Re: [PATCH iommufd v3 2/9] iommu: Add iommu_group_has_isolated_msi()

From: Baolu Lu <baolu.lu@linux.intel.com>
To: Jason Gunthorpe <jgg@nvidia.com>,
	Alexander Gordeev <agordeev@linux.ibm.com>,
	Alex Williamson <alex.williamson@redhat.com>,
	Christian Borntraeger <borntraeger@linux.ibm.com>,
	Cornelia Huck <cohuck@redhat.com>,
	David Woodhouse <dwmw2@infradead.org>,
	Gerald Schaefer <gerald.schaefer@linux.ibm.com>,
	Vasily Gorbik <gor@linux.ibm.com>,
	Heiko Carstens <hca@linux.ibm.com>,
	iommu@lists.linux.dev, Joerg Roedel <joro@8bytes.org>,
	kvm@vger.kernel.org, linux-s390@vger.kernel.org,
	Marc Zyngier <maz@kernel.org>,
	Robin Murphy <robin.murphy@arm.com>,
	Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>,
	Sven Schnelle <svens@linux.ibm.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Will Deacon <will@kernel.org>
Cc: Bharat Bhushan <bharat.bhushan@nxp.com>,
	Christian Borntraeger <borntraeger@de.ibm.com>,
	Eric Auger <eric.auger@redhat.com>,
	Eric Farman <farman@linux.ibm.com>,
	Kevin Tian <kevin.tian@intel.com>,
	Marc Zyngier <marc.zyngier@arm.com>,
	Matthew Rosato <mjrosato@linux.ibm.com>,
	Tomasz Nowicki <tomasz.nowicki@caviumnetworks.com>,
	Will Deacon <will.deacon@arm.com>
Subject: Re: [PATCH iommufd v3 2/9] iommu: Add iommu_group_has_isolated_msi()
Date: Fri, 6 Jan 2023 19:28:46 +0800	[thread overview]
Message-ID: <2c12143b-eaa9-2f6b-d367-e55d6f1e180d@linux.intel.com> (raw)
In-Reply-To: <2-v3-3313bb5dd3a3+10f11-secure_msi_jgg@nvidia.com>

On 1/6/2023 3:33 AM, Jason Gunthorpe wrote:
> Compute the isolated_msi over all the devices in the IOMMU group because
> iommufd and vfio both need to know that the entire group is isolated
> before granting access to it.
>
> Tested-by: Matthew Rosato <mjrosato@linux.ibm.com>
> Reviewed-by: Kevin Tian <kevin.tian@intel.com>
> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
> ---
>   drivers/iommu/iommu.c | 26 ++++++++++++++++++++++++++
>   include/linux/iommu.h |  1 +
>   2 files changed, 27 insertions(+)
>
> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
> index de91dd88705bd3..7f744904e02f4d 100644
> --- a/drivers/iommu/iommu.c
> +++ b/drivers/iommu/iommu.c
> @@ -30,6 +30,7 @@
>   #include <linux/cc_platform.h>
>   #include <trace/events/iommu.h>
>   #include <linux/sched/mm.h>
> +#include <linux/msi.h>
>   
>   #include "dma-iommu.h"
>   
> @@ -1897,6 +1898,31 @@ bool device_iommu_capable(struct device *dev, enum iommu_cap cap)
>   }
>   EXPORT_SYMBOL_GPL(device_iommu_capable);
>   
> +/**
> + * iommu_group_has_isolated_msi() - Compute msi_device_has_isolated_msi()
> + *       for a group
> + * @group: Group to query
> + *
> + * IOMMU groups should not have differing values of
> + * msi_device_has_isolated_msi() for devices in a group. However nothing
> + * directly prevents this, so ensure mistakes don't result in isolation failures
> + * by checking that all the devices are the same.
> + */
> +bool iommu_group_has_isolated_msi(struct iommu_group *group)
> +{
> +	struct group_device *group_dev;
> +	bool ret = true;
> +
> +	mutex_lock(&group->mutex);
> +	list_for_each_entry(group_dev, &group->devices, list)
> +		ret &= msi_device_has_isolated_msi(group_dev->dev) ||
> +		       device_iommu_capable(group_dev->dev,
> +					    IOMMU_CAP_INTR_REMAP);
> +	mutex_unlock(&group->mutex);
> +	return ret;
> +}
> +EXPORT_SYMBOL_GPL(iommu_group_has_isolated_msi);
> +
>   /**
>    * iommu_set_fault_handler() - set a fault handler for an iommu domain
>    * @domain: iommu domain
> diff --git a/include/linux/iommu.h b/include/linux/iommu.h
> index 46e1347bfa2286..9b7a9fa5ad28d3 100644
> --- a/include/linux/iommu.h
> +++ b/include/linux/iommu.h
> @@ -455,6 +455,7 @@ static inline const struct iommu_ops *dev_iommu_ops(struct device *dev)
>   extern int bus_iommu_probe(struct bus_type *bus);
>   extern bool iommu_present(struct bus_type *bus);
>   extern bool device_iommu_capable(struct device *dev, enum iommu_cap cap);
> +extern bool iommu_group_has_isolated_msi(struct iommu_group *group);

This lacks a static inline definition when CONFIG_IOMMU_API is false?

>   extern struct iommu_domain *iommu_domain_alloc(struct bus_type *bus);
>   extern struct iommu_group *iommu_group_get_by_id(int id);
>   extern void iommu_domain_free(struct iommu_domain *domain);

Others look good to me. With above addressed,

Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>

--

Best regards,

baolu