robclark changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Chrome OS Trogdor Devices - Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630 - Lenovo ThinkPad X13s - and various other snapdragon laptops) - https://oftc.irclog.whitequark.org/aarch64-laptops
todi1 has quit [Remote host closed the connection]
todi has joined #aarch64-laptops
KhazAkar has quit [Quit: Connection closed for inactivity]
matthias_bgg has quit [Ping timeout: 480 seconds]
matthias_bgg has joined #aarch64-laptops
Ablu is now known as Guest8265
Ablu has joined #aarch64-laptops
Guest8265 has quit [Ping timeout: 480 seconds]
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
exit70_ has quit []
exit70 has joined #aarch64-laptops
exit70 has quit []
todi has quit [Remote host closed the connection]
todi has joined #aarch64-laptops
martiert has quit [Quit: WeeChat 4.1.1]
martin has joined #aarch64-laptops
martin is now known as Guest8286
Guest8286 has quit []
martin1 has joined #aarch64-laptops
iivanov has joined #aarch64-laptops
KhazAkar has joined #aarch64-laptops
srinik has quit [Killed (NickServ (Too many failed password attempts.))]
<jhovold>
- fix usb dwc3 resource leaks on probe deferral
<jhovold>
- fix usb dwc3 broken wakeup after probe deferral
<jhovold>
- fix a690 gpu register values
<jhovold>
- fix bad asoc merge
<jhovold>
Note that the pops and clicks artefacts are completely gone with 6.7-rc1 (and soon also with 6.5.3 and 6.5.13) after a seemingly unrelated fix made the remaining pops go away.
<Jasper[m]>
Sounds good
<craftyguy>
_[m]1: there are ground pins on the m.2 connector, what do you mean?
<jhovold>
That was supposed to say *6.6.3* and 6.5.13
<clover[m]>
jhovold: Nope, right after a fresh boot. Downgraded to 6.6.2-0 and issue went away. I have to do some work now but can check dmesg logs later
<jhovold>
steev: no need to remove asound.state for the pops and clicks fixes, only needed after updating the ucm files to address the speaker volume/distortion issues
<jhovold>
clover[m]: probably something different then, looks different, my screen becomes completly garbled
<jhovold>
oh, and you said external monitor, i see it on the eDP, sorry for the noise
<jhovold>
clover[m]: did it happen on every boot or just once? if so it should be straight forward to track down the change in 6.6.3 that caused it
<clover[m]>
jhovold: every boot. for clarity, i was running 6.6.2-1 (whatever is currently in steev's 6.6.y branch). and things are working again after downgrading to 6.6.2-0
<steev>
clover[m]: i probably screwed up the backport of the a690 regs
<steev>
you can drop that patch and see if it fixes it
<jhovold>
oh, ok, shouldn't blame stable then
<robclark>
clover[m]: cat /sys/kernel/debug/drm/0/state ? That kinda looks like display thinks the buffer is linear but gpu thinks it is UBWC .. which would probably be a compositor bug.. if you could run the compositor with FD_MESA_DEBUG=noubwc and that "fixes" it, that would confirm
<robclark>
hmm, or it could be messed up ubwc setting
<jhovold>
fwiw, external display still works with 6.7-rc3 and the v2 of the a690 fixes here, only using 1920x1200 though
<steev>
definitely not on 6.6, but i probably screwed it up, either way, i'll drop it
<steev>
just tested here and i just get... gnome crashing
<steev>
actually unplugging and plugging my adapter back in and it works fine
<steev>
Oddly though, the gnome crashes seemed to have reset my font settings, and my alacritty is *very* readable from a distance
<steev>
clover[m]: pushed again, dropping the patches
<_[m]1>
can I conclud the corsair one is not working for x13s?!
<_[m]1>
s/conclud/conclude/
<_[m]1>
I was trying debian maybe I'll go for ye olde ironrobin arch iso 🙂
rfs613- is now known as rfs613
<robclark>
steev: I noticed that g-s crashing make things fall back to Xorg.. which might be what is happening to you.. `ps auxw | grep X` and see if you have Xorg vs Xwayland?
<steev>
robclark: it was neither; it was crashing and falling back to "_" in the uppper left corner (and yes X is installed, and normally that's what happens) - either way, i'm not gonna back port the reg changes to stable, i'll leave that to the ~~professionals~~people better than me
<steev>
there is definitely something odd going on though, because i don't always seem to get my other sessions
<steev>
oh, i bet i'm in X only and we reinstalled something that disabled my damn wayland sessions
<steev>
wtf? i can't get wayland on 6.7 only on 6.6?
<steev>
oh, i bet its mutter
laine has joined #aarch64-laptops
<steev>
that also explains the "font" changes... because alacritty uses different font sizing between xorg and wayland
<steev>
jhovold: one thing i notice is, every so often, logging out and back in (gnome sessions testing xorg vs wayland), audio will just break; another logout cycle usually fixes it, but it's kinda weird
<_[m]1>
lol I don't understand nothing,when I put another m2 nvme the debian installer passes into installation after loading dtb, I can boot into archiso and I see the block device and can partition and write ext4 fs but somehow it won't continue debian installer grr
<robclark>
steev: "because X" :-P ... actually I bet that is coming from the trick that xorg tries to do for "smooth" transition from gdm to X.. when it starts it grabs the current scanout buffer and memcpy's it into the buffer used for modeset. It probably ignores the modifier in the process (since that code pre-dates modifiers) so you end up with transient UBWC confusion until gnome-shell steps in and starts rendering
<robclark>
I think the answer is "don't X" (but ofc someone needs to fix the g-s crash first)
<steev>
it's not gs crash, it's mutter, but yeah :(
<steev>
there's a workaround, but i hate having to carry it forward
<steev>
_[m]1: which debian installer? i think you need testing/sid for the 6.5 kernel
<robclark>
yeah, ok, I tend to use those names interchangeably ;-)
<robclark>
I just wish fedora would carry the workaround
iivanov has quit [Quit: Leaving]
<steev>
weird they don't, i thought they wrote it
<robclark>
no, I think it was someone from linaro
<steev>
ah, for some reason i was thinking danielt was at fedora
<steev>
i haven't tried it myself, since i didn't buy the WD and now it's on less sale (they moved it from $70 off to $20 off for cyber monday??)
<clover[m]>
i bought a WD, full price. its been flawless so far
<clover[m]>
i agree with hdkr that the extender isnt really necessary. it fits SUPER snug
<steev>
well, if you wanna buy me a christmas present, i'm fine with that :P
<_[m]1>
<steev> "i haven't tried it myself, since..." <- wait for a 2tb disk? i thgouth the extender thing w t f
<clover[m]>
wait wtf. i bought back in july for 250, and now its supposedly on "sale" for 250? wack
<_[m]1>
yeah now I saw the wd is on alixpress grr for 120 something
<HdkR>
clover[m]: I still recommend a little bit of capton tape if you don't use an extender, for that vibration resist
<HdkR>
kapton even
<steev>
85.60 is the cheapest i'm seeing it on aliexpress.... that is tempting but still... i need to do adult things with my money (aka i need to move soon)
<steev>
oh wait, that's only the 500gb
<steev>
the 2tb is 232 on there
<clover[m]>
<steev> "well, if you wanna buy me a..." <- maybe, but you'd have to dedicate at least 1 TB to arch linux
<_[m]1>
at least !! 😄
<_[m]1>
I honestly thought you needed some sort of ground point on the end screw point, honestly still unsure - or is kapton tape inductive?
<Jasper[m]>
I paid 200€/~219$ for my 2TB SN770M
<Jasper[m]>
On WD's site
<_[m]1>
oh yeah 260 here, was looking at wrong one
<_[m]1>
cheaper huh ? but it ships from usa?
<craftyguy>
_[m]1: ah, i thought the metal on the screw part was just so you don't grind/stress the pcb when tightening the screw. it would be weird if that was the only connection to ground...
<Jasper[m]>
_[m]1: Ehhh, from The Hague so fairly local really
<HdkR>
_[m]1: Screw is just there to hold it captive, the connector has other ways to provide a ground to the drive. Latest motherboards are using little plastic retention systems these days even
<_[m]1>
HdkR but why the copper rhaaa
<_[m]1>
myeah so mm fckd
<_[m]1>
anywya I usually don't do all that much IO on my internal disk
<_[m]1>
going to testdrive quartz64a now, anyone experience with setting up NAS n such?
<_[m]1>
s/n/on/
<clover[m]>
no but im sure you could follow instructions for e.g. a raspberry pi
<HdkR>
_[m]1: Yea, as craftyguy, the plating there while also likely going to the ground plane, is also for screw relief :)
<HdkR>
Those SSDs look like 870 QVOs which are quite nice. The best $/GB of any SATA SSD
<Jasper[m]>
QLC does have it's issues
<_[m]1>
I took WD red disks out of the book product
<Jasper[m]>
But otherwise yeah
<Jasper[m]>
_[m]1: I did the same, paid like 169€ a pop at the time for 2x8TB drives
<_[m]1>
so mm zfs or btrfs?
<Jasper[m]>
Worked OOTB too, no weird 3.3V sense wire
<_[m]1>
I prefer btrfs because it's in tree
<Jasper[m]>
_[m]1: zfs, but I'm not on a stable kernel yet. gonna use kwizarts copr for that soon
<_[m]1>
so you like updating huh not
<Jasper[m]>
<_[m]1> "so you like updating huh not" <- Fedora is kinda quick
<Jasper[m]>
Too quick for openzfs
<HdkR>
ZFS is also battling a data corruption bug right now. Might be good to wait a month for that to smooth over :)
<steev>
Jasper[m]: in the US, we can't buy it from WD site, it's Best Buy exclusive here, unfortunately (or we can buy on aliexpress for slightly more)
<craftyguy>
I went with a QLC 2TB drive from team group. I've had the same one in my steam deck for a few months and it seems fine. I try to avoid WD :P
<HdkR>
WD has proven on multiple occasions that it doesn't have the consumer's best interest in mind. Hopefully that changes at some point
<HdkR>
But also Seagate has issues, so choices start getting harder and harder
<_[m]1>
<Jasper[m]> "Too quick for openzfs" <- is disabled per default but I read lol nice timing also only in latest
<_[m]1>
HdkR: just do raid 6 + 2 dunno
<HdkR>
_[m]1: Yea, I do RAIDZ2, which is a bit overkill on SSDs but oh well :)
<_[m]1>
why overkill on SSDs? it's more likely 2 will fail at same time than HDD?
<HdkR>
Mostly because reconstructing an array of SSDs is significantly faster than platter drives, so the period of being vulnerable to data loss from a second driver failure is less of a concern. Especially since reading doesn't cause much wear on SSDs unlike platters
<_[m]1>
hmyes makes sense
<HdkR>
Sadly large capacity U.2/U.3 SSDs aren't quite as cheap as SATA SSDs yet, which would significantly improve that reconstructing performance.
<HdkR>
Intel leaving this space is probably one of the most sad developments but maybe Kioxia will get us something comparable at some point.
<_[m]1>
isn't btrfs still default fs in fedora but like it got ripped out of rhel?
<_[m]1>
sometimes I seriously don't get redhat
djakov has quit [Remote host closed the connection]