Re: [PATCH v1 03/24] vfio-user: add container IO ops vector

["Philippe Mathieu-Daudé" <philmd@linaro.org>]

On 9/12/22 17:10, Cédric Le Goater wrote:
> Hello John,
> 
> On 11/9/22 00:13, John Johnson wrote:
>> Used for communication with VFIO driver
>> (prep work for vfio-user, which will communicate over a socket)
>>
>> Signed-off-by: John G Johnson <john.g.johnson@oracle.com>
>> ---
>>   hw/vfio/common.c              | 126 
>> ++++++++++++++++++++++++++++--------------
>>   include/hw/vfio/vfio-common.h |  33 +++++++++++
>>   2 files changed, 117 insertions(+), 42 deletions(-)
>>
>> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
>> index ace9562..83d69b9 100644
>> --- a/hw/vfio/common.c
>> +++ b/hw/vfio/common.c
>> @@ -432,12 +432,12 @@ static int vfio_dma_unmap_bitmap(VFIOContainer 
>> *container,
>>           goto unmap_exit;
>>       }
>> -    ret = ioctl(container->fd, VFIO_IOMMU_UNMAP_DMA, unmap);
>> +    ret = CONT_DMA_UNMAP(container, unmap, bitmap);
> 
> I am not sure these macros are very useful, compared to :
> 
>      container->ops->dma_unmap(container, unmap, bitmap);

I was going to report the same.


>> +/*
>> + * The next 2 ops vectors are how Devices and Containers
>> + * communicate with the server.  The default option is
>> + * through ioctl() to the kernel VFIO driver, but vfio-user
>> + * can use a socket to a remote process.
>> + */
>> +
>> +struct VFIOContIO {
> 
> VFIOContainerOps seems more adequate with the current VFIO terminology
> in QEMU.

Yes please, abbreviated "Cont" is not helpful.