02:17 airlied: gfxstrand: we don't have any variable length nir intrinsics do we?
02:22 airlied: gonna need a bigger boat
06:22 tzimmermann: sima, daniels, any issues with force-pushing into drm-misc-next-fixes? i intent to fixup commit tags in 216d6d913861 ("drm/panfrost: Print RSS for tiler heap BO's in debugfs GEMS file")
06:22 tzimmermann: mlankhorst ^
06:23 tzimmermann: see https://lore.kernel.org/dri-devel/20250812081415.7afe541e@canb.auug.org.au/T/#u and https://lore.kernel.org/dri-devel/CAPj87rPdJS_pZppfc4BnLWLT+BfikzKhpZJkU2uL8V_OHKYk7g@mail.gmail.com/T/#m291d9b0f27a96a58596c5e00ec30ce282eb908c9
06:23 daniels: tzimmermann: fine with me, I thought from what you said yesterday that we were waiting for you to revert it so I could reapply properly
06:24 tzimmermann: daniels, i thought you'd just revert yourself :)
06:25 tzimmermann: looking at it now, i'd like to replace that commit if possible
06:27 MoeIcenowy: I am trying to write a new kms driver (using bridge and bridge connector), but I got Unable to handle kernel NULL pointer dereference because of a call to NULL happens at `drm_connector_free_work_fn+0x64/0x78`
06:27 MoeIcenowy: anyone knows the possible problem?
06:28 MoeIcenowy: oh I am testing rmmod'ing it
06:29 daniels: tzimmermann: oh … sorry, I was waiting for you! yes, please feel free to replace if that’s what you think is best
06:30 tzimmermann: daniels, ok, deadlock resolved :)
06:32 daniels: thanks a lot!
06:33 MoeIcenowy: looks like drivers/gpu/drm/display/drm_bridge_connector.c's connector funcs lacks a destroy() func?
06:48 tzimmermann: imre, FYI the fbdev-format patches have been sent now
07:45 jani: daniels: so how did that happen, did dim fail?
08:19 mlankhorst: tzimmermann: I just made it read-only yesterday, it's yours. :)
08:20 tzimmermann: mlankhorst, great. PR sent out
08:28 MoeIcenowy: well, no destroy() is intended
08:32 daniels: jani: I did dim checkout -> b4 shazam (instead of dim b4-shazam) -> dim ub -> dim push
08:32 daniels: I suspect the dim ub (also unnecessary; I guess I was trying to subconsciously compensate for not using dim when I should've with one unnecessary command) meant that dim push didn't warn about the missing tags?
08:32 jani: daniels: huh, so dim didn't check the sob at push time
08:33 jani: I don't think dim ub should matter there
08:41 MoeIcenowy: okay drm_mode_config_cleanup seems to be not needed (and should not) be explicitly called
09:07 imre: tzimmermann: ok, thx
11:49 karolherbst: airlied: in what sense? number of sources or components?
11:50 karolherbst: or you mean "get the size of this variable"?
12:23 zmike: mareko: any chance you can pass this to someone internal https://bugzilla.kernel.org/show_bug.cgi?id=219732
12:23 zmike: perf has been broken on 5000 series threadrippers for a long time now
14:11 agd5f: zmike, passed it along
14:11 zmike: agd5f: thanks!
16:26 krathul: Hello, I am new here. I am interested in contributing to kernel gpu drivers.
16:29 krathul: To get started I though I would send some code style patches. I noticed some codestyle errors and warnings while running the checkpatch script on the panfrost driver source. I am on the drm-misc-next branch. Which branch should I target for my patches ?
17:07 jenatali: Ugh auth is painful
17:07 jenatali: airlied: re variable length intrinsics, you could follow what I did for printf which is to pass a struct?