Windows 10: SESSION_HAS_VALID_VIEWS_ON_EXIT before login screen

Hi, all.

This problem's a bit of a pain and likely to be a pain to solve, not least of all because it's wrapped up to do with some dodgy hardware. However, said dodgy hardware had been working mostly fine for 7 months until today, so I hope to at least be able to return to a semblance of working order. I haven't replaced the hardware yet because I'm not sure whether it's the motherboard or graphics card which is responsible and haven't had a chance to swap anything out.

So anyway, the context: today while gaming, I got flickering squares all over the screen and the game stopped. This is the symptom of the dodgy hardware (obviously dodgy because the problem can be reproduced/temporarily fixed by wiggling the graphics card in its slot) What was different today was that after giving the PC a nudge, it went to BSOD with a page fault in atikmdag.sys. I attempted to reboot, but instead of a login screen I get a black screen. However, moving the mouse causes the cursor to appear - for about three seconds, whereupon it disappears again. Moving it continuously yields a cursor most of the time, which every three seconds blinks back to the centre of the screen with a "busy" spinning animation, before shortly turning to a normal pointer.

There are no system restore points, startup repair just fails with a generic error message, resetting windows from the recovery environment got to about 34% before failing with another generic error message, and booting into any kind of safe mode except a bare command prompt yields the same problem as booting normally - a black screen with a mouse cursor disappearing after 3 seconds. A look at the boot log suggests this is because the system's trying to load dxgkrnl.sys repeatedly, and failing - as it appears with "DRIVER_NOT_LOADED" several times at the end of the boot log.

Since the graphics card is very suspect, I disabled it in the BIOS and switched my monitor to the onboard GPU. This is how I get the titular SESSION_HAS_VALID_VIEWS_ON_EXIT BSOD, which is apparently also often caused by a rogue video driver. However, I don't know what's going on: the graphics card is disabled, the AMD driver should just not run. The onboard GPU is not suspect at all. Everything works via either the onboard or separate graphics until the log-in screen or until booting into safe mode.

I've tried running sfc from the command prompt, but it always says there's a system repair pending. I don't know what's going on with that, I suppose it could also be why the other repair operations failed.

Now, since there's a question of hardware damage you might all be tempted to treat this is a lost cause, but I'm hoping someone has a better idea since a) I'm short on cash and b) it was working at least acceptably until today. By that I mean it was perfectly happy running furmark and prime95, or running taxing games at full settings, except occasionally squares would go everywhere and the graphics driver would crash. Therefore I wonder if this latest fit has involved a software issue, too - perhaps the driver went and corrupted itself while everything was going haywire - but I can't be sure. Nonetheless I clearly have fairly restricted options (basically only stuff which can be run from the command prompt from the recovery environment).

Cheers, all.

:)

 

I get error when shutting computer down session_has_valid_views_on_exit then restarts.

I get error when shutting computer down session_has_valid_views_on_exit then restarts.

 

Session_has_valid_views_on_exit error

I have EXACTLY the same problem.

I upgraded my Sony Vaio Ultrabook from Win 8.1 to Win 10 a few days ago, and it seemed fine.

Last night I left it powered on overnight, and this morning I have EXACTLY the same problem.

Microsoft you'd better tell us the solution RIGHT AWAY !!!!!

 

Stop 0xBA is a very rare bugcheck. I haven't noticed much of this. As per my best knowledge, it is any one of these three:

• A display related issue
• A font caused issue
• A system software corruption

We can eliminate all the three possibilities, though.

If it is a display related issue: Remove the graphics card. Apply the display cable to the motherboard's end. Boot the computer in normal mode. Does it work?
If yes, it is a display related issue, and then the originating cause of the issue is to be identified. It may be the graphics card, or the instilled display driver, or anything influencing it.
If it is font related issue: Boot into a bootable Live Linux Distro, say Puppy; or use the Windows 10 Recovery Tools.
Browse to C:\Windows\Fonts, and delete the fonts those you have lately installed before the BSODs started to occur.

Then rebuild the font cache. Go to C:\Windows\System32\ and delete the "Fntcache.dat" file there. Then restart the computer. On restart, a fresh font cache will be built.
If both of the above failed: Accept it as a system software corruption and go for a clean reinstall.
Hope it helps.

 

Thanks for the reply. In the end I decided I was probably dealing with a corrupted driver or something and reinstalled. I hadn't realised it could be font related, but I hadn't done anything to do with fonts recently. In either case, the system is now working as it did before - i.e. generally stable, but with occasional graphical glitches due to the hardware damage.

By the way, for anyone interested, here is a screenshot I took a while ago of what the artifacts look like:



Anyway, that's beside the point. Assuming it was indeed software corruption, though, I'm quite curious: do you have any idea how this might have occurred? I'm working under the assumption that the only component that's damaged is the graphics card, or if the motherboard, only the interface between it and the card - since the only issues that have surfaced prior to this one have been glitches and crashes occurring together with the glitches - never anything like general memory errors (prime95 runs fine for example) or general instability that might lead to something being clobbered on the drives. But this time, a graphical glitch and crash seems to have messed up a system driver - very strange. I'd like to know whether there is any mechanism that a fault with the graphics card/PCI-e slot could nobble stuff on the hard drive, or if I'm actually dealing with more serious damage.

 

The screen is clearly indicating to display problem, which goes with the probable causes of stop 0xBA.

From your post it is appearing that you want to shortlist the problem between a failing graphics card and a failing motherboard. Generally a stress test tells it all, but I doubt that you will be able to run Furmark successfully in this situation. You will need to do elimination tests here.

• Your motherboard has two slots that can take the graphics card. Apply the card in the other one. Does it behave the same? If yes, it eliminates the possibility of a single bad PCIe port.
• Borrow a graphics card from friends/family and apply it to the computer. Does it behave the same? If yes, tend to think that the board/ elements of the board is originating the issue.
• Apply the graphics card to any other computer. Does it behave the same? If so, decide that the card is itself failing.

You cannot decide about the memory based on a Prime95 test. It is designed to test the CPU. To test the memory, use Memtest86+. But I would suggest to leave those till you arrive at a concrete decision about the display.

 

Hi again, thanks to a new job I have enough money to start replacing stuff, but I want to prioritise: graphics card, or motherboard, CPU and RAM first.

I am looking for someone with a spare graphics card to test in my machine, but in case that doesn't work out, I was wondering what people thought was most likely going on here: the machine has clearly sustained damage; joggling the graphics card or the case in general throws artifacts all over the screen. This phenomenon is focused on the graphics card, as mostly what happens is the display driver crashes and, once everything's all sorted, it's just fine from then on.

But then there's the other phenomenon, which has now happened twice: the machine throws a wobbler, BSODs and when it comes back it's unbootable. In both instances I just had to reinstall Windows - SFC and co were no help (though different errors occurred, I suspect the root cause was "system is too boned") The optimistic part of me wants to believe both of these issues stem from the same physical problem: either broken motherboard, broken GPU or broken slot. (BTW my motherboard is mini-ITX so I only have one) but if it's the motherboard, why are the main problems so much focused on graphics? (In fact the smell exactly like broken VRAM to be honest) and if it's the card, how did windows get corrupted?

If anyone can propose a mechanism for this, it would soothe my worries about replacing the card; I don't really want to do that only to have windows break itself again in a month.