<chewitt>
last count was around 45 boards .. only the VIM4 will be out-of-bounds for now
<f_>
chewitt: nice!
<chewitt>
Mrs H would like me to reduce the number a little
<f_>
Who's Mrs H?
<f_>
anyway, how many gxbb + gxl boards?
<chewitt>
Mrs Hewitt (wife)
<f_>
ah
<f_>
but yeah plenty of boards to test on!
<chewitt>
I agree with her .. but I also have the hoarding gene so I struggle to throw stuff out
cyrozap has quit [Quit: ZNC 1.8.2+deb3.1 - https://znc.in]
<chewitt>
I don't have many GXBB boards, but quite a few GXL ones
<f_>
Sure
<chewitt>
and a few GXM too
<f_>
You got a WP2 and Hub
<f_>
SPL should work out of the box on WP2
<chewitt>
also C2
<f_>
C2 too
<f_>
ah well no
<f_>
Still have to finish DRAM init lol
<f_>
but if you want to get it working on C2 DRAM init shouldn't need much mods
<chewitt>
LE has lots of C2 users so it would be handy..
cyrozap has joined #linux-amlogic
<f_>
I want to upstream this eventually
<f_>
with complete DRAM init and gxl support
<f_>
So don't worry :)
cyrozap has quit [Remote host closed the connection]
cyrozap has joined #linux-amlogic
ldevulder has quit [Quit: Leaving]
<f_>
chewitt: it boots linux, yes, but I'd say until this goes upstream consider it a WIP still
<chewitt>
sure
<f_>
At least until I create a for-master branch
<chewitt>
btw, one feature that would be _really_ nice to have is a method for setting the IR remote wake code for boxes
<chewitt>
this is normally hardcoded/hacked in bl301
<f_>
Probably won't go in SPL
<f_>
SPL is only ran when the box powers on
<lvrp16>
chewitt: this is already doable, just have the IR read the wake code from a certain part of memory
<f_>
and when you turn it off, and back on, SPL is reran again as if you just did a cold boot
<lvrp16>
I don't know the Linux wake system well enough to do this myself
<f_>
But BL30 will go in my TODO list to reverse-engineer one day
<chewitt>
as I understand it, it's more for suspend/resume
<f_>
There should be some CPU init code there as well
<chewitt>
so box cold-boots, the wake value is setup, then something is watching for it still once the box is placed into a suspend (or low-power) state
<chewitt>
to resume/wake is always a warm-boot event, not a reset + cold-boot
<f_>
SPL isn't aware of all that
<chewitt>
right .. but for future bl30 reversing magic :)
<f_>
:)
<f_>
Already came up with a name for a perhaps future SCP firmware
<chewitt>
one of the current 'features' in the upstream u-boot using LE images is that 'shutdown' on Hub/WP2 results in a restart, so you can't power-off the box without pulling power
<f_>
It's ridiculous, just like the other names I came up with before
<f_>
chewitt: Ahh that
<f_>
It doesn't always result in a restart
<chewitt>
it's a little beyond my knowledge to diagnose
<f_>
happens on my box too but if you leave it on for a long time it can power off
<f_>
I think it also happens on lepotato w/ pmOS
<f_>
but pmOS uses outdated binaries
<lvrp16>
chewitt: this is already fixed for potato.
<chewitt>
is there a link/repo?
<lvrp16>
We got a new bl30 from Amlogic that implements PSCI.
<chewitt>
ahh
<f_>
Yeah seems like BL30 was pretty buggy back then
<chewitt>
usable with a gxbb board?
<f_>
I think it's like the vendor BSP
<lvrp16>
No
<chewitt>
not the solution I'm looking for then :)
<lvrp16>
gxl only.
<chewitt>
gxm?
<f_>
it works for the intended use case, but when you're doing something else it breaks spectacularly :)
<lvrp16>
Maybe, never tested.
mripard has quit [Remote host closed the connection]
jacobk has quit [Ping timeout: 246 seconds]
cyrozap has quit [Quit: ZNC 1.8.2+deb3.1 - https://znc.in]
cyrozap has joined #linux-amlogic
cyrozap has quit [Quit: ZNC 1.8.2+deb3.1 - https://znc.in]
chewitt has quit [Read error: Connection reset by peer]
chewitt_ has joined #linux-amlogic
ungeskriptet9 has joined #linux-amlogic
LucasTanure has quit [Quit: Leaving]
ungeskriptet has quit [Ping timeout: 244 seconds]
ungeskriptet9 is now known as ungeskriptet
vagrantc has joined #linux-amlogic
cyrozap has joined #linux-amlogic
jacobk has joined #linux-amlogic
jacobk has quit [Ping timeout: 240 seconds]
<f_>
Have no idea why but trying to run the scp firmware fails half of the time
cyrozap has quit [Quit: ZNC 1.8.2+deb3.1 - https://znc.in]
<f_>
jokes aside, postmarketOS has been running on my set-top box for a while and so far hasn't crashed
<f_>
Ahh yesterday
<f_>
I misread, sorry
<f_>
but yeah it boots linux now
<f_>
Am just cleaning up that mess
<Kwiboo>
hade some issues with bl30 and bl301 yesterday because of sizes diff, and .text/.data missaligned, some bl31 will have 40kb/13kb, others 41kb/12kb or 44kb/9kb, wrong combo will not work correctly
<f_>
hm
<f_>
you mean s/bl31/bl30/?
<f_>
Kwiboo: BTW have you modified TF-A BL31?
luka177 has joined #linux-amlogic
<Kwiboo>
also there was a commit in amlogic bl33 repo that changed some user task id, they where off by one, not sure if they changed in bl30 or just was wrong from start, so may be the thread loop for secure/high/low fails
<f_>
hm
<Kwiboo>
re bl31: only what I wrote earlier: <Kwiboo> bl31 for gxbb should probably switch to use bl31_params_parse_helper(arg0, NULL, &bl33_image_ep_info), instead of redefine gxbb_bl31_param that is an exact copy of the v1 params struct
<f_>
ah ok
<f_>
Should perhaps send a patch to arm once we're done
<Kwiboo>
maybe that was required for it to work at all?, did not look closer only that it looked like it could use existing helper and changed it to use it
<Kwiboo>
Ahh, missed that check, possible I had that issue, but I initially run a release build, only tested a debug build after it got stuck because of missing scp :-)
<Kwiboo>
Yeah, nice to have it working with only one closed source blob (+ one hard to build bl301.bin)
<Kwiboo>
f_: Have you tested anything on gxl yet?, I am interested to know if it enters SPL at EL1 och EL3
<Kwiboo>
seems my c4/sm1 enters SPL/bl2 at EL3 instead of EL1
<Kwiboo>
btw, BINMAN_INDIRS= is a nice build param to use, put all external blobs in one device specific folder (amlogic-boot-fip) and binman will pick up correct blobs with BINMAN_INDIRS=path/to/amlogic-boot-fip/<device>
<Kwiboo>
so there is a mess with bl30/bl301 combo, a possible central repo with scp_task code for each board would be nice, and update all boards to use same bl30, for c2 I picked the one for p200 in amlogic-boot-fip
<Kwiboo>
some newer builds would print something about unsupported thermal when bl31 calls aml_thermal_unknown()