netdev - Re: [RFC PATCH net-next v2 1/2] virtio-net: support zerocopy multi buffer XDP in mergeable

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <c75005c0-7e1a-48f0-b39b-b6310642ad4a@gmail.com>
Date: Thu, 29 May 2025 10:42:27 +0700
From: Bui Quang Minh <minhquangbui99@...il.com>
To: ALOK TIWARI <alok.a.tiwari@...cle.com>, netdev@...r.kernel.org
Cc: "Michael S. Tsirkin" <mst@...hat.com>, Jason Wang <jasowang@...hat.com>,
 Xuan Zhuo <xuanzhuo@...ux.alibaba.com>, Eugenio Pérez
 <eperezma@...hat.com>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
 Alexei Starovoitov <ast@...nel.org>, Daniel Borkmann <daniel@...earbox.net>,
 Jesper Dangaard Brouer <hawk@...nel.org>,
 John Fastabend <john.fastabend@...il.com>, virtualization@...ts.linux.dev,
 linux-kernel@...r.kernel.org, bpf@...r.kernel.org
Subject: Re: [RFC PATCH net-next v2 1/2] virtio-net: support zerocopy multi
 buffer XDP in mergeable

On 5/28/25 23:44, ALOK TIWARI wrote:
>
>
> On 27-05-2025 21:49, Bui Quang Minh wrote:
>> Currently, in zerocopy mode with mergeable receive buffer, virtio-net
>> does not support multi buffer but a single buffer only. This commit adds
>> support for multi mergeable receive buffer in the zerocopy XDP path by
>> utilizing XDP buffer with frags.
>>
>> Signed-off-by: Bui Quang Minh <minhquangbui99@...il.com>
>> ---
>>   drivers/net/virtio_net.c | 123 +++++++++++++++++++++------------------
>>   1 file changed, 66 insertions(+), 57 deletions(-)
>>
>> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
>> index e53ba600605a..a9558650f205 100644
>> --- a/drivers/net/virtio_net.c
>> +++ b/drivers/net/virtio_net.c
>> @@ -45,6 +45,8 @@ module_param(napi_tx, bool, 0644);
>>   #define VIRTIO_XDP_TX        BIT(0)
>>   #define VIRTIO_XDP_REDIR    BIT(1)
>>   +#define VIRTNET_MAX_ZC_SEGS    8
>> +
>>   /* RX packet size EWMA. The average packet size is used to 
>> determine the packet
>>    * buffer size when refilling RX rings. As the entire RX ring may 
>> be refilled
>>    * at once, the weight is chosen so that the EWMA will be 
>> insensitive to short-
>> @@ -1232,65 +1234,53 @@ static void xsk_drop_follow_bufs(struct 
>> net_device *dev,
>>       }
>>   }
>>   -static int xsk_append_merge_buffer(struct virtnet_info *vi,
>> -                   struct receive_queue *rq,
>> -                   struct sk_buff *head_skb,
>> -                   u32 num_buf,
>> -                   struct virtio_net_hdr_mrg_rxbuf *hdr,
>> -                   struct virtnet_rq_stats *stats)
>> +static int virtnet_build_xsk_buff_mrg(struct virtnet_info *vi,
>> +                      struct receive_queue *rq,
>> +                      u32 num_buf,
>> +                      struct xdp_buff *xdp,
>> +                      struct virtnet_rq_stats *stats)
>>   {
>> -    struct sk_buff *curr_skb;
>> -    struct xdp_buff *xdp;
>> -    u32 len, truesize;
>> -    struct page *page;
>> +    unsigned int len;
>>       void *buf;
>>   -    curr_skb = head_skb;
>> +    if (num_buf < 2)
>> +        return 0;
>> +
>> +    while (num_buf > 1) {
>> +        struct xdp_buff *new_xdp;
>>   -    while (--num_buf) {
>>           buf = virtqueue_get_buf(rq->vq, &len);
>> -        if (unlikely(!buf)) {
>> -            pr_debug("%s: rx error: %d buffers out of %d missing\n",
>> -                 vi->dev->name, num_buf,
>> -                 virtio16_to_cpu(vi->vdev,
>> -                         hdr->num_buffers));
>> +        if (!unlikely(buf)) {
>
> if (unlikely(!buf)) { ?

Thanks, I'll fix this in the next version.

>
>> +            pr_debug("%s: rx error: %d buffers missing\n",
>> +                 vi->dev->name, num_buf);
>>               DEV_STATS_INC(vi->dev, rx_length_errors);
>
> Thanks,
> Alok