On Fri, Feb 26, 2021 at 05:38:04AM +0000, Christoph Hellwig wrote:

On Mon, Feb 22, 2021 at 09:50:35AM -0700, Alex Williamson wrote:

quoted

By linking all the device fds we provide to userspace to an
address space through a new pseudo fs, we can use tools like
unmap_mapping_range() to zap all vmas associated with a device.

Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Alex Williamson <redacted>

Adding Al:

I hate how we're are growing these tiny file systems just to allocate an
anonymous inode all over.  Shouldn't we allow to enhance fs/anon_inodes.c
to add a new API to allocate a new specific inode from anon_inodefs
instead?

+1 when I was researching this I also felt this was alot of
boilerplate to just get an inode. With vfio and rdma getting this it
is at least 5 places now.

Jason