anessen973383701 has quit [Read error: Connection reset by peer]
anessen973383701 has joined #linux-amlogic
paddymahoney has quit [Ping timeout: 248 seconds]
paddymahoney has joined #linux-amlogic
rockosov has quit [Ping timeout: 252 seconds]
<f_>
I have a neat little trick for you. Want to crash my u-boot spl wip stuff easily and in a reproducible way? You can do so fairly quickly in a few easy steps. Step 1: load it from USB. Step 2: there is no step 2.
<f_>
But be aware that this will work for a limited time only!
<f_>
Yes, I am trying to get DFU up and running.
<f_>
I stripped down pyamlboot to only load a binary to memory and let the bootrom jump to it
<f_>
(SPI debugging is postponed for now.)
paulk has quit [Ping timeout: 248 seconds]
<f_>
And ugh don't tell me the DRAM clock was causing the thing to be unreliable....
<f_>
I knew something was a bit off lol
<f_>
so I guess lafrite now has okay DRAM init
rockosov has joined #linux-amlogic
<f_>
lvrp16: in configs/gxl_p241_v1.h on the downstream u-boot tree, there's a `#define CONFIG_DDR4_CLK 1080`, however the BL2 logs say `DDR4 chl: Rank0 16bit @ 1008MHz`, and I also had slightly more reliable init with that
<f_>
out of date tree or?
<lvrp16>
Lollllll
<lvrp16>
I think 1080 uses looser timings and 1008 uses tighter timings.
<lvrp16>
Both should be OK since it is 1pcs DDR4.
<lvrp16>
I stopped pushing my luck with the IC lottery a while back.
<f_>
both do init but the former sometimes doesn't work
<lvrp16>
Some boards will run 1200MHz fine for days. Others, 1080. DDR3 it was 912.
<f_>
I've a 512MB lafrite FWIW
<lvrp16>
Iso timing or different timing? If iso timing, 1008 us much better I imagine.
<f_>
possible I mixed up the timings though
<f_>
¯\_(ツ)_/¯
pisen4859 has joined #linux-amlogic
linkmauve has left #linux-amlogic [#linux-amlogic]