[Bug]: RTMemCacheAllocEx SIGSEGV in VBoxRT.so

[DEHughes]

Version

7.2.6

Host OS Type

Linux

Host OS name + version

Snipe5 Leap 16.0

Host Architecture

amd64

Guest OS Type

Windows

Guest Architecture

x86

Guest OS name + version

Wren11B Windows 11

Component

Host Support

What happened?

After approximately 46 hours and 34 minutes of continuous operation, the VirtualBoxVM host process
(PID 1184861) terminated with SIGSEGV. The fault occurred inside RTMemCacheAllocEx in
VBoxRT.so during an async virtual-disk write triggered by PDMR3BlkCacheFlush. The guest
Windows 11 VM was running unattended at the time. No prior BSOD or guest-side fault was logged;
the guest simply lost its host process and became unresponsive.

The core dump is available, but will not upload. Here is a summary:

root@Snipe5:~ > coredumpctl info 1184861 2>/dev/null | head -40
PID: 1184861 (VirtualBoxVM)
UID: 0 (root)
GID: 0 (root)
Signal: 11 (SEGV)
Timestamp: Tue 2026-06-09 21:30:53 EST (10h ago)
Command Line: /usr/lib/virtualbox/VirtualBoxVM --comment Wren11B --startvm 9cb5ceaa-6625-4a64-94e1-776164f35999 --no-startvm-errormsgbox
Executable: /usr/lib/virtualbox/VirtualBoxVM
Control Group: /user.slice/user-0.slice/session-3.scope
Unit: session-3.scope
Slice: user-0.slice
Session: 3
Owner UID: 0 (root)
Boot ID: 72c55451402543cd97646b0901cfcdf8
Machine ID: 3ddf1048c1de40209eb5133578edd6ae
Hostname: Snipe5
Storage: /var/lib/systemd/coredump/core.VirtualBoxVM.0.72c55451402543cd97646b0901cfcdf8.1184861.1781058653000000.zst (present)
Size on Disk: 36.2M
Message: Process 1184861 (VirtualBoxVM) of user 0 dumped core.

            Stack trace of thread 1184926:
            #0  0x00007fb30feb16bb RTMemCacheAllocEx (VBoxRT.so + 0xb16bb)
            #1  0x00007fb30feb1930 RTMemCacheAlloc (VBoxRT.so + 0xb1930)
            #2  0x00007fb304b35759 n/a (VBoxDDU.so + 0x13759)
            #3  0x00007fb304b77d33 n/a (VBoxDDU.so + 0x55d33)
            #4  0x00007fb304b38c11 n/a (VBoxDDU.so + 0x16c11)
            #5  0x00007fb304b34f01 n/a (VBoxDDU.so + 0x12f01)
            #6  0x00007fb304b382e1 n/a (VBoxDDU.so + 0x162e1)
            #7  0x00007fb304b39e44 n/a (VBoxDDU.so + 0x17e44)
            #8  0x00007fb304b44a91 VDAsyncWrite (VBoxDDU.so + 0x22a91)
            #9  0x00007fb2cf6c77a9 n/a (VBoxDD.so + 0xc77a9)
            #10 0x00007fb304510f88 n/a (VBoxVMM.so + 0x110f88)
            #11 0x00007fb304511070 n/a (VBoxVMM.so + 0x111070)
            #12 0x00007fb3045131a5 PDMR3BlkCacheFlush (VBoxVMM.so + 0x1131a5)
            #13 0x00007fb2cf6c767c n/a (VBoxDD.so + 0xc767c)
            #14 0x00007fb2cf6cf028 n/a (VBoxDD.so + 0xcf028)
            #15 0x00007fb2cf78f8ee n/a (VBoxDD.so + 0x18f8ee)
            #16 0x00007fb30452797a n/a (VBoxVMM.so + 0x12797a)
            #17 0x00007fb30ffb70dc n/a (VBoxRT.so + 0x1b70dc)
            #18 0x00007fb310073ddd n/a (VBoxRT.so + 0x273ddd)
            #19 0x00007fb310281ed9 start_thread (libc.so.6 + 0x91ed9)

How can we reproduce this?

The crash was not manually triggered. Conditions at time of crash:

1. Start a Windows 11 guest with the configuration described above.
2. Allow the VM to run unattended for an extended period (crash occurred at ~46 h).
3. The guest has Vipre antivirus (SBAMSvc.exe) running, performing background real-time protection
   and periodic definition updates.
4. JungleDisk cloud backup service is also running in the guest.
5. No user interaction was occurring at the time of the crash.
6. VirtualBox host process terminates with SIGSEGV. The guest becomes unresponsive. Heartbeat
   flatline is logged after ~4 seconds.

OrLogs.zip

Did you upload all of your necessary log files, screenshots, etc.?

• Yes, I've uploaded all pertinent files to this issue.

[alflr]

The logs show an amd64 host, whereas the issue hints an ARM host -- did you mix up something there?

[DEHughes]

Yes, that should have been an amd64 host instead of ARM.