#
…

2025-08-22

01:15hdbngr has joined #asahi-dev
01:24hdbngr has quit [Ping timeout: 480 seconds]
01:29Pokemod97736 has quit []
01:48rivendell has quit [Remote host closed the connection]
01:49hdbngr has joined #asahi-dev
01:57hdbngr has quit [Ping timeout: 480 seconds]
02:04Pokemod97736 has joined #asahi-dev
02:17beastd has joined #asahi-dev
02:18beastd|2 has quit [Ping timeout: 480 seconds]
02:27hexdump0815 has joined #asahi-dev
02:29hexdump02 has quit [Ping timeout: 480 seconds]
02:40tobhe has joined #asahi-dev
02:47tobhe_ has quit [Ping timeout: 480 seconds]
03:01skipwich has quit [Remote host closed the connection]
03:02skipwich has joined #asahi-dev
03:09nickchanjannau: Apple TV HD uses USB-C
03:09nickchanhas A8 chip
03:10nickchanBut I haven't need to handle that any different from iPhone 6
03:15hdbngr has joined #asahi-dev
03:24hdbngr has quit [Ping timeout: 480 seconds]
03:46hdbngr has joined #asahi-dev
03:55hdbngr has quit [Ping timeout: 480 seconds]
03:59nickchanand T2 is a newer SoC than A11 that's still exploitable though T2 is a mix of old and new hw blocks
03:59nickchanso I suppose A11 has the most latest tech apple SoC that's still exploitable
04:00nickchanand of course the T2s has USB-C ports but I haven't seen the need to deal with them any different from iPhone 7
04:00nickchanother than hotplug not working
04:04nora has joined #asahi-dev
04:06nora_ has quit [Ping timeout: 480 seconds]
04:29hdbngr has joined #asahi-dev
04:37hdbngr has quit [Ping timeout: 480 seconds]
04:38nickchandid anyone work on the ancient jpeg block?
04:39nickchan(in Linux)
05:00hdbngr has joined #asahi-dev
05:08hdbngr has quit [Ping timeout: 480 seconds]
06:20jannaunickchan: I was mustly wondering if there are devices with cd321x without thunderbolt or displayport when reading the commit description of "[PATCH RFC 14/22] usb: typec: tipd: Register DisplayPort and Thunderbolt altmodes for cd321x"
06:22jannauI don't think anyone started writing a linux driver for the jpeg block
06:24jannauit's third-party IP so there might be a linux driver. not upstream though (checked around 2022)
06:25sveniirc someone (alyssa or maybe eiln) reverse engineered it a long while ago
06:26svennever wrote a driver though
06:27svensame for the aes block, there’s some script in m1n1 but I never bothered to write a driver because it’s slower than the main cpu with the optimized arm64 implementation Linux has
06:27jannauI think R reversed it
06:27svenoh yeah, that sounds like something they would’ve done
06:30hdbngr has joined #asahi-dev
06:38hdbngr has quit [Ping timeout: 480 seconds]
06:44chadmed has joined #asahi-dev
06:53chadmedjannau: i think one of the more recent iphone models uses cd3217
06:53chadmedi think it was the 15 pro, which has usb-c but can only do usb 2.0
06:54chadmednot sure if the 16 pro can do tbt or just usb 3.2 gen 2 2x2 2.0 2
06:54chadmedor whatever usb-if are calling it this quarter
06:56chadmedchaos_princess: i was gonna look at libfdt either while im away or when i get back
06:56chadmedanyone got any strong opinions on what the rust api should look like? i dont think it should really look like how it looks now
06:56jannauI think the iphone 15 pro uses the spmi variant, sn201202x. also not sure if it makes sense to worry about devices which likely will never run linux
06:57chaos_princess15 pro should do 3.0
06:57chaos_princess(If it is a pro and has usb, it can do 3.0)
06:57chadmedno there was some quirk
06:57chadmedthe first pro which had usb-c could still only do 2.0
06:58chadmedthere was a dwc2 behind the phy or something stupid like that
06:58hdbngr has joined #asahi-dev
06:59chadmedah nvm it was the base 15
06:59chadmedi thought the 15 pro was the only one that had usb-c for that gen
07:00pb17 has quit [Ping timeout: 480 seconds]
07:03pb17 has joined #asahi-dev
07:03svenmy 15 pro at least does display port via a dell thunderbolt dock
07:03svenI assume that is dp altmode but dunno
07:03svencould be dp over tbt as well
07:07hdbngr has quit [Ping timeout: 480 seconds]
07:08svenbut anyway, I wouldn’t worry about devices that will likely never run Linux
07:08svenand only deal with that situation if it really happens
07:20jannauiphone (pro) have so far only usb3 with dp-altmode. I wouldn't be surprised if A family SoCs lack apciec if the saved die space is worth the effort. even the ipad airs with m2/m3 do not advertise usb4
07:31hdbngr has joined #asahi-dev
07:39hdbngr has quit [Ping timeout: 480 seconds]
07:51___nick___ has joined #asahi-dev
07:51svenohh.. that glue.h stuff the dwc3 maintainer suggested looks interesting. maybe I can move all the crap this thing needs to dwc3-apple.c
07:54svenbut I guess that would mean dropping the snps,dwc3 compatible
07:57jannauyes, I was thinking the same. added in https://lore.kernel.org/all/20250414-dwc3-refactor-v7-0-f015b358722d@oss.qualcomm.com/ (April 2025) but started in at the end of 2023
07:59svenyeah, but that looks like it’ll make everything much cleaner
07:59svenalso confused why the dt checks are failing, I remember running them before submitting and I didn’t get the errors :(
08:00svenbut I think I can even more the role switch mess into dwc3-apple.c with that
08:05___nick___ has quit []
08:07___nick___ has joined #asahi-dev
08:07___nick___ has quit []
08:10nickchanchadmed: base iphone 15 has dwc2 and dwc3 but still can only do usb 2.0
08:10___nick___ has joined #asahi-dev
08:11nickchandwc2's in A-series up to A16 and for the bigger ipad chips up to A12X/A12Z
08:12nickchandwc3 is added in A13/A12X/A12Z
08:12nickchanin any case there's no need to worry about cd321x without tbt stuff besides iPhones and iPads there's only adp3,2 and that doesn't count
08:13nickchanI mean besides iPads
08:17hdbngr has joined #asahi-dev
08:17nickchanalso ive checked on my m2 ipad pro i don't think it advertises thunderbolt/usb4
08:18nickchanit has jailbreak in case anyone wants to check stuff
08:19pb17 has quit [Ping timeout: 480 seconds]
08:20chadmedslightly ot but sometimes i wonder why mobile device manufacturers bother hard-locking bootloaders to the degree that they do. the vast vast vast majority of people arent going to know or care what the bootloader is and how to put linux on these things anyway
08:20chadmedand the like 10 of us who do are not going to take your revenue if you let us do stuff to it
08:21pb17 has joined #asahi-dev
08:22chaos_princessif you allow people to unlock bootloaders, they will screengrab a 4k version of generic-netflix-slop-2025-q3
08:22sven„just“ do it like Apple: disable all drm keys if booting unsigned code
08:23chadmedi can already get generic.netflix.slop.2025.q3.2160p.10-bit.x265.webrip.amzn.mkv
08:23chadmedbut yeah ik thats the "real" answer
08:26hdbngr has quit [Ping timeout: 480 seconds]
08:27nickchani don't actually know if it's possible to run linux on those iphone/ipads with EL2 with only the exploits out there tbh
08:28nickchanEL2 is CTRR protected but it might just be possible to put Linux on EL1 after manipulating HCR_EL2 and pagetables in just the right way
08:28svendoes that thing already have gxf?
08:28nickchansven: they do have gxf
08:29svencause then stuff like iommus are protected by gl2 and you can map that mmio region
08:29svenunless you have a gxf bypass that allows pagetable manipulation ofc
08:29opticron has quit [Remote host closed the connection]
08:29nickchanoh that's a basic assumption, required for jailbreak on 15.2+
08:29svenah okay
08:30svenso maybe if can drop to el1 and disable SPRR/GXF there then and map all of physical memory
08:30opticron has joined #asahi-dev
08:30nickchanthe vmsa lock register will not allow GXF and SPRR to be disabled
08:31sventalking about el1 here
08:32hdbngr has joined #asahi-dev
08:32nickchanone can go from EL1 -> GL1 in a Linux VM under macOS
08:33nickchannow of course macOS handle this case gracefully to the point where if the vmm don't have a certain entitlement it wouldn't continue running that vm
08:40hdbngr has quit [Ping timeout: 480 seconds]
09:08rivendell has joined #asahi-dev
09:08hdbngr has joined #asahi-dev
09:16rivendell has quit [Ping timeout: 480 seconds]
09:17hdbngr has quit [Ping timeout: 480 seconds]
09:17rivendell has joined #asahi-dev
09:25svenuh, gl1 require sprr in el1 to be enabled and that would really mess any Linux vm up
09:26svenand when I originally reverse engineered sprr that just wasn’t enabled in el1 vms under macOS unless you had a special apple entitlement
09:33hdbngr has joined #asahi-dev
09:40hdbngr has quit [Remote host closed the connection]
09:40hdbngr has joined #asahi-dev
09:49hdbngr has quit [Ping timeout: 480 seconds]
10:02hdbngr has joined #asahi-dev
10:10hdbngr has quit [Ping timeout: 480 seconds]
10:34hdbngr has joined #asahi-dev
10:51hdbngr has quit [Ping timeout: 480 seconds]
10:57__nick__ has joined #asahi-dev
11:00___nick___ has quit [Ping timeout: 480 seconds]
11:01cylm has joined #asahi-dev
11:35hdbngr has joined #asahi-dev
11:43hdbngr has quit [Ping timeout: 480 seconds]
11:50hdbngr has joined #asahi-dev
12:25cylm_ has joined #asahi-dev
12:32cylm has quit [Ping timeout: 480 seconds]
13:37hdbngr has quit [Ping timeout: 480 seconds]
13:48hdbngr has joined #asahi-dev
13:53rivendell has quit [Remote host closed the connection]
14:41svenjust started prototyping with glue.h and this already looks much nicer so far
15:23cylm has joined #asahi-dev
15:30cylm_ has quit [Ping timeout: 480 seconds]
15:31Halian has quit [Quit: I'll be back.]
16:42pb17 has quit [Ping timeout: 480 seconds]
16:45pb17 has joined #asahi-dev
17:16perichus has joined #asahi-dev
17:29perichus has quit [Quit: perichus]
17:39cylm_ has joined #asahi-dev
17:41cylm has quit [Ping timeout: 480 seconds]
18:01dev-muhammad-adel has joined #asahi-dev
18:02dev-muhammad-adel has quit [Remote host closed the connection]
18:59cylm has joined #asahi-dev
19:00cylm_ has quit [Ping timeout: 480 seconds]
19:46chadmed has quit [Quit: Konversation terminated!]
19:47chadmed has joined #asahi-dev
19:50chadmed has quit []
19:50chadmed has joined #asahi-dev
20:07__nick__ has quit [Remote host closed the connection]
20:18opoku has quit [Ping timeout: 480 seconds]
20:59rivendell has joined #asahi-dev
21:16rivendell has quit [Ping timeout: 480 seconds]
21:18rivendell has joined #asahi-dev
21:22chaos_princessuhh, github probably did not show you the line i was targeting
21:23chadmedyeah it just took me to the makefile diff
21:24chadmedohhh right just left over detritus from when it was all based on offsets
21:26cylm_ has joined #asahi-dev
21:26chadmedthere was a line somewhere (i think in get reg) that checked if the first offset was 0 and returned early
21:27chaos_princessnot the issue
21:27chaos_princessi was getting a trace for, say /arm-io/sgx and getting [arm-io, sgx, root node].
21:27chaos_princesswhen it roundtrips via c, root node is gone
21:27chaos_princessbut it confuses rust
21:28chadmedoh right its the last one
21:28cylm has quit [Ping timeout: 480 seconds]
21:29chadmedoh you know what
21:29chadmedit is now very obvious in hindsight but i was so locked in on getting rid of offsets
21:30chadmedthis was probably supposed to be a null terminator and not actually a pointer to the root node
21:30chaos_princessprobably what it was
22:17nyx_o has quit []
22:30nyx_o has joined #asahi-dev
23:28hdbngr has quit [Ping timeout: 480 seconds]
23:30rivendell has quit [Quit: rivendell]
…