Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
Schimsalabim has quit [Read error: Connection reset by peer]
linkmauve has left #linux-sunxi [#linux-sunxi]
linkmauve has joined #linux-sunxi
Schimsalabim has joined #linux-sunxi
vagrantc has quit [Quit: leaving]
bauen1 has joined #linux-sunxi
ftg has quit [Read error: Connection reset by peer]
apritzel has quit [Ping timeout: 480 seconds]
<MasterR3C0RD>
Finally some free time; going to take a fine toothed comb through the A527 DRAM code and see if I can catch anything...
aggi has quit [Quit: connection closed.]
aggi has joined #linux-sunxi
<MasterR3C0RD>
jernej: If I were a betting man, I'd say the issues we're running into on the Cubie are related to the DRAM special case
<MasterR3C0RD>
s/DRAM/address mapping/
<MasterR3C0RD>
I mentioned it before here on the A133, but essentially it shifts around some of the column/row bits in the mapping. It was necessary to get working DRAM init on the 4GB board that was being used by someone else
<MasterR3C0RD>
Hmm wait... I might be wrong
hexdump01 has joined #linux-sunxi
hexdump0815 has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
<MasterR3C0RD>
Okay no, the special case on A527 is just for 3GiB models it seems. Strange; on the A133 there was definitely something strange with address mapping
gnarface has quit [Quit: Leaving]
gnarface has joined #linux-sunxi
<MasterR3C0RD>
Okay; I've patched together a few things and I can confirm it is 100% memory corruption. I took the DRAM test I used on the A133 (which was copied from some other repo) and threw it into the A523 DRAM code. Doesn't seem to be aliasing
<jernej>
MasterR3C0RD: have you compared register values between BSP and your code?
<jernej>
besides some delay fine tunning, it should match perfectly
<jernej>
I have recieved cubie board, so I can also start experimenting soon
<MasterR3C0RD>
I haven't yet, no; I was running into issues even getting U-Boot from the vendor image working initially, though I finally got it working (dumb PhoenixCard tool made bad SD image, as did OpenixCard initially; still can't boot into Linux)
<jernej>
well, for entering U-Boot, you don't need PhoenixCard. Last time it worked that I just concatenated GPT, boot0 and U-Boot parts together and flashed that.
<MasterR3C0RD>
Ahhh... wait... it appears that the issues I was running into were entirely my fault
<MasterR3C0RD>
I had reset the changes I made for MR14 so I could test my Avaota, and forgot to switch it back
<MasterR3C0RD>
No more memory test errors
<jernej>
great! what's the next issue? :)
Schimsalabim has quit [Read error: Connection reset by peer]
<MasterR3C0RD>
Not getting past "trying to boot from MMC1"; investigating...
<jernej>
are you booting from SD or FEL?
Schimsalabim has joined #linux-sunxi
<MasterR3C0RD>
SD card, FEL doesn't work at all
<jernej>
well, it works if you use 32-bit SPL and then for jumping to TF-A you use reset64 command
aggi_ has joined #linux-sunxi
aggi has quit [Remote host closed the connection]
<MasterR3C0RD>
I'm running into some issues compiling ATF now actually... not entirely sure why
<MasterR3C0RD>
Figured it out; environment variables
<MasterR3C0RD>
It's been a long week LOL
apritzel has joined #linux-sunxi
chewitt has joined #linux-sunxi
dsimic is now known as Guest6942
dsimic has joined #linux-sunxi
Guest6942 has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
colinsane has joined #linux-sunxi
colinsane has quit [Ping timeout: 480 seconds]
colinsane has joined #linux-sunxi
<jernej>
which version of cubie a5e board did you receive? I have v1.1
apritzel has joined #linux-sunxi
apritzel has quit []
apritzel has joined #linux-sunxi
<apritzel>
jernej: I have v1.1 as well. I "diff'ed" both versions of the schematics in gimp, there are not many differences:
<apritzel>
1) PJ25 is EEPROM write detect (though I am not sure there is an EEPROM even)?
<apritzel>
2) the SD card detect is different - maybe it's the push/push vs. push/pull difference? Might create a different polarity for the CD pin then: active low vs. active high
<apritzel>
3) R75 on page 14, on the VBUS_5V to 5V_USB regulator, is 6.8K vs 3.6K
colinsane has quit [Ping timeout: 480 seconds]
<apritzel>
(EEPROM write *protect above, of course)
<apritzel>
those are the only differences, so not much to worry about, I think. If the SD CD pin is indeed different polarity, it might become a slightly annoying DT problem, though
<jernej>
apritzel: ah, thanks, I was just about to start diff'ing
<jernej>
well, it wouldn't be the first time to have multiple DTs for different revisions
<jernej>
EEPROM can be very unassuming, so it's entirely possible that's there
<apritzel>
yeah, but they would differ in just one bit, literally, that's sounds annoying. Maybe we can somehow detect it, and patch the DT at runtime
<apritzel>
jernej: are you in the U-Boot shell on the Cubie?
<jernej>
no, I haven't had time to start playing with it
<apritzel>
no worries, I just think MasterR3C0RD reported something like this, and it would help to check those things (SD CD polarity, EEPROM)
<apritzel>
I am making slow, but steady progress with cleaning up the A523 U-Boot patches. Rewrote most of the H6 PLL clock code yesterday, to accommodate all of H6, H616, T113/D1, A523, aligning with the manual
<apritzel>
jernej: your PLL code in your "avaota WIP" patch was very helpful, though I ended up changing it quite a bit, to include H6/H616
<jernej>
oh, that's nice
<jernej>
apritzel: do you know what frequency should DSU have in comparison to little or big cores?
<apritzel>
and I got rid of "struct sunxi_ccm_reg" entirely, which has always bugged me. At least for H6 and friends
<apritzel>
not from the top of my head, but I can find that out, I guess there might be hints in the A55 TRM
<apritzel>
jernej: did you look at what the BSP does there?
<jernej>
not yet, no
<jernej>
I mean, DSU clock init code is similar to boot0
<jernej>
but I haven't checked kernel code yet
<jernej>
apritzel: SD card difference really looks like inverter
<jernej>
apritzel: now that I'm looking to schematic, it looks like there is no pull up on sd-det pin, so it has to be enabled internally
<loki666>
Thanks, should I also creates a axp717.dtsi or embed it in board DTS like it was done for h700 ? Or this completely irrelevant now
Schimsalabim has quit [Ping timeout: 480 seconds]
wingrime-ww has quit [Quit: WeeChat 4.5.1]
bauen1 has quit [Ping timeout: 480 seconds]
wingrime-ww has joined #linux-sunxi
Schimsalabim has joined #linux-sunxi
bauen1 has joined #linux-sunxi
<apritzel>
no AXP .dtsi, do it like H700 does it
<hlauer>
<apritzel> "or use ®_dldo1, which is..." <- Tried and failed ;-)
<hlauer>
But noticed mmc1 (used by the wifi interface) uses dldo1 too, and disabled that and (e)mmc2
<hlauer>
And with all other interfaces disabled, mmc0 seems to be stable so far - at least just survived a testing update. Thanks!
apritzel has quit [Ping timeout: 480 seconds]
apritzel has joined #linux-sunxi
chewitt has quit [Quit: Zzz..]
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
<apritzel>
hlauer: that sounds very odd! Does it help to just disable MMC1?
<apritzel>
and what is the actual voltage of DLDO1? The DT is very unhelpful (wrong?) here by saying "between 1.8V and 3.3V", which probably means it stays at the reset default
<MasterR3C0RD>
It'll need to be ported to work with U-Boot and SPL, but you'll certainly need to use the BSP as a reference since there's no reference manual available for the TCS
<apritzel>
what DRAM parameters did you use? the ones shown in the BSP bootlog, or the ones extracted from the image, via sunxi-fw
<loki666>
sunxi-fw
<loki666>
it seems to happens since I declared mmc2
<loki666>
but even when it detect the ram, ok it still doesn't boot from mmc2
<apritzel>
in U-Boot's drivers/mmc/sunxi_mmc.c, change the line "cfg->host_caps = MMC_MODE_8BIT;" to use MMC_MODE_4BIT instead
<apritzel>
the SPL assumes that MMC2 is eMMC, and always uses 8-bit
<apritzel>
which won't work on an SD card
<apritzel>
loki666: did you manage to extract a DT from the BSP image? Do you have a link to some image file?
<MasterR3C0RD>
loki666: That's a RAM misdetection; it happens sometimes for some reason. I'll try and port jernej's patches for RAM detection at some point to see if that improves things