#asahi · 2021-07-29 07:04
svendart offset + 0x4000 or so#asahi · 2021-07-29 07:04
svenmarcan: just after the DART#asahi · 2021-07-29 07:02
svenit'll just send a crashlog and continue#asahi · 2021-07-29 07:02
svenwith 0x40 it even survives#asahi · 2021-07-29 07:02
sventhat's "force crash"#asahi · 2021-07-29 07:02
svenmarcan: send 0x22 or 0x40 to the crashlog ep#asahi · 2021-07-29 06:40
sven(at least for the system endpoints)#asahi · 2021-07-29 06:39
sveni think if you don't start some it'll hang#asahi · 2021-07-29 06:36
svenand 3 is this kdebug thing#asahi · 2021-07-29 06:36
sveni think crash is 1#asahi · 2021-07-29 06:35
svenmarcan: i think at least 3 is wrong there fwiw#asahi · 2021-07-29 06:26
sven(might be larger, but who knows)#asahi · 2021-07-29 06:26
svenmarcan: 31, 24#asahi · 2021-07-29 06:25
sven0x80 + 0x20 = 0xa0#asahi · 2021-07-29 06:25
sven+ the header which is always 0x20#asahi · 2021-07-29 06:25
svenand the 0x80 is the msg size#asahi · 2021-07-29 06:24
svenmarcan :further down#asahi · 2021-07-29 06:23
sven#define APPLE_RTKIT_SYSLOG_MSG_SIZE GENMASK(31, 24)#asahi · 2021-07-29 06:23
sven#define APPLE_RTKIT_SYSLOG_N_ENTRIES GENMASK(7, 0)#asahi · 2021-07-29 06:22
svendoesn't matter for DCP unless they change it ;)#asahi · 2021-07-29 06:21
svenbut it's different for other asc#asahi · 2021-07-29 06:21
svenit's hardcoded for dcp#asahi · 2021-07-29 06:20
svenit used to be constant 0xa0 in there#asahi · 2021-07-29 06:20
svenoh, is it? but then you fixed it at some point#asahi · 2021-07-29 06:19
svenhttps://github.com/AsahiLinux/linux/blob/asc/dev/drivers/mailbox/apple-asc-mailbox.c#L419#asahi · 2021-07-29 06:19
svenit's not just constant 0xa0 or whatever#asahi · 2021-07-29 06:18
sventhe init message sends a max number of entries and the length of the syslog messages#asahi · 2021-07-29 06:18
svenand i think it'll be fine for dcp probably, but syslog is a little bit more complex#asahi · 2021-07-29 06:17
svenalso, please double check the endpoint id -> class assignment. iirc they were wrong at some point and i'm not sure i fixed them#asahi · 2021-07-29 06:16
svenkdebug is more fun i think. it should contain tracepoints somehow#asahi · 2021-07-29 06:15
svenjust that get buffer request is the same#asahi · 2021-07-29 06:15
svenno#asahi · 2021-07-29 06:14
svensomething like that, yes#asahi · 2021-07-29 06:14
svenit's called Log just because it also needs a buffer#asahi · 2021-07-29 06:14
svenmarcan: no, it isn't. that's where the hack part comes in ;)#asahi · 2021-07-28 20:30
svenvery helpful :P#asahi · 2021-07-28 20:27
svenfirst i need to clean up the quirks/fixes required to the nvme core. and then i'll have to figure out with arnd how to submit this#asahi · 2021-07-28 20:26
svenit's still a hack. and getting this upstream will take a while :)#asahi · 2021-07-28 20:06
svenaaaand nvme is back. this time with less hacks :)#asahi · 2021-07-28 20:06
sven[ 0.421379] nvme0n1: p1 p2 p3 p4 p5 p6#asahi · 2021-07-28 20:06
sven[ 0.402106] apple-mailbox 277400000.mbox: syslog message: cmd.c:5414: boot mode normal#asahi · 2021-07-28 19:26
svenfrom what i can tell ANS receives messages on endpoint 0x20 but replies to them using a different mechanism on endpoint 0x21#asahi · 2021-07-28 19:25
svenmarcan: i think you may be right with that iop communication theory#asahi · 2021-07-28 17:45
sveni've only managed to crash it by doing what mac os calls "sending a NMI" (makes sense :D), by sending invalid commands (who wouldn't panic there?!) and by telling to switch to a power state it's already in#asahi · 2021-07-28 17:44
svenunless you actively try to crash it it's kinda hard#asahi · 2021-07-28 17:43
svenit's not great :D#asahi · 2021-07-28 17:43
svenyeah... i've done that to ANS#asahi · 2021-07-28 17:43
sven(or maybe there's a bit somewhere to disable resetting them, who knows)#asahi · 2021-07-28 17:43
sveni think that's just leftovers as well