Re: [PATCH net-next v2 2/4] vsock/virtio: support to send non-linear skb

[Arseniy Krasnov <avkrasnov@sberdevices.ru>]

On 19.07.2023 10:36, Stefano Garzarella wrote:
> On Wed, Jul 19, 2023 at 07:46:05AM +0300, Arseniy Krasnov wrote:
>>
>>
>> On 18.07.2023 23:27, Michael S. Tsirkin wrote:
>>> On Tue, Jul 18, 2023 at 09:02:35PM +0300, Arseniy Krasnov wrote:
>>>> For non-linear skb use its pages from fragment array as buffers in
>>>> virtio tx queue. These pages are already pinned by 'get_user_pages()'
>>>> during such skb creation.
>>>>
>>>> Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
>>>> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
>>>> ---
>>>>  net/vmw_vsock/virtio_transport.c | 40 +++++++++++++++++++++++++++-----
>>>>  1 file changed, 34 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c
>>>> index e95df847176b..6cbb45bb12d2 100644
>>>> --- a/net/vmw_vsock/virtio_transport.c
>>>> +++ b/net/vmw_vsock/virtio_transport.c
>>>> @@ -100,7 +100,9 @@ virtio_transport_send_pkt_work(struct work_struct *work)
>>>>      vq = vsock->vqs[VSOCK_VQ_TX];
>>>>
>>>>      for (;;) {
>>>> -        struct scatterlist hdr, buf, *sgs[2];
>>>> +        /* +1 is for packet header. */
>>>> +        struct scatterlist *sgs[MAX_SKB_FRAGS + 1];
>>>> +        struct scatterlist bufs[MAX_SKB_FRAGS + 1];
>>>>          int ret, in_sg = 0, out_sg = 0;
>>>>          struct sk_buff *skb;
>>>>          bool reply;
>>>> @@ -111,12 +113,38 @@ virtio_transport_send_pkt_work(struct work_struct *work)
>>>>
>>>>          virtio_transport_deliver_tap_pkt(skb);
>>>>          reply = virtio_vsock_skb_reply(skb);
>>>> +        sg_init_one(&bufs[out_sg], virtio_vsock_hdr(skb),
>>>> +                sizeof(*virtio_vsock_hdr(skb)));
>>>> +        sgs[out_sg] = &bufs[out_sg];
>>>> +        out_sg++;
>>>> +
>>>> +        if (!skb_is_nonlinear(skb)) {
>>>> +            if (skb->len > 0) {
>>>> +                sg_init_one(&bufs[out_sg], skb->data, skb->len);
>>>> +                sgs[out_sg] = &bufs[out_sg];
>>>> +                out_sg++;
>>>> +            }
>>>> +        } else {
>>>> +            struct skb_shared_info *si;
>>>> +            int i;
>>>> +
>>>> +            si = skb_shinfo(skb);
>>>> +
>>>> +            for (i = 0; i < si->nr_frags; i++) {
>>>> +                skb_frag_t *skb_frag = &si->frags[i];
>>>> +                void *va = page_to_virt(skb_frag->bv_page);
>>>>
>>>> -        sg_init_one(&hdr, virtio_vsock_hdr(skb), sizeof(*virtio_vsock_hdr(skb)));
>>>> -        sgs[out_sg++] = &hdr;
>>>> -        if (skb->len > 0) {
>>>> -            sg_init_one(&buf, skb->data, skb->len);
>>>> -            sgs[out_sg++] = &buf;
>>>> +                /* We will use 'page_to_virt()' for userspace page here,
>>>
>>> don't put comments after code they refer to, please?
>>>
>>>> +                 * because virtio layer will call 'virt_to_phys()' later
>>>
>>> it will but not always. sometimes it's the dma mapping layer.
>>>
>>>
>>>> +                 * to fill buffer descriptor. We don't touch memory at
>>>> +                 * "virtual" address of this page.
>>>
>>>
>>> you need to stick "the" in a bunch of places above.
>>
>> Ok, I'll fix this comment!
>>
>>>
>>>> +                 */
>>>> +                sg_init_one(&bufs[out_sg],
>>>> +                        va + skb_frag->bv_offset,
>>>> +                        skb_frag->bv_len);
>>>> +                sgs[out_sg] = &bufs[out_sg];
>>>> +                out_sg++;
>>>> +            }
>>>>          }
>>>>
>>>>          ret = virtqueue_add_sgs(vq, sgs, out_sg, in_sg, skb, GFP_KERNEL);
>>>
>>>
>>> There's a problem here: if there vq is small this will fail.
>>> So you really should check free vq s/gs and switch to non-zcopy
>>> if too small.
>>
>> Ok, so idea is that:
>>
>> if (out_sg > vq->num_free)
>>    reorganise current skb for copy mode (e.g. 2 out_sg - header and data)
>>    and try to add it to vq again.
>>
>> ?
>>
>> @Stefano, I'll remove net-next tag (guess RFC is not required again, but not net-next
>> anyway) as this change will require review. R-b I think should be also removed. All
>> other patches in this set still unchanged.
> 
> It's still a new feature so we have net-next tree as the target, right?
> 
> I think we should keep net-next. Even if patches require to be
> re-reviewed, net-next indicates the tree where we want these to be merge
> and for new features is the right one.
> 
> Ack for not putting RFC again and for R-b removal for this patch.

Ok,

Thanks, Arseniy

> 
> Thanks,
> Stefano
>