Re: [PATCH 5/6] iommu: init pasid array while doing domain_replace and iopf is active

From: Joel Granados
Date: Wed Sep 11 2024 - 06:57:26 EST


On Thu, Sep 05, 2024 at 11:30:05AM +0800, Baolu Lu wrote:
> On 9/4/24 9:17 PM, Joel Granados via B4 Relay wrote:
> > From: Joel Granados<j.granados@xxxxxxxxxxx>
> >
> > iommu_report_device_fault expects a pasid array to have an
> > iommu_attach_handle when a fault is detected.
>
> The iommu_attach_handle is expected only when an iopf-capable domain is
> attached to the device or PASID. The iommu_report_device_fault() treats
> it as a fault when a fault occurs, but no iopf-capable domain is
> attached.
>
> > Add this handle when the
> > replacing hwpt has a valid iommufd fault object. Remove it when we
> > release ownership of the group.
>
> The iommu_attach_handle is managed by the caller (iommufd here for
> example). Therefore, before iommu_attach_handle tries to attach a domain
> to an iopf-capable device or pasid, it should allocate the handle and
> pass it to the domain attachment interfaces.
What do you mean here?
1. Do you want to move the iommufd_init_pasid_array call up to
iommufd_hwpt_replace_device?
2. Do you want to move it further up to iommufd_device_do_replace?

Note that all this implemented on a call to replace HWPT. So a
non-iopf-capable VFIO_DEVICE_ATTACH_IOMMUFD_PT ioctl has already been
completed before the one that calls iommufd_device_do_replace.

> Conversely, the handle can
> only be freed after the domain is detached.
Do you have a function in specific where you would put the free handle
logic?

Best

--

Joel Granados