qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Cédric Le Goater" <clg@redhat.com>
To: "Duan, Zhenzhong" <zhenzhong.duan@intel.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>
Cc: Eric Auger <eric.auger@redhat.com>,
	Alex Williamson <alex.williamson@redhat.com>,
	Nicholas Piggin <npiggin@gmail.com>,
	Harsh Prateek Bora <harshpb@linux.ibm.com>
Subject: Re: [PATCH for-9.0 05/10] vfio/container: Introduce a VFIOIOMMU legacy QOM interface
Date: Mon, 11 Dec 2023 15:00:50 +0100	[thread overview]
Message-ID: <6f47ba3d-2b87-4bf6-84af-e0eb5e78b19a@redhat.com> (raw)
In-Reply-To: <SJ0PR11MB6744BFFFC1507716EF630EEB928FA@SJ0PR11MB6744.namprd11.prod.outlook.com>

On 12/11/23 07:14, Duan, Zhenzhong wrote:
> 
> 
>> -----Original Message-----
>> From: Cédric Le Goater <clg@redhat.com>
>> Sent: Friday, December 8, 2023 4:46 PM
>> Subject: [PATCH for-9.0 05/10] vfio/container: Introduce a VFIOIOMMU
>> legacy QOM interface
>>
>> Convert the legacy VFIOIOMMUOps struct to the new VFIOIOMMU QOM
>> interface. The set of of operations for this backend can be referenced
>> with a literal typename instead of a C struct. This will simplify
>> support of multiple backends.
>>
>> Signed-off-by: Cédric Le Goater <clg@redhat.com>
>> ---
>> include/hw/vfio/vfio-common.h         |  1 -
>> include/hw/vfio/vfio-container-base.h |  1 +
>> hw/vfio/common.c                      |  6 ++-
>> hw/vfio/container.c                   | 59 +++++++++++++++++++++++----
>> 4 files changed, 56 insertions(+), 11 deletions(-)
>>
>> diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-
>> common.h
>> index
>> b8aa8a549532442a31c8e85ce385c992d84f6bd5..14c497b6b0a79466e8f56
>> 7aceed384ec2c75ea90 100644
>> --- a/include/hw/vfio/vfio-common.h
>> +++ b/include/hw/vfio/vfio-common.h
>> @@ -210,7 +210,6 @@ typedef QLIST_HEAD(VFIOGroupList, VFIOGroup)
>> VFIOGroupList;
>> typedef QLIST_HEAD(VFIODeviceList, VFIODevice) VFIODeviceList;
>> extern VFIOGroupList vfio_group_list;
>> extern VFIODeviceList vfio_device_list;
>> -extern const VFIOIOMMUOps vfio_legacy_ops;
>> extern const VFIOIOMMUOps vfio_iommufd_ops;
>> extern const MemoryListener vfio_memory_listener;
>> extern int vfio_kvm_device_fd;
>> diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-
>> container-base.h
>> index
>> 81d49fe562d3840859096dd8a62ac38d62314939..a31fd9c2e3b9a571083e
>> a8987ac27e91b332c170 100644
>> --- a/include/hw/vfio/vfio-container-base.h
>> +++ b/include/hw/vfio/vfio-container-base.h
>> @@ -95,6 +95,7 @@ void vfio_container_destroy(VFIOContainerBase
>> *bcontainer);
>> typedef struct VFIOIOMMU VFIOIOMMU;
>>
>> #define TYPE_VFIO_IOMMU "vfio-iommu"
>> +#define TYPE_VFIO_IOMMU_LEGACY TYPE_VFIO_IOMMU "-legacy"
>>
>> #define VFIO_IOMMU(obj) INTERFACE_CHECK(VFIOIOMMU, (obj),
>> TYPE_VFIO_IOMMU)
>> DECLARE_CLASS_CHECKERS(VFIOIOMMUClass, VFIO_IOMMU,
>> TYPE_VFIO_IOMMU)
>> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
>> index
>> 49dab41566f07ba7be1100fed1973e028d34467c..2329d0efc8c1d617f0bfee
>> 5283e82b295d2d477d 100644
>> --- a/hw/vfio/common.c
>> +++ b/hw/vfio/common.c
>> @@ -1503,13 +1503,17 @@ retry:
>> int vfio_attach_device(char *name, VFIODevice *vbasedev,
>>                         AddressSpace *as, Error **errp)
>> {
>> -    const VFIOIOMMUClass *ops = &vfio_legacy_ops;
>> +    const VFIOIOMMUClass *ops =
>> +
>> VFIO_IOMMU_CLASS(object_class_by_name(TYPE_VFIO_IOMMU_LEGACY));
>>
>> #ifdef CONFIG_IOMMUFD
>>      if (vbasedev->iommufd) {
>>          ops = &vfio_iommufd_ops;
>>      }
>> #endif
>> +
>> +    assert(ops);
>> +
>>      return ops->attach_device(name, vbasedev, as, errp);
>> }
>>
>> diff --git a/hw/vfio/container.c b/hw/vfio/container.c
>> index
>> f4a0434a5239bfb6a17b91c8879cb98e686afccc..fdf4e116570013732d4807
>> 1a5122d25b02da715c 100644
>> --- a/hw/vfio/container.c
>> +++ b/hw/vfio/container.c
>> @@ -369,10 +369,30 @@ static int vfio_get_iommu_type(VFIOContainer
>> *container,
>>      return -EINVAL;
>> }
>>
>> +/*
>> + * vfio_get_iommu_ops - get a VFIOIOMMUClass associated with a type
>> + */
>> +static const VFIOIOMMUClass *vfio_get_iommu_class(int iommu_type,
>> Error **errp)
>> +{
>> +    ObjectClass *klass = NULL;
> 
> No need to nullify?

well, I am not sure. Some compilers might complain. I will check.


>> +
>> +    switch (iommu_type) {
>> +    case VFIO_TYPE1v2_IOMMU:
>> +    case VFIO_TYPE1_IOMMU:
>> +        klass = object_class_by_name(TYPE_VFIO_IOMMU_LEGACY);
>> +        break;
>> +    default:
>> +        g_assert_not_reached();
>> +    };
>> +
>> +    return VFIO_IOMMU_CLASS(klass);
>> +}
>> +
>> static int vfio_init_container(VFIOContainer *container, int group_fd,
>>                                 VFIOAddressSpace *space, Error **errp)
>> {
>>      int iommu_type, ret;
>> +    const VFIOIOMMUClass *vioc = NULL;
> 
> No need to nullify?

No need indeed.
  
>>
>>      iommu_type = vfio_get_iommu_type(container, errp);
>>      if (iommu_type < 0) {
>> @@ -401,7 +421,14 @@ static int vfio_init_container(VFIOContainer
>> *container, int group_fd,
>>      }
>>
>>      container->iommu_type = iommu_type;
>> -    vfio_container_init(&container->bcontainer, space, &vfio_legacy_ops);
>> +
>> +    vioc = vfio_get_iommu_class(iommu_type, errp);
>> +    if (!vioc) {
>> +        error_setg(errp, "No available IOMMU models");
>> +        return -EINVAL;
>> +    }
>> +
>> +    vfio_container_init(&container->bcontainer, space, vioc);
>>      return 0;
>> }
>>
>> @@ -1098,12 +1125,26 @@ out_single:
>>      return ret;
>> }
>>
>> -const VFIOIOMMUOps vfio_legacy_ops = {
>> -    .dma_map = vfio_legacy_dma_map,
>> -    .dma_unmap = vfio_legacy_dma_unmap,
>> -    .attach_device = vfio_legacy_attach_device,
>> -    .detach_device = vfio_legacy_detach_device,
>> -    .set_dirty_page_tracking = vfio_legacy_set_dirty_page_tracking,
>> -    .query_dirty_bitmap = vfio_legacy_query_dirty_bitmap,
>> -    .pci_hot_reset = vfio_legacy_pci_hot_reset,
>> +static void vfio_iommu_legacy_class_init(ObjectClass *klass, void *data)
>> +{
>> +    VFIOIOMMUClass *vioc = VFIO_IOMMU_CLASS(klass);
>> +
>> +    vioc->dma_map = vfio_legacy_dma_map;
>> +    vioc->dma_unmap = vfio_legacy_dma_unmap;
>> +    vioc->attach_device = vfio_legacy_attach_device;
>> +    vioc->detach_device = vfio_legacy_detach_device;
>> +    vioc->set_dirty_page_tracking = vfio_legacy_set_dirty_page_tracking;
>> +    vioc->query_dirty_bitmap = vfio_legacy_query_dirty_bitmap;
>> +    vioc->pci_hot_reset = vfio_legacy_pci_hot_reset;
>> };
>> +
>> +static const TypeInfo types[] = {
>> +    {
>> +        .name = TYPE_VFIO_IOMMU_LEGACY,
>> +        .parent = TYPE_VFIO_IOMMU,
>> +        .class_init = vfio_iommu_legacy_class_init,
>> +        .class_size = sizeof(VFIOIOMMUClass),
> 
> Inherit parent class_size is enough? Otherwise,

No need to define class_size again.


Thanks,

C.



> 
> Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
> 
> Thanks
> Zhenzhong
> 
>> +    },
>> +};
>> +
>> +DEFINE_TYPES(types)
>> --
>> 2.43.0
> 



  reply	other threads:[~2023-12-11 14:01 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-12-08  8:45 [PATCH for-9.0 00/10] vfio: Introduce a VFIOIOMMUClass Cédric Le Goater
2023-12-08  8:45 ` [PATCH for-9.0 01/10] vfio/spapr: Extend VFIOIOMMUOps with a release handler Cédric Le Goater
2023-12-11  5:51   ` Duan, Zhenzhong
2023-12-08  8:45 ` [PATCH for-9.0 02/10] vfio/container: Introduce vfio_legacy_setup() for further cleanups Cédric Le Goater
2023-12-11  5:53   ` Duan, Zhenzhong
2023-12-08  8:45 ` [PATCH for-9.0 03/10] vfio/container: Initialize VFIOIOMMUOps under vfio_init_container() Cédric Le Goater
2023-12-11  5:59   ` Duan, Zhenzhong
2023-12-11 13:57     ` Cédric Le Goater
2023-12-12  4:15       ` Duan, Zhenzhong
2023-12-08  8:45 ` [PATCH for-9.0 04/10] vfio/container: Introduce a VFIOIOMMU QOM interface Cédric Le Goater
2023-12-11  6:08   ` Duan, Zhenzhong
2023-12-11 13:58     ` Cédric Le Goater
2023-12-08  8:45 ` [PATCH for-9.0 05/10] vfio/container: Introduce a VFIOIOMMU legacy " Cédric Le Goater
2023-12-11  6:14   ` Duan, Zhenzhong
2023-12-11 14:00     ` Cédric Le Goater [this message]
2023-12-08  8:45 ` [PATCH for-9.0 06/10] vfio/container: Intoduce a new VFIOIOMMUClass::setup handler Cédric Le Goater
2023-12-11  6:17   ` Duan, Zhenzhong
2023-12-08  8:45 ` [PATCH for-9.0 07/10] vfio/spapr: Introduce a sPAPR VFIOIOMMU QOM interface Cédric Le Goater
2023-12-11  6:25   ` Duan, Zhenzhong
2023-12-11 14:01     ` Cédric Le Goater
2023-12-08  8:45 ` [PATCH for-9.0 08/10] vfio/iommufd: Introduce a VFIOIOMMU iommufd " Cédric Le Goater
2023-12-11  6:32   ` Duan, Zhenzhong
2023-12-08  8:45 ` [PATCH for-9.0 09/10] vfio/spapr: Only compile sPAPR IOMMU support when needed Cédric Le Goater
2023-12-11  6:34   ` Duan, Zhenzhong
2023-12-08  8:46 ` [PATCH for-9.0 10/10] vfio/iommufd: Remove CONFIG_IOMMUFD usage Cédric Le Goater
2023-12-11  6:44   ` Duan, Zhenzhong

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=6f47ba3d-2b87-4bf6-84af-e0eb5e78b19a@redhat.com \
    --to=clg@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=eric.auger@redhat.com \
    --cc=harshpb@linux.ibm.com \
    --cc=npiggin@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --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).