ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at
KREYREN_ has joined #linux-sunxi
KREYREN_oftc has quit [Remote host closed the connection]
montjoie_ has joined #linux-sunxi
montjoie has quit [Ping timeout: 480 seconds]
KREYREN_ has quit [Remote host closed the connection]
<ItsKaitlyn03> taking a bit of a break for now zzz
ItsKaitlyn03 has quit [Remote host closed the connection]
<Jookia> o/
JohnDoe_71Rus has joined #linux-sunxi
hexdump0815 has joined #linux-sunxi
chuang[m] has quit []
hexdump01 has quit [Ping timeout: 480 seconds]
exkc has quit []
psydroid[m] has quit []
<Jookia> yes my driver is subtly broken yay
junari has joined #linux-sunxi
junari has quit []
warpme has joined #linux-sunxi
<Jookia> wait, it's not broken. the documentation is just wrong. so the best kind of borken!
<jernej> Jookia: which driver? pwm?
<Jookia> yeah. i've spent too much time reading this documentation but i think it's highly misleading
<Jookia> the descriptions of counters and timings all say (ENTIRE_CYCLES + 1) which i took to mean 'if i say there's 1 entire cycles, i get 2'
<Jookia> but it actually means the register's encoded value + 1, where 0 = 1 cycle, 1 = 2 cycles
<jernej> ah, yes
<Jookia> so my documentation says there's 65537 cycles and talks about an extra cycle
<Jookia> so while the code works fine, the documentation needs fixing
<Jookia> i'll make a quick patch today and just note it down in the mailing list some other time since there's plenty of review to do it sometime
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC:]
jernej has joined #linux-sunxi
<Jookia> i have dyscalculia so this is really just a big struggle for me :(
<Jookia> ok i think i've mentally untangled why this is this way and i'll write it up tomorrow
<tokyovigilante> Jookia: will be worth it.
<Jookia> hopefully
<tokyovigilante> For me, I have just realised I've been constantly pushing between a DT in upstream and legacy for u-boot, and then wondering why my patches kept disappearing off the device... sigh
<tokyovigilante> still, I now have a *very* tidy upstream DT
<Jookia> i'm not a hardware designer but i think they may have done this to avoid arithmetic math in their counter
<tokyovigilante> loki666: sorry about the mess in my u-boot tree, have just pushed a totally revised set, with what should be an entirely correct upstream DT in u-boot, rebased on jernej's latest DRAM patches and apritzel's PMIC driver, so should be the absolute latest and best version
<tokyovigilante> suppose I'd better test it myself now that I've pushed it...
<jernej> tokyovigilante: I just updated DRAM patches again less than an hour ago :)
<jernej> and yes, it was for LPDDR4
<jernej> but I think I'll stop now
<jernej> tokyovigilante: did those fixes help in any way with DRAM size detection issue?
<tokyovigilante> I'm just building a fresh kernel and u-boot and will test just that
<tokyovigilante> thanks, good timing!
<tokyovigilante> dts/upstream/src/arm64/allwinner/sun50i-h700-anbernic-rg35xx-h.dtb: Warning (graph_endpoint): /panel/port/endpoint: graph connection to node '/soc/tcon-top@6510000/ports/port@1/endpoint@2' is not bidirectional
<tokyovigilante> No kidding ;) Still have no idea what I'm doing for the LCD, but realistically will need the PWM for that to, so will get these DRAM changes tested
<jernej> you have hackish H616 PWM driver too in my branch
<tokyovigilante> oh right, cool. I think that makes 3 ;) macromorgan's and at least one other about
<tokyovigilante> ok, so first boot detected 1024MB correctly, any other way to test than keeping an eye out for misdetection?
<jernej> not really
<jernej> note that I didn't bother that much with PWM driver, good enough for backlight and AC200 clock source
JohnDoe_71Rus has quit [Quit: KVIrc KVIrc Quasar 5.2.2, revision: 5.2.0+git-7584-200e7bcec, build type: debug, sources date: 20160102, built on: 2024-05-13 20:04:01 UTC 5.2.0+git-7584-]
<tokyovigilante> makes sense, ta
<tokyovigilante> will hopefully be up and running then swap out for one of the others
<tokyovigilante> Ok, have rebased my kernel tree against torvalds/master, so pulled in cpufreq, GPADC, NMI etc that got in this window, so should be good to go
<tokyovigilante> 3 boots, so far so good
<jernej> good to hear
<loki666> tokyovigilante : no problem which Linux branch should be good to test ?
warpme has quit []
<tokyovigilante> anbernic-hdmi for linux, and rg35xx-upstream-extras should be good to go, they are both rebased and some bugs fixed. Need to just recheck the issue you found, but my tree has just built here
<tokyovigilante> │ Symbol: SUN50I_IOMMU [=n]
<tokyovigilante> ah, that is probably why
<loki666> Yes
<tokyovigilante> sorry, have fixed that in my tree against jernej's patches and repushed, you should be ok with my tree and IOMMU enabled now
<loki666> Ok I'll give it a spin
<tokyovigilante> actually not quite sorry
<tokyovigilante> every time I think i've mastered git, I haven't
warpme has joined #linux-sunxi
<tokyovigilante> sorry, fixed now
warpme has quit []
<tokyovigilante> Interestingly I'm now getting a kernel panic very early in boot with the IOMMU enabled
<tokyovigilante> [ 1.440839] sun50i_de2_bus_probe+0x44/0x90
<tokyovigilante> [ 1.133711] sun8i-mixer 1100000.mixer: Adding to iommu group 1
<tokyovigilante> [ 1.140165] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000078
<jernej> how much RAM has your board?
<tokyovigilante> 1GB
<tokyovigilante> I had the IOMMU 4GB enable patch wrong, but wasn't seeing it because I had the IOMMU disabled. Could that be the problem?
<jernej> for 1GB, you don't need 4 GB patch
<jernej> and you can skip IOMMU altogether, if you won't play videos
<tokyovigilante> makes sense, I did think I'd had HDMI working previously without it.
<tokyovigilante> I guess ideally the patch would support up to 4G and ideally I will play video at some point, but ok for now without
<tokyovigilante> loki666: ok, repushed with that 4G patch deleted, HDMI should work either with IOMMU disabled or the current branch now with your config
<tokyovigilante> getting *slightly* closer to lcd too
<tokyovigilante> dts/upstream/src/arm64/allwinner/sun50i-h700-anbernic-rg35xx-h.dtb: Warning (reg_format): /soc/lcd-controller@6511000/ports:reg: property has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1)
<tokyovigilante> but the DT compiles and I think I have everything I need driver-wise, minus some potentially bogus panel timings. Will take any feedback on how to wire the backlight too :)
<tokyovigilante> dts/upstream/src/arm64/allwinner/sun50i-h700-anbernic-rg35xx-h.dtb: Warning (reg_format): /soc/lcd-controller@6511000/ports:reg: property has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1)
<tokyovigilante> good news is no DRAM misdetection on about 10 reboots
<Jookia> yay!
<loki666> I'll try your branches u-boot rg35xx-upstream-extras + Linux anbernic-hdmi
dsimic is now known as Guest6984
dsimic has joined #linux-sunxi
Guest6984 has quit [Ping timeout: 480 seconds]
<jernej> tokyovigilante: do you have this patch in entirety?
Schimsalabim has quit [Read error: Connection reset by peer]
JohnDoe_71Rus has joined #linux-sunxi
Schimsalabim has joined #linux-sunxi
Robot_ has joined #linux-sunxi
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
Schimsalabim has quit [Ping timeout: 480 seconds]
KREYREN_oftc has joined #linux-sunxi
<KREYREN_oftc> does anyone have any idea to why i can't get into FEL?
<KREYREN_oftc> it literally gets requested the address 0x10000 and it boots uboot again
Schimsalabim has joined #linux-sunxi
warpme has joined #linux-sunxi
ftg has joined #linux-sunxi
warpme has quit []
warpme has joined #linux-sunxi
warpme has quit []
warpme has joined #linux-sunxi
warpme has quit []
KREYREN_oftc has quit [Remote host closed the connection]
sddddd has joined #linux-sunxi
sddddd has quit [Remote host closed the connection]
warpme has joined #linux-sunxi
warpme has quit []
warpme has joined #linux-sunxi
warpme has quit []
<loki666> tokyovigilante: so I tried, your update branches, it boots (still no wifi), but no hdmi, I tried on two screens, a 1080p TV and a 1080p monitor
<loki666> on TV no signal, on monitor, i have a screen with garbage
<loki666> should I call a fbset to change the resolution ?
<jernej> loki666: add drm.debug=0x1e to kernel arguments and see what it prints in dmesg
<jernej> you should see messages that hdmi is connected and all supported resolutions on your monitor
<jernej> if you have 4k monitor, it may be also issue with the cable or with driver
<loki666> that's the TV
<loki666> i also have random crash just after [ 2.433977] sun4i-drm display-engine: bound 6515000.lcd-controller (ops 0xffff800080f1faf0)
<jernej> something is wrong with connection detection
<jernej> you can force it to always connected with kernel argument, but I don't remember exact command from top of my head
<loki666> that's the monitor log
<loki666> it does show resolutions
<jernej> did you plug out hdmi before you checked dmesg?
<loki666> yes i did
<loki666> without wifi, I can't check while plugged id, my monitor is not next to my uart
<jernej> your monitor should be good in theory, but there are certainly issues with TV
<jernej> but if you don't have image with monitor, then I don't know. All pieces seem to be in place
<loki666> Well TV does works with anbernic OFW
JohnDoe_71Rus has quit [Quit: KVIrc 5.2.2 Quasar]
<jernej> which tv is that?
<tokyovigilante> loki666: thats very annoying! maybe try to fix wifi first, that has been working for a couple of months.
<tokyovigilante> i can see bluetooth but not wifi in your dmesg, maybe check your wifi module is enabled in your kconfig? you will also need the firmware in your image
<tokyovigilante> i
<tokyovigilante> your bluetooth is not loading even with the driver built so i suspect missing firmware
<loki666> both firmware and rtw88 seems to be there
<tokyovigilante> not sure if thats easy in buildroot, i would suggest using a distro userspace so you hsve more tools available.
<tokyovigilante> is it compressed?
<tokyovigilante> you need support for compressed modules then.
<loki666> philippe@bob:/media/philippe/rootfs/lib/firmware/rtw88$ ls
<loki666> rtw8723d_fw.bin rtw8821c_fw.bin rtw8822b_fw.bin rtw8822c_fw.bin rtw8822c_wow_fw.bin
<loki666> philippe@bob:/media/philippe/rootfs/lib/firmware/rtl_bt$ ls
<loki666> rtl8821a_fw.bin rtl8822b_config.bin rtl8822cs_fw.bin rtl8851bu_config.bin rtl8852au_fw.bin rtl8852cu_config.bin
<loki666> rtl8821a_config.bin rtl8821c_fw.bin rtl8822cs_config.bin rtl8822cu_fw.bin rtl8852au_config.bin rtl8852bu_fw.bin rtl8852cu_fw_v2.bin
<loki666> rtl8812ae_fw.bin rtl8821c_config.bin rtl8822b_fw.bin rtl8822cu_config.bin rtl8851bu_fw.bin rtl8852bu_config.bin rtl8852cu_fw.bin
<tokyovigilante> you're missing 8821cs i think, will check mine
dok_ has joined #linux-sunxi
<tokyovigilante> but you should have driver messages regardless, and it would warn about missing firmware
<tokyovigilante> oh no all there
<tokyovigilante> [ 16.905032] platform pwrseq: deferred probe pending: pwrseq_simple: external clock not ready
<tokyovigilante> This is your problem, the system can't turn the wifi regulator on
<tokyovigilante> although clearly detecting a HDMI connector subsequently
<tokyovigilante> also I recommend you blacklist panfrost for now, that is probably why it isn't working, the GPU driver is crashing because of (another) power issue
dok has quit [Ping timeout: 480 seconds]
<loki666> ok gonna try to blacklist panfrost
<loki666> jernej: tv is an old LCD philips 42" model : 42pfl7606h
<loki666> tokyovigilante: panfrost blacklisted but still no wifi
<loki666> could you share you kconfig ?
<loki666> also, don't know why, but anbernic OFW boot-0 refuse to boot after booting mainline... need to unplug the power completly (unplug battery)
<loki666> as if mainline powerdown, still keeps something in a reg/mem
<tokyovigilante> sure, will post it ths evening, sorry at work now
<tokyovigilante> there is a PMIC reset that I think we are not quite handling correctly, which is blocking the anbernic u-boot which has a much older version, and probably poor support for the AXP
Robot_ has quit [Ping timeout: 480 seconds]
ftg has quit [Read error: Connection reset by peer]