#
…

2023-03-11

00:31stipa has quit [Read error: Connection reset by peer]
00:33stipa has joined #asahi-dev
00:57blazra has quit [Ping timeout: 480 seconds]
01:01blazra has joined #asahi-dev
02:20Emantor has quit [Quit: ZNC - http://znc.in]
02:20Emantor has joined #asahi-dev
02:35zef_ has joined #asahi-dev
02:42zef has quit [Ping timeout: 480 seconds]
02:55bluetail9 has joined #asahi-dev
03:00bluetail has quit [Ping timeout: 480 seconds]
03:13pbsds3 has quit []
03:14pbsds has joined #asahi-dev
03:14D-Spirits has quit [Ping timeout: 480 seconds]
03:18chadmed_ has joined #asahi-dev
03:25chadmed has quit [Ping timeout: 480 seconds]
03:33chadmed_ has quit [Read error: No route to host]
03:33chadmed_ has joined #asahi-dev
03:35avocicltb^ has quit [Remote host closed the connection]
03:40Mary6 has quit []
03:41Mary6 has joined #asahi-dev
05:05chadmed_ has quit [Remote host closed the connection]
06:45robinp_ has joined #asahi-dev
06:50chadmed has joined #asahi-dev
06:52robinp has quit [Ping timeout: 480 seconds]
07:26chadmed_ has joined #asahi-dev
07:26robinp_ has quit [Remote host closed the connection]
07:26robinp has joined #asahi-dev
07:32chadmed has quit [Ping timeout: 480 seconds]
07:39eiln has quit [Quit: Page closed]
08:04eiln has joined #asahi-dev
08:12eiln has quit [Remote host closed the connection]
08:18eiln has joined #asahi-dev
08:19eilntop hacker news post is about the neural engine... currently testing if those LMs can run
08:36eilnthey all convert successfully.. lets see if it runs
08:37eiln has quit [Quit: Page closed]
08:49hightower2 has quit [Remote host closed the connection]
08:52eiln has joined #asahi-dev
09:02chadmed_ has quit [Remote host closed the connection]
09:03chadmed has joined #asahi-dev
09:04eilnshaves off a solid 0.01s too
09:04eilnalso should get faster once i get to hardware queue sched
09:04eilncode's copy-pasted from the compute examples i pushed yesterday
09:12Ry_Darcy has joined #asahi-dev
09:13Ry_DarcyI have not seen any progress on either Arch or Debian working on an M2 Mini. Any developments here?
09:16Bey0ndB1nary has joined #asahi-dev
09:26chadmed has quit [Remote host closed the connection]
09:28chadmed has joined #asahi-dev
09:38eiln has quit [Quit: Page closed]
09:47jannaumaz: pushed a preliminary fix for wifi
09:50Ry_Darcy has quit [Remote host closed the connection]
10:06Ry_Darcy has joined #asahi-dev
10:16roxfan2 has joined #asahi-dev
10:20Ry_Darcy has quit [Remote host closed the connection]
10:21roxfan has quit [Ping timeout: 480 seconds]
10:37mazjannau: now building.
10:43mazjannau: success!
10:45mazjannau: feel free to add my TB tag to it when you post it (given that the original patch has a Cc stable, it is going to quickly trickle down older kernels and wreck havoc...)
10:53marcanI'm going to throw on some stuff for the 6.2 branch and cut a release, if that works we might as well go straight to 6.3-rc1
11:01povikmarcan: throw in b4b422579339
11:02marcandone, thanks
11:20marcanjannau: that pull is just 4 commits right? (GH is being silly)
11:21jannauyes, two commit from me and two from lina on top of bits/200-dcp
11:22marcanjannau: nothing in that branch for 6.2, right?
11:23jannauno, I assume we do not want dcp multi fw support yet
11:23marcanyeah, not yet
11:24marcanwe've accumulated too much stuff for 6.2, let's get it out the door and then we can think about all the M2 stuff
11:25jannaudo you want the bits/000-devicetree rebase on top of my upstream changes?
11:25marcanI did pull that (probably not pushed yet)
11:26ChaosPrincesswhat do you think of pulling the z2 series in?
11:31marcanChaosPrincess: not very useful without screen tbh :/
11:32marcanI can pull it and leave the config off?
11:32marcanwhat's your latest branch?
11:32ChaosPrincessyea, screen is wip, and im not yet sure if i have the skill to even re it
11:33ChaosPrincessi guess lets keep it as an open mr for now until screen happens
11:33marcanup to you, I just don't want it enabled for users because it adds potential for confusion without being useful
11:36marcanjannau: stuff should be pushed now
11:43chadmed has quit [Remote host closed the connection]
11:45Mary6 has quit []
11:45Mary6 has joined #asahi-dev
11:52chadmed has joined #asahi-dev
11:56jannaumarcan: smoke test of asahi-6.2-5 with agx/release from tuesday passes on j493
12:00joshtaylor has quit [Quit: Connection closed for inactivity]
12:02marcanpushed everything to asahi-dev, please test :)
12:03marcanasahi/mesa:main is agx/release rebased on current mesa with some stuff replaced with commits from a MR alyssa has open
12:26marcansven, povik: please check/test :)
12:26DarkShadow44 has quit [Quit: ZNC - https://znc.in]
12:27jannauuboot fails to load efi binaries with "Reading file would overwrite reserved memory" on j493
12:27DarkShadow44 has joined #asahi-dev
12:29___nick___ has joined #asahi-dev
12:29marcanuboot regression?
12:29marcangrub works for me on j314
12:29marcancc kettenis
12:30marcanor something with the DCP changes?
12:30chadmedlatest u-boot + sd-boot works fine on j314 here too
12:31chadmed has quit [Remote host closed the connection]
12:34marcanjannau: 12.3 or 13.2?
12:35jannauI was thinking m1n1 bug, reserving random memory from malloc in m1n1
12:35marcanyeah, could be a m1n1 bug
12:36marcanlooks sensible on j314 at least
12:37chadmed has joined #asahi-dev
12:38chadmedasahi-6.2-5 has that pmgr failure on j314 here. dts are correct and everything built fine
12:39marcanwhich one?
12:39marcan[ 0.138139] apple-pmgr-pwrstate 292280000.power-management:power-controller@a0: always-on domain atc3_usb_aon is not on at boot
12:39marcanthis is a devicetree bug
12:39chadmedno not that one
12:39chadmedit fails to boot, cant find the power domain for (seems to be) nvme
12:40jannau dcp_txt@80020c000 looks broken
12:41marcanthat looks right to me, unless the length is excessive
12:42marcanchadmed: I'm confused, worked fine for me? did you update everything including all the device trees etc?
12:43chadmedyeah new kernel, new initramfs, ran update-m1n1 pointing at the correct dts
12:43jannauiirc the dcp reserved mem is all at the end
12:43chadmedlet me try again hang on im in macos
12:43jannauI saw that on thursday as well
12:43chadmed has quit [Quit: Page closed]
12:44marcanoh hm ok
12:44marcandid something change in the driver core for this?
12:44marcanjannau: text should be in lowmem, covered by CTRR
12:46jannauah, yes. only t8112 has txt mapped via dart
12:47___nick___ has quit []
12:49___nick___ has joined #asahi-dev
12:50___nick___ has quit [Remote host closed the connection]
12:52___nick___ has joined #asahi-dev
12:53marcanisn't the ans2 hierarchy backwards in t600x? at least it's backwards from how it is on t8103
12:54chadmed has joined #asahi-dev
12:54marcanlet me check the adt...
12:54marcanyeah lol
12:54marcanwe forgot to fix that
12:54marcanwait no it's right
12:54jannauae8ac19655e0cd7ca5ac2f45837d386745aaed7f looks relevant
12:55marcanps_ans2 os !on but ps_apcie_st_sys and ps_apcie_st1_sys are <- that is definitely an illegal combination since there is a dep
12:55marcanso this sounds like a driver core regression
12:56chadmedapple-pmgr-pstate 28e080000.power-management:power-controller@2XX: failed to add to parent domain: -22
12:56chadmedthats what im getting
12:56marcanwait hold on
12:57marcanI bet it's e7d90cfac5510f8c94baa18f9f3f7808859c8332 + ae8ac19655e0cd7ca5ac2f45837d386745aaed7f
12:57marcanor something related
12:57nyilas has joined #asahi-dev
12:58marcanhm but that doesn't make sense
12:59marcanare we probing racily?
12:59marcanhere's the thing
12:59marcanps_apcie_st_sys probes, gets deferred because ps_ans2 hasn't probed
12:59marcanps_ans2 probes
12:59marcansome random mailbox for nvme probes
12:59marcanmailbox supports runtime-pm now
12:59marcanrpm suspends ps_ans2
13:00marcanps_apcie_st_sys probes again
13:00marcanfails because it finds itself on but ps_ans2 off
13:00marcanthis is, again, the problem of domains getting shut down greedily as soon as *one* consumer has probed
13:00marcanit is particularly bad with subdomains
13:01marcanwe need to do something to make sure all the subdomains probe before anything else can start shutting things down
13:01marcanmaybe using syscon/simple-mfd for this was a bad idea...
13:04marcanargh it's because it's a different pmgr altogether
13:04marcangoddammit we need some real solution for this mess
13:06marcanI think we can probably fix it with some ugly custom module init stuff in apple-pmgr-pwrstate, so that it greedily keeps scanning for and probing devices at load time until everything is probed, instead of letting EPROBE_DEFER go back to the driver core
13:07jannauuboot issue seems to CONFIG_LMB_MAX_REGIONS=8 being too small
13:08marcanthat's easy then
13:12jannaunice, the sd-boot menu is fixed
13:13kettenismailine u-boot already bumped CONFIG_LMB_MAX_REGIONS to 16
13:13kettenisdo we anticipate needing more?
13:13marcanI think what we *really* want here is something in the genpd core to make it refuse to shut down any PDs that have consumers in OF that have not probed yet
13:14marcankaprests: I think we may need way more for dcpext on e.g. M1 Ultra? if this is about reserved-regions
13:14marcansince there's like 8 of those
13:16marcanblergh this PD thing is going to suck to fix
13:16marcanlet's just make ans2 always-on for now
13:20chadmed has quit [Quit: Page closed]
13:22jannaumarcan: https://github.com/AsahiLinux/u-boot/pull/9 for a defconfig chnage for CONFIG_LMB_MAX_REGIONS
13:22marcandone, thanks
13:23jannaukettenis: hard to say, depends on the regions being consectutive
13:24jannau8 worked so far only because u-boot combines consectutive regions
13:25Cyrinux9Hey, marcan, I just update my m2 air, I can't boot anymore https://ibb.co/g9GD2GP
13:25marcanCyrinux9: too late, already on it
13:25Cyrinux9If this can help. Should I update macosx blob? Ok sorry
13:25kettenismaybe I should just bump it to 64 for m1; it is not as if we're starved for memory
13:25jannaupovik: are you sure that increasing CONFIG_LMB_MAX_REGIONS didn't fix the sio issues
13:25marcanCyrinux9: unless you intend your system to break, you shouldn't be using asahi-dev
13:26marcan(so this is on you to fix, we don't support asahi-dev)
13:26jannaukettenis: I'll send a patch
13:26Cyrinux9Don't worry :) I will fix from recovery when fix ready
13:26marcanCyrinux9: reverting m1n1/boot.bin to boot.bin.old from recovery will probably work enough to boot to linux and update again
13:27kettenisjannau: cool
13:27Cyrinux9OK, will do now
13:29kettenisjannau: did you reach a conclusion about pcie?
13:29MajorBiscuit has joined #asahi-dev
13:30MajorBiscuit has quit []
13:30MajorBiscuit has joined #asahi-dev
13:33jannaukettenis: no, only thing which worked are workarounds: 1. do not init already up ports and 2. skip ports with pwren-gpios in u-boot
13:34marcanrobher: is OF probe order across buses/sub-buses defined at all? we have a problem with pmgr nodes, where some nodes probe before their power-domain parents, so they get deferred. then some other device probes, and as the first consumer goes into suspend and the parent gets powered off. then the deferred child probes, and bails when it finds itself on but the parent off, which is illegal.
13:34marcanthis is a deeper problem we also have with e.g. DCP (where its PD is illegal to turn off until the DCP driver probes), but that's a bit of a special case
13:34marcan(we hack around that one by marking it always-on and then the DCP driver unmarks it)
13:34marcanhowever, PD dependencies not being honored are a much bigger problem
13:35jannaumarcan: booted with updated u-boot from asahi-dev
13:35marcanso basically we need some way for the pmgr driver to synchronously ensure the entire hierarchy it is responsible for is probed, not just a single pass with stuff ending up deferred
13:37chadmed has joined #asahi-dev
13:37marcanI know OF already has some stuff to pre-populate fwnode links based on power-domains properties, maybe what we should do is make the PM core refuse to power off a domain if there are any links that haven't probed yet?
13:38chadmedaaand we're back, apple,always-on in ps_ans2 sorted it
13:38marcanwe don't really want that since it makes nvme PM worse, but for now it will do
13:38Cyrinux9the workaround works, thanks marcan
13:39chadmedodd that it didnt affect your machine, is that code path racey?
13:39marcanit would depend on probe order, yes
13:40svenlovely :(
13:41marcanwe can also do the same thing DCP does for now and unmark that flag in the nvme driver
13:42marcanI'll try doing it "properly" with some fwlink based thing in the PM core
13:42marcanbut if people don't like that then hack it is
13:42marcannew kernels should be up too
13:44marcanthe "proper" fix might have to wait until later though, since I have way too much stuff on my plate first
13:44Bey0ndB1nary has quit []
13:44* svenhas been busy with $work that actually pays my bills as well :D
13:45chadmedcouple of other pcie niggles seem to have cropped up too https://tpaste.us/RBEv
13:46gladiac has joined #asahi-dev
13:46jannauthat warning shouldn't be there
13:47marcanI think I saw the patch for that
13:48marcanjannau: added your fix to bits/030-misc
13:48kettenishah: [ 0.000000] efi: EFI v2.100 by Das U-Boot
13:48kettenisthat version number isn't printed quite right
13:48jannauchadmed: do you have d57163d701f2 and aad9b063352e?
13:49marcanoh wait I had that in 130 already
13:49marcanI do get the warnings though
13:50DarkShadow44 has quit [Quit: ZNC - https://znc.in]
13:50jannaume too, I'll look into it
13:50chadmedjannau: yeah they made it into asahi-6.2-5
13:51chadmedill pull down the new tag and test tomorrow anyway
13:52jannauno need, I see it with asahi-6.2-6 as well
13:55MajorBiscuit has quit [Ping timeout: 480 seconds]
13:58jannauI blame that on u-boot's pcie code. I don't see those warnings when booted directly through m1n1
14:00jannauI disn't realize that until now that we use u-boot with PCIe support
14:01pzc has joined #asahi-dev
14:01DarkShadow44 has joined #asahi-dev
14:01jannaumarcan: did you already test on mini, imac or studio? I expect ethernet and usb xhci (on imac, studio)
14:01robhermarcan: only parent-child dependencies. The device that defers should have a dependency on the PD provider and not attempt to probe. But how is the PD turned off if its driver hasn't probed yet?
14:02marcanrobher: PDs turn off when their *first* dependency probes and idles itself
14:02marcanso then another dependency comes along that expected that not to happen and it breaks
14:03veeyee has joined #asahi-dev
14:03robherThat's probably broken for any provider. Or if you defer for provider A, and then provider B is disabled.
14:04marcanthis is a particular problem for PDs because hardware has interesting rules for that, like you can't just turn off some power domains that are on at boot without first hooking up their dependencies
14:05marcan(because hardware is already on at the bootloader and you can't just yank power from it, or because it's outright a dependent PD)
14:05chadmed_ has joined #asahi-dev
14:05robherA shared clock (or parent) would have the same issue.
14:05marcanso I'm thinking the genpd core should refuse to power off PDs if there are any unprobed/unfulfilled fwlinks
14:06marcanI don't particularly care about clocks for this platform so someone else can worry about that :-)
14:06robherAnd if it was on already at boot.
14:07marcanif that sounds like a sensible solution to you I'll go with that
14:07marcan(but not now, this is going on my queue for later)
14:07robherI think there's been some binding proposal for 'boot clocks'. It's been a while back.
14:08povikjannau: re max_regions. i will double check. i haven't tried that other option kettenis suggested
14:09jannaumarcan: confirmed, broken ethernet and usb pcie xhci on the studio. easiest workaround is disabling pcie in the u-boot defconfig
14:09marcanyeah okay let's do that
14:09marcanwant to send the patch?
14:10jannauotherwise I have workaround patches for linux and u-boot
14:10marcanat this point I just want to get this out ASAP, we've delayed this release long enough
14:10marcanmore involved fixes can wait
14:10chadmed_ has quit [Remote host closed the connection]
14:11marcanI'll send the wifi fix
14:11chadmed has quit [Read error: No route to host]
14:11jannauok, I'll prepare the u-boot config change
14:11robhermarcan: That could leave domains powered forever if there is not a driver. It's a common problem because we never know 'all drivers are loaded' except for !CONFIG_MODULES
14:12chadmed has joined #asahi-dev
14:12marcanrobher: yeah, but what are we going to do? if it was on at boot, and the DT describes devices that depend on it, we can't know it's safe to power it off
14:12marcanI would say that's a platform bug
14:12marcaneither ship the drivers or have the bootloader turn it off
14:13marcaneither that or we add a new DT property to mark devices that may require PDs active at boot...
14:15robhermarcan: It might be better to have some hint in the DT as to what domains are critical. Maybe simple-framebuffer node needs a power-domains property (though the display phandle link could provide that).
14:17marcanyes, that is the case for us, though that's just one instance of an issue
14:17marcanbut then really every domain that has PD children is effectively critical, since turning off a parent before children is a bad idea
14:22marcannew u-boot in asahi-dev
14:24marcanwarnings gone
14:25robhermarcan: I suppose the PD driver could prevent that assuming it's all one driver and instance. Otherwise, if your refcounts and actual h/w state are mismatched, kind of hard for s/w to do the right thing,
14:26marcanthe way we bind the PD driver it's one instance per PD, which is what is causing the issue (but actually even if it were a whole driver for a set of PDs, we have a cross-PD-set dependency and those would probably be separate instances no matter what)
14:26marcan(the bigger chips have multiple PMGRs)
14:27marcanit is all one driver though, so I think it could prevent that by using a custom module init function that makes triple sure everything is probed until nothing defers, and the driver core never sees the defers (or something like that)
14:27marcanbut it feels really ugly
14:29jannauethernet+xhci restored on the studio
14:29marcanOK, I *think* we dealt with all the explosions so far
14:29marcanI should get some dinner+sleep
14:30marcanplease do continue to test stuff
14:30marcanideally I'd like to release this next week or so
14:30marcanI'll give it a spin on some more machines myself
14:41roxfan has joined #asahi-dev
14:45roxfan2 has quit [Ping timeout: 480 seconds]
14:47Hibyehello_ has joined #asahi-dev
14:47roxfan2 has joined #asahi-dev
14:48c10l has quit [Quit: Bye o/]
14:49roxfan has quit [Ping timeout: 480 seconds]
14:49Hibyehello has quit [Ping timeout: 480 seconds]
14:51c10l has joined #asahi-dev
14:53veeyee has quit [Quit: Leaving.]
14:54veeyee has joined #asahi-dev
15:01c10l5 has joined #asahi-dev
15:03c10l has quit [Read error: Connection reset by peer]
15:13veeyee has quit [Ping timeout: 480 seconds]
15:23Bey0ndB1nary has joined #asahi-dev
15:25Bey0ndB1nary has quit [Max SendQ exceeded]
15:28Bey0ndB1nary has joined #asahi-dev
17:33Glanzmann has joined #asahi-dev
17:36Hibyehello_ has quit [Remote host closed the connection]
17:36Hibyehello has joined #asahi-dev
17:41bpye has quit [Quit: The Lounge - https://thelounge.chat]
17:42bpye has joined #asahi-dev
18:04GlanzmannDisabling freedreno let me compile mesa. Kernel boots, xorg starts.
18:14MajorBiscuit has joined #asahi-dev
18:14Hibyehello_ has joined #asahi-dev
18:14Glanzmannmarcan: You probably won't do anything about it, but with xorg and xfeerdp, I have still display artefacts. The rosa tiles when I log in for a fraction of the second as with jannaus video and when I open firefox in a rdp session and zoom in/out the result is: tg.st/u/screenshot-air-2023-03-11-19_10_58.png
18:17GlanzmannI use this command line: xfreerdp +clipboard:use-selection:PRIMARY /u:Administrator /p:'Pa$$w0rd' -grab-keyboard /cert-ignore /f /bpp:32 /v:zh17w2k22.glanzmann.de -gfx /rfx
18:18GlanzmannIf you or someone else needs a rdp account to reproduce, let me know.
18:21Hibyehello has quit [Ping timeout: 480 seconds]
18:21Glanzmannmarcan: A regression for me: When I type 'reboot' in Linux. The screen turns off, but I think the system does not. It does not reboot. But when I press the power button for a long time, release it, press it again, the system boots again.
18:21GlanzmannNothing in /var/log/kern.log unfortunately.
18:22Bey0ndB1nary has quit [Remote host closed the connection]
18:23Bey0ndB1nary has joined #asahi-dev
18:23Glanzmannhalt works.
18:32nyilas has quit [Remote host closed the connection]
19:12Bey0ndB1nary has quit []
19:20roxfan2 has quit [Ping timeout: 480 seconds]
19:30Bey0ndB1nary has joined #asahi-dev
19:49Hibyehello has joined #asahi-dev
19:51Hibyehello_ has quit [Ping timeout: 480 seconds]
20:14GlanzmannTHe debs are here: https://tg.st/u/asahi-debian-2023-03-11.tar - I also push to 'm1-debian'.
20:49MajorBiscuit has quit [Quit: WeeChat 3.6]
21:03sam_- is now known as sam_
21:06___nick___ has quit [Remote host closed the connection]
21:07___nick___ has joined #asahi-dev
21:14roxfan has joined #asahi-dev
21:39hightower2 has joined #asahi-dev
21:48jordy_ has joined #asahi-dev
22:47jordy_ has quit [Remote host closed the connection]
22:53chadmed has quit [Remote host closed the connection]
23:18chadmed has joined #asahi-dev
23:28Bey0ndB1nary has quit []
23:46bluetail9 has quit [Ping timeout: 480 seconds]
23:57chadmedthe workaround for the ANS power domains seems to siphon about 20 mins of battery life all else being equal
23:58chadmeddid lina end up fixing the gpu pstate thing for this release or are we still stuck in a single pstate
…