06:21 <travmurav[m]> jlinton: interesting... AFAICT after a quick look, btmgmt public-addr should read BE ("normal order") and converts it into LE bdaddr_t internally. Then it's passed on unchanged until it's sent to qca with qca_set_bdaddr_rome(). Notably on OLD qca based bt controller, mac address is set correctly. (mostly the same code path is used afaict)

06:25 <travmurav[m]> which means we can't just flip it in the kernel before sending - this will break other things

08:11 <Jasper[m]> I think I've asked this before, but does anyone else have issues with roaming and connecting to other ap's with the x13s?

08:11 <Jasper[m]> My range is terrible while I'm fine on my phone

08:12 <Jasper[m]> dmesg is also full of authenticate associate deauthenticate lines

11:15 <neobrain> Looks like my x13s can't charge from a usb-c dock. Is that expected?

11:49 <exeat> My x13s is usually plugged into an "Anker 543 USB-C Hub (6-in-1, Slim)" and charging seems to work as expected

11:58 <exeat> IIRC there were a couple of early (1.2x) bios versions with weird usb issues but it's been plain sailing ever since (though no >1G ethernet or external displays here)

12:09 <HdkR> My X13s is also connected to a dock and charges correctly

12:09 <Jasper[m]> Mine too, a cable matters dock

12:09 <HdkR> When you plug in the power does the power LED blink three times without dock and with dock?

12:11 <Jasper[m]> Afaik yes, will test it again in a sec

12:12 <javierm> jhovold, ajhalaney[m]: most mutters devs are already on PTO for the holiday season, but I'll ping them again in early Jan

12:13 <javierm> another option is to ask your distros to backport Jonas' patches to the stable released packages

12:16 <Jasper[m]> @dgilmore I have another addition suggestion for the x13s package. Make wifi mac static aswell. It switches around too apprently

12:22 <neobrain> hmm guess it must be this weird dell dock then

12:24 <neobrain> actually no, my anker 6-in-1 hub doesn't charge it either, weird

12:28 <HdkR> pd-mapper running?

12:29 <HdkR> If it charges when shutdown which you can see if the power LED blinks three times when plugging it then it's a userspace or kernel problem

12:47 <Jasper[m]> <Jasper[m]> "@dgilmore I have another..." <- I had to delete like 12 DHCP ip reservations from my router with its hostname hahahaha

12:47 <Jasper[m]> Solved it with a udev rule: `ACTION=="add", SUBSYSTEM=="net|pci", KERNELS=="0006:01:00.0", RUN+="/usr/sbin/ip link set dev $name address AA:BB:CC:DD:EE:FF`

12:47 <Jasper[m]> * address AA:BB:CC:DD:EE:FF"`

12:48 <neobrain> HdkR: yeah, charging directly with a usb-c cable works fine

12:49 <neobrain> will look at shutdown behavior later, maybe it's actually charging but just not telling userspace about it

12:50 <Jasper[m]> <Jasper[m]> "My range is terrible while I'm..." <- Okay I checked, it keeps connecting to the 5GHz network of the sole wifi 6 capable AP on the other side of the house. Is there a way to prefer range instead or is it like that for everyone?

12:56 <jhovold> javierm: thanks, it'd be best if upstream could do the backport, but asking the distros may be an option if for some reason that doesn't happen

12:58 <jhovold> Jasper[m]: a semi random MAC is better than a static non-unique MAC in a distro

13:00 <Jasper[m]> jhovold: These change on reboot right?

13:00 <jhovold> yeah

13:00 <jhovold> I think so anyway

13:00 <Jasper[m]> Ah okay that's not too bad then

13:01 <Jasper[m]> I stuck it to something so I don't have weird issues, but yeah

13:01 <jhovold> makes perfect sense, but the user needs to provide a unique MAC (preferrably the one assigned to the device)

13:02 <jhovold> until we've figured out how to do that from Linux

13:02 <jhovold> s/do that/read out the assigned MAC/

13:03 <jhovold> neobrain: I've head of at least one dock that fails to charge (an Apple one), but many seems to work

13:05 <jhovold> the exact reason is yet to be determined, but an hypothesis is that it's related to role-switching not yet being supported

13:05 <Jasper[m]> Biggest problem I currently have with wifi is that it refuses to connect to the ap with the best signal

13:05 <Jasper[m]> Which is quite annoying since it fails to connect to the one it tries to connect to aswell

13:08 <Jasper[m]> A different (Mediatek) nic works fine, don't think it's a configuration issue

13:11 <danielt> Interestingly the Fedora folks are talking about defaulting to stable-but-randomly-assigned MACs everywhere (e.g. each SSID will see your machine with a different MAC): https://fedoraproject.org/wiki/Changes/WifiMACRandomization

13:11 <danielt> It's not intended to workaround unassigned hardware MACs but will solve unstable DHCP addresses anyway·

13:15 <Jasper[m]> Would at least solve my problem

13:15 <jhovold> interesting indeed

13:16 <konradybcio> that's the default on android nowadays, i think

13:16 <konradybcio> and ios

13:16 <Jasper[m]> Yes

13:28 <neobrain> jhovold: ah, interesting

13:35 <neobrain> of course, I can't actually reproduce it anymore now. reboot magically fixed it, though I'd definitely seen this other times before as well

13:37 <jhovold> heh, ok, good to hear still I guess :)

13:39 <neobrain> What I still see is that Plasma's systray only reacts to plug-in/out events like half of the time, hence not updating the "charging" indicator. But if you open the associated menu, it clearly does know whether it's charging or not. So it's a bit messy apparently, but at least not outright broken :D

14:57 <Jasper[m]> @HdkR am I correct in assuming that installing packages in the respective fex-emu rootfs is impossible? You should be able to install it directly from rpmfusion in a fedora install

15:00 <Jasper[m]> unbreak_chroot does not seem to work with the fexrootfsfetcher install of fedora 38

15:03 <HdkR> Jasper[m]: Yea, unbreak_chroot.sh only works with the ubuntu rootfs on an ubuntu/debian system due to multi-arch shenanigans. I haven't solved that for Fedora or Arch rootfs images

15:03 <HdkR> Would need to chroot in with an x86 host and copy the changes over

15:03 <Jasper[m]> rip

15:03 <Jasper[m]> Well, I'll download steam by hand then from rpmfusion instead of tencent hahaha

15:36 <Jasper[m]> <HdkR> "Would need to chroot in with..." <- Any tips on getting a game running? There's the list obviously, but I'd love to test some stuff

15:39 <HdkR> Jasper[m]: I usually just launch games through Steam

15:41 <Jasper[m]> HdkR: Well I tried that and I either got greeted by nothing or a black screen

15:41 <Jasper[m]> Assuming it crashed along the way

15:41 <Jasper[m]> It's ULTRAKILL (awesome game btw)

15:44 <HdkR> Huh, Steam black screening or the game?

15:45 <Jasper[m]> The game

15:45 <Jasper[m]> It did crash on its own now instead of hanging, terminal is full of stuff now

15:47 <HdkR> I can't remember, do you have TSO emulation enabled in the config?

15:47 <Jasper[m]> HdkR: `TSO Enabled` is checked

15:47 <Jasper[m]> It's on x13s so I'm assuming it's emulation then?

15:47 <HdkR> Yea, it's emulation

15:48 <HdkR> Could be that Unity game is abusing something that we aren't expecting

15:48 <Jasper[m]> Want some logs from somewhere?

15:49 <HdkR> I'm guessing the logs would be a backtrace in the terminal?

15:50 <Jasper[m]> HdkR: Well it's a bunch of ld.so related errors, but that's about the overlay

15:50 <HdkR> Oh, that one is benign

15:51 <Jasper[m]> I'll throw it up somewhere, do you want me to open an issue or just dm you here?

15:51 <HdkR> That's just glibc complaining about stuff that doesn't matter

15:51 <HdkR> You can throw me a gist or create an issue

15:52 <HdkR> Unity games are fairly spicy, the garbage collector can make the JIT upset right now

15:58 <HdkR> There are some plans to make that more stable but I haven't had a chance to get back to it. Deferring signals from the kernel is a bit spicy

16:01 <Jasper[m]> HdkR: Oh yeah that's fair. It seemed like something that would probably run on this hw so it's not really that urgent

16:01 <Jasper[m]> What did wonder if streaming could work in some way

16:02 <Jasper[m]> Without ripping the pi steam link package lmfao

16:02 <HdkR> Last I tested the Steam Link app worked fine, although it relies on CPU decoding of course

16:02 <HdkR> Obviously wouldn't be wired up to use v4l2

16:03 <Jasper[m]> <HdkR> "Could be that Unity game is..." <- Got a tip for downloading based on an appid, Steam doesn't allow downloading demo's from the interface if you bought it.

16:04 <Jasper[m]> s/,/?/

16:04 <HdkR> No idea on that one. I'm just a normal Steam user like anyone else :P

16:04 <travmurav[m]> perhaps can download the depot via steam console?

16:05 <Jasper[m]> HdkR: Yet? :p

16:05 <Jasper[m]> travmurav[m]: A workaround would be to use "open" and use it as a link but that's not working I think

16:13 <Jasper[m]> I am somehow having trouble shutting it down to restart with console enabled

16:37 <Jasper[m]> @HdkR guessing you need host mesa version for the video driver?

16:39 <HdkR> Without messing with thunk configs then it will be using the mesa version shipped in the rootfs. Which will be 23.3.0

17:07 <Jasper[m]> <HdkR> "Without messing with thunk..." <- https://github.com/FEX-Emu/FEX/issues/3347

17:14 <HdkR> Very interesting that it didn't throw error messages, maybe the game redirects stderr,stdout

17:15 <Jasper[m]> Could be yeah

17:15 <HdkR> I can probably test that soon-ish though and maybe get some more information

17:16 <Jasper[m]> If you need anything from me in the meantime, feel free to @ me here or on the issue

17:16 <Jasper[m]> or dm

17:17 <Jasper[m]> Could even try tweeting at the gamedev/publisher, I can see them answering there

17:17 <Jasper[m]> They're pretty good

17:18 <HdkR> Pestering gamedevs about a WIP emulator isn't quite nice, especially when we know that we have Unity problems :P

17:19 <Jasper[m]> Oh yeah sure, though I know they appreciate getting that stuff to run on anything.

17:19 <Jasper[m]> They were quick with steamdeck and proton support too

17:20 <Jasper[m]> Can see them sending over some debug build maybe, who knows

18:28 <Jasper[m]> Funny, Dusk does actually seem to work. Also a Unity game

18:29 <Jasper[m]> Works pretty well too, apart from the visual corruption I see in general

18:34 <HdkR> Yea, it heavily depends on the game

20:36 <neobrain> Jasper[m]: steamcmd can download games by appid, not sure it would work on demos in your case though

20:37 <Jasper[m]> neobrain: I got it working with the console, downloading worked. Game not so much though hahaha

20:37 <neobrain> ah great...ish I guess, heh