Re: [PATCH v5 0/3] io_uring: add restrictions to support untrusted applications and guests

[Jens Axboe]

On 8/27/20 8:41 AM, Stefano Garzarella wrote:
> On Thu, Aug 27, 2020 at 08:10:49AM -0600, Jens Axboe wrote:
>> On 8/27/20 8:10 AM, Stefano Garzarella wrote:
>>> On Thu, Aug 27, 2020 at 07:50:44AM -0600, Jens Axboe wrote:
>>>> On 8/27/20 7:40 AM, Stefano Garzarella wrote:
>>>>> v5:
>>>>>  - explicitly assigned enum values [Kees]
>>>>>  - replaced kmalloc/copy_from_user with memdup_user [kernel test robot]
>>>>>  - added Kees' R-b tags
>>>>>
>>>>> v4: https://lore.kernel.org/io-uring/20200813153254.93731-1-sgarzare@xxxxxxxxxx/
>>>>> v3: https://lore.kernel.org/io-uring/20200728160101.48554-1-sgarzare@xxxxxxxxxx/
>>>>> RFC v2: https://lore.kernel.org/io-uring/20200716124833.93667-1-sgarzare@xxxxxxxxxx
>>>>> RFC v1: https://lore.kernel.org/io-uring/20200710141945.129329-1-sgarzare@xxxxxxxxxx
>>>>>
>>>>> Following the proposal that I send about restrictions [1], I wrote this series
>>>>> to add restrictions in io_uring.
>>>>>
>>>>> I also wrote helpers in liburing and a test case (test/register-restrictions.c)
>>>>> available in this repository:
>>>>> https://github.com/stefano-garzarella/liburing (branch: io_uring_restrictions)
>>>>>
>>>>> Just to recap the proposal, the idea is to add some restrictions to the
>>>>> operations (sqe opcode and flags, register opcode) to safely allow untrusted
>>>>> applications or guests to use io_uring queues.
>>>>>
>>>>> The first patch changes io_uring_register(2) opcodes into an enumeration to
>>>>> keep track of the last opcode available.
>>>>>
>>>>> The second patch adds IOURING_REGISTER_RESTRICTIONS opcode and the code to
>>>>> handle restrictions.
>>>>>
>>>>> The third patch adds IORING_SETUP_R_DISABLED flag to start the rings disabled,
>>>>> allowing the user to register restrictions, buffers, files, before to start
>>>>> processing SQEs.
>>>>>
>>>>> Comments and suggestions are very welcome.
>>>>
>>>> Looks good to me, just a few very minor comments in patch 2. If you
>>>> could fix those up, let's get this queued for 5.10.
>>>>
>>>
>>> Sure, I'll fix the issues. This is great :-)
>>
>> Thanks! I'll pull in your liburing tests as well once we get the kernel
>> side sorted.
> 
> Yeah. Let me know if you'd prefer that I send patches on io-uring ML.
> 
> About io-uring UAPI, do you think we should set explicitly the enum
> values also for IOSQE_*_BIT and IORING_OP_*?
> 
> I can send a separated patch for this.

No, I actually think that change was a little bit silly. If you
inadvertently renumber the enum in a patch, then tests would fail left
and right. Hence I don't think this is a real risk. I'm fine with doing
it for the addition, but doing it for the others is just going to cause
stable headaches for patches.

-- 
Jens Axboe