#
…

2026-11-02

00:01Youwes09Hey, heads up for when you move past 7.3. Mainline 7.4 changed how Rust drivers hold DMA memory, so a buffer can now only exist while its driver is bound to the device. The Rust RTKit layer keeps driver state in something that has to outlive the driver, so PMP can't keep its buffers there anymore and won't compile. The C side is fine though.
00:02Youwes09I got around it by having the driver own the RTKit connection and shut it down on unbind (buffers live inside it). PMP and the GPU driver would need some small changes for that. I was wondering how you were planning to handle it though and whether my approach makes sense?
00:19salva_ has joined #asahi-dev
00:21salva has quit [Ping timeout: 480 seconds]
00:23salva_ is now known as salva
00:26ddxtanx has joined #asahi-dev
00:58Ardu0us_ has joined #asahi-dev
01:02myriaWoohoo, 7.4-rc1 is out!
01:02Ardu0us has quit [Ping timeout: 480 seconds]
01:05Youwes09 has quit [Ping timeout: 480 seconds]
01:06Youwes09 has joined #asahi-dev
01:19JamesCalligeros[m]Yay, i can send macaudio v4
01:24ohxorud has quit [Ping timeout: 480 seconds]
01:29chadmed has quit [Read error: Connection reset by peer]
01:29chadmed has joined #asahi-dev
01:38MichaelReeves[m]Same with dchid v5. Will have to sort out when feeling better
01:48fatcullen has quit [Quit: This computer has gone to sleep]
01:50fatcullen has joined #asahi-dev
02:02Youwes09 has quit [Remote host closed the connection]
02:32fatcullen has quit [Quit: This computer has gone to sleep]
02:32fatcullen has joined #asahi-dev
02:34myriawill also send out M3/M4 perf soon
03:00yurix has quit [Remote host closed the connection]
03:00yurix has joined #asahi-dev
03:03fatcullen has quit [Quit: This computer has gone to sleep]
03:08myriaHey, for people with an A18/M4 device and time on their hands, can they try and test https://github.com/sempiternal-aurora/linux/pull/1?
03:09myriaI didn't really have a dedicated testing methodology in mind. The main plan was just using the linux `perf` tool to set multiple raw event counters at a time, and ensure that multiple counters of the same type count the same thing, that no errors occurs when trying to run perf counts
03:09KimTalmo_ has joined #asahi-dev
03:09myriaThis is usually done with `perf -e r%EVENT_HEX_CODE%:u -- TOOL TO RUN`. `sleep 5` is what I've been using to test the counters over
03:11nicolas17I have access to an A18 device but I don't know where I'd begin loading a Linux kernel there :p
03:11myriaEvent hardware counter tables can be found at `drivers/perf/apple_m1_cpu_pmu.c`, under the `m3` or `m4` pmu events enum
03:13myrianicolas17: https://asahilinux.org/docs/sw/tethered-boot/ might help. Though A18 really only works on the macbook neo, since others I don't think we can run untrusted code on yet
03:17myria(For running multiple counters, this can be achieved with multiple `-e rXXXX:u` arguments to perf)
03:22tobhe_ has joined #asahi-dev
03:27fatcullen has joined #asahi-dev
03:29tobhe has quit [Ping timeout: 480 seconds]
04:13umo is now known as Guest4541
04:13umo has joined #asahi-dev
04:20Guest4541 has quit [Ping timeout: 480 seconds]
04:23nora has joined #asahi-dev
04:25nora_ has quit [Ping timeout: 480 seconds]
05:01chaos_princessnicolas17: derive key to feed to aea and decrypt root header to know the size in advance
05:35jnn has joined #asahi-dev
05:38jn has quit [Ping timeout: 480 seconds]
06:02[Kalisto] has quit [Quit: Ping timeout (120 seconds)]
06:02[Kalisto] has joined #asahi-dev
06:02[Kalisto] is now known as Guest4545
06:05mattia013 has quit [Quit: mattia013]
06:05mattia013 has joined #asahi-dev
06:43runxiyu_ is now known as runxiyu
06:51Youwes09 has joined #asahi-dev
06:51sven has quit [Quit: The Lounge - https://thelounge.chat]
06:52sven has joined #asahi-dev
06:55Youwes09myria: I have an M4 Air (J713) on 7.4-rc1, I can test the perf PR today
07:01myriaah, that would be amazing thank you!
07:33umo has quit [Remote host closed the connection]
07:33umo has joined #asahi-dev
08:00mattia013 has quit [Remote host closed the connection]
08:04AdityaGarg8[m]jannau: I had flagged informed here when we had that on T2
08:04AdityaGarg8[m]Someone said iommu issue isn't on arm Macs
08:05AdityaGarg8[m]Regardless, we have sent this upstream
08:05AdityaGarg8[m]It is in the Linux wireless tree
08:08AdityaGarg8[m]* I had informed here when we had that on T2
09:01Bhyve has joined #asahi-dev
09:29chadmed has quit [Read error: Connection reset by peer]
09:29chadmed has joined #asahi-dev
09:31chadmed has quit [Read error: Connection reset by peer]
09:31chadmed has joined #asahi-dev
09:32MichaelReeves[m]maz: did you want changes for the kvm errata, sorry I might have forgotten? checked 7.4-rc1 and it didn't seem they were in there
09:33chadmed has quit [Read error: Connection reset by peer]
09:34chadmed has joined #asahi-dev
09:34mazMichaelReeves[m]: it was sent way too late for being taken into the initial PR. I'll see if I can squeeze it in as a fix.
09:34mazbut that's rather low priority compared to the rest of the shit I need to get merged first.
09:49MichaelReeves[m]no worries, thanks for confirming!
09:51linuxgemini has quit [Quit: Ping timeout (120 seconds)]
09:51linuxgemini has joined #asahi-dev
10:14jannauAdityaGarg8[m]: thanks, that fixes things on bcm4388 as well
10:16ddxtanx has quit [Quit: Konversation terminated!]
10:19ddxtanx has joined #asahi-dev
10:21aissen_ has quit []
10:21aissen has joined #asahi-dev
10:25Bhyve has quit []
10:32lion328 has quit [Remote host closed the connection]
11:15ninelore has quit []
11:21ninelore has joined #asahi-dev
11:46lion328 has joined #asahi-dev
13:08apc is now known as Guest4559
13:08apc has joined #asahi-dev
13:12fatcullen has quit [Quit: This computer has gone to sleep]
13:14Guest4559 has quit [Ping timeout: 480 seconds]
13:14fatcullen has joined #asahi-dev
13:16fatcullen has quit []
13:28fatcullen has joined #asahi-dev
13:29bgb has quit [Remote host closed the connection]
13:30fatcullen has quit []
13:37bgb has joined #asahi-dev
13:41bgb has quit [Remote host closed the connection]
13:47bgb has joined #asahi-dev
13:48Bhyve has joined #asahi-dev
13:50fatcullen has joined #asahi-dev
14:02fatcullen has quit [Quit: This computer has gone to sleep]
14:10fatcullen has joined #asahi-dev
14:10svenok, is there anything I missed picking up from last cycle? the m5 device trees need a v2 I think
14:11svenah, I haven't picked the m4 pro/max trees either, have I?
14:32jannauM3* nvme support but I can send a v2
14:34apc is now known as Guest4563
14:34apc has joined #asahi-dev
14:37svenjust the message id is also fine
14:39Guest4563 has quit [Ping timeout: 480 seconds]
14:49KimTalmo_ has quit [Quit: bye]
15:06myriaM5 sart might be good? dt has been acked, not sure about soc requirements though
15:07myriaA18 I2C and pinctrl is waiting on review
15:07fatcullen has quit [Quit: This computer has gone to sleep]
15:07svenif dt is acked I can review SART myself and just merge it
15:08svendo you have a link or message id?
15:08svenfor the dt-style-check: does that need a new version because the checker changed or is it still fine?
15:12myriaalso, gentle poke to resend #7 of https://lore.kernel.org/asahi/20260829-b4-tbt-fixes-v3-0-e1fab6ac54fe@kernel.org/ now that 7.4-rc1 is out, if you have the time
15:12yukaCheck the links in this pad
15:12yukaOr the google docs patch tracker
15:16fatcullen has joined #asahi-dev
15:23croissant` has joined #asahi-dev
15:28bgb has quit [Remote host closed the connection]
15:30croissant has quit [Ping timeout: 480 seconds]
15:31bgb has joined #asahi-dev
15:43croissant has joined #asahi-dev
15:43Guest3859 has quit [Quit: Bridge terminating on SIGTERM]
15:43rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
15:49croissant` has quit [Ping timeout: 480 seconds]
15:50croissant` has joined #asahi-dev
15:53croissant has quit [Ping timeout: 480 seconds]
15:54svenoh right, totally forgot that existed! thanks!
15:56Jamie has joined #asahi-dev
15:56Jamie is now known as Guest4575
15:56Youwes09myria: tested the PR on my M4 Air (J713), 7.4-rc1. all 60 M4 events count on both P and E cores, two counters on the same event always match exactly, all 10 counters at once work, and perf record works on both clusters!
15:57rhysmdnz has joined #asahi-dev
15:57Youwes09myria: found one bug in "Support per-implementation event size". PMU_FORMAT_ATTR(event_8b, ...) names the sysfs file event_8b instead of event, so perf can't parse event= anymore and the cycles/instructions aliases break on every chip, M1 too. raw rXXXX still works which is why sleep 5 didn't show it...
15:58Youwes09myria: the 16-bit table also points at event_8b, so M3/M4 only show config:0-7 and perf warns on events like r182. I fixed it here by keeping both named event with __ATTR(event, 0444, event_16b_show, NULL), cycles and instructions work after that. Happy to send the diff if you want it!
15:59croissant` has quit []
15:59croissant has joined #asahi-dev
16:04LucaT[m] has joined #asahi-dev
16:08Youwes09 has quit [Ping timeout: 480 seconds]
16:11hdbngr_ has joined #asahi-dev
16:16hdbngr__ has joined #asahi-dev
16:17hdbngr has quit [Ping timeout: 480 seconds]
16:19hdbngr_ has quit [Ping timeout: 480 seconds]
16:23mattia013 has joined #asahi-dev
16:25hardy94 has joined #asahi-dev
16:33Bhyve has quit []
16:39umo has quit [Remote host closed the connection]
16:39umo has joined #asahi-dev
16:44beastd|2 has joined #asahi-dev
16:50beastd has quit [Ping timeout: 480 seconds]
16:57___nick___ has joined #asahi-dev
16:57___nick___ has quit []
16:59___nick___ has joined #asahi-dev
17:17MikeM64 has quit [Ping timeout: 480 seconds]
17:41jannauI've updated asahi-wip and the bits branches to v7.3 and picked devicetree changes up. bits/000-devicetree has almost 250 commits
17:44jannauI dropped a few patches which I believe are no longer necessary. power-domain or less likelydart issue might be caused by that
17:47jannaualso switched to the v3 upstream avd DTS and fixed 1 compatibility issue with the current fw, there might be more, I did not test thouroughly
18:05cy8aer has quit [Remote host closed the connection]
18:10junefish has quit [Ping timeout: 480 seconds]
18:12cy8aer has joined #asahi-dev
18:12remexre has quit [Ping timeout: 480 seconds]
18:14cy8aer has quit [Remote host closed the connection]
18:19cy8aer has joined #asahi-dev
18:22luca020400 has quit [Remote host closed the connection]
18:22luca020400 has joined #asahi-dev
18:50iwuvmika has joined #asahi-dev
18:50nullbert[m] has joined #asahi-dev
19:00bgb has quit [Remote host closed the connection]
19:00bgb has joined #asahi-dev
19:16sofus has joined #asahi-dev
19:31jannauflokli: no that was done before we settled on the change in the firmware, https://github.com/AsahiLinux/linux/commit/faf8d1233a82cff11182dc7e83bcdd356f5feb83 does read from 0x400
19:32jannauhttps://forge.fedoraproject.org/asahi/avd-fw has compattibility patches to provide the config at the place asahi-7.2.* kernels expect it
19:32flokliAh thanks, so not in a tagged kernel yet. I was just very confused
19:33flokliI might just wait for that to get picked, none of this is currently exposed yet.
19:35flokli(on our side)
19:49amonakov has joined #asahi-dev
20:00junefish has joined #asahi-dev
20:37svenhttps://github.com/AsahiLinux/docs/pull/287 tested on m5, a18 pro will probably panic somewhere in early boot before snooze6214[m]'s fix
20:39svenif anyone wants to shave a yak: it might be useful to write a small python script that calls into ipsw --remote to extract all those blobs from an .ipsw file
20:39svenor at least collects them from the preboot volume automatically
20:43jannau`python3 -m urlcache $IPSW $FILES` in asahi-installer/src should work
20:53mattia013 has quit [Quit: mattia013]
20:58umo has quit [Read error: Connection reset by peer]
20:58umo has joined #asahi-dev
20:59sveneven better. "just" needs some logic to find the right files inside the ipsw then
21:01svenI bet they are just listed in some plist or whatever
21:02nicolas17exclavecore?
21:02nicolas17they should be in the BuildManifest.plist yeah
21:03svenyeah, that and something called "trustcache" and "integrity_catalog"
21:04sventhere's a bunch of trustcaches but only one integrity_catalog
21:05svenright now I just take the trustcache that has the same prefix as the single integrity thingy but I bet that's just listed in BuildManifest as well then
21:07___nick___ has quit [Ping timeout: 480 seconds]
21:10sofus has quit [Quit: leaving]
21:18hardy94 has quit [Remote host closed the connection]
21:19fatcullen has quit [Quit: This computer has gone to sleep]
21:28iwuvmika has quit [Ping timeout: 480 seconds]
21:35bgb has quit [Remote host closed the connection]
21:36veloek_ has joined #asahi-dev
21:38veloek has quit [Ping timeout: 480 seconds]
21:40bgb has joined #asahi-dev
21:58mulderje[mds] has quit [Ping timeout: 480 seconds]
21:58Pen[mds] has quit [Ping timeout: 480 seconds]
21:58tumblingweed has quit [Ping timeout: 480 seconds]
21:59faranith[mds] has quit [Ping timeout: 480 seconds]
21:59JPeisach has quit [Ping timeout: 480 seconds]
21:59Q_ has quit [Ping timeout: 480 seconds]
21:59Finxx[mds] has quit [Ping timeout: 480 seconds]
21:59root[mds] has quit [Ping timeout: 480 seconds]
21:59mjolnir[mds] has quit [Ping timeout: 480 seconds]
21:59travltux[mds] has quit [Ping timeout: 480 seconds]
22:04josiedotlol_ has quit [Remote host closed the connection]
22:05veloek has joined #asahi-dev
22:07veloek_ has quit [Ping timeout: 480 seconds]
22:11iwuvmika has joined #asahi-dev
22:19fatcullen has joined #asahi-dev
22:31mjolnir[mds] has joined #asahi-dev
22:36josiedotlol_ has joined #asahi-dev
23:02iwuvmika has quit [Ping timeout: 480 seconds]
23:04iwuvmika has joined #asahi-dev
23:05hardy94 has joined #asahi-dev
23:22hardy94 has quit [Remote host closed the connection]
23:41myriajannau: Does the new avd patches require the new firmware, have we moved more of the work onto the CM3? I think I vaguely remember sofus talking about getting more work to be done there
23:42myriaYouwes09: though you seem to have left the channel, I would love to see your diff, and thank you so much for testing!
23:42myria(so if you ever see this, please send it through)
23:43myriaCurrently, it looks like the naming of the format attrs means that you can only have one per file, otherwise they'll name clash, or you need to expand some of the macros themselves
23:44myriadoesn't seem like older m2 devices should have problems if I just declare the the first 16 bytes are the config, so I may just use `config:0-15` for everyone
23:47myriaAlso, can you send me through a name/email if you see this and are comfortable with that? Just to add a tested-by tag if you're happy otherwise
23:59JamesCalligeros[m]Michael Reeves: would you mind sharing the FNV screenshots again? sorry.
…