linux-image-5.4.0-37-generic freezes on idle. linux-image-5.4.0-33-generic works fine.

[Emilio]

UPDATE:
linux-image-5.4.0-39-generic Works OK
==========
Original report:
linux-image-5.4.0-33-generic Works OK
linux-image-5.4.0-37-generic Freezes when leaving idle for some time.

Ubuntu 20.04LTS 64 bits. Fully updated.

Hardware:
AMD Ryzen 7 1800X
32GB RAM (two sticks: Part number KHX2666C16/16G)
Asus PRIME A320M-K - latest BIOS installed
AMD Radeon R9 390X 8GB
NVMe SSD 512GB (I don't recall the brand)

Tried AMDGPU and RADEON drivers for the video card. The issue persists.

I was not able to find any logs nor kernel panics on the usual log files. Booting the older kernel makes the machine work perfectly stable as before.

Most of the time I come back to a black screen. Other times is partially dimmed, like when entering into power save mode, when the screensaver kicks in to shut down the monitor.

WORKAROUND:
Disabling the ability to shut down the monitor "fixes" the problem.

In the xscreensaver settings I disabled quick power-off and also disabled power management, so it never tries to power off the monitor. Now the system is stable again.

[Emilio (emilio-moretti)]

After some more testing I verified this only happens when it tries to power off the monitor. I can confirm that disabling the quick monitor power-off and the power management in the screensaver settings makes the kernel work stable. Again, this feature works OK on linux-image-5.4.0-33-generic

[Emilio (emilio-moretti)]

Doing more testing, and now that I see this "feature" disabled, I don't think it always turned off the monitor. Looks like half of the time linux-image-5.4.0-33-generic failed to turn it off, but it still did NOT hang the entire system like linux-image-5.4.0-37-generic does.

Monitor details

$ sudo get-edid | parse-edid
This is read-edid version 3.0.2. Prepare for some fun.
Attempting to use i2c interface
No EDID on bus 1
No EDID on bus 2
No EDID on bus 3
No EDID on bus 4
No EDID on bus 5
No EDID on bus 6
No EDID on bus 7
No EDID on bus 8
2 potential busses found: 0 9
Will scan through until the first EDID is found.
Pass a bus number as an option to this program to go only for that one.
Bus 0 doesn't really have an EDID...
128-byte EDID successfully retrieved from i2c bus 9
Looks like i2c was successful. Have a good day.
Checksum Correct

Section "Monitor"
        Identifier "DELL S2409W"
        ModelName "DELL S2409W"
        VendorName "DEL"
        # Monitor Manufactured week 22 of 2009
        # EDID version 1.3
        # Digital Display
        DisplaySize 530 300
        Gamma 2.20
        Option "DPMS" "true"
        Horizsync 30-83
        VertRefresh 50-76
        # Maximum pixel clock is 170MHz
        #Not giving standard mode: 1152x864, 75Hz
        #Not giving standard mode: 1280x1024, 60Hz
        #Not giving standard mode: 1920x1080, 60Hz
        Modeline "Mode 0" 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync
EndSection

[Emilio (emilio-moretti)]

This bug no longer happens on 5.4.0-39. I turned on power management again and the screen is correctly turned off by xscreensaver