2026-08-11
00:06jsk has quit [Read error: No route to host]
00:06jsk has joined #asahi-dev
00:18Rafay has quit [Remote host closed the connection]
01:19jsk_ has joined #asahi-dev
01:24jsk has quit [Ping timeout: 480 seconds]
01:28jsk_ is now known as jsk
01:40catred has quit [Remote host closed the connection]
01:45pb17 has quit [Ping timeout: 480 seconds]
01:50pb17 has joined #asahi-dev
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
03:43pthariensflame has joined #asahi-dev
03:54pthariensflame has quit [Remote host closed the connection]
03:57nora has joined #asahi-dev
03:59nora_ has quit [Ping timeout: 480 seconds]
04:29eluks has quit [Remote host closed the connection]
04:30eluks has joined #asahi-dev
05:03WindowPain9 has joined #asahi-dev
05:07WindowPain has quit [Ping timeout: 480 seconds]
05:11WindowPain9 has quit [Ping timeout: 480 seconds]
05:27pb17 has quit [Ping timeout: 480 seconds]
05:33pb17 has joined #asahi-dev
07:00hdbngr has joined #asahi-dev
07:08hdbngr has quit [Ping timeout: 480 seconds]
07:27chadmedis there a better way to test local mesa builds than meson devenv? it's kinda flakey inside muvm
07:43Pokemod9773620847128183300 has quit [Ping timeout: 480 seconds]
07:59mixiyou should be able to set a --prefix in your home with meson, install mesa there, and then manually use a LD_LIBRARY_PATH pointing to that prefix
08:01hdbngr has joined #asahi-dev
08:09hdbngr has quit [Ping timeout: 480 seconds]
08:16BrandiWhite[m] has joined #asahi-dev
08:26picnoir has quit [Quit: WeeChat 4.10.0]
08:27picnoir has joined #asahi-dev
08:39xal has quit []
08:41hdbngr has joined #asahi-dev
08:41xal has joined #asahi-dev
08:43chaos_princesssven: are both ends of your connection on linux
08:43chaos_princessno worky here
08:43chaos_princessalso checked via `cat /sys/module/thunderbolt_net/parameters/e2e`
08:44chaos_princesswhich side is the ping "source"
08:44chaos_princessare both devices on ethernet
08:45chaos_princessweird
08:46svenlinux does two writes to the ring options, first one setting the enable bit and the second one then setting the e2e bit. let me merge that and see if that also works here
08:55BrandiWhite[m]Get in touch with this platform for greatness you’ll definitely thank me later
ℹ️❤️
👇🏻👇🏻👇🏻
08:55BrandiWhite[m]Get in touch with this platform for greatness you’ll definitely thank me later
ℹ️❤️
👇🏻👇🏻👇🏻
08:55BrandiWhite[m] has left #asahi-dev [Kicked by @svenpeter:matrix.org : spam]
08:59svenchaos_princess: https://f.svpe.de/1be370d90ef99e1985432b83f6c5c89ef5bf4ab84861dfb8a014c73ac5159e0c_tbt.patch seems to work for me
09:01chaos_princessok, will test in a bit, i will probably need to re-merge it on top of asahi-wip
09:07hdbngr has quit [Ping timeout: 480 seconds]
09:25chaos_princessstill does not work
09:28svenhttps://f.svpe.de/0c266ac77d9403c8bc714e105b8e4ba6c14c54e35e22b58fbe972f63729a723f_tbtnet.txt i don't see how this can be anything else than usb4net
09:36sri has joined #asahi-dev
09:36Rafay has joined #asahi-dev
09:41Rafaysven: These are the logs, just tbt-reset-wip at commit 98588460927272c731c6018af25463986fdf9e0e and the patch you gave me. I didn't add the later patch you sent here and I tried with e2e=0 with no luck. https://pastebin.com/JxrwH0wK
09:42hdbngr has joined #asahi-dev
09:51hdbngr has quit [Ping timeout: 480 seconds]
09:55ansg191 has quit [Remote host closed the connection]
09:55svenchaos_princess: last difference i see, but that's related to interrupts and i'd be surprised if this makes a difference: https://f.svpe.de/52ff3d5be413a43da7d17b374a2f6cd5625844b8ce6f276043395437a034db77_hack.patch
10:04ansg191 has joined #asahi-dev
10:05chaos_princessstill no
10:06svenso it's either that macos/linux works for some reason and linux/linux doesn't or the non-base M1 machines do something very weird that i don't see in the mmio traces
10:08svenif M1/M1 works the initial series will be M1 only and the pro/max/etc. variants are going to be someone else's problem. ideally someone who actually has that hardware :D
10:08chaos_princessdoes that look suspicious?
10:16chaos_princesst6000
10:21yuka has quit [Remote host closed the connection]
10:21svenchaos_princess: can you read32 0x701fd0000 and 0x701f1401c after you started the ping from the hv?
10:30yuka has joined #asahi-dev
10:35chaos_princesssven: both zero
10:35chaos_princessyuka: just send it, m3 does not need 15
10:37chaos_princesswe haven't tried the stream 15 nonsense again, right?
10:38jannauI plan to send more M3 hw enabled patches once 7.3-rc1 is out. I haven't done anything yet with adding missing dt-binding patches and separating device tree modifications
10:38WindowPain has joined #asahi-dev
10:39svenchaos_princess: also, which firmware is that? already the one where acio is a bit more talkative over its syslog?
10:40chaos_princessyes, whatever comes with latest 27 beta
10:41jannaustill not sure how to send it. I guess for keeping every ones mailbox clear I should send a series for each hw with dt-binding change and dts changes. They probably need to depend on each other for the dts changes to apply
10:43chaos_princessanything suspicious there?
10:43jannaumaz: I need to check how the PCIe power control changes fir in with the mac pro PCIe support with the big PCIe switch in the middle
10:43yukamaz: I prepared a new version of PCIe power control, but it's gotten very messy and I don't feel too good about sending it out like this. I can send it if you really want, but I don't expect it to go anywhere.
10:45svenhttps://f.svpe.de/878d6f8e2b179d9a8833af140b0775c5c600bb3ede38c05ac08fd4dc88f3c51e_tbt-dmesg.txt
10:45mazyuka: no point in sending something you're not confident about. once you have something that is in a reviewable state, I'll have a look. in the meantime, I'll stop tracking this in my inbox.
10:46chaos_princessmine is from later, once it actually sees another host and starts to negotiate usb4net
10:47svenhttps://f.svpe.de/69d6594c21cdc03d46f2489c0b46b1a29ecffa1195566a5d0d1a9822f88aa396_tbt-dmesg.txt
10:57svendunno... guess i either need to give up on non-M1 for now or write a tracer for the ring 0 traffic
10:58chaos_princesslet's see if m1 can talk to other m1 first
10:58hdbngr has joined #asahi-dev
10:59yukajannau: if you send the M3 nvme stuff through the soc tree, then I can only do M4 nvme in 7.4, or it will result in merge conflicts. any idea how to deal with this?
11:00svenmaybe it's time to write dt-bindings. or look at the horror that lurk inside pcie hotplug again
11:07hdbngr has quit [Ping timeout: 480 seconds]
11:10jannauyuka: you can either depend on change or trust that linux maintainers are able to resolve simple one-line conflicts in dt-binding changes
11:11svenit's kinda close to the 7.3 merge window anyway, not sure if anything will still be picked up
11:14yukathe main point being that my change has actual driver changes, so it with its accompanying dt-bindings change would go through the nvme tree
11:21yukaI guess apple,t6020-nvme-ans2 and apple,t6000-nvme-ans2 are only explicitly listed because one of them is in the old base compatible
11:31ieyasu_ has quit [Read error: Connection reset by peer]
11:32ieyasu has joined #asahi-dev
11:34hdbngr has joined #asahi-dev
11:46hdbngr has quit [Ping timeout: 480 seconds]
11:59hdbngr has joined #asahi-dev
12:02yukaconfirmed linux-next + https://termbin.com/8dyl + "idle=nop arm64.nowfxt" can boot on M4 smp
12:07hdbngr has quit [Ping timeout: 480 seconds]
12:15hdbngr has joined #asahi-dev
12:16___nick___ has joined #asahi-dev
12:19mripard has quit [Remote host closed the connection]
12:22liv[m]i have an OWC Thunderbolt Pro Dock and HP Thunderbolt Dock G2, I think the HP does ethernet over USB but the OWC does it over something more complicated than that because of 10G capabilities, so would it be useful if I compiled a few patches and dropped a dmesg every once in a while?
12:24mripard has joined #asahi-dev
12:26___nick___ has quit []
12:28___nick___ has joined #asahi-dev
12:29___nick___ has quit []
12:31___nick___ has joined #asahi-dev
12:32hdbngr has quit [Ping timeout: 480 seconds]
12:34hdbngr has joined #asahi-dev
12:36liv[m]i'm running nixos so i can easily - if i have kernel patches - rebuild and boot a new kernel once the time is there :)
12:42hdbngr has quit [Ping timeout: 480 seconds]
12:49hdbngr has joined #asahi-dev
12:49LarstiQ has quit [Ping timeout: 480 seconds]
12:54jannauno KWIN environment set. testing your surface / plane PR (modified) but drm_info looks as expected
12:56chadmedplane 2 gets the cursor, plane 1 gets nothing without KWIN_USE_OVERLAYS, plane 0 gets the desktop
12:57sri has quit [Quit: Leaving]
13:03mripard has quit [Remote host closed the connection]
13:04mripard has joined #asahi-dev
13:04chadmed8076 got merged so i think we should be okay from a kwin perspective to enable overlays unconditionally no
13:05chadmedi need to look at colour pipelines again soon, we do hit the colour bug mentioned in the comments on that MR under certain conditions
13:09hdbngr has quit [Ping timeout: 480 seconds]
13:11jannauwhy would we hit that bug without supported colour pipelines? It would be either a kwin bug/weirdness or some additional blending restrictions in DCP which we don't understand
13:12chadmednot sure, all ive noticed is that colours change slightly on some content between compositing and scanout
13:17chadmed has quit [Quit: Konversation terminated!]
13:17hdbngr has joined #asahi-dev
13:21chadmed has joined #asahi-dev
13:21svenoh ffs, there's some weird race condition that appears and claims that "port is locked" when i remove the debug prints in apple.c
13:26hdbngr has quit [Ping timeout: 480 seconds]
13:34hdbngr has joined #asahi-dev
13:44bgb_ has joined #asahi-dev
13:44bgb has quit [Remote host closed the connection]
13:48hdbngr has quit [Ping timeout: 480 seconds]
13:55bgb_ has quit [Remote host closed the connection]
13:59bgb has joined #asahi-dev
14:14your_nick has joined #asahi-dev
14:16your_nick has quit [Remote host closed the connection]
14:16thequantumnargle has joined #asahi-dev
14:26Rakshit[m] has joined #asahi-dev
14:33bennofs30 has joined #asahi-dev
14:35bennofs3 has quit [Ping timeout: 480 seconds]
14:35bennofs30 is now known as bennofs3
14:37jannauI get "high_speed_lane.c:289: Gen2/3 link error. lane=[01], error=" spam. I guess I'll try different cables
14:43sveni suspect half of the bit descriptions are wrong but the only way to fix them is to have enough (PDVDO or EUDO),(what macos writes) pairs
14:46jannauthunderbolt 3 cable (according to vendor markings) somewhat works but I don't see any devices in boltctl list. dmesg: https://paste.debian.net/hidden/91917d40
14:49chaos_princesshost to host?
14:53hdbngr has joined #asahi-dev
14:53sveni think that's correct because it gets far enough, but can you check what macos would write? let me find the debug arguments you need to get the print
14:54svenit already gets far enough, it should find whatever is connected. is that a device or another host
14:54svendebug=0x14e serial=3 apcie=0xfffffffe wdt=-1 AppleSynopsysDRDUSBXHCI-debug=0xffff AppleT8103TypeCPhy-debug=0xffff AppleT8103USB40XHCITypeCPort-debug=0xffff AppleT8103USBXDCI-debug=0xffff AppleTypeCPhy-debug=0xffff AppleTypeCPhyCIOInterface-debug=0xffff AppleTypeCPhyDisplayPortInterface-debug=0xffffffff AppleTypeCPhyInterface-debug=0xffff
15:04jannaugot further along this time but boltctl list is still empty. I see the othewr device announced in dmesg though
15:06chadmed has quit [Quit: Konversation terminated!]
15:07jannauyes, thunderbolt0 net interface is there. I was expecting to see something in boltctl for host2host
15:07chadmed has joined #asahi-dev
15:08chadmed has quit [Read error: Connection reset by peer]
15:08chadmed has joined #asahi-dev
15:09chadmed has quit [Read error: Connection reset by peer]
15:09chadmed has joined #asahi-dev
15:09Moderation[m] has joined #asahi-dev
15:12chaos_princessit didn't for me either
15:13thequantumnargle has quit [Remote host closed the connection]
15:24sventblog=0xffffffff tblogmode=3 tbdp=0xffffffff tbip=0xffffffff are the relevant bootargs I used to enable all verbose debug spam I think
15:31hdbngr has quit [Ping timeout: 480 seconds]
15:34sveni accidentally added a two or three level indirect pointer lookup plus jump to the hot path so it's not surprising we don't get the full speed, but 2gbps seems low
15:36sventhere's also that magic 0x186 write next to irq enable which might just be some tweaking to when that irq fires
15:37jannauerr, I typoed the size resulting in fragmentation. `ping -s 1472 -f IP` is now at ~28 gbps
15:50sven"In our tests, 13.6 Gbps is within a realistic range. Although this may seem below the raw capacity, real-world factors (including half-duplex operation) usually limit effective throughput." lol
15:58rany has quit [Remote host closed the connection]
16:00bgb has quit [Remote host closed the connection]
16:00bgb has joined #asahi-dev
16:01rany has joined #asahi-dev
16:06sveni'm writing a tracer for traffic on ring 0 anyway due track down that weird race where some ports are locked and then xdomain doesn't work at all if i remove enough debug prints in apple.c
16:06Rafaysven: Here's the trace under linux. https://gist.github.com/rafayahmed317/d7d8768cec63178c0c15a8983af7d553 . Just to remind you, on my setup, thunderbolt_net doesn't create a device and xdomain doesn't initialize. You already have the mac trace: https://gist.github.com/rafayahmed317/e083de8e478830dc37969ca0b36750e7
16:14qyliss has quit [Quit: bye]
16:16rany has quit [Remote host closed the connection]
16:16qyliss has joined #asahi-dev
16:16Guest15226 has quit [Quit: Bridge terminating on SIGTERM]
16:16rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
16:17rany has joined #asahi-dev
16:21chaos_princessyuka: is m4 smp reliable w/ that, or still a dice roll?
16:22qyliss has quit [Quit: bye]
16:22qyliss has joined #asahi-dev
16:25WindowPain has quit [Ping timeout: 480 seconds]
16:28pb17 has quit [Ping timeout: 480 seconds]
16:37bgb has quit [Remote host closed the connection]
16:38nnym has quit [Ping timeout: 480 seconds]
16:43bgb has joined #asahi-dev
16:45pb17 has joined #asahi-dev
16:48rany has quit [Remote host closed the connection]
16:48rany has joined #asahi-dev
16:53__nick__ has joined #asahi-dev
16:53Rafaysven: Is this lspcie output expected? I think some nodes are missing. https://pastebin.com/PeSMcLTQ
16:56___nick___ has quit [Ping timeout: 480 seconds]
16:57rhysmdnz has joined #asahi-dev
16:57svenchaos_princess: can you run https://f.svpe.de/f9237287e3af548f79fb0d8ceaa85c49b7f8f6ed27a36d07d15d6f86d1292b21_trace_acio.py.txt under the hv up until the point where you run ping and the packets then disappear?
16:58Guest15459 has joined #asahi-dev
16:59chaos_princessack, give me a bit
17:02Carlos has quit [Ping timeout: 480 seconds]
17:03Carlos has joined #asahi-dev
17:05rany has quit [Remote host closed the connection]
17:05rany has joined #asahi-dev
17:05rany has quit [Remote host closed the connection]
17:06rany has joined #asahi-dev
17:08Guest15459 has quit [Quit: Bridge terminating on SIGTERM]
17:08rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
17:09nnym has joined #asahi-dev
17:15apc has quit [Quit: Ping timeout (120 seconds)]
17:15apc has joined #asahi-dev
17:15ralim89 has joined #asahi-dev
17:15msimonelli has quit [Read error: Connection reset by peer]
17:15msimonelli has joined #asahi-dev
17:15rhysmdnz has joined #asahi-dev
17:16Jamie1 has joined #asahi-dev
17:16ralim8 has quit [Read error: Connection reset by peer]
17:16ralim89 is now known as ralim8
17:18Rafaysven: here's your tracer on my machine if it helps. Linux: https://pastebin.com/V8kfgXPH Mac: https://pastebin.com/rGZNU87Y
17:37hdbngr has joined #asahi-dev
17:45hdbngr has quit [Ping timeout: 480 seconds]
18:00chaos_princessshould be disabled unless i did sth wrong
18:02sven[cpu1] [ACIOTracer@/arm-io/acio1] TX ring 1 flags = 0x80000000 (ENABLE=1, RAW=0, NO_SNOOP=0, E2E_FLOW_CONTROL=0, ISOCH_ENABLE=0)
18:03chaos_princessshould i make one w/ e2e enabled?
18:04sventhere has to be some difference somewhere and the closer the trace are the easier it will be to stop
18:15svenRafay: are you running the latest commits from the tbt-reset-wip branch? if yes, do you see a "TBT cable: cable_mode " entry in dmesg?
18:16sven[cpu4] [ACIOTracer@/arm-io/acio1] CTL TX WRITE route=0x0 port=0 space=SWITCH offset=0x36 len=1 seq=0 data=[0x00000603] <-- that's the VSE being written and after that nothing happens
18:16svenand macos writes [cpu2] [ACIOTracer@/arm-io/acio1] CTL TX WRITE route=0x0 port=0 space=SWITCH offset=0x36 len=1 seq=2 data=[0x00000413]
18:17hdbngr has joined #asahi-dev
18:22rany has quit [Remote host closed the connection]
18:23msimonelli5 has joined #asahi-dev
18:23msimonelli has quit [Read error: Connection reset by peer]
18:23rany has joined #asahi-dev
18:23msimonelli5 is now known as msimonelli
18:26hdbngr has quit [Ping timeout: 480 seconds]
18:32Rafay[ 193.553395] thunderbolt-apple-acio 501ac0000.nhi: TBT cable: cable_mode 0x0, device_mode 0x0, enter_vdo 0x3, orientation 2 -> cable info 0x603
18:34sventhat actually looks like either a bug in the tipd driver or my tbt-switch thing because those modes don't make sense
19:04hdbngr has joined #asahi-dev
19:07rany has quit [Ping timeout: 480 seconds]
19:08rany has joined #asahi-dev
19:12hdbngr has quit [Ping timeout: 480 seconds]
19:15svenchaos_princess: https://f.svpe.de/f1881bb46acce261fdbb1865e5d3093f5157cd31169501371ab71ce7f8bcd9c2_0001-HACK.patch
19:16chaos_princessshould i try this w/ or w/o e2e?
19:17blazra has quit [Quit: The Lounge - https://thelounge.chat]
19:17blazra6 has joined #asahi-dev
19:18hdbngr has joined #asahi-dev
19:20jsk has quit [Remote host closed the connection]
19:20jsk has joined #asahi-dev
19:23chaos_princessno change
19:27hdbngr has quit [Ping timeout: 480 seconds]
19:34svenchaos_princess: now this really shouldn't make a difference, but i'm kinda out of idea at this point: https://f.svpe.de/99fe30ff297fe3d18e7d347080279f27036bffe01b4f1bfc0500fc57bfa187f9_0002-HACK.patch
19:41hdbngr has joined #asahi-dev
19:47chaos_princessno change
19:55hdbngr has quit [Ping timeout: 480 seconds]
19:55chaos_princessno, one is macos
19:56jannausven: cables are all working now. the issues before were probably caused by smc's charging handling breaking things when running HV and connecting a power source into the other port
19:56vup has quit []
19:57sveni think everything that doesn't use EUDO is very broken right now though. there's a bug somewhere in tipd that i'll have to figure out
20:00vup has joined #asahi-dev
20:07__nick__ has quit [Remote host closed the connection]
20:09svenchaos_princess: can you add return -ENODEV or whatever at the beginning of tb_xdomain_lane_bonding_enable in xdomain.c?
20:18rany has quit [Ping timeout: 480 seconds]
20:19chaos_princessgets stuck in connect/disconnect loop
20:19hdbngr has joined #asahi-dev
20:19chaos_princess"failed to add NVM devices: -19"
20:20rany has joined #asahi-dev
20:22svenugh, sorry. should've tested that first locally but i already shut down my test rig. anyway, maybe lane bonding is somehow broken because i don't see that with my m1 setup but i think yours tries to do that
20:28hdbngr has quit [Ping timeout: 480 seconds]
20:32hdbngr has joined #asahi-dev
20:41hdbngr has quit [Ping timeout: 480 seconds]
20:42bgb has quit [Remote host closed the connection]
20:42bgb has joined #asahi-dev
20:44hdbngr has joined #asahi-dev
20:54hdbngr has quit [Ping timeout: 480 seconds]
20:56hdbngr has joined #asahi-dev
21:05hdbngr has quit [Ping timeout: 480 seconds]
21:12Yushi has quit [Remote host closed the connection]
21:14jole_ has quit [Remote host closed the connection]
21:15Yushi has joined #asahi-dev
21:15jole has joined #asahi-dev
21:19hdbngr has joined #asahi-dev
21:31hdbngr has quit [Ping timeout: 480 seconds]
21:34hdbngr has joined #asahi-dev
21:46hdbngr has quit [Ping timeout: 480 seconds]
21:59hdbngr has joined #asahi-dev
22:16hdbngr has quit [Ping timeout: 480 seconds]
22:19rany has quit [Remote host closed the connection]
22:19rany has joined #asahi-dev
22:31rany has quit [Remote host closed the connection]
22:33rany has joined #asahi-dev
23:37hdbngr has joined #asahi-dev
23:38pb17 has quit [Ping timeout: 480 seconds]
23:45hdbngr has quit [Ping timeout: 480 seconds]
23:53pb17 has joined #asahi-dev
23:54hdbngr has joined #asahi-dev