16:44 cheakoirccloud: Is there any interest in the SaftyNet Layer for Vulkan I mentioned?
18:10 anarsoul: It looks like CI is broken -- all the runners for arm_build and arm_test are down
22:00 Sesse: hi. when I'm calling vaMapBuffer() (cfl if it matters, I'm doing mjpeg encoding), I can seemingly get my program into an infinite loop:
22:00 Sesse: ioctl(28, DRM_IOCTL_I915_GEM_MMAP, 0x7f9661ad96e0) = -1 EAGAIN (Resource temporarily unavailable)
22:00 Sesse: ioctl(28, DRM_IOCTL_I915_GEM_MMAP, 0x7f9661ad96e0) = -1 EAGAIN (Resource temporarily unavailable)
22:01 Sesse: ioctl(28, DRM_IOCTL_I915_GEM_MMAP, 0x7f9661ad96e0) = -1 EAGAIN (Resource temporarily unavailable)
22:01 Sesse: is there a way to have it, like... give up?
22:01 Sesse: what could cause EAGAIN anyway?
22:01 imirkin: afaik the intel va-api is done by not-this-team
22:01 Sesse: what's the best way to get into contact with... not-this-team?
22:02 imirkin: sorry, dunno. i don't think they have a public presence.
22:02 Sesse: interesting, thanks
22:02 Sesse: this is the old intel va-api driver, too, since the new one just errors out right away for me :-)
22:03 Sesse: a bumpy ride
22:04 kisak: maybe ##intel-media? (from https://01.org/linuxmedia/vaapi )
22:04 Sesse: let's see
22:04 Sesse: might very well be it
22:04 imirkin: ooh, that's new
22:05 Sesse: there are so many... #dri-devel, #intel-3d, freenode, oftc :-)
22:05 imirkin: looks like that page hasn't been updated in a bit...
22:09 Sesse: heh, as luck might have it, I stumbled upon the solution on my own
22:09 Sesse: my program was simply out of RAM
22:10 Sesse: and the ioctl responds to that by, er, not ENOMEM