#asahi-re · 2022-04-16 10:39
marcanheh, neat#asahi-re · 2022-04-16 10:39
rqou_(by samsung jpeg i mean one of the ones supported by this https://github.com/torvalds/linux/blob/master/drivers/media/platform/samsung/s5p-jpeg/jpeg-regs.h)#asahi-re · 2022-04-16 10:38
rqou_the EJPEG version looks closer to the AJPEG version, but idk what's different#asahi-re · 2022-04-16 10:38
rqou_AppleJPEGHal looks like a samsung jpeg block but iirc has a few extra registers#asahi-re · 2022-04-16 10:38
rqou_there's also a AppleEJPEGHal and AppleJPEGHal#asahi-re · 2022-04-16 10:37
rqou_the version in the m1 is controlled by AppleAJPEGHal#asahi-re · 2022-04-16 10:37
rqou_there's also at least three major versions of the jpeg block if you look inside the driver#asahi-re · 2022-04-16 10:37
marcanyup, that would make sense#asahi-re · 2022-04-16 10:35
rqou_before/after 0x1000#asahi-re · 2022-04-16 10:35
rqou_it would explain why the registers seem to come in two groups#asahi-re · 2022-04-16 10:35
rqou_and that color space conversion, tiling, agx compression, etc. are stuff apple wrapped around it#asahi-re · 2022-04-16 10:34
rqou_i actually assume that only the "DCT, entropy coding" parts come from shikino#asahi-re · 2022-04-16 10:32
marcanI love it#asahi-re · 2022-04-16 10:32
rqou_iirc "SHIKINO KJN-7GI 0001" if i remember correctly#asahi-re · 2022-04-16 10:31
marcanthat's amazing#asahi-re · 2022-04-16 10:31
marcanand it says shikino?#asahi-re · 2022-04-16 10:31
marcanHA#asahi-re · 2022-04-16 10:31
rqou_i brute-force read all the registers, and there's ascii text at that address#asahi-re · 2022-04-16 10:30
marcanI'm assuming then these cores are https://www.shikino.co.jp/eng/products/product-ip_core_products_1.php#asahi-re · 2022-04-16 10:30
marcancurious, where'd that come from?#asahi-re · 2022-04-16 10:30
marcanrqou_: > SHIKINO_VERSION_MAGIC0#asahi-re · 2022-04-15 08:09
chadmed_expected for the polling/keep alive messages (which ive also figured out) but idk why the wake and sleep messages repeat so much #asahi-re · 2022-04-15 08:09
chadmed_theres a lot of weird repitition though#asahi-re · 2022-04-15 08:08
chadmed_i think ive figured out the on, wake and sleep messages#asahi-re · 2022-04-15 07:47
jannauit's possible that I only map the iova for the keyboard#asahi-re · 2022-04-15 07:47
jannauI only have a rudimentary understanding of the SIO protocol and I stopped as soon as it was good enough for the keyboard#asahi-re · 2022-04-15 07:41
chadmed_ive pretty reliably worked out what it does during power on but the data it spits out after the OS has woken it up is garbage#asahi-re · 2022-04-15 07:39
chadmed_fixing the tracer to see what exactly _it_ is a bit beyond my skills though#asahi-re · 2022-04-15 07:39
chadmed_it seems to do something with this during calibration#asahi-re · 2022-04-15 07:38
Jamie[m]1132x112 at 12 bits per pixel#asahi-re · 2022-04-15 07:38
Jamie[m]122K is the exact size of the fingerprint sensor image i’ve reversed before#asahi-re · 2022-04-15 07:10
chadmed_0x19 also does something similar when macos is trying to read your fingerprint but it blows up before the tracer can even get the message out of it#asahi-re · 2022-04-15 07:10
chadmed_0x18* sorry#asahi-re · 2022-04-15 07:09
chadmed_0x19 also sends an enormous 22kB message which blows up the tracer if you try looking at it#asahi-re · 2022-04-15 07:08
chadmed_then after that its just a bunch of variable length messages#asahi-re · 2022-04-15 07:08
chadmed_it has a power on sequence where it gets kicked with the same 7 byte message until it replies with a specific 23 byte message#asahi-re · 2022-04-15 07:07
chadmed_mesa is capable of a few message formats it seems#asahi-re · 2022-04-14 05:04
AdministratorFREENODE AND LIBERA MOVE TO irc.worldhacker.org is open for public #0day , [ 0day (xc) Our ] WhiteHat Hacker Team And Dev Official Public Channel : https://worldhacker.org , REQUEST PROGRAM , SOURCE CODE , OR ANYTHING HERE ... . NEED STORAGE OR VPS ??? Free #asahi-re · 2022-04-13 06:32
marcanyeah#asahi-re · 2022-04-13 06:25
svenproxyclient is deliberately full of hacks but at least merging the dart variants into a single file should be doable :D#asahi-re · 2022-04-13 06:24
svenmight make sense to integrate it into dart.py#asahi-re · 2022-04-13 06:18
rqou_especially since the original dart code wasn't particularly well parameterized over device type#asahi-re · 2022-04-13 06:16
rqou_also i ended up removing the L0 logic because t8110 obviously didn't have it, and afterwards i was lazy and didn't want to figure out the appropriate place to stick if statements to unify the code#asahi-re · 2022-04-13 06:14
rqou_i then realized it didn't, at least not in the t600x#asahi-re · 2022-04-13 06:14
rqou_i duplicated the code before i was sure about the page table format (and suspected it had additional page table levels)#asahi-re · 2022-04-13 06:14
rqou_reason? yes. good reason? probably not#asahi-re · 2022-04-13 06:10
jannaurqou_: is there a reason why it's no integrated into dart.py? it probably uses the same pte format as dart-t6000 are #asahi-re · 2022-04-13 05:58
rqou_e.g. that's how i figured out that bit1 of the PTE is probably "uncacheable"#asahi-re · 2022-04-13 05:57
rqou_the code itself is basically all m1n1 code, but i got a bunch of register names and bit positions by dumping the DWARF data from the KDK#asahi-re · 2022-04-13 05:55
rqou_feel free to yell at me for doing a huge copypasta