goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
user982492 has joined #asahi-dev
c10l1 has joined #asahi-dev
c10l has quit [Remote host closed the connection]
c10l1 is now known as c10l
ma4 has joined #asahi-dev
ma3 has quit [Ping timeout: 480 seconds]
robinp has quit [Read error: Connection reset by peer]
lawrence has quit [Remote host closed the connection]
tobhe has joined #asahi-dev
tobhe_ has quit [Ping timeout: 480 seconds]
user982492 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jtf has quit [Ping timeout: 480 seconds]
SSJ_GZ has joined #asahi-dev
jtf has joined #asahi-dev
goldsoultheory has joined #asahi-dev
goldsoultheory has quit []
<sven>
huh, that looks like it tries to call typec_enter twice
<sven>
I thought that couldn’t happen ::
chadmed_ has joined #asahi-dev
goldsoultheory has joined #asahi-dev
goldsoultheory has quit []
povik has joined #asahi-dev
goldsoultheory has joined #asahi-dev
<_jannau_>
I don't think I've seen .enter being called more than once. .vdm is called multiple times
<sven>
hmm.. that makes more sense
chadmed_ has quit [Ping timeout: 480 seconds]
eiln has joined #asahi-dev
<eiln>
for dt are we doing name_dart or dart_name
goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
eiln has quit []
goldsoultheory has joined #asahi-dev
goldsoultheory has quit []
c10l has quit [Quit: Bye o/]
chadmed_ has joined #asahi-dev
c10l has joined #asahi-dev
MajorBiscuit has joined #asahi-dev
goldsoultheory has joined #asahi-dev
SSJ_GZ has quit [Ping timeout: 480 seconds]
goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
goldsoultheory has joined #asahi-dev
<marcan>
eiln: looks like name_dart, except dart_sio slipped in
<marcan>
(I don't care as long as we're consistent, looks like name_dart wins here)
Cyrinux has quit [Server closed connection]
Cyrinux has joined #asahi-dev
goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<marcan>
sent a fixup patch
<marcan>
also sent out the DART suspend/resume patch, in preparation for the t8110 series (because it comes first in git)
<thasti>
any chance we can get CONFIG_PKCS8_PRIVATE_KEY_PARSER enabled in future kernel releases? I fail to connect to eduroam (wpa enterprise network) since iwd can't read private keys without the pkcs8_key_parser module..
TheLink has quit [Server closed connection]
TheLink has joined #asahi-dev
chadmed_ has quit [Remote host closed the connection]
goldsoultheory has joined #asahi-dev
<marcan>
is *that* why wpa-enterprise fails?
<marcan>
seriously? I thought it was something between iwd and the wifi driver...
<marcan>
it's been in my back burner forever. sigh.
<marcan>
thanks for figuring that out.
<thasti>
:)
foxtrot has quit [Server closed connection]
foxtrot has joined #asahi-dev
<thasti>
neither networkmanager nor iwd are very explicit with their error messages, so that turned out to be quite the rabbit hole...
chadmed_ has joined #asahi-dev
lewurm_ has quit [Server closed connection]
lewurm_ has joined #asahi-dev
jix__ has quit [Server closed connection]
jix__ has joined #asahi-dev
nico_32 has quit [Server closed connection]
nico_32 has joined #asahi-dev
rmk has quit [Server closed connection]
rmk has joined #asahi-dev
<marcan>
povik, sven, ChaosPrincess, jannau: FYI, I fixed a bunch of kernel test robot nits in audio, atc, fpwm, and spi-hid stuff because I was getting tired of the endless emails :p
<marcan>
it got squashed in the rebase though, but you might want to know
<sven>
i think i already fixed the atc ones in my current wip branch
<marcan>
cool
<sven>
(and introduce a new one in the typec mess \o/)
<sven>
+d
<marcan>
lol
<thasti>
on a different note: for another (unencrypted) network, I'm getting 'brcmf_set_key_mgmt: invalid akm suite (1027090)' + 'brcmf_cfg80211_connect: wl_set_key_mgmt failed (-22)'
<maz>
marcan: thanks for having had a look at the IPI stuff.
<sven>
atcphc-20221130 fwiw, you'll need that one once the dwc3 race fix hits mainline
flying_sausages has quit [Server closed connection]
flying_sausages has joined #asahi-dev
<marcan>
maz: it wasn't a very close look, but then again it's just moving code as far as I can tell
<sven>
(except for the top commit, that's the typec mess)
<marcan>
only "regression" is the IPIs get sent once by one instead of potentially several at once for pre-fastipi SoCs, but whatever
<maz>
marcan: yeah, it only took 15 revisions to get there...
<sven>
:D
<marcan>
glad I'm not the only one suffering the pain :p
<marcan>
sven: should I pull that now? or wait until the race fix?
<sven>
that branch contains the race fix fwiw
<marcan>
ah, cool
<sven>
so, uh, whatever you prefer :)
<maz>
marcan: I'm not too worried about the single-IPI behaviour. if you trace them, you'll see that multi-target IPIs are super rare.
<marcan>
sven: I mean, is it stable material or needs more testing?
<marcan>
I want to push out a stable release, like, tomorrow or so with what we've fixed since the other week
<marcan>
maz: yeah, I figured as much
<sven>
without the typec commit it just needs some smoke testing on everything that's not a t8103
<marcan>
cool
<marcan>
I'll roll it in
<sven>
the changes are really that dwc3 race fix + trickle down crap to atcphy from that
<sven>
and preparing it for the typec subsytem altmode stuff
<sven>
so it should fix the quick plug/unplug issue and work the same otherwise
<_jannau_>
the dwc3 race fix works without issues on t600x, running it since last week
leitao has joined #asahi-dev
<marcan>
sven: do I need that svid thing in all the instances?
<sven>
that's for displayport and usb4/thunderbolt
<sven>
if you drop the typec commit it's unused
<marcan>
ah, cool
<marcan>
I dropped the last 2 commits (assume both are unnecessary)
<_jannau_>
yes
<sven>
yup
<sven>
there's apparently this weird setup where some typec ports have a mux for usb2/3 and another mux for dp-altmode or something
<sven>
so it uses that svid thing to route the mux_set to the correct mux
<sven>
don't ask me how that works though :D
chadmed_ has quit [Quit: Konversation terminated!]
chadmed_ has joined #asahi-dev
<povik>
marcan: thanks!
<marcan>
thasti: that ciphersuite is unsupported in the driver. I can try to add it (apparently it's in bcmdhd, no idea if the firwmare supports it but I'm guessing yes?)
<marcan>
this is WPA3 I assume?
chadmed_ has quit [Remote host closed the connection]
<marcan>
thasti: looks like that's a FT-FILS cipher suite, but the driver doesn't support FILS at all
<marcan>
yet another case of "broadcom doesn't care about the linux driver"
<marcan>
I could probably implement it if I had a way of testing it but...
lewurm has quit [Read error: Connection reset by peer]
lewurm has joined #asahi-dev
<marcan>
thasti: kernel in asahi-dev should have that PKCS8 option enabled now
chadmed_ has joined #asahi-dev
<marcan>
(and also the atcphy updates/etc)
SSJ_GZ has joined #asahi-dev
steffen[m] has quit [Server closed connection]
steffen[m] has joined #asahi-dev
akemin_dayo has quit [Server closed connection]
akemin_dayo has joined #asahi-dev
tumblingweed has quit [Server closed connection]
tumblingweed has joined #asahi-dev
leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<thasti>
marcan: thanks, I confirm WPA3 works fine now!
Retr0id has quit [Server closed connection]
<thasti>
WPA enterprise*
Retr0id has joined #asahi-dev
maximbaz has quit [Server closed connection]
maximbaz has joined #asahi-dev
<ChaosPrincess>
Is there a list somewhere of everything that can he hardware en/decoded on m1 series? Googling results in conflicting info
_jannau_ has quit [Server closed connection]
_jannau__ has joined #asahi-dev
<_jannau__>
I guess one has to query videotoolbox in macos to be sure. h.265, h264, vp9 (probably decode only), mpeg4 part2 and mpeg2 where supported on ios devices (probably decode only as well)
<_jannau__>
and prores on t600x and m2
<ChaosPrincess>
Ty
leitao has joined #asahi-dev
<_jannau__>
and prores is separate of avd/ave
VinDuv has quit [Ping timeout: 480 seconds]
<chadmed_>
could you coax the jpeg block to do mjpeg :P
<_jannau__>
h.263 (decode) as well. yes, you could use the hw jpeg block for m-jpeg
_jannau__ is now known as _jannau_
chadmed_ has quit [Remote host closed the connection]
beeblebrox has joined #asahi-dev
leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
leitao has joined #asahi-dev
ChaosPrincess is now known as ChaosPrincess[n]
<sven>
jannau: can you log all dp sid changes when you see that warn? i can't reproduce it with my screens here
<sven>
but I suspect that yours either first negotiates 2-lane dp and then changes pin configuration or that yours is just a bit slow and it only looks like it renegotiates the configuration
<sven>
either way i'll probably have to actually implement a better fake dp altmode state machine *sigh*
<sven>
that typec subsystem is kind meant for the situation where you there's more control over the VDOs actually sent while this tps chip does all that automatically
ChaosPrincess[n] is now known as ChaosPrincess
<sven>
I think I even added a tracepoints for DP sid changes
<sven>
-s
<marcan>
sven: Linux subsystems with poor abstractions? color me shocked :p
<sven>
yeah.. it was written for some typec controller called tcpm where Linux has to do everything afaict
<sven>
then there’s ucsi which already implemented a fake DisplayPort state machine that I more or less copied :D
<sven>
it’ll also fail in funny ways when there are two altmodes like tbt and DP at the same time
<Cyrinux>
Hi, marcan, can you enable as module PKCS8_PRIVATE_KEY_PARSER by default in kernel? It's commonly use to WPA auth over tls. For both standard and edge. It's like this in Linux and Linux--lts.Thank for your work. :)
<j`ey>
Cyrinux: he already did that this morning!
<jannau>
Cyrinux: 12:26 < marcan> thasti: kernel in asahi-dev should have that PKCS8 option enabled now
<blazra>
I don't know if this is known or not, but the way to reproduce the "spi-hid-apple-of spi0.0: Read package crc mismatch" message in dmesg is to type on the internal keyboard and moving cursor using the touchpad at the same time.
systwi has quit [Ping timeout: 480 seconds]
<jannau>
despite the warning dcpext seems to work most of the time
<sven>
that’s, uh, surprising :D
systwi has joined #asahi-dev
<sven>
maybe I’ll have to redo my mode transition tests now with atcphy being stable
<sven>
could be that e.g. usb3 to usb/DP transitions don’t require a dwc3 reset
<jannau>
blazra: thanks, now when you say that it seems to match what I see. still doesn't explain it though
<povik>
could touchpad events be interjecting the keyboard ones in some weird prioritizing scheme?
<povik>
do we see the crc mismatches when tracing macos?
yuyichao has quit [Ping timeout: 480 seconds]
systwi_ has quit [Ping timeout: 480 seconds]
<jannau>
no, the packet payload is mostly zeros
<povik>
hm? as in, the bad packets are mostly zeroed out?
<tpw_rules>
yeah
<povik>
and macos gets it too in tracer?
<jannau>
macos uses sio for spi, packets I traced from there are all ok
<povik>
right, hm
<jannau>
so no idea what sio does, it could simple ignore broken packets
<sven>
maybe they have a hack in SIO to workaround a race condition in their keyboard that sends packets too early ;p
<blazra>
is it like two devices on one SPI bus or one device which sends both touchpad and keyboard data?
yuyichao has joined #asahi-dev
bluetail5 has joined #asahi-dev
bluetail has quit [Read error: Connection reset by peer]
bluetail5 is now known as bluetail
<jannau>
it's single device integrating kbd, touchpad, ...
<blazra>
I am looking at the spihid_transfer_packet struct: Could there be more messages in one packet? Like a combined keyboard and touchpad data maybe?
<jannau>
I don't think so. as I said the payload is mostly 0x00 and we certainly are not using keyboard events
user982492 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]