#
✕ nick:sven
#asahi-dev · 2021-12-12 17:33
svenj`ey: yeah, so?
#asahi-dev · 2021-12-12 17:33
sveni think https://github.com/AsahiLinux/linux/blob/fd7c0ced9b5e58028c66dde6cb54fa61a19283b9/drivers/soc/apple/rtkit.c is the latest rtkit code i have
#asahi-dev · 2021-12-12 17:33
svenuhh.... dunno?
#asahi-dev · 2021-12-12 17:32
svenwhat do you mean 'what happened'?
#asahi-dev · 2021-12-12 17:04
sventhat should be the correct domain
#asahi-dev · 2021-12-12 16:19
sveni think pwrstate is almost upstream
#asahi-dev · 2021-12-11 12:25
svenyup, i missed that :D
#asahi-dev · 2021-12-11 12:22
svenlet me check
#asahi-dev · 2021-12-11 12:22
svenhuh, did i miss that?
#asahi-dev · 2021-12-10 10:23
sventhere's also the entire gpu iommu node that's just used to create a small shim driver that just calls back to the main gpu driver
#asahi-dev · 2021-12-10 10:22
svenyeah, there are some horrible hacks in some drivers
#asahi-dev · 2021-12-10 10:16
svenlol
#asahi-dev · 2021-12-09 17:19
sven(except for all those HACKs ofc ;))
#asahi-dev · 2021-12-09 17:19
svenpovik: i briefly looked at it and left some general comments. i don't know anything about those subsystems but it mostly already looks good to me
#asahi-dev · 2021-12-09 16:26
svenit also looks like this nvme controller just doesn't support the abort command :/
#asahi-dev · 2021-12-09 16:21
svenhah :D
#asahi-dev · 2021-12-09 16:21
sven(guenter roeck pointed that out in my watchdog v1)
#asahi-dev · 2021-12-09 16:20
svenotherwise you could in theory disable the clock before the spi controller is removed
#asahi-dev · 2021-12-09 16:20
svenand this doesn't matter here because the refclock is constant but i think you should use devm_add_action_or_reset to clk_disable_unprepare the clock.
#asahi-dev · 2021-12-09 16:19
sventhere's also devm_spi_alloc_master
#asahi-dev · 2021-12-09 16:18
sveni think you want to get rid of the "retries++;"
#asahi-dev · 2021-12-09 16:16
svenmarcan: nice!
#asahi-dev · 2021-12-07 08:46
sveni can also include the commit that adds the watchdog node while i'm at it
#asahi-dev · 2021-12-07 08:45
svenmarcan: sure, i'll do it this evening
#asahi-dev · 2021-12-06 21:12
sveni've only seen DAPF used for DCP to map the firmware and prevent us from unmapping it
#asahi-dev · 2021-12-06 21:08
sveni briefly looked into SIO and the DART looked like every other DART
#asahi-dev · 2021-12-06 21:07
svenyeah, they are uninitialized garbage
#asahi-dev · 2021-12-06 21:07
svenbut for SIO you can just ignore it i think
#asahi-dev · 2021-12-06 21:07
svenDAPF
#asahi-dev · 2021-12-06 21:06
svennope
#asahi-dev · 2021-12-06 21:06
svensart is only for the nvme co-processor
#asahi-dev · 2021-12-06 21:06
sven:D
#asahi-dev · 2021-12-06 21:06
svendart-sio shouldn't be behind SART
#asahi-dev · 2021-12-06 20:57
sven:D
#asahi-dev · 2021-12-06 20:52
svenso some entries in the last level are weird?
#asahi-dev · 2021-12-06 20:52
svenok
#asahi-dev · 2021-12-06 20:50
svenis this the t6000 dart? i think there are also some darts with another compatible there but i thought that was only for thunderbolt
#asahi-dev · 2021-12-06 20:26
svenon the M1 max subpage protection is always on fwiw
#asahi-dev · 2021-12-06 20:26
svenpretty sure the VALID/VALID2 is from before I knew about subpage protection
#asahi-dev · 2021-12-06 20:15
svenBIT(0) or BIT(1) is "disable subpage protection". don't remember which one by heart
#asahi-dev · 2021-12-05 09:25
svenmaybe it’s time to reverse that power thread in the rtkit firmware and see what exactly happens when 0x6 and 0xb is sent.
#asahi-dev · 2021-12-05 08:18
svenhuh
#asahi-dev · 2021-12-05 07:49
svenno
#asahi-dev · 2021-12-05 07:43
svenSo dunno. I’d expect some way to reset them as well but I’m out of ideas at this point
#asahi-dev · 2021-12-05 07:42
svenit didnt work previously for me because I thought it also send 0xb…1
#asahi-dev · 2021-12-05 07:42
sveniboot does this 0xb…10 followed by 0x6…1 dance
#asahi-dev · 2021-12-04 13:37
svenso this reset bit is some kind of soft reset at best. Can’t even recover from a crash with it
#asahi-dev · 2021-12-03 18:37
svenand only then does the reset work :/
#asahi-dev · 2021-12-03 18:37
svenso it looks like we need to send 0xb0000000000010 followed by 0x60000000000001 to shut it down cleanly
#asahi-dev · 2021-12-03 18:37
svenhah, got it to reset successfully i think

plus ›