[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZTgEApUgriFj1dKa@Asurada-Nvidia>
Date: Tue, 24 Oct 2023 10:50:58 -0700
From: Nicolin Chen <nicolinc@...dia.com>
To: Jason Gunthorpe <jgg@...dia.com>
CC: Yi Liu <yi.l.liu@...el.com>, <joro@...tes.org>,
<alex.williamson@...hat.com>, <kevin.tian@...el.com>,
<robin.murphy@....com>, <baolu.lu@...ux.intel.com>,
<cohuck@...hat.com>, <eric.auger@...hat.com>,
<kvm@...r.kernel.org>, <mjrosato@...ux.ibm.com>,
<chao.p.peng@...ux.intel.com>, <yi.y.sun@...ux.intel.com>,
<peterx@...hat.com>, <jasowang@...hat.com>,
<shameerali.kolothum.thodi@...wei.com>, <lulu@...hat.com>,
<suravee.suthikulpanit@....com>, <iommu@...ts.linux.dev>,
<linux-kernel@...r.kernel.org>, <linux-kselftest@...r.kernel.org>,
<zhenzhong.duan@...el.com>, <joao.m.martins@...cle.com>
Subject: Re: [PATCH v6 07/10] iommufd: Add a nested HW pagetable object
On Tue, Oct 24, 2023 at 02:31:39PM -0300, Jason Gunthorpe wrote:
> On Tue, Oct 24, 2023 at 10:28:45AM -0700, Nicolin Chen wrote:
> > On Tue, Oct 24, 2023 at 02:18:10PM -0300, Jason Gunthorpe wrote:
> > > On Tue, Oct 24, 2023 at 08:06:06AM -0700, Yi Liu wrote:
> > > > @@ -195,6 +279,10 @@ int iommufd_hwpt_alloc(struct iommufd_ucmd *ucmd)
> > > > if (pt_obj->type == IOMMUFD_OBJ_IOAS) {
> > > > struct iommufd_hwpt_paging *hwpt_paging;
> > > >
> > > > + if (cmd->data_type != IOMMU_HWPT_DATA_NONE) {
> > > > + rc = -EINVAL;
> > > > + goto out_put_pt;
> > > > + }
> > > > ioas = container_of(pt_obj, struct iommufd_ioas, obj);
> > > > mutex_lock(&ioas->mutex);
> > > > hwpt_paging = iommufd_hwpt_paging_alloc(ucmd->ictx, ioas, idev,
> > >
> > > ?? What is this?
> > >
> > > Ah something went wrong earlier in "iommu: Pass in parent domain with
> > > user_data to domain_alloc_user op"
> > >
> > > Once we added the user_data we should flow it through to the op
> > > always.
> >
> > Hmm, iommufd_hwpt_paging_alloc doesn't take (or need) user_data,
> > but we could pass in a dummy one if that looks better?
>
> The point is for the user_data to always be available, the driver
> needs to check it if it is passed.
>
> This should all be plumbed to allow drivers to also customize their
> paging domains too.
We don't have a use case of customizing the paging domains.
And our selftest isn't covering this path. Nor the case is
supported by the uAPI:
458- * A kernel-managed HWPT will be created with the mappings from the given
459- * IOAS via the @pt_id. The @data_type for this allocation must be set to
460: * IOMMU_HWPT_DATA_NONE. The HWPT can be allocated as a parent HWPT for a
461- * nesting configuration by passing IOMMU_HWPT_ALLOC_NEST_PARENT via @flags.
462- *
Also, if we do passing in the data, we'd need to...
280-static struct iommu_domain *
281-mock_domain_alloc_user(struct device *dev, u32 flags,
282- struct iommu_domain *parent,
283: const struct iommu_user_data *user_data)
284-{
285- struct mock_iommu_domain *mock_parent;
286- struct iommu_hwpt_selftest user_cfg;
287- int rc;
288-
289: if (!user_data) { /* must be mock_domain */
...change this to if (!parent)...
290- struct mock_dev *mdev = container_of(dev, struct mock_dev, dev);
291- bool has_dirty_flag = flags & IOMMU_HWPT_ALLOC_DIRTY_TRACKING;
292- bool no_dirty_ops = mdev->flags & MOCK_FLAGS_DEVICE_NO_DIRTY;
293-
294- if (parent)
295- return ERR_PTR(-EINVAL);
...and drop this.
296- if (has_dirty_flag && no_dirty_ops)
297- return ERR_PTR(-EOPNOTSUPP);
298- return __mock_domain_alloc_paging(IOMMU_DOMAIN_UNMANAGED,
299- has_dirty_flag);
300- }
301-
302- /* must be mock_domain_nested */
303: if (user_data->type != IOMMU_HWPT_DATA_SELFTEST)
304- return ERR_PTR(-EOPNOTSUPP);
Thanks
Nicolin
Powered by blists - more mailing lists