#asahi-dev · 2026-08-30 11:45
chadmedill fix up the dcp driver up tomorrow at some point#asahi-dev · 2026-08-30 11:44
chadmedeyyyyy lets gooo#asahi-dev · 2026-08-30 11:44
sofusyay#asahi-dev · 2026-08-30 11:43
sofusand that gets me a red :p#asahi-dev · 2026-08-30 11:43
sofusmetadata offset also needs +offest!#asahi-dev · 2026-08-30 11:17
sofusthat did not change anything, im gonna double check im not doing anything stupid from gstreamer#asahi-dev · 2026-08-30 11:13
JamesCalligeros[m]the only thing i can see that we dont set is unk_1 #asahi-dev · 2026-08-30 11:13
sofustile_{w,h}, tiles_{w,h} and tile_bytes all match, so i think its correct#asahi-dev · 2026-08-30 11:08
JamesCalligeros[m]actually no the UV is going to be 2 bytes isnt it#asahi-dev · 2026-08-30 11:01
JamesCalligeros[m]dont mind me it's getting late :P #asahi-dev · 2026-08-30 11:01
JamesCalligeros[m]oh yes#asahi-dev · 2026-08-30 11:01
sofusbit depth should be 8 so it should already be tile_dim * tile_dim right?#asahi-dev · 2026-08-30 11:00
JamesCalligeros[m]it looks like the tile_bytes field of the compression info is the same#asahi-dev · 2026-08-30 10:59
JamesCalligeros[m](youll need to guard that behind an if (yuv) otherwise RGB images will probably break #asahi-dev · 2026-08-30 10:59
JamesCalligeros[m]did you try making tile_size just tile_dim * tile_dim instead of tile_dim * tile_dim * bytes per pix?#asahi-dev · 2026-08-30 10:58
sofussurf->planes[i].offset (like the dump)#asahi-dev · 2026-08-30 10:57
JamesCalligeros[m]what did you make .data_offset? #asahi-dev · 2026-08-30 10:57
sofusstill the same DECOMP fail#asahi-dev · 2026-08-30 10:53
JamesCalligeros[m]oh yeah so we do, that wont work for planar formats lol#asahi-dev · 2026-08-30 10:53
sofusbut not for the second, we always set it to 0#asahi-dev · 2026-08-30 10:53
JamesCalligeros[m]that looks fine#asahi-dev · 2026-08-30 10:52
JamesCalligeros[m]it is for the first plane#asahi-dev · 2026-08-30 10:52
sofusdata offset is also not zero, not sure if thats a macos thing#asahi-dev · 2026-08-30 10:51
JamesCalligeros[m]the only thing i can see that is sussy is that tile_size is calculated without reference to the bit depth of the image here#asahi-dev · 2026-08-30 10:50
JamesCalligeros[m]ah yeah i see the stupid aliased fourcc #asahi-dev · 2026-08-30 10:50
sofusyes#asahi-dev · 2026-08-30 10:50
sofusthis was withouth your pr, i had something less verbose already#asahi-dev · 2026-08-30 10:49
JamesCalligeros[m]thats from macos? #asahi-dev · 2026-08-30 10:48
sofushttps://pastebin.com/U2HMdDmh#asahi-dev · 2026-08-30 09:58
jannauI've tagged asahi-7.1.12-1#asahi-dev · 2026-08-30 09:57
JamesCalligeros[m]but otherwise no there isnt#asahi-dev · 2026-08-30 09:57
JamesCalligeros[m]pull my m1n1 PR and it should be a little easier. it will (very verbosely) dump every swap request to the log file you pass to run_guest -l#asahi-dev · 2026-08-30 09:55
sofuscan i compare the IOSurface withouth m1n1? The dcp tracer is kinda hard to run#asahi-dev · 2026-08-30 09:54
sofusmaybe that was i long reach, gstreamer has the correct data#asahi-dev · 2026-08-30 09:52
JamesCalligeros[m]run linux under the hypervisor and try to dump the buffer maybe#asahi-dev · 2026-08-30 09:52
JamesCalligeros[m]hmm, have you put your source through macos and compared the IOSurface its generating with the one we're generating? #asahi-dev · 2026-08-30 09:52
sofusis there a way i can check the surface actually has the data i expect?#asahi-dev · 2026-08-30 09:51
sofusi dont get the crash anymore, but still getting the DECOMP fail#asahi-dev · 2026-08-30 09:51
sofusvery nice :)#asahi-dev · 2026-08-30 09:49
jannausofus: that fixes 1 sample (net, reports now 77/135 ass well)#asahi-dev · 2026-08-30 09:47
JamesCalligeros[m]Think Different#asahi-dev · 2026-08-30 09:47
JamesCalligeros[m]which is counterintuitive but whatever#asahi-dev · 2026-08-30 09:47
JamesCalligeros[m]yeah if that is getting set to 32 by macos then its based on the image/data tile size and not the compression subtiles #asahi-dev · 2026-08-30 09:46
sofusiirc i saw layout.meta_tile_bytes being sent as 32 so im gonna try that first#asahi-dev · 2026-08-30 09:46
sofusJamesCalligeros[m]: you are right, the calculation with layout.meta_tile_bytes=8 is the same as what i calculate in avd driver#asahi-dev · 2026-08-30 09:44
sofusjannau: would you mind testing with the write to AVD_V3_VP_INSN_FIFO_MASK being set to 0? (avd-hw.c)#asahi-dev · 2026-08-30 09:43
jannausofus: t8103 h.264 reports 76/135#asahi-dev · 2026-08-30 09:43
JamesCalligeros[m]but keep the DIV_ROUND_UP(roundup_pow_of_two(...)) stuff, just change the denominator of the division to tile_dim and bytes_per_tile dependent on the format and plane#asahi-dev · 2026-08-30 09:42
JamesCalligeros[m]sofus: can you try with the metadata buffer size calculated based on the number of image tiles, with 32 bytes per tile#asahi-dev · 2026-08-30 09:38
jannausofus: my fault, that system had still incomplete h.264 test samples