On Mon, Feb 10, 2025 at 04:56:43PM -0800, Keith Busch wrote:
> From: Keith Busch <kbusch@xxxxxxxxxx>
>
> Provide an interface for the kernel to leverage the existing
> pre-registered buffers that io_uring provides. User space can reference
> these later to achieve zero-copy IO.
>
> User space must register an empty fixed buffer table with io_uring in
> order for the kernel to make use of it.
>
> Signed-off-by: Keith Busch <kbusch@xxxxxxxxxx>
> ---
...
>
> +int io_buffer_register_bvec(struct io_ring_ctx *ctx, struct request *rq,
> + void (*release)(void *), unsigned int index)
> +{
> + struct io_rsrc_data *data = &ctx->buf_table;
> + struct req_iterator rq_iter;
> + struct io_mapped_ubuf *imu;
> + struct io_rsrc_node *node;
> + struct bio_vec bv;
> + u16 nr_bvecs;
> + int i = 0;
> +
> + lockdep_assert_held(&ctx->uring_lock);
> +
> + if (!data->nr)
> + return -EINVAL;
> + if (index >= data->nr)
> + return -EINVAL;
> +
> + node = data->nodes[index];
> + if (node)
> + return -EBUSY;
> +
> + node = io_rsrc_node_alloc(IORING_RSRC_KBUFFER);
> + if (!node)
> + return -ENOMEM;
> +
> + node->release = release;
> + node->priv = rq;
> +
> + nr_bvecs = blk_rq_nr_phys_segments(rq);
> + imu = kvmalloc(struct_size(imu, bvec, nr_bvecs), GFP_KERNEL);
> + if (!imu) {
> + kfree(node);
> + return -ENOMEM;
> + }
> +
> + imu->ubuf = 0;
> + imu->len = blk_rq_bytes(rq);
> + imu->acct_pages = 0;
> + imu->nr_bvecs = nr_bvecs;
> + refcount_set(&imu->refs, 1);
> + node->buf = imu;
request buffer direction needs to be stored in `imu`, for READ,
the buffer is write-only, and for WRITE, the buffer is read-only,
which isn't different with user mapped buffer.
Meantime in read_fixed/write_fixed side or buffer lookup abstraction
helper, the buffer direction needs to be validated.
Thanks,
Ming
