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
DarkNeutrino has joined #linux-sunxi
<DarkNeutrino> apritzel_: so the random null dereference was linux-next. Stable boots flawesly :)
apritzel_ has left #linux-sunxi [#linux-sunxi]
apritzel has joined #linux-sunxi
<apritzel> DarkNeutrino: that's still odd, I guess we need to keep an eye on that
<DarkNeutrino> Well it was a tip from a friend. If weird stuff happens on next try stable. If it works its next issue. If it doesnt switch back to next and fix it :)
<apritzel> well, -next will be tomorrow's stable, so this will come back and haunt us ...
<DarkNeutrino> I got the issue on 18.1.2023 next. So we shall see
cnxsoft has joined #linux-sunxi
<apritzel> jernej: smaeul: wens: I get "Unevaluated properties are not allowed ('phy-supply' was unexpected)", seemingly for all DTs with an sun8i EMAC node.
<apritzel> I see that mentioned in the D1 DT series review from last summer, did anyone ever address this?
<apritzel> and is this (relatively) new? I don't remember seeing this before ...
<apritzel> ah, I even reviewed it, I now remember :-D
sunshavi has joined #linux-sunxi
DarkNeutrino has quit [Ping timeout: 480 seconds]
DarkNeutrino has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
evgeny_boger has joined #linux-sunxi
evgeny_boger has quit [Remote host closed the connection]
vagrantc has quit [Quit: leaving]
macromorgan has quit [Read error: Network is unreachable]
macromorgan has joined #linux-sunxi
jason123onirc has quit [Quit: ZNC 1.8.2+deb2 - https://znc.in]
sunshavi has quit [Remote host closed the connection]
sunshavi has joined #linux-sunxi
jason123onirc has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
codekipper has joined #linux-sunxi
cnxsoft has quit []
pmp-p is now known as Guest1915
pmp-p has joined #linux-sunxi
hlauer has joined #linux-sunxi
Guest1915 has quit [Ping timeout: 480 seconds]
<szemzoa> On V853/V851S it's indeed possible to use the RISC-V core. On ARM, I have a linux running, and the RISC-V is running a BSP RT-Thread firmware (600MHz) with a console on another uart, what I uploaded from my bootloader now: https://pastebin.com/GtjRu6AB
<gamiee> Oh cool!
apritzel has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
bauen1 has joined #linux-sunxi
seb_ has quit [Remote host closed the connection]
JohnDoe_71Rus has quit []
DarkNeutrino has quit [Ping timeout: 480 seconds]
DarkNeutrino has joined #linux-sunxi
evgeny_boger has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
evgeny_boger has quit [Ping timeout: 480 seconds]
<mps> looks like sun4i drm doesn't work if CMA is not enabled?
<smaeul> you need either CMA or an IOMMU
<mps> does IOMMU works on allwinner A64
<mps> because I have IOMMU_SUPPORT=y but without CMA display is blank
<mps> and I don't know which specific iommu needs to be enabled for A64 SoC
codekipper has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
grming has joined #linux-sunxi
<apritzel> mps: there is no IOMMU on the A64
<apritzel> only the newer ones have it: H6, H616, D1 & family
<mps> apritzel: thanks for info
cnxsoft has quit []
<apritzel> mps: what's the problem with CMA? You can adjust the size with a command line parameter, if that's a concern
<mps> apritzel: it is problematic in some cases especially for distro kernels
<mps> select size which fits all use cases is practically impossible
<apritzel> yeah, don't envy the distros on that ;-)
<mps> Greg Kroa-Hartman advised me to disable it on distro kernels
<apritzel> is that a practical problem for A64? The boards should have enough DRAM to cope with some decent sized CMA?
<mps> but some drm drivers doesn't work without CMA, etnaviv, sun4i and maybe yet another which I can't test
<apritzel> so the arm64 defconfig sets that to 32MB, if in doubt I would go with that
<mps> no, it is not problem for A64 if the kernel is built for it specicaly
<mps> specifically*
<apritzel> I meant for any arm64 board, really
<mps> yes, 32MB is default but some big machines needs it to be at least 64MB, ampere altra for example
<apritzel> I see
<mps> and it is not needed at all on ampere altra
<apritzel> was just wondering ...
codekipper has joined #linux-sunxi
<mps> on all my arm64 machines drivers works without CMA, but one alpine users asked me yesterday to enable it because of sun4i drm. so I'm in doubt
<apritzel> what's the problem with the Altra, specifically? I see Ubuntu is using 32MB, but it seems to work fine here
<mps> maybe they added different size on kernel cmdline
<mps> apritzel: here is copy of my talk with gregkh https://gitlab.alpinelinux.org/alpine/aports/-/issues/14492
<apritzel> mps, thanks, I see
<apritzel> so I would assume that Alpine is not so much focused on the big irons, is it? But rather those smaller boards?
<apritzel> I see those CMA messages in the Altra dmesg here as well, but that doesn't seem to be fatal?
<mps> alpine runs everywhere from big servers to really small boards
<apritzel> I understand that ;-)
<mps> it is not fatal ofc
<mps> problem is to find default size compromise
<apritzel> if you ask me: I'd go with the defconfig default as set by the arch maintainers: 32MB. Especially if you have the choice between some mildly annoying messages on a fancy machine with 5 users ...
<apritzel> ... and missing a core feature for those hundreds of users with affordable boards
<mps> apritzel: thanks for advice, I'm thinking to do that
<mps> "one size fits all, but doesn't fit anywhere well"
<apritzel> indeed
<apritzel> I guess it's actually more a question of the memory size, isn't it?
hlauer has quit [Ping timeout: 480 seconds]
<mps> I think so
<apritzel> my box here says 512G, I wonder if that's what lets the kernel allocate bigger buffers?
<mps> heh, I'm not employed by ARM Inc.
<apritzel> or maybe it's more a question of the number of cores?
<mps> though I have account on ampere altra with 512GB ram
bauen1_ has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
<apritzel> yeah, I don't own this thing either, but can login ;-)
Danct12 has joined #linux-sunxi
grming has quit [Ping timeout: 480 seconds]
bauen1_ has quit [Ping timeout: 480 seconds]
codekipper has quit [Ping timeout: 480 seconds]
vagrantc has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
bauen1 has joined #linux-sunxi
<jernej> mps: CMA is needed on AW SoCs without IOMMU (either not in SoC or no driver available) for display & multimedia.
<jernej> so DRM, video decoding, deinterlacing, rotation core, etc.
<jakllsch> CMA? contiguious physical memory?
<jernej> possibly ISP (camera input too)
<jernej> yes
<jernej> If you want to support standard video decoding (1080p), assing around 128 MiB to CMA
<jernej> for 4k HEVC decoding, you need more like 320 MiB
<mps> jernej: yes, I know that it is needed for some SoCs but don't have list of such SoCs
<jernej> currently only SoC with IOMMU supported is H6
<jernej> and soon D1
<jernej> for all others, including A64, CMA is mandatory for display & multimedia
<mps> jakllsch: contiguous memory allocations
<jernej> H616 has IOMMU, but it's not implemented
<mps> jernej: I saw iommu in dmesg output on A64, but not sure it works
<jernej> there is no iommu in A64
<mps> so this is just kernel trying to find it
<jernej> rule of thumb - if there are "iommus" properties in DT, there is IOMMU, otherwise it's not
<mps> jernej: this is good to remember
<jakllsch> jernej: well, at least assuming the DT is an accurate description of the HW
<mps> grep show iommu only in arch/arm64/boot/dts/allwinner/sun50i-h6.dtsi
<jernej> mps: correct
<jernej> as I said
<mps> good, thank for info
<mps> thanks*
<jernej> D1 and H616 will gain support for it, but most currently supported SoCs simply don't have it
<jernej> jakllsch: I don't see how this would be inaccurate. Either there is iommu node ind DT and it's used or not (either physically not present or just unimplemented).
<jakllsch> depends if you're asking the question "does it work" or "could it work"
<jernej> if you ask question "could it work" then none of the sunxi DTs is accurate, because there are some peripherals which nobody implemented due to niche use
<jernej> but they are accurate for checking what's implemented
<gamiee> jernej: this reminds me that A64 have MPEG-TS core heh
<jernej> exactly what
<jernej> I had in mind
<gamiee> I wonder if anyone actually used it for DVB-T/S receiver. Also, there is SIM peripheral, probably for DVB-S ?
<jernej> I think I heard for one board where MPEG-TS was used
<gamiee> heh, also, I would understand to have this on H-series, or on "H64", but on A64, which targets to tablets... D:
bauen1_ has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
<DarkNeutrino> jernej: a question. Does the hdmi driver you have in your repo work ? I tried it on my device here but i got no output from it.
<gamiee> DarkNeutrino: is it related to A64 or similar SOC?
<DarkNeutrino> H616
<jernej> Yes, it works on T95 and OrangePi Zero2 boards that I have and tested
<DarkNeutrino> Dang. Is there anything i need to add to boot params or such other then the DT changes?
<jernej> possible reasons why it doesn't work for you are missed drivers and/or too small CMA
JohnDoe_71Rus has quit []
<DarkNeutrino> Which drivers?
<DarkNeutrino> Cause the hdmi pipeline and DE is there
<jernej> de clocks
<jernej> I don't have config anymore
<jernej> anyway, you could enable drm.debug=0x1e and check what dmesg says
<DarkNeutrino> Good idea :)
<jernej> you can post it here and I can check for clues
<DarkNeutrino> Sure thing. Will finish making some food and will get on that :)
vagrantc has quit [Quit: leaving]
apritzel has joined #linux-sunxi
megi has quit [Quit: WeeChat 3.6]
megi has joined #linux-sunxi
hlauer has joined #linux-sunxi
vagrantc has joined #linux-sunxi
bauen1 has joined #linux-sunxi
bauen1_ has quit [Ping timeout: 480 seconds]
hlauer has quit [Ping timeout: 480 seconds]
vagrantc has quit [Quit: leaving]