<yawniek>
hmm if i could find a correct cable i might even be able to attach the display without controller to the rock64pro, that would be interesting
<Lyude>
working on getting it to run es2_gears on this vim2 now
<Lyude>
(also chewitt ^ :)
<hanetzer>
hehe, force pushes :P
BenG83 has quit [Remote host closed the connection]
BenG83 has joined #panfrost
<yawniek>
hmm seems the eDP connector is 30pin 2 lanes only. no UHD
TheCycoONE has quit [Ping timeout: 244 seconds]
tlwoerner has quit [Ping timeout: 268 seconds]
TheCycoONE has joined #panfrost
tlwoerner has joined #panfrost
tlwoerner has quit [Ping timeout: 252 seconds]
<Lyude>
alyssa_: quick q-wasn't there something else you had to do at XDC to get the gears to come up onscreen
<Lyude>
like, modify a size somewhere
<HdkR>
There's a hardcoded backbuffer dimensions
<Lyude>
any idea where?
<Lyude>
i realized most times I launch the gears they don't actually hit that error
<HdkR>
_screen maybe?
<Lyude>
so i want to make sure it's not actually secretely rendering things
<Lyude>
oh! I think I mayyyyyy have found it?
<Lyude>
doesn't seem to have made a change
<Lyude>
welp, was worth a shot
<Lyude>
also apparently mali_kbase just dies when a job dies
<Lyude>
nice
<Lyude>
chewitt: actually-do you know if anyone's tested GPU resets on the vim2?
<Lyude>
if that's not working it would explain why the driver dies after one invalid job
<alyssa_>
"also apparently mali_kbase just dies when a job dies" Huh? My kernel driver is resilient
<alyssa_>
Sounds like a kernel regression
<chewitt>
@Lyude: not sure what you mean, but there's a working Armbian (using Xorg) image so I guess most functions have been road-tested
tlwoerner has joined #panfrost
<urjaman>
chewitt: a gpu reset is not something that would normally happen tho... atleast shouldnt, but likely would eventually given life :P (that is it is error recovery stuff)
<urjaman>
(also, does the image like have the ARM gpu blobs?)
<Lyude>
urjaman: i mean
<Lyude>
it is if you're writing a driver
<Lyude>
;)
<Lyude>
which is exactly why I I'd like to try to get such a thing to work, rebooting this board is slow
<urjaman>
yeah it's very normal in our case, but i wouldnt expect it to be quickly tested by running random userspace with xorg ... even if it had the ARM libmali hooked up
<urjaman>
since like when is the last time a normal person ran a GLES thing in Xorg :P
<Lyude>
oh-that's what you meant, whoops
<Lyude>
yeah i agree
<Lyude>
the driver does have an error injection thing, no idea if you need the ddk to make that useful though
<alyssa_>
Lyude: the code for error injection is missing, at least from my sourcr tree
<alyssa_>
like, you can't compile the kernel with support enabled
chewitt has quit [Ping timeout: 240 seconds]
cwabbott has quit [Quit: cwabbott]
cwabbott has joined #panfrost
cwabbott has quit [Client Quit]
cwabbott has joined #panfrost
<Lyude>
alyssa_: interesting
BenG83 has quit [Ping timeout: 272 seconds]
BenG83 has joined #panfrost
BenG83 has quit [Remote host closed the connection]
BenG83 has joined #panfrost
<alyssa_>
Gah this linked listification patch set should not be so hard :P
BenG83 has quit [Ping timeout: 252 seconds]
BenG83 has joined #panfrost
BenG83 has quit [Read error: Connection reset by peer]
<Lyude>
reason being the headers from arm's drivers aren't anywhere near the mess they were when we started panfrost; they don't do that weird thing with the 32/64 bit pointers and all that other junk
<Lyude>
and it's going to be easier for us to maintain
<Lyude>
granted-I don't think we are planning on maintaining it for all that long
anarsoul|2 has joined #panfrost
<alyssa_>
Lyude: Alright..
<Lyude>
yeah, the only real oddness nowadays with the headers is the annoying naming scheme they have
<Lyude>
but the idea is that in the future we can just update to the newest uapi by just copying in new headers, and everything should still work
yawniek has quit [Quit: WeeChat 2.3]
afaerber has quit [Quit: Leaving]
unoccupied has quit [Ping timeout: 260 seconds]
unoccupied has joined #panfrost
<alyssa_>
"should"
unoccupied has quit [Ping timeout: 252 seconds]
unoccupied has joined #panfrost
paulk-leonov has quit [Ping timeout: 246 seconds]
paulk-leonov has joined #panfrost
<Lyude>
alyssa_: yep
<Lyude>
"should"...
<Lyude>
either way-even if it doesn't, it's still going to be easier to work with
adjtm_ has joined #panfrost
alyssa_ has quit [Quit: ZNC 1.7.1+deb2 - https://znc.in]