VMware Communities
mbarrientos
Contributor
Contributor

Bug Report: Fusion 4.1.3 + Linux 3.1.0 or 3.4.6 + hgfs == guest oops/lockup

VMware:

     Fusion 4.1.3 (730298)

Host:

     Macbook Air 2012/8GB RAM/512 GB SSD

     OS X 10.7.4 (Lion)

Guest:

     Fedora Core 15 + all updates, tested with either a rebuild of:

          kernel-3.1.0-7.fc16.src.rpm (http://archive.fedoraproject.org/pub/fedora/linux/releases/16/Everything/source/SRPMS/kernel-3.1.0-7...)

     or

          kernel-3.4.6-2.fc17.src.rpm (http:////archive.fedoraproject.org/pub/fedora/linux/updates/17/SRPMS/kernel-3.4.6-2.fc17.src.rpm)

          module-init-tools-3.16-5.fc17.src.rpm (http://archive.fedoraproject.org/pub/fedora/linux/releases/17/Everything/source/SRPMS/m/module-init-...)

VMWare tools:

     VMwareTools-8.8.4-730257

     for 3.1.0, installed without modification

     for 3.4.6, install needed a patch to source/vmhgfs-only/filesystem.c due to a interface name change:

          + #if LINUX_VERSION_CODE < KERNEL_VERSION(3, 4, 0)
          + tempRootDentry = d_alloc_root(rootInode);
          + #else
             tempRootDentry = d_make_root(rootInode);
          + #endif
Hitting the bug:
     1. created a hgfs share, and put a large source tree in the share.
     2. compile with multiple jobs (e.g. make -j4)
     then, the VM locks up and you get repeated oops to the console.
    
     I hit the bug in the same way on either 3.4.6 or 3.1.0 and is is 100% reproducible.
Attached is a screencapture of the fist oops messages (which never makes it to disk).  The image is of two captures pasted together, so the times to not match up (I left a blank line between the two captures so you can see where you go from one to the other), but they are from the same kernel (3.1.0) and contain all of the console messages in the correct order.
Also attached is my vm conf file.
Thanks,
Michael
0 Kudos
0 Replies