Re: [Qemu-devel] [PATCH v3 3/3] IOMMU: Integrate between VFIO and vIOMMU to support device assignment

[Alex Williamson <alex.williamson@redhat.com>]

On Sat, 21 May 2016 19:19:50 +0300
"Aviv B.D" <bd.aviv@gmail.com> wrote:

> From: "Aviv Ben-David" <bd.aviv@gmail.com>
> 

Some commentary about the changes necessary to achieve $SUBJECT would
be nice here.

> Signed-off-by: Aviv Ben-David <bd.aviv@gmail.com>
> ---
>  hw/i386/intel_iommu.c          | 69 ++++++++++++++++++++++++++++++++++++++++--
>  hw/i386/intel_iommu_internal.h |  2 ++
>  hw/vfio/common.c               | 11 +++++--
>  include/hw/i386/intel_iommu.h  |  4 +++
>  include/hw/vfio/vfio-common.h  |  1 +
>  5 files changed, 81 insertions(+), 6 deletions(-)
> 
> diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
> index 410f810..128ec7c 100644
> --- a/hw/i386/intel_iommu.c
> +++ b/hw/i386/intel_iommu.c
> @@ -43,6 +43,9 @@ static int vtd_dbgflags = VTD_DBGBIT(GENERAL) | VTD_DBGBIT(CSR);
>  #define VTD_DPRINTF(what, fmt, ...) do {} while (0)
>  #endif
>  
> +static int vtd_dev_to_context_entry(IntelIOMMUState *s, uint8_t bus_num,
> +                                    uint8_t devfn, VTDContextEntry *ce);
> +
>  static void vtd_define_quad(IntelIOMMUState *s, hwaddr addr, uint64_t val,
>                              uint64_t wmask, uint64_t w1cmask)
>  {
> @@ -126,6 +129,22 @@ static uint32_t vtd_set_clear_mask_long(IntelIOMMUState *s, hwaddr addr,
>      return new_val;
>  }
>  
> +static int vtd_get_did_dev(IntelIOMMUState *s, uint8_t bus_num, uint8_t devfn, uint16_t * domain_id)
> +{
> +    VTDContextEntry ce;
> +    int ret_fr;
> +
> +    assert(domain_id);
> +
> +    ret_fr = vtd_dev_to_context_entry(s, bus_num, devfn, &ce);
> +    if (ret_fr){
> +        return -1;
> +    }
> +
> +    *domain_id =  VTD_CONTEXT_ENTRY_DID(ce.hi);
> +    return 0;
> +}
> +
>  static uint64_t vtd_set_clear_mask_quad(IntelIOMMUState *s, hwaddr addr,
>                                          uint64_t clear, uint64_t mask)
>  {
> @@ -724,9 +743,6 @@ static int vtd_dev_to_context_entry(IntelIOMMUState *s, uint8_t bus_num,
>      }
>  
>      if (!vtd_context_entry_present(ce)) {
> -        VTD_DPRINTF(GENERAL,
> -                    "error: context-entry #%"PRIu8 "(bus #%"PRIu8 ") "
> -                    "is not present", devfn, bus_num);
>          return -VTD_FR_CONTEXT_ENTRY_P;
>      } else if ((ce->hi & VTD_CONTEXT_ENTRY_RSVD_HI) ||
>                 (ce->lo & VTD_CONTEXT_ENTRY_RSVD_LO)) {
> @@ -1033,18 +1049,58 @@ static void vtd_iotlb_domain_invalidate(IntelIOMMUState *s, uint16_t domain_id)
>                                  &domain_id);
>  }
>  
> +static void vtd_iotlb_page_invalidate_vfio(IntelIOMMUState *s, uint16_t domain_id,
> +                                      hwaddr addr, uint8_t am)
> +{
> +    VFIOGuestIOMMU * giommu;
> +

VT-d parsing VFIO private data structures, nope this is not a good idea.

> +    QLIST_FOREACH(giommu, &(s->giommu_list), iommu_next){
> +        VTDAddressSpace *vtd_as = container_of(giommu->iommu, VTDAddressSpace, iommu);

VT-d needs to keep track of its own address spaces and call the iommu
notifier, it should have no visibility whatsoever that there are vfio
devices attached.

> +        uint16_t vfio_domain_id; 
> +        int ret = vtd_get_did_dev(s, pci_bus_num(vtd_as->bus), vtd_as->devfn, &vfio_domain_id);
> +        int i=0;
> +        if (!ret && domain_id == vfio_domain_id){
> +            IOMMUTLBEntry entry; 
> +            
> +            /* do vfio unmap */
> +            VTD_DPRINTF(GENERAL, "Remove addr 0x%"PRIx64 " mask %d", addr, am);
> +            entry.target_as = NULL;
> +            entry.iova = addr & VTD_PAGE_MASK_4K;
> +            entry.translated_addr = 0;
> +            entry.addr_mask = ~VTD_PAGE_MASK(VTD_PAGE_SHIFT_4K + am);
> +            entry.perm = IOMMU_NONE;
> +            memory_region_notify_iommu(giommu->iommu, entry);
> +       
> +            /* do vfio map */
> +            VTD_DPRINTF(GENERAL, "add addr 0x%"PRIx64 " mask %d", addr, am);
> +            /* call to vtd_iommu_translate */
> +            for (i = 0; i < (1 << am); i++, addr+=(1 << VTD_PAGE_SHIFT_4K)){
> +                IOMMUTLBEntry entry = s->iommu_ops.translate(giommu->iommu, addr, IOMMU_NO_FAIL); 
> +                if (entry.perm != IOMMU_NONE){
> +                    memory_region_notify_iommu(giommu->iommu, entry);
> +                }
> +            }
> +        }
> +    }
> +}
> +
>  static void vtd_iotlb_page_invalidate(IntelIOMMUState *s, uint16_t domain_id,
>                                        hwaddr addr, uint8_t am)
>  {
>      VTDIOTLBPageInvInfo info;
>  
>      assert(am <= VTD_MAMV);
> +    
>      info.domain_id = domain_id;
>      info.addr = addr;
>      info.mask = ~((1 << am) - 1);
> +    
>      g_hash_table_foreach_remove(s->iotlb, vtd_hash_remove_by_page, &info);
> +
> +    vtd_iotlb_page_invalidate_vfio(s, domain_id, addr, am);
>  }
>  
> +
>  /* Flush IOTLB
>   * Returns the IOTLB Actual Invalidation Granularity.
>   * @val: the content of the IOTLB_REG
> @@ -1912,6 +1968,13 @@ static Property vtd_properties[] = {
>      DEFINE_PROP_END_OF_LIST(),
>  };
>  
> +void vtd_register_giommu(VFIOGuestIOMMU * giommu)
> +{
> +    VTDAddressSpace *vtd_as = container_of(giommu->iommu, VTDAddressSpace, iommu);
> +    IntelIOMMUState *s = vtd_as->iommu_state;
> +    
> +    QLIST_INSERT_HEAD(&s->giommu_list, giommu, iommu_next);
> +}
>  
>  VTDAddressSpace *vtd_find_add_as(IntelIOMMUState *s, PCIBus *bus, int devfn)
>  {
> diff --git a/hw/i386/intel_iommu_internal.h b/hw/i386/intel_iommu_internal.h
> index ae40f73..102e9a5 100644
> --- a/hw/i386/intel_iommu_internal.h
> +++ b/hw/i386/intel_iommu_internal.h
> @@ -339,6 +339,8 @@ typedef struct VTDIOTLBPageInvInfo VTDIOTLBPageInvInfo;
>  #define VTD_PAGE_SHIFT_1G           30
>  #define VTD_PAGE_MASK_1G            (~((1ULL << VTD_PAGE_SHIFT_1G) - 1))
>  
> +#define VTD_PAGE_MASK(shift)        (~((1ULL << (shift)) - 1))
> +
>  struct VTDRootEntry {
>      uint64_t val;
>      uint64_t rsvd;
> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> index 88154a1..54fc8bc 100644
> --- a/hw/vfio/common.c
> +++ b/hw/vfio/common.c
> @@ -35,6 +35,9 @@
>  #endif
>  #include "trace.h"
>  
> +#include "hw/sysbus.h"
> +#include "hw/i386/intel_iommu.h"
> +
>  struct vfio_group_head vfio_group_list =
>      QLIST_HEAD_INITIALIZER(vfio_group_list);
>  struct vfio_as_head vfio_address_spaces =
> @@ -315,12 +318,12 @@ static void vfio_iommu_map_notify(Notifier *n, void *data)
>  out:
>      rcu_read_unlock();
>  }
> -
> +#if 0
>  static hwaddr vfio_container_granularity(VFIOContainer *container)
>  {
>      return (hwaddr)1 << ctz64(container->iova_pgsizes);
>  }
> -
> +#endif


Clearly this is unacceptable, the code has a purpose.

>  static void vfio_listener_region_add(MemoryListener *listener,
>                                       MemoryRegionSection *section)
>  {
> @@ -384,11 +387,13 @@ static void vfio_listener_region_add(MemoryListener *listener,
>          giommu->n.notify = vfio_iommu_map_notify;
>          QLIST_INSERT_HEAD(&container->giommu_list, giommu, giommu_next);
>  
> +        vtd_register_giommu(giommu);

vfio will not assume VT-d, this is why we register the notifier below.

>          memory_region_register_iommu_notifier(giommu->iommu, &giommu->n);
> +#if 0
>          memory_region_iommu_replay(giommu->iommu, &giommu->n,
>                                     vfio_container_granularity(container),
>                                     false);
> -
> +#endif

Clearly this also has a purpose.

>          return;
>      }
>  
> diff --git a/include/hw/i386/intel_iommu.h b/include/hw/i386/intel_iommu.h
> index b024ffa..22f3f83 100644
> --- a/include/hw/i386/intel_iommu.h
> +++ b/include/hw/i386/intel_iommu.h
> @@ -23,6 +23,7 @@
>  #define INTEL_IOMMU_H
>  #include "hw/qdev.h"
>  #include "sysemu/dma.h"
> +#include "hw/vfio/vfio-common.h"

No.  This header probably should not have been put under include, VT-d
has no business walking our guest IOMMU list.

>  
>  #define TYPE_INTEL_IOMMU_DEVICE "intel-iommu"
>  #define INTEL_IOMMU_DEVICE(obj) \
> @@ -123,6 +124,8 @@ struct IntelIOMMUState {
>      MemoryRegionIOMMUOps iommu_ops;
>      GHashTable *vtd_as_by_busptr;   /* VTDBus objects indexed by PCIBus* reference */
>      VTDBus *vtd_as_by_bus_num[VTD_PCI_BUS_MAX]; /* VTDBus objects indexed by bus number */
> +
> +    QLIST_HEAD(, VFIOGuestIOMMU) giommu_list;
>  };
>  
>  /* Find the VTD Address space associated with the given bus pointer,
> @@ -130,4 +133,5 @@ struct IntelIOMMUState {
>   */
>  VTDAddressSpace *vtd_find_add_as(IntelIOMMUState *s, PCIBus *bus, int devfn);
>  
> +void vtd_register_giommu(VFIOGuestIOMMU * giommu);
>  #endif
> diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h
> index eb0e1b0..bf56a1d 100644
> --- a/include/hw/vfio/vfio-common.h
> +++ b/include/hw/vfio/vfio-common.h
> @@ -92,6 +92,7 @@ typedef struct VFIOGuestIOMMU {
>      MemoryRegion *iommu;
>      Notifier n;
>      QLIST_ENTRY(VFIOGuestIOMMU) giommu_next;
> +    QLIST_ENTRY(VFIOGuestIOMMU) iommu_next;

No.  Use the existing interfaces, create your own address space
tracking in VT-d, we are not going to host a list for VT-d to use.
Also note that there's no consideration of hot-unplug support in these
changes.  vfio already works with guest iommus on powerpc, so any
change to vfio needs to be justified and generalized to a common
guest iommu api.  Thanks,

Alex

>  } VFIOGuestIOMMU;
>  
>  typedef struct VFIODeviceOps VFIODeviceOps;