ChanServ changed the topic of #linux-msm to:
<bryanodonoghue> latest next and 6.2-rc2 have the same failure
<bryanodonoghue> oddly I can associate with an station - so basic tx/rx works..
<bryanodonoghue> Regression somewhere between * 830b3c68c1fb1 - (tag: v6.1) Linux 6.1 (4 weeks ago) and 6.2-rc2
<bryanodonoghue> and if I cherry-pick my latest dtsi and dts onto that base - everything works
Danct12 has joined #linux-msm
Danct12 has quit [Ping timeout: 480 seconds]
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]
latex_ has joined #linux-msm
latex has quit [Ping timeout: 480 seconds]
<steev> bryanodonoghue: as in, you get like 32KB/s max?
<steev> or can you not pass traffic at all
<mal> konradybcio: hi, sorry to bother you but about so the ordering of nodes is based on the pwm in "backlight_pwm: pwm {" not based on backlight_pwm?
<konradybcio> yes, it's based on the node name and not the label
<konradybcio> LABEL: node_name {
<mal> ok, will change it then
<mal> konradybcio: about the order of pinctrl-names and pinctrl-0, in pretty much all examples in binding documentation the order is pinctrl-names and then pinctrl-0 like I have
<konradybcio> yes, but it was only recently discovered that we've been doing a big derp all along
<konradybcio> for all other properties (clocks, reset, power-domains) we've always had X and then X-names
<mal> ok
jhovold has quit [Quit: WeeChat 3.6]
jhovold has joined #linux-msm
latex_ has quit [Quit: leaving]
<bryanodonoghue> steev as in it takes *forever* to get an IP address and then traffic throughput is extraordinarly low
<bryanodonoghue> probably related to a790cc3a4fa ("wifi: mac80211: add wake_tx_queue callback to drivers")
<aka_[m]> so reverting it helps
<aka_[m]> ?
<aka_[m]> might try that on my msm8976 crap
<aka_[m]> bryanodonoghue:
<aka_[m]> also if i were smarter and not dumb fuck i might be able to know why imps breaks wifi
<aka_[m]> sadly im not
<bryanodonoghue> aka_[m] are you running bluetooth concurrently ?
<aka_[m]> by default BT and WIFI is enabled
<aka_[m]> tried with postmarketos rootfs
<aka_[m]> not like i want to have bt connected with wifi
<aka_[m]> still i only tried to connect to wifi in order to download unixbench, had to resort to reinstalling rootfs at the end because wifi even after i connected i was unable to install package with apk
<aka_[m]> whatever mainline is so sad lately on these old socs
<aka_[m]> too many things get easily broken
<aka_[m]> a5xx is dead right now
<bryanodonoghue> right what about cat /sys/kernel/debug/ieee80211/phy0/wcn36xx/firmware_feat_caps
<aka_[m]> and if i run software render cpu is getting really fucked and phone rebooting with some random watchdog after a while
<aka_[m]> do we need wcn debug for that?
<bryanodonoghue> should be on by default
<aka_[m]> last time you told me i have wcn3680 instead of 3660b
<aka_[m]> based on caps
<aka_[m]> it has 5Ghz AC
<bryanodonoghue> Ah yes it exported DOT11AC
<bryanodonoghue> ?
<aka_[m]> yes
<bryanodonoghue> 3660 has 5ghz
<aka_[m]> 3660b
<aka_[m]> i have it on my msm8953 device
<bryanodonoghue> I can't find an email from hyou on 8976
<bryanodonoghue> can you ping my
<aka_[m]> no email
<aka_[m]> we discussed it here
<bryanodonoghue> heh - I don't have in the backbuffer
<bryanodonoghue> contents of /sys/kernel/debug/ieee80211/phy0/wcn36xx/firmware_feat_caps and maybe a recommendation on a phone or devkit I can pick up with an 8976 for reasonable money
<bryanodonoghue> I can add it to my stack of random buys ;)
<aka_[m]> i sent you this
<aka_[m]> back in time based on matrix backlog
<aka_[m]> wrong
<aka_[m]> not this log
<aka_[m]> i hate this
<aka_[m]> somehow my windows going as retard as me and this log is indeed one with caps
<aka_[m]> <bryanodonoghue> "I can add it to my stack of..." <- 8976 isnt that worth money, its quite broken soc and for wcn featured device you might want to try look into leeco Le2 or Redmi Note 3 Pro, Sony offerings using broadcom module instead of WCN
<aka_[m]> quite sad SoC
<bryanodonoghue> sorry and again you are excluding IMPS but not BMPS
<aka_[m]> 3.8k in geekbench with some shit patches for cpufreq but these are too shit to ever be accepted 😐️
<bryanodonoghue> is that correct ?
<aka_[m]> after reverting IMPS it not crashing
<aka_[m]> * IMPS it's not
<bryanodonoghue> do you have a trace of the crash ?
<aka_[m]> still connecting nowadays is harder
<aka_[m]> its pretty much like on screenshot
<aka_[m]> remoteproc returning this haltACK error
<aka_[m]> and if it triggers few times it reboots
<aka_[m]> this appears to be some conf file
<aka_[m]> gEnableImps=1 so it should not fail
<aka_[m]> i feel like giving up im too dumb to sort it out.
<aka_[m]> Guys like Angelo are leagues above why would i even bother.
<aka_[m]> log says last caps appears to be EXT_SCAN_ENHANCED does it mean later ones are not available?
<aka_[m]> so it might not like ANTENNA_DIVERSITY_SELECTION
<bryanodonoghue> the ini file is not parsed
<bryanodonoghue> try "iw wlan0 set power_save off"
<bryanodonoghue> you definitely have a different version of firmware to me
<aka_[m]> for sure older
<aka_[m]> company is ded for long
<aka_[m]> fw_info says something about LA.BR.1.0.1
<aka_[m]> thats really old
<bryanodonoghue> so if we can't determine a firmware feature bit, we will have to quirk IMPS with a dts setting
marvin24_ has joined #linux-msm
marvin24 has quit [Ping timeout: 480 seconds]
minecrell has quit [Remote host closed the connection]
minecrell has joined #linux-msm
<aka_[m]> bryanodonoghue: it appears redmi note 3 pro ships with bit newer firmware but i would have to get someone to test.
<aka_[m]> my leeco is: CNSS-PR-3-0-c2-00095
<aka_[m]> redmi note 3 pro: CNSS-PR-3-0-c2-00121
<aka_[m]> based on google random search it even brought me some asus device shipping CNSS-PR-3-0-c2-00085
<aka_[m]> maybe i notice it wrong but
<aka_[m]> msm8916 -PR-2 msm8976 -PR-3 msm8937/msm8953 -PR4
<aka_[m]> some timeouts, quite old topic
jhovold has quit [Ping timeout: 480 seconds]