Re: [PATCH v2 1/3] iommufd: Add devices_users to track the hw_pagetable usage by device

[Nicolin Chen]

On Sun, Jan 29, 2023 at 09:23:05AM +0000, Tian, Kevin wrote:
> > From: Nicolin Chen <nicolinc@xxxxxxxxxx>
> > Sent: Sunday, January 29, 2023 5:18 AM
> >
> > From: Yi Liu <yi.l.liu@xxxxxxxxx>
> >
> > Currently, hw_pagetable tracks the attached devices using a device list.
> > When attaching the first device to the kernel-managed hw_pagetable, it
> > should be linked to IOAS. When detaching the last device from this hwpt,
> > the link with IOAS should be removed too. And this first-or-last device
> > check is done with list_empty(hwpt->devices).
> >
> > However, with a nested configuration, when a device is attached to the
> > user-managed stage-1 hw_pagetable, it will be added to this user-managed
> > hwpt's device list instead of the kernel-managed stage-2 hwpt's one. And
> > this breaks the logic for a kernel-managed hw_pagetable link/disconnect
> > to/from IOAS/IOPT. e.g. the stage-2 hw_pagetable would be linked to IOAS
> > multiple times if multiple device is attached, but it will become empty
> > as soon as one device detached.
> >
> > Add a devices_users in struct iommufd_hw_pagetable to track the users of
> 
> device_users

I assume you are suggesting a rename right? I can do that.

> > hw_pagetable by the attached devices. Make this field as a pointer, only
> > allocate for a stage-2 hw_pagetable. A stage-1 hw_pagetable should reuse
> > the stage-2 hw_pagetable's devices_users, because when a device attaches
> > to a stage-1 hw_pagetable, linking the stage-2 hwpt to the IOAS is still
> > required. So, with a nested configuration, increase the devices_users on
> > the stage-2 (parent) hwpt, no matter a device is attached to the stage-1
> > or the stage-2 hwpt.
> 
> Above is very confusing w/o seeing the full series of nesting support.
> 
> As a preparatory step this should focus on existing code and what this
> series tries to achieve. e.g. I'd not make device_users a pointer here.
> Do that incrementally when the nesting support comes.

OK. I will shift that part to the nesting series.

Thanks
Nic