#asahi-dev · 2022-06-21 10:55
marcan:p#asahi-dev · 2022-06-21 10:53
marcanmaz: "under control" is defined as "325 messages but at least I have some categories to make sense of them"#asahi-dev · 2022-06-21 10:43
marcanI *think* I managed to get my inbox mostly under control... mostly...#asahi-dev · 2022-06-20 15:28
marcan(punting to tomorrow since it got late, but I'll spend the day on that, so expect new packages then)#asahi-dev · 2022-06-20 07:14
marcanjannau: I want to get it done today ideally#asahi-dev · 2022-06-19 12:11
marcankeep with the format, avoid hardcoding nonsense in the driver#asahi-dev · 2022-06-19 12:11
marcanwe should just have the renamer do s/MUR/m/ s/USI/u/#asahi-dev · 2022-06-19 12:11
marcanit's the module vendor#asahi-dev · 2022-06-19 12:11
marcanu = no idea but probably starts with USI#asahi-dev · 2022-06-19 12:10
marcanm = murata#asahi-dev · 2022-06-19 12:10
marcansven: V=m = MUR, V=u = USI#asahi-dev · 2022-06-19 12:09
marcansven: yes, that's the wifi otp format#asahi-dev · 2022-06-13 07:21
marcanthat will go out with the pushed packages#asahi-dev · 2022-06-13 07:20
marcanof course#asahi-dev · 2022-06-13 07:19
marcanyeah, I guess that's probably the sanest option#asahi-dev · 2022-06-13 07:18
marcan(or just tell the few people affected to manually download the packages)#asahi-dev · 2022-06-13 07:17
marcanmaybe there is some horrible hack, like using one of the SMC sideband mailboxes meant for other ASCs as long as they don't panic on the unexpected reply?#asahi-dev · 2022-06-13 07:17
marcanwonder if there's any good way to fix that though, short of hacking together an SMC client in m1n1 to flip that GPIO on :/#asahi-dev · 2022-06-13 07:15
marcanjannau: isn't pcie going to be a problem for people who want to update without ethernet?#asahi-dev · 2022-06-13 07:14
marcanjannau: merged and pushed :)#asahi-dev · 2022-06-13 07:13
marcanlet me do it now before I forget#asahi-dev · 2022-06-13 07:13
marcanah yes, I was going to do that#asahi-dev · 2022-06-13 02:25
marcankettenis: it's more correct on all hardware, it's just Apple's firmware is specifically buggy with the existing brute force version :)#asahi-dev · 2022-06-12 11:04
marcanack#asahi-dev · 2022-06-12 08:44
marcaneven if it doesn't right now#asahi-dev · 2022-06-12 08:43
marcaneither way I know for sure not having that patch can cause breakage#asahi-dev · 2022-06-12 08:43
marcanor maybe I just saw it with m1n1 experiments?#asahi-dev · 2022-06-12 08:42
marcanjannau: I don't remember which boot stage that was breaking but it was definitely breaking one of them, so yes (even if it doesn't currently break u-boot)#asahi-dev · 2022-06-11 16:50
marcankettenis: I think any one of those failing would bail anyway? but sure, if some codepaths could have that 100ms is fine#asahi-dev · 2022-06-11 16:08
marcan:)#asahi-dev · 2022-06-11 16:06
marcanhaving them at 10ms serves no useful purpose and is just a dumb way for it to break in the future#asahi-dev · 2022-06-11 16:06
marcanjannau: also while you're at it can you bump all the timeouts across the board to at least 100ms, if not 1s?#asahi-dev · 2022-06-11 16:05
marcanah, if it's just sart then sven, yeah#asahi-dev · 2022-06-11 16:03
marcanif git blame says it's just folks from around here I'm sure relicensing won't be an issue#asahi-dev · 2022-06-11 16:03
marcansame, not sure if anyone other than us touched rtkit?#asahi-dev · 2022-06-11 10:17
marcanI agree#asahi-dev · 2022-06-11 10:03
marcannice#asahi-dev · 2022-06-11 06:02
marcannice :)#asahi-dev · 2022-06-08 18:14
marcanthat's... surprising#asahi-dev · 2022-06-07 14:08
marcangood to hear that!#asahi-dev · 2022-06-07 13:31
marcanor more...#asahi-dev · 2022-06-07 12:15
marcantrue but... :p#asahi-dev · 2022-06-07 12:15
marcanu-boot should just handle and ack syslog messages (doesn't even need to read the ring buffer) and check for it on every nvme command just in case; it's just a single bit to poll, won't hurt performance.#asahi-dev · 2022-06-07 12:15
marcanwas never a fan of that one either#asahi-dev · 2022-06-07 12:15
marcanthat too...#asahi-dev · 2022-06-07 12:15
marcanlast time we were hitting the one in the main pwrstate wait; this time I bet it's one of the earlier ones due to it taking longer than 10ms to boot/proceed#asahi-dev · 2022-06-07 12:14
marcanre nvme#asahi-dev · 2022-06-07 12:14
marcankettenis: I suspect some of the timeouts are still too low, but I also have reservations about the linear rtkit init code (as opposed to having one big state machine/reply switch that supports more paths)#asahi-dev · 2022-06-06 19:09
marcanthe kind of weirdness I saw without that didn't match what I'd expect with unreliable i2c, so...#asahi-dev · 2022-06-06 07:50
marcanthank you for figuring that out, I'm glad I didn't submit it as-is then :)