[vbox-dev] Implementation of export_operations in vboxsf

Klaus Espenlaub klaus.espenlaub at oracle.com
Mon Feb 5 16:07:15 UTC 2018


Hi Greg,

On 03.02.2018 20:33, Grzegorz Zdanowski wrote:
> Hi,
> 
> Fiddling around I discovered vboxsf shares cannot be exported via NFS and the only workaround is bindfs AFAIK. 8 years ago someone created a ticket (#5295) describing this exact issue, but it didn’t get any love.

Probably because everyone instinctively thought that this doesn't sound
like a good idea :D

Re-exporting filesystems is in the general case a security concern.
Which you may or may not consider an issue in your use case. There's no
universal truth in this area, but it still makes devs hesitate.

> I’m a programmer, but I never fiddled around the kernel. From what I found Linux kernel will require defining "s_export_op“ in sf_read_super_aux() with minimum of „fh_to_dentry" implementation in order to enable NFS export. This seem to be a really small requirement, but I don’t really know where to start.
> 
> Can you give me some hint or maybe someone more knowledgeable in kernel internals can add s_export_op to superblock definition?

Sorry, not my area either...
Klaus

> 
> Thanks,
> --
> Greg


More information about the vbox-dev mailing list