<libv>
but perhaps you can recycle some existing kms test app, and find some way in which to make this happen
<libv>
but beware, if you find a way, it might be extremely stupid
uis has quit [Client Quit]
<libv>
like what fun i had when trying to disable planes
uis has joined #linux-sunxi
<plaes>
well, I would just like to same as `xset dpms force off`
<libv>
if only my hiring at intel had not been blocked, and i had driven the team manager to give me and ville the time we needed, then ville would have done more than just plant to seed of atomic, and we would've had something an order of magnitude saner than what we have now
<libv>
they would control the h and v sync lines on the vga connector, and depending on which was off, the monitor either went into sleep mode or in full powerdown
<libv>
and even though Xfree86-4.0 was supposed to have been rid of vga (spec) specifics... for ages, dpms was still poking those two vga registers directly
matthias_bgg has joined #linux-sunxi
<libv>
it only got fixed with randr1.2, where keithp took my modesetting ideas (the ones he shot down a year earlier), and ignored the lessons learned with xfree86-4.0 (like the use of helpers, which only got rediscovered with kms atomic), and then badly implemented something where one was not able to tell why modes were culled or where modes came from during validation
<libv>
in this round, the vga stuff got removed, but was only solved on the driver side
<libv>
vga sync stuff
uis has quit [Client Quit]
<libv>
from then on, a modeset would occur, and then raw memory content would be shown before X had had a chance to draw anything into it
uis has joined #linux-sunxi
<libv>
and even though keithp, the great master of the xorg protocol, who never cared about drivers or driver infrastructure in his life, was driving this randr1.2, it never got fixed
<libv>
i had a suse manager panick and ask why his lcd on his laptop was being destroyed in Q3 2007, because there was old memory content on his laptop lcd
<libv>
days before a suse release
<libv>
the fix, by adam jackson, just null the memory content before doing anything...
<libv>
so much for history
<libv>
you would probably need to go through the whole song and dance of kms, and then find some convoluted way of doing things...
<libv>
which then probably would die with your application dying
<apritzel>
does anyone here have one of those "X96 Mate" TV boxes (w/ H616)? jernej?
<apritzel>
I have one here now and was just wondering if anyone has already some DT for that, so I don't need to start from scratch ...
<apritzel>
the Android kernel thankfully provides regulator_summary in debugfs, but the relation between devices and AXP rails probably requires trial and error, I guess
<jernej>
apritzel: I have T95 H616 box
<apritzel>
right, saw that older branch in your github ...
<jernej>
and yes, regulators are pain
<apritzel>
I guess they are not that much different
<apritzel>
well, at least I know the voltages
<jernej>
apritzel: did you figure out USB?
<apritzel>
jernej: I think so, smaeul's hint of USB2 reset being required everywhere looks good so far
<jernej>
that's from boot rom, right?
<apritzel>
but then I said that several times during the last posts ;-)
<apritzel>
jernej: yes, I think. Not sure also from the BSP code, maybe?
<apritzel>
the neat thing is that it's just a DT fix
<apritzel>
jernej: can you say whether the RTC clocks are different between the H6 and H616?
<apritzel>
I am using a separate compatible now, but just use the H6 recipe for the clocks part
<jernej>
I forgot already, I have to check manuals
vagrantc has joined #linux-sunxi
<apritzel>
jernej: no worries, it's not urgent ...
<apritzel>
I just always get confused over those RTC clocks ...
<jernej>
apritzel: there are some changes - check register 0x60 description in both manuals
<apritzel>
right, I see
<jernej>
although I think this can be handled inside rtc driver based on the fact if external 32k oscillator is defined in DT or not
<apritzel>
did anyone hear anything from the Allwinner guys doing the A100 support? They seemed to have vanished after the first round of patches, or at least after the first post of the 2nd round?
<apritzel>
a lot of H616 IP changes seems to have actually been introduced with the A100 already (USB PHY, for instance), so I wonder if this could be tackled in one go
rzerres has joined #linux-sunxi
victhor has quit [Ping timeout: 246 seconds]
victhor has joined #linux-sunxi
<jernej>
apritzel: do you have any solution for copying protected regions from DT changed by TF-A to loaded from SD in U-Boot for H616?
<apritzel>
jernej: ah, right, that one
<apritzel>
I looked it up the other day, IIRC U-Boot can enter protected regions, if it know about them itself
<apritzel>
so we would need to teach U-Boot to parse /reserved-memory and enter that into its own data structure
matthewcroughan has quit [Quit: No Ping reply in 180 seconds.]
<apritzel>
although I think I booted without problems lately, while loading the DTB from TFTP, let me check that ...
matthewcroughan has joined #linux-sunxi
rzerres has quit [Ping timeout: 240 seconds]
<jernej>
yeah, but probably reset doesn't work
<jernej>
basically anything that needs TF-A
lucas_ has joined #linux-sunxi
_0x5eb_ has quit [Ping timeout: 268 seconds]
_0x5eb_ has joined #linux-sunxi
<apritzel>
... for instance SMP bringup. Turns out I still had the /reserved-memory node in my external DTB, if I remove that, it indeed hangs during Linux boot
<apritzel>
mmh, bdinfo lists the region from the TF-A patched DT already, so there must be more to it
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
lucas_ has quit [Ping timeout: 260 seconds]
specing has joined #linux-sunxi
netlynx has quit [Quit: Ex-Chat]
rzerres has joined #linux-sunxi
elros1 has quit [Remote host closed the connection]
aballier has joined #linux-sunxi
aballier has joined #linux-sunxi
lucas_ has joined #linux-sunxi
lucas_ has quit [Ping timeout: 246 seconds]
ElBarto has quit [Quit: leaving]
ElBarto has joined #linux-sunxi
jbrown has quit [Ping timeout: 260 seconds]
lucas_ has joined #linux-sunxi
<MoeIcenowy>
apritzel: U-Boot DT is patched but Linux DT is not?
<apritzel>
MoeIcenowy: Trusted Firmware finds the DT that is attached to the U-Boot binary, and patches that
<apritzel>
so U-Boot proper sees the patched version
<apritzel>
but if you load a new DTB in U-Boot proper, from some storage, then it's too late for TF-A to patch that
<apritzel>
but the reserved memory regions should be preserved
<apritzel>
U-Boot scans its DT for reserved-memory regions, and it looks like the EFI loaders translates that into the EFI memory map, as a reserved region
<apritzel>
but for booti it seems to miss that
<apritzel>
there is fdtdec_add_reserved_memory(), which has all the code, but it doesn't seem to be called, AFAICT
macc24 has quit [Ping timeout: 250 seconds]
JohnDoe_71Rus has quit [Quit: KVIrc KVIrc Aria 5.0.1, revision: 5.0.1+git-7433-0df9f22f2, build type: debug, sources date: 20160102, built on: 2019-12-08 19:19:20 UTC 5.0.1+git-7433-0df9f22f2 http://www.kvirc.net/]