#asahi-dev · 2023-04-11 14:25
marcan(and why does that affect mesa?)#asahi-dev · 2023-04-11 14:25
marcanoh, did we grow cursor plane support?#asahi-dev · 2023-04-11 14:07
marcanand gpu#asahi-dev · 2023-04-11 14:07
marcanso... wifi left?#asahi-dev · 2023-04-11 14:03
marcanjannau: feel free to refactor as you see fit, I left the commits broken out so you can see what I did (except the DT squash)#asahi-dev · 2023-04-11 14:01
marcanit just needs not to straddle an ias boundary (so the size needs to be right)#asahi-dev · 2023-04-11 14:01
marcanhopefully I got the DART logic right but I think I did#asahi-dev · 2023-04-11 14:00
marcan(even if ias is downgraded in locked mode)#asahi-dev · 2023-04-11 14:00
marcan(even outside the ias)#asahi-dev · 2023-04-11 14:00
marcanthe dma-range can contain upper bits which does what we want here#asahi-dev · 2023-04-11 14:00
marcanI think we can get rid of the asc-dram-mask stuff entirely now, assuming this also works out on M2#asahi-dev · 2023-04-11 13:59
marcanpushed!#asahi-dev · 2023-04-11 13:54
marcanok, let me commit :)#asahi-dev · 2023-04-11 13:54
marcanworks!#asahi-dev · 2023-04-11 13:53
marcanwe pick the dva#asahi-dev · 2023-04-11 13:53
marcanwe just map it there and give it the dva#asahi-dev · 2023-04-11 13:53
marcanit's an output#asahi-dev · 2023-04-11 13:53
marcanno no#asahi-dev · 2023-04-11 13:50
marcaninstead of requesting addrs, just map it by index#asahi-dev · 2023-04-11 13:50
marcanmakes sense, this is the design fix for their stupid vuln#asahi-dev · 2023-04-11 13:50
marcanwe're supposed to just map it there#asahi-dev · 2023-04-11 13:50
marcanthe new u64 in the dcpep_cb_map_reg is the dva#asahi-dev · 2023-04-11 13:49
marcanohh I think I get it#asahi-dev · 2023-04-11 13:43
marcanbut, that's not the issue, the issue is dma_map_resource() is failing#asahi-dev · 2023-04-11 13:43
marcanI'm using the dma-range stuff now#asahi-dev · 2023-04-11 13:43
marcanactually just 0#asahi-dev · 2023-04-11 13:33
marcanI think it's missing the dma mask stuff#asahi-dev · 2023-04-11 13:29
marcanit's the RT thing I think#asahi-dev · 2023-04-11 13:23
marcanbut I don't know if macOS ever does a real modeset? maybe after DPMS?#asahi-dev · 2023-04-11 13:23
marcanso it dies during a real modeset#asahi-dev · 2023-04-11 13:22
marcanI'm going to be annoyed if this is an outright firmware bug that macOS just never hits...#asahi-dev · 2023-04-11 13:22
marcanit's dying switching to 60Hz mode#asahi-dev · 2023-04-11 13:22
marcanso it works in 120Hz mode#asahi-dev · 2023-04-11 13:22
marcanjannau: ha.#asahi-dev · 2023-04-11 13:18
marcanI fixed the RT thing#asahi-dev · 2023-04-11 13:18
marcanjannau: it's dying in set_digital_out_mode#asahi-dev · 2023-04-11 13:08
marcanfixing the RT bandwidth stuff now#asahi-dev · 2023-04-11 12:58
marcanjannau: still the same crash, wasn't the dual pipe thing#asahi-dev · 2023-04-11 12:51
marcanso custom#asahi-dev · 2023-04-11 12:51
marcannah, needs to return true and write 0 to the argument#asahi-dev · 2023-04-11 12:41
marcanand that's right after the callback, ok#asahi-dev · 2023-04-11 12:40
marcanok so this is wrong, macOS gets disabled there#asahi-dev · 2023-04-11 12:40
marcan[ 0.454192] apple-dcp 389c00000.dcp: RTKit: syslog message: PPipeDCP_H13P.cpp:17276: Dual pipe enabled offset 1#asahi-dev · 2023-04-11 12:25
marcanyes#asahi-dev · 2023-04-11 12:23
marcanlots of stuff got inserted there#asahi-dev · 2023-04-11 12:23
marcanD113 is... get_tiling_state?#asahi-dev · 2023-04-11 12:22
marcansorry#asahi-dev · 2023-04-11 12:22
marcanI'm dumb#asahi-dev · 2023-04-11 12:22
marcanah wait no#asahi-dev · 2023-04-11 12:21
marcanand D112 is still set_idle_caching_state_ap