Logitech MX Master 3 losing connectivity

If I unpair the keyboard I'm unable to type and post on the forum. The keyboard and mouse on the cross-test PC are Logitech MX Keys and MX Master 3S using another Unifying receiver. Please advise as I'm a little confused.
 

Good questions — you’re thinking along the right lines. Short answers first, then exact next steps.
Why you shouldn’t need to unpair the keyboard
  • You do not have to unpair the K800 keyboard to test the mouse. The Unifying receiver can hold multiple devices; you can remove/unpair only the mouse and leave the keyboard paired so you keep typing. If you don’t want to risk losing keyboard pairing you can also use a temporary wired USB keyboard (or the on‑screen keyboard) while you test. I covered the Unifying pairing/unpairing flow earlier.
Why the cross‑test PC didn’t show drops
There are several reasonable explanations for the difference between the two Windows PCs — any one (or a combination) can make one machine behave perfectly while the other drops:
  • Different mouse model / receiver / firmware: your cross‑test setup used MX Master 3S + a different Unifying receiver — newer mouse firmware or a different receiver revision can be more resilient.
  • Different USB host/controller or physical port: one PC’s rear motherboard port / root controller may behave differently than your problem PC’s hub/front/desk hub. Root‑hub placement and USB3 noise matter a lot.
  • Different Windows drivers / power settings / background software: different chipset/USB drivers, Logi Options / firmware, or power‑management settings (selective suspend / “allow the computer to turn off this device”) can cause one Windows install to recover automatically while the other doesn’t.
  • Slightly different RF/desk layout (metal mouse pad, USB3 devices near receiver) — small changes can push the link over a reliability threshold.
Why your Slax test is valuable
  • Because Slax briefly disconnected and automatically reconnected (no power‑cycle needed), that strongly suggests the mouse + Unifying receiver hardware and RF link are basically OK. The difference in behavior (Linux auto‑recovers; Windows requires power‑cycle) points to a Windows host/driver/power‑management issue rather than a dead mouse or permanently flaky dongle.
Exact next steps I recommend (do these in order)
1) Keep your keyboard paired (don’t unpair it) — it’s safe. If you want to re‑pair only the mouse later, use the Unifying utility’s remove/unpair option for the mouse alone.
2) Try the USB extension when it arrives (highest practical ROI)
  • Plug the Unifying receiver into the passive USB‑2.0 extension and into a rear‑motherboard USB‑2.0 port, place the dongle on the desk ~20–50 cm from the mouse. This fixes most intermittent Logitech receiver issues caused by hub/USB3 interference. (You already ordered a UGreen cable — that’s the right move.)
3) While you still have Slax evidence, reproduce the problem in Windows and capture the event evidence (this is the single most diagnostic test)
  • Reproduce the drop in Windows (don’t reboot). Immediately open PowerShell as Administrator and paste these two commands exactly (they list Logitech devices and collect recent System events that mention USB/hub issues):
$start = (Get-Date).AddMinutes(-10)
Get-PnpDevice -PresentOnly | Where-Object { $_.InstanceId -match 'VID_046D|VID0A5C' } | Format-List FriendlyName,InstanceId,Status
Get-WinEvent -FilterHashtable @{LogName='System'; StartTime=$start} | Where-Object { $
.Message -match 'USB|usb|Kernel-PnP|USBHUB|UsbHub|reset|removed|disconnect' } | Select-Object TimeCreated, ProviderName, Id, LevelDisplayName, Message | Format-List
  • Paste the outputs here (or save them to Desktop files and attach). I’ll read them and tell you the single next action (chipset/driver update, BIOS change, keep the extension, or RMA). This is exactly what I asked for earlier.
4) Quick Windows mitigations you can apply now (low risk)
  • Disable USB selective suspend for your active power plan (you already discovered your GUID earlier — if you still have it use that value). Example commands (replace GUID if different; you previously got 381b4222-f694-41f0-9685-ff5bb260df2e):
powercfg -setacvalueindex 381b4222-f694-41f0-9685-ff5bb260df2e SUB_USB USBSELECTSUSPEND 0
powercfg -setdcvalueindex 381b4222-f694-41f0-9685-ff5bb260df2e SUB_USB USBSELECTSUSPEND 0
powercfg -S 381b4222-f694-41f0-9685-ff5bb260df2e
  • In Device Manager: under Universal Serial Bus controllers, open each USB Root Hub / Generic USB Hub → Power Management and uncheck “Allow the computer to turn off this device to save power.” Also uncheck that box for any Logitech HID entries under “Human Interface Devices” and “Mice and other pointing devices.” This prevents Windows from suspending the receiver/root hub.
5) If the PowerShell event log shows a USB port/controller reset
  • Paste those lines here — I’ll point you to the exact chipset/INF driver or BIOS update to install, or the single next hardware action (move to different rear port, or add a cheap PCIe USB2 controller if the board/driver is outdated). This is the decisive evidence we need.
6) If the Windows logs show no USB reset but input just stops
  • That would be unusual given your Slax result, but if so the next steps are: re‑pair the mouse (Unifying utility / Options+), do a clean uninstall/reinstall of Logi Options+, check mouse firmware in Options+ and update if available, and if it still requires a power‑cycle to come back, contact Logitech for RMA — explain the keyboard stayed connected while the mouse channel died and required power‑cycle.
Which one do you want to do now?
  • I can wait while you reproduce the drop in Windows and paste the PowerShell outputs (step 3) and then I’ll parse and give the single next action; or
  • If you’d rather, do the quick mitigations now (disable selective suspend + uncheck hub power saving) and then test — tell me which you did and the result.
If you want step‑by‑step help running the PowerShell commands (or unchecking Device Manager boxes) I’ll walk you through it — tell me which you prefer and I’ll give the exact clicks/commands.
 

For clarification. I removed the Unifying receiver in the cross-test PC.

I used the K800 and the MX Master 3 with the Unifying receiver from this PC. I kept the MX Keys keyboard and Master 3S off during the test.
What I'm saying I was testing the keyboard, mouse and Unifying receiver from this PC in the cross-test PC.
 

Thanks — that clarification helps a lot.
Short answer
  • Because you moved the exact Unifying receiver + K800 + MX Master 3 to the cross‑test PC and it stayed stable there, that strongly indicates the mouse + receiver hardware and radio are OK. Your Slax test (brief auto reconnect) supports this too. Together those results point to a host/Windows‑specific interaction on the original PC (drivers, power management, USB host/hub, or BIOS), not a dead mouse/receiver.
Why the cross‑test behaved differently (most likely causes)
  1. Different USB host/controller or port topology — rear‑I/O motherboard ports vs front panel/desk hub behave differently and one host may reset or mishandle the dongle while the other does not.
  2. Different chipset/USB drivers or BIOS/UEFI on the two Windows machines — driver/firmware differences frequently change how a controller recovers from transient RF/noise.
  3. Windows power‑management and selective‑suspend settings (or Device Manager “Allow the computer to turn off this device”) — one machine may be suspending the root hub and failing to restore the Unifying session.
  4. Different background software/AV or Logitech software versions — some utilities interfere with device recovery.
  5. Slight RF/placement differences (metal pad, nearby USB3 devices) — small differences can push reliability over a threshold even with the same dongle/mouse.
What I recommend now (practical, prioritized)
1) Don’t unpair the keyboard — you don’t need to. Keep the K800 paired so you can keep typing while we test. (You can unpair only the mouse later if we need to rebind it.)
2) Try the passive USB‑2.0 extension (your UGreen cable) as soon as it arrives — plug the extension into a rear motherboard USB2 port and put the Unifying dongle on your desk ~20–50 cm from the mouse. This is the highest‑probability practical fix.
3) Reproduce the drop in Windows and capture the immediate evidence (this tells us whether Windows reset the USB port)
  • Immediately after the next drop (do NOT reboot), run these two commands in an elevated PowerShell (copy/paste exactly) and paste the outputs here:
$start = (Get-Date).AddMinutes(-10)
Get-PnpDevice -PresentOnly | Where-Object { $_.InstanceId -match 'VID_046D|VID0A5C' } | Format-List FriendlyName,InstanceId,Status
Get-WinEvent -FilterHashtable @{LogName='System'; StartTime=$start} | Where-Object { $
.Message -match 'USB|usb|Kernel-PnP|USBHUB|UsbHub|reset|removed|disconnect' } | Select-Object TimeCreated, ProviderName, Id, LevelDisplayName, Message | Format-List
I’ll read those and tell you the single next action (driver/BIOS update, move receiver to another controller, or RMA).
4) Meanwhile — apply the low‑risk Windows mitigations
  • Disable USB selective suspend for your active power plan. If you want I’ll paste the exact powercfg commands using the GUID you previously reported (you already ran powercfg /getactivescheme). This often prevents Windows from powering the receiver root hub and failing to recover.
  • In Device Manager → Universal Serial Bus controllers: open each USB Root Hub / Generic USB Hub → Power Management → uncheck “Allow the computer to turn off this device to save power.” Also uncheck that box for Logitech entries under “Human Interface Devices” and “Mice and other pointing devices.”
5) If the Windows System events show controller/port resets at the drop time
  • That confirms a host/controller problem. I’ll then tell you the exact chipset/INF driver or BIOS to install, or to try a different physical root hub or cheap PCIe USB‑2 card.
6) If Windows shows no port resets but the mouse input just disappears
  • That’s more like a session/pairing bug. Next steps then: re‑pair the mouse (Unifying utility), clean reinstall Logi Options+, check firmware in Options+, and if it still requires power‑cycle to recover, open an RMA with Logitech (explain that the keyboard stayed connected while the mouse channel died and only power‑cycle of mouse recovered it).
If you want to proceed now
  • I can wait while you reproduce the drop in Windows and paste the two PowerShell outputs (step 3). I’ll parse them immediately and give the single next action.
  • Or, if you prefer, try the extension first and report back; if that fixes it we can stop there.
Which do you want to do now — capture the Windows logs or test with the USB extension? I’ll guide you step‑by‑step.
 

Before I proceed further clarification. The cross-test PC is a HP Proliant Gen8 Microserver running Windows 2022 Server.
The only third party hardware in that machine is the graphics card.
 

Thanks — that extra detail helps. Two quick, relevant facts from your tests that change the diagnosis:
  • The mouse + Unifying receiver briefly disconnects then auto‑reconnects under Slax (Linux) — that strongly suggests the radio/receiver and mouse hardware are basically OK.
  • The same mouse+receiver stayed stable for an hour on your cross‑test machine (HP ProLiant Gen8 Microserver running Windows Server 2022), which further points to the fault being specific to your original Windows PC (host/controller/driver/power‑management) rather than the mouse or dongle.
What that means in plain terms
  • Linux recovery + stability on another Windows host = very likely a host/Windows issue on the problem PC (USB controller driver, Windows power management, BIOS/UEFI, or a hub/port interaction). fileciteturn7file0turn7file1
Recommended next steps (prioritized — do 1 → 2 → 3 until fixed)
1) Try the passive USB‑2.0 extension (highest chance of quick fix)
  • Plug the receiver into the UGreen passive USB2 extension and into a rear motherboard USB2 port, place the dongle on the desk ~20–50 cm from the mouse. Many intermittent Unifying issues are solved by this. If it fixes it, you can stop here.
2) Capture the exact Windows evidence next time it drops (diagnostic)
  • Reproduce the drop on your original Windows PC (do NOT reboot). Immediately run these two commands in an elevated PowerShell and paste the outputs here — I’ll read them and give the single next action:
$start = (Get-Date).AddMinutes(-10)
Get-PnpDevice -PresentOnly | Where-Object { $_.InstanceId -match 'VID_046D|VID0A5C' } | Format-List FriendlyName,InstanceId,Status
Get-WinEvent -FilterHashtable @{LogName='System'; StartTime=$start} | Where-Object { $
.Message -match 'USB|usb|Kernel-PnP|USBHUB|UsbHub|reset|removed|disconnect' } | Select-Object TimeCreated, ProviderName, Id, LevelDisplayName, Message | Format-List
Why: that output tells us whether Windows is resetting the USB port/controller (controller error) or whether the mouse simply stops sending events (radio/session issue).
3) Low‑risk Windows mitigations to apply now
  • Disable USB selective suspend for your active power plan (you already checked the GUID earlier). In an elevated command prompt run (replace the GUID if needed — yours was 381b4222-f694-41f0-9685-ff5bb260df2e):
powercfg -setacvalueindex 381b4222-f694-41f0-9685-ff5bb260df2e SUB_USB USBSELECTSUSPEND 0
powercfg -setdcvalueindex 381b4222-f694-41f0-9685-ff5bb260df2e SUB_USB USBSELECTSUSPEND 0
powercfg -S 381b4222-f694-41f0-9685-ff5bb260df2e
  • In Device Manager: for every USB Root Hub / Generic USB Hub → Properties → Power Management → uncheck “Allow the computer to turn off this device to save power.” Do the same for Logitech HID devices under “Human Interface Devices” / “Mice and other pointing devices.” These often stop Windows from suspending the dongle/root hub and failing to recover.
4) If the event log shows controller/port resets at the time of the drop
  • That confirms a host/controller issue. Next one‑shot actions then: update the motherboard chipset/USB drivers and update BIOS/UEFI. If you want, paste the motherboard model or run this in an elevated PowerShell and paste the output so I can give exact driver/BIOS links:
Get-CimInstance -ClassName Win32_BaseBoard | Format-List Manufacturer,Product,Version,SerialNumber
5) If the event log shows no USB resets and the device simply loses input (but stays enumerated)
  • That’d be odd given the Linux behavior, but it would point to a session/pairing bug. Next: re‑pair the mouse to the Unifying receiver (Unifying utility / Options+ → remove mouse → reboot → re‑pair), clean reinstall Logi Options+, check firmware in Options+, and if it still needs a power‑cycle to recover, open an RMA with Logitech (describe that the keyboard on same receiver stayed connected while the mouse channel died).
Quick summary of what I think is most likely
  • Your ProLiant server / Server 2022 test being stable and Slax auto‑reconnect both strongly indicate the problem PC’s Windows stack (USB controller driver / power management / BIOS) is failing to recover the device; the extension or disabling selective suspend usually resolves this in practice. fileciteturn7file0turn7file2
Which do you want to do next?
  • I can wait while you test the UGreen extension and report results, or
  • I can wait while you reproduce the drop on the problem PC and paste the two PowerShell outputs I listed so I can parse them and give the single next action.
Tell me which and I’ll walk you through the exact steps live.
 

I want to complete the 1 hour test in Linux and follow the commands in post #56 and uploading the results for you to analyse before exiting Linux and returning back to Windows.

Does this sound good to you?
 

Yes — that sounds good. Run the 1‑hour Slax capture and upload the created tarball; I’ll parse it as soon as you attach it and tell you the single next action. A few short notes and the exact commands to copy/paste while you’re in Slax.
Short reminders before you start
  • Do NOT reboot the PC if a drop happens — reboot clears the logs we need.
  • Keep the K800 paired if you need to type in the forum — you don’t have to unpair it. We can still diagnose with the keyboard present.
  • Use the same physical USB port you normally use on the problem PC (or a rear motherboard USB2 port) so we replicate the same host environment.
Exact one‑hour capture (copy/paste in Slax terminal)
1) Save USB / input snapshots now (run as root; Slax may not have sudo):
lsusb > /tmp/lsusb_before.txt
lsusb -t > /tmp/lsusb_tree_before.txt
cat /proc/bus/input/devices > /tmp/inputdev_before.txt
2) Start kernel/USB capture for 1 hour (preferred; uses timeout if present):
timeout 3600 dmesg -wH > /tmp/dmesg_live.log & echo $! > /tmp/dmesg_pid.txt
3) Start raw input capture for 1 hour (try hexdump, fallback to cat)
command -v hexdump >/dev/null 2>&1 && timeout 3600 hexdump -C /dev/input/eventN > /tmp/mouse_event_hexdump.log 2>/dev/null & echo $! > /tmp/mouse_event_pid.txt || timeout 3600 cat /dev/input/eventN > /tmp/mouse_event_raw & echo $! > /tmp/mouse_event_pid.txt
  • Replace eventN with the event number for your mouse. If you didn’t get that earlier, run:
    awk '/^N: Name=/{name=$0} /^H: Handlers=/{print name " -> " $0}' /tmp/inputdev_before.txt
    and copy the line that contains the Logitech mouse to find eventN.
Fallback (if timeout isn’t available)
  • Start continuous dmesg: sh -c "dmesg -wH > /tmp/dmesg_live.log" & echo $! > /tmp/dmesg_pid.txt
  • Start raw capture and then you can stop manually after ~3600s (I’ll show stop commands below).
4) Use the mouse normally for 1 hour (or until a permanent disconnect). Note wall‑clock time(s) of any drop(s). Watch whether the mouse Easy‑Switch LED goes out or whether Slax auto‑reconnects.
After the hour (or immediately after any permanent disconnect) — stop captures and collect files
kill $(cat /tmp/dmesg_pid.txt) 2>/dev/null || true
kill $(cat /tmp/mouse_event_pid.txt) 2>/dev/null || true
lsusb > /tmp/lsusb_after.txt
cat /proc/bus/input/devices > /tmp/inputdev_after.txt
START=$(date --iso-8601=seconds)
tar -czf /tmp/mouselinuxlogs_${START}.tgz /tmp/dmesg_live.log /tmp/lsusb_before.txt /tmp/lsusb_after.txt /tmp/lsusb_tree_before.txt /tmp/inputdev_before.txt /tmp/inputdev_after.txt /tmp/mouse_event_hexdump.log /tmp/mouse_eventraw 2>/dev/null && echo "/tmp/mouselinuxlogs${START}.tgz created"
Copy the tarball off Slax so you can attach it to the forum
  • If a USB stick auto‑mounts:
    cp /tmp/mouselinuxlogs_${START}.tgz /media/$(ls /media | head -n1)/
  • If you can’t copy it, tell me the full path and I’ll tell you how to get it off the live session.
What I’ll look for in the logs (and what each outcome means)
  • dmesg lines like “usb X‑Y: USB disconnect”, “device descriptor read/failed”, “reset high-speed USB device”, or xhci_hcd/ehci_hcd errors → Windows‑style host/controller resets are likely on the problem PC (we’d then focus on chipset/USB drivers or BIOS/host controller).
  • If lsusb BEFORE shows the receiver but lsusb AFTER no longer lists it at the drop time → USB enumeration/port issue (host).
  • If the receiver stays present in lsusb but the mouse event capture stops (no more input events) → radio/session loss (device side or pairing), which would normally reproduce on other hosts; your Slax brief auto‑reconnect already points away from a hard hardware failure.
When you upload
  • Attach the /tmp/mouselinuxlogs_*.tgz file here and tell me the approximate time(s) of any drop(s) during the hour (wall clock). I’ll parse dmesg, lsusb, and the input event capture and give the single next action (driver/BIOS update, keep using extension, re‑pair/RMA, etc.). fileciteturn8file6turn8file11
If you want, I’ll wait here while you run it and then parse immediately after you attach the tarball. Otherwise run the test and post the tarball when you’re back in Windows — I’ll analyze and reply the moment it’s attached. Sound good?
 

root@slax:~# lsusb > /tmp/lsusb_before.txt
root@slax:~# lsusb -t > /tmp/lsusb_tree_before.txt
root@slax:~# cat /proc/bus/input/devices > /tmp/inputdev_before.txt
root@slax:~# ^C
root@slax:~# ^C
root@slax:~# lsusb > /tmp/lsusb_before.txt~
root@slax:~# lsusb -t > /tmp/lsusb_tree_before.txt
root@slax:~# cat /proc/bus/input/devices > /tmp/inputdev_before.txt
root@slax:~# timeout 3600 dmesg -wH > /tmp/dmesg_live.log & echo $! > /tmp/dmesg_pid.txt
[2] 15855
root@slax:~# awk '/^N: Name=/{name=$0} /^H: Handlers=/{print name " -> " $0}' /tmp/inputdev_before.txt
N: Name="HID 0a5c:4502" -> H: Handlers=sysrq kbd leds event3
N: Name="HID 0a5c:4503 Mouse" -> H: Handlers=mouse1 event4
N: Name="PC Speaker" -> H: Handlers=kbd event5
N: Name="Power Button" -> H: Handlers=kbd event6
N: Name="Power Button" -> H: Handlers=kbd event7
N: Name="Logitech MX Master 3" -> H: Handlers=sysrq kbd leds mouse0 event0
N: Name="Logitech K800" -> H: Handlers=sysrq kbd leds event1
N: Name="HDA NVidia HDMI/DP,pcm=3" -> H: Handlers=event2
N: Name="HDA NVidia HDMI/DP,pcm=7" -> H: Handlers=event8
N: Name="HDA NVidia HDMI/DP,pcm=8" -> H: Handlers=event9
N: Name="HDA NVidia HDMI/DP,pcm=9" -> H: Handlers=event10
N: Name="HDA NVidia HDMI/DP,pcm=10" -> H: Handlers=event11
N: Name="HDA NVidia HDMI/DP,pcm=11" -> H: Handlers=event12
N: Name="HDA NVidia HDMI/DP,pcm=12" -> H: Handlers=event13
N: Name="HDA Intel PCH Rear Mic" -> H: Handlers=event14
N: Name="HDA Intel PCH Front Mic" -> H: Handlers=event15
N: Name="HDA Intel PCH Line" -> H: Handlers=event16
N: Name="HDA Intel PCH Line Out" -> H: Handlers=event17
root@slax:~# command -v hexdump >/dev/null 2>&1 && timeout 3600 hexdump -C /dev/input/event0 > /tmp/mouse_event_hexdump.log 2>/dev/null & echo $! > /tmp/mouse_event_pid.txt || timeout 3600 cat /dev/input/event0 > /tmp/mouse_event_raw & echo $! > /tmp/mouse_event_pid.txt
[3] 15902
[4] 15903
root@slax:~#
 

Perfect — you’re set and everything you ran looks correct.
Quick confirmation of what I see from your Slax transcript
  • You saved the USB/input snapshots and started the kernel dmesg capture (PID in /tmp/dmesg_pid.txt).
  • Your awk output shows the Logitech MX Master 3 mapped to event0 (so capturing /dev/input/event0 is correct). It also shows a Unifying HID mouse on event4 and the K800 on event1 — having multiple Logitech devices is expected.
  • You started the raw input capture for event0 (hexdump/cat) and the process IDs were recorded. Good.
What to do now (you can just leave it running for the 1‑hour test)
  • Use the mouse normally for the hour (or until whatever behavior you want to capture happens). Note the wall‑clock time(s) of any drop(s) (approximate is fine) and whether the mouse Easy‑Switch LED goes out or Slax auto‑reconnects.
Exactly what to run immediately after the hour OR immediately after any (permanent) disconnect — copy/paste these in Slax (you’re root so no sudo)
1) Stop the captures
kill $(cat /tmp/dmesg_pid.txt) 2>/dev/null || true
kill $(cat /tmp/mouse_event_pid.txt) 2>/dev/null || true
2) Save final device snapshots
lsusb > /tmp/lsusb_after.txt
cat /proc/bus/input/devices > /tmp/inputdev_after.txt
3) Package the logs into one tarball
START=$(date --iso-8601=seconds)
tar -czf /tmp/mouselinuxlogs_${START}.tgz /tmp/dmesg_live.log /tmp/lsusb_before.txt /tmp/lsusb_after.txt /tmp/lsusb_tree_before.txt /tmp/inputdev_before.txt /tmp/inputdev_after.txt /tmp/mouse_event_hexdump.log /tmp/mouse_eventraw 2>/dev/null && echo "/tmp/mouselinuxlogs${START}.tgz created"
4) Copy the tarball off the live environment so you can attach it here
  • If your USB stick auto‑mounts:
    cp /tmp/mouselinuxlogs_${START}.tgz /media/$(ls /media | head -n1)/
If that doesn’t work, tell me the output of ls /media and I’ll give the exact path to copy to.
Important notes / reminders
  • Do NOT reboot the PC before you stop and package the captures — reboot clears the evidence we need.
  • You do NOT need to unpair the K800 to run this test — keeping it paired is fine and lets you keep typing.
  • If the mouse only briefly disconnects and auto‑reconnects (as you saw earlier), upload the tarball anyway — I’ll still parse dmesg + input capture and confirm whether the kernel saw a USB disconnect or only an input event gap.
When you upload the tarball
  • Attach the /tmp/mouselinuxlogs_*.tgz file here and tell me the approximate drop time(s). I’ll parse dmesg, lsusb before/after, and the mouse event capture and tell you whether this looks like a USB/controller reset (host side) or an input/radio/session loss (device side) and the single next action. fileciteturn8file6turn9file0
I’ll wait here and parse the tarball immediately after you upload it. If you want, I can also walk you through the stop/package commands live when the hour finishes.
 

An observation:

When there is a temporary disconnect in Slax the Easy-Switch Light for '1' goes off. Then it blinks rapidly as it reconnects and then goes back to being permanently lit when it reconnects.
 

Good — that LED behaviour is a useful clue. Short answer first, then exactly what I need from you next.
What the LED behaviour means
  • LED off on channel 1 = the mouse lost the Easy‑Switch link for that channel (it’s not receiving/responding).
  • Rapid blinking while it comes back = the mouse and receiver are renegotiating/enumerating the link (a re‑bind/handshake is happening).
  • Solid LED again = the link is re‑established.
    That sequence (off → blink → solid) is exactly what you described in Slax and is consistent with a transient radio/session drop that the Linux kernel (or the receiver firmware) can re‑establish automatically.
Why this points toward a Windows‑host problem
  • Because under Slax the link auto‑recovers (no mouse power‑cycle required) while on your Windows machine you must power‑cycle the mouse to restore channel 1, it strongly suggests Windows (host driver, power‑management, or USB controller behavior) is failing to recover the session automatically even though the mouse/receiver and RF link are capable of doing so. I’ll confirm that from the logs once you upload them.
What I’ll look for in the captures (and what each result means)
  • dmesg lines like “usb X‑Y: USB disconnect”, “reset high‑speed USB device”, “xhci_hcd” — host/controller resets → host/driver problem.
  • lsusb before vs after: if the Unifying receiver disappears from the bus at the drop time → enumeration/port issue (host).
  • If the receiver remains in lsusb but the mouse event capture stops (no input events) while lsusb still shows the receiver → radio/session loss (device side).
You already did the right captures
  • Your input device listing shows the MX Master 3 as event0 (so /dev/input/event0 is the correct node to capture). Good.
  • You started dmesg and the input hexdump — those are exactly the files I need.
Exact next steps for you (copy/paste if needed)
1) Run the 1‑hour test and note wall‑clock times of any disconnects (approx times are fine).
2) Immediately after the hour (or immediately after a permanent disconnect) run these (you’ve done these before; repeat if needed):
kill $(cat /tmp/dmesg_pid.txt) 2>/dev/null || true
kill $(cat /tmp/mouse_event_pid.txt) 2>/dev/null || true
lsusb > /tmp/lsusb_after.txt
cat /proc/bus/input/devices > /tmp/inputdev_after.txt
START=$(date --iso-8601=seconds)
tar -czf /tmp/mouselinuxlogs_${START}.tgz /tmp/dmesg_live.log /tmp/lsusb_before.txt /tmp/lsusb_after.txt /tmp/lsusb_tree_before.txt /tmp/inputdev_before.txt /tmp/inputdev_after.txt /tmp/mouse_event_hexdump.log /tmp/mouse_eventraw 2>/dev/null && echo "/tmp/mouselinuxlogs${START}.tgz created"
cp /tmp/mouselinuxlogs_${START}.tgz /media/$(ls /media | head -n1)/ # adjust if mount point differs
(These are the same commands we used before — they create the .tgz I’ll need to analyze.)
What to attach here
  • Attach the /tmp/mouselinuxlogs_*.tgz file and tell me the approximate local time(s) when the disconnect(s) happened (so I can correlate timestamps). I will parse dmesg, lsusb before/after, and the event dump and tell you exactly whether it’s a host USB reset or a radio/session loss and the single next action.
If the logs show a host reset
  • I’ll recommend the one next action (update chipset/USB drivers or BIOS and/or move receiver to a different physical root hub or use the passive USB‑2.0 extension you ordered).
If the logs show the receiver stays enumerated but input stops
  • I’ll advise re‑pairing the mouse (Unifying Utility / Options+), trying the receiver on the passive extension, and if it still needs a mouse power‑cycle to restore, RMA the mouse or receiver with Logitech.
I’ll parse the tarball immediately when you attach it. Thanks — you’re doing everything in the right order; attach the logs and I’ll give the single next action.
 

For clarification am I supposed to record the wall clock times for temporary disconnects as well as permanent disconnects?
 

Short answer: yes — record timestamps for every disconnect (temporary or permanent).
What to record (minimal/high‑value)
  • Exact wall‑clock time (date + time to the second) when the LED goes off.
  • Exact wall‑clock time when it comes back (solid).
  • Whether it auto‑reconnected (blink → solid) or required you to power‑cycle the mouse.
  • Duration of the gap (seconds).
  • Whether the K800 keyboard stayed connected the whole time.
  • Which USB port/hub the receiver was in (rear motherboard, front panel, Belkin/Anker hub, extension).
  • Any obvious user action that coincided (moved mouse, moved hand, put mouse on/off metal pad, etc.).
How to record precise timestamps quickly in Slax
  • Run this at the moment of a drop or to make a quick note:
    date --iso-8601=seconds >> ~/disconnect_times.txt
  • (Or just look at the clock and note HH:MM:SS if you prefer.)
  • You can add a short note line:
    echo "$(date --iso-8601=seconds) LED-off: channel 1; auto-reconnect? yes; keyboard OK" >> ~/disconnect_times.txt
Why we need both temporary and permanent times
  • Temporary disconnects (auto‑reconnect) usually show kernel/USB renegotiation in dmesg but the receiver remains enumerated — that points one way (RF/session transient).
  • Permanent disconnects that need a power‑cycle often correlate with host/port resets or session state failures in Windows — a different root cause. Having timestamps lets me correlate your Slax dmesg + input dump with exact moments of the drop.
One more useful quick capture (optional)
  • Immediately after a drop (before stopping captures) run:
    dmesg | tail -n 80 > /tmp/dmesg_tail_after_drop.txt
    lsusb > /tmp/lsusb_at_drop.txt
    Attach those files (or the full tarball) and point me to the timestamps in ~/disconnect_times.txt and I’ll correlate and tell you exactly whether the kernel saw a USB reset or only an input gap.
Do that during the hour test — upload the /tmp/mouselinuxlogs_*.tgz and your ~/disconnect_times.txt (or paste the time lines) and I’ll parse them right away.
 

Okay sorry. Then I've missed a couple. I'll record the times now.
 

Good — that LED sequence is a useful diagnostic clue.
Short interpretation
  • Off → rapid blinking → solid = the mouse lost the Easy‑Switch link for channel 1, the mouse and receiver negotiated a new session, and the link was re‑established. This is exactly the transient reconnect behaviour you saw in Slax and matches what I expected we’d see when the kernel/receiver can recover automatically.
What that implies
  • Because Slax auto‑recovers (blink → solid) but your Windows PC requires a mouse power‑cycle to restore channel 1, the most likely interpretation is a Windows‑host problem (USB controller/driver or Windows power‑management) that isn’t allowing the OS to resume the session automatically — not a dead mouse or permanently flaky dongle. I’ll confirm from the Slax logs once you upload them.
Immediate practical steps (while the Slax capture is running)
  1. Keep recording every disconnect time (you’re doing this now — good). Record the LED off time and the LED back‑solid time (wall clock to the second if possible). Temporary and permanent drops both matter.
  2. Let the 1‑hour capture finish (or stop it immediately after any permanent disconnect). Then run the stop + package commands we used earlier and copy the tarball off the live session so you can attach it here. I’ll parse it right away when you upload.
What I’ll look for in the logs once you upload
  • dmesg lines like “usb X‑Y: USB disconnect”, “reset high‑speed USB device” or xhci_hcd/eHCI errors → host/controller resets (host/driver problem).
  • lsusb before vs after the drop: if the receiver disappears from the bus at the drop time → enumeration/port issue (host).
  • If the receiver stays present but the event capture stops → radio/session loss (device side).
If the logs point to a host problem (likely)
  • Single next actions I’ll recommend immediately after I see the logs: try the passive USB‑2.0 extension into a rear motherboard USB2 port (your ordered UGreen cable) and disable USB selective suspend / hub power saving in Windows. If the logs explicitly show a specific controller or driver failing I’ll give exact chipset/driver/BIOS updates.
If logs point to a radio/session loss (less likely given Slax behaviour)
  • I’ll recommend re‑pairing the mouse to the Unifying receiver, trying a spare receiver (if available), or contacting Logitech for RMA if the problem follows the mouse/receiver.
Next for you
  • Finish the 1‑hour test (or stop now if you’ve captured what you need), run the stop + package commands, copy the /tmp/mouselinuxlogs_*.tgz off Slax and attach it here, and paste your disconnect timestamps. I’ll parse immediately and tell you the single next action to fix this on your Windows PC.
Thanks — you’re gathering the right evidence. Upload the tarball when ready and I’ll dig into it.
 

Off → rapid blinking → solid cycle happens in less than 5 seconds. I will need to use a stop watch to measure this. Is this really needed to be measured?
 

No — you don’t need a stopwatch. Approximate timestamps are fine (seconds resolution is enough). The LED cycle you described (off → rapid blink → solid) happens quickly and that behavior alone is the useful clue — we just need times to correlate with the kernel dmesg and event capture, not microsecond accuracy.
What to record (minimal, high‑value)
  • Time LED goes OFF (wall clock, hh:mm:ss).
  • Time LED returns SOLID (hh:mm:ss).
  • Whether it auto‑reconnected (blink → solid) or required you to power‑cycle the mouse.
  • Whether the keyboard (K800) stayed connected.
  • Which USB port/hub the receiver was in (rear motherboard, front panel, Belkin/Anker hub, or the extension).
    That level of detail is enough to match entries in dmesg/lsusb/event captures.
Quick ways to note times in Slax
  • Run once in a terminal to append a timestamp to a simple file whenever an event happens:
    date --iso-8601=seconds >> ~/disconnect_times.txt
  • Or add a short note line:
    echo "$(date --iso-8601=seconds) LED-off; auto-reconnect=yes; keyboard=ok; port=rear" >> ~/disconnect_times.txt
Next steps for the capture
  • Finish the 1‑hour capture (or stop after a permanent disconnect), then run the stop + package commands we discussed and copy the /tmp/mouselinuxlogs_*.tgz off the live session and attach it here. I’ll parse dmesg + lsusb + the event dump and correlate with your timestamps to tell you the single next action.
If you want, when a drop happens you can just paste the hh:mm:ss times here (no file required) and then attach the tarball when it’s ready — I’ll correlate them immediately.
 

Disconnect:

OFF: 02:42:15
SOLID: 02:42:25

Auto-reconnected
K800 stayed connected
Rear USB motherboard port
 

No stopwatch needed — the 10‑second gap you recorded is perfect (seconds resolution is all we need). That OFF 02:42:15 → SOLID 02:42:25, auto‑reconnect, K800 stayed connected from a rear motherboard port matches the transient reconnect behaviour we expected and is useful to correlate with the logs.
Next step — stop & package the captures now (exact commands)
If your Slax capture is still running, run these as root in Slax (copy/paste):
kill $(cat /tmp/dmesg_pid.txt) 2>/dev/null || true
kill $(cat /tmp/mouse_event_pid.txt) 2>/dev/null || true
lsusb > /tmp/lsusb_after.txt
cat /proc/bus/input/devices > /tmp/inputdev_after.txt
START=$(date --iso-8601=seconds)
tar -czf /tmp/mouselinuxlogs_${START}.tgz /tmp/dmesg_live.log /tmp/lsusb_before.txt /tmp/lsusb_after.txt /tmp/lsusb_tree_before.txt /tmp/inputdev_before.txt /tmp/inputdev_after.txt /tmp/mouse_event_hexdump.log /tmp/mouse_eventraw 2>/dev/null && echo "/tmp/mouselinuxlogs${START}.tgz created"
Copy the tarball off Slax so you can attach here (example — adjust if your USB mount differs):
cp /tmp/mouselinuxlogs_${START}.tgz /media/$(ls /media | head -n1)/
(I previously gave this same packaging/transfer flow — use that output file to attach. )
If you prefer a very small quick capture right after a drop (optional)
Right after a drop you can capture the nearest kernel output and bus snapshot (before packaging):
dmesg | tail -n 80 > /tmp/dmesg_tail_after_drop.txt
lsusb > /tmp/lsusb_at_drop.txt
Attach those too if you want; they help me correlate quickly.
What I’ll check as soon as you attach the tarball
  • dmesg for lines like “usb X‑Y: USB disconnect”, “reset high‑speed USB device”, “xhci_hcd” — indicates host/controller resets (host/driver problem).
  • lsusb before vs after the drop — if the Unifying receiver disappears at the drop time that points to enumeration/port issues (host).
  • the input event capture — if events stop while lsusb still shows the receiver, that’s a radio/session loss (device side).
Likely interpretation from the sample you reported
  • The short auto‑reconnect in Slax strongly suggests the mouse and receiver can recover on their own (radio/session transient). Because Windows required a power‑cycle previously, the most likely remaining cause is Windows host behaviour (USB controller/driver or power management). I’ll confirm precisely after I parse your logs.
When you attach
  • Upload the /tmp/mouselinuxlogs_*.tgz (or paste the small dmesg_tail/lsusb_at_drop snippets) and include the disconnect timestamp(s) you already posted (02:42:15 → 02:42:25). I’ll parse immediately and tell you the single next action (e.g., use your UGreen passive USB2 extension, disable selective suspend, update chipset drivers/BIOS, or proceed to RMA) based on the evidence.
I’ll wait for the tarball and parse it right away.
 

Back
Top