Desktop freeze?!

[Hedon James]

okay....Debian LXDE has no issues with xclients using up all connections while sleeping.  I think it's safe to say this is an LXQT issue.  And just for good measure, I noted that my LXQT remix uses LightDM as the display manager, while stock Debian LXQT uses sddm as the display manager.  I now conclude this is an LXQT issue.  Or possibly an issue with the combination of LXQT and this BIOS.

 

Ray, your suggestion of EndeavorOS LXQT testing would be interesting, just to see if this is an LXQT issue....or an issue of Debian implementation of LXQT.  But after all this....I'm burnt.  I'm just going to attempt to flash the BIOS and see if that doesn't do the trick.  I'm out of time, and have so much else to do!

 

My refurbed Lenovo m910q with Intel CPU is supposed to arrive tomorrow, and you can be damn sure I'll be testing the LiveUSB before installation on that one!  Jeezus.....

[Hedon James]

latest report...BIOS flashed...which wasn't nearly as painful as I feared it to be.  After the password/disk lock fiasco, I feared whatever else might be lurking in that BIOS.  But it was actually quite easy, and went smoothly.  If I had known, I would've done it sooner.

 

Unfortunately, it would've led to the same outcome.  Even with updated BIOS, I'm getting zombie Xclient connections occurring in the background during sleep.  I am now clueless.  I guess I'll focus on the new machine arriving tomorrow and put this one on the shelf until I acquire that golden nugget of knowledge that makes me say "aha...I gotcha now!"  sigh...

[raymac46]

I wonder if when LXQt runs with Wayland properly all this will be moot. I imagine EndeavourOS might roll this way.

[securitybreach]

It depends, some of us are still using Xorg.

[sunrat]

Just out of curiosity, I tried xrestop on my machine. 36 clients and it's been up for 8 days.

Totally different system though, Bullseye w/ KDE Plasma, i5-6500, Nvidia graphics.

In case you want to try a newer kernel, linux-image-6.5.0-0.deb12.1-amd64-unsigned (or 6.4.4-3~bpo12+1 if you need a signed one)  is available in bookworm-backports now, or try a Liquorix 6.5 kernel.

[Hedon James]

40 minutes ago, sunrat said:

Just out of curiosity, I tried xrestop on my machine. 36 clients and it's been up for 8 days.

Totally different system though, Bullseye w/ KDE Plasma, i5-6500, Nvidia graphics.

In case you want to try a newer kernel, linux-image-6.5.0-0.deb12.1-amd64-unsigned (or 6.4.4-3~bpo12+1 if you need a signed one)  is available in bookworm-backports now, or try a Liquorix 6.5 kernel.

i've got around 40-something clients on my Debian 10 machine, with LXQT, and it's been running for several weeks.  I'd have a go at the liquorix kernel, but not sure how that might address an LXQT issue?  Based on my testing, it seems to be localized to Debian LXQT.  Maybe other LXQTs, but I haven't confirmed that.  But I have confirmed that live sessions of Debian LXDE and OpenSUSE KDE do NOT have these issues, on the same hardware.  I'll probably file a bug report with LXQT devs, but only after I have a stable machine to migrate to.

[raymac46]

I have just tried the LXQt desktop on Debian Trixie on my Thinkpad T430 with ancient Intel graphics. Xrestop shows 25 clients and after a short Suspend (maybe 10 minutes) I was right back with 25 clients again. I don't know how long you need to suspend. I don't have a powersaver suspend here, just suspended through the menu. For what it's worth.

ray@ray-debian-T430:~$ inxi -Fxz
System:
  Kernel: 6.5.0-3-amd64 arch: x86_64 bits: 64 compiler: gcc v: 13.2.0
    Desktop: LXQt v: 1.3.0 Distro: Debian GNU/Linux trixie/sid
Machine:
  Type: Laptop System: LENOVO product: 2347H76 v: ThinkPad T430
    serial: <superuser required>
  Mobo: LENOVO model: 2347H76 serial: <superuser required>
    UEFI-[Legacy]: LENOVO v: G1ET94WW (2.54 ) date: 04/22/2013
Battery:
  ID-1: BAT0 charge: 40.2 Wh (90.1%) condition: 44.6/55.5 Wh (80.4%)
    volts: 12.1 min: 11.1 model: SANYO 45N1023 status: discharging
CPU:
  Info: dual core model: Intel Core i5-3320M bits: 64 type: MT MCP
    arch: Ivy Bridge rev: 9 cache: L1: 128 KiB L2: 512 KiB L3: 3 MiB
  Speed (MHz): avg: 1249 high: 1400 min/max: 1200/3300 cores: 1: 1197
    2: 1200 3: 1200 4: 1400 bogomips: 20752
  Flags: avx ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
Graphics:
  Device-1: Intel 3rd Gen Core processor Graphics vendor: Lenovo driver: i915
    v: kernel arch: Gen-7 bus-ID: 00:02.0
  Device-2: Bison Integrated Camera driver: uvcvideo type: USB
    bus-ID: 1-1.6:4
  Display: x11 server: X.Org v: 1.21.1.8 with: Xwayland v: 23.2.1 driver: X:
    loaded: modesetting unloaded: fbdev,vesa dri: crocus gpu: i915
    resolution: 1366x768~60Hz
  API: EGL v: 1.5 drivers: crocus,swrast platforms:
    active: gbm,x11,surfaceless,device inactive: wayland
  API: OpenGL v: 4.2 vendor: intel mesa v: 23.2.1-1 glx-v: 1.4
    direct-render: yes renderer: Mesa Intel HD Graphics 4000 (IVB GT2)
Audio:
  Device-1: Intel 7 Series/C216 Family High Definition Audio vendor: Lenovo 7
    driver: snd_hda_intel v: kernel bus-ID: 00:1b.0
  API: ALSA v: k6.5.0-3-amd64 status: kernel-api
  Server-1: PipeWire v: 0.3.83 status: active
Network:
  Device-1: Intel 82579LM Gigabit Network vendor: Lenovo driver: e1000e
    v: kernel port: 5080 bus-ID: 00:19.0
  IF: enp0s25 state: down mac: <filter>
  Device-2: Intel Centrino Advanced-N 6205 [Taylor Peak] driver: iwlwifi
    v: kernel bus-ID: 03:00.0
  IF: wlp3s0 state: up mac: <filter>
Drives:
  Local Storage: total: 119.24 GiB used: 40.92 GiB (34.3%)
  ID-1: /dev/sda vendor: Samsung model: MZ7PC128HAFU-000H1 size: 119.24 GiB
Partition:
  ID-1: / size: 109.23 GiB used: 40.92 GiB (37.5%) fs: ext4 dev: /dev/sda1
Swap:
  ID-1: swap-1 type: partition size: 7.7 GiB used: 0 KiB (0.0%) dev: /dev/sda5
Sensors:
  System Temperatures: cpu: 37.0 C mobo: N/A
  Fan Speeds (rpm): fan-1: 0
Info:
  Processes: 224 Uptime: 28m Memory: total: 12 GiB note: est.
  available: 11.39 GiB used: 1.9 GiB (16.7%) Init: systemd
  target: graphical (5) Compilers: gcc: 13.2.0 Packages: 2928 Shell: Bash
  v: 5.2.15 inxi: 3.3.30
ray@ray-debian-T430:~$ 

 

[raymac46]

Note wayland graphics appears to be inactive.

Edited October 31, 2023 by raymac46

[V.T. Eric Layton]

You're a glutton for punishment, @Hedon James. Just install Slackware w/ Xfce and enjoy your computer life again.

[Hedon James]

41 minutes ago, V.T. Eric Layton said:

You're a glutton for punishment, @Hedon James. Just install Slackware w/ Xfce and enjoy your computer life again.

I am.  A little S&M never hurt anyone....at least not in a BAD way.  LOL!

[V.T. Eric Layton]

If you were REALLY into S&M, you'd be running Gentoo. 

[securitybreach]

2 hours ago, V.T. Eric Layton said:

If you were REALLY into S&M, you'd be running Gentoo. 

 

Or Linux From Scratch

[Hedon James]

7 hours ago, V.T. Eric Layton said:

If you were REALLY into S&M, you'd be running Gentoo. 

I just dabble.  on occasion.  when the mood is forced upon me. 

[V.T. Eric Layton]

 

 

And for those of you who may prefer the less intrusive Invidious version.

[raymac46]

I assume you have seen this thread:

https://forum.artixlinux.org/index.php/topic,3514.0.html

[Hedon James]

8 hours ago, raymac46 said:

I assume you have seen this thread:

https://forum.artixlinux.org/index.php/topic,3514.0.html

Yes I have.  That's the one that sent me down the rabbit hole of xclient connections, which eventually lead to my workaround/bandaid.  But I appreciate you looking!

[Hedon James]

Well this just continues to get all clustered up....

 

I posted on the LXQT forum, and devs quickly determined it's not an LXQT issue, but a configuration issue.  Environment is suspected (this is important....I'll circle back).  I posted results of requested information, but haven't heard back from the devs.

 

In the meantime, I ordered a Lenovo ThinkCentre Tiny m910x....basically the same machine as before, but with Intel iCore7 6700T processor and Intel HD 530 graphics.  i915 driver.  I tested this with LiveMedia for a few hours with no issues.  Feeling confident, I installed.  And I have graphics issues.  Not the same as the Ryzen5.  When the Core i7 wakes, my tint2 has re-positioned from right hand center to the top left; and one of my lxpanels repositions slightly; and SOME windows launch without decorations (notably Firefox and qterminal)....until I restart Fluxbox, and everything gets restored, and working fine until the next idle/sleep session.

 

With the "problem" Ryzen5 disconnected from monitor, I sat it on a credenza and dug out an old Dell 15" VGA with 4:3 ratio.  I figured this would save me the trouble of disconnecting and reconnecting monitors, and booting & rebooting (hot swaps are not recognized by these Lenovos) while I was receiving assistance from LXQT devs.  I moused the machine today to wake it up, and there are NO PROBLEMS with the monitor!  No zombie xclient connections getting used up...just holding steady at 40 clients...same as my Debian 10 setup for 3+ years now!

 

I do believe the LXQT dev is onto something with his preliminary thoughts.  It seems that problems follow the 27" 1920x1080 monitor, not the machine!  But before anyone suspects the monitor, this monitor was connected to my Debian 10 desktop for 3+ years with NO ISSUES, and was connected to my Ubuntu 16.04 desktop for about 3 years before that, with NO ISSUES.  I have to believe the monitor is fine.  But something has obviously changed between Debian 10 and Debian 12.  Or between LXQT 0.14 and LXQT 1.20.  The only other variable I can see, besides the monitor, is connection method.  Both Lenovos require DisplayPort monitor connections.  But neither monitor has that port.  So the 27" 1920x1080 "problem" monitor is an HDMI>DisplayPort adaptor>Lenovo connection; while the 15" 4:30 "good" monitor is a VGA>DisplayPort adaptor>Lenovo connection.  Could HDMI vs VGA be the culprit?  Or maybe the DisplayPort adaptor to the respective monitor connector?

 

Ideally, I'd like to use the Ryzen 5 machine.  I just prefer AMD...personal preference.  But I won't be too upset having to use a Core i7.  I'll take the low-hanging fruit, for now!

 

Considering that my zombie client issues disappeared on the Ryzen simply by changing monitors, where should I be looking now?  Seems like it should be a quick & easy fix, if I only knew how, or where to look?

OTOH, any idea how to address the weird behavior with the Core i7?  It just seems like something simple & stupid is "off" with both machines.  Which would be easier to address?

 

Anyone have experience with either of these issues?  Is it something with Xorg.conf?  Which Debian no longer provides, apparently?  here I am shrugging my shoulders again...

Edited November 1, 2023 by Hedon James

[raymac46]

OK I'll bite. Does the VGA monitor work OK with the new Intel desktop - ie no shifted tint2 or lxpanels, no missing decorations? 

[Hedon James]

22 minutes ago, raymac46 said:

OK I'll bite. Does the VGA monitor work OK with the new Intel desktop - ie no shifted tint2 or lxpanels, no missing decorations? 

hmmm.....interesting question.  hadn't thought of testing that?  let me reconfigure and hookup hardware. 

 

Ironically, we're looking at the same things again, but different angles...posted here for future reference....not quite my problem, but perhaps related?  A displayport issue perhaps?

https://forums.debian.net/viewtopic.php?t=156021

 

let's run down the VGA question first.  that may provide a clue.  be back shortly...

Edited November 1, 2023 by Hedon James

[Hedon James]

okay....question answered....the VGA monitor with DisplayPort adaptor works just fine with the new Intel machine!

 

So it's gotta be the monitor, OR the HDMI/DisplayPort connection.  And since that monitor was connected via HDMI to my Debian desktop for 3+ years with NO ISSUES, I'm going to rule out the monitor.  I think it's the HDMI/DisplayPort connector.  and the post above seems to indicate that an HDMI/DisplayPort cable solved his problem.

 

I wish I could hookup my large monitor via VGA/DisplayPort because that would confirm that HDMI/DP is the only "bad connection" combo.  And I still don't KNOW if a new cable will solve the problem, but an $8 cable is a preferred solution over a $150+ monitor, if the cable is an option!  LOL!

[raymac46]

Most monitors have a couple of HDMI ports usually. Does yours, and have you tried connecting via a second HDMI port?

[Hedon James]

14 minutes ago, raymac46 said:

Most monitors have a couple of HDMI ports usually. Does yours, and have you tried connecting via a second HDMI port?

Yep, 2 HDMI connections.  tried 'em both.  and 3 DP ports on the Lenovos.  tried them all too.  and every combination.  same results.

 

But....I did locate a recessed (almost hidden) VGA port!  So I got the Ryzen hooked up to the "problem" monitor and got it booted with a nomodeset parameter in GRUB (didn't need that for the other VGA monitor?!) and the desktop loaded up.  Got a terminal open to monitor xclient connections.  Give me enough time for the machine to idle/sleep and possibly build up more xclient connections (or NOT!).

 

If the xclients build up, it's gotta be the monitor configuration.  if they do not, it HAS TO be the HDMI/DP connection referenced in the post above.  Stay tuned...

[raymac46]

AAAND of course you gotta hook up that monitor via VGA to the Intel and see if you need nomodeset and if it works.

Nomodeset can often get you into the VESA driver. If that happens you may still have driver issues.

Edited November 1, 2023 by raymac46

[Hedon James]

it's been 27 minutes since my last post, and xclients holding steady.  Looks like the culprit is confirmed to be the HDMI/DP combination!

 

With 2 machines, each with 3DP connections; and with 2 monitors, 1 with VGA only, and 1 with 2 HDMI and 1 VGA, I have 24 possible connections.  12 of those connections are HDMI/DP and 12 of them are VGA/DP.  All 12 HDMI/DP combinations yielded graphics problems, although different problems for the Ryzen and Intel.  9 of the 12 VGA/DP connections have been tested, and none have yielded any graphical artefact issues, nor xclient zombie spawns.  I feel confident in saying that the last 3 untested combinations will yield the same results.  And if I get what I want, the Ryzen will be connected to the 27" 'problem', not the Intel.  So I think I'm done testing.

 

I've got a new HDMI/DP cable coming tomorrow.  Gawd I love Amazon.  And if it works, I WON!  and if it doesn't, it appears this VGA/DP connection works just fine, so I STILL WIN....just not the way I wanted to.  I gotta thank you Ray!  Once again, we were looking at the same issue, but from different angles, and we still arrived at the (likely) same conclusion.  But I got there a LOT faster with your insights!  THANK YOU!

[Hedon James]

49 minutes ago, raymac46 said:

AAAND of course you gotta hook up that monitor via VGA to the Intel and see if you need nomodeset and if it works.

Nomodeset can often get you into the VESA driver. If that happens you may still have driver issues.

Good call on the Vesa driver!  That's exactly what happened!  (but nomodeset was NOT required on the "good" monitor, so presumably loaded with amdgpu(?) or radeon(?) or whatever the default AMD driver is for VGA connections)  Here's my Inxi  -Fxz on this machine, and there's the Vesa driver in effect!:

inxi -Fxz
System:
  Kernel: 6.1.0-13-amd64 arch: x86_64 bits: 64 compiler: gcc v: 12.2.0
    Desktop: LXQt v: 1.2.1 Distro: Debian GNU/Linux 12 (bookworm)
Machine:
  Type: Mini-pc System: LENOVO product: 10VHS17600 v: ThinkCentre M715q
    serial: <superuser required>
  Mobo: LENOVO model: 3130 v: SDK0J40697 WIN 3305183174933
    serial: <superuser required> UEFI: LENOVO v: M1XKT62A date: 06/21/2023
CPU:
  Info: quad core model: AMD Ryzen 5 PRO 2400GE w/ Radeon Vega Graphics
    bits: 64 type: MT MCP arch: Zen rev: 0 cache: L1: 384 KiB L2: 2 MiB
    L3: 4 MiB
  Speed (MHz): avg: 1905 high: 3001 min/max: 1600/3200 boost: enabled
    cores: 1: 2081 2: 2081 3: 1325 4: 1386 5: 1324 6: 1575 7: 3001 8: 2474
    bogomips: 51103
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3
    svm
Graphics:
  Device-1: AMD Raven Ridge [Radeon Vega Series / Radeon Mobile Series]
    vendor: Lenovo driver: N/A arch: GCN-5 bus-ID: 04:00.0
  Display: x11 server: X.Org v: 1.21.1.7 driver: X: loaded: vesa
    unloaded: fbdev,modesetting,radeon dri: swrast gpu: N/A
    resolution: 1920x1080~77Hz
  API: OpenGL v: 4.5 Mesa 22.3.6 renderer: llvmpipe (LLVM 15.0.6 256
    bits) direct-render: Yes
Audio:
  Device-1: AMD Raven/Raven2/Fenghuang HDMI/DP Audio driver: snd_hda_intel
    v: kernel bus-ID: 04:00.1
  Device-2: AMD ACP/ACP3X/ACP6x Audio Coprocessor driver: snd_pci_acp3x
    v: kernel bus-ID: 04:00.5
  Device-3: AMD Family 17h/19h HD Audio vendor: Lenovo
    driver: snd_hda_intel v: kernel bus-ID: 04:00.6
  API: ALSA v: k6.1.0-13-amd64 status: kernel-api
  Server-1: PulseAudio v: 16.1 status: active
Network:
  Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet
    vendor: Lenovo driver: r8169 v: kernel port: fc00 bus-ID: 01:00.0
  IF: eth0 state: up speed: 1000 Mbps duplex: full mac: <filter>
  Device-2: Intel Wireless-AC 9260 driver: iwlwifi v: kernel
    bus-ID: 02:00.0
  IF: wlan0 state: up mac: <filter>
Bluetooth:
  Device-1: Intel Wireless-AC 9260 Bluetooth Adapter type: USB
    driver: btusb v: 0.8 bus-ID: 4-1.4:3
  Report: hciconfig ID: hci0 rfk-id: 1 state: up address: <filter>
    bt-v: 3.0 lmp-v: 5.1
Drives:
  Local Storage: total: 238.47 GiB used: 21.29 GiB (8.9%)
  ID-1: /dev/nvme0n1 vendor: Samsung model: MZVLB256HAHQ-000L7
    size: 238.47 GiB temp: 37.9 C
Partition:
  ID-1: / size: 233.38 GiB used: 21.27 GiB (9.1%) fs: ext4
    dev: /dev/nvme0n1p2
  ID-2: /boot/efi size: 299.4 MiB used: 20 MiB (6.7%) fs: vfat
    dev: /dev/nvme0n1p1
Swap:
  ID-1: swap-1 type: file size: 512 MiB used: 0 KiB (0.0%) file: /swapfile
  ID-2: swap-2 type: zram size: 13.84 GiB used: 0 KiB (0.0%)
    dev: /dev/zram0
Sensors:
  System Temperatures: cpu: 51.5 C mobo: N/A
  Fan Speeds (RPM): N/A
Info:
  Processes: 252 Uptime: 1h 25m Memory: 14.57 GiB used: 2.96 GiB (20.3%)
  Init: systemd target: graphical (5) Compilers: gcc: 12.2.0 Packages: 2084
  Shell: Bash v: 5.2.15 inxi: 3.3.26

 

[raymac46]

I was afraid of that. You are getting no hardware graphics at all from your AM Vega solution. You have VESA and LLVMPipe graphics. That is OK if you don't need any serious graphical  muscle but it's a terrible waste of a good graphics card.

I'm still recommending hooking up the monitor via VGA to the Intel desktop if your new HDMI-DP cable still gives problems.

[raymac46]

Passive and active DP adapters.

The reason I bring this up is that if you have a passive (no chips) adapter then you must connect to a Thunderbolt/DP++ port on the desktop.

Edited November 2, 2023 by raymac46

[Hedon James]

1 hour ago, raymac46 said:

I was afraid of that. You are getting no hardware graphics at all from your AM Vega solution. You have VESA and LLVMPipe graphics. That is OK if you don't need any serious graphical  muscle but it's a terrible waste of a good graphics card.

I'm still recommending hooking up the monitor via VGA to the Intel desktop if your new HDMI-DP cable still gives problems.

now you're telling me something I don't know.  I don't have a 2nd VGA cable, so I can only run 1 at a time.  When the new HDMI/DP cable comes tomorrow, I'll test it out.  And when I'm testing that out on the Ryzen, I'll move the VGA/DP cable to the Intel.  But I didn't need to nomodeset the Ryzen on that monitor, so I'm hoping I won't have to nomodeset the Intel.  But if I do, I'm guessing the Intel will revert from HD530 to vesa also.  Not necessarily a problem, as it was my backup/safety net machine.  And if the new HDMI/DP cable resolves the Ryzen issues, another cable is only $8 with next day delivery.  I can live with that!

 

I'll check out your Passive/Active DP adapter link.  These came with the Lenovo refurbs, so I'm sure they are the cheapest adapters the seller could get (I'm assuming "passive" no chip is cheaper than a chip adapter).  I'm pretty sure my Lenovo ports are just DP, not DP++, but I'll read up and verify.  Would've been handy knowledge BEFORE this fiasco, but that's my Linux experience....knowledge obtained the hard way!  LOL!

[Hedon James]

Okay, the Lenovo DP ports are all DP++.  My adaptor is passive.  In theory, they shouldn't be a problem.  In reality, something is wrong with that HDMI/DP conversion.  Hoping the single piece cable addresses the issue.  But I'm less confident, now that I know the difference.  Seems like I solved the problem by "downgrading" the graphics driver.

 

But you know what, that's a MUCH more acceptable solution than pkill-ing zombie spawned xclient connections, or increasing the number of allowed connections.  Those are temporary band-aid solutions.  It's not ideal, but running the VGA driver gives me infinite staying power, until I figure out the HDMI/DP issue.  And I still have the LXQT devs looking into things on the LXQT side.  There's still the possibility of a bug in lxqt-config-monitor, or something similar.

 

But in the meantime, I'm functional again, and can move forward with my household migration.  Bottleneck mitigated.  Tomorrow will be interesting.  Fingers crossed that the new cable serves to resolve the issue!

[raymac46]

Well if you're happy, I am happy. This is a lot more difficult issue than we originally thought. I am still not sure it's a hardware issue totally but hopefully it is. AMD has great open source graphics drivers but they can be glitchy. Also I am not that sure if the later AMD hardware really supports VGA anymore. Hence your "downgrade."

Any help I gave was certainly the lame leading the blind, or vice-versa. I have never hooked anything up wth Display Port.

The LXQt guys sound very helpful and knowledgeable so maybe they can help nail this down. As you know I *do* run LXQt but I have a later kernel, my WM is XFWM4 and it's with an ancient Intel graphics solution. I'm not seeing any issues with zombie Xclients and for sure I am using the Intel driver, not VESA/LLVMPipie.