narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - Publicly Logged on https://irclog.whitequark.org/
<Ely>
Quick question, I've been using https://github.com/akuster/meta-odroid to build poky for my C1, but I just realized it's not the ""official"" meta-meson. It is however very up to date and working quite well (at leastfor odroids). What's the reasoning for not using this as a base ?
<narmstrong>
Because it’s odroid centric
<narmstrong>
Meta-meson is amlogic centric for mainline only
<Ely>
fair enough
<narmstrong>
iin the fact we share some recipes !
<gab>
Hi, I have a question about DDR detection in the amlogic s912's bootloader: I got a box advertised as having 3GB DDR (beelink GT1 ultimate), but on serial console BL2 says "DDR3 chl: Rank0+1 @ 912Mhz - PASS" followed by "Rank0: 1024MB(auto)-2T-13 / Rank1: 1024MB(auto)-2T-13", while Rank0 should have 2048MB
<gab>
I checked the DDR chip references on the board and they look like proper DDR4 chips with correct sizes
<gab>
could there be something wrong on the box's flash, or did I just buy a box with fake DDR chips ?
<gab>
from looking at freaktab.com forums, it appears this issue is quite common with this board model
<gab>
also, when I load a 3GB device tree, it resets immediately, same thing if I try to access addresses above 0x80000000 with md.b in uboot
<gab>
another strange thing is that the kernel's /proc/meminfo shows 2840744 MemTotal, but zoneinfo says 524032 spanned pages
<gab>
(and dmesg "Memory:" line shows only 2GB)
<narmstrong>
the DDR init initialized bby the FIP fw loaded before u-boot, and is dependant on the board config
<narmstrong>
check on the PCB you have 3GB !
<gab>
I checked the DDR chips on the PCB, and they look like legit DDR4 chips: 2x K4A4G16 5WE BCRC + 2x K4A8G16 5WB BCRC
<gab>
so if I understand well, this may just be a software config mistake made by the OEM ?
<gab>
the MemTotal value looks fake though, that makes it very suspicious
<Elpaulo_m>
gab: There may be a manufacturer error. Can be an individual flashing for the defect of components. In addition, the kernel is often fixed to display the correct size.
<gab>
okay
<gab>
so is there any way to fix that through flashing ?
<gab>
and if possible without bricking the box while trying do to it ;)
<gab>
I checked the stock firmware's DDR.USB file, and it matches the one on my board's emmc
<gab>
even uboot says "DRAM: 3GiB", appears to select the 3G dtb, but when I check in /proc/device-tree after the android image has booted, linux,usable-memory is 0x7ff00000 (ie. the one from the 2g dtb)
<gab>
so I doubt this is individual flashing
<Elpaulo_m>
if you remove the power from emmc or another signal, you can experiment with booting with sd. The firmware can always be filled with USB. I have s905, but in your case I think, too, there is no problem with recovery and firmware. IMHO
yann has quit [Ping timeout: 276 seconds]
<Elpaulo_m>
gab.I'm not an expert on this platform, but it would be convenient for everyone to see a link to the full log of your system's boot.
<gab>
that dtb_read_shortcut line looks suspicious
Barada has joined #linux-amlogic
Barada has left #linux-amlogic [#linux-amlogic]
Barada has joined #linux-amlogic
afaerber has joined #linux-amlogic
dsd_ has joined #linux-amlogic
dsd_ has quit [Quit: leaving]
dsd_ has joined #linux-amlogic
Elpaulo_m2 has joined #linux-amlogic
Elpaulo_m has quit [Ping timeout: 260 seconds]
Elpaulo_m has joined #linux-amlogic
Elpaulo_m2 has quit [Ping timeout: 248 seconds]
Elpaulo_m has quit [Quit: Quit]
<dlan>
jbrunet: do you prefer I send the aoclk series along with the patch - drop the CLK_IGNORE_UNUSED flag? but queue the patch for another cycle (or whatever make it goes in *after* the DT patch which change UART_AO clk)
sputnik_ has quit [Remote host closed the connection]