VM Ware Fusion potentially causes macOS 10.15.6 to crash

[Alex_Ma]

I recently upgraded to macOS 10.15.6. Now if I leave VM Ware Fusion running while leaving my mac for some time, e.g. - for a night, the system crashes. This did not happen with the older macOS 10.15.5 version.

VM Ware Fusion version: Professional Version 11.5.5 (16269456)

macOS Catalina version: 10.15.6 (19G73)

[JayAllison]

It's working great for me after the patch!

[dariusd]

Hi ktalley_tn,

I decoded your panic log (as much as I could do without proper debugging symbols) and there isn't much to see.  It appears that Google Chrome Helper is in the midst of quitting normally, and while the kernel is handling that operation it encounters memory corruption.  There is no robust evidence to suggest that this is or isn't related to the issues we have seen elsewhere in this thread, but I would say that the balance of probabilities is that it is probably unrelated.  If you are provided with the opportunity to send the report to Apple, I would suggest that you do so.  Check that your McAfee AV is up-to-date.  If in doubt, perhaps consider running hardware diagnostics on your Mac, since the panic report does not rule out the possibility that it resulted from a hardware problem.

Gory details, since we have an above-average complement of kernel people in the thread so I ain't gonna get out of it lightly: Fault is at unix_syscall64➞exit➞exit1➞exit_with_reason➞task_terminate_internal➞ipc_space_terminate➞ipc_right_terminate+193 (inlined: ➞ip_lock➞io_lock➞ipc_object_validate➞io_otype), source location is osfmk/ipc/ipc_object.c line 498, inlined through osfmk/ipc/ipc_right.c line 646, where we find that io=0xffffff8000040203 is an invalid pointer (page not present) when attempting to reference (io)->io_bits, hence the page fault with error code zero.  entry->ie_bits is 0x3020000 which seems valid to my somewhat-superficial understanding, but entry->ie_object (0xffffff8000040203) is clearly invalid... Given that the invalid part is the first few bytes of the entry struct, it might have been overwritten by an overrun from an adjacent object in memory, or might be used after free, or it might be bad memory or even (a remote chance) a defective CPU (in roughly decreasing order of probability).  Memory pressure (such as we saw earlier in this thread) is just one of many potential indirect causes of software-induced write overrun or use-after-free errors.  I don't know what more to make of it.

--

Darius

[alphasupremicus]

Ditto. I ran a large Window VM on Fusion overnight and it's been rock solid with no memory leakage. I think this problem is fixed (finally).

[ktalley_tn]

Thanks for the analysis, @dariusd!  I appreciate you taking the time to look at my panic log screenshot.  After my crash yesterday morning, I tried to clean up my Mac a bit.  As I mentioned, I uninstalled Virtualbox completely, and also uninstalled a few other remote desktop tools (and a few other odd & end apps that I don't use).  I have been running Fusion non-stop since, and have not had anything close to a crash.  That comment may have jinxed the whole deal, but so be it! 

I will report back if I have another crash OR if I can get thru another day or two without incident...

Thanks again, dariusd...

(BTW...  I did run hardware diagnostics a couple of weeks ago, and it reported that all was good...)

[NOMOS]

Do you run MS Teams (for Mac) concurrently  with Fusion?

[rossco_au]

I run Fusion alongside Teams on a daily basis - sometimes multiple VMs and XCode and a simulator while I'm on a Teams call with video so I'm definitely "giving it some stick" as we say here in Oz.  I'm currently on 10.15.5 - haven't upgraded to 10.15.6 yet and applied the supplemental update yet.  Once bitten twice shy.

[blurred_vision]

For what it is worth, I stumbled upon this thread having the same "Unapproved caller. SecurityAgent may only be invoked by Apple software" issue when returning to my machine after the screen saver was running..

While I have Fusion installed, I have not even opened it since upgrading to Catalina.

I *WAS* also having issues with MS Teams randomly soiling itself and hard locking up my machine with the un-resolveable "Unapproved caller. SecurityAgent may only be invoked by Apple software"

To resolve, I needed to reboot...

Advice on memory pressure was heeded (I also had iStats handy, so that was easy) and have seen no issues since I have started rebooting once Memory Pressure registered at the 50% mark...

I will likely wait another 48hrs to monitor people in here's results on the supplemental update, and then install...

[sunsmoke]

Hi dariusd,

(Edit: fixed version number)

I saw the message about the supplemental update for 10.15.6. It's great that there's a potential solution. You said earlier that you had created a reduced test case to provoke the bug for Apple. Have you tried that test case with the supplemental update, and does the test case no longer cause the issue? It would just be nice to see that explicitly.

[dariusd]

Yes, I've tested with the latest Supplemental Update applied – macOS 10.15.6 (19G2021) – and the minimal test case no longer causes any problem.

--

Darius

[bzdura]

I was hoping that latest patch released by Apple will solve my problem described in following post Problem when host (Mac) is going to sleep

Unfortunately problem still exist what is really frustrating -  when my guest system (Kali Linux) is running and I close the host MacBook lid then after waking up my MacBook the guest system Is down.

It's happening in ~70-80% cases - in other 20-30% when I open the lid guest system is up and running what is a standard behaviour on MS VM Workstation

[kpfarrell33]

Hello, I am running fusion on a mac, with 2 ubuntu servers.  I started getting frequent panic/crashes after running smoothly for years.  I upgraded to 10.15.6 and it got slightly better, ran for about 6 or 8 hours and crashed again, and kept crashing.  I turned off one of the ubuntu vms, which lowered my memory pressure to about 50% (from 75-80%).  I have 32 GBs on the server.  I found that if the VM ran at 50% or less, it would not crash. I have since moved the 2nd vm to another computer, running them both independently and all is stable for last 6 days.

It defeats the purpose of running the vms, but it is a work around until this gets resolved, probably with Big Sur.   Seems there are a lot of changes around kernel protection in Big Sur.

I wanted to share, as I tried very very many different approaches.  I was hopeful to see 10.15.6 and how it "addresses vm stability" ... but, it did not for me.

This has worked so far.

Good Luck.

[ColoradoMarmot]

Make sure it's 10.15.6 and the supplemental update (build 19F101)

[Mikero]

My friends, I think it's time to lock this thread.

The original issue was analyzed, troubleshot, and resolved.

Normally I don't mind keeping old threads open, but as folks pile on in the future with 'me too' and 'this didn't fix my issue', there's no sense pinging hundreds of followers here with new updates when the issue itself is resolved through the 10.15.6 supplemental update.

If folks encounter similar issues, we'd like them to start a new thread, as they would be different issues than the one here.

Thanks again to everyone who helped us dig in here and get this sorted out with Apple!