RE: [PATCH v4 26/41] backends/iommufd: Introduce the iommufd object

[Duan, Zhenzhong]

>-----Original Message-----
>From: Cédric Le Goater <clg@redhat.com>
>Sent: Wednesday, November 8, 2023 5:41 PM
>Subject: Re: [PATCH v4 26/41] backends/iommufd: Introduce the iommufd object
>
>>>> +                              hwaddr iova, ram_addr_t size)
>>>> +{
>>>> +    int ret;
>>>> +    struct iommu_ioas_unmap unmap = {
>>>> +        .size = sizeof(unmap),
>>>> +        .ioas_id = ioas_id,
>>>> +        .iova = iova,
>>>> +        .length = size,
>>>> +    };
>>>> +
>>>> +    ret = ioctl(be->fd, IOMMU_IOAS_UNMAP, &unmap);
>>>> +    trace_iommufd_backend_unmap_dma(be->fd, ioas_id, iova, size, ret);
>>>> +    /*
>>>> +     * TODO: IOMMUFD doesn't support mapping PCI BARs for now.
>>>> +     * It's not a problem if there is no p2p dma, relax it here
>>>> +     * and avoid many noisy trigger from vIOMMU side.
>>>
>>> Should we add a warn_report() ?
>>
>> The purpose of checking "ret && errno == ENOENT" is to avoid many
>> error_report() for PCI BARs, If we add warn_report(), there will still be
>> many print for PCI BARs.
>
>a trace event then ?

Good idea, will do.

Thanks
Zhenzhong