth...There is no GPU...
<steev> yeah, FD630 on the c630
<robclark> steev: check file perms on /dev/dri/* perhaps?
<robclark> bamse: no GPU.. that sounds like a sad reality :-P
<steev> there is no /dev/dri heh
<robclark> oh, well there's your problem :-P
<steev> i'm assuming i'm missing some kinda kernel config option for the sc8180x but i dunno what
<robclark> if -rc kernel (I didn't read whole backlog), do you have bamse's reverts?
<steev> i do indeed
<robclark> oh, 8c[x].. hmm, not sure
<steev> actually i believe the reverts made it in to rc4
<robclark> \o/
Obsdark has joined #aarch64-laptops
Obsdark has left #aarch64-laptops [#aarch64-laptops]
<steev> i'm not doing anything special on the kernel command line, maybe i should throw in some drm debug output to see wtf
<bamse> steev: the two issues described in the panel/bridge revert discussion would apply to edp and dp
<bamse> steev: but i have a feeling that there's something related ot the port@ of the dp nodes which causes it to fail in dp_parse
<steev> that's possible
<steev> so, i was curious what would happen
<steev> i did modprobe -r msm;modprobe msm; and i get https://paste.debian.net/1239376
<bamse> i think you have a memory corruption
<bamse> steev: but i don't see any printouts from msm or dtm in that kernel log...
<steev> same, but it does show the modules as being loaded
<bamse> is your /sys/kernel/debug/devices_deferred empty?
<steev> nope
<steev> that is likely our problem
<steev> and yeah, your 2 reverts are in rc4, so they are in there
<steev> near the "8 days ago"
<bamse> CONFIG_SM_GPUCC_8150?
<bamse> given that 2ca0000.iommu is waiting for a clock controller...and you have &gcc
<steev> no i don't have that
<steev> wtf
<bamse> also check that you have CONFIG_SM_DISPCC_8250 because there was something done to that recently
<steev> hm, that's not enabled either but it *should* be
<steev> CONFIG_SM_DISPCC_8250=y is in my defconfig i used
<bamse> time to get some sleep here, soon tomorrow...
<bamse> ttyl
<steev> so i need DISPCC_8250 but GPUCC *8150* , okay
<steev> thanks, i'll poke, i figured it was a config option but had no idea what, and that pointer to devices_deferred is awesome
<bamse> yeah, the dispcc is the same for 8150, 8250 and 8180...while the gpucc is different between 8150 and 8250
<steev> go sleep, i'm gonna do the same after this test
derzahl has joined #aarch64-laptops
jhovold has joined #aarch64-laptops
derzahl has quit [Quit: auf wiedersehen]
jenneron_ has joined #aarch64-laptops
jenneron has quit [Ping timeout: 480 seconds]
<steev> bamse: while those options were probably needed, i now end up with cpu0 stuck
<bamse> steev: wonder if it relates to the callstack...i thought i disabled bandwidth voting from the cpu again
<bamse> steev: because i kept running into issues like the system crashing occasionally when cpuidle kicked in
<steev> possibly?
<bamse> what kernel are you on?
matthias_bgg has joined #aarch64-laptops
<bamse> steev: what's the oops at the top of the screen? the same?
<bamse> steev: and is that message in the middle saying that it failed to mount the rootfs?
<steev> Pretty sure yeah
<steev> And yes, the cpu gets stuck
<steev> Cpu get stuck and rootfs ends up not mounting
matthias_bgg has quit []
<steev> It’s really weird
<steev> shouldn't i have hit that regardless of SM_GPUCC_8150 being enabled?
<bamse> how did you cook up this branch?
<bamse> steev: one way would be if you undervolt the mmcx power-domain momentarily, then you get something looking like this
<bamse> and that would only happen if you actually probe dispcc
<steev> bamse: i took what i have for c630, applied a few other things from pwcli on top, and then took everything from your take2 branch and applied that
<steev> i need to learn to use annotate like jstultz does/did
<bamse> steev: i thought you where on the flex now?
<steev> the one change i had to make is the distro_defconfig from shawn's kernel had CONFIG_PHY_QCOM_QMP=y but that makes the build fail if typec is M
<steev> bamse: i am
<bamse> steev: looking at my tree i see the interconnect driver is broken upstream, i'll post the fixes for that...
<steev> ah, that could definitely cause issues
<steev> stupid question, if i just disabled interconnect for now... would that stop the bandwidth voting, or would it just not work at all
<bamse> it might work, unless there's something that actually require you to turn on some bus
<steev> one way to find out!
<steev> one other minor annoying thing i've found; there doesn't seem to be any way to get recovery media from lenovo for this thing
<bamse> ouch, that's good to know
<steev> without interconnect, it does boot
<steev> but a bunch of stuff missing
<bamse> okay, i've trashed my tree in a rebase...will figure out how to get out of that and see what can be posted
<steev> but yeah... bunch of deferred
<steev> https://paste.debian.net/1239448/ so interconnect is needed :(
<bamse> hehe, it's only all the hardware blocks you care about ;)
<bamse> ahh storage forgot to defer, because we don't yet vote for the bus there
<bamse> i think we'll be fixing that soon
<steev> ah, that would definitely do it
<steev> bamse: whenever they get posted, if i could get a CC to know to test, that would be great but i'll try to keep an eye out; we're gearing up for our 2022.2 release though so my free time will be limited again
iivanov has quit []
<bamse> steev: will do
jhovold has quit [Ping timeout: 480 seconds]