12:38 <DavidHeidelberg> msm driver for REFGEN; as recently added to be dep of mdss, a bad things happen when panel (any) is in-kernel and refgen gets in as a module

12:38 <DavidHeidelberg> could be there solution which would pull REFGEN when panel driver is compiled-in?

17:20 <konradybcio> mani_s: I get correctable AERs on hamoa nvme on today's next

17:21 <mani_s> konradybcio, did you test other next tags for this release or just today's?

17:22 <konradybcio> just today

17:22 <mani_s> Ok. Not sure if that's due to ASPM enablement. Try booting with pcie_aspm=off

17:24 <konradybcio> mani_s: gone

17:25 <mani_s> konradybcio, hmm, we just enable L0s and L1 as of -rc3, but the NVMe might not be playing well with it

17:26 <konradybcio> pci 00...: ASPM: default states L1

17:26 <mani_s> please share 'sudo lspci -vv'

17:30 <konradybcio> i also can't echo 1 > remove on the wlan card on today's next so let me find a thumb drive to move the file over :)

17:30 <konradybcio> (it segfaults)

17:31 <mani_s> oops

17:31 <mani_s> that's weird

17:33 <konradybcio> mani_s: https://pastebin.com/raw/mFrEJrtQ

17:35 <mani_s> konradybcio, this dump is with cmdline passed?

17:35 <konradybcio> no

17:35 <konradybcio> do you want that instead?

17:36 <mani_s> no, was wondering how L1 gets enabled

17:38 <konradybcio> just to be clear, pcie_aspm=off does not print that line in dmesg

17:39 <mani_s> ok, pcie_aspm=off just retains the fw config. And the firmware has enabled L1 already for you

17:40 <mani_s> but still not sure how that option fixed the AERs

17:41 <mani_s> none of the Root Ports support L0s on Hamoa, we disable the L0s CAP. So only L1 would be enabled

17:41 <mani_s> s/enabled/available

17:42 <mani_s> konradybcio, maybe, share 'sudo lspci -vv' without that option for comparison

17:43 <konradybcio> with*?

17:44 <mani_s> without pcie_aspm=off so that we can see what goes with config space during AERs

17:44 <konradybcio> I believe that's what I sent above

17:49 <mani_s> sorry, I rebooted my machine in between and got only your "with*' message

17:52 <konradybcio> the pastebin link which contains the output of lspci -vv which you looked at is without the pcie_aspm=off cmdline option added

17:53 <mani_s> konradybcio, ah ok. But I do not see any AER errors logged

17:54 <mani_s> can you share the dmesg snippet of the AERs?

17:55 <konradybcio> rest assured they're there.. perhaps i managed to lspci -vv before the first one appeared

17:56 <mani_s> could be

17:56 <mani_s> better to capture it after the AERs

17:57 <konradybcio> dmesg https://gist.githubusercontent.com/konradybcio/af4b3c219b08a073fb558182c5206f7c/raw/3c334e046a681d2f3b9eebba9e14fd6db69fa411/gistfile1.txt

17:59 <konradybcio> sudolspci-vvv https://gist.githubusercontent.com/konradybcio/d4aa77b9b6e934603dfc7b13f9f4a022/raw/a705f530713434125258c087bbaf727485afb625/gistfile1.txt

17:59 <konradybcio> notably still lacking errors

18:00 <mani_s> wow

18:04 <mani_s> konradybcio, try this: https://pastebin.com/raw/trtgdVvv

18:10 <konradybcio> yeah printing on just the nvme as expected and no more AERs

18:10 <mani_s> konradybcio, cool. I will spin up a patch tomorrow

18:11 <mani_s> just sent out the series for the day

18:11 <mani_s> I'm done

18:14 <konradybcio> mani_s: so our best guess is that this specific nvme can't do L1?

18:16 <BhushanShah[m]> adrian: aklimov: few days ago one of you mentioned patch in lineageos to set initial master gain, is it done on DSP side or on awinic? Basically we're trying to figure out missing piece for this on mainline