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
bauen1 has joined #linux-sunxi
ftg has quit [Read error: Connection reset by peer]
apritzel has quit [Ping timeout: 480 seconds]
hexdump01 has joined #linux-sunxi
hexdump0815 has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
junari has joined #linux-sunxi
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
jernej has quit [Remote host closed the connection]
jernej has joined #linux-sunxi
norton has joined #linux-sunxi
anarsoul has joined #linux-sunxi
anarsoul|2 has quit [Read error: Connection reset by peer]
aat596_2- has quit []
aat596_2 has joined #linux-sunxi
kuba2k2 has joined #linux-sunxi
dsimic is now known as Guest10570
dsimic has joined #linux-sunxi
Guest10570 has quit [Ping timeout: 480 seconds]
warpme has joined #linux-sunxi
apritzel has joined #linux-sunxi
gsz has joined #linux-sunxi
Schreiber has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
kuba2k2 has quit [Ping timeout: 480 seconds]
kuba2k2 has joined #linux-sunxi
<jernej> junari: I think AW DRAM driver does that based on one configuration bit and that's the reason it's done in orangepi sources too.
<jernej> Do you have any idea how to detect those 1.5 GiB sizes? IIRC mctl_auto_detect_dram_size() detects 2 GiB, right?
<junari> jernej: now I'm trying to determine this through checking
<junari> My TVbox looks working, now I'm trying to build something heavier than openwrt for testing
<jernej> hm... any particular reason for offset-0x10?
<MoeIcenowy> junari/jernej: how is the 1.5G size organized?
<MoeIcenowy> does it forbid some of address bits to be 1 at the same time to reduce 1/4 of capacity?
<junari> I was debugging the code on a device with 1.5GB and did this for reliability to get into the last quarter of RAM
<MoeIcenowy> I remember reading some i.MX document, and seeing some bit in DW DRAM controller that enables such kind of DRAM that forbid two bits to be 11
<Jookia> that reminds me, at some point i have to submit a patch that fixes the i.mx6 u-boot DRAM calibration
<jernej> it a long time since I last worked on this, but IIRC rank, width, rows and cols are properly detected, like for 2G version, it's just that last 0.5G is unusable
<MoeIcenowy> i.MX 7Dual Applications Processor Reference Manual Rev. 1, 01/2018 pg.2304 DDRC_ADDRMAP6 BIT 31
<MoeIcenowy> (I remember H6 DDR controller should have a similar register map
<MoeIcenowy> well what will happen when accessing the last 1/4?
<MoeIcenowy> some kind of wrapping around?
<junari> In zynq register reference there are one bit in addrmap for that "Set this to 1 if there is an LPDDR3 SDRAM 6Gb or 12Gb device in use.
<junari> - 0 - non-LPDDR3 6Gb/12Gb device in use. All addresses are valid"
<junari> - 1 - LPDDR3 SDRAM 6Gb/12Gb device in use. Every address having row[14:13]==2b11 is considered as invalid
<junari> But it is not used in the original AW u-boot code
<jernej> I wonder what is the purpose of such bit. If you avoid non-existing memory region, it shouldn't make any difference
<MoeIcenowy> yes
<MoeIcenowy> and I think you must put the row address to the highest bits
<MoeIcenowy> to really prevent sending out addresses with row[14:13] == 2b11
<junari> I was doing some tweaking to match AW u-boot values
<MoeIcenowy> Reading JESD209-3C: `No memory present at addresses with R13=R14=HIGH. ACT command with R13=R14=HIGH is ignored (NOP). Write to R13=R14=HIGH is ignored (NOP)`
<MoeIcenowy> if the ACT command is ignored, what will READ command return?
<MoeIcenowy> READing from the previous address?
<junari> MoeIcenowy: Is there any way I can check this?
apritzel has joined #linux-sunxi
warpme has quit []
kuba2k2 has quit [Ping timeout: 480 seconds]
<junari> jernej: Maybe you can tell me why HDMI may not work on my devices?
<jernej> I have no idea, would need some info
<jernej> which device, which driver (code), resolution and refresh rate, clk_summary content at the minimum
kuba2k2 has joined #linux-sunxi
<jernej> and also one anecdote - I thought that there is something wrong with the driver while in reality, I had a crap cable. Once I replaced it with higher quality one, it works fine.
<junari> jernej: Could you tell me how I can diagnose the cause of problems with hdmi output? I'm using your latest h616-val branch, image output does not work with some displays on h616/h618 devices.
<junari> If I add to the cmdline video=HDMI-A-1:1920x1080@60e I get sun8i-dw-hdmi 6000000.hdmi: PHY PLL failed to lock
<junari> without that sun4i-drm display-engine: [drm] Cannot find any crtc or sizes
<junari> devices are h616 tvbox and orangepi zero3
<junari> android works with same cable and displays on my tv and monitor. Zero3 works only with spi image
<junari> If you have any thoughts, write to me even if I'm offline
<jernej> I've never seen PHY PLL lock issue
<jernej> but the reason you need to add video argument could be that EDID is not read correctly
<jernej> can you show your mainline and Android DT files?
<junari> it's android DT from TVbox, and patches that I applied to your branch
<jernej> I guess HDMI type should be set to "a", but other than that, hdmi looks to be set properly
<jernej> anyway, I have zero3 now too, so I can take a look into that
asriel has quit [Quit: Don't drink the water. They put something in it to make you forget.]
Danct12 has quit [Quit: ZNC 1.8.2 - https://znc.in]
Danct12 has joined #linux-sunxi
<jernej> junari: btw, on zero3, hdmi is powered by bldo1
<jernej> eh, sorry, schematic is a bit confusing and convoluted, aldo1 is correct one
ftg has joined #linux-sunxi
<junari> jernej: It is possible that this will not appear on your devices, since I connected zero3 to another TV (neighbor's) and it worked
<jernej> hm... I have 1080p monitor, which doesn't work with Tanix TX6 (H6) box.
<jernej> maybe you experience similar issue, just with different soc
<junari> there will be an opportunity to test the new orangepi zero3 on it :D
asriel has joined #linux-sunxi
<jernej> I have plenty of different HDMI monitors for testing purposes that I collected over the years
<jernej> but I only recently discovered that issue
<junari> Haven't you added h616 support to libreelec yet? I think the number of reports will increase after this
<jernej> it's wip
junari has quit [Ping timeout: 480 seconds]
junari has joined #linux-sunxi
ftg has quit [Ping timeout: 480 seconds]
warpme has joined #linux-sunxi
norton has quit [Quit: WeeChat 4.1.2]
<jernej> you already have my r-b tag :D
<jernej> and it was merged upstream
<jernej> oh, it's applied
<apritzel> yes, I pushed it, so that OPi-Zero3 support is complete. Also I felt guilty because I got that wrong in the initial DT submission ;-)
junari has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
ftg has joined #linux-sunxi
JohnDoe_71Rus has quit []
<jernej> don't, ethernet timing is tricky sometimes, and I'm just as guilty not checking it
<jernej> anyway, latest U-Boot master boots on my zero3 just fine and ethernet is pretty quick for netbooting. I like that
<jernej> junari: it works fine for me on zero3 too (4K@60)
<jernej> I'll test on "problematic" monitor
kuba2k2 has quit [Ping timeout: 480 seconds]
<jernej> it works there too
gamiee has joined #linux-sunxi
apritzel has joined #linux-sunxi
warpme has quit []
vagrantc has joined #linux-sunxi
kuba2k2 has joined #linux-sunxi
kuba2k2 has quit [Ping timeout: 480 seconds]
<apritzel> So does anyone have the BananaPi M4 Berry, with the H618? Anyone cares to either create a DT for mainline, or at least be able to test this? Or does anyone know who Wiki user Rayknight is?
<apritzel> junari: for an experiment: maybe you can try to ignore the 1.5G for a second, let it initialise as 2GB, then try to write from U-Boot to the gap and see what happens?
<apritzel> might need to tweak CONFIG_SUNXI_DRAM_MAX_SIZE to 1GB to prevent U-Boot from relocating itself into that gap (it uses the end of DRAM, normally)
<apritzel> so it might be as easy as to write something, and check whether you can read that back
<apritzel> in a second step figure out if this gets aliases somewhere else, we might be able to give some educated guesses where to look at
<apritzel> junari: oh, also, what TV box do you have? Something with H618 & AXP313? do you have some DT somewhere? I pushed one for mainline, can you check or compare this? Would be curious to see how similar they are
colinsane has quit [Ping timeout: 480 seconds]
vagrantc has quit [Quit: leaving]