#asahi-dev · 2023-01-27 22:28
chadmed_interestingly i cant find a commit in the PKGBUILD config where CONFIG_BLK_RAM was changed, so idk why it wasnt probing with the stable kernel#asahi-dev · 2023-01-27 22:13
chadmed_but i think i noticed it drop out of the config before that#asahi-dev · 2023-01-27 22:13
chadmed_marcan: i asked about it last month https://oftc.irclog.whitequark.org/asahi-dev/search?q=brd #asahi-dev · 2023-01-27 08:01
chadmed_marcan: -dev was/is necessary to create a ramdisk for building the root image for the installer using brd. if we could get brd in the default kernel config like it used to be i can drop -dev as a requirement. i try to test every release to make sure nothing breaks but i mustve missed something, sorry#asahi-dev · 2023-01-22 12:09
chadmed_or can we just say "if dmi table use this codepath if of use this other codepath" #asahi-dev · 2023-01-22 12:09
chadmed_would it be worth just not unifying them, but factoring out the backend stuff we can and still have an independent driver for the intel machines#asahi-dev · 2023-01-22 12:07
chadmed_ah... yeah...#asahi-dev · 2023-01-22 12:07
chadmed_does the existing intel driver work for t2 macs?#asahi-dev · 2023-01-22 12:05
chadmed_yeah the way they do it is odd too#asahi-dev · 2023-01-22 12:04
chadmed_ChaosPrincess: 90% of that applesmc-intel driver can be factored out into an x86 backend for smc core #asahi-dev · 2023-01-22 09:57
chadmed_idek if thats possible though, i need to read up on how dts are compiled#asahi-dev · 2023-01-22 09:56
chadmed_then you could just walk the table and build the required structures in the driver#asahi-dev · 2023-01-22 09:56
chadmed_or something like that#asahi-dev · 2023-01-22 09:56
chadmed_have an element for each key with smc-key and smc-label props #asahi-dev · 2023-01-22 09:55
chadmed_i was thinking something like the opp table #asahi-dev · 2023-01-22 09:51
chadmed_building the hwmon_channel_info array based on dt info is possible just annoying#asahi-dev · 2023-01-22 09:51
chadmed_jeffmiw: yeah i was just playing around with how that would look#asahi-dev · 2023-01-21 22:16
chadmed_povik: ack no worries#asahi-dev · 2023-01-21 13:13
chadmed_povik: is there anything i can cherry pick to get V/ISENSE working in some capacity? i want to wire up and start testing the safety daemon at some point next week if possible#asahi-dev · 2023-01-20 13:54
chadmed_society if there was no broadcom: https://imgur.com/a/5XcLfJm#asahi-dev · 2023-01-20 13:51
chadmed_mkurz: itll get there soon enough, better to release something that works reliably than a hacky janky mess#asahi-dev · 2023-01-20 13:50
chadmed_no, hence why HDMI Out on the feature page is WIP #asahi-dev · 2023-01-20 13:49
chadmed_yes, DCP is not the only thing involved here#asahi-dev · 2023-01-20 13:49
chadmed_DCP is just the display controller, theres a whole ballet dance that needs to happen between it, the DP routing circuitry, the usb-c mux chip, the ATCPHY... #asahi-dev · 2023-01-20 13:48
chadmed_DCP works, but the code to route a dp signal from it to the type-c phy is not yet in the kernel #asahi-dev · 2023-01-20 13:43
chadmed_theres a dp to hdmi chip on the board #asahi-dev · 2023-01-20 13:43
chadmed_mkurz: the hdmi port on the laptops is actually a displayport on an atcphy #asahi-dev · 2023-01-20 10:44
chadmed_j`ey: the patch number actually went down for the first time ever with 6.2 :P #asahi-dev · 2023-01-19 12:22
chadmed_would bypass be necessary to pass through externally connected devices to an x86 vm or something like that?#asahi-dev · 2023-01-19 12:21
chadmed_im guessing thats been "fixed" for t600x and newer given they're allowed to call them tb4#asahi-dev · 2023-01-16 20:53
chadmed_its the kb that doesnt work on my machine#asahi-dev · 2023-01-16 20:53
chadmed_marcan: mb, wake on lid does work#asahi-dev · 2023-01-14 21:29
chadmed_well no i dont wonder, its a janky hack :P #asahi-dev · 2023-01-14 21:29
chadmed_joske: i wonder why its unreliable... #asahi-dev · 2023-01-13 12:31
chadmed_yeah i dont think intel would be stupid enough to try to say "oh btw to use tb4 you have to have vt-d which we dont license lol soz" and if they were that stupid theyd get immediately ground into a fine paste by competition watchdogs the world over#asahi-dev · 2023-01-13 12:29
chadmed_i believe its "real" TB4 because the spec pages for t8103 and t8112 products make a note of saying "Thunderbolt 3/USB4" instead#asahi-dev · 2023-01-13 12:28
chadmed_yeah the studio spec page mentions it a few times#asahi-dev · 2023-01-13 12:26
chadmed_intel integrate thunderbolt controllers on their socs whereas board vendors have to chuck an intel controller on the pcie bus for amd (expensive and takes away pcie lanes from things people actually use like nvme) #asahi-dev · 2023-01-13 12:15
chadmed_“I wouldn’t read it as it’s only Intel, because of VT-d,” Ziller said. “If [at AMD] there was an equivalent technology that supports DMA protection, that supports prevention against physical attacks, then that would be the requirement.”#asahi-dev · 2023-01-13 12:13
chadmed_a company we work with tries to sell MS SQL Server 2019 as their own branded solution... until the implementers install it on client servers of course ;) #asahi-dev · 2023-01-13 12:12
chadmed_arnd: being that it's a press deck id read any reference to "Intel VT-d" as being "some implementation of DMA remapping and virtualisation"#asahi-dev · 2023-01-13 12:09
chadmed_theres an intel "paper" floating around somewhere where they use the terms VT-d/AMD IOMMU/ARM SMMU interchangeably when talking about dma hardening in the uefi env#asahi-dev · 2023-01-13 12:08
chadmed_the spec page for the mac studio mentions TB4 explicitly a few times, and they sell TB4 cables #asahi-dev · 2023-01-13 12:07
chadmed_Video Support: Thunderbolt 4 digital video output#asahi-dev · 2023-01-13 11:54
chadmed_there's a few commits it _might_ be#asahi-dev · 2023-01-13 11:43
chadmed_mps: on the headphone jack? #asahi-dev · 2023-01-13 11:33
chadmed_plus intel isnt really in a strategic position to be making enemies out of anyone atm#asahi-dev · 2023-01-13 11:31
chadmed_which wouldnt stop amd, and theyd be braindead to try and thwart apple given the co-designed the spec out of what was left of lightpeak... #asahi-dev · 2023-01-13 11:28
chadmed_arnd: iiuc they don't mandate specifically vt-d specifically, but support for dma virt/directed io in general #asahi-dev · 2023-01-12 08:53
chadmed_the fact that the capacity factor in the dt is called dmips/mhz doesnt mean anything, its just treated as a dimensionless constant and the scheduler just uses the ratios between the values in a given system