Re: [PATCH v8 08/28] vfio: add region cache

["Cédric Le Goater" <clg@redhat.com>]

On 2/19/25 15:48, John Levon wrote:
> From: Jagannathan Raman <jag.raman@oracle.com>
> 
> Instead of requesting region information on demand with
> VFIO_DEVICE_GET_REGION_INFO, maintain a cache: this will become
> necessary for performance for vfio-user, where this call becomes a
> message over the control socket, so is of higher overhead than the
> traditional path.
> 
> Originally-by: John Johnson <john.g.johnson@oracle.com>
> Signed-off-by: Elena Ufimtseva <elena.ufimtseva@oracle.com>
> Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
> Signed-off-by: John Levon <john.levon@nutanix.com>
> ---
>   hw/vfio/ccw.c                 |  5 -----
>   hw/vfio/common.c              | 12 ++++++++++++
>   hw/vfio/container.c           | 10 ++++++++++
>   hw/vfio/helpers.c             | 21 ++++++++++++++++-----
>   hw/vfio/igd.c                 |  8 ++++----
>   hw/vfio/pci.c                 |  8 ++++----
>   include/hw/vfio/vfio-common.h |  1 +
>   7 files changed, 47 insertions(+), 18 deletions(-)
> 
> diff --git a/hw/vfio/ccw.c b/hw/vfio/ccw.c
> index 67bc137f9b..22378d50bc 100644
> --- a/hw/vfio/ccw.c
> +++ b/hw/vfio/ccw.c
> @@ -510,7 +510,6 @@ static bool vfio_ccw_get_region(VFIOCCWDevice *vcdev, Error **errp)
>   
>       vcdev->io_region_offset = info->offset;
>       vcdev->io_region = g_malloc0(info->size);
> -    g_free(info);
>   
>       /* check for the optional async command region */
>       ret = vfio_get_dev_region_info(vdev, VFIO_REGION_TYPE_CCW,
> @@ -523,7 +522,6 @@ static bool vfio_ccw_get_region(VFIOCCWDevice *vcdev, Error **errp)
>           }
>           vcdev->async_cmd_region_offset = info->offset;
>           vcdev->async_cmd_region = g_malloc0(info->size);
> -        g_free(info);
>       }
>   
>       ret = vfio_get_dev_region_info(vdev, VFIO_REGION_TYPE_CCW,
> @@ -536,7 +534,6 @@ static bool vfio_ccw_get_region(VFIOCCWDevice *vcdev, Error **errp)
>           }
>           vcdev->schib_region_offset = info->offset;
>           vcdev->schib_region = g_malloc(info->size);
> -        g_free(info);
>       }
>   
>       ret = vfio_get_dev_region_info(vdev, VFIO_REGION_TYPE_CCW,
> @@ -550,7 +547,6 @@ static bool vfio_ccw_get_region(VFIOCCWDevice *vcdev, Error **errp)
>           }
>           vcdev->crw_region_offset = info->offset;
>           vcdev->crw_region = g_malloc(info->size);
> -        g_free(info);
>       }
>   
>       return true;
> @@ -560,7 +556,6 @@ out_err:
>       g_free(vcdev->schib_region);
>       g_free(vcdev->async_cmd_region);
>       g_free(vcdev->io_region);
> -    g_free(info);
>       return false;
>   }
>   
> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> index 4434e0a0a2..1866b3d3c5 100644
> --- a/hw/vfio/common.c
> +++ b/hw/vfio/common.c
> @@ -1569,6 +1569,16 @@ retry:
>       return info;
>   }
>   
> +static void vfio_get_all_regions(VFIODevice *vbasedev)
> +{
> +    struct vfio_region_info *info;
> +    int i;
> +
> +    for (i = 0; i < vbasedev->num_regions; i++) {
> +        vfio_get_region_info(vbasedev, i, &info);
> +    }
> +}
> +
>   void vfio_prepare_device(VFIODevice *vbasedev, VFIOContainerBase *bcontainer,
>                            VFIOGroup *group, struct vfio_device_info *info)
>   {
> @@ -1586,6 +1596,8 @@ void vfio_prepare_device(VFIODevice *vbasedev, VFIOContainerBase *bcontainer,
>       }
>   
>       QLIST_INSERT_HEAD(&vfio_device_list, vbasedev, global_next);
> +
> +    vfio_get_all_regions(vbasedev);
>   }
>   
>   bool vfio_attach_device_by_iommu_type(const char *iommu_type, char *name,
> diff --git a/hw/vfio/container.c b/hw/vfio/container.c
> index 37a3befbc5..36cd245c92 100644
> --- a/hw/vfio/container.c
> +++ b/hw/vfio/container.c
> @@ -886,6 +886,16 @@ static bool vfio_get_device(VFIOGroup *group, const char *name,
>   
>   static void vfio_put_base_device(VFIODevice *vbasedev)
>   {
> +    if (vbasedev->regions != NULL) {
> +        int i;
> +
> +        for (i = 0; i < vbasedev->num_regions; i++) {
> +            g_free(vbasedev->regions[i]);
> +        }
> +        g_free(vbasedev->regions);
> +        vbasedev->regions = NULL;
> +    }
> +
>       if (!vbasedev->group) {
>           return;
>       }
> diff --git a/hw/vfio/helpers.c b/hw/vfio/helpers.c
> index 4b255d4f3a..3c923d23b9 100644
> --- a/hw/vfio/helpers.c
> +++ b/hw/vfio/helpers.c
> @@ -345,7 +345,7 @@ static int vfio_setup_region_sparse_mmaps(VFIORegion *region,
>   int vfio_region_setup(Object *obj, VFIODevice *vbasedev, VFIORegion *region,
>                         int index, const char *name)
>   {
> -    g_autofree struct vfio_region_info *info = NULL;
> +    struct vfio_region_info *info = NULL;
>       int ret;
>   
>       ret = vfio_get_region_info(vbasedev, index, &info);
> @@ -562,6 +562,17 @@ int vfio_get_region_info(VFIODevice *vbasedev, int index,
>   {
>       size_t argsz = sizeof(struct vfio_region_info);
>   
> +    /* create region cache */
> +    if (vbasedev->regions == NULL) {
> +        vbasedev->regions = g_new0(struct vfio_region_info *,
> +                                   vbasedev->num_regions);
> +    }
> +    /* check cache */
> +    if (vbasedev->regions[index] != NULL) {
> +        *info = vbasedev->regions[index];
> +        return 0;
> +    }
> +

why not populate vbasedev->regions[index] in vfio_get_all_regions() ?


Thanks,

C.



>       *info = g_malloc0(argsz);
>   
>       (*info)->index = index;
> @@ -581,6 +592,9 @@ retry:
>           goto retry;
>       }
>   
> +    /* fill cache */
> +    vbasedev->regions[index] = *info;
> +
>       return 0;
>   }
>   
> @@ -599,7 +613,6 @@ int vfio_get_dev_region_info(VFIODevice *vbasedev, uint32_t type,
>   
>           hdr = vfio_get_region_info_cap(*info, VFIO_REGION_INFO_CAP_TYPE);
>           if (!hdr) {
> -            g_free(*info);
>               continue;
>           }
>   
> @@ -611,8 +624,6 @@ int vfio_get_dev_region_info(VFIODevice *vbasedev, uint32_t type,
>           if (cap_type->type == type && cap_type->subtype == subtype) {
>               return 0;
>           }
> -
> -        g_free(*info);
>       }
>   
>       *info = NULL;
> @@ -621,7 +632,7 @@ int vfio_get_dev_region_info(VFIODevice *vbasedev, uint32_t type,
>   
>   bool vfio_has_region_cap(VFIODevice *vbasedev, int region, uint16_t cap_type)
>   {
> -    g_autofree struct vfio_region_info *info = NULL;
> +    struct vfio_region_info *info = NULL;
>       bool ret = false;
>   
>       if (!vfio_get_region_info(vbasedev, region, &info)) {
> diff --git a/hw/vfio/igd.c b/hw/vfio/igd.c
> index b1a237edd6..b5425ba9c0 100644
> --- a/hw/vfio/igd.c
> +++ b/hw/vfio/igd.c
> @@ -490,10 +490,10 @@ void vfio_probe_igd_bar0_quirk(VFIOPCIDevice *vdev, int nr)
>   
>   void vfio_probe_igd_bar4_quirk(VFIOPCIDevice *vdev, int nr)
>   {
> -    g_autofree struct vfio_region_info *rom = NULL;
> -    g_autofree struct vfio_region_info *opregion = NULL;
> -    g_autofree struct vfio_region_info *host = NULL;
> -    g_autofree struct vfio_region_info *lpc = NULL;
> +    struct vfio_region_info *rom = NULL;
> +    struct vfio_region_info *opregion = NULL;
> +    struct vfio_region_info *host = NULL;
> +    struct vfio_region_info *lpc = NULL;
>       VFIOQuirk *quirk;
>       VFIOIGDQuirk *igd;
>       PCIDevice *lpc_bridge;
> diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
> index 83fe329474..a4f99fc5e0 100644
> --- a/hw/vfio/pci.c
> +++ b/hw/vfio/pci.c
> @@ -879,7 +879,7 @@ static void vfio_update_msi(VFIOPCIDevice *vdev)
>   
>   static void vfio_pci_load_rom(VFIOPCIDevice *vdev)
>   {
> -    g_autofree struct vfio_region_info *reg_info = NULL;
> +    struct vfio_region_info *reg_info = NULL;
>       uint64_t size;
>       off_t off = 0;
>       ssize_t bytes;
> @@ -2665,7 +2665,7 @@ static VFIODeviceOps vfio_pci_ops = {
>   bool vfio_populate_vga(VFIOPCIDevice *vdev, Error **errp)
>   {
>       VFIODevice *vbasedev = &vdev->vbasedev;
> -    g_autofree struct vfio_region_info *reg_info = NULL;
> +    struct vfio_region_info *reg_info = NULL;
>       int ret;
>   
>       ret = vfio_get_region_info(vbasedev, VFIO_PCI_VGA_REGION_INDEX, &reg_info);
> @@ -2730,7 +2730,7 @@ bool vfio_populate_vga(VFIOPCIDevice *vdev, Error **errp)
>   static bool vfio_populate_device(VFIOPCIDevice *vdev, Error **errp)
>   {
>       VFIODevice *vbasedev = &vdev->vbasedev;
> -    g_autofree struct vfio_region_info *reg_info = NULL;
> +    struct vfio_region_info *reg_info = NULL;
>       struct vfio_irq_info irq_info = { .argsz = sizeof(irq_info) };
>       int i, ret = -1;
>   
> @@ -3177,7 +3177,7 @@ static void vfio_realize(PCIDevice *pdev, Error **errp)
>   
>       if (!vdev->igd_opregion &&
>           vdev->features & VFIO_FEATURE_ENABLE_IGD_OPREGION) {
> -        g_autofree struct vfio_region_info *opregion = NULL;
> +        struct vfio_region_info *opregion = NULL;
>   
>           if (vdev->pdev.qdev.hotplugged) {
>               error_setg(errp,
> diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h
> index ae3ecbd9f6..304030e71d 100644
> --- a/include/hw/vfio/vfio-common.h
> +++ b/include/hw/vfio/vfio-common.h
> @@ -151,6 +151,7 @@ typedef struct VFIODevice {
>       IOMMUFDBackend *iommufd;
>       VFIOIOASHwpt *hwpt;
>       QLIST_ENTRY(VFIODevice) hwpt_next;
> +    struct vfio_region_info **regions;
>   } VFIODevice;
>   
>   struct VFIODeviceOps {