#
#asahi-dev · 2026-08-02 15:08
svenuh, also cwd12 isn't prp
#asahi-dev · 2026-08-02 14:54
svens/ping/long/
#asahi-dev · 2026-08-02 14:54
svenI also brought this up in a very old firmware ping before we had the installer (12.1 or so)
#asahi-dev · 2026-08-02 14:54
svenbut I very much remember that it doesn’t like something being NULL
#asahi-dev · 2026-08-02 14:53
svenmaybe that just happens on m1
#asahi-dev · 2026-08-02 14:28
yukasince we changed cwd12 = 1 -> cwd12 = 0, nvme_read() works fine without APPLE_ANS_PRP_NULL_CHECK on old and new firmware. which does not make sense with the comment in the kernel driver?
#asahi-dev · 2026-08-02 14:07
yukaand if not having it is strictly a problem
#asahi-dev · 2026-08-02 14:06
yukabut what APPLE_ANS_PRP_NULL_CHECK _actually_ does is still a mystery to me
#asahi-dev · 2026-08-02 14:05
yukaEverything points to 15.x+ fw being simply missing APPLE_ANS_UNKNOWN_CTRL and behaving as 14.x without clearing APPLE_ANS_PRP_NULL_CHECK
#asahi-dev · 2026-08-02 14:04
yukanvme is annoying too. The comment on APPLE_ANS_UNKNOWN_CTRL in the kernel driver is very confusing, and does not match what I can observe anywhere.
#asahi-dev · 2026-08-02 14:03
yukasomehow b4 sent out an old version of the cover letter. the changes probably lost when doing the patch dependencies things.
#asahi-dev · 2026-08-02 13:15
sofusi updated my pr, i think clang-format makes it readable enough, and i would prefer not having to do the formatting by hand
#asahi-dev · 2026-08-02 12:07
chaos_princessultra will be later
#asahi-dev · 2026-08-02 12:07
chaos_princesswe can
#asahi-dev · 2026-08-02 12:07
NickChan[m]can't exactly commit to 14.8.3 anyways because m3 ultra mac studio needs at least 15.1
#asahi-dev · 2026-08-02 12:03
svenI keep forgetting that m3 runs 14.8
#asahi-dev · 2026-08-02 11:58
chadmedyeah
#asahi-dev · 2026-08-02 11:58
chaos_princessso not having dcp ready is fine
#asahi-dev · 2026-08-02 11:57
chaos_princesswell, yes, we are doing boot framebuffer until gpu becomes a thing
#asahi-dev · 2026-08-02 11:57
chadmededid and hdmi audio arent important but i dont want to regress hw cursors on plasma so we need to sort out the surface clearing at the very least
#asahi-dev · 2026-08-02 11:57
chadmedi still dont think gpu-less dcp is very useable but if everyone else is fine with it i wont nack a release
#asahi-dev · 2026-08-02 11:56
chaos_princessisp on base m3, getting stuff merged and tested.
#asahi-dev · 2026-08-02 11:55
chadmedgpu aside, whats actually left?
#asahi-dev · 2026-08-02 11:54
chaos_princesssven: i think we should release m3 on 14.8.3 sooner rather than later
#asahi-dev · 2026-08-02 11:54
svenno
#asahi-dev · 2026-08-02 11:54
Tachyonray[m]can test it on the M3 too if that would confirm something
#asahi-dev · 2026-08-02 11:54
chaos_princessyou shouldn't
#asahi-dev · 2026-08-02 11:53
Tachyonray[m]did run the boot flags fix on my M1, but haven’t updated to 26.6
#asahi-dev · 2026-08-02 11:53
svenand unless we release the installer with m3. I’d prefer if people didn’t have to go through a macOS round trip later if we can avoid it
#asahi-dev · 2026-08-02 11:52
chaos_princessas long as we aren't starting gpu/dcp work, we don't really need to commit to specific versions
#asahi-dev · 2026-08-02 11:49
Tachyonray[m]chadmed: i like tachyons
#asahi-dev · 2026-08-02 11:48
svenwould be good to confirm it works before we settle on 26.6 or whatever
#asahi-dev · 2026-08-02 10:51
chadmedwe also have a waning moon, so theres fewer tachyons in the atmosphere
#asahi-dev · 2026-08-02 10:50
yukaMaybe because the heat wave is over, the quantum effects have come down to a level that is favorable (/s)
#asahi-dev · 2026-08-02 10:43
yukasven: I haven't seen the "trying to write to read-only [...]" since a few weeks now, but it depends very much on the kernel config used. I guess I would need to try some kernel with fedora config to make a definite statement
#asahi-dev · 2026-08-02 10:35
chadmedbut thats so gross...
#asahi-dev · 2026-08-02 10:35
chadmedwhich is why i suspect ^
#asahi-dev · 2026-08-02 10:35
chadmedi cant tell. i took some very very large and verbose traces of me typing so that the I-beam disappears and nothing changes. its just a normal swap
#asahi-dev · 2026-08-02 10:34
chaos_princesswhat does macos do?
#asahi-dev · 2026-08-02 10:33
chadmedi am really hoping and praying the answer is not "send a transparent framebuffer and keep it alive until the surface is needed again"
#asahi-dev · 2026-08-02 10:33
chadmeddue to 14971a2eb042 your machine will eventually hang but its good enough to enable tracing an diagnostics until we figure out how to get this thing to clear a surface
#asahi-dev · 2026-08-02 10:31
chadmedeverything we already supported except edid, hdmi audio and for some reason clearing surfaces works reliably
#asahi-dev · 2026-08-02 10:30
chadmedanyway WIP dcp 14.7 MR https://github.com/AsahiLinux/linux/pull/562
#asahi-dev · 2026-08-02 10:30
chadmedmight be worth making it shared?
#asahi-dev · 2026-08-02 10:29
chadmedlol
#asahi-dev · 2026-08-02 10:29
sofusi think there are only to other stateless av1 drivers
#asahi-dev · 2026-08-02 10:29
chadmedif this is stuff that is inherent to the spec and required by multiple drivers im surprised it isnt shared already
#asahi-dev · 2026-08-02 10:28
sofusjust thought it could be interesting for upstream, maybe they would like to make a shared thing or something, but i suppose i could fix the style things in a different commit?
#asahi-dev · 2026-08-02 10:26
LarstiQwill it have updates to be incorporated?
#asahi-dev · 2026-08-02 10:25
chadmed(not saying futzing with the preprocessor does to be clear)

plus ›