#
…

2021-09-24

00:20quarkyalice_ has joined #asahi-dev
00:21quarkyalice has quit [Ping timeout: 480 seconds]
00:41marcanalyssa: the i2c stuff should AIUI
00:41marcanas in i2c over one of the other endpoints
00:42marcanbut also someone should fix GNOME
00:42marcannot everything has or will have an EDID
00:42marcanI worked out the ring buffer format for the other endpoints (see the tracer), but I never looked at the actual messages beyond a cursory glance
00:42marcanit's a completely different RPC to the main dcpep
00:43marcan(but at least I think it's the same for all the *other *eps)
00:43alyssasimpler, I hope?
00:43marcanat least I don't think it has the multichannel asynchronous callback stuff
00:43marcanand I think in the case of those they moved entire drivers outright
00:43marcanso the API is more like "driver front side", not a cut down the middle
00:43marcanbut I'm not sure
00:43alyssamoan. will punt on I2C etc for later on I guess
00:43alyssawith latest m1n1
00:44marcanbacktrace?
00:44alyssaoh err uhhhh sec
00:45marcanalso uh that's weird
00:45marcanwhat are you running again, exactly?
00:45alyssa11.4 macOS..
00:45marcan11.4 no third digit?
00:46alyssado i look like a math major
00:46marcanthat looks like the stack pointer went bananas and ended up in IO space
00:47marcaneither that or something's really bugged in the HV
00:48alyssalets see if I get a backtrace this time
00:49marcancan you also p.hv_translate(hv.ctx.sp[1], True)
00:51alyssa0x211fffcc0
00:51marcanthat's like a stack overflow or something
00:51alyssagoing to say something's really bugged in the HV
00:52alyssa(breakage since SMP stuff)
00:52marcanworks for me on newer versions though :/
00:52marcanand with linux
00:52marcanalso it's breaking when spinning up the extra CPUs
00:52alyssasven: can you reproduce?
00:52marcanso it's not a regression, it's something broken there in SMP support
00:53alyssaah, fair.
00:53alyssathat's.. marginally better
00:54alyssado I have enough energy to cherrypick apple-pmgr-pstate and rebase my tree ... maybe i can start
00:57alyssa'mfd: syscon: Use of_iomap() instead of ioremap()' already landed, yes?
00:58marcanI believe so
00:59alyssaack, that explains why I already had it in my tree 😅
01:00alyssacherrypicked, now to deal with conflicts with the other dozen drivers in here
01:01alyssaand.. fix up the DTs I guess.
01:04alyssahow fun.
01:09alyssaWARNING: unmet direct dependencies detected for PM_GENERIC_DOMAINS Depends on [n]: PM [=n]
01:09alyssa^ might be a Kconfig bug or I just botched the cherrypick
01:10alyssaand of course enabling PM=y will rebuild my whole kernel..
01:16yuyichao has joined #asahi-dev
01:43marcanohh wait
01:44marcanno, I am doing that,wtf
01:44marcannevermidn me
02:29phiologe has quit [Ping timeout: 480 seconds]
05:59svenalyssa: works for me
06:32aleasto has joined #asahi-dev
08:18X-Scale has joined #asahi-dev
08:23X-Scale` has quit [Ping timeout: 480 seconds]
09:24robinp_ is now known as robinp
09:30chadmed has joined #asahi-dev
11:38alyssasven: mumble. ok
11:39sveni got a different exception at first but that was just because i forgot to chainload fwiw
11:46sveni'm also still one some old mac os version
12:49marcansven: I fixed the hv thing
12:49marcanit was a dev kernel assert that I'd fixed for UP but borked on SMP by accident, and never tested on dev since then
12:49marcanalso I really need to add a "get version" thing to m1n1 and a big fat blinking red text warning to the HV on mismatch
12:50marcanor just refuse to run really, or even just auto-chainload
13:42svenoh, that explains why it just worked here :D
14:36yuyichao has quit [Ping timeout: 480 seconds]
14:52yuyichao has joined #asahi-dev
14:53jbowen has quit [Quit: leaving]
15:03brinly_ has joined #asahi-dev
15:08brinly has quit [Ping timeout: 480 seconds]
15:08brinly_ is now known as brinly
16:42phiologe has joined #asahi-dev
17:10roxfan has joined #asahi-dev
18:44yuyichao has quit [Ping timeout: 480 seconds]
20:02aleasto has quit [Ping timeout: 480 seconds]
20:28d4ve has quit [Ping timeout: 480 seconds]
20:28d4ve has joined #asahi-dev
22:36alyssahey, if anyone wants to start testing the DCP driver, let me know
22:36alyssathis week's changes in theory mean it should work on other people's computers than my own
22:37alyssait's still far from done, but I've been dogfooding it for my CS theory homework and it's pretty stable
22:38alyssaSo far only tested on Mac Mini
22:39alyssano idea if it'll work on the MacBooks/iMac.
22:39alyssa(If it does work there, it's only for the internal screen, not external monitors. Yet)
…