#
#asahi-dev · 2024-08-11 09:35
cy8aerOk, the new one is better. Needed to change the room. And the basses are better too.
#asahi-dev · 2024-08-11 09:34
chadmedthe IRs are loaded into pipewire's memory when it parses the filter chain and the only way to change them is reload the entire filter chain (restart pipewire)
#asahi-dev · 2024-08-11 09:34
chadmedunfortunately you cant do that
#asahi-dev · 2024-08-11 09:20
cy8aer(?? not sure anymore) for having a good A-B it would be nice to have a script which live changes the necessary parameters
#asahi-dev · 2024-08-11 09:18
cy8aerbut somehow less bass (ghost in the shell drums)
#asahi-dev · 2024-08-11 09:13
cy8aerFor me it is more 3d room in the new one (tested with: https://yt.drgnz.club/watch?v=s_VH36ChGXw and https://invidious.incogniweb.net/watch?v=H3v9unphfi0). But my second subjective tester had doubts and A-B failed once 😕
#asahi-dev · 2024-08-11 08:59
chaos_princess(though i still like the headphone sound better :P)
#asahi-dev · 2024-08-11 08:59
chaos_princessto my very much not audiophile-quality ears - it is better than before.
#asahi-dev · 2024-08-11 01:17
chadmedim pretty sure ive finally fixed the cruddy sound
#asahi-dev · 2024-08-11 01:17
chadmedcan someone with a 14" mbp do me a favour and give feedback on asahi-audio/main
#asahi-dev · 2024-08-10 21:25
amarioguyso yea back to figuring out how to get this working properly :/
#asahi-dev · 2024-08-10 21:24
amarioguynot even a 32 bit read works lol
#asahi-dev · 2024-08-10 21:21
amarioguylooks like smth is wrong with PCI controller startup after all
#asahi-dev · 2024-08-10 21:21
amarioguynope
#asahi-dev · 2024-08-10 21:12
svencan you read any bar at all without an serror?
#asahi-dev · 2024-08-10 21:11
amarioguyi've not had problems with ECAM space so far fwiw, mainly just writing to/reading from BAR registers
#asahi-dev · 2024-08-10 21:11
amarioguyjannau: doesn't seem like it? at least edk2 can enumerate devices on the bus and ECAM space can be written to, but i'm not 100% sure if this means that it isn't power/clock gated in some way
#asahi-dev · 2024-08-10 21:10
amarioguyi wouldn't be surprised honestly
#asahi-dev · 2024-08-10 21:00
svenyeah, this very much sounds like something‘s wrong with the pci controller itself to me
#asahi-dev · 2024-08-10 20:58
jannauis pcie clock or power-gated?
#asahi-dev · 2024-08-10 20:50
amarioguyi adapted the u-boot asmedia xhci code for what i'm doing, 16 bit write and all but who knows if i missed smth
#asahi-dev · 2024-08-10 20:48
amarioguyi'm going to try that actually
#asahi-dev · 2024-08-10 20:47
svenbut I guess 32bit wide access works for you? I.e. it’s not just a broken pci init sequence?
#asahi-dev · 2024-08-10 20:47
amarioguyhmm odd, i've been combing u-boot traces/source to see what you're doing that isn't working on my end, not sure why the 16-bit write to the same register isn't being accepted in my case
#asahi-dev · 2024-08-10 20:46
svenno idea why you get an serror though
#asahi-dev · 2024-08-10 20:45
svenpci isn’t *that* broken, iirc 8/16bit wide access has always worked there
#asahi-dev · 2024-08-10 20:45
amarioguyprime example being writing ASMT_REG_ADDR during fw upload
#asahi-dev · 2024-08-10 20:44
amarioguyat least to those registers
#asahi-dev · 2024-08-10 20:44
amarioguyindicating i think that the bus only allows 32 bit writes per the asahi docs
#asahi-dev · 2024-08-10 20:44
amarioguyhmm how is u-boot able to do 16bit/8bit writes to some PCI BAR registers on Apple chips? trying those same writes myself results in an SError with L2C_ERR_INF=0x2
#asahi-dev · 2024-08-10 20:31
lily10kdocumentation note: it would be helpful to link to https://github.com/AsahiLinux/docs/wiki/Tethered-Boot-Setup-%28For-Developers%29 from https://github.com/AsahiLinux/docs/wiki/m1n1:User-Guide#proxy-mode
#asahi-dev · 2024-08-09 14:48
sventhough i don't remember if it was usb3 that broke that
#asahi-dev · 2024-08-09 14:48
svenusb 2 also seems to work with devices plugged in at boot already
#asahi-dev · 2024-08-09 14:47
sveni just started over and copy/pasted stuff :D
#asahi-dev · 2024-08-09 14:47
jannaunice. I assumed you were working on the bits branch(es)
#asahi-dev · 2024-08-09 14:46
svennext step is took hook usb 3 back up. hopefully without that annoying synchronization this time
#asahi-dev · 2024-08-09 14:45
svenjannau: that's my current tree https://github.com/AsahiLinux/linux/tree/20240806-atcphy-rewrite
#asahi-dev · 2024-08-09 14:39
svenhttps://paste.debian.net/1325943/ that's what i have so far
#asahi-dev · 2024-08-09 14:35
jannaunice. did you already undo the 2 dwc3 reverts?
#asahi-dev · 2024-08-09 14:33
svenlooks like usb 2 works again with this rewrite. let's see about usb3 now
#asahi-dev · 2024-08-09 14:29
jannau6.10 has cleanup support for of_node
#asahi-dev · 2024-08-09 14:04
svensweet, linux has supported stuff like guard(mutex)(&atcphy->lock); with linux/cleanup.h for a while now which removes quite a bit of if (ret) goto fail_whatever; boilerplate
#asahi-dev · 2024-08-09 13:46
jannauI I I aslo think that we want/need to map the whole memory there
#asahi-dev · 2024-08-09 13:43
jannauI strongly suspect we need to abstract the size away to ensure that have a consistent execution environment for macos asfter chaninloading and in the hypervisor
#asahi-dev · 2024-08-09 13:33
emxlasking because i dont quite get the point of using mem_size_actual over ram_size when the other mappings encompassing the entire memory already use "ram_size"
#asahi-dev · 2024-08-09 13:32
emxloh ok, is changing the mapping to be "ram_size" large an acceptable fix or are you planning on abstracting the boot_args struct somehow and still using mem_size_actual for this?
#asahi-dev · 2024-08-09 13:27
jannauemxl: apple changed the size of boot args. mem_size_actual is quite bit further down. iirc ~200-300 bytes. Ican check on a different computer. I haven't come up with a good way to handle this
#asahi-dev · 2024-08-09 12:43
emxlHey, I was wondering why m1n1 uses "mem_size_actual" here (https://github.com/AsahiLinux/m1n1/blob/main/src/memory.c#L414) instead of ram_size? "mem_size_actual" is 0 on my system (t6030).
#asahi-dev · 2024-08-09 10:18
chadmed_finally got REW to play nice with a sane version of openjdk
#asahi-dev · 2024-08-07 12:21
chadmedyeah i think we can ship this. doesnt seem possible to melt anything with the current behaviour

plus ›