#asahi-dev · 2026-11-06 19:24
yukagood that I now have a gen3 trace for comparison#asahi-dev · 2026-11-06 19:23
yukagen4#asahi-dev · 2026-11-06 19:23
yukagen3 is the old one, I need to test this on the M5 Pro#asahi-dev · 2026-11-06 19:23
yukaah wait#asahi-dev · 2026-11-06 19:23
yukafwiw I'm looking into the spmi gen3 interrupt regs#asahi-dev · 2026-11-06 18:00
svenlol, sashiko hallucinates the old dt-check-style apparently :D#asahi-dev · 2026-11-06 17:57
svenack#asahi-dev · 2026-11-06 17:57
yukaThen also wait with M5 please, I'll send a v3#asahi-dev · 2026-11-06 17:56
yukathanks#asahi-dev · 2026-11-06 17:56
yukaAck#asahi-dev · 2026-11-06 17:55
sveneasiest way unfortunately is a separate series (possibly with fixes tags) that i'll just pick up on top the dt-7.5 branch#asahi-dev · 2026-11-06 17:54
svennot once they've been picked up from the ML#asahi-dev · 2026-11-06 17:42
yukaLike fixup! for mixed indent#asahi-dev · 2026-11-06 17:40
yukaIs there any unbureaucratic way I can apply some style fixes to the t604x dts before they bubble up further?#asahi-dev · 2026-11-06 16:43
cyberheadIs there any way I can support Asahi Linux on M4?#asahi-dev · 2026-11-06 10:13
s0ullight[m]OpenCL gives similar results fyi#asahi-dev · 2026-11-06 10:11
s0ullight[m]I don't have the setup for that#asahi-dev · 2026-11-06 10:10
chaos_princessmight be interesting to do the same from metal and see if there is additional pmp traffic#asahi-dev · 2026-11-06 10:09
s0ullight[m]Also 20 dispatches in one command buffer and submit#asahi-dev · 2026-11-06 10:08
s0ullight[m]Ah, no a 256mb surface without overlap#asahi-dev · 2026-11-06 10:05
chaos_princessyou are clearing a large memory block, not having every single thread slam into the same address#asahi-dev · 2026-11-06 10:04
s0ullight[m]v2, what do you mean by different one?#asahi-dev · 2026-11-06 10:00
chaos_princessalso, is that pmpv2 or pmpv1?#asahi-dev · 2026-11-06 10:00
chaos_princessdifferent one?#asahi-dev · 2026-11-06 09:58
s0ullight[m]a 16 byte vec4 in 65536 workgroups and 256 threads each. gpu local memory, and 20 submits.#asahi-dev · 2026-11-06 09:45
chaos_princessalso, yes, pmp has "bw-req" areas, which the dcp, isp, and gpu get pointers to#asahi-dev · 2026-11-06 09:42
chaos_princesshow are you measuring write throughput#asahi-dev · 2026-11-06 09:41
chaos_princessgpu should autonomously report to pmp#asahi-dev · 2026-11-06 09:31
s0ullight[m]Came across "but it likely involves controlling Apple Fabric power and clocking, among other things" regarding pmp in the 7.0 progress report. Enabled pmp and it impacted gpu write throughput, as in it got reduced. I guess it's currently not fully understood and needs more work ¯_(ツ)_/¯#asahi-dev · 2026-11-06 07:01
nicolas17ooh AWDL#asahi-dev · 2026-11-06 05:56
JamesCalligeros[m]It’s worth#asahi-dev · 2026-11-06 05:55
myriaIs it worth replying to the T2 BRCM patch to point out the older BRCMEXT patches that chaos_princess sent, or do we want to just leave it?#asahi-dev · 2026-11-06 01:03
MichaelReeves[m]oh i forgot I have to resend dchid for 7.4-rc1, I keep forgetting. I'll (hopefully) remember to do it later today#asahi-dev · 2026-11-06 00:59
MichaelReeves[m]oh yay wifi stuff in linux-next now. I guess I'll figure out how to get my phone to do 6GHz hotspot and work on that next#asahi-dev · 2026-11-05 21:23
jannauoops, I thought I had updated asahi-wip already to my v7.3.1 a couple of days ago#asahi-dev · 2026-11-05 19:13
svenor more specifically, i don't think we're allowed to send PRs that introduce new dt-check-style --mode=strict warnings#asahi-dev · 2026-11-05 19:10
svenbut no need to fix everything or make everything consistent#asahi-dev · 2026-11-05 19:10
sven+1 on eventually consistent: try to avoid introducing new warnings, etc.#asahi-dev · 2026-11-05 18:57
jannauthanks. I think "eventually correct" (by fixing and avoiding to introducing new issuses) is more important than "consistent"#asahi-dev · 2026-11-05 18:19
yukadon't want it inconsistent within one file#asahi-dev · 2026-11-05 18:19
yukawell then I will add a commit in front to fix it in the entire dt before adjusting it in the new additions#asahi-dev · 2026-11-05 18:16
jannauif patchwork / b4 prep --check complains about you're likely introducing the issue. for example with breaking the pmgr "compatible"s. please fix that for resubmission#asahi-dev · 2026-11-05 18:01
yukaunderstood, but the question remains: would this a blocker or a followup#asahi-dev · 2026-11-05 17:36
jannauyuka: the issue is that you have more than 7 spaces. kernel wants (x / 8) tabs followed by (x % 8) spaces#asahi-dev · 2026-11-05 17:25
yukait's "wrong" in all the dts I sent#asahi-dev · 2026-11-05 17:25
yukaOk, I'll send a v2 for t8132 dt-check-style. Do you really want the mixed-indent-chars lint too?#asahi-dev · 2026-11-05 17:15
sventhat would've been very convenient for e.g. dt-check-style or dt_binding_check#asahi-dev · 2026-11-05 17:15
svenhrm, trying b4 review tui again and ci checks are nice but they unfortunately just take the mail message as input and there doesn't seem to be any support for running checks on a tree#asahi-dev · 2026-11-05 17:06
svenlet's keep them together then#asahi-dev · 2026-11-05 16:52
jannauchanged by c08fe4da4acc ("dtc: dt-check-style: Relax property ordering rules (drop alphabetical)") in v7.4-rc1