From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Shiva Shankar Kommula <kshankar@marvell.com>,
"dev@dpdk.org" <dev@dpdk.org>,
"chenbox@nvidia.com" <chenbox@nvidia.com>
Cc: "david.marchand@redhat.com" <david.marchand@redhat.com>,
Jerin Jacob <jerinj@marvell.com>,
Nithin Kumar Dabilpuram <ndabilpuram@marvell.com>,
Srujana Challa <schalla@marvell.com>
Subject: Re: [EXTERNAL] Re: [PATCH v2] net/virtio: add virtio hash report feature
Date: Mon, 3 Mar 2025 11:25:13 +0100 [thread overview]
Message-ID: <b9844729-a2a0-45f9-b0c5-387d0d1e03f2@redhat.com> (raw)
In-Reply-To: <DM4PR18MB42690E8143C44B3893DC09D9DFCD2@DM4PR18MB4269.namprd18.prod.outlook.com>
On 2/27/25 10:44 AM, Shiva Shankar Kommula wrote:
>> -----Original Message-----
>> From: Maxime Coquelin <maxime.coquelin@redhat.com>
>> Sent: Tuesday, February 25, 2025 9:39 PM
>> To: Shiva Shankar Kommula <kshankar@marvell.com>; dev@dpdk.org;
>> chenbox@nvidia.com
>> Cc: david.marchand@redhat.com; Jerin Jacob <jerinj@marvell.com>; Nithin
>> Kumar Dabilpuram <ndabilpuram@marvell.com>; Srujana Challa
>> <schalla@marvell.com>
>> Subject: [EXTERNAL] Re: [PATCH v2] net/virtio: add virtio hash report feature
>>
>> Hi, On 1/27/25 8: 41 AM, Kommula Shiva Shankar wrote: > This patch adds
>> virtio hash report feature, which is > supported in packet queue mode with
>> scalar version Why only restrict to packed ring? isn't it also compatible with
>> split ring?
>>
>> Hi,
>>
>> On 1/27/25 8:41 AM, Kommula Shiva Shankar wrote:
>>> This patch adds virtio hash report feature, which is supported in
>>> packet queue mode with scalar version
>>
>> Why only restrict to packed ring? isn't it also compatible with split ring?
> Hi Maxim,
> Yes, it is. I can add split ring support later if it's alright.
Ok, I checked the split path and it should not break. But please add its
support in next release.
I see one thing missing in virtio_ethdev.c:
int
eth_virtio_dev_init(struct rte_eth_dev *eth_dev)
{
struct virtio_hw *hw = eth_dev->data->dev_private;
uint32_t speed = RTE_ETH_SPEED_NUM_UNKNOWN;
int vectorized = 0;
int ret;
if (sizeof(struct virtio_net_hdr_mrg_rxbuf) > RTE_PKTMBUF_HEADROOM) {
PMD_INIT_LOG(ERR,
"Not sufficient headroom required = %d, avail = %d",
(int)sizeof(struct virtio_net_hdr_mrg_rxbuf),
RTE_PKTMBUF_HEADROOM);
return -1;
}
It should now be checked against virtio_net_hdr_hash_report struct size.
Regards,
Maxime
>>
>>>
>>> Signed-off-by: Kommula Shiva Shankar <kshankar@marvell.com>
>>> ---
>>> drivers/net/virtio/virtio.h | 2 ++
>>> drivers/net/virtio/virtio_ethdev.c | 20 ++++++++++++-
>>> drivers/net/virtio/virtio_ethdev.h | 1 +
>>> drivers/net/virtio/virtio_rxtx.c | 30 +++++++++++++++++++
>>> .../net/virtio/virtio_user/virtio_user_dev.c | 1 +
>>> drivers/net/virtio/virtqueue.h | 21 +++++++++++++
>>> 6 files changed, 74 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/net/virtio/virtio.h b/drivers/net/virtio/virtio.h
>>> index ef5827c5f5..c2a0fd477c 100644
>>> --- a/drivers/net/virtio/virtio.h
>>> +++ b/drivers/net/virtio/virtio.h
>>> @@ -30,6 +30,7 @@
>>> #define VIRTIO_NET_F_GUEST_ANNOUNCE 21 /* Guest can
>> announce device on the network */
>>> #define VIRTIO_NET_F_MQ 22 /* Device supports Receive
>> Flow Steering */
>>> #define VIRTIO_NET_F_CTRL_MAC_ADDR 23 /* Set MAC address */
>>> +#define VIRTIO_NET_F_HASH_REPORT 57 /* Supports hash report */
>>> #define VIRTIO_NET_F_RSS 60 /* RSS supported */
>>>
>>> /*
>>> @@ -187,6 +188,7 @@ struct virtio_hw {
>>> uint8_t started;
>>> uint8_t weak_barriers;
>>> uint8_t vlan_strip;
>>> + uint8_t has_hash_report;
>>> bool rx_ol_scatter;
>>> uint8_t has_tx_offload;
>>> uint8_t has_rx_offload;
>>> diff --git a/drivers/net/virtio/virtio_ethdev.c
>>> b/drivers/net/virtio/virtio_ethdev.c
>>> index 70d4839def..caacbce57a 100644
>>> --- a/drivers/net/virtio/virtio_ethdev.c
>>> +++ b/drivers/net/virtio/virtio_ethdev.c
>>> @@ -1796,7 +1796,9 @@ virtio_init_device(struct rte_eth_dev *eth_dev,
>> uint64_t req_features)
>>> eth_dev->data->dev_flags &= ~RTE_ETH_DEV_INTR_LSC;
>>>
>>> /* Setting up rx_header size for the device */
>>> - if (virtio_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF) ||
>>> + if (virtio_with_feature(hw, VIRTIO_NET_F_HASH_REPORT))
>>> + hw->vtnet_hdr_size = sizeof(struct
>> virtio_net_hdr_hash_report);
>>> + else if (virtio_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF) ||
>>> virtio_with_feature(hw, VIRTIO_F_VERSION_1) ||
>>> virtio_with_packed_queue(hw))
>>> hw->vtnet_hdr_size = sizeof(struct
>> virtio_net_hdr_mrg_rxbuf); @@
>>> -2181,6 +2183,10 @@ virtio_dev_configure(struct rte_eth_dev *dev)
>>> (1ULL << VIRTIO_NET_F_GUEST_TSO4) |
>>> (1ULL << VIRTIO_NET_F_GUEST_TSO6);
>>>
>>> + if (rx_offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH)
>>> + req_features |=
>>> + (1ULL << VIRTIO_NET_F_HASH_REPORT);
>>> +
>>> if (tx_offloads & (RTE_ETH_TX_OFFLOAD_UDP_CKSUM |
>>> RTE_ETH_TX_OFFLOAD_TCP_CKSUM))
>>> req_features |= (1ULL << VIRTIO_NET_F_CSUM); @@ -2233,6
>> +2239,9 @@
>>> virtio_dev_configure(struct rte_eth_dev *dev)
>>> if (rx_offloads & RTE_ETH_RX_OFFLOAD_VLAN_STRIP)
>>> hw->vlan_strip = 1;
>>>
>>> + if (rx_offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH)
>>> + hw->has_hash_report = 1;
>>> +
>>> hw->rx_ol_scatter = (rx_offloads & RTE_ETH_RX_OFFLOAD_SCATTER);
>>>
>>> if ((rx_offloads & RTE_ETH_RX_OFFLOAD_VLAN_FILTER) && @@ -
>> 2285,6
>>> +2294,12 @@ virtio_dev_configure(struct rte_eth_dev *dev)
>>> "disabled packed ring vectorized rx for
>> TCP_LRO enabled");
>>> hw->use_vec_rx = 0;
>>> }
>>> +
>>> + if (rx_offloads & RTE_ETH_RX_OFFLOAD_RSS_HASH) {
>>> + PMD_DRV_LOG(INFO,
>>> + "disabled packed ring vectorized rx for
>> RSS_HASH enabled");
>>> + hw->use_vec_rx = 0;
>>> + }
>>> }
>>> } else {
>>> if (virtio_with_feature(hw, VIRTIO_F_IN_ORDER)) { @@ -
>> 2669,6
>>> +2684,9 @@ virtio_dev_info_get(struct rte_eth_dev *dev, struct
>> rte_eth_dev_info *dev_info)
>>> dev_info->flow_type_rss_offloads = 0;
>>> }
>>>
>>> + if (host_features & (1ULL << VIRTIO_NET_F_HASH_REPORT))
>>> + dev_info->rx_offload_capa |=
>> RTE_ETH_RX_OFFLOAD_RSS_HASH;
>>> +
>>> if (host_features & (1ULL << VIRTIO_F_RING_PACKED)) {
>>> /*
>>> * According to 2.7 Packed Virtqueues, diff --git
>>> a/drivers/net/virtio/virtio_ethdev.h
>>> b/drivers/net/virtio/virtio_ethdev.h
>>> index 7be1c9acd0..7aa771fd2b 100644
>>> --- a/drivers/net/virtio/virtio_ethdev.h
>>> +++ b/drivers/net/virtio/virtio_ethdev.h
>>> @@ -46,6 +46,7 @@
>>> 1u << VIRTIO_NET_F_CSUM | \
>>> 1u << VIRTIO_NET_F_HOST_TSO4 | \
>>> 1u << VIRTIO_NET_F_HOST_TSO6 | \
>>> + 1ULL << VIRTIO_NET_F_HASH_REPORT | \
>>> 1ULL << VIRTIO_NET_F_RSS)
>>>
>>> extern const struct eth_dev_ops virtio_user_secondary_eth_dev_ops;
>>> diff --git a/drivers/net/virtio/virtio_rxtx.c
>>> b/drivers/net/virtio/virtio_rxtx.c
>>> index b67f063b31..35bd61350c 100644
>>> --- a/drivers/net/virtio/virtio_rxtx.c
>>> +++ b/drivers/net/virtio/virtio_rxtx.c
>>> @@ -40,6 +40,19 @@
>>> #define VIRTIO_DUMP_PACKET(m, len) do { } while (0)
>>> #endif
>>>
>>> +static const uint32_t vhdr_hash_report_to_mbuf_pkt_type[] = {
>>> + RTE_PTYPE_UNKNOWN,
>>> + RTE_PTYPE_L3_IPV4,
>>> + RTE_PTYPE_L3_IPV4 | RTE_PTYPE_L4_TCP,
>>> + RTE_PTYPE_L3_IPV4 | RTE_PTYPE_L4_UDP,
>>> + RTE_PTYPE_L3_IPV6,
>>> + RTE_PTYPE_L3_IPV6 | RTE_PTYPE_L4_TCP,
>>> + RTE_PTYPE_L3_IPV6 | RTE_PTYPE_L4_UDP,
>>> + RTE_PTYPE_L3_IPV6_EXT,
>>> + RTE_PTYPE_L3_IPV6_EXT | RTE_PTYPE_L4_TCP,
>>> + RTE_PTYPE_L3_IPV6_EXT | RTE_PTYPE_L4_UDP, };
>>> +
>>> void
>>> vq_ring_free_inorder(struct virtqueue *vq, uint16_t desc_idx, uint16_t
>> num)
>>> {
>>> @@ -901,6 +914,16 @@ virtio_discard_rxbuf_inorder(struct virtqueue *vq,
>> struct rte_mbuf *m)
>>> }
>>> }
>>>
>>> +static inline void
>>> +virtio_rx_update_hash_report(struct rte_mbuf *m, struct
>>> +virtio_net_hdr_hash_report *hdr) {
>>> + if (likely(hdr->hash_report)) {
>>> + m->packet_type = vhdr_hash_report_to_mbuf_pkt_type[hdr-
>>> hash_report];
>>> + m->hash.rss = hdr->hash_value;
>>> + m->ol_flags |= RTE_MBUF_F_RX_RSS_HASH;
>>> + }
>>> +}
>>> +
>>> /* Optionally fill offload information in structure */
>>> static inline int
>>> virtio_rx_offload(struct rte_mbuf *m, struct virtio_net_hdr *hdr) @@
>>> -1134,6 +1157,9 @@ virtio_recv_pkts_packed(void *rx_queue, struct
>> rte_mbuf **rx_pkts,
>>> hdr = (struct virtio_net_hdr *)((char *)rxm->buf_addr +
>>> RTE_PKTMBUF_HEADROOM - hdr_size);
>>>
>>> + if (hw->has_hash_report)
>>> + virtio_rx_update_hash_report(rxm,
>>> + (struct
>> virtio_net_hdr_hash_report *)hdr);
>>> if (hw->vlan_strip)
>>> rte_vlan_strip(rxm);
>>>
>>> @@ -1617,6 +1643,10 @@ virtio_recv_mergeable_pkts_packed(void
>> *rx_queue,
>>> if (hw->vlan_strip)
>>> rte_vlan_strip(rx_pkts[nb_rx]);
>>>
>>> + if (hw->has_hash_report)
>>> + virtio_rx_update_hash_report(rxm,
>>> + (struct
>> virtio_net_hdr_hash_report *)header);
>>> +
>>> seg_res = seg_num - 1;
>>>
>>> /* Merge remaining segments */
>>> diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
>>> b/drivers/net/virtio/virtio_user/virtio_user_dev.c
>>> index 2997d2bd26..187f81b066 100644
>>> --- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
>>> +++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
>>> @@ -726,6 +726,7 @@ virtio_user_free_vrings(struct virtio_user_dev *dev)
>>> 1ULL << VIRTIO_F_RING_PACKED | \
>>> 1ULL << VIRTIO_F_NOTIFICATION_DATA | \
>>> 1ULL << VIRTIO_F_ORDER_PLATFORM | \
>>> + 1ULL << VIRTIO_NET_F_HASH_REPORT | \
>>> 1ULL << VIRTIO_NET_F_RSS)
>>>
>>> int
>>> diff --git a/drivers/net/virtio/virtqueue.h
>>> b/drivers/net/virtio/virtqueue.h index 60211a40c9..3b468bd4f4 100644
>>> --- a/drivers/net/virtio/virtqueue.h
>>> +++ b/drivers/net/virtio/virtqueue.h
>>> @@ -257,6 +257,27 @@ struct virtio_net_hdr_mrg_rxbuf {
>>> uint16_t num_buffers; /**< Number of merged rx buffers */
>>> };
>>>
>>> +/**
>>> + * This is the version of the header to use when the HASH_REPORT
>>> + * feature has been negotiated.
>>> + */
>>> +struct virtio_net_hdr_hash_report {
>>> + struct virtio_net_hdr_mrg_rxbuf hdr;
>>> + uint32_t hash_value;
>>> +#define VIRTIO_NET_HASH_REPORT_NONE 0
>>> +#define VIRTIO_NET_HASH_REPORT_IPV4 1
>>> +#define VIRTIO_NET_HASH_REPORT_TCPV4 2
>>> +#define VIRTIO_NET_HASH_REPORT_UDPV4 3
>>> +#define VIRTIO_NET_HASH_REPORT_IPV6 4
>>> +#define VIRTIO_NET_HASH_REPORT_TCPV6 5
>>> +#define VIRTIO_NET_HASH_REPORT_UDPV6 6
>>> +#define VIRTIO_NET_HASH_REPORT_IPV6_EX 7
>>> +#define VIRTIO_NET_HASH_REPORT_TCPV6_EX 8
>>> +#define VIRTIO_NET_HASH_REPORT_UDPV6_EX 9
>>> + uint16_t hash_report;
>>> + uint16_t pad_reserved;
>>> +};
>>> +
>>> /* Region reserved to allow for transmit header and indirect ring */
>>> #define VIRTIO_MAX_TX_INDIRECT 8
>>> struct virtio_tx_region {
>
prev parent reply other threads:[~2025-03-03 10:25 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-01-26 19:34 [PATCH] " Kommula Shiva Shankar
2025-01-27 7:41 ` [PATCH v2] " Kommula Shiva Shankar
2025-02-25 16:09 ` Maxime Coquelin
2025-02-27 9:44 ` [EXTERNAL] " Shiva Shankar Kommula
2025-03-03 10:25 ` Maxime Coquelin [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b9844729-a2a0-45f9-b0c5-387d0d1e03f2@redhat.com \
--to=maxime.coquelin@redhat.com \
--cc=chenbox@nvidia.com \
--cc=david.marchand@redhat.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=kshankar@marvell.com \
--cc=ndabilpuram@marvell.com \
--cc=schalla@marvell.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).