public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Eric Auger <eric.auger@redhat.com>
To: Yi Liu <yi.l.liu@intel.com>, Kevin Tian <kevin.tian@intel.com>,
	Zhenyu Wang <zhenyuw@linux.intel.com>,
	Zhi Wang <zhi.a.wang@intel.com>,
	Jani Nikula <jani.nikula@linux.intel.com>,
	Joonas Lahtinen <joonas.lahtinen@linux.intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>,
	Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>,
	David Airlie <airlied@linux.ie>, Daniel Vetter <daniel@ffwll.ch>,
	Eric Farman <farman@linux.ibm.com>,
	Matthew Rosato <mjrosato@linux.ibm.com>,
	Halil Pasic <pasic@linux.ibm.com>,
	Vineeth Vijayan <vneethv@linux.ibm.com>,
	Peter Oberparleiter <oberpar@linux.ibm.com>,
	Heiko Carstens <hca@linux.ibm.com>,
	Vasily Gorbik <gor@linux.ibm.com>,
	Alexander Gordeev <agordeev@linux.ibm.com>,
	Christian Borntraeger <borntraeger@linux.ibm.com>,
	Sven Schnelle <svens@linux.ibm.com>,
	Tony Krowiak <akrowiak@linux.ibm.com>,
	Jason Herne <jjherne@linux.ibm.com>,
	Harald Freudenberger <freude@linux.ibm.com>,
	Diana Craciun <diana.craciun@oss.nxp.com>,
	Alex Williamson <alex.williamson@redhat.com>,
	Cornelia Huck <cohuck@redhat.com>,
	Longfang Liu <liulongfang@huawei.com>,
	Shameer Kolothum <shameerali.kolothum.thodi@huawei.com>,
	Jason Gunthorpe <jgg@ziepe.ca>, Yishai Hadas <yishaih@nvidia.com>,
	Kirti Wankhede <kwankhede@nvidia.com>,
	Leon Romanovsky <leon@kernel.org>,
	Abhishek Sahu <abhsahu@nvidia.com>,
	intel-gvt-dev@lists.freedesktop.org,
	intel-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org,
	linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org,
	kvm@vger.kernel.org
Subject: Re: [PATCH v2 15/15] vfio: Add struct device to vfio_device
Date: Thu, 8 Sep 2022 11:39:07 +0200	[thread overview]
Message-ID: <4c9350cd-c2ce-dc84-9a29-210907d2a2a2@redhat.com> (raw)
In-Reply-To: <546463b8-54fa-2071-6a9a-e4087eb8bb2c@intel.com>



On 9/8/22 11:17, Yi Liu wrote:
> On 2022/9/8 17:06, Eric Auger wrote:
>> Hi Kevin,
>>
>> On 9/1/22 16:37, Kevin Tian wrote:
>>> From: Yi Liu <yi.l.liu@intel.com>
>>>
>>> and replace kref. With it a 'vfio-dev/vfioX' node is created under the
>>> sysfs path of the parent, indicating the device is bound to a vfio
>>> driver, e.g.:
>>>
>>> /sys/devices/pci0000\:6f/0000\:6f\:01.0/vfio-dev/vfio0
>>>
>>> It is also a preparatory step toward adding cdev for supporting future
>>> device-oriented uAPI.
>>>
>>> Add Documentation/ABI/testing/sysfs-devices-vfio-dev.
>>>
>>> Also take this chance to rename chardev 'vfio' to 'vfio-group' in
>>> /proc/devices.
>>>
>>> Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
>>> Signed-off-by: Yi Liu <yi.l.liu@intel.com>
>>> Signed-off-by: Kevin Tian <kevin.tian@intel.com>
>>> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
>>> ---
>>>   .../ABI/testing/sysfs-devices-vfio-dev        |  8 +++
>>>   drivers/vfio/vfio_main.c                      | 67
>>> +++++++++++++++----
>>>   include/linux/vfio.h                          |  6 +-
>>>   3 files changed, 66 insertions(+), 15 deletions(-)
>>>   create mode 100644 Documentation/ABI/testing/sysfs-devices-vfio-dev
>>>
>>> diff --git a/Documentation/ABI/testing/sysfs-devices-vfio-dev
>>> b/Documentation/ABI/testing/sysfs-devices-vfio-dev
>>> new file mode 100644
>>> index 000000000000..e21424fd9666
>>> --- /dev/null
>>> +++ b/Documentation/ABI/testing/sysfs-devices-vfio-dev
>>> @@ -0,0 +1,8 @@
>>> +What:         /sys/.../<device>/vfio-dev/vfioX/
>>> +Date:         September 2022
>>> +Contact:     Yi Liu <yi.l.liu@intel.com>
>>> +Description:
>>> +         This directory is created when the device is bound to a
>>> +         vfio driver. The layout under this directory matches what
>>> +         exists for a standard 'struct device'. 'X' is a unique
>>> +         index marking this device in vfio.
>>> diff --git a/drivers/vfio/vfio_main.c b/drivers/vfio/vfio_main.c
>>> index bfa675d314ab..141f55c3faf5 100644
>>> --- a/drivers/vfio/vfio_main.c
>>> +++ b/drivers/vfio/vfio_main.c
>>> @@ -46,6 +46,8 @@ static struct vfio {
>>>       struct mutex            group_lock; /* locks group_list */
>>>       struct ida            group_ida;
>>>       dev_t                group_devt;
>>> +    struct class            *device_class;
>>> +    struct ida            device_ida;
>>>   } vfio;
>>>     struct vfio_iommu_driver {
>>> @@ -483,12 +485,13 @@ static struct vfio_device
>>> *vfio_group_get_device(struct vfio_group *group,
>>>    * VFIO driver API
>>>    */
>>>   /* Release helper called by vfio_put_device() */
>>> -void vfio_device_release(struct kref *kref)
>>> +static void vfio_device_release(struct device *dev)
>>>   {
>>>       struct vfio_device *device =
>>> -            container_of(kref, struct vfio_device, kref);
>>> +            container_of(dev, struct vfio_device, device);
>>>         vfio_release_device_set(device);
>>> +    ida_free(&vfio.device_ida, device->index);
>>>         /*
>>>        * kvfree() cannot be done here due to a life cycle mess in
>>> @@ -498,7 +501,6 @@ void vfio_device_release(struct kref *kref)
>>>        */
>>>       device->ops->release(device);
>>>   }
>>> -EXPORT_SYMBOL_GPL(vfio_device_release);
>>>     /*
>>>    * Alloc and initialize vfio_device so it can be registered to vfio
>>> @@ -546,6 +548,13 @@ int vfio_init_device(struct vfio_device
>>> *device, struct device *dev,
>>>   {
>>>       int ret;
>>>   +    ret = ida_alloc_max(&vfio.device_ida, MINORMASK, GFP_KERNEL);
>>> +    if (ret < 0) {
>>> +        dev_dbg(dev, "Error to alloc index\n");
>>> +        return ret;
>>> +    }
>>> +
>>> +    device->index = ret;
>>>       init_completion(&device->comp);
>>>       device->dev = dev;
>>>       device->ops = ops;
>>> @@ -556,11 +565,15 @@ int vfio_init_device(struct vfio_device
>>> *device, struct device *dev,
>>>               goto out_uninit;
>>>       }
>>>   -    kref_init(&device->kref);
>>> +    device_initialize(&device->device);
>>> +    device->device.release = vfio_device_release;
>>> +    device->device.class = vfio.device_class;
>>> +    device->device.parent = device->dev;
>>>       return 0;
>>>     out_uninit:
>>>       vfio_release_device_set(device);
>>> +    ida_free(&vfio.device_ida, device->index);
>>>       return ret;
>>>   }
>>>   EXPORT_SYMBOL_GPL(vfio_init_device);
>>> @@ -657,6 +670,7 @@ static int __vfio_register_dev(struct
>>> vfio_device *device,
>>>           struct vfio_group *group)
>>>   {
>>>       struct vfio_device *existing_device;
>>> +    int ret;
>>>         if (IS_ERR(group))
>>>           return PTR_ERR(group);
>>> @@ -673,16 +687,21 @@ static int __vfio_register_dev(struct
>>> vfio_device *device,
>>>           dev_WARN(device->dev, "Device already exists on group %d\n",
>>>                iommu_group_id(group->iommu_group));
>>>           vfio_device_put_registration(existing_device);
>>> -        if (group->type == VFIO_NO_IOMMU ||
>>> -            group->type == VFIO_EMULATED_IOMMU)
>>> -            iommu_group_remove_device(device->dev);
>>> -        vfio_group_put(group);
>>> -        return -EBUSY;
>>> +        ret = -EBUSY;
>>> +        goto err_out;
>>>       }
>>>         /* Our reference on group is moved to the device */
>>>       device->group = group;
>>>   +    ret = dev_set_name(&device->device, "vfio%d", device->index);
>>> +    if (ret)
>>> +        goto err_out;
>>> +
>>> +    ret = device_add(&device->device);
>>> +    if (ret)
>>> +        goto err_out;
>>> +
>>>       /* Refcounting can't start until the driver calls register */
>>>       refcount_set(&device->refcount, 1);
>>>   @@ -692,6 +711,12 @@ static int __vfio_register_dev(struct
>>> vfio_device *device,
>>>       mutex_unlock(&group->device_lock);
>>>         return 0;
>>> +err_out:
>>> +    if (group->type == VFIO_NO_IOMMU ||
>>> +        group->type == VFIO_EMULATED_IOMMU)
>>> +        iommu_group_remove_device(device->dev);
>>> +    vfio_group_put(group);
>>> +    return ret;
>>>   }
>>>     int vfio_register_group_dev(struct vfio_device *device)
>>> @@ -779,6 +804,9 @@ void vfio_unregister_group_dev(struct
>>> vfio_device *device)
>>>       group->dev_counter--;
>>>       mutex_unlock(&group->device_lock);
>>>   +    /* Balances device_add in register path */
>>> +    device_del(&device->device);
>>> +
>>>       if (group->type == VFIO_NO_IOMMU || group->type ==
>>> VFIO_EMULATED_IOMMU)
>>>           iommu_group_remove_device(device->dev);
>>>   @@ -2145,6 +2173,7 @@ static int __init vfio_init(void)
>>>       int ret;
>>>         ida_init(&vfio.group_ida);
>>> +    ida_init(&vfio.device_ida);
>>>       mutex_init(&vfio.group_lock);
>>>       mutex_init(&vfio.iommu_drivers_lock);
>>>       INIT_LIST_HEAD(&vfio.group_list);
>>> @@ -2160,12 +2189,20 @@ static int __init vfio_init(void)
>>>       vfio.class = class_create(THIS_MODULE, "vfio");
>>>       if (IS_ERR(vfio.class)) {
>>>           ret = PTR_ERR(vfio.class);
>>> -        goto err_class;
>>> +        goto err_group_class;
>>>       }
>>>         vfio.class->devnode = vfio_devnode;
>>>   -    ret = alloc_chrdev_region(&vfio.group_devt, 0, MINORMASK + 1,
>>> "vfio");
>>> +    /* /sys/class/vfio-dev/vfioX */
>>> +    vfio.device_class = class_create(THIS_MODULE, "vfio-dev");
>>> +    if (IS_ERR(vfio.device_class)) {
>>> +        ret = PTR_ERR(vfio.device_class);
>>> +        goto err_dev_class;
>>> +    }
>>> +
>>> +    ret = alloc_chrdev_region(&vfio.group_devt, 0, MINORMASK + 1,
>>> +                  "vfio-group");
>>>       if (ret)
>>>           goto err_alloc_chrdev;
>>>   @@ -2181,9 +2218,12 @@ static int __init vfio_init(void)
>>>   err_driver_register:
>>>       unregister_chrdev_region(vfio.group_devt, MINORMASK + 1);
>>>   err_alloc_chrdev:
>>> +    class_destroy(vfio.device_class);
>>> +    vfio.device_class = NULL;
>>> +err_dev_class:
>>>       class_destroy(vfio.class);
>>>       vfio.class = NULL;
>>> -err_class:
>>> +err_group_class:
>>>       misc_deregister(&vfio_dev);
>>>       return ret;
>>>   }
>>> @@ -2195,8 +2235,11 @@ static void __exit vfio_cleanup(void)
>>>   #ifdef CONFIG_VFIO_NOIOMMU
>>>       vfio_unregister_iommu_driver(&vfio_noiommu_ops);
>>>   #endif
>>> +    ida_destroy(&vfio.device_ida);
>>>       ida_destroy(&vfio.group_ida);
>>>       unregister_chrdev_region(vfio.group_devt, MINORMASK + 1);
>>> +    class_destroy(vfio.device_class);
>>> +    vfio.device_class = NULL;
>>>       class_destroy(vfio.class);
>>>       vfio.class = NULL;
>>>       misc_deregister(&vfio_dev);
>>> diff --git a/include/linux/vfio.h b/include/linux/vfio.h
>>> index f03447c8774d..5c13f74da1bb 100644
>>> --- a/include/linux/vfio.h
>>> +++ b/include/linux/vfio.h
>>> @@ -45,7 +45,8 @@ struct vfio_device {
>>>       struct kvm *kvm;
>>>         /* Members below here are private, not for driver use */
>>> -    struct kref kref;    /* object life cycle */
>>> +    unsigned int index;
>>> +    struct device device;    /* device.kref covers object life
>>> circle */
>>>       refcount_t refcount;    /* user count on registered device*/
>>>       unsigned int open_count;
>>>       struct completion comp;
>> I am not totally clear about remaining 'struct device *dev;' in
>> vfio_device struct. I see it used in some places. Is it supposed to
>> disappear at some point?
>
> no, Eric. *dev will not disappear, it stores the dev pointet passed in by
> caller of vfio_init_device().

yeah I see but you have device->device.parent = device->dev;

Eric
>
>>> @@ -154,10 +155,9 @@ struct vfio_device *_vfio_alloc_device(size_t
>>> size, struct device *dev,
>>>   int vfio_init_device(struct vfio_device *device, struct device *dev,
>>>                const struct vfio_device_ops *ops);
>>>   void vfio_free_device(struct vfio_device *device);
>>> -void vfio_device_release(struct kref *kref);
>>>   static inline void vfio_put_device(struct vfio_device *device)
>>>   {
>>> -    kref_put(&device->kref, vfio_device_release);
>>> +    put_device(&device->device);
>>>   }
>>>     int vfio_register_group_dev(struct vfio_device *device);
>>
>> Thanks
>>
>> Eric
>>
>


  reply	other threads:[~2022-09-08  9:39 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-01 14:37 [PATCH v2 00/15] Tidy up vfio_device life cycle Kevin Tian
2022-09-01 14:37 ` [PATCH v2 01/15] vfio: Add helpers for unifying " Kevin Tian
2022-09-06  9:41   ` Christoph Hellwig
2022-09-07  0:43     ` Tian, Kevin
2022-09-07 11:55       ` Christoph Hellwig
2022-09-07 12:11         ` Jason Gunthorpe
2022-09-07 19:28   ` Eric Auger
2022-09-08  6:19     ` Tian, Kevin
2022-09-08  9:08       ` Eric Auger
2022-09-01 14:37 ` [PATCH v2 02/15] vfio/pci: Use the new device life cycle helpers Kevin Tian
2022-09-01 14:37 ` [PATCH v2 03/15] vfio/mlx5: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 04/15] vfio/hisi_acc: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 05/15] vfio/mdpy: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 06/15] vfio/mtty: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 07/15] vfio/mbochs: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 08/15] drm/i915/gvt: " Kevin Tian
2022-09-07  3:17   ` Zhenyu Wang
2022-09-01 14:37 ` [PATCH v2 09/15] vfio/ap: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 10/15] vfio/fsl-mc: " Kevin Tian
2022-09-01 14:37 ` [PATCH v2 11/15] vfio/platform: " Kevin Tian
2022-09-07 19:28   ` Eric Auger
2022-09-01 14:37 ` [PATCH v2 12/15] vfio/amba: " Kevin Tian
2022-09-01  7:37   ` Tian, Kevin
2022-09-07 19:32   ` Eric Auger
2022-09-01 14:37 ` [PATCH v2 13/15] vfio/ccw: " Kevin Tian
2022-09-08  7:19   ` Tian, Kevin
2022-09-08 20:50     ` Eric Farman
2022-09-09  1:52       ` Tian, Kevin
2022-09-01 14:37 ` [PATCH v2 14/15] vfio: Rename vfio_device_put() and vfio_device_try_get() Kevin Tian
2022-09-07 19:35   ` Eric Auger
2022-09-01 14:37 ` [PATCH v2 15/15] vfio: Add struct device to vfio_device Kevin Tian
2022-09-01  7:40   ` Tian, Kevin
2022-09-08  9:06   ` Eric Auger
2022-09-08  9:17     ` Yi Liu
2022-09-08  9:39       ` Eric Auger [this message]
2022-09-08 12:37         ` Jason Gunthorpe
2022-09-09  3:09           ` Tian, Kevin

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=4c9350cd-c2ce-dc84-9a29-210907d2a2a2@redhat.com \
    --to=eric.auger@redhat.com \
    --cc=abhsahu@nvidia.com \
    --cc=agordeev@linux.ibm.com \
    --cc=airlied@linux.ie \
    --cc=akrowiak@linux.ibm.com \
    --cc=alex.williamson@redhat.com \
    --cc=borntraeger@linux.ibm.com \
    --cc=cohuck@redhat.com \
    --cc=daniel@ffwll.ch \
    --cc=diana.craciun@oss.nxp.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=farman@linux.ibm.com \
    --cc=freude@linux.ibm.com \
    --cc=gor@linux.ibm.com \
    --cc=hca@linux.ibm.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-gvt-dev@lists.freedesktop.org \
    --cc=jani.nikula@linux.intel.com \
    --cc=jgg@ziepe.ca \
    --cc=jjherne@linux.ibm.com \
    --cc=joonas.lahtinen@linux.intel.com \
    --cc=kevin.tian@intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=kwankhede@nvidia.com \
    --cc=leon@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=liulongfang@huawei.com \
    --cc=mjrosato@linux.ibm.com \
    --cc=oberpar@linux.ibm.com \
    --cc=pasic@linux.ibm.com \
    --cc=rodrigo.vivi@intel.com \
    --cc=shameerali.kolothum.thodi@huawei.com \
    --cc=svens@linux.ibm.com \
    --cc=tvrtko.ursulin@linux.intel.com \
    --cc=vneethv@linux.ibm.com \
    --cc=yi.l.liu@intel.com \
    --cc=yishaih@nvidia.com \
    --cc=zhenyuw@linux.intel.com \
    --cc=zhi.a.wang@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