2022-12-21 00:44
ChanServ changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630)
01:07
<
clover[m] >
Are any of y'all interested in getting a pinetab 2?
02:46
<
amstan >
woah, that's a nice price
02:47
<
clover[m] >
pinetab 1 was painfully slow, but this one looks like quite an improvement
03:00
hightower3 has joined #aarch64-laptops
03:07
hightower2 has quit [Ping timeout: 480 seconds]
03:43
matthias_bgg has quit [Ping timeout: 480 seconds]
03:47
<
clover[m] >
steev, works!
03:48
<
clover[m] >
now if only i can get displaylink working
03:48
<
clover[m] >
i can actually use these monitors that have been collecting dust for like a year
04:31
hexdump01 has joined #aarch64-laptops
04:33
hexdump0815 has quit [Ping timeout: 480 seconds]
04:36
<
clover[m] >
Wow I managed to break gnome
04:36
<
clover[m] >
Good time to switch to sway?
04:38
<
clover[m] >
I get this when trying to start gdm
04:55
<
clover[m] >
oh i accidentally deleted libevdev lmao
04:55
<
clover[m] >
trying to start sway told me
04:55
<
clover[m] >
all good i think
04:57
<
travmurav[m] >
steev: I think what saved you is that you've added the node into DP port node and not the DPU node as the original patch :P
05:04
<
steev >
oh, that's probably it yeah... i think i actually tried the dpu first and it didn't work...
05:04
<
travmurav[m] >
which was the intended behavior probably :D
05:06
<
clover[m] >
i get a kernel null pointer deference thing when i plug in my displaylink monitor
05:08
<
steev >
i blame evdi
05:09
<
clover[m] >
should i tell them, or the prob dont care?
05:10
<
steev >
should probably report it
05:10
<
steev >
i have no idea about evdi
05:11
<
travmurav[m] >
sounds like sad glue to proprietary userspace blobs
05:20
<
amstan >
it's possible it's the kind of bug that we might be interested to fix for chromeos, but not really sure how high priority it would be
05:20
<
amstan >
dianders's hobby is to look at these tricky usb bugs
05:32
<
HdkR >
I hear other DRM drivers also have problems with displaylink, so it might not be limited to qcom
05:50
<
clover[m] >
hdkr i haven't had much time to test your request. will try tomorrow
05:57
<
HdkR >
clover[m]: no problem
06:08
<
clover[m] >
i can drive two displays now, cool
06:08
<
clover[m] >
ok bedtime
07:01
matthias_bgg has joined #aarch64-laptops
07:13
systwi has quit [Ping timeout: 480 seconds]
07:16
<
amstan >
ah, so not a usb problem
07:25
systwi has joined #aarch64-laptops
08:37
irctian1 has joined #aarch64-laptops
08:46
irctian1 has quit [Remote host closed the connection]
08:46
irctian1 has joined #aarch64-laptops
11:16
Erisa has quit [Remote host closed the connection]
11:17
Erisa has joined #aarch64-laptops
11:48
falk689 has quit [Ping timeout: 480 seconds]
11:55
falk689 has joined #aarch64-laptops
13:22
<
abelvesa >
can't repro it easily yet
13:56
hightower3 has quit [Ping timeout: 480 seconds]
14:41
hightower2 has joined #aarch64-laptops
17:09
<
clover[m] >
my bluetooth headsets are not showing up as Input devices, only output devices. i'd like to be able to use the mic in my headset. this used to work. any advice?
17:15
<
clover[m] >
nvm, i think can use this jabra usb dongle thing
17:28
matthias_bgg has quit [Ping timeout: 480 seconds]
17:37
matthias_bgg has joined #aarch64-laptops
18:11
<
steev >
clover[m]: i don't know what would have broken in bluetooth - show me dmesg | grep Bluetooth
18:11
<
steev >
i think, with the new stuff, i don't override the board id anymore, and you mgiht need to symlink the nvm patch
18:11
<
steev >
or just rename it
18:12
<
steev >
[ 7.566587] Bluetooth: hci0: qca_generate_nvm_name: nvm name is qca/hpnv21.8c
18:12
<
steev >
[ 7.566590] Bluetooth: hci0: QCA Downloading qca/hpnv21.8c
18:13
<
steev >
but that's shouldn't break it?
19:14
<
clover[m] >
Its not a Bluetooth issue. Its a Firefox issue
19:14
<
clover[m] >
Zoom audio works in chromium. Zoom video works on Firefox. FML
19:25
<
steev >
ah, good deal
19:25
<
steev >
btw, i just noticed srini cc'd me on a couple more asoc patches
19:25
<
steev >
it's my week for bug tracker so all bug emails go to my inbox this week :D
19:30
<
gwolf >
clover[m]: Does the device show in pulseaudio? (I use pavucontrol)
19:30
<
gwolf >
if firefox opens the audio-in sink, you can force via pulseaudio which device it is taken from
19:58
hexdump01 has quit [Quit: WeeChat 1.9.1]
19:58
hexdump0815 has joined #aarch64-laptops
20:01
<
clover[m] >
now i see why chromium video is not working on zoom. no WebGL
20:01
<
clover[m] >
the test passes with Firefox. wierd
20:05
<
hexdump0815 >
seems to be sdm850 with 8gb ram and 128gb ufs and goes new for 200 euro and sometimes a bit below in germany - sounds fair to me - just in case anyone wants to enter the "build a dts from the limited acpi info" game with some yet different hardware :)
20:21
<
clover[m] >
if i start chromium with --disable-gpu-compositing zoom video in starts working
20:49
<
steev >
i wonder if our hack for mutter is rbeaking it?
20:50
<
clover[m] >
iirc it's been broken like that for a while
20:50
<
clover[m] >
its weird cause vivaldi works lol
20:50
<
clover[m] >
vivaldi is giving best zoom experience in browser right now
20:51
<
steev >
oh, you might be right
20:51
<
steev >
iirc there's a bug open somewhere about chromium being broken since ~95
20:51
<
steev >
but the person the bug is assigned to... has never even shown up on the bug tracker
20:51
<
steev >
(for that bug)
20:52
<
steev >
oh, no, that's a different one
20:53
<
clover[m] >
that might explain why vivaldi works. i think vivaldi is based on an older version of chrome
20:53
<
steev >
not afaik, i thought they were based on 100, but they do different stuff with media/graphics stack
20:54
<
clover[m] >
im on vivaldi stable 3.8, upgrading to 5.7 now. lets see if it breaks
21:06
<
steev >
i'm just saying i *think*; i'm in no way an expert, and i may be thinking of Arc
21:06
<
steev >
i recently uninstalled vivaldi, but yeah... i was using it for a while because it was the only browser that actually worked when everything else was broken
21:14
<
clover[m] >
upgrading broke proprietary media support lol
21:20
matthias_bgg has quit [Ping timeout: 480 seconds]
21:29
matthias_bgg has joined #aarch64-laptops
21:52
<
clover[m] >
vivaldi looking good
21:54
<
steev >
yeah, it is pgood
21:54
<
steev >
it surprises me that linux chromium is so broken when cros is literally based on linux
21:58
<
robclark >
for good measure, try the arg to disable gpu sandbox?
21:59
<
robclark >
and yeah, there is probably too much #ifdef CHROMEOS type stuff with things needed to fix things for arm linux
22:02
<
HdkR >
#ifdef
__aarch64__ && APPLE #define PAGESIZE 16384
22:11
<
clover[m] >
HdkR why do you think vulkan is disabled?
22:11
<
HdkR >
clover[m]: What do you mean?
22:12
<
clover[m] >
on vivaldi://gpu its showing vulkan as disabled
22:12
<
HdkR >
Did you enable it when you built mesa?
22:12
<
HdkR >
I don't even know what vivaldi is
22:13
<
clover[m] >
its a browser lol
22:13
<
HdkR >
Then I guess ensure you built the correct branch of mesa with the vulkan driver enabled and hope that browser doesn't block turnip
22:17
<
HdkR >
And of course if this is an a690 device, hope that the GPU doesn't crash.
22:18
<
HdkR >
es2_info, glxinfo, and vulkaninfo outside of the browser to ensure you have a working hardware driver of course.
22:22
<
steev >
i definitely enabled it here, and just noticed that yeah... it says not compatible
22:22
<
steev >
more specifically, failed with ERROR_INCOMPATIBLE_DRIVER (vulkaninfo)
22:23
<
steev >
that's mesa as of... sometime today, with rob's 3 patches on top?
22:23
<
steev >
or did i do something stupid when i meson built it telling it to use freedreno and i shoulda told it turnip?
22:24
<
HdkR >
`prefix : /opt/freedreno` Not `/usr`?
22:24
<
clover[m] >
idk what turnip is, it doesnt seem to be a meson option
22:24
<
steev >
turnip is the name of the vulkan driver
22:25
<
HdkR >
The loaders hardcode the prefix path, so if you're copying it over to `/usr` later then you're breaking stuff :P
22:25
<
steev >
why would you think i would?
22:25
<
HdkR >
It's just a weird path
22:25
<
steev >
steev@wintermute:~/git/mesa$ cat /etc/ld.so.conf.d/0-freedreno.conf
22:25
<
steev >
/opt/freedreno/lib/aarch64-linux-gnu
22:25
<
HdkR >
So maybe the loader is breaking
22:25
<
steev >
opt isn't a weird path
22:26
<
HdkR >
ls `/usr/share/vulkan/icd.d/`
22:26
<
steev >
that doesn't exist
22:26
<
robclark >
you should use something along the lines of -Dgallium-drivers=freedreno -Dvulkan-drivers=freedreno
22:26
<
steev >
it does exist in /opt/freedreno/share/vulkan/icd.d tho
22:27
<
steev >
seems odd that it would hardcode it since you're supposed to be able to devenv mesa for a while now
22:27
<
HdkR >
strace the thing and see if it is not finding the icd file I guess? :)
22:28
<
robclark >
the loader (not part of mesa) needs to know where to look for icd files.. should be env var to override, IIRC
22:28
<
steev >
that is exactly part of it
22:28
<
steev >
okay, so i need to do that as well i suppose
22:28
<
robclark >
export VK_ICD_FILENAMES=$p/release/b/share/vulkan/icd.d/freedreno_icd.aarch64.json
22:31
<
clover[m] >
do we also need to install vulkan mesa layers?
22:32
<
steev >
no idea what those are... but if you're building manually and not installing them? yes
22:33
<
szclsya[m] >
my build of mesa works with Vulkan just fine
22:35
<
steev >
i would just patch mesa itself, but... debian is in freeze, so it's gonna be a while til 23 rotates in
22:37
<
clover[m] >
how can you tell Leo Shen ?
22:37
<
steev >
vulkaninfo doesn't error out :P
22:38
Mathew has joined #aarch64-laptops
22:38
mcbridematt has quit [Read error: Connection reset by peer]
22:43
<
clover[m] >
ok i can enable it in vivaldi but it does that for zoom, and video playback stops working in youtube
22:45
<
szclsya[m] >
<clover[m]> "how can you tell Leo Shen ?" <- vkcube
23:00
falk689 has quit [Remote host closed the connection]
23:00
falk689 has joined #aarch64-laptops
23:01
<
steev >
yeah, vkcube runs fine here
23:06
<
clover[m] >
i tried emulating x86_64 Zoom with box64, it was so slow, i gave up trying to launch it after it hung for 15 minutes
23:07
<
clover[m] >
idk if its just super heavy or what
23:07
<
HdkR >
It's a browser, so yes
23:13
<
szclsya[m] >
its actually qt, but heavy indeed
23:14
<
HdkR >
Oh, is it not actually an embedded browser?
23:14
<
clover[m] >
i noticed it required qt-tools
23:14
<
szclsya[m] >
yeah I'm surprises as well lol
23:14
<
HdkR >
no, I just ran it and it uses CEF
23:16
<
HdkR >
You probably need Box64's `steam_chrome` branch if you want CEF to work correctly there
23:17
<
clover[m] >
oh, nice. will try that later
23:17
<
clover[m] >
that branch seems pretty active. might be best to wait
23:20
<
clover[m] >
will post results