qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Cédric Le Goater" <clg@redhat.com>
To: Zhenzhong Duan <zhenzhong.duan@intel.com>, qemu-devel@nongnu.org
Cc: alex.williamson@redhat.com, mst@redhat.com, jasowang@redhat.com,
	yi.l.liu@intel.com, clement.mathieu--drif@eviden.com,
	eric.auger@redhat.com, joao.m.martins@oracle.com,
	avihaih@nvidia.com, xudong.hao@intel.com,
	giovanni.cabiddu@intel.com, mark.gross@intel.com,
	arjan.van.de.ven@intel.com
Subject: Re: [PATCH 1/5] vfio/iommufd: Add framework code to support getting dirty bitmap before unmap
Date: Fri, 19 Sep 2025 11:30:51 +0200	[thread overview]
Message-ID: <ee5c9d7c-a85e-4da4-bd5a-a59951a19a06@redhat.com> (raw)
In-Reply-To: <20250910023701.244356-2-zhenzhong.duan@intel.com>

Hello Zhenzhong

On 9/10/25 04:36, Zhenzhong Duan wrote:
> Currently we support device and iommu dirty tracking, device dirty
> tracking is preferred.
> 
> Add the framework code in iommufd_cdev_unmap_one() to choose either
> device or iommu dirty tracking, just like vfio_legacy_dma_unmap_one().

I wonder if commit 567d7d3e6be5 ("vfio/common: Work around kernel
overflow bug in DMA unmap") could be removed now to make the code
common to both VFIO IOMMU Type1 and IOMMUFD backends.

I asked Alex and Peter in another thread.

Thanks,

C.




> 
> Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
> Tested-by: Xudong Hao <xudong.hao@intel.com>
> Tested-by: Giovannio Cabiddu <giovanni.cabiddu@intel.com>
> ---
>   hw/vfio/iommufd.c | 56 +++++++++++++++++++++++++++++++++++++----------
>   1 file changed, 45 insertions(+), 11 deletions(-)
> 
> diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c
> index 48c590b6a9..b5d6e54c45 100644
> --- a/hw/vfio/iommufd.c
> +++ b/hw/vfio/iommufd.c
> @@ -58,33 +58,67 @@ static int iommufd_cdev_map_file(const VFIOContainerBase *bcontainer,
>                                           iova, size, fd, start, readonly);
>   }
>   
> -static int iommufd_cdev_unmap(const VFIOContainerBase *bcontainer,
> -                              hwaddr iova, ram_addr_t size,
> -                              IOMMUTLBEntry *iotlb, bool unmap_all)
> +static int iommufd_cdev_unmap_one(const VFIOContainerBase *bcontainer,
> +                                  hwaddr iova, ram_addr_t size,
> +                                  IOMMUTLBEntry *iotlb)
>   {
>       const VFIOIOMMUFDContainer *container =
>           container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer);
> +    bool need_dirty_sync = false;
> +    Error *local_err = NULL;
> +    IOMMUFDBackend *be = container->be;
> +    uint32_t ioas_id = container->ioas_id;
> +    int ret;
> +
> +    if (iotlb && vfio_container_dirty_tracking_is_started(bcontainer)) {
> +        if (!vfio_container_devices_dirty_tracking_is_supported(bcontainer) &&
> +            bcontainer->dirty_pages_supported) {
> +            /* TODO: query dirty bitmap before DMA unmap */
> +            return iommufd_backend_unmap_dma(be, ioas_id, iova, size);
> +        }
> +
> +        need_dirty_sync = true;
> +    }
> +
> +    ret = iommufd_backend_unmap_dma(be, ioas_id, iova, size);
> +    if (ret) {
> +        return ret;
> +    }
>   
> +    if (need_dirty_sync) {
> +        ret = vfio_container_query_dirty_bitmap(bcontainer, iova, size,
> +                                                iotlb->translated_addr,
> +                                                &local_err);
> +        if (ret) {
> +            error_report_err(local_err);
> +            return ret;
> +        }
> +    }
> +
> +    return 0;
> +}
> +static int iommufd_cdev_unmap(const VFIOContainerBase *bcontainer,
> +                              hwaddr iova, ram_addr_t size,
> +                              IOMMUTLBEntry *iotlb, bool unmap_all)
> +{
>       /* unmap in halves */
>       if (unmap_all) {
>           Int128 llsize = int128_rshift(int128_2_64(), 1);
>           int ret;
>   
> -        ret = iommufd_backend_unmap_dma(container->be, container->ioas_id,
> -                                        0, int128_get64(llsize));
> +        ret = iommufd_cdev_unmap_one(bcontainer, 0, int128_get64(llsize),
> +                                     iotlb);
>   
>           if (ret == 0) {
> -            ret = iommufd_backend_unmap_dma(container->be, container->ioas_id,
> -                                            int128_get64(llsize),
> -                                            int128_get64(llsize));
> +            ret = iommufd_cdev_unmap_one(bcontainer, int128_get64(llsize),
> +                                         int128_get64(llsize), iotlb);
>           }
>   
>           return ret;
>       }
>   
> -    /* TODO: Handle dma_unmap_bitmap with iotlb args (migration) */
> -    return iommufd_backend_unmap_dma(container->be,
> -                                     container->ioas_id, iova, size);
> +    return iommufd_cdev_unmap_one(bcontainer, iova, size, iotlb);
>   }
>   
>   static bool iommufd_cdev_kvm_device_add(VFIODevice *vbasedev, Error **errp)



  reply	other threads:[~2025-09-19  9:31 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-09-10  2:36 [PATCH 0/5] vfio: relax the vIOMMU check Zhenzhong Duan
2025-09-10  2:36 ` [PATCH 1/5] vfio/iommufd: Add framework code to support getting dirty bitmap before unmap Zhenzhong Duan
2025-09-19  9:30   ` Cédric Le Goater [this message]
2025-09-22  3:17     ` Duan, Zhenzhong
2025-09-22  8:27       ` Cédric Le Goater
2025-09-23  2:45         ` Duan, Zhenzhong
2025-09-23  7:06           ` Cédric Le Goater
2025-09-23  9:50             ` Duan, Zhenzhong
2025-09-10  2:36 ` [PATCH 2/5] vfio/iommufd: Query dirty bitmap before DMA unmap Zhenzhong Duan
2025-09-10  2:36 ` [PATCH 3/5] vfio/iommufd: Add IOMMU_HWPT_GET_DIRTY_BITMAP_NO_CLEAR flag support Zhenzhong Duan
2025-09-19 10:27   ` Joao Martins
2025-09-22  5:49     ` Duan, Zhenzhong
2025-09-22 16:02       ` Cédric Le Goater
2025-09-22 16:06         ` Joao Martins
2025-09-22 17:01           ` Joao Martins
2025-09-23  2:50             ` Duan, Zhenzhong
2025-09-23  9:17               ` Joao Martins
2025-09-23  9:55                 ` Duan, Zhenzhong
2025-09-23 10:06                   ` Duan, Zhenzhong
2025-09-23  2:47         ` Duan, Zhenzhong
2025-10-09 10:20         ` Duan, Zhenzhong
2025-10-09 12:43           ` Cédric Le Goater
2025-10-10  4:09             ` Duan, Zhenzhong
2025-09-10  2:37 ` [PATCH 4/5] intel_iommu: Optimize unmap_bitmap during migration Zhenzhong Duan
2025-10-12 10:31   ` Yi Liu
2025-10-13  2:50     ` Duan, Zhenzhong
2025-10-13 12:56       ` Yi Liu
2025-10-14  2:31         ` Duan, Zhenzhong
2025-10-14  6:46           ` Yi Liu
2025-09-10  2:37 ` [PATCH 5/5] vfio/migration: Allow live migration with vIOMMU without VFs using device dirty tracking Zhenzhong Duan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ee5c9d7c-a85e-4da4-bd5a-a59951a19a06@redhat.com \
    --to=clg@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=arjan.van.de.ven@intel.com \
    --cc=avihaih@nvidia.com \
    --cc=clement.mathieu--drif@eviden.com \
    --cc=eric.auger@redhat.com \
    --cc=giovanni.cabiddu@intel.com \
    --cc=jasowang@redhat.com \
    --cc=joao.m.martins@oracle.com \
    --cc=mark.gross@intel.com \
    --cc=mst@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=xudong.hao@intel.com \
    --cc=yi.l.liu@intel.com \
    --cc=zhenzhong.duan@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).