03:43 airlied: robclark: woah, climbing mount doom :-)
06:49 lucaceresoli: vsyrjala: oops, thanks for the report! Do you have a link to see the failure log?
06:49 lucaceresoli: vsyrjala: a defconfig for me to reproduce?
06:54 lucaceresoli: vsyrjala: nevermind, I reproduced it: error: redefinition of ‘devm_drm_put_bridge’
07:14 lucaceresoli: vsyrjala: fix sent: https://lore.kernel.org/all/20250522-devm_drm_put_bridge-fix-non-of-build-v1-1-a05234dea046@bootlin.com/
08:12 pq: zmike, thank you for the surfaceless DRI_PRIME! :-)
09:33 tjaalton: any idea how to get mesa use rustc binary which has the version attached to it, like 'rustc-1.78' (from a backport)?
09:57 jannau: tjaalton: `RUSTC=rustc-1.78 meson ...`. might need RUSTC_LD as well
09:57 tjaalton: jannau: oh.. thanks, sounds familiar
10:01 pepp: tjaalton: you should be able to specify it in the native file used by meson (rust = 'rustc-1.78')
10:02 lucaceresoli: vsyrjala: fix applied to drm-misc-next
10:02 tjaalton: hmm okay, exporting the variable is easier though, and does seem to work
12:18 robclark: airlied: heheh
14:14 zmike: pq: 🤝
14:19 pinchartl: does DRM support limited-range RGB ?
14:21 emersion: there is a non-standard "Broadcast RGB" property
14:23 pinchartl: emersion: thanks
14:23 pinchartl: how about limited-range RGB contents in planes ? I don't see any property to support that
14:24 emersion: oh, it seems like it's a standard prop now
14:25 emersion: https://drmdb.emersion.fr/properties/3233857728/Broadcast%20RGB
14:25 emersion: for planes there is https://drmdb.emersion.fr/properties/4008636142/COLOR_RANGE
14:25 emersion: but it's for YUV only
14:26 emersion: it'll soon get superseded by color pipelines
14:29 zamundaaa[m]: pinchartl: you can use (DE)GAMMA_LUT or the upcoming color pipelines to scale colors in whatever way you need
14:31 pinchartl: maybe I'll wait for the color pipeline properties and see :-)
14:32 pinchartl: zamundaaa[m]: my hardware hardcodes RGB <-> YUV conversion with a selection of preset matrices, the LUT properties won't help
16:41 lumag: robclark, karolherbst, alyssa: I've run into infinite loop in the compiler. https://gitlab.freedesktop.org/mesa/mesa/-/issues/13223 :-D
16:46 alyssa: oops
16:48 alyssa: lumag: could you show a bit more nir_print?
16:48 alyssa: I can't tell what op nir_lower_pack is actually lowering here
16:48 alyssa: oh pack_64_4x16
16:48 alyssa: nvm
16:49 alyssa: so opt_algebraic creates that op unless options->lower_pack_64_4x16
16:49 alyssa: but nir_lower_packing checks instead skip_lower_packing_ops
16:49 alyssa: ostensibly this is a freedreno bug for not setting that option but also wat, why are those separate options
16:50 alyssa: oh i see how we got here.
16:50 alyssa: gah
16:50 alyssa: that pass used to lower everything, then d43c5003fc8 ("nir: add skip_lower_packing_ops shader compile option") happened
16:51 karolherbst: lumag: did you try with main?
16:51 karolherbst: I think it got fixed already
16:51 karolherbst: alyssa: ohh.. that's on freedreno? mhhh
16:52 alyssa: dad5ee10392 ("intel/brw, lp: enable lower_pack_64_4x16") papered over this
16:52 karolherbst: I fixed it for other dirvers 🙃
16:52 karolherbst: yeah, that's mine
16:52 alyssa: yeah that just swept the issue under the rug
16:52 alyssa: the actual problem here is having duplicated nir options
16:52 karolherbst: the proper solution is to get rid of lower_pack
16:52 alyssa: well, or that
16:52 alyssa: I can start typing this afternoon, currently dealing with a $situation
16:52 karolherbst: the option situations is a bit of a disaster honestly
16:53 alyssa: no argument
16:53 alyssa: ()
16:53 alyssa: (void)
16:55 lumag: alyssa: that would be really great, thanks :-) I can try checking if I can adapt Intel solution for now.
16:56 lumag: alyssa: karolherbst: indeed setting .lower_pack_64_4x16 fixes the issue for fdo
16:57 lumag: thanks for the quick help and if necessary, I can test a proper solution anytime
17:34 hch12907: can anyone help with this? https://gitlab.freedesktop.org/mesa/demos/-/issues/48
17:35 hch12907: me not owning any nvidia cards and also optimus laptops kinda makes it impossible to debug on my side
17:36 hch12907: (basically the only possibility is me staring at the code for hours)
22:02 alarumbe: Hi eric_engestrom, I made some changes to src/mesa/main/objectlabel.c. It seems the file hasn't received many changes over the years, but latest was yours, so I was wondering if you could quickly glance over the MR
22:02 alarumbe: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34224
22:03 alarumbe: The commit that touches on that file is here: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34224/diffs?commit_id=bc28666abedba69749e2f778dad4780cb2893a02
22:16 demarchi: humn... apparently a bad merge went in drm-tip?
22:16 demarchi: void devm_drm_put_bridge(struct device *dev, struct drm_bridge *bridge);
22:16 demarchi: defined both inside and outside the ifdef
22:17 demarchi: humn... or simply a wrong patch
22:18 demarchi: drm/bridge: add devm_drm_put_bridge() ...
22:28 demarchi: ugh... drm-tip was failing to fetch for me, so I wasn't getting an update
22:28 demarchi: commit ee1855582e5e0f8d73428875b91a30785bcc4b7c fixes it