#asahi-dev · 2022-11-22 18:35
rqou_nice work on the release / progress report!#asahi-dev · 2022-11-22 18:34
rqou_sven: re backlog question about usb3 in device mode: when you had first posted atcphy months and months ago, gadget mode did *not* work (neither usb2 nor usb3). i have not tested the current version#asahi-dev · 2022-10-12 00:20
rqou_the memory descriptors are accessed behind a dart, which is what `mapper-avd0-piodma` appears to describe#asahi-dev · 2022-10-12 00:18
rqou_i have tested the AVD one with a hardware test, but i have not personally tested the scaler one (but the "shape" of the register interface is definitely identical)#asahi-dev · 2022-10-12 00:17
rqou_for AVD this is actually used, but for the scaler it does not appear to be used (but the driver contains code for it)#asahi-dev · 2022-10-12 00:17
rqou_idk if this is the same thing or just a naming clash, but "piodma" in the context of AVD and the hardware scaler refers to an actual piece of hardware that can copy data to/from hardware registers according to memory descriptors#asahi-dev · 2022-10-01 00:04
rqou_the hardware is simply "feed it a frame (and a dma descriptor) and the hardware tells you when it's done"#asahi-dev · 2022-10-01 00:01
rqou_i've also been intermittently working on AVD but don't have much to show for it yet#asahi-dev · 2022-10-01 00:00
rqou_rcombs: i might not be consistently online, but i read logs periodically#asahi-dev · 2022-09-30 23:59
rqou_rcombs: i've done enough reverse engineering of the hardware that it should be possible to start developing a linux-side driver, but i don't know anything about what would be needed on the kernel-userspace interface or above#asahi-dev · 2022-09-30 23:58
rqou_rcombs: (reading logs) i would like to poke you wrt ProRes#asahi-dev · 2022-07-18 17:53
rqou_congrats on the release!#asahi-dev · 2022-07-04 21:35
rqou_anybody know anybody with the seemoo-lab folks who might be interested in this driver?#asahi-dev · 2022-07-04 21:34
rqou_i wonder if the BT firmware is signed and how deep that rabbit hole can go#asahi-dev · 2022-07-04 21:28
rqou_i thought you had something much fancier#asahi-dev · 2022-07-04 21:28
rqou_oh#asahi-dev · 2022-07-04 21:27
rqou_heh, i need to set up the tooling to patch stuff from the hv#asahi-dev · 2022-07-04 21:25
rqou_not sure how you figured out the beamforming bits#asahi-dev · 2022-07-04 21:25
rqou_yeah i saw that too#asahi-dev · 2022-07-04 21:22
rqou_hope the prototype was useful#asahi-dev · 2022-07-04 21:22
rqou_nice job btw (although i haven't tested it yet)#asahi-dev · 2022-07-04 21:21
rqou_well, if it works ¯\_(ツ)_/¯#asahi-dev · 2022-07-04 21:20
rqou_oh, not using 0x154 at all but using 0x174 in all cases?#asahi-dev · 2022-07-04 21:17
rqou_did you ever figure out the SCO doorbell properly?#asahi-dev · 2022-07-04 21:16
rqou_wtf#asahi-dev · 2022-07-04 21:15
rqou_guess not#asahi-dev · 2022-07-04 21:15
rqou_i was wondering whether or not that would work#asahi-dev · 2022-07-04 21:15
rqou_yeah, i did see that in the logs#asahi-dev · 2022-07-04 21:13
rqou_... huh#asahi-dev · 2022-07-04 21:13
rqou_*iirc the tracer also said it doesn't#asahi-dev · 2022-07-04 21:13
rqou_iirc what the tracer said#asahi-dev · 2022-07-04 21:13
rqou_the plist definitely doesn't say virtual#asahi-dev · 2022-07-04 21:11
rqou_anyways, in case you were wondering why the ACL D2H pipe doesn't set the virtual flag, this is why#asahi-dev · 2022-07-04 21:10
rqou_i do know that the loopback test command appears to crash the firmware#asahi-dev · 2022-07-04 21:10
rqou_idk, never tried that command#asahi-dev · 2022-07-04 21:05
rqou_(also, sorry for not having been very active, Life(TM) happened)#asahi-dev · 2022-07-04 21:05
rqou_iirc i needed to do this in order for OBEX to work#asahi-dev · 2022-07-04 21:05
rqou_and which scenario happened is indicated by bits in flags#asahi-dev · 2022-07-04 21:04
rqou_if the data doesn't fit in the completion ring footer, it goes into the buffer you queued up#asahi-dev · 2022-07-04 21:04
rqou_and, if the data fits in the completion ring footer, it'll appear there#asahi-dev · 2022-07-04 21:04
rqou_tl;dr you queue up a buffer in the *transfer* ring for the D2H ACL pipe#asahi-dev · 2022-07-04 21:04
rqou_sven: (reading logs) the logic around pipe6_iobuf_off in my hacky PoC takes care of the "ACL too big to fit into completion ring" situation#asahi-dev · 2022-04-13 18:49
rqou_unfortunately it's not very clear (at least to me) what any of it means#asahi-dev · 2022-04-13 18:48
rqou_https://github.com/geohot/tinygrad/tree/master/accel/ane#asahi-dev · 2022-04-13 18:48
rqou_chadmed_: "i dont think anyone's had a real look at the webcams, SEP or neural engine" <-- geohot has done some work on neural engine independent of the asahi project#asahi-dev · 2022-04-09 17:07
rqou_sven: oh ok, that makes sense. thanks for that explanation#asahi-dev · 2022-04-09 07:22
rqou_but not super important#asahi-dev · 2022-04-09 07:22
rqou_anyways, i'm not familiar with the linux dma-iommu api, but vfio's interface is quite high-level and "should" be possible to get working with mismatched page sizes with "just" some coding#asahi-dev · 2022-04-09 07:18
rqou_once we have tbt it's possible!#asahi-dev · 2022-04-09 07:18
rqou_idk if anybody would be nuts enough to try that on m1