"The instruction at 0x... referenced memory at 0x0. The memory could not be read." on boot

Discussions about using Windows guests in VirtualBox.
Post Reply
Bats
Posts: 2
Joined: 5. May 2019, 03:41

"The instruction at 0x... referenced memory at 0x0. The memory could not be read." on boot

Post by Bats »

I'm running a Win 7x64 SP1 guest (+guest additions) in a freshly updated VirtualBox 5.2.28 on top of a (suspiciously similar) Win 7 x64 SP1 host - or trying to, at any rate.

Things were working fine the last time I spun it up a few weeks back (and, aside from a few browser updates, there'd been no significant hardware or software changes in the meantime), but today I can't boot up the guest without bombing and returning an application error reading (as you may have guessed from the subject) "The instruction at 0xda07a08f referenced memory at 0x00000000. The memory could not be read.". The second address is always 0x0 while the first has changed a few times in the course of my troubleshooting, but the log always seems to suggest that it's pointing at C:\Windows\system32\atig6txx.dll.

After running into the problem I updated both VBox and the host's video drivers (both of which were substantially out of date) but neither made a difference (not entirely surprising, as things had been working previously), nor did uninstalling & reinstalling the guest additions package. The two things that do clear up the problem are disabling 2D & 3D acceleration (not entirely surprising, with the error giving dirty looks at the ATI drivers) or booting the guest in safe mode, but obviously neither is much of a solution in itself.

So... any guesses what I broke? Or, better still, how I might unbreak it?


-Bats
(if there's any other info I can provide, just ask)
Attachments
VBox Logs.zip
VBox.log + VBoxHardening.log
(52.72 KiB) Downloaded 20 times
socratis
Site Moderator
Posts: 27330
Joined: 22. Oct 2010, 11:03
Primary OS: Mac OS X other
VBox Version: PUEL
Guest OSses: Win(*>98), Linux*, OSX>10.5
Location: Greece

Re: "The instruction at 0x... referenced memory at 0x0. The memory could not be read." on boot

Post by socratis »

I'm afraid it's still your ATI drivers. From the VBox.log:
00:00:01.684297 supR3HardenedErrorV: supR3HardenedScreenImage/NtCreateSection:
        rc=VERR_BAD_EXE_FORMAT fImage=1 fProtect=0x7 fAccess=0x2
        \Device\HarddiskVolume2\Windows\System32\drivers\atikmdag.sys:
Do NOT send me Personal Messages (PMs) for troubleshooting, they are simply deleted.
Do NOT reply with the "QUOTE" button, please use the "POST REPLY", at the bottom of the form.
If you obfuscate any information requested, I will obfuscate my response. These are virtual UUIDs, not real ones.
Bats
Posts: 2
Joined: 5. May 2019, 03:41

Re: "The instruction at 0x... referenced memory at 0x0. The memory could not be read." on boot

Post by Bats »

That was my suspicion. In addition to the logs pointing in that direction, I've also had a long-running but highly intermittent bluescreen (a couple times a year, with no obvious repro) that also points at atikmdag.sys, but I'm at something of a loss as to actually fixing it (fixing either issue, really, but the VirtualBox crash specifically - the BSOD is strictly a host-side headache).

The obvious route - a clean driver install, and experimenting with various driver versions - didn't make much of a difference (although I did discover some completely unrelated problems with the latest version), so I'm open to other suggestions.

The other interesting twist is that I ended up discovering (some time after posting) that the crash only happens when the guest window is on one particular monitor - and it's the one being run off the internal video (Intel HD 4600), not either of the screens on the ATI card (AMD Radeon HD 7970). If I boot up the guest on either of the ATI/AMD-driven monitors and then drag it over to the Intel one it'll work for a while, but certain actions - most reliably, opening Windows Experience Index or the System panel (which pings WEI) - will trigger the same crash, making it puzzling that the ATI driver would be at fault.

Not that I find it unlikely, mind you - just puzzling.

-Bats
Post Reply