Kernel Panic on x64 ISO

[VastOne]

Since the debian kernel reached the 4.5.0-1 and 4.5.0-2 releases in sid, I have not been able to build a bootable x64 ISO ... the image below shows the kernel panic that is happening


 


As you can see from the google scans this is a known issue dating back several years.  I have tried so many of the suggestions I have lost count and lost my desire to troubleshoot it even more (been at this for 3 weeks now)

I need help...

The ISO is on this Download Test Page ... Thanks for the help

[Snap]

First thing springing in my head.... A workaround. Why not stuck to a LTS kernel for a while into the isos?

[PackRat]

^ This - or wait for 4.6 to make it to sid.

The 4.5 kernel may be a dog - I have it running on both machines (Debian Sid and Void 64-bit) and it throws a usb error (related to power apparently) on each; throws the same error if I compile my own. On my laptop, there is also a tendency to go to a blank screen and apparently dead keayboard on logout. I though it may be systemd related so I switched to Void, same issue.

You may want to stay with the last 4.4 kernel in the repos.

[VastOne]

Thanks for the advice and input.. there will not be a release until this is resolved.

I understand these 'glitches' when using sid and latest kernels, what gets me is the x32 working fine.. that makes no sense to me

[VastOne]

This seems to be the exact same issue on Arch from last winter

I do build on a AMD Phenom II X4 965 Processor and the issue is all about AMD

There is a patch I may try, or I might fsarchive the x64 up and send it to one of you Intel users to try to build and recreate the issue to confirm

[Snap]

Another AMD (FX-8350) user here. Sorry.

PS, Thinking of it... I've never had such kernel panic and I use VBox daily with some machines currently having 4.5 kernels (though I usually go LTS). Maybe the issue is processor specific and not affecting all AMD series?

PS2, That last winter's kernel 4.0 issues affected my VMs but it was related to the Xserver, maybe the CPU too. Booting without X and replacing the kernel solved the issues back then. That's why my first thought was replacing the current kernel for a known working LTS.

PS3, Wait a minute. My Vsido VM recently went south. After an update the X server stopped starting anymore. It was the only (rolling) VM having this issue while updating all my VMs that day. I tend to think that is a similar issue like that former 4.0 as you said before. for some reason VBox, the X and the kernel don't play well, but why only in Vsido? Maybe because the Guest additions are pre-installed? Yesterday I've borked a fresh PCLinuxOS LXQt VM trying to install the Guest Additions. So long Xserver. Not working anymore.  Just thinking out loud.

[PackRat]

why only in Vsido?

I'm thinking that should be "why only in Debian Sid?" - do you have any other Sid spins in VM?

And just to close my loop - both my computers are intel i3 with intel graphics. Last time I had any similar issue was when the kernel team was re-coding the drivers for the SD card reader on my laptop. Took a couple incremental releases to sort it out.

My retired desktop is an old AMD dual core, I'll try the iso linked above later if I have some time. The md5sum checks out so your 64-bit iso is good.

[VastOne]

Couple of things...

The current ISO's for anyone to get and use are not affected... This is only stopping me from releasing a new ISO and only in x64

In that Arch thread the main supplier of the issue was also using a AMD Phenom II X4 965, the same thing I have... I could try to build on another AMD machine I have that is a Phenom 955

For me this issue will be resolved only when the kernel is fixed and that may not be until 4.6 is released

Thanks for the input.. 

[PackRat]

An upgrade for kernel 4.5.0-2 and systemd came through on my system this morning.

Limited use so far, but I have been able to exit back to the login manager from xfce, fluxbox, and dwm without getting the blank screen or dead keyboard.

I still get the usb error:

Code Select Expand

usb 1-1.4: device descriptor read/64, error -110

but that appears to be a separate issue. I have been using the 4.4.8 kernel (custom compile) and get that error, but no issues exiting.

@vastone - maybe your fix has come through too?

[VastOne]

Nyet... Built and tested as soon as this came through last night and the same failure is there