01:03 sgm_: incredible that I lived to the times where amdgpu drivers are at the current state
01:03 sgm_: fglrx experience was horrid
01:04 sgm_: I've just slept upon last 10-yrs of amdgpu development. Anyone would be so kind to tell me how this happened? ;-)
01:04 sgm_: you people are incredible
01:05 sgm_: I remember times when it was like "get nvidia, they at least have a blob driver that 'works'"
02:47 cheako: sgm_: Both drivers work until they don't, that includes the Intel ones. What's missing is for there to be any remedy for when the drivers don't work.
16:34 adavy: Hi, has anyone an idea of where to start to look for https://gitlab.freedesktop.org/mesa/mesa/-/issues/8716 ?
16:34 adavy: I'm not sure why there is z fighting that is fixed when hyperz is disabled
16:35 adavy: is there a possible path where somehow depth is not cleared if some conditions are violated ?
16:57 adavy: In the case of nine, fb->layers and fb->samples were never set. So before https://gitlab.freedesktop.org/mesa/mesa/-/commit/bea0897878e0188b7cc34d33c0ab54a2e160ed4a - which did cause the regression - we always hit the fb_too_small condition
16:57 adavy: setting them doesn't help though
22:46 Lyude: agd5f_: btw, do you guys have someone looking at this? https://lore.kernel.org/dri-devel/fed2a3f9111713cc619cbd54d7c1be0987c7da7b.camel@kernel.org/
22:46 Lyude: (would have poked hwentlan but they don't seem to be around atm?)
23:14 Lyude: agd5f_: nevermind, looks like that person is actually looking at it already whoops (at least that's the impression i get)