Re: [PATCH] vfio: platform: use vfio_iommu_group_get/put

[Eric Auger]

Hi Peng,
On 05/20/2016 10:36 AM, Peng Fan wrote:
> Hi,
> 
> On Fri, May 20, 2016 at 10:20:32AM +0200, Baptiste Reynal wrote:
>> Hi,
>>
>> On Fri, May 20, 2016 at 9:58 AM, Eric Auger <eric.auger@xxxxxxxxxx> wrote:
>>>
>>> Hi,
>>> On 05/19/2016 11:01 PM, Alex Williamson wrote:
>>>> On Tue, 10 May 2016 15:40:28 +0800
>>>> Peng Fan <van.freenix@xxxxxxxxx> wrote:
>>>>
>>>>> Hi Alex,
>>>>>
>>>>> On Mon, May 09, 2016 at 09:32:38AM -0600, Alex Williamson wrote:
>>>>>> On Mon,  9 May 2016 18:01:43 +0800
>>>>>> Peng Fan <van.freenix@xxxxxxxxx> wrote:
>>>>>>
>>>>>>> Use vfio_iommu_group_get and vfio_iommu_group_put, but not
>>>>>>> iommu_group_get or iommu_group_put.
>>>>>>
>>>>>> I assume you're trying to duplicate the vfio_pci changes from commit
>>>>>> 03a76b60f8ba to enable no-iommu mode.  That would be really relevant
>>>>>> information for the commit log.
>>>>>
>>>>> This is not to support non-iommu for vfio platform. I just think
>>>>> vfio_iommu_group_get/put is vfio core API and should be used by
>>>>> vfio-pci and vfio-platform.
>>>>
>>>> Hi Peng,
>>>>
>>>> I suppose I would consider this an optional part of the internal vfio
>>>> API, it's only real purpose it to provide the ability to create fake
>>>> groups which are only used for no-iommu.  It's perfectly legitimate to
>>>> use iommu_group_get/put if there is no desire to enable no-iommu.
>>>> Baptiste, Eric, do you have an opinion whether enabling no-iommu in
>>>> vfio/platform is something we should do?  Thanks,
>>> I think it would make sense to introduce that no-iommu feature for
>>> vfio-platform because it allows potential users to get familiar with
>>> VFIO platform without having the proper HW. I also thought the primary
>>> purpose of this patch was to introduce that support.
> 
> Yeah. Initially I would like to try vfio on my i.MX6/7 boards which does not
> have HW iommu. I am new to this, so just wrote this simple patch when I was
> reading the source code.
> 
>>
>> Since safety issues have already been addressed for the PCI patch
>> series, I don't see any reason not to enable it. You might extend the
>> commit message to remind those issues.
> 
> Eric, Baptiste,
> 
> Is vfio-platform no-iommu on your TODO list? I am happy to help test.
> If not, I can try work on this if you could guide me -:)
No it is not at the moment. shouldn't it work already with this very
patch? I will be happy to bring support to you if needed.

Best Regards

Eric
> 
> Thanks,
> Peng.
>>
>> Best regards,
>> Baptiste
>>
>>>
>>> Best Regards
>>>
>>> Eric
>>>>
>>>> Alex
>>>>
>>>>>>>
>>>>>>> Signed-off-by: Peng Fan <van.freenix@xxxxxxxxx>
>>>>>>> Cc: Baptiste Reynal <b.reynal@xxxxxxxxxxxxxxxxxxxxxx>
>>>>>>> Cc: Alex Williamson <alex.williamson@xxxxxxxxxx>
>>>>>>> ---
>>>>>>>  drivers/vfio/platform/vfio_platform_common.c | 6 +++---
>>>>>>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>>>>>>
>>>>>>> diff --git a/drivers/vfio/platform/vfio_platform_common.c b/drivers/vfio/platform/vfio_platform_common.c
>>>>>>> index e65b142..582885e 100644
>>>>>>> --- a/drivers/vfio/platform/vfio_platform_common.c
>>>>>>> +++ b/drivers/vfio/platform/vfio_platform_common.c
>>>>>>> @@ -561,7 +561,7 @@ int vfio_platform_probe_common(struct vfio_platform_device *vdev,
>>>>>>>
>>>>>>>    vdev->device = dev;
>>>>>>>
>>>>>>> -  group = iommu_group_get(dev);
>>>>>>> +  group = vfio_iommu_group_get(dev);
>>>>>>>    if (!group) {
>>>>>>>            pr_err("VFIO: No IOMMU group for device %s\n", vdev->name);
>>>>>>>            return -EINVAL;
>>>>>>> @@ -569,7 +569,7 @@ int vfio_platform_probe_common(struct vfio_platform_device *vdev,
>>>>>>>
>>>>>>>    ret = vfio_add_group_dev(dev, &vfio_platform_ops, vdev);
>>>>>>>    if (ret) {
>>>>>>> -          iommu_group_put(group);
>>>>>>> +          vfio_iommu_group_put(group);
>>>>>>>            return ret;
>>>>>>>    }
>>>>>>>
>>>>>>> @@ -589,7 +589,7 @@ struct vfio_platform_device *vfio_platform_remove_common(struct device *dev)
>>>>>>>
>>>>>>>    if (vdev) {
>>>>>>>            vfio_platform_put_reset(vdev);
>>>>>>> -          iommu_group_put(dev->iommu_group);
>>>>>>> +          vfio_iommu_group_put(dev->iommu_group, dev);
>>>>>>>    }
>>>>>>>
>>>>>>>    return vdev;
>>>>>>
>>>>
>>>