Ubuntu kernel > 4.4.0-51 reboots w/VGA connection

Asked by Tom on 2017-02-13

Short version: All kernels after 4.4.0-51 reboot almost immediately when running with a VGA cable. Earlier kernels are all fine. Changing the monitor cable from VGA to a DVI cable (on the same monitor) solved the problem. Why? Is this a feature or bug or ???

I have an MSI motherboard, A78M-E45 (MS-7721) and a Dell 20" LCD monitor which until today has always been connected with a VGA cable. The motherboard & monitor both support DVI connections but I had used a VGA cable on an even older monitor that only supported VGA and I was just lazy and kept using the same cable even with the Dell monitor.

For years the system has been running just fine with many past versions of Ubuntu & kernel. With Ubuntu 16.04 (current/up-to-date), everything was great through kernel 4.4.0-51.

After kernel 4.4.0-51, each subsequent kernel (4.4.0-57, 4.4.0-59, and now 4.4.0-62) I installed through the standard distribution would almost immediately reboot. I did not get any kern.log or even syslog information so I was at a complete loss to figure it out.

With each of those later kernels that would reboot, I was able to use grub and select the recovery mode of each & every kernel with no problem, and then continue the boot process. The monitor was always locked in a 1024x768 resolution even thought it's native resolution is 1600x1200 but I assumed that was just a function of recovery mode. Finally today I thought I'd try using a DVI cable instead. Immediately every kernel started booting normally without any problems, not requiring recovery mode.

I'm happy that I can now boot normally with the more recent kernels, but I'm at a loss to explain why a VGA cable wouldn't work on the most recent kernels. Perhaps there is some issue with EDID over a VGA cable, but does anyone know enough of what changed after kernel 4.4.0-51 that would cause this, and is there some setting I might change that would allow me to continue to use a VGA cable? I have no issue running with a DVI cable forever, but I'm just bothered that something else is wrong that may come back to haunt me later.

I still have all the kernels mentioned above installed so I can easily use grub to select a different one for testing any theories.

Thanks in advance.

Question information

English Edit question
Ubuntu linux Edit question
No assignee Edit question
Last query:
Last reply:

Does the original kernel boot OK and run all your hardware?

Tom (web-l) said : #2

If by "original" you mean the 4.4.0-51, yes it continues to work fine with the VGA cable. I have not removed that kernel so it's easy to use grub and manually select that one which works fine. Selecting any other kernel causes the "infinite loop of reboots". So the presence of later kernels has not caused earlier kernels to exhibit the problem. Earlier kernels continue to work/boot just fine.

Tom (web-l) said : #3

I don't know if it's relevant but this morning I just saw these messages (while using the DVI cable). I'll have to investigate this further and see if it might be related or not.

Feb 12 14:51:57 thermaltake kernel: [ 1.483676] [drm:radeon_init [radeon]] *ERROR* No UMS support in radeon module!
Feb 12 14:51:57 thermaltake kernel: [ 50.412773] [drm:radeon_init [radeon]] *ERROR* No UMS support in radeon module!

Great, use the web to see what it means

Tom (web-l) said : #5

I tracked down the "No UMS" that ultimate led to a fix. I needed to set radeon.dpm=0 in the GRUB_CMDLINE_LINUX_DEFAULT.

Can someone explain why the kernel updates caused that to be required? Is this a case of, "Oh, of course because <something> changed that became more strict"? I'm glad to have the answer, but I'd like to understand better. Can someone offer an explanation for what changed sometime shortly after kernel 4.4.0-51?

Can you help with this problem?

Provide an answer of your own, or ask Tom for more information if necessary.

To post a message you must log in.