#asahi · 2024-05-03 11:05
PaulFertserPipewire#asahi · 2024-05-03 11:05
PaulFertserfrek: I guess that means that the kernel evdi out of tree driver increases the latency of the whole system beyond what's normally seen on these boards. Are you using wireplumber for audio?#asahi · 2024-05-03 11:04
PaulFertserfrek: does it affect not only the integrated sound card but also audio via bluetooth?#asahi · 2024-05-02 20:26
PaulFertserSo it should be relatively easy to at least put the blame.#asahi · 2024-05-02 20:26
PaulFertserProbably a bug in that evdi driver then. I guess we now have good facilities for finding latency bottlenecks even in the kernel land.#asahi · 2024-05-02 20:25
PaulFertserj`ey: yes but why and how would it be affecting sound?#asahi · 2024-05-02 17:24
PaulFertsernicolas17: and if not?#asahi · 2024-05-02 17:18
PaulFertsernicolas17: btw, what simple explanation would that be? Doesn't USB Audio class use isochronous transmissions with reserved bandwidth?#asahi · 2024-05-02 17:10
PaulFertserSo yes, that dock has sound card integrated but does frek really use it?#asahi · 2024-05-02 17:08
PaulFertsernicolas17: we have full dmesg luckily https://privatebin.net/?0fed0c57cc48f6cd#J1VZsBJKAAPyRRhijqNAXS9MVCUKCE4fauteqXDhCKC9#asahi · 2024-05-02 17:07
PaulFertserOK, so Performance it is.#asahi · 2024-05-02 17:07
PaulFertserfrek: https://github.com/AsahiLinux/docs/issues/21#issuecomment-1226845524#asahi · 2024-05-02 16:43
PaulFertserfrek: are those the power cores or not?#asahi · 2024-05-02 16:33
PaulFertserfrek: what about just moving mouse cursor around manually? It still leads to sound stuttering, right? And you say DisplayLinkManager consumes all of its core while at it.#asahi · 2024-05-02 16:30
PaulFertsertoots: that doesn't work after shutdown, just after power failure.#asahi · 2024-05-02 16:29
PaulFertsertoots: probably back up all NVRAM contents , then do that from macOS, then compare what it changed? https://www.reddit.com/r/osx/comments/w8vy7v/comment/ihspc3l/#asahi · 2024-05-02 16:27
PaulFertserfrek: so what kind of core is getting hogged by that process?#asahi · 2024-05-02 16:22
PaulFertser(not tested, just trying to show what I had in mind)#asahi · 2024-05-02 16:22
PaulFertserfrek: probably something like this in bash: while true; do alarcritty & sleep .2; kill $!; sleep .2; done#asahi · 2024-05-02 16:17
PaulFertserfrek: you should be able to see in /proc/cpuinfo which core is which#asahi · 2024-05-02 16:15
PaulFertserWith "taskset"#asahi · 2024-05-02 16:15
PaulFertserfrek: I wonder if pinning that process to one of "performance" cores makes a difference.#asahi · 2024-05-02 16:12
PaulFertserfrek: hm, probably just opening and closing another terminal window there in a loop to make it fully automated and reproducible?#asahi · 2024-05-02 16:11
PaulFertserfrek: also interesting test would probably be to use a fast SS storage in the same port and trying to write or read data there to see if it might affect sound the same way.#asahi · 2024-05-02 16:10
PaulFertserfrek: I think the most interesting usecase is running a regular sound player on M1 display and some simple animation on external display.#asahi · 2024-05-02 15:46
PaulFertserfrek: but it's just one core getting hogged?#asahi · 2024-05-02 15:30
PaulFertserfrek: I would start "top" (or other kind of top if you prefer to see what kind of CPU load (per-core) is getting high during movement of cursor on the external display. And what processes stand out.#asahi · 2024-05-02 15:28
PaulFertserfrek: ok, so it reports SS there. And evdi driver is used.#asahi · 2024-05-02 15:08
PaulFertserAnd what drivers get used etc.#asahi · 2024-05-02 15:08
PaulFertserfrek: so please paste all of dmesg as it's not visible what device you really have attached there.#asahi · 2024-05-02 14:44
PaulFertserfrek: how about not grepping and pasting full dmesg? ;)#asahi · 2024-05-02 14:33
PaulFertserfrek: ok, wanted to make it clear. So hm, what about HS vs SS, can you confirm it's working as SS?#asahi · 2024-05-02 14:28
PaulFertserfrek: affects audio that an application that's not showing anything on the external display is outputting?#asahi · 2024-05-02 14:27
PaulFertserAnd what is that driver you mention, is it just https://github.com/DisplayLink/evdi packaged?#asahi · 2024-05-02 14:22
PaulFertserfrek: is it shown to be connected as a Super Speed device in dmesg?#asahi · 2024-05-02 14:20
PaulFertserfrek: do you mean you move any application there and it affects audio playing elsewhere?#asahi · 2024-04-30 15:04
PaulFertserIt's not going to fix itself so power to your elbow :)#asahi · 2024-04-30 14:51
PaulFertserHolding Shift was good in LILO times.#asahi · 2024-04-30 14:50
PaulFertserAnd the regular way works.#asahi · 2024-04-30 14:50
PaulFertserAh, so GRUB is there just with hidden menu#asahi · 2024-04-30 14:50
PaulFertserMik_: please just wait for someone who actually runs Fedora on Apple Silicon to answer. There must be a trivial way to boot in text mode, I just do not know specific instructions.#asahi · 2024-04-30 14:49
PaulFertsertacitPanda: it'll be available as a boot option on startup#asahi · 2024-04-30 14:48
PaulFertserThe DFU remark was an answer to a specific question (a risk of bricking), not a suggestion to recover in that specific case.#asahi · 2024-04-30 14:48
PaulFertserbluetail: good thing in Mik_'s case nothing is unbootable.#asahi · 2024-04-30 14:47
PaulFertserMik_: isn't u-boot booting GRUB?#asahi · 2024-04-30 14:46
PaulFertserJust one messed up config is a strange excuse to reinstall the whole OS :)#asahi · 2024-04-30 14:45
PaulFertserAlso, why do not you just boot Fedora to text mode and modify /etc/sddm.conf back?#asahi · 2024-04-30 14:45
PaulFertserMik_: if you can still boot recovery and macOS there's no need for idevicerestore really.#asahi · 2024-04-25 11:46
PaulFertseriyes: even if it doesn't you can get a cheap USB sound card. Or use Bluetooth headset, if they're good enough for a smartphone then it should be OK for your VoIP too, right?#asahi · 2024-04-24 15:37
PaulFertserI wonder if tpw_rules intended it :)