simon-perretta-img_ has quit [Remote host closed the connection]
simon-perretta-img_ has joined #powervr
simon-perretta-img__ has joined #powervr
simon-perretta-img has quit [Ping timeout: 480 seconds]
simon-perretta-img_ has quit [Ping timeout: 480 seconds]
guptaxpn8 has quit [Read error: Connection reset by peer]
guptaxpn8 has joined #powervr
Dinolek[m] has joined #powervr
simon-perretta-img__ has quit []
simon-perretta-img has joined #powervr
jluthra has quit []
<enunes>
and these ones get to run as well with a similar patch to my very old one https://ernunes.fedorapeople.org/bxs2.png (and actually also one small fix to weston vulkan)
<paulk>
so this is using a weston vulkan backend with gbm and weston vulkan wsi for apps?
<paulk>
wayland vulkan wsi for apps*
<simon-perretta-img>
Very nice! This also confirms that the generated store/load state shaders work properly haha, I wasn't super confident testing it with d2d, but if two vk apps are running at once then... :D
<enunes>
yes for vulkan wsi and weston vulkan, but no for gbm, the trick is that unfortunately gbm still goes through gallium which would for now require zink, so this is actually vulkan renderer with the output result captured over vnc with weston vnc backend
<enunes>
(more details and hopefully not too much rant on my XDC talk :)
<MoeIcenowy>
I think gbm won't work here
<MoeIcenowy>
dumb gem + vulkan import dma seems to be needed now
<paulk>
enunes: thanks for the clarification! so no dma-buf import at the moment
<paulk>
looking forward to your talk too :)
<enunes>
paulk: it does dmabuf import, if for example you export a vulkan buffer and import it, or import something else, just no gbm allocations
<enunes>
the vulkan wsi clients are dmabuf clients
<paulk>
oh cool :)
<MoeIcenowy>
I think currently the compositor needs to do what gallium renderonly code does when allocating gbm buffer