#asahi-re · 2022-11-28 17:47
eilndriver is responsible for packing the BAR #asahi-re · 2022-11-28 17:47
eilnwhich has a very specific order on mapping the input/output dvas #asahi-re · 2022-11-28 17:47
eilnin each hw queue there's a BAR slot #asahi-re · 2022-11-28 17:46
eilnwait no, I misread #asahi-re · 2022-11-28 17:45
marcancan ane programs decide where to do data reads or are the input/output buffers programmed by the driver always?#asahi-re · 2022-11-28 17:45
marcanwhat is the kernel responsible for as far as deciding where data buffers live#asahi-re · 2022-11-28 17:45
marcanok, you probably need to make sure you understand that fairly well before starting to design the linux driver since it completely changes the design#asahi-re · 2022-11-28 17:44
eilnI actually have no idea #asahi-re · 2022-11-28 17:44
marcanbut are there separate DART contexts?#asahi-re · 2022-11-28 17:44
eilnbut only one main queue that triggers the circuit #asahi-re · 2022-11-28 17:44
eilnthere are 8 hw queue slots which can be stacked #asahi-re · 2022-11-28 17:43
marcanif it's shared then the kernel driver needs to deal with all input/output/etc addresses, which I'm guessing is the case here? (no addresses in untrusted code/kernels from userspace)#asahi-re · 2022-11-28 17:43
marcanlike can it have different user contexts/VMs or is it all one shared address space?#asahi-re · 2022-11-28 17:43
marcanone important question there is how ANE handles multiprocessing#asahi-re · 2022-11-28 17:43
marcaneiln: also just to give you an idea: AIUI the main thing you care about re the DRM/linux interface is how you do memory management and how you do work submission from userspace.#asahi-re · 2022-11-28 17:42
sven:D#asahi-re · 2022-11-28 17:42
eilnyou can see him transition to ane lol#asahi-re · 2022-11-28 17:42
eilnhttps://patents.google.com/?inventor=Christopher+L.+Mills#asahi-re · 2022-11-22 18:51
millenialhackermagic! #asahi-re · 2022-11-22 18:51
svenall changes #asahi-re · 2022-11-22 18:51
millenialhackersven, what poll does? it shows the changes or just the new contents? #asahi-re · 2022-11-22 18:50
svenjust call poll whenever a message between ap and isp is exchanged #asahi-re · 2022-11-22 18:50
sventhatll work for normal memory as well#asahi-re · 2022-11-22 18:50
sventhere regmonitor or however it’s called#asahi-re · 2022-11-22 18:50
millenialhackerAny tip you think may be helpful on this situation? #asahi-re · 2022-11-22 18:49
millenialhackerI want to understand how a particular memory region with a defined size changes over time so I can guess more or less what kind of struct it host#asahi-re · 2022-11-22 18:49
millenialhackerI can try to watch the memory but I'm not clear if that's possible with M1N1. #asahi-re · 2022-11-22 18:49
millenialhackerbut I have no idea how to approach struct RE #asahi-re · 2022-11-22 18:48
millenialhackerI guess those allocations are for two things: Internal structs and data (maybe images)#asahi-re · 2022-11-22 18:48
millenialhackerHi folks, I'm looking for a RE tip. ISP has a channel for ISP to request memory allocations and also to free them. ISP requests AP let's say 500 bytes, and AP returns in the channel the address of that memory requested. #asahi-re · 2022-11-15 08:42
marcan(with a gap between atcphy0 and 1)#asahi-re · 2022-11-15 08:41
sventhanks. will check later #asahi-re · 2022-11-15 08:41
marcanI didn't check the atcphy registers but it sounds like it should all be the same#asahi-re · 2022-11-15 08:41
marcanafaict all the bit widths should be the same and all fields contiguous in file order#asahi-re · 2022-11-15 08:41
marcanplease check my DT for obvious fails :p#asahi-re · 2022-11-15 08:40
marcananyway, as long as the code works with a non-workaround atcphy1 there's no issue#asahi-re · 2022-11-15 08:39
marcanweeeeird#asahi-re · 2022-11-15 08:39
marcansven: yeah, you're right, it totally does that#asahi-re · 2022-11-15 08:34
svenat least I think so, jannau knows the details ;)#asahi-re · 2022-11-15 08:34
marcancool#asahi-re · 2022-11-15 08:34
svenyup#asahi-re · 2022-11-15 08:34
marcanand t6002 is the same with the fuses split across both dies I'm guessing?#asahi-re · 2022-11-15 08:34
svenR did that #asahi-re · 2022-11-15 08:34
svenyes#asahi-re · 2022-11-15 08:34
marcanwe have t6000 and t6001?#asahi-re · 2022-11-15 08:33
sveniirc that was the case for t6000 as well#asahi-re · 2022-11-15 08:33
svenNice, thanks!#asahi-re · 2022-11-15 08:33
marcanfor each atcphy#asahi-re · 2022-11-15 08:33
marcanthe fuses are exactly the same, just with different start offsets#asahi-re · 2022-11-15 08:33
marcansven: either way this is what I get: https://mrcn.st/p/x9z9sIx9