[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <25102c92-1831-be52-677d-60bbf2e11772@linux.ibm.com>
Date: Thu, 9 Feb 2023 12:59:58 -0500
From: Matthew Rosato <mjrosato@...ux.ibm.com>
To: Yi Liu <yi.l.liu@...el.com>, joro@...tes.org,
alex.williamson@...hat.com, jgg@...dia.com, kevin.tian@...el.com,
robin.murphy@....com
Cc: cohuck@...hat.com, eric.auger@...hat.com, nicolinc@...dia.com,
kvm@...r.kernel.org, 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,
baolu.lu@...ux.intel.com
Subject: Re: [PATCH 06/17] iommufd/hw_pagetable: Use domain_alloc_user op for
domain allocation
On 2/8/23 11:31 PM, Yi Liu wrote:
> This converts iommufd to use iommu_domain_alloc_user() for iommu_domain
> creation.
>
> Suggested-by: Jason Gunthorpe <jgg@...dia.com>
> Signed-off-by: Nicolin Chen <nicolinc@...dia.com>
> Signed-off-by: Yi Liu <yi.l.liu@...el.com>
> ---
> drivers/iommu/iommufd/hw_pagetable.c | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c
> index 43d473989a06..08d963ee38c7 100644
> --- a/drivers/iommu/iommufd/hw_pagetable.c
> +++ b/drivers/iommu/iommufd/hw_pagetable.c
> @@ -30,6 +30,7 @@ struct iommufd_hw_pagetable *
> iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
> struct device *dev)
> {
> + const struct iommu_ops *ops;
> struct iommufd_hw_pagetable *hwpt;
> int rc;
>
> @@ -37,7 +38,13 @@ iommufd_hw_pagetable_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
> if (IS_ERR(hwpt))
> return hwpt;
>
> - hwpt->domain = iommu_domain_alloc(dev->bus);
> + ops = dev_iommu_ops(dev);
> + if (!ops || !ops->domain_alloc_user) {
> + rc = -EOPNOTSUPP;
> + goto out_abort;
> + }
Hi Yi,
This seems to break the iommufd vfio container support for any iommu that hasn't implemented domain_alloc_user yet.
I noticed it using vfio-pci on s390 with
CONFIG_IOMMUFD=m
CONFIG_IOMMUFD_VFIO_CONTAINER=y
CONFIG_VFIO_GROUP=y
Not sure if the intent is to make domain_alloc_user a hard requirement for using iommufd (if so then the commit description really should highlight that). Otherwise, conditionally calling iommu_domain_alloc(dev->bus) when !ops->domain_alloc_user (instead of returning -EOPNOTSUPP) seems to restore the prior functionality for me.
Thanks,
Matt
Powered by blists - more mailing lists