#
…

2022-12-03

00:00sventhere’s dart.ioread or something like that
00:01sventhat one just walks the pagetable to find the physical address and then just reads that
00:01chengsun_ has joined #asahi-re
00:02amarioguyrn i'm just trying to figure out how SEP knows where the gigalocker is because i'm not seeing any shared writes
00:02amarioguystrings on apple sep manager talk about "OOL buffers"
00:02chengsun has quit [Ping timeout: 480 seconds]
00:02amarioguyperhaps those have smth to do with it
00:03amarioguyor maybe i have this all wrong and AP just queries the locker on demand whenever it needs to decrypt it
00:22roxfan has quit [Remote host closed the connection]
00:24roxfan has joined #asahi-re
00:28amarioguyokay so the CRCs that XNU is saying the lockers have don't match up at all with the gigalocker CRCs...
00:48chengsun_ has quit [Ping timeout: 480 seconds]
01:19amarioguyalso apparently single user mode data accesses do not trigger SEP - likely because of how it's decrypted on the fly through the storage controller ephemeral key
01:26amarioguyLOL setting a breakpoint while the framebuffer comes on causes snow effects like old TVs
01:26amarioguy(or rather right after SEP changes the lock state)
01:36amarioguyokay so one of the absolute most verbose sep operations is pfk_data_unwrap and it's associated reply in the tracer
01:36amarioguy(yes i am just dumping what i find as i find it how did you know :) )
01:37* amarioguysure hopes i'm doing this mmio tracing thing right
01:49amarioguyhuh so during the unlock sequence after the first unlock, XNU logs the replies on the AP side with 0xffff in the upper two bytes
01:50amarioguythen it logs as "saving" the xart with a 2 byte crc
01:50amarioguyon the tracer side though, the actual mailbox message has that two byte CRC at the start
01:50amarioguy(the upper two bytes i mean)
01:53amarioguythink i'll want to look at the gigalocker file i have very closely tomorrow
01:54amarioguyto match up the xnu "crc" with the GL CRC
02:20Emantor_ has quit []
02:20Emantor has joined #asahi-re
02:38user982492 has joined #asahi-re
02:57ma has joined #asahi-re
03:03ma4 has quit [Ping timeout: 480 seconds]
03:37user982492 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
03:41mini0n has joined #asahi-re
04:01marcaneiln: also you may know this already, but ane1/ane3 are fused off on Max/Ultra devices
04:01marcanI don't know why, but iBoot filters them out
04:01marcanthey seem to be hard shut down
04:02marcanmy theory is it's either broken silicon or they decided not to implement it due to power envelope reasons
04:03marcanconcurrency *should* be used on M1 Ultra for the ANE across each die
04:03marcanmacOS has some kind of balancer/dispatcher kext for this
04:04marcanit was probably intended for M1 Max too, I don't think the reason they ditched the second ANE is because it wasn't ready in time, though it could be
04:06marcanand yes, geohot's "research" is, well... yeah what you said tracks
04:31mini0n has quit []
04:39nicolas17 has quit [Ping timeout: 480 seconds]
05:03eiln has joined #asahi-re
06:09SSJ_GZ has joined #asahi-re
06:39eilnohh I thought max/pros had ane1 enabled
06:39eilnso "device with two neural processor circuits" == just ultra
06:39eilnthat makes a lot more sense
06:39eilnmarketing speak says "32-core Neural Engine" for ultra
06:39eilnand "16-core Neural Engine" for all else
06:39eilnassuming core == ne*8, that checks out
06:39eilnseems they are pretty set on that decision
06:39eilnmy question is why they'd add more cores to ultra if 3 max models had issues with exactly that?
06:40eilni guess m2+ lines would give the definite answer
06:40eilnh14 firmware is definitely different tho if that signals anything
06:40eilnalso what exactly is meant by "iBoot filters them out"?
06:40eilnassociated clock/power node doesn't go through?
07:30mattgirv has quit [Server closed connection]
07:30mattgirv_ has joined #asahi-re
07:38eiln has quit [Quit: Page closed]
07:49arekm has quit [Server closed connection]
07:49arekm has joined #asahi-re
08:37n1c has quit [Server closed connection]
08:37n1c has joined #asahi-re
08:41robinp has joined #asahi-re
09:29nuup has quit [Server closed connection]
09:29nuup has joined #asahi-re
09:38svenIboot removes them from the template device tree iirc
09:39svenand I think marcan tried turning them on in pmgr and that just didn’t work
10:45chengsun has joined #asahi-re
10:52corion has joined #asahi-re
10:58vup has joined #asahi-re
11:18Dcow has quit [Ping timeout: 480 seconds]
11:22Dcow has joined #asahi-re
12:27corion has quit [Quit: Page closed]
12:52goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
13:20chadmed_ has joined #asahi-re
13:25yamii has joined #asahi-re
13:28goldsoultheory has joined #asahi-re
14:14marcaneiln: it's 16 cores per neural engine
14:14marcaneverything has one functional one per die, Ultra has two dies
14:14marcanultra is just two max dies
14:14marcanif ane0 works on max then ane2 will work on ultra since it's just two of the same thing
14:15marcanane1 is the broken one, which means also ane3 on ultra
14:15marcanpro does not have ane1 at all
14:15marcanpro only has one ane
14:15marcanso there is only one die, max, with two anes, and one is always disabled
14:15marcanhence my theory about silicon bug or power delivery issues
14:34goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
14:54chadmed_ has quit [Remote host closed the connection]
15:34goldsoultheory has joined #asahi-re
15:41goldsoultheory has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:14SSJ_GZ has quit [Ping timeout: 480 seconds]
16:18SSJ_GZ has joined #asahi-re
16:49corion has joined #asahi-re
16:52user982492 has joined #asahi-re
16:56corion has quit [Quit: Page closed]
17:01Dcow has quit [Remote host closed the connection]
17:02Dcow has joined #asahi-re
17:22os has quit [Server closed connection]
17:22os has joined #asahi-re
17:30paddatrapper_ has quit [Server closed connection]
17:30paddatrapper_ has joined #asahi-re
20:07nicolas17 has joined #asahi-re
20:57goldsoultheory has joined #asahi-re
21:30ChaosPrincess has quit [Quit: WeeChat 3.7.1]
21:32ChaosPrincess has joined #asahi-re
22:25Tom__ has quit [Read error: Connection reset by peer]
23:21SSJ_GZ has quit [Ping timeout: 480 seconds]
23:37Dcow has quit [Remote host closed the connection]
…