ChanServ changed the topic of #lima to: Development channel for open source lima driver for ARM Mali4** GPUs - Kernel driver has landed in mainline, userspace driver is part of mesa - Logs at https://oftc.irclog.whitequark.org/lima/
Daanct12 has joined #lima
Daanct12 has quit [Read error: Connection reset by peer]
Daanct12 has joined #lima
camus has joined #lima
<anarsoul> enunes: could you remind me what's missing from your performance counters MR?
<enunes> anarsoul: need to figure out a backwards compatible solution to adding a new field to the submit struct https://gitlab.freedesktop.org/enunes/linux/-/commit/2a5d90ac4f4c97b7e25f94497398794d5f0206f4#note_581894
<enunes> I just never bothered to look into that much after this
Bentley has joined #lima
Bentley has quit [Remote host closed the connection]
marex has joined #lima
marex has quit []
marex has joined #lima
freemangordon has quit [Ping timeout: 480 seconds]
erle has joined #lima
freemangordon has joined #lima
Daanct12 has quit [Ping timeout: 480 seconds]
camus has quit []
<anarsoul> enunes: can you please share a command (or results) for failed spec/!opengl 1.1/depthstencil-default_fb-* tests?
marex has quit [Ping timeout: 480 seconds]
drod has joined #lima
<enunes> anarsoul: I think this should work PIGLIT_SOURCE_DIR="$HOME/piglit" PIGLIT_PLATFORM="gbm" $HOME/piglit/bin/fbo-depthstencil blit default_fb -samples=2 -auto
<anarsoul> OK, thanks!
<anarsoul> did drawpixels variant pass before?
<anarsoul> and why we aren't running it in CI?
<anarsoul> blit requires depth/stencil support in lima blitter for msaa resolve, I just added it
<anarsoul> drawpixels doesn't work, it fails for stencil
<enunes> no specific reason, I bundled it in skip to lower runtime
<enunes> I just replied in the MR, we could consider reevaluating how much runtime those add, maybe fine tune the skip list better if something is still low cost/benefit
<enunes> this is the regression list from that https://paste.centos.org/view/raw/269825a3
<anarsoul> drawpixels fail in a lot of other drivers, so I don't think it's lima issue
uis has quit []
uis has joined #lima
mmind00_ has joined #lima
clifontaive[m] has quit [reticulum.oftc.net helix.oftc.net]
drod has quit [reticulum.oftc.net helix.oftc.net]
mmind00 has quit [reticulum.oftc.net helix.oftc.net]
cyrozap has quit [reticulum.oftc.net helix.oftc.net]
drod has joined #lima
cyrozap has joined #lima
<anarsoul> oh, actually it's a flake
<anarsoul> not a failure
<anarsoul> same as on raspberry
<anarsoul> i.e. on vc4 and v3d
<anarsoul> so likely a bug in u_transfer_helper
<anarsoul> let me try if it fails on panfrost :)
<anarsoul> well, it passes on panfrost
mmind00_ has left #lima [#lima]
mmind00 has joined #lima
marex has joined #lima
<anarsoul> so looks like blit job races with something
<anarsoul> if I wait for blit job completion, it fails reliably, otherwise it fails intermittently
<anarsoul> enunes: any ideas how to debug it? it fails reliably with LIMA_DEBUG=dump, so I can't get a dump of passed job
<anarsoul> LIMA_DEBUG=dump synchronizes it, since it waits for completion of each job
clifontaive[m] has joined #lima
drod has quit [Remote host closed the connection]
Daanct12 has joined #lima
Daanct12 has quit [Remote host closed the connection]
<anarsoul> so looks like all the drivers that hit draw_stencil_pixels() fallback in st_DrawPixels() fail
<anarsoul> i.e. all the drivers that do not expose PIPE_CAP_SHADER_STENCIL_EXPORT