00:53 airlied: robclark: dim: 49c2e01be19c ("drm/msm: fix -Wformat-security warnings"): Subject in fixes line doesn't match referenced commit:
00:54 airlied: dim: 1a40bb31fcf1 ("drm/msm/dpu: Add VBIF to DPU snapshot") # and others
00:54 airlied: are those real problems?
00:54 airlied: dim: 402074f0105e ("opp: core: Fix off by one in dev_pm_opp_get_bw()"): SHA1 in fixes line not found:
00:54 airlied: dim: d78653dcd8bf ("opp: core: implement dev_pm_opp_get_bw")
01:07 robclark: airlied: the opp ones, well I guess that is a pm/opp tree problem? I'll check the other but I think it is probably right subject wrong sha?
01:25 airlied: robclark: if both those are in the opp tree that is a pita since you've pulled from it and it shouldn't have changes it's shas
01:26 robclark: only the `opp: core:` ones are... can't really do much about that.. lumag would have merged the other
01:47 Venemo: what is the right way to indicate that I would like to get a MR in the next Mesa branch point?
01:47 Venemo: I thought that was what GitLab milestones are for, but I don't see any relevant one
01:50 kisak: https://docs.mesa3d.org/submittingpatches.html#nominating-a-commit-for-a-stable-branch
01:50 kisak: Prefered -> Fixes: <git commit> "human readable commit name"
01:51 kisak: secondary new: Backport to: 24.3 24.2 || secondary old: CC: mesa-stable
01:51 kisak: if there's a problem, then make a merge request against staging/##.#
01:52 kisak: Oh, of course I typed that all out before realizing I misread the question.
01:53 kisak: yes, usually there's a tracking issue report. If there isn't then it might make sense to make one that looks like one of the older iterations and ping the upcoming package maintainer.
01:55 kisak: s/issue report/milestone/
02:50 hybridmelon: you are debiliated people and to shut you off fucking human trash i needed to pull stunts and put my own wisdom at risk, but for such scamming you get entirely destroid too.
02:51 hybridmelon: destroyed
02:53 hybridmelon: and about this eesti garbage who changed the world to a scam evil and abnoxious place to live for all stars and special people under umbrella of europe that human is human , we gonna take care very brutally.
02:54 karenw: 🍿
03:02 dcbaker: Venemo: there will be a milestone for 25.0 when it gets branched in a week or so. For 24.3
03:02 dcbaker: What kisak said
03:05 hybridmelon: i told you all what is going to happen when my grannies assets were used to endorse people who never achieved anything but absurd lies when my family said i am crazy and financed those groups with tech that was snatched from me, people without any hesitation spammed the world with wank tunes, and as to how retarded they are has no measure, its shocking such people pull as to chaos and we
03:05 hybridmelon: get rid of them no matter what.
03:07 robclark: airlied: so 1a40bb31fcf1 is in the history of msm-next... so that one is, I think, dim being... dim
03:08 robclark: airlied: I guess the other one viresh must have rebased pm/opp next branch at some point.. I don't think we can do much about that
03:10 airlied: robclark: cool thanks!
03:20 hybridmelon: i know those people they got a set of laptops to pull shit into media and delay my life , so to block the evolution i arrived 10 years later hence, cause 2008 i got my first mobile device after escaping their terror for 3years, my life is over despite as to how big fighter i am cause of such clowns.
03:24 hybridmelon: werner is such a dj who goes to other countries telling stories of how big bitcoin magnat he is, having no clue whatsover what is technonlogy and makes tunes as to how genetically ill man is a wanker schizophrenic, as dumb as a human can be spins dj vinyls and thinks life works so wank tunes, wtf. i ask never made any work and he thinks genuinly that he is someone
03:25 Venemo: kisak, dcbaker sorry, I don't mean that I want to backport something to a stable branch. I mean that I want to make sure a MR gets into the next release
03:25 Venemo: so, that would be 25.0, I think
03:27 Venemo: in other words, what I mean is, I want it merged before 25.0 is branched
03:30 dcbaker: Venemo: I think just ping eric_engestrom on that
03:30 Venemo: okay, I will
04:55 airlied: mripard: that merge failed to build due to the change dropping the int ret = 0; to int ret;
04:55 airlied: clang correctly points out it leads to an uninit ret
05:11 airlied: mripard: if you fix that I can pull it still
08:13 mripard: airlied: on it
09:02 mripard: airlied: done
09:12 dolphin: airlied, sima: Sent the drm-intel-gt-next PR. Haswell engine reset improvement and the rest is smaller tweaks.
13:43 eric_engestrom: Venemo: hakzsam told me there might be a blocker for the 25.0 branchpoint but I think he needed to talk to someone, that might be what you're referring to as well? either way, for blocking/pushing back the branchpoint, there's no tracker issue/milestone set up, it's just "talk to the release manager" :)
13:43 eric_engestrom: if you want to make sure something is merged and it's already in an MR, give me the link and I'll keep an eye on it and ping as appropriate; if it's something that can be public yet (or just hasn't been done yet), then let me know (in private if you want) and we can see when to push back the branchpoint
13:43 eric_engestrom: just let me know by the end of tuesday (14th) :)
14:26 Venemo: eric_engestrom: thanks, will do
14:26 Venemo: eric_engestrom: I'm referring to something different
14:35 eric_engestrom: Venemo: ack
14:42 eric_engestrom: Venemo: talked to hakzsam just now, I'll delay the branchpoint by 2 weeks, so Jan 29 instead of Jan 15, hopefully that's also enough for what you need :)
14:42 eric_engestrom: (I'll update the calendar in a bit)
17:25 glehmann: why does gallivm have the two soa/aos modes? is that just technical debt, perf optimization or required for some features?
17:35 alyssa: glehmann: iirc, perf
17:35 alyssa: IIRC, one of them is a fast path for simple 2D workloads
17:35 alyssa: i don't remember details
17:37 zmike: // AOS is used for vectors of uint8[16]
17:40 HdkR: Small vector element size moment
17:52 alyssa: ..Why do we have both IS_POT(_NONZERO) and util_is_power_of_two_(non/or)zero(64)
17:52 alyssa: what is happening here t_t
17:52 zmike: macro vs util
17:52 zmike: the eternal struggle
17:53 alyssa: Can we sed the util_* into the macro? Well.. maybe not if there's some driver out there relying on 64-bit -> 32-bit truncation behaviour
17:53 alyssa: alwayswas.jpg
17:54 jenatali: glehmann: Looks like the AOS version operates on unorm8 as fixed-point integer (4 channels in 2 SIMD registers) where SOA is floating point (4 channels in 4 SIMD registers)
17:54 jenatali: Makes sense, WARP has a similar optimization
18:30 DemiMarie: alyssa: what about replacing both with a macro that calls a `_Generic ` static inline?
19:31 jenatali: Would be fine by me, MSVC finally supports _Generic and C11
22:02 agd5f: does arrow lake support HDMI 2.1 natively or does it use a PCON?
22:03 Sachiel: noticing a bunch of "Reviewed-by: Kenneth Graunke <None>" stuff in logs now, is marge adding RBs on its own now based on MR approvals?
22:04 jenatali: Yes
22:05 Sachiel: interesting. Was it intended?
22:05 jenatali: Yes
22:38 Kayden: interesting, I'll stop using approvals then