_lore_ has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
_lore_ has joined #linux-msm
<lumag>
marc|gonzalez, apq8096-db820c has working HDMI audio
<lumag>
aka_[m], it might be, I don't think we have GFX CPR nowhere close
<lumag>
So maybe we have to bump GFX voltage
<steev>
lumag: have you happened to use 6.11+ on the c630? it, as well as the flex5g seem to no longer be starting up(?) the remoteprocs for the wifi (i'm not entirely sure of how this all works so i have no idea how to dig into what is happening. i see the "ath10k_snoc XXXXXX.wifi: Adding to iommu group XX" but that's it. no wifi device shows up
<lumag>
steev, I've been testing 6.11-rc if I'm not mistaken
<lumag>
chec that you have pd-mapper running (in-kernel or userspace one), check that tqftpserv is running
<lumag>
and of course status in /sys/class/remoteproc/remoteprocN/state
<steev>
tqftpserv is running, everything seems running just no wifi
<lumag>
steev,try disabling in-kernel pd-mapper (Kconfig symbol) and then starting the userspace one
<lumag>
also check if there is anything in syslog/journal from tqftp
<steev>
hm, okay it might be tqftpserv. i'm seeing a "stat failed: no such file or directory" and then a "invalid path /readonly/vendor/firmware/wlanmdsp.bin, rejecting" - at least on the flex 5g, i'll have to check the c630 in a bit
<lumag>
steev, also make sure that it passes the right wlanmdsp.bin - i.e. Lenovo-signed
<lumag>
I think we need to improve logging
<steev>
that's the one that is in WCN3990?
<lumag>
steev, yes
<lumag>
steev, but both c630 and 5g have wcn3990
<steev>
right
<lumag>
steev, I don't have c630 (or 5g) at hand, but I should be able to test miix630.
<lumag>
It is currently running 6.11-rc5
<lumag>
marc|gonzalez, btw: any plans for cpufreq on msm8998?
<steev>
there's no rush on it, these devices are my backups (and i have external wifi cards galore), with the x13s being my main, but it doesn't show the same issue (probably because it's a pci device?)
<lumag>
yep, x13s has pcie wifi (ath11k), no tqftpserv, no remoteprocs
<lumag>
I'd really suggest instrumenting tqftpserv to dump more information (so that it states, which file is being sent, etc).
<lumag>
also could you check if it was upgraded recently?
<lumag>
(and once you know which files are being sent, check that they are Lenovo ones)
<steev>
i was showing 3990's wlanmdsp.bin from jan 4th (and the one in my flex's directory being from 2021); this definitely only seemed to start happening this month - according to the debian tracker there *was* an update to it on the 2nd of this month...
<lumag>
steev, maybe that's the case, try tqftpserv from snapshot.debian.org
<steev>
exactly what i was loading :D
<lumag>
:-D
<steev>
yep!
<steev>
okay the 1.1-1 update is borked, thanks, i'll file a bug
<lumag>
steev, nice!
<lumag>
konradybcio, ^^
<lumag>
let's understand what and why.
<lumag>
I'll check what do I have on miix
<steev>
quick guess after glancing at the debian packaging... the zstd patch may mean that it's only checking for the compressed file now
<lumag>
steev, no, I don't think so
<steev>
yeah i tried compressing to look, and it's not that
<lumag>
The patch was to fallback to zstd
<lumag>
if required file isn't present
<lumag>
check, which wlanmds..mbn gets loaded
<lumag>
There were changes in the path lookup
<steev>
well, with 1.1 none of them do :P
<lumag>
ok, which one is attempted
<steev>
yeah, i'll try to look into it, but tomorrow, it's 2:30am here and i *should* be responsible and get some sleep
<steev>
well, it tries to read /sys/module/firmware_class/parameters/path and on my system, that seems to be... empty
<steev>
which, i assume means it should fall back to the FIRMWARE_BASE of /lib/firmware and debian is now merged usr so that is a symlink to /usr/lib/firmware and maybe that trips it up
<steev>
okay i really should go to bed, but i'll try to look tomorrow after work
<lumag>
steev, thanks!
<lumag>
If I find something in the meantime, I'll ping you on the IRC
<steev>
no rush or anything, i can always apt-mark hold tqftpserv ;)
<steev>
and yeah, downgrading on both causes it to work
<lumag>
hmm, ok
<lumag>
I'll check which version do I have on miix in 10-15 minutes
srinik has joined #linux-msm
pespin has joined #linux-msm
srinik has quit [Remote host closed the connection]
srinik has joined #linux-msm
srinik has quit [Remote host closed the connection]
srinik has joined #linux-msm
srinik has quit [Remote host closed the connection]
srinik has joined #linux-msm
srinik has quit [Ping timeout: 480 seconds]
srinik has joined #linux-msm
<TJ[m]>
Are the QCA6390 support suspend in mainline? Testing in RB5 board and also other SM8250, the board will restart or failed to wakeup from suspend when ATH11K enabled in kernel.
srinik has quit [Ping timeout: 480 seconds]
<lumag>
TJ[m], which kernel, what kind of suspend/resume/hibernate are you using?
<lumag>
Why do you think that it's related to qca6390?
srinik has joined #linux-msm
<TJ[m]>
<lumag> "TJ, which kernel, what kind of..." <- 6.6 from codelinaro and also with latest 6.11 have same problem
<TJ[m]>
<lumag> "Why do you think that it's..." <- Because suspend/resume work when I don’t use QCA6390, example when using v6.10, the power sequence patch are not exist. Or when using v6.11 with QCA6390 related disabled in device tree
<TJ[m]>
I test this release in RB5, the suspend/resume broken
<lumag>
ok, nice. brgl, any chance you can take try suspend/resume on 6.11?
<lumag>
if that's related to pwrseq, we might need to fix it.
srinik has quit [Ping timeout: 480 seconds]
<brgl>
lumag: yeah will do. Last I tested suspend/resume, it worked fine
<lumag>
brgl, great :-)
<brgl>
TJ[m] lumag: seems to work fine here, how do you reproduce it?
<TJ[m]>
Now I’m in RB5, in Linaro release of Debian Bookworm above with v6.6, just press power button to suspend. Then press power button again, it takes time to resume (about 5-10 second) then the WiFi and BT are failed to resume.
<TJ[m]>
Previously using v6.11, RB5 will rebooted after pressing power button from suspend
<brgl>
So it corresponds to rtcwake's suspend to disk?
<brgl>
I only have remote access ATM
<brgl>
need to reproduce it from console
<lumag>
brgl, I think rtcalarm can be used to wake it up
<brgl>
lumag: I can use rtcwake alright but it works fine for both -m disk and -m mem
<lumag>
hmm
<lumag>
TJ[m], could you possibly share your cmdline and .config?
<brgl>
lumag: but I'm running initramfs remotely so maybe we need suspend to disk with a real on-disk rootfs
<lumag>
brgl, I'd lean to something like clk_ignore_unused / pd_ignore_unused
<TJ[m]>
lumag: I use default rootfs and boot.img in latest Linaro release above, no changed at all.