> From: Alex Williamson <alex.williamson@xxxxxxxxxx> > Sent: Wednesday, November 20, 2019 4:08 PM > > On Wed, 20 Nov 2019 14:11:08 -0400 > Jason Gunthorpe <jgg@xxxxxxxx> wrote: > > > I feel like mdev is suffering from mission creep. I see people > > proposing to use mdev for many wild things, the Mellanox SF stuff in > > the other thread and this 'virtio subsystem' being the two that have > > come up publicly this month. > > Tell me about it... ;) > Initial Mellanox sub function proposal was done using dedicated non-mdev subdev bus in [1] because mdev looked very vfio-ish. Along the way mdev proposal was suggested at [2] by mdev maintainers to use. The bus existed that detached two drivers (mdev and vfio_mdev), there was some motivation to attach other drivers. After that we continued discussion and mdev extension using alias to have persistent naming in [3]. So far so good, but when we want to have actual use of mdev driver, it doesn't look right. :-) > > Putting some boundaries on mdev usage would really help people know > > when to use it. My top two from this discussion would be: > > > > - mdev devices should only bind to vfio. It is not a general kernel > > driver matcher mechanism. It is not 'virtual-bus'. > So yes, we must define the scope of mdev and have right documentation to capture that. If mdev is not supposed to be extended beyond vfio, why do you even need a bus? For iommu attachment? [1] https://lkml.org/lkml/2019/3/1/19 [2] https://lkml.org/lkml/2019/3/7/696 [3] https://lkml.org/lkml/2019/8/26/854
