Re: [PATCH v3 03/10] vfio/iommufd: Return errno in iommufd_cdev_attach_ioas_hwpt()

[Joao Martins]

On 08/07/2024 16:32, Joao Martins wrote:
> On 08/07/2024 16:28, Cédric Le Goater wrote:
>> Hello Joao,
>>
>> On 7/8/24 4:34 PM, Joao Martins wrote:
>>> In preparation to implement auto domains have the attach function
>>> return the errno it got during domain attach instead of a bool.
>>>
>>> -EINVAL is tracked to track domain incompatibilities, and decide whether
>>> to create a new IOMMU domain.
>>
>> Please leave the return value as a bool unless there is a very
>> good reason not to.
>>
> 
> Error* doesn't store the errno, and thus I can't actually test the number of
> errno to know when to bail to the next hwpt. Maybe the commit message wasn't
> clear enough there. But not sure if we have an alternative here? Or maybe 
> Error
> does store errno, and I totally missed it.

Or I can just use 'errno' and keep the same return value as bool. But I didn't
do that because we are purposedly calling error_set*(errp, errno) with the errno
value that it felt the right thing to just return it. Also, considering how hard
the code in util/error.c saves/restores errno in all the helpers.

        Joao