Re: [PATCH v27 14/17] vfio: Dirty page tracking when vIOMMU is enabled

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

On Thu, 22 Oct 2020 16:42:04 +0530
Kirti Wankhede <kwankhede@nvidia.com> wrote:

> When vIOMMU is enabled, register MAP notifier from log_sync when all
> devices in container are in stop and copy phase of migration. Call replay
> and get dirty pages from notifier callback.
> 
> Suggested-by: Alex Williamson <alex.williamson@redhat.com>
> Signed-off-by: Kirti Wankhede <kwankhede@nvidia.com>
> ---
>  hw/vfio/common.c              | 95 ++++++++++++++++++++++++++++++++++++++++---
>  hw/vfio/trace-events          |  1 +
>  include/hw/vfio/vfio-common.h |  1 +
>  3 files changed, 91 insertions(+), 6 deletions(-)
> 
> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> index 2634387df948..98c2b1f9b190 100644
> --- a/hw/vfio/common.c
> +++ b/hw/vfio/common.c
> @@ -442,8 +442,8 @@ static bool vfio_listener_skipped_section(MemoryRegionSection *section)
>  }
>  
>  /* Called with rcu_read_lock held.  */
> -static bool vfio_get_vaddr(IOMMUTLBEntry *iotlb, void **vaddr,
> -                           bool *read_only)
> +static bool vfio_get_xlat_addr(IOMMUTLBEntry *iotlb, void **vaddr,
> +                               ram_addr_t *ram_addr, bool *read_only)
>  {
>      MemoryRegion *mr;
>      hwaddr xlat;
> @@ -474,8 +474,17 @@ static bool vfio_get_vaddr(IOMMUTLBEntry *iotlb, void **vaddr,
>          return false;
>      }
>  
> -    *vaddr = memory_region_get_ram_ptr(mr) + xlat;
> -    *read_only = !writable || mr->readonly;
> +    if (vaddr) {
> +        *vaddr = memory_region_get_ram_ptr(mr) + xlat;
> +    }
> +
> +    if (ram_addr) {
> +        *ram_addr = memory_region_get_ram_addr(mr) + xlat;
> +    }
> +
> +    if (read_only) {
> +        *read_only = !writable || mr->readonly;
> +    }
>  
>      return true;
>  }
> @@ -485,7 +494,6 @@ static void vfio_iommu_map_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb)
>      VFIOGuestIOMMU *giommu = container_of(n, VFIOGuestIOMMU, n);
>      VFIOContainer *container = giommu->container;
>      hwaddr iova = iotlb->iova + giommu->iommu_offset;
> -    bool read_only;
>      void *vaddr;
>      int ret;
>  
> @@ -501,7 +509,9 @@ static void vfio_iommu_map_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb)
>      rcu_read_lock();
>  
>      if ((iotlb->perm & IOMMU_RW) != IOMMU_NONE) {
> -        if (!vfio_get_vaddr(iotlb, &vaddr, &read_only)) {
> +        bool read_only;
> +
> +        if (!vfio_get_xlat_addr(iotlb, &vaddr, NULL, &read_only)) {
>              goto out;
>          }
>          /*
> @@ -899,11 +909,84 @@ err_out:
>      return ret;
>  }
>  
> +static void vfio_iommu_map_dirty_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb)
> +{
> +    VFIOGuestIOMMU *giommu = container_of(n, VFIOGuestIOMMU, dirty_notify);
> +    VFIOContainer *container = giommu->container;
> +    hwaddr iova = iotlb->iova + giommu->iommu_offset;
> +    ram_addr_t translated_addr;
> +
> +    trace_vfio_iommu_map_dirty_notify(iova, iova + iotlb->addr_mask);
> +
> +    if (iotlb->target_as != &address_space_memory) {
> +        error_report("Wrong target AS \"%s\", only system memory is allowed",
> +                     iotlb->target_as->name ? iotlb->target_as->name : "none");
> +        return;
> +    }
> +
> +    rcu_read_lock();
> +
> +    if (vfio_get_xlat_addr(iotlb, NULL, &translated_addr, NULL)) {
> +        int ret;
> +
> +        ret = vfio_get_dirty_bitmap(container, iova, iotlb->addr_mask + 1,
> +                                    translated_addr);
> +        if (ret) {
> +            error_report("vfio_iommu_map_dirty_notify(%p, 0x%"HWADDR_PRIx", "
> +                         "0x%"HWADDR_PRIx") = %d (%m)",
> +                         container, iova,
> +                         iotlb->addr_mask + 1, ret);
> +        }
> +    }
> +
> +    rcu_read_unlock();
> +}
> +
>  static int vfio_sync_dirty_bitmap(VFIOContainer *container,
>                                    MemoryRegionSection *section)
>  {
>      ram_addr_t ram_addr;
>  
> +    if (memory_region_is_iommu(section->mr)) {
> +        VFIOGuestIOMMU *giommu;
> +        int ret = 0;
> +
> +        QLIST_FOREACH(giommu, &container->giommu_list, giommu_next) {
> +            if (MEMORY_REGION(giommu->iommu) == section->mr &&
> +                giommu->n.start == section->offset_within_region) {
> +                Int128 llend;
> +                Error *err = NULL;
> +                int idx = memory_region_iommu_attrs_to_index(giommu->iommu,
> +                                                       MEMTXATTRS_UNSPECIFIED);
> +
> +                llend = int128_add(int128_make64(section->offset_within_region),
> +                                   section->size);
> +                llend = int128_sub(llend, int128_one());
> +
> +                iommu_notifier_init(&giommu->dirty_notify,
> +                                    vfio_iommu_map_dirty_notify,
> +                                    IOMMU_NOTIFIER_MAP,
> +                                    section->offset_within_region,
> +                                    int128_get64(llend),
> +                                    idx);
> +                ret = memory_region_register_iommu_notifier(section->mr,
> +                                                  &giommu->dirty_notify, &err);
> +                if (ret) {
> +                    error_report_err(err);
> +                    break;
> +                }
> +
> +                memory_region_iommu_replay(giommu->iommu,
> +                                           &giommu->dirty_notify);
> +
> +                memory_region_unregister_iommu_notifier(section->mr,
> +                                                        &giommu->dirty_notify);


Is it necessary to do the register/unregister?  It seemed to me that
perhaps we could do a replay independent of those.

I'd also be tempted to move dirty_notify to a temporary object rather
than store it on the giommu for such a brief usage, ie. define:

struct giommu_dirty_notfier {
    IOMMUNotifier n;
    VFIOGuestIOMMU *giommu;
}

struct giommu_dirty_notfier n = { .giommu = giommu };

iommu_notifier_init(&n,...);

memory_region_iommu_replay(giommu->iommu, &n);
...

struct giommu_dirty_notfier *ndnotifier = container_of(n, struct giommu_dirty_notfier, n);
VFIOGuestIOMMU *giommu = n->giommu;

It's nice that we remove the extra bloat of the list/tree entirely with
this approach.  Thanks,

Alex

> +                break;
> +            }
> +        }
> +        return ret;
> +    }
> +
>      ram_addr = memory_region_get_ram_addr(section->mr) +
>                 section->offset_within_region;
>  
> diff --git a/hw/vfio/trace-events b/hw/vfio/trace-events
> index c3b35aa2cce8..d9cb8998a228 100644
> --- a/hw/vfio/trace-events
> +++ b/hw/vfio/trace-events
> @@ -163,3 +163,4 @@ vfio_load_device_config_state(const char *name) " (%s)"
>  vfio_load_state(const char *name, uint64_t data) " (%s) data 0x%"PRIx64
>  vfio_load_state_device_data(const char *name, uint64_t data_offset, uint64_t data_size) " (%s) Offset 0x%"PRIx64" size 0x%"PRIx64
>  vfio_get_dirty_bitmap(int fd, uint64_t iova, uint64_t size, uint64_t bitmap_size, uint64_t start) "container fd=%d, iova=0x%"PRIx64" size= 0x%"PRIx64" bitmap_size=0x%"PRIx64" start=0x%"PRIx64
> +vfio_iommu_map_dirty_notify(uint64_t iova_start, uint64_t iova_end) "iommu dirty @ 0x%"PRIx64" - 0x%"PRIx64
> diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h
> index b1c1b18fd228..92872fae59ee 100644
> --- a/include/hw/vfio/vfio-common.h
> +++ b/include/hw/vfio/vfio-common.h
> @@ -99,6 +99,7 @@ typedef struct VFIOGuestIOMMU {
>      IOMMUMemoryRegion *iommu;
>      hwaddr iommu_offset;
>      IOMMUNotifier n;
> +    IOMMUNotifier dirty_notify;
>      QLIST_ENTRY(VFIOGuestIOMMU) giommu_next;
>  } VFIOGuestIOMMU;
>