#asahi-dev · 2023-02-14 07:26
marcanI think I'm going to work on speakers again a bit today#asahi-dev · 2023-02-13 09:35
marcanand I might beat you to that :p#asahi-dev · 2023-02-13 09:35
marcanchadmed: yes, j314 was the plan :)#asahi-dev · 2023-02-13 08:19
marcan("the tablet chips", "the desktop chips")#asahi-dev · 2023-02-13 08:19
marcanwhich is neither numeric nor release order but made some sense to me at the time :p#asahi-dev · 2023-02-13 08:19
marcanre compatible order, yeah, I've ended up with t8xxx, t6xxx each individually numerically sorted#asahi-dev · 2023-02-13 08:17
marcanand after that probably spend time on Fedora Asahi :)#asahi-dev · 2023-02-13 08:17
marcankettenis: I think my plan is to wrap up speakers (at least on one model, but properly done - we can add variants while working on other stuff) and then wrap up whatever bringup isn't done yet#asahi-dev · 2023-02-12 13:30
marcanbus factor and all that#asahi-dev · 2023-02-12 13:30
marcanI can't be the only one who knows how to do this, you know :-)#asahi-dev · 2023-02-12 13:30
marcanshould I take it from this that y'all want another bringup stream tomorrow to fix most of this stuff?#asahi-dev · 2023-02-12 13:30
marcanpcie will need a new fuse table for sure#asahi-dev · 2023-02-11 09:34
marcanthe iboot interface is a bit special, but iboot itself also uses AFK/EPIC#asahi-dev · 2023-02-11 09:29
marcanbut yeah#asahi-dev · 2023-02-11 09:29
marcanI'm still hoping it works because iBoot needs enough to bring up DP on the laptops still, and the dp2hdmi stuff is controlled via EPIC and not dcpep as far as I know#asahi-dev · 2023-02-11 09:28
marcanugh.#asahi-dev · 2023-02-09 13:48
marcankettenis: nice! :D#asahi-dev · 2023-02-09 13:19
marcanif that is triggering IPIs we need to fix that#asahi-dev · 2023-02-09 13:19
marcanthe only thing that should see the IRQ bloat is the SPI host driver itself, since it has to IRQ every 8 bytes (current threshold)#asahi-dev · 2023-02-09 13:18
marcanjannau: no, this can't be the SPI-HID interface using threading because it (if it's not broken) can't be causing more than single-digit IRQs per update give or take#asahi-dev · 2023-02-09 13:16
marcanyeah that's not right#asahi-dev · 2023-02-09 13:07
marcanbut still, even with all that silliness, there's no way you add up to 10k without something being broken#asahi-dev · 2023-02-09 13:06
marcane.g. all the TX IRQs are completely unnecessary and burning power for zero reason#asahi-dev · 2023-02-09 13:06
marcanmaybe I should finally fix that next week#asahi-dev · 2023-02-09 13:06
marcanthere's also the fact that mailbox is stupid, of course#asahi-dev · 2023-02-09 13:06
marcanbeing generous and saying 10 mailbox messages, that's no more than 600 on top#asahi-dev · 2023-02-09 13:05
marcan60 display updates per *second*, no?#asahi-dev · 2023-02-09 13:02
marcanso I maintain there's a bug :)#asahi-dev · 2023-02-09 13:02
marcanjannau: everything else has no excuse to have more than ~1 IRQ per update on top of that, SPI is really the worst case bit#asahi-dev · 2023-02-09 12:52
marcan(unless you're lina and you use Rust and it takes care of that for you)#asahi-dev · 2023-02-09 12:52
marcanjeffmiw: which module? I'm not in the least surprised that there's bad OF refcounting somewhere, as far as I can tell ~everyone gets OF refcounting wrong, all the time. it's impossible to get right.#asahi-dev · 2023-02-09 12:49
marcanit should be on the order of 256 / 16 * update rate (let's say 100Hz) = ~1.6k IRQ/s#asahi-dev · 2023-02-09 12:49
marcanand SPI has a FIFO depth of 16 so that's a bug#asahi-dev · 2023-02-09 12:46
marcanpovik, jannau: 10k IRQ/s is either an IRQ for each byte of SPI, or a bug#asahi-dev · 2023-02-08 16:57
marcankh: As far as I know, yes, certainly without significant discussion first.#asahi-dev · 2023-02-08 03:23
marcanpovik: pushed the testtone/notebook stuff#asahi-dev · 2023-02-07 10:20
marcanthanks :)#asahi-dev · 2023-02-07 04:58
marcanyup, saw that :D#asahi-dev · 2023-02-06 14:24
marcanassuming it doesn't change like several times per second, it doesn't really matter if the safety daemon doesn't sync perfectly#asahi-dev · 2023-02-06 14:24
marcanor just communicate it via a control?#asahi-dev · 2023-02-06 14:20
marcanotherwise you'd fail if you try to change rates on the TX path too fast#asahi-dev · 2023-02-06 14:19
marcanso I think we need to make the capture PCM rate effectively able to change arbitrarily#asahi-dev · 2023-02-06 14:19
marcanI think we're always going to have races if we don't solve that in the kernel#asahi-dev · 2023-02-06 14:18
marcanahh#asahi-dev · 2023-02-06 14:17
marcanpovik: what's the current plan for syncing device usage between the tx/rx sides? I saw that speaker ready flag thing#asahi-dev · 2023-02-06 14:15
marcan:D#asahi-dev · 2023-02-06 14:14
marcanawesome :)#asahi-dev · 2023-02-06 14:01
marcanso maybe tmag increasing by less than 1/30 of the tvc delta is something we can just ignore#asahi-dev · 2023-02-06 14:01
marcanOTOH those ratios are like 2:60 for the tweeters and even higher for the woofers#asahi-dev · 2023-02-06 14:00
marcansince that's related to thermal mass