ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at https://oftc.irclog.whitequark.org/linux-sunxi
Daanct12 has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
macromorgan has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
chewitt has joined #linux-sunxi
macromorgan has joined #linux-sunxi
chewitt has quit [Quit: Zzz..]
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
<wens> apritzel: (reading logs) the entry for the display engine in DT is the display-engine node. This sits outside of the OF graph but points to phandles for the beginning of all pipelines
hexdump0815 has joined #linux-sunxi
hexdump01 has quit [Ping timeout: 480 seconds]
<wens> oh I see you already figured it out
gsz has joined #linux-sunxi
<Jookia> uh-oh. i found a bug in my panel code that can't be easily fixed in u-boot :\
<Jookia> the nv3052c code requires setting a register based on bit depth or the screen is dim, but u-boot has no mechanism of doing thing
<Jookia> maybe using the bus format is the right way...
<Jookia> i'll have to think about this.
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
<loki666> tokyovigilante: did you tried my build ?
ungeskriptet has quit [Quit: Contact links: https://david-w.eu]
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
<tokyovigilante> loki666: not yet sorry, have been working on the u-boot display code
<tokyovigilante> Jookia: have got the power supply sorted and got the backlight on in u-boot, but seeing this now: nv3052c_panel panel@0: pinctrl_select_state_full: pinctrl_config_one: err=-38
<tokyovigilante> must be getting close..
<tokyovigilante> video 0 [ + ] sunxi_de2 |-- sunxi_de2
<tokyovigilante> vidconsole 0 [ + ] vidconsole0 | `-- sunxi_de2.vidconsole0
pg12_ has joined #linux-sunxi
pg12 has quit [Ping timeout: 480 seconds]
ungeskriptet has joined #linux-sunxi
pg12_ is now known as pg12
<loki666> ok let know, how it goes, because I fear we are pushing dvfs for 6.11 that will need a hotfix
<loki666> if you have a stable Kconfig for your h700-cpufreq-fix branch, I can give it a shot, but I don't think it's related to kconfig choices, either dvfs is enabled or it is not
<tokyovigilante> will do, have downloaded it. I have an alpine build you could try too, https://xserve.testtoast.com/Public/muos-mainline-alpha0-1.img.bz2
<loki666> ok thanks
<tokyovigilante> loki666: what's your image's login?
gsz has quit [Ping timeout: 480 seconds]
<tokyovigilante> I have successfully run a login prompt for about 10mins ;)
<loki666> root/linux
<loki666> you need to modprobe sunxi nvmem cpufreq driver
<loki666> and stress-ng installed
<wens> loki666: we can always revert stuff during -rc
bauen1_ has quit [Ping timeout: 480 seconds]
<loki666> tokyovigilante: your image hangs on my device, before login prompt
apritzel has joined #linux-sunxi
<loki666> second boot worked, https://pastebin.com/1v3fViSd testing
<loki666> what can I run to stress the cpu?
<tokyovigilante> thanks, if you can get into iwd and set up wifi you could just apk add stress-ng, assuming it's in the repos
Daanct12 has quit [Quit: WeeChat 4.3.2]
<tokyovigilante> your build seems stable-ish here, what stress-ng iteration were you using?
<tokyovigilante> From : To
<tokyovigilante> # cat trans_table
<tokyovigilante> : 480000 720000 1008000 1032000 1104000 1200000 1416000 1512000
<tokyovigilante> 720000: 2 0 0 0 0 0 0 1
<tokyovigilante> 480000: 0 0 26 1 0 0 1 787
<tokyovigilante> 1008000: 22 1 0 1 2 1 0 2
bauen1 has joined #linux-sunxi
<loki666> the example in the stress-ng help
<loki666> from my tests, you don't need to run for long, usually it crash early, so try mulitple runs of 10s
<loki666> also my cpu freq never returned to 480000, but maybe it's because of shedutil gov
<loki666> tokyovigilante how do get root on your image (to install stress-ng)
bauen1 has quit [Ping timeout: 480 seconds]
<tokyovigilante> doas for sudo
<tokyovigilante> u/p is muos/muos
<tokyovigilante> I can crash your image pretty readily with just --cpu 2, trying the same test on the alpine build
<tokyovigilante> stress-ng: info: [1179] failed: 0
<tokyovigilante> stress-ng: info: [1179] passed: 4: cpu (4)
<tokyovigilante> stress-ng: info: [1179] metrics untrustworthy: 0
<tokyovigilante> stress-ng: info: [1179] successful run completed in 2 mins, 0.16 secs
<tokyovigilante> Seems ok
<tokyovigilante> I'll get you an up-to-date defconfig for my kernel
lvrp16_ has quit [Ping timeout: 480 seconds]
palmer has quit [Ping timeout: 480 seconds]
yang has quit [Ping timeout: 480 seconds]
arnd has quit [Ping timeout: 480 seconds]
benettig has quit [Ping timeout: 480 seconds]
narmstrong has quit [Ping timeout: 480 seconds]
<loki666> the issue is scaling_gov, as soon as I set your alpine build to schedutil
<loki666> it crash
NishanthMenon has quit [Ping timeout: 480 seconds]
<tokyovigilante> Is there any reason to change it?
<loki666> no, it was just set by default by arm64 defconfig, and seems to be the new recommended scaling gov
<loki666> but the real question, is why does it crash
bauen1 has joined #linux-sunxi
<tokyovigilante> I would suggest defaulting to conservative, which will still scale to 1.5GHz as required
<loki666> sure, but this is more like a workaround
<tokyovigilante> mine defaults to conservative, I don't think I ever changed it. Buildroot may default to schedutil I suppose
<tokyovigilante> Would be worth feeding that back on the list, but I feel like that's more of a thing to document rather than a reason to revert or not have the 1.5GHz OPP available. The bug may be in the schedutil governor, or a combination of the two
narmstrong has joined #linux-sunxi
<loki666> now that we have a lead, i'll test the other governors
benettig has joined #linux-sunxi
lvrp16_ has joined #linux-sunxi
NishanthMenon has joined #linux-sunxi
<tokyovigilante> thanks, that's good detective work
arnd has joined #linux-sunxi
yang has joined #linux-sunxi
<loki666> I can't hack the kernel, but I can a least test it
<tokyovigilante> it's only a small step from there to here ;)
yang is now known as Guest10736
palmer has joined #linux-sunxi
<tokyovigilante> Do you see instability with schedutil + the 6.9 DVFS code? ie just the lower frequencies?
<loki666> it was stableis when I removed latestes 2 freq bins
<tokyovigilante> ok, good to know
<loki666> I think 6.9 didn't work at all, it was missing the hw-support in the DTS
<loki666> and the code patch you made
<tokyovigilante> Ah sorry, yeah 6.10. I've been running too many patches on top
<tokyovigilante> interestingly the vendor BSP runs schedutil with the 1.5GHz bin and seems fine
<tokyovigilante> on a much older kernel of course
<tokyovigilante> I would definitely feed it back to the list
<loki666> oh, I thought shedutil didn't exists yet on 6.4
<loki666> it using ondemand
<tokyovigilante> that's according to my CFW, I'll dig into it
Guest10736 is now known as yang
<loki666> looks like on schedutil is causing issues, conservative, ondemand, powersave and performance are working fine
<loki666> didn't test userspace
<apritzel> loki666: that's good info for the list!
<loki666> I report on your patch serie ?
<loki666> I mean on tokyovigilante serie ?
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
jernej has joined #linux-sunxi
<apritzel> yeah, for example. Most important is to include the right people, tokyovigilante's series should have them already
bauen1 has quit [Ping timeout: 480 seconds]
gsz has joined #linux-sunxi
gsz has quit [Ping timeout: 480 seconds]
<wens> IIRC from issues at work, schedutil requires two passive trip points
bauen1 has joined #linux-sunxi
JohnDoe3 has joined #linux-sunxi
<Jookia> tokyovigilante: not sure what the pinctrl thing is
dsimic is now known as Guest10751
dsimic has joined #linux-sunxi
Guest10751 has quit [Ping timeout: 480 seconds]
<Jookia> tokyovigilante: could you tell me if the rg35xxx has a nice pmic with proper reset? as in the rails go down them up on reboot?
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
electricworry has joined #linux-sunxi
Gabriel has joined #linux-sunxi
Gabriel has quit [Remote host closed the connection]
Gabriel has joined #linux-sunxi
Gabriel has quit [Remote host closed the connection]
haluka has joined #linux-sunxi
warpme has quit []
haluka has quit [Remote host closed the connection]
<apritzel> Jookia: IIRC the AXP717 has reset functionality, though we don't use that atm
<apritzel> reboot goes via PSCI, though, so we could change TF-A to use the PMIC for reboot, instead of the watchdog
<Jookia> i've found that these panels need proper power cycling if uncleanly reset :\
<Jookia> i'm going to try troubleshooting the reset logic a bit more
<Jookia> otherwise the panel introduces artifacts such as flickering or lack of channels
<apritzel> yeah, that's an interesting problem: the normal Linux reboot only affects the SoC, not anything on the board (or beyond)
<apritzel> Jookia: is that for some T113-s3 board? Those typically don't have a PMIC, do they?
<Jookia> yeah. so i was wondering if the rg35xx pmic resets the LCD
<Jookia> things like that
<Jookia> it could also be that this is specific to the panel i have
<apritzel> is there some extra pin on the panel to reset it, maybe?
<Jookia> yes, there's a reset gpio and reset instructions that does reset it. but something is not getting reset. so i'll have to experiment
warpme has joined #linux-sunxi
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
warpme has quit []
<macromorgan> Jooka: Hold the panel in reset in your panel disable/unprepare routine?
<macromorgan> I think I do that on every panel driver I've done thus far since issues like to pop up
<macromorgan> also, starting over and narrowing things down, it looks like the i2c_write() is what's failing when dcdc2 and dcdc3 get set on boot for the AXP717 in U-Boot SPL
<macromorgan> digging into that a bit more
<macromorgan> looks like i2c_write is writing the value but not receiving an acknowledgement
<loki666> wens: what do you mean?
JohnDoe3 has quit []
warpme has joined #linux-sunxi
gsz has joined #linux-sunxi
warpme has quit []
Schimsalabim has quit [Read error: Connection reset by peer]
apritzel has joined #linux-sunxi
Schimsalabim has joined #linux-sunxi
ftg has joined #linux-sunxi
vagrantc has joined #linux-sunxi
gsz has quit [Ping timeout: 480 seconds]
electricworry has quit [Quit: Leaving]
vagrantc has quit [Ping timeout: 480 seconds]
ftg has quit [Read error: Connection reset by peer]
dliviu has quit [Quit: Going away]
dliviu has joined #linux-sunxi