On Sun, Jun 14, 2020 at 09:52:30AM -0600, Jens Axboe wrote:
> On 6/9/20 8:24 AM, Stefano Garzarella wrote:
> > Hi Jens,
> > Stefan and I have a proposal to share with io_uring community.
> > Before implementing it we would like to discuss it to receive feedbacks and
> > to see if it could be accepted:
> >
> > Adding restrictions to io_uring
> > =====================================
> > The io_uring API provides submission and completion queues for performing
> > asynchronous I/O operations. The queues are located in memory that is
> > accessible to both the host userspace application and the kernel, making it
> > possible to monitor for activity through polling instead of system calls. This
> > design offers good performance and this makes exposing io_uring to guests an
> > attractive idea for improving I/O performance in virtualization.
> >
> > PoC and preliminary benchmarks
> > ---------------------------
> > We realized a PoC, using QEMU and virtio-blk device, to share io_uring
> > CQ and SQ rings with the guest.
> > QEMU initializes io_uring, registers the device (NVMe) fd through
> > io_uring_register(2), and maps the rings in the guest memory.
> > The virtio-blk driver uses these rings to send requests instead of using
> > the standard virtqueues.
> >
> > The PoC implements a pure polling solution where the application is polling
> > (IOPOLL enabled) in the guest and the sqpoll_kthread is polling in the host
> > (SQPOLL and IOPOLL enabled).
> >
> > These are the encouraging results we obtained from this preliminary work;
> > we used fio (rw=randread bs=4k) to measure the kIOPS on a NVMe device:
> >
> > - bare-metal
> > iodepth
> > | fio ioengine | 1 | 8 | 16 | 32 |
> > |-------------------------------------------|----:|----:|----:|----:|
> > | io_uring (SQPOLL + IOPOLL) | 119 | 550 | 581 | 585 |
> > | io_uring (IOPOLL) | 122 | 502 | 519 | 538 |
> >
> > - QEMU/KVM guest (aio=io_uring)
> > iodepth
> > | virtio-blk | fio ioengine | 1 | 8 | 16 | 32 |
> > |-----------------------|-------------------|----:|----:|----:|----:|
> > | virtqueues | io_uring (IOPOLL) | 27 | 144 | 209 | 266 |
> > | virtqueues + iothread | io_uring (IOPOLL) | 73 | 264 | 306 | 312 |
> > | io_uring passthrough | io_uring (IOPOLL) | 104 | 532 | 577 | 585 |
> >
> > All guest experiments are using the QEMU io_uring backend with SQPOLL and
> > IOPOLL enabled. The virtio-blk driver is modified to support blovk io_poll
> > on both virtqueues and io_uring passthrough.
> >
> > Before developing this proof-of-concept further we would like to discuss
> > io_uring changes required to restrict rings since this mechanism is a
> > prerequisite for real-world use cases where guests are untrusted.
> >
> > Restrictions
> > ------------
> > This document proposes io_uring API changes that safely allow untrusted
> > applications or guests to use io_uring. io_uring's existing security model is
> > that of kernel system call handler code. It is designed to reject invalid
> > inputs from host userspace applications. Supporting guests as io_uring API
> > clients adds a new trust domain with access to even fewer resources than host
> > userspace applications.
> >
> > Guests do not have direct access to host userspace application file descriptors
> > or memory. The host userspace application, a Virtual Machine Monitor (VMM) such
> > as QEMU, grants access to a subset of its file descriptors and memory. The
> > allowed file descriptors are typically the disk image files belonging to the
> > guest. The memory is typically the virtual machine's RAM that the VMM has
> > allocated on behalf of the guest.
> >
> > The following extensions to the io_uring API allow the host application to
> > grant access to some of its file descriptors.
> >
> > These extensions are designed to be applicable to other use cases besides
> > untrusted guests and are not virtualization-specific. For example, the
> > restrictions can be used to allow only a subset of sqe operations available to
> > an application similar to seccomp syscall whitelisting.
> >
> > An address translation and memory restriction mechanism would also be
> > necessary, but we can discuss this later.
> >
> > The IOURING_REGISTER_RESTRICTIONS opcode
> > ----------------------------------------
> > The new io_uring_register(2) IOURING_REGISTER_RESTRICTIONS opcode permanently
> > installs a feature whitelist on an io_ring_ctx. The io_ring_ctx can then be
> > passed to untrusted code with the knowledge that only operations present in the
> > whitelist can be executed.
> >
> > The whitelist approach ensures that new features added to io_uring do not
> > accidentally become available when an existing application is launched on a
> > newer kernel version.
> >
> > The IORING_REGISTER_RESTRICTIONS opcode takes an array of struct
> > io_uring_restriction elements that describe whitelisted features:
> >
> > #define IORING_REGISTER_RESTRICTIONS 11
> >
> > /* struct io_uring_restriction::opcode values */
> > enum {
> > /* Allow an io_uring_register(2) opcode */
> > IORING_RESTRICTION_REGISTER_OP,
> >
> > /* Allow an sqe opcode */
> > IORING_RESTRICTION_SQE_OP,
> >
> > /* Only allow fixed files */
> > IORING_RESTRICTION_FIXED_FILES_ONLY,
> >
> > /* Only allow registered addresses and translate them */
> > IORING_RESTRICTION_BUFFER_CHECK
> > };
> >
> > struct io_uring_restriction {
> > __u16 opcode;
> > union {
> > __u8 register_op; /* IORING_RESTRICTION_REGISTER_OP */
> > __u8 sqe_op; /* IORING_RESTRICTION_SQE_OP */
> > };
> > __u8 resv;
> > __u32 resv2[3];
> > };
> >
> > This call can only be made once. Afterwards it is not possible to change
> > restrictions anymore. This prevents untrusted code from removing restrictions.
> >
> > Limiting access to io_uring operations
> > --------------------------------------
> > The following example shows how to whitelist IORING_OP_READV, IORING_OP_WRITEV,
> > and IORING_OP_FSYNC:
> >
> > struct io_uring_restriction restrictions[] = {
> > {
> > .opcode = IORING_RESTRICTION_SQE_OP,
> > .sqe_op = IORING_OP_READV,
> > },
> > {
> > .opcode = IORING_RESTRICTION_SQE_OP,
> > .sqe_op = IORING_OP_WRITEV,
> > },
> > {
> > .opcode = IORING_RESTRICTION_SQE_OP,
> > .sqe_op = IORING_OP_FSYNC,
> > },
> > ...
> > };
> >
> > io_uring_register(ringfd, IORING_REGISTER_RESTRICTIONS,
> > restrictions, ARRAY_SIZE(restrictions));
> >
> > Limiting access to file descriptors
> > -----------------------------------
> > The fixed files mechanism can be used to limit access to a set of file
> > descriptors:
> >
> > struct io_uring_restriction restrictions[] = {
> > {
> > .opcode = IORING_RESTRICTION_FIXED_FILES_ONLY,
> > },
> > ...
> > };
> >
> > io_uring_register(ringfd, IORING_REGISTER_RESTRICTIONS,
> > restrictions, ARRAY_SIZE(restrictions));
> >
> > Only requests with the sqe->flags IOSQE_FIXED_FILE bit set will be allowed.
>
> I don't think this sounds unreasonable, but I'd really like to see a
> prototype hacked up before rendering any further opinions on it :-)
Yeah :-) I'll be back with a prototype of this changes ASAP.
Thanks for you feedback,
Stefano
