jacobk has quit [Quit: No Ping reply in 180 seconds.]
vagrantc has quit [Quit: leaving]
jacobk has joined #linux-amlogic
jacobk has quit [Quit: No Ping reply in 180 seconds.]
jacobk has joined #linux-amlogic
jacobk has quit [Ping timeout: 240 seconds]
Daanct12 has quit [Quit: WeeChat 4.0.4]
Daanct12 has joined #linux-amlogic
Daanct12 has quit [Client Quit]
Daanct12 has joined #linux-amlogic
jacobk has joined #linux-amlogic
jacobk has quit [Ping timeout: 240 seconds]
jacobk has joined #linux-amlogic
Daanct12 has quit [Read error: Connection reset by peer]
Daanct12 has joined #linux-amlogic
naoki has quit [Quit: naoki]
Daanct12 has quit [Quit: WeeChat 4.0.4]
Daanct12 has joined #linux-amlogic
Daanct12 has quit [Quit: WeeChat 4.0.4]
Daanct12 has joined #linux-amlogic
mripard has joined #linux-amlogic
camus1 has joined #linux-amlogic
camus has quit [Ping timeout: 240 seconds]
camus1 is now known as camus
Daanct12 has quit [Quit: WeeChat 4.0.4]
khilman has quit [Server closed connection]
khilman has joined #linux-amlogic
buzzmarshall has joined #linux-amlogic
f_ has joined #linux-amlogic
f_[xmpp] has quit [Ping timeout: 240 seconds]
<f_>
hi
<f_>
Disabled MMC support in TPL for now
<f_>
Kwiboo: I don't see anything wrong in my DRAM init code
<f_>
so it could probably be LTO causing bugs
<f_>
anyway, I'll try your suggestion of having a TPL which would do DRAM init and load SPL
vagrantc has joined #linux-amlogic
<f_>
TPL: reading 0x1000000 returns -784017578
<f_>
Error: test failed. pattern 0x1 != data
<f_>
Kwiboo: seems like it's my code after all
dlan has quit [Read error: Connection reset by peer]
<Kwiboo>
f_: strange, maybe some value that is configured is somehow wrong? I also tried to go over everything and it does seem to match the bl2 code and did not find anything sticking out
<Kwiboo>
maybe we need to do a reg dump using old working atf bl2 code and compare to a reg dump after using u-boot spl dram init code
<f_>
Funny fact; because we now have access to a (somewhat incomplete) decompilation of newer BL2 we could actually modify it
<f_>
Shouldn't be too hard to dump the regs in new BL2
<f_>
Kwiboo: Oh wait you meant my old port of their old code?
<f_>
It resets multiple times on my box because of a wrong channel setting (RANK01_SAME v.s. RANK01_DIFF) but otherwise works, yes
<f_>
Other regs should be fine. It's really the ones modified by BL2 in pre_init and everything else in the ACS
<f_>
But let's try.
<f_>
Kwiboo: Have you tested that code?
<Kwiboo>
I will give your new commit a try, tried to compare all values with hardkernel bl2 decompile and all non-computed values was at least "correct"
<lvrp16>
speaking of headbangs, we have been debugging manufacturing test USB issues for 3 days and it turns out sandisk controller craps out while 5 different other vendors do not. it seems the only American company that can get flash right is Intel.
<f_>
oh lol
<lvrp16>
every Crucial SSD (12) I've ever owned has died
<f_>
Samsung apparently make good and durable SSDs
<lvrp16>
25% of my SanDisk MicroSD cards have failed
<f_>
So far haven't got any problems with their SSD in my laptop
<lvrp16>
Samsung is solid.
<lvrp16>
I had 2 MicroSD cards out of 50 die. But I'm not exactly the average user.
<f_>
In general, never trust MicroSD cards for sensitive stuff.
<f_>
Same for USBs I'd say.
<lvrp16>
I don't trust flash. I only trust Google Cloud :D
<f_>
What if Google Cloud ran on flash
<lvrp16>
I've lost way too much.
<lvrp16>
5N redudancy
<f_>
Use cottonwood as NAS and make automatic backups :D
f_[xmpp] has joined #linux-amlogic
<f_>
lvrp16: In the meantime I recently had to replace a dead Samsung HDD
<lvrp16>
we have another board for NAS coming next year with ECC.
<lvrp16>
I wouldn't do NAS without ECC
<f_>
Nice.
<f_>
Will it be based on cottonwood?
<f_>
Kwiboo: Even that last commit which syncs the code with BL2 doesn't work
<f_>
so doubt it would work on your C2
<f_>
Time for me to dump every single PCTL/PUB register
<Kwiboo>
Tried to decompile u-boot spl and looks like it inject a dmb before each writel, something the bl2 will not do
<f_>
hm
<Kwiboo>
cache should be disabled so guess it should not matter, but at least something that is different