Re: [PATCH v4 2/2] qapi: add '@fdset' feature for BlockdevOptionsVirtioBlkVhostVdpa

[Stefano Garzarella]

On Fri, May 26, 2023 at 04:20:14PM -0500, Jonathon Jongsma wrote:
> On 5/26/23 10:03 AM, Stefano Garzarella wrote:
> > The virtio-blk-vhost-vdpa driver in libblkio 1.3.0 supports the fd
> > passing through the new 'fd' property.
> >
> > Since now we are using qemu_open() on '@path' if the virtio-blk driver
> > supports the fd passing, let's announce it.
> > In this way, the management layer can pass the file descriptor of an
> > already opened vhost-vdpa character device. This is useful especially
> > when the device can only be accessed with certain privileges.
> >
> > Add the '@fdset' feature only when the virtio-blk-vhost-vdpa driver
> > in libblkio supports it.
> >
> > Suggested-by: Markus Armbruster <armbru@redhat.com>
> > Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
> > ---
> >
> > Notes:
> >     v4:
> >     - added this patch to allow libvirt to discover we support fdset [Markus]
> >
> >  meson.build          | 4 ++++
> >  qapi/block-core.json | 8 +++++++-
> >  2 files changed, 11 insertions(+), 1 deletion(-)
> >
> > diff --git a/meson.build b/meson.build
> > index 78890f0155..8ea911f7b4 100644
> > --- a/meson.build
> > +++ b/meson.build
> > @@ -2108,6 +2108,10 @@ config_host_data.set('CONFIG_LZO', lzo.found())
> >  config_host_data.set('CONFIG_MPATH', mpathpersist.found())
> >  config_host_data.set('CONFIG_MPATH_NEW_API', mpathpersist_new_api)
> >  config_host_data.set('CONFIG_BLKIO', blkio.found())
> > +if blkio.found()
> > +  config_host_data.set('CONFIG_BLKIO_VHOST_VDPA_FD',
> > +                       blkio.version().version_compare('>=1.3.0'))
> > +endif
> >  config_host_data.set('CONFIG_CURL', curl.found())
> >  config_host_data.set('CONFIG_CURSES', curses.found())
> >  config_host_data.set('CONFIG_GBM', gbm.found())
> > diff --git a/qapi/block-core.json b/qapi/block-core.json
> > index 98d9116dae..1538d84ef4 100644
> > --- a/qapi/block-core.json
> > +++ b/qapi/block-core.json
> > @@ -3955,10 +3955,16 @@
> >  #
> >  # @path: path to the vhost-vdpa character device.
> >  #
> > +# Features:
> > +# @fdset: Member @path supports the special "/dev/fdset/N" path (since 8.1)
> > +#
> >  # Since: 7.2
> >  ##
> >  { 'struct': 'BlockdevOptionsVirtioBlkVhostVdpa',
> > -  'data': { 'path': 'str' },
> > +  'data': { 'path': { 'type': 'str',
> > +                      'features': [ { 'name' :'fdset',
> > +                                      'if': 'CONFIG_BLKIO_VHOST_VDPA_FD' } ]
> > +            } },
> >    'if': 'CONFIG_BLKIO' }
> >  ##
>
>
> Take this for what it's worth and do what's best for qemu, but... It's 
> easier for libvirt if the 'features' are on the object rather than on 
> the 'path' member. Our schema parsing code already supports object 
> features but does not yet support features on builtin types.

I had done it that way in the first instance :-), then I saw that the 
members themselves could have their own functionality, and it seemed 
better.

However I agree that if it's easier for libvirt, then better to move the 
feature to the whole object.

I'll change that in v5.

Thanks,
Stefano

> i.e. something like this just works without any refactoring in libvirt:
>
> diff --git a/qapi/block-core.json b/qapi/block-core.json
> index 1538d84ef4..78cfd10cdb 100644
> --- a/qapi/block-core.json
> +++ b/qapi/block-core.json
> @@ -3961,11 +3961,11 @@
> # Since: 7.2
> ##
> { 'struct': 'BlockdevOptionsVirtioBlkVhostVdpa',
> -  'data': { 'path': { 'type': 'str',
> -                      'features': [ { 'name' :'fdset',
> -                                      'if': 
> 'CONFIG_BLKIO_VHOST_VDPA_FD' } ]
> -            } },
> -  'if': 'CONFIG_BLKIO' }
> +  'data': { 'path': 'str' },
> +  'features': [ { 'name' :'fdset',
> +                'if': 'CONFIG_BLKIO_VHOST_VDPA_FD' } ],
> +  'if': 'CONFIG_BLKIO'
> + }
>
> ##
> # @IscsiTransport: