#
…

2023-01-17

00:03hightower4 has quit [Ping timeout: 480 seconds]
00:54hertz has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
01:12seeeath has joined #asahi-dev
01:29hightower4 has joined #asahi-dev
01:36hightower3 has quit [Ping timeout: 480 seconds]
01:52hightower3 has joined #asahi-dev
01:53seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
01:55seeeath has joined #asahi-dev
01:59hightower4 has quit [Ping timeout: 480 seconds]
02:00seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
02:16amarioguy has joined #asahi-dev
02:21hightower4 has joined #asahi-dev
02:28hightower3 has quit [Ping timeout: 480 seconds]
02:47seeeath has joined #asahi-dev
02:49seeeath has quit []
03:00seeeath has joined #asahi-dev
03:00seeeath has quit []
03:28marcanyeah, wake on keyboard is not implemented
03:29marcanIRQ behavior on s2idle is opt-in, right now we opt in all mailboxes under the assumption that mostly everything but SMC will be shut down anyway
03:30marcanbut not SPI (and then the driver would need a codepath to trigger the wakeup anyway)
04:10hertz has joined #asahi-dev
04:36nyx_o has quit [Quit: WeeChat 3.5]
05:00c10l has quit [Quit: Bye o/]
05:00c10l has joined #asahi-dev
05:27faruk has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
05:29nyx_o has joined #asahi-dev
07:43zzywysm has joined #asahi-dev
08:42amw has quit [Ping timeout: 480 seconds]
09:17c10l1 has joined #asahi-dev
09:24c10l has quit [Ping timeout: 480 seconds]
09:38leitao has joined #asahi-dev
09:43tim has joined #asahi-dev
09:43tim is now known as Guest1569
09:52leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
09:59leitao has joined #asahi-dev
10:00leitao has quit []
10:09leitao has joined #asahi-dev
10:30leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
10:41faruk has joined #asahi-dev
10:43faruk has quit []
10:44leitao has joined #asahi-dev
10:44jluthra has quit [Remote host closed the connection]
10:45jluthra has joined #asahi-dev
10:54eiln has joined #asahi-dev
10:57eiln has quit [Remote host closed the connection]
10:59eiln has joined #asahi-dev
11:24leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
11:25leitao has joined #asahi-dev
12:01leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
12:04povikhaha, so the I2S ports are hooked up to GPIO after all
12:05poviki could have made such good use of this when i was figuring out MCA
12:05povikwhy didnt i try mon.add(gpio_base, xxx) and enable/disable the I2S clocks earlier?
12:05sveni feel like almost everything is somehow hooked up as GPIO as well :D
12:06hightower3 has joined #asahi-dev
12:13hightower4 has quit [Ping timeout: 480 seconds]
12:17hightower3 has quit [Ping timeout: 480 seconds]
12:32zalyx has quit [Read error: Connection reset by peer]
12:33zalyx has joined #asahi-dev
12:37leitao has joined #asahi-dev
12:39chadmed_ has quit [Remote host closed the connection]
12:40svenlol, so thunderbolt hotplug seems to work pretty well. it just doesn't detect the adapter the first time for some reason
12:40svenpcie is still kinda broken though
12:41sven0001:00:00.0 PCI bridge: Apple Inc. Apple Silicon USB4/Thunderbolt PCI Express Root Port
12:41sven0001:01:00.0 PCI bridge: Intel Corporation DSL2210 Thunderbolt Controller [Port Ridge 1C 2011]
12:41svenpretty sure the second thing is a lie :D
12:41chadmed_ has joined #asahi-dev
12:41svenoh... wait... pcie has tunables as well, doesn't it?
12:47hightower2 has joined #asahi-dev
12:53svenwait.. or maybe that's the 3->2 adapter showing up
12:55c10l1 has quit []
12:56c10l has joined #asahi-dev
13:05eiln has quit [Remote host closed the connection]
13:14jannausven DSL2210 seems to be a thunderbolt 1 device controller so that is might already be on the ethernet adapter
13:15svenhttps://paste.debian.net/hidden/2c372cad/ looks like i might've messed up something when describing the bus in the dt, let me double check that
13:15jannauit is on the thunderbolt to firewire adapter
13:16faruk has joined #asahi-dev
13:20bisko has joined #asahi-dev
13:30leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
13:32leitao has joined #asahi-dev
13:40faruk has quit [Quit: Textual IRC Client: www.textualapp.com]
13:47chadmed_ has quit [Remote host closed the connection]
13:48sven0001:00:00.0 PCI bridge: Apple Inc. Apple Silicon USB4/Thunderbolt PCI Express Root Port
13:48sven0001:01:00.0 PCI bridge: Intel Corporation DSL2210 Thunderbolt Controller [Port Ridge 1C 2011]
13:48sven0001:02:00.0 PCI bridge: Intel Corporation DSL2210 Thunderbolt Controller [Port Ridge 1C 2011]
13:48sven0001:03:00.0 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM57762 Gigabit Ethernet PCIe
13:48svenDMA is obviously broken since there's no IOMMU
13:48svenand I need to read code to understand how the kernel handles hotplug management because right now this is only works with even more ugly hacks
13:48eiln has joined #asahi-dev
13:48chadmed_ has joined #asahi-dev
13:53mazsven: awesome!
13:58mazdo you get the interrupt on the RP when plugging in the device? this should be what triggers the rescan.
13:59svenit's a bit more ugly. I need to wait for the Thunderbolt NHI to activate the PCIe link and only then clear some "tunnel reset" bit and then trigger the link-training and only then the "port up" interrupt appears
14:02maztwo state machine intersecting. what's not to love? :-/
14:04svenit gets even better. the DART MMIO range only starts to work once PERST is disabled
14:04___nick___ has joined #asahi-dev
14:04povik"product state space"
14:04___nick___ has quit []
14:05svenand ofc this is all triggered by that lovely PD controller
14:05mazsven: "insert DART enable >here<"
14:06___nick___ has joined #asahi-dev
14:06bisko has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
14:13povikoh hello
14:14* mazstarts counting pennies
14:14povikno streamed event this time?
14:14poviki have learned to like them
14:14* sventries to resists the temptation
14:15ChaosPrincessoh, hey wifi6e finally
14:15jannauwith wifi 6e and hdmi 2.1 apparently
14:16j`eyhm £2000 for one with upgraded CPU/RAM
14:17jannau19/38 gpu cores
14:18jannauapparently also new mac mini
14:18j`ey8 core with 24GB ram is 1200gbp.. tough choice
14:19j`ey(what I was quoting was the mac mini)
14:19jannauwith m2 pro
14:19j`eyis it 8p4e with the 12 core?
14:20jannau10 core is 6p4e
14:20* mazstarts looking at ebay for a second-hand M1 mini...
14:23jannaum2 mini tech specs say thunderbolt 4 ports and supports two displays over thunderbolt
14:23svenso maybe the also used that tiled display support to support two displays? :D
14:23faruk has joined #asahi-dev
14:23svenor is there a pro/max in that mini as well?
14:24jannauno, for the base m2
14:24jannauit looks like they can somehow route dcp over thunderbolt
14:24ChaosPrincesssven: it appears they route the hdmi output to tb now
14:24svenohh.. fancy
14:25svenalso a pity though
14:25ChaosPrincesscause it is thunderbolt + thunderbolt or hdmi
14:25jannauthe second thunderbolt display seems to exclude hdmi
14:25svenso, uh, do they have another display crossbar on the m2?
14:25leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
14:28eiln has quit [Remote host closed the connection]
14:28leitao has joined #asahi-dev
14:29* amarioguyhopes they solved the serror bug
14:31mazhere's to hope!
14:31mazthough TBH, it isn't too hard to work around.
14:32ketteniswouldn't get my hopes up
14:33ChaosPrincesswould be nice to see device trees for those machines
14:33mazwhat I'd really like is a mask on the vgic maintenance interrupt. can't imagine that happening either (and I wouldn't know where to look)
14:33ChaosPrincessbut ipsws aren't up yet
14:33faruk has quit [Quit: Textual IRC Client: www.textualapp.com]
14:34ketteniswonder if they stayed with the Broadcom BCM5776x for the 1G Ethernet on the mini
14:36maztime to start building a 4th serial dongle, just in case...
14:37ketteniswhat did you use to build those?
14:37mazRPi Pico, FUSB302 on an adapter board, various level shifters for the uart and VBUS.
14:38mazI have half of a proper PCB designed, but need time to finish it...
14:40jannau8k 60 Hz HDMI seems to require two dcpext
14:41chadmed_ has quit [Ping timeout: 480 seconds]
14:43svenlet me know if you finish that pcb. i'd be interested in a real serial adapter as well
14:43kettenisa bit neater than what bluerise rigged together for me
14:45mazwhat I like about it is that the pico is the serial adapter. and in theory, one adapter could deal with two FUSB302 (one per I2C bus), which is useful when you have multiple boxes.
14:46mazthe expensive part here was the QFN adapter being delivered from Canada.
14:46mazthe PCB would solve that problem.
14:47nicolas17 has joined #asahi-dev
14:53nicolas17ChaosPrincess: speculation was product announcement 11am, betas 1pm
14:56sven[ 3.460700] thunderbolt-apple-cio 501f00000.nhi: Port 6: 5ac:2000 (Revision: 0, TB Version: 1, Type: DP/HDMI (0xe0101))
14:56svenaaaand there's the thunderbolt DP tunnel
14:56svenwhich also seems to come up at least on the NHI side when I connected my old thunderbolt display via apple's 3->2 adapter
14:57svenso I guess we'll have to come up with some kind of NHI -> {pcie, dpin0-phy, dpin1-phy} notification system anyway
14:58svenon intel/amd controller this all happens without additional software notifications :(
15:02eiln has joined #asahi-dev
15:02eilnholy shit
15:02eilniovas are a function of access distance
15:03eilni'm getting a 4x speedup with a capped iova limit
15:04eiln0.002398727 per batch for inception_v3
15:22nate8 has joined #asahi-dev
15:26jannauI wonder whether the display out switching will work on the m2 laptops too. 2 external displays in clamshell mode could be interesting. I doubt apple will support that in macos
15:32eiln has quit [Remote host closed the connection]
15:44seeeath has joined #asahi-dev
15:48svenhuh, that would be useful
15:48nicolas17 has quit [Ping timeout: 480 seconds]
15:49seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
15:50seeeath has joined #asahi-dev
15:55nate8 has quit [Quit: Leaving]
15:55nate8 has joined #asahi-dev
15:59vivithecanine2 external in clamshell works fine on mbp m1 max
15:59vivithecanine3 works fine even
16:00vivithecanineI think the max does a total of 5 displays?
16:00svenfor the pro/max chips, not for the base m1/m2 variant
16:00jannauthis would be for the m2 mb air and and pro 13" which support only 1 external display officially
16:04marcan< sven> pretty sure the second thing is a lie :D
16:04marcanno that tracks
16:05svenyeah, nvm, i'm just an idiot
16:05marcanthat intel chip should be what's in the dongle bridging to PCIe
16:05sventhe config space was just completely messed up because I unplugged the adapter and that's why I assumed it was broken
16:06sventhis currently runs on essentially duct tape so i'll have to clean that up and then figure out how to get that DART up sanely
16:06marcanalso, for the DART, vague idea I had was to play with runtime-pm somehow? like have the DART driver defer doing anything and start up in rpm sleep state, only to be woken up when the PCIe dependency needs it. or something something.
16:06marcaneither that or maybe trigger on IOMMU domain attach? not sure if that would work
16:06svenyeah, i have some ideas.
16:07sveni'll see what happens if I just fake a clock controller for now just to see if that breaks already
16:07svenit might just work because it technically doesn't need the DART before the first device comes up
16:07marcanhow would that work?
16:07marcanyou mean have the DART depend on a clock supplied by PCIe?
16:08marcancan you bring that up independent of any thunderbolt/hotplug/etc stuff?
16:08sventurns out I can just leave the thunderbolt stuff up the entire time because it's much less broken than dwc3
16:09marcansure, as long as we can still sanely power sequence it to off for power saving
16:09marcanbut that should work ish with the rpm stuff I think?
16:09marcanthough I still need to figure out rpm for the DARTs properly
16:09jannauruntime pm might be annoying, the internal pcie devices are runtime pm enabled on probe_device
16:09svenI actually haven't figured out how to bring it down without resetting the entire soc yet :D
16:10svenbut it should be possible, just need to figure out the right atcphy/acio sequence there
16:10svenI think I know it but since this is all duct taped together I can't easily test it
16:11svenif you mess that one up though you either get SErrors on pcie/acio MMIO access or some internal watchdog just resets everything
16:14leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:15jannaumarcan: with https://github.com/jannau/linux/commit/e941a3e8a1dd7ecdcaa032d1fe75b2fb97ccff3c there is at least no breakage with DART runtime PM, not sure if runtime PM would work correctly if any of the pcie drivers supported it
16:16leitao has joined #asahi-dev
16:22nicolas17 has joined #asahi-dev
16:28leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:29leitao has joined #asahi-dev
16:36yamii has quit [Remote host closed the connection]
16:36yamii has joined #asahi-dev
16:39Guest1569 has quit [Ping timeout: 480 seconds]
16:42seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:42seeeath has joined #asahi-dev
16:59povikfinally got some IV sense curves: https://ibb.co/5KWrpg1
16:59kettenis_ has joined #asahi-dev
17:00povikthe curves are roughly overlaid because TI chose the units in a ~3.7 ohm ratio (understandably)
17:01cylm has joined #asahi-dev
17:01kettenis has quit [Ping timeout: 480 seconds]
17:03marcanpovik: nice! :D
17:03marcankinda surprising it's that jagged though, unless the input wasn't a sine wave?
17:05leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
17:07leitao has joined #asahi-dev
17:18svenokay, removing and plugging in again works when I manually set that tunnel reset but again inbetween as well
17:23faruk has joined #asahi-dev
17:24jannauany idea how to wake from s2idle in upstream v6.2-rc4 (without macsmc)? I tried enabling wakeup for usb devices and ttySAC0 but both don't appear to work
17:28bcrumb has joined #asahi-dev
17:32bcrumb_ has joined #asahi-dev
17:33bcrumb_ has quit []
17:33bcrumb_ has joined #asahi-dev
17:37bcrumb has quit [Ping timeout: 480 seconds]
17:39povikmarcan: this was a sine wave, played not too loud
17:40povikhuh, datasheet says noise should be at -60 dB, which i would expect to look less jagged
17:43povikthis is the spectral density: https://ibb.co/kSzwzst
18:00povikyeah, but the -60 dB figure is quoted at 5 W, and if i didn't miscalculate i only drove 10 mW to it
18:05bcrumb_ has quit [Quit: WeeChat 3.8]
18:07faruk has quit [Read error: Connection reset by peer]
18:39leitao has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
18:56seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
18:56seeeath has joined #asahi-dev
18:57eiln has joined #asahi-dev
19:02faruk has joined #asahi-dev
19:02faruk has quit []
19:08eiln has quit [Remote host closed the connection]
19:45nate8 has quit [Remote host closed the connection]
19:46nate8 has joined #asahi-dev
20:27seeeath has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
20:27seeeath has joined #asahi-dev
20:36eiln has joined #asahi-dev
20:37eilnrough benchmark & progress https://pastebin.com/UGyeNH6T :)
21:08___nick___ has quit [Ping timeout: 480 seconds]
21:35abby has quit [Remote host closed the connection]
21:35abby has joined #asahi-dev
21:43psykose has quit [Remote host closed the connection]
21:44psykose has joined #asahi-dev
21:57jannausigh, 16GB to 32GB m2 pro RAM upgrade cost as much as the 32GB to 64GB
21:59j`eyweird that you can only get pro m2 mini and not max
22:00nicolas17max is reserved for Mac Studio 2023 maybe?
22:01nicolas17btw looks like MBP 14" M2 Pro is "j414"
22:01jannauj414s probably
22:02nicolas17the raw thing I found (from a regulatory listing) was "J414 A2779 MBP 14 SG/10C/16C GPU/16G/512G-USA~||~MPHF3"
22:02jannaustill no new 13.2 beta/rc
22:03nicolas17I wonder what "SG" means
22:03nicolas17Space Gray maybe
22:05bluetail has quit [Quit: The Lounge - https://thelounge.chat]
22:06nicolas17FCC wireless testing documents also mentioned these new machines running macOS 22A8375 and 22A12320r (ie. macOS 13.0),
22:06nicolas17sometimes I need another reminder of how long hardware production cycles are :p
22:08bluetail has joined #asahi-dev
22:45jannaudid anyone already ran ventura 13.2 beta or earlier in the HV? panics here immediately with "trust cache module 0 has unsupported version 2 @trustcache.c:113" and some AppleImage4 errors before
22:50karolherbst has quit [Remote host closed the connection]
…