2021-09-29
00:05bgb_ has quit [Ping timeout: 480 seconds]
00:15bgb_ has joined #asahi
00:24bgb_ has quit [Ping timeout: 480 seconds]
00:33bgb_ has joined #asahi
00:33bdju has quit [Ping timeout: 480 seconds]
00:38marcanGlanzmann: when it asks you to press enter to continue, are you sure you're not pressing enter twice?
00:55bgb_ has quit [Ping timeout: 480 seconds]
00:58marcanafter you pick the boot disk, it's supposed to show you another message, then shut down after you press enter again
00:58marcanthe only way it can shut down without you seeing that is if you pressed enter twice and it buffered the input
01:00DragoonAethis has quit [Quit: hej-hej!]
01:00marcanfirefox317: I think there is some weird situation where it caches things... I'm not entirely sure what's up with that
01:00DragoonAethis has joined #asahi
01:01roxfansven: nubs are explained here: https://developer.apple.com/library/archive/documentation/DeviceDrivers/Conceptual/IOKitFundamentals/ArchitectOverview/ArchitectOverview.html
01:03bgb_ has joined #asahi
01:04marcanI'm pretty sure they're either overloading the term or using it to describe something "similar" in hardware
01:12bgb_ has quit [Ping timeout: 480 seconds]
01:19bgb_ has joined #asahi
01:24jhweruyuw has quit [Remote host closed the connection]
01:28yuyichao_ has joined #asahi
01:45bgb has joined #asahi
01:47bgb_ has quit [Ping timeout: 480 seconds]
01:49tylo has joined #asahi
01:54bgb_ has joined #asahi
01:58___nick___ has joined #asahi
01:59bgb has quit [Ping timeout: 480 seconds]
02:39PhilippvK has joined #asahi
02:42phiologe has quit [Ping timeout: 480 seconds]
03:18marvin24 has joined #asahi
03:22marvin24_ has quit [Ping timeout: 480 seconds]
03:31marvin24_ has joined #asahi
03:34marvin24 has quit [Ping timeout: 480 seconds]
03:35bgb has joined #asahi
03:37Gue___________________________ has quit [Quit: Textual IRC Client: www.textualapp.com]
03:38bgb_ has quit [Ping timeout: 480 seconds]
04:06psykose has quit [Ping timeout: 480 seconds]
04:07psykose has joined #asahi
04:15marcanjust updated the installer at mrcn.st/alxsh, now it properly supports multiple macOS installs in the same APFS container, and it guards against people pressing enter too many times
04:42___nick___ has quit [Ping timeout: 480 seconds]
05:00bgb has quit [Remote host closed the connection]
05:01bgb has joined #asahi
05:21lethalbit has quit [Quit: kill -9 -1]
05:21lethalbit has joined #asahi
05:21hexchain has quit [Quit: Goodbye!]
05:25hexchain has joined #asahi
05:42Glanzmannmarcan: I probably only pressed enter once, but I tried at least three times, so on these three times, I only pressed it once at least one time. :-)
05:47Glanzmannmarcan: Bad news. Boot picker does not show m1n1: https://ab34.de/u/Screenshot_2021-09-29_at_07.46.53.png
05:49Glanzmannmarcan: I do the following: I delete the old m1n1 partition, than it also increases my macos partition, so I use the diskutil command to make 50 GB of space.
05:49GlanzmannThan I run your installer, but the bootpicker only shows me the macos, not the m1n1 as with my mac mini.
05:50GlanzmannIf you want I can also stream my desktop while doing it or make a recording and send it to you, maybe I'm doing something else wrong.
05:56Glanzmannmarcan: I'll send you the output, but this one has two recovery os already on it. And it only offers me to install 11.4 boot firmware.
06:02Glanzmannmarcan: On my wifes laptop, the one that I never tinkered with, same thing: https://ab34.de/u/Bildschirmfoto_2021-09-29_um_08.00.42.png
06:47nsklaus_maybe i'm wrong but i was under the impression macos and the whole system itself, used that partition as a both a bios and a bootloader too
06:50bgb_ has joined #asahi
06:52marcanyou're going to have to flash with Apple Configurator 2; a "revive" *might* fix it, if it doesn't you'll have to do a full erase
06:53marcanhttps://support.apple.com/guide/apple-configurator-2/revive-or-restore-a-mac-with-apple-silicon-apdd5f3c75ad/mac <- official docs
06:54marcanas for the new OS not showing up in the boot volume selector, there's some weird caching thing
06:54bgb has quit [Ping timeout: 480 seconds]
06:54marcanI also suspect it might have something to do with deleting m1n1 then re-creating it in the same boot
06:55aleasto has joined #asahi
06:56Glanzmannmarcan: As you said, it shows up: https://ab34.de/u/Screenshot_2021-09-29_at_08.55.32.png
06:56marcanplease experiment to see if you can figure out something to do to make it show up without a reboot
07:02GlanzmannHowever the installer is really nice even if we need to reboot inbetween to make m1n1 partition show up.
07:05GlanzmannBut it was the most recent stable release. My laptop was not and my wifes laptop is on an even older release, so probably the issue is already gone.
07:19Glanzmannfirefox317: In case you missed, reboot your system, go to system preferences and pick the m1n1 parition to boot from, execute step2 and you should be good to go.
07:22FireFox317Glanzmann, i actually figured this out myself yesterday ;p
07:23FireFox317should have tagged you, but i did say something about that on irc last night
07:33Glanzmannalyssa: Thank you for turning my mac mini in a linux desktop, I'm using it all day, and everything works like a charm. Had to wake it up with xrandr :-). I have to find my bag with usb device though so that I get sound.
07:38FireFox317roxfan, i think the iokit documentation link you posted might be very useful for alyssa (the dcp god) ;p
07:47everslick has quit [Remote host closed the connection]
07:47everslick has joined #asahi
07:47Glanzmannmarcan: Do you plan on extending the installer to install a complete linux distribution?
07:59FireFox317marcan, killing storagekitd before opening the ´Startup Disk´ fixes it?
08:00FireFox317Okay, I can try in a bit :)
08:00GlanzmannI'm debian guy for 10 years, but to be honest the installer could easily support multiple distributions, just put another rootfs.
08:00marcanI'm looking into storagekitd itself, if it lets me connect to the service I might be able to bypass the picker and just do everything internally
08:00marcanthe installer will support multiple distros, to the extent that people are willing to maintain them :)
08:01marcan(but I expect a bit more than a rootfs; whatever we support needs to have proper package builds and add-on repos and such)
08:02FireFox317marcan, are you sure you pushed? last changes seem to be 4 hours ago
08:05bgb has joined #asahi
08:06FireFox317i was btw creating the installer.tar.gz myself, such that I could get the latest version of m1n1 installed
08:06FireFox317probably doesnt matter that much, since I can just chainload the new m1n1 instead, but whatever xd
08:07bgb_ has quit [Ping timeout: 480 seconds]
08:10Glanzmannfirefox317: Depends on your scenario, but it was also getting on my nerves. I should have put the one from alyssa with the patch.
08:14Glanzmannmarcan: Did you buy a mini with a 10 Gbit/s nic? Do we have a working driver for that?
08:22marcanj_ey: according to the drhystone benchmarks I ran yesterday to get dmips/mhz numbers for the device tree, they matter less than 20% as much as the big cores
08:24aleasto has quit [Quit: Konversation terminated!]
08:25aleasto has joined #asahi
08:26marcanit's about 32% of the performance at the same clock, but the p-cores run more than 50% faster on top of that
08:37FireFox317did this not happen before
08:42FireFox317marcan, hmm are you sure? I managed to break it twice in succession
08:42FireFox317let me try one more time
08:43FireFox317with power attached
08:47chadmedi am very very very excited to see how fast the gpu is specifically, the fact that the 7 core AGX in the air can run binary translated games at 100+ fps is incredible
08:50FireFox317marcan, okay yeah that was flaky internet. Also the Asahi Linux showed up in the 'Startup Disk' this time, however when clicking on the logo it did not actually change the boot mode.
08:51FireFox317I had to go back one to the main menu of the Settings, and then it asked me to authenticate again. Then I went back to the 'Startup Disk' and I could change it and it would take effect and the script continues.
08:59sorearbinary translation won't sandbag graphics workloads (unless you have precompiled shaders you're binary translating, but that's very rare outside console games)
09:02chadmedplus the source engine doesnt even have a native metal, vulkan or opengl pipeline. it uses metalvk on macos and togl on gnu/linux
09:03chadmedso still extremely impressive that the cpu can feed the agx enough draw calls to get that much performance out of it, and just impressive in general that they can eke that much performance out of suhc a low power chip
09:06FireFox317Glanzmann, yes! It's just that i had to reautenticate in the previous screen for whatever reason
09:15FireFox317Maybe it has something to do with TouchID.
09:19FireFox317marcan, yeah that wouldn't be suprising either
09:19FireFox317at the end of the day it wasnt really that hard to install m1n1, so thats super nice
09:45marcanI *think*, unless there's an entitlement to block me, I can use StorageKit.framework to talk to storagekitd and request a bless/setboot (which is what the System Preferences pref pane does behind the scenes)
09:46FireFox317marcan, lets hope that works. That would be a very nice solution.
10:09bgb_ has joined #asahi
10:13bgb has quit [Ping timeout: 480 seconds]
10:23bgb has joined #asahi
10:29bgb_ has quit [Ping timeout: 480 seconds]
10:32linearcannon has quit [Remote host closed the connection]
10:32bgb_ has joined #asahi
10:34linearcannon has joined #asahi
10:37bgb has quit [Ping timeout: 480 seconds]
10:46tylo has quit [Ping timeout: 480 seconds]
10:49tylo has joined #asahi
10:57tylo has quit [Ping timeout: 480 seconds]
11:45marcansigh, working out private frameworks is such a PITA... seems the automated dump tools can't work out block arg signatures :(
11:50alyssaGlanzmann: :-) spent some time last night investigating the hotplug breakage with i3/your wm but not gnome/fbcon. hopefully will figure out a fix soon
11:51Glanzmannalyssa: If you have something, let me know and I'll test it. I'm working the whole day and have no issues, I have even sound, however some kernel messages: https://pbot.rmdir.de/W7urrfwnn23H1NmE9_kRsA
11:52GlanzmannAlso I have to benchmark the nvme under macos and linux and compare, because the nvme appears to be slow on random operations: 80 jobs 4k random write: Jobs: 80 (f=80): [w(80)][0.0%][w=5052KiB/s][w=1263 IOPS][eta 01d:17h:46m:43s]
11:52alyssaGlanzmann: right, the second WARN_ON about drm_atomic_connector_commit_dpms is related to the same bug as Xorg hotplug
11:53GlanzmannI see. Perfect. I like your soundcard handy adapter solution, but I probably threw mine away when I bought a mobile phone with a 3.5 jack.
11:55alyssa(shutdown, wait for everything to go down and NVMe to finish up syncing and all, and then hold down power button 10 secs to force it off is the workaround)
11:56GlanzmannI see, I used ssh to wake it up using xrandr. And I think the mode change and rotation is very fast. I have an intel nuc that one is slower when I do these operations.
11:56FireFox317marcan, hmm that is annoying
11:57alyssaGlanzmann: Dealing with this is on the todo list but it's out of my usual depths... Corellium has patches but they're not suitable to be cherry-picked even into my downstream. sven has bigger fish to fry. So that sounds like a marcan problem maybe :-p
12:03alyssa(Time permitting, I'd like to do the GPU kernel driver after marcan does the needed r/e. But we'll see what's involved.)
12:04Glanzmannalyssa: I don't know anything about graphics, so once the gpu is reverse engineered, will it be possible to offload mpeg strems to the graphic card?
12:06FireFox317Glanzmann, encoding and decoding of a/v is actually done by another coprocessor in the m1 chip
12:06FireFox317the gpu is basically optimized to to 3d rendering
12:08GlanzmannSo once the gpu is reverse engineered, we need a opengl layer and than an xserver that uses that opengl layer?
12:09chadmedi think you misunderstood. r/e-ing the GPU will not help with video encode/decode. for that we need the hardware block responsible for encode/decode
12:10FireFox317Glanzmann, basically yes. But there is a really nice project called mesa (https://gitlab.freedesktop.org/mesa/mesa), which already implement abstractions for opengl etc and only a backend has to be written (which alyssa already started with)
12:11FireFox317klange, true lmao (english isnt my native language xd)
12:14sorear3d graphics and mpeg decode are handled in different hardware blocks on PCs too, but there are cultural differences in whether people use the term "GPU" in a wide or narrow sense
12:15sorearwell card would be physically inaccurate if you have integrated graphics, which is the situation here
12:16sorearmobile socs always have integrated graphics, so there's no frequent need to talk about the "rendering engine, media codec, and display controller" as a group
12:17chadmedbut it is PC users that this platform is trying to win over, so theyre going to come with their own ideas and abstractions in mind which is where the confusion is
12:18chadmedas a tangent its pretty awesome that these macs are basically the first viable _and popular_ non-x86/amd64 workstations in a very long time
12:34marcangiving up on pyobjc, I had to hack their binding generator to even get it to work, and now it's triggering internal cpython errors
13:02sorearyou can get raw disk access but you can't change the default boot partition without an entitlement?
13:03marcanmore specifically, you can't bless a volume without an entitlement, because that requires creating a boot policy which requires talking to the SEP
13:17sventhe SEP gets send some cdhash which is apparently the hash of a signature of the binary or something but if you're running in XNU kernel context you can just fake that
13:17svenand actually creating a policy for m1n1 as a custom boot object requires you to be booted in to 1TR
13:18sorearenjoying all the implicit trust on the window server, GUI libraries, keyboard and display drivers...
13:19sventhe critical security boundary is 1TR. and you don't get into 1TR without long-pressing the power button
13:19marcanthe annoying thing here is that you also don't even get to bless the volume normally without an entitlement
13:20marcanbut it looks like we're going to have to stick with that hideous boot picker instruction + race thing for now
13:20FireFox317there is a bless tool on my macbook air tho, but that can only be used on intel macbooks?
13:21chadmedim inclined to ask if the recent code signature translocation exploit could help here but its probably not worth relying on exploits that can and will be patched
13:23chadmedyeah exactly, its like following clues left for us by the engineers rather than "hacking" in the colloquial sense
13:24svendon't get me wrong, those are pretty impressive. i've just burnt out from the whole secrecy and cat-mouse-game BS
13:26chadmedyeah true. given the recent developments wrt the bug bounty program seems like this particular cat has gone for its siesta though ;)
13:38chadmed+2 though the crushing guilt walking into the gym after having a massive poutine is *almost* not worth it
13:48chadmedreally tempted to post that copypasta re the kid that gets ridiculed by their class for using linux lmao
13:51chadmed has quit [Quit: Konversation terminated!]
13:53yuyichao_ has quit [Ping timeout: 480 seconds]
13:59tylo has joined #asahi
14:04marcanand yeah, I started this project, among other things, because I get to work on reversing hardware *without* exploits
14:04marcanand I'd like to keep it that way for future hardware, so I also try not to do things that might annoy Apple ;)
14:08chadmed has joined #asahi
14:16marcan(it's why I can say my software ships by default in ubuntu, or at least used to at some point)
14:20yuyichao_ has joined #asahi
14:58patience[m] has joined #asahi
15:16linearcannon has quit [Read error: Connection reset by peer]
15:21linearcannon has joined #asahi
15:24linearcannon has quit [Remote host closed the connection]
15:24linearcannon has joined #asahi
15:43linearcannon has quit [Read error: No route to host]
15:44linearcannon has joined #asahi
16:10linearcannon has quit [Remote host closed the connection]
16:11linearcannon has joined #asahi
16:24linearcannon has quit [Remote host closed the connection]
16:24linearcannon has joined #asahi
16:35linearcannon has quit [Read error: Connection reset by peer]
16:36linearcannon has joined #asahi
17:23nsklaus_maybe this question have already been asked, but, since M1 arch is spanning over a few hardware products at apple, would there be a chance to get asahi linux running on ipad and iphone, someday ? or is there some fundamental difference/roadblock preventing it ?
17:24i509vcb[m]I imagine the ios devices are locked down much more than the macs
17:24i509vcb[m]Would some of the work we have done on hardware and drivers be congruent, possibly
17:24i509vcb[m]s/we have//
17:25nsklaus_hmm i see.. maybe later on then.. if apple smells the coffee and decide to allow it, like it does now on mac..
17:27i509vcb[m]iOS has been locked down like this for nearly a decade, so I imagine it isn't going to change like the macs
17:27i509vcb[m]* for nearly (or more I can't recall) a decade,
17:28i509vcb[m]Not going to deny that running Linux on the iDevices would be cool
17:29i509vcb[m]But it isn't an easy task either
17:30alyssaAlso even with exploits, there's not really a... point in Linux on the iDevices like there is with the M1 Macs
17:32nsklaus_one has to wonder why on earth apple would decide to make such a difference across its products, i mean, why introduce such a new possibility on mac ? logically i would have expected they would have pushed toward some for of coherence across all the product lines
17:34i509vcb[m]I recall the Windows for ARM being a legal nightmare at the moment for licensing, so the firmware may have been designed for bootcamp, but even then the fact apple wants you to use the virtualization in macos conflicts with that
17:34i509vcb[m]yeah just speculation on my part
17:34kdrag0n[m]hi, the drivers mentioned in https://twitter.com/marcan42/status/1442871737419321347?s=19 haven't been pushed anywhere yet, right?
17:35kdrag0n[m]I saw pmgr but nothing related to cpufreq
17:35kdrag0n[m]asking because I ran some benchmarks for capacity-aware scheduling, but frequencies above 2.988 GHz don't seem to do anything with Corellium's cpufreq driver
17:38nehsou^ has quit [Ping timeout: 480 seconds]
17:38nehsou^ has joined #asahi
17:38Glanzmannmarcan: Thank you for the instructions. I try to revive my macbook air now. Btw. in order to boot into recovery mode, one need only need to press the power button. Not the whole left control option left shift thing.
17:45linearcannon has quit [Read error: Connection reset by peer]
17:51dottedmagalyssa: I can envision using an old iPad (one that can't even get any apps installed from App Store) as a very fine controller for a smart home. One can't really get iOS to use a Z-Wave dongle, can one?
17:52dottedmagWeb interfaces are quite clunky, and I'm not sure how to sideload native apps on iOS properly without an exploit.
18:00opticrondottedmag, the best solution there that I've seen is to attach the dongle to a pi running hass.io and then you get a pretty web interface
18:00opticronmost people don't want the control interface to be the actual controller for the z-wave/zigbee/etc network
18:04kdrag0n[m]j_ey: ah ok
18:11Glanzmannmarcan: Revive failed, going for restore now. https://ab34.de/u/Screenshot_2021-09-29_at_20.09.58.png
18:19X-Scale` has joined #asahi
18:23X-Scale has quit [Ping timeout: 480 seconds]
18:53Glanzmannmarcan: I did it wrong, you have to press the three buttons. Now I'm in DFU mode, hopefully this allows me to revive or restore. Keep you posted.
19:05erincandescent has quit [Remote host closed the connection]
19:06erincandescent has joined #asahi
19:27tylo has quit [Ping timeout: 480 seconds]
19:41tylo has joined #asahi
19:51_yt_ has joined #asahi
19:51Glanzmannmarcan: Thank you for the instructions, I was able to restore the mac mini. During the process it filled up 180 GB of disk space on the mac mini because of failed tries. :-)
20:16mort_ has quit [Quit: Ping timeout (120 seconds)]
20:16mort_ has joined #asahi
20:17dottedmag_ has joined #asahi
20:19d4ve has quit [Ping timeout: 480 seconds]
20:19d4ve has joined #asahi
20:19maximus64 has quit [Remote host closed the connection]
20:20vup2 has joined #asahi
20:21nehsou^ has quit [Remote host closed the connection]
20:22dottedmag has quit [Ping timeout: 480 seconds]
20:22dottedmag_ is now known as dottedmag
20:22Esmil has quit [Ping timeout: 480 seconds]
20:22vup has quit [Ping timeout: 480 seconds]
20:25Esmil has joined #asahi
20:26yuyichao_ has quit [resistance.oftc.net weber.oftc.net]
20:26sjg1 has quit [resistance.oftc.net weber.oftc.net]
20:26esden has quit [resistance.oftc.net weber.oftc.net]
20:26sorear has quit [resistance.oftc.net weber.oftc.net]
20:26philpax has quit [resistance.oftc.net weber.oftc.net]
20:26guan has quit [resistance.oftc.net weber.oftc.net]
20:26kenzie has quit [resistance.oftc.net weber.oftc.net]
20:26austriancoder has quit [resistance.oftc.net weber.oftc.net]
20:26ihaveamac has quit [resistance.oftc.net weber.oftc.net]
20:26Techcable has quit [resistance.oftc.net weber.oftc.net]
20:26cz3 has quit [resistance.oftc.net weber.oftc.net]
20:26Ariadne has quit [resistance.oftc.net weber.oftc.net]
20:26tardyp has quit [resistance.oftc.net weber.oftc.net]
20:26alyssa has quit [resistance.oftc.net weber.oftc.net]
20:26enomem has quit [resistance.oftc.net weber.oftc.net]
20:26tmlind has quit [resistance.oftc.net weber.oftc.net]
20:26tbodt has quit [resistance.oftc.net weber.oftc.net]
20:26_jannau_ has quit [resistance.oftc.net weber.oftc.net]
20:26mjg59 has quit [resistance.oftc.net weber.oftc.net]
20:26eichin has quit [resistance.oftc.net weber.oftc.net]
20:26maximus64 has joined #asahi
20:29yuyichao_ has joined #asahi
20:29esden has joined #asahi
20:29sjg1 has joined #asahi
20:29austriancoder has joined #asahi
20:29sorear has joined #asahi
20:29kenzie has joined #asahi
20:29philpax has joined #asahi
20:29ihaveamac has joined #asahi
20:29guan has joined #asahi
20:29Ariadne has joined #asahi
20:29tardyp has joined #asahi
20:29cz3 has joined #asahi
20:29Techcable has joined #asahi
20:29enomem has joined #asahi
20:29eichin has joined #asahi
20:29tbodt has joined #asahi
20:29_jannau_ has joined #asahi
20:29alyssa has joined #asahi
20:29tmlind has joined #asahi
20:29mjg59 has joined #asahi
20:39aleasto has quit [Ping timeout: 480 seconds]
20:55linearcannon has joined #asahi
20:57crucifix has joined #asahi
21:57crucifix has quit [Quit: Leaving]