Re: [PATCH v7 06/23] iommu: Introduce bind/unbind_guest_msi

[Auger Eric]

Hi Robin,
On 5/8/19 3:59 PM, Robin Murphy wrote:
> On 08/04/2019 13:18, Eric Auger wrote:
>> On ARM, MSI are translated by the SMMU. An IOVA is allocated
>> for each MSI doorbell. If both the host and the guest are exposed
>> with SMMUs, we end up with 2 different IOVAs allocated by each.
>> guest allocates an IOVA (gIOVA) to map onto the guest MSI
>> doorbell (gDB). The Host allocates another IOVA (hIOVA) to map
>> onto the physical doorbell (hDB).
>>
>> So we end up with 2 untied mappings:
>> ÂÂÂÂÂÂÂÂÂ S1ÂÂÂÂÂÂÂÂÂÂÂ S2
>> gIOVAÂÂÂ ->ÂÂÂ gDB
>> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ hIOVAÂÂÂ ->ÂÂÂ hDB
>>
>> Currently the PCI device is programmed by the host with hIOVA
>> as MSI doorbell. So this does not work.
>>
>> This patch introduces an API to pass gIOVA/gDB to the host so
>> that gIOVA can be reused by the host instead of re-allocating
>> a new IOVA. So the goal is to create the following nested mapping:
>>
>> ÂÂÂÂÂÂÂÂÂ S1ÂÂÂÂÂÂÂÂÂÂÂ S2
>> gIOVAÂÂÂ ->ÂÂÂ gDBÂÂÂÂ ->ÂÂÂ hDB
>>
>> and program the PCI device with gIOVA MSI doorbell.
>>
>> In case we have several devices attached to this nested domain
>> (devices belonging to the same group), they cannot be isolated
>> on guest side either. So they should also end up in the same domain
>> on guest side. We will enforce that all the devices attached to
>> the host iommu domain use the same physical doorbell and similarly
>> a single virtual doorbell mapping gets registered (1 single
>> virtual doorbell is used on guest as well).
>>
>> Signed-off-by: Eric Auger <eric.auger@xxxxxxxxxx>
>>
>> ---
>> v6 -> v7:
>> - remove the device handle parameter.
>> - Add comments saying there can only be a single MSI binding
>> ÂÂ registered per iommu_domain
>> v5 -> v6:
>> -fix compile issue when IOMMU_API is not set
>>
>> v3 -> v4:
>> - add unbind
>>
>> v2 -> v3:
>> - add a struct device handle
>> ---
>> Â drivers/iommu/iommu.c | 37 +++++++++++++++++++++++++++++++++++++
>> Â include/linux/iommu.h | 23 +++++++++++++++++++++++
>> Â 2 files changed, 60 insertions(+)
>>
>> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
>> index 6d6cb4005ca5..0d160bbd6f81 100644
>> --- a/drivers/iommu/iommu.c
>> +++ b/drivers/iommu/iommu.c
>> @@ -1575,6 +1575,43 @@ static void __iommu_detach_device(struct
>> iommu_domain *domain,
>> ÂÂÂÂÂ trace_detach_device_from_domain(dev);
>> Â }
>> Â +/**
>> + * iommu_bind_guest_msi - Passes the stage1 GIOVA/GPA mapping of a
>> + * virtual doorbell
>> + *
>> + * @domain: iommu domain the stage 1 mapping will be attached to
>> + * @iova: iova allocated by the guest
>> + * @gpa: guest physical address of the virtual doorbell
>> + * @size: granule size used for the mapping
>> + *
>> + * The associated IOVA can be reused by the host to create a nested
>> + * stage2 binding mapping translating into the physical doorbell used
>> + * by the devices attached to the domain.
>> + *
>> + * All devices within the domain must share the same physical doorbell.
>> + * A single MSI GIOVA/GPA mapping can be attached to an iommu_domain.
>> + */
>> +
>> +int iommu_bind_guest_msi(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t giova, phys_addr_t gpa, size_t size)
>> +{
>> +ÂÂÂ if (unlikely(!domain->ops->bind_guest_msi))
>> +ÂÂÂÂÂÂÂ return -ENODEV;
>> +
>> +ÂÂÂ return domain->ops->bind_guest_msi(domain, giova, gpa, size);
>> +}
>> +EXPORT_SYMBOL_GPL(iommu_bind_guest_msi);
>> +
>> +void iommu_unbind_guest_msi(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t iova)
>> +{
>> +ÂÂÂ if (unlikely(!domain->ops->unbind_guest_msi))
>> +ÂÂÂÂÂÂÂ return;
>> +
>> +ÂÂÂ domain->ops->unbind_guest_msi(domain, iova);
>> +}
>> +EXPORT_SYMBOL_GPL(iommu_unbind_guest_msi);
>> +
>> Â void iommu_detach_device(struct iommu_domain *domain, struct device
>> *dev)
>> Â {
>> ÂÂÂÂÂ struct iommu_group *group;
>> diff --git a/include/linux/iommu.h b/include/linux/iommu.h
>> index 7c7c6bad1420..a2f3f964ead2 100644
>> --- a/include/linux/iommu.h
>> +++ b/include/linux/iommu.h
>> @@ -192,6 +192,8 @@ struct iommu_resv_region {
>> ÂÂ * @attach_pasid_table: attach a pasid table
>> ÂÂ * @detach_pasid_table: detach the pasid table
>> ÂÂ * @cache_invalidate: invalidate translation caches
>> + * @bind_guest_msi: provides a stage1 giova/gpa MSI doorbell mapping
>> + * @unbind_guest_msi: withdraw a stage1 giova/gpa MSI doorbell mapping
>> ÂÂ * @pgsize_bitmap: bitmap of all possible supported page sizes
>> ÂÂ */
>> Â struct iommu_ops {
>> @@ -243,6 +245,10 @@ struct iommu_ops {
>> ÂÂÂÂÂ int (*cache_invalidate)(struct iommu_domain *domain, struct
>> device *dev,
>> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ struct iommu_cache_invalidate_info *inv_info);
>> Â +ÂÂÂ int (*bind_guest_msi)(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t giova, phys_addr_t gpa, size_t size);
>> +ÂÂÂ void (*unbind_guest_msi)(struct iommu_domain *domain, dma_addr_t
>> giova);
>> +
>> ÂÂÂÂÂ unsigned long pgsize_bitmap;
>> Â };
>> Â @@ -356,6 +362,11 @@ extern void iommu_detach_pasid_table(struct
>> iommu_domain *domain);
>> Â extern int iommu_cache_invalidate(struct iommu_domain *domain,
>> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ struct device *dev,
>> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ struct iommu_cache_invalidate_info *inv_info);
>> +extern int iommu_bind_guest_msi(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t giova, phys_addr_t gpa, size_t size);
>> +extern void iommu_unbind_guest_msi(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t giova);
>> +
>> Â extern struct iommu_domain *iommu_get_domain_for_dev(struct device
>> *dev);
>> Â extern struct iommu_domain *iommu_get_dma_domain(struct device *dev);
>> Â extern int iommu_map(struct iommu_domain *domain, unsigned long iova,
>> @@ -812,6 +823,18 @@ iommu_cache_invalidate(struct iommu_domain *domain,
>> ÂÂÂÂÂ return -ENODEV;
>> Â }
>> Â +static inline
>> +int iommu_bind_guest_msi(struct iommu_domain *domain,
>> +ÂÂÂÂÂÂÂÂÂÂÂÂ dma_addr_t giova, phys_addr_t gpa, size_t size)
>> +{
>> +ÂÂÂ return -ENODEV;
>> +}
>> +static inline
>> +int iommu_unbind_guest_msi(struct iommu_domain *domain, dma_addr_t
>> giova)
>> +{
>> +ÂÂÂ return -ENODEV;
> 
> It's less of a problem than mismatching the other way round, but for
> consistency this should return void like the real version.
thank you for spotting this.

Eric
> 
> Robin.
> 
>> +}
>> +
>> Â #endif /* CONFIG_IOMMU_API */
>> Â Â #ifdef CONFIG_IOMMU_DEBUGFS
>>