Skip to contentSkip to conversation

Broken RSS hash computation on Intel 82574L

DISCUSSION Martin Drašar 5 participants 7 messages Aug 31, 2015 – Sep 1, 2015

lore
Share & tools
lore
#0 lore raw

Hi,

I am using RSS to get identical hash for both directions of a flow. Everything is working as it should with the Intel 82599ES, but when I run the same code with the Intel 82574L, then the hash is not calculated correctly and returns 0 as a hash value for all flows.

I could use the Toeplitz hash implementation from 2.1.0, but I would rather do it in hardware, because according to specification, 82574L can do RSS in hardware.

Any ideas what might be wrong?

Thanks in advance,
Martin

conf:
#define RSS_HASH_KEY_LENGTH 40

static uint8_t hash_key[RSS_HASH_KEY_LENGTH] = {
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A };

static const struct rte_eth_conf port_conf = {
.link_speed = 0,
.link_duplex = 0,

  .rxmode = {
    .mq_mode        = ETH_RSS,
    .max_rx_pkt_len = 0,
    .split_hdr_size = 0,
    .header_split   = 0,
    .hw_ip_checksum = 0,
    .hw_vlan_filter = 0,
    .hw_vlan_strip  = 0,
    .hw_vlan_extend = 0,
    .jumbo_frame    = 0,
    .hw_strip_crc   = 0,

},
.txmode = {
.mq_mode = ETH_DCB_NONE,
},
.lpbk_mode = 0,
.rx_adv_conf = {
.rss_conf = {
.rss_key = hash_key,
.rss_key_len = RSS_HASH_KEY_LENGTH,
.rss_hf = ETH_RSS_PROTO_MASK,
},
},
};

#1 lore raw

Hi Martin,

Quoted 45 lines

-----Original Message-----
From: dev [mailto:[email protected]] On Behalf Of Martin Drašar
Sent: Monday, August 31, 2015 1:16 PM
To: [email protected]
Subject: [dpdk-dev] Broken RSS hash computation on Intel 82574L

Hi,

I am using RSS to get identical hash for both directions of a flow. Everything is working as it should with the Intel 82599ES, but when I run the same code with the Intel 82574L, then the hash is not calculated correctly and returns 0 as a hash value for all flows.

I could use the Toeplitz hash implementation from 2.1.0, but I would rather do it in hardware, because according to specification, 82574L can do RSS in hardware.

Any ideas what might be wrong?

Thanks in advance,
Martin

conf:
#define RSS_HASH_KEY_LENGTH 40

static uint8_t hash_key[RSS_HASH_KEY_LENGTH] = {
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A,
        0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A, 0x6D, 0x5A };

static const struct rte_eth_conf port_conf = {
.link_speed = 0,
.link_duplex = 0,

  .rxmode = {
    .mq_mode        = ETH_RSS,
    .max_rx_pkt_len = 0,
    .split_hdr_size = 0,
    .header_split   = 0,
    .hw_ip_checksum = 0,
    .hw_vlan_filter = 0,
    .hw_vlan_strip  = 0,
    .hw_vlan_extend = 0,
    .jumbo_frame    = 0,
    .hw_strip_crc   = 0,
Quoted 13 lines

},
.txmode = {
.mq_mode = ETH_DCB_NONE,
},
.lpbk_mode = 0,
.rx_adv_conf = {
.rss_conf = {
.rss_key = hash_key,
.rss_key_len = RSS_HASH_KEY_LENGTH,
.rss_hf = ETH_RSS_PROTO_MASK,
},
},
};

82574L NIC uses em PMD, which does not support more than 1 queue. Therefore RSS is disabled in the NIC and then you cannot have RSS hashes.

Thanks,
Pablo

#2 lore raw

Dne 1.9.2015 v 15:45 De Lara Guarch, Pablo napsal(a):

Quoted 5 lines

82574L NIC uses em PMD, which does not support more than 1 queue. Therefore RSS is disabled in the NIC and then you cannot have RSS hashes.

Thanks,
Pablo

Hi Pablo,

that is an interesting information. I read the rationale in em_ethdev.c and I was wondering, what would have to be done to enable RSS hash computation on that card. I can live with just one RX queue, but hashes would help me a lot. The computer which is using those NICs is not that powerful and every bit of offloaded computation counts...

Thanks,
Martin

#3 lore raw
Quoted 9 lines · Martin Drašar

Dne 1.9.2015 v 15:45 De Lara Guarch, Pablo napsal(a):

Quoted 5 lines

82574L NIC uses em PMD, which does not support more than 1 queue. Therefore RSS is disabled in the NIC and then you cannot have RSS hashes.

Thanks,
Pablo

Hi Pablo,

Quoted 8 lines

that is an interesting information. I read the rationale in em_ethdev.c and I was wondering, what would have to be done to enable RSS hash computation on that card. I can live with just one RX queue, but hashes would help me a lot. The computer which is using those NICs is not that powerful and every bit of offloaded computation counts...

Thanks,
Martin

RSS calculations are used to direct packets across multiple RX queues. With only one RX queue it cannot possibly increase performance by enabling it.

Matthew.

#4 lore raw
Quoted 15 lines · Matthew Hall
Quoted 14 lines · Martin Drašar

Dne 1.9.2015 v 15:45 De Lara Guarch, Pablo napsal(a):

Quoted 5 lines

82574L NIC uses em PMD, which does not support more than 1 queue. Therefore RSS is disabled in the NIC and then you cannot have RSS hashes.

Thanks,
Pablo

Hi Pablo,

that is an interesting information. I read the rationale in em_ethdev.c and I was wondering, what would have to be done to enable RSS hash computation on that card. I can live with just one RX queue, but hashes would help me a lot. The computer which is using those NICs is not that powerful and every bit of offloaded computation counts...

Quoted 5 lines
Quoted 2 lines

Thanks,
Martin

RSS calculations are used to direct packets across multiple RX queues. With only one RX queue it cannot possibly increase performance by enabling it.

As an example, seastar uses the RSS hash computed by the NIC to select a core to process on, if the number of hardware queues is smaller than the number of cores.

#5 lore raw

Dne 1.9.2015 v 16:50 Avi Kivity napsal(a):

Quoted 9 lines · Matthew Hall
Quoted 4 lines

RSS calculations are used to direct packets across multiple RX queues. With
only one RX queue it cannot possibly increase performance by enabling it.

As an example, seastar uses the RSS hash computed by the NIC to select a core to process on, if the number of hardware queues is smaller than the number of cores.

Yep, we use the hash for pairing of flows and other more complicated packet processing.

#6 lore raw
In reply to Matthew Hall
Quoted 19 lines · Matthew Hall
Quoted 15 lines · Martin Drašar

Dne 1.9.2015 v 15:45 De Lara Guarch, Pablo napsal(a):

Quoted 5 lines

82574L NIC uses em PMD, which does not support more than 1 queue. Therefore RSS is disabled in the NIC and then you cannot have RSS hashes.

Thanks,
Pablo

Hi Pablo,

that is an interesting information. I read the rationale in em_ethdev.c and I was wondering, what would have to be done to enable RSS hash computation on that card. I can live with just one RX queue, but hashes would help me a lot. The computer which is using those NICs is not that powerful and every bit of offloaded computation counts...

Quoted 5 lines
Quoted 2 lines

Thanks,
Martin

RSS calculations are used to direct packets across multiple RX queues. With only one RX queue it cannot possibly increase performance by enabling it.

Well, there are other things you could use that hw-provided hash for…

(but point-taken)