net/i40e: fix incorrect checksum flag of L4 checksum
PATCH 1/1 net/i40e: fix incorrect checksum flag of L4 checksum
When send tunneled packet that inner L4 checksum value is correct, the test_pmd output log shows 'ol_flags' value is 'PKT_RX_L4_CKSUM_UNKNOWN', but expected value is 'PKT_RX_L4_CKSUM_GOOD'.
Add the 'PKT_RX_L4_CKSUM_GOOD' to 'l3_l4e_flags' for sse and 'l3_l4_flags_shuf' for avx2 to ensure that the 'ol_flags' can match correct flags.
--- drivers/net/i40e/i40e_rxtx_vec_avx2.c | 40 ++++++++++++++++----------- drivers/net/i40e/i40e_rxtx_vec_sse.c | 20 ++++++++------ 2 files changed, 35 insertions(+), 25 deletions(-)
drivers/net/i40e/i40e_rxtx_vec_avx2.c +24 -16
@@ -342,24 +342,32 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts,*/ const __m256i l3_l4_flags_shuf =_mm256_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 ,/* second 128-bits */ 0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 ); + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 ); const __m256i cksum_mask = _mm256_set1_epi32( PKT_RX_IP_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD |
drivers/net/i40e/i40e_rxtx_vec_sse.c +11 -9
@@ -254,16 +254,18 @@ desc_to_olflags_v(struct i40e_rx_queue *rxq, volatile union i40e_rx_desc *rxdp,const __m128i l3_l4e_flags =_mm_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD | -PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 ); + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 );/* Unpack "status" from quadword 1, bits 0:32 */ vlan0 =_mm_unpackhi_epi32 (descs[0 ], descs[1 ]);
Signature · 1 line
2.17.1
Hi, murphy
Quoted 9 lines
-----Original Message-----
From: Murphy Yang <[email protected]>
Sent: Wednesday, November 11, 2020 5:11 PM
To: [email protected]
Cc: Yang, Qiming <[email protected]>; Yang, SteveX <[email protected]>; Xing, Beilei <[email protected]>; Guo, Jia <[email protected]>; Yang, MurphyX <[email protected]>
Subject: [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum
Quoted 8 lines
When send tunneled packet that inner L4 checksum value is correct, the
test_pmdoutput log shows 'ol_flags' value is 'PKT_RX_L4_CKSUM_UNKNOWN', but expected value is 'PKT_RX_L4_CKSUM_GOOD'.Add the '
PKT_RX_L4_CKSUM_GOOD' to 'l3_l4e_flags' for sse and 'l3_l4_flags_shuf' for avx2 to ensure that the 'ol_flags' can match correct flags.
Seems that 'PKT_RX_L4_CKSUM_GOOD' is previous there but not set correctly, so maybe it should not say " Add the 'PKT_RX_L4_CKSUM_GOOD' to 'l3_l4e_flags' .... "
Add more, could you please to check if the other rx vec path also need it, such as vec_altivec and vec_neon?
Quoted 91 lines
Fixes 9966a00a0688 ("net/i40e: enable bad checksum flags in vector Rx")Fixes dafadd73762e ("net/i40e: add AVX2 Rx function")Signed-off-by Murphy Yang <[email protected]>--- drivers/net/i40e/i40e_rxtx_vec_avx2.c | 40 ++++++++++++++++----------- drivers/net/i40e/i40e_rxtx_vec_sse.c | 20 ++++++++------ 2 files changed, 35 insertions(+), 25 deletions(-)drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-8@@ -342,24 +342,32 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue*rxq, struct rte_mbuf **rx_pkts, */ const __m256i l3_l4_flags_shuf =_mm256_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD) >> 1, - (PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD) >> 1, - PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1/* second 128-bits */ 0, 0, 0, 0, 0, 0, 0, 0, - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD|PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-2- PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
Quoted 2 lines
PKT_RX_IP_CKSUM_GOOD) >> 1);
Could you double check if it is reasonable that the " PKT_RX_EIP_CKSUM_BAD" is always be set, but no " PKT_RX_EIP_CKSUM_GOOD "?
Quoted 53 lines
drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-0const __m256i cksum_mask = _mm256_set1_epi32( PKT_RX_IP_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD| diff --gita/drivers/net/i40e/i40e_rxtx_vec_sse.cb/drivers/net/i40e/i40e_rxtx_vec_sse.c
index4b2b6a28fc..0bcb48e24e100644drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1@@ -254,16 +254,18 @@ desc_to_olflags_v(struct i40e_rx_queue *rxq,volatile union i40e_rx_desc *rxdp, const __m128i l3_l4e_flags = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-2- PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-2- PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1/* Unpack "status" from quadword 1, bits 0:32 */ vlan0 =_mm_unpackhi_epi32 (descs[0 ], descs[1 ]); --2.17.1
PATCH v2 1/1 net/i40e: fix incorrect checksum flag of L4 checksum
In reply to Murphy YangWhen send tunneled packet that inner L4 checksum value is correct, the test_pmd output log shows 'ol_flags' value is 'PKT_RX_L4_CKSUM_UNKNOWN', but expected value is 'PKT_RX_L4_CKSUM_GOOD'.
If the inner l4 checksum is correctly, mark the 'PKT_RX_L4_CKSUM_GOOD' flag to 'l3_l4e_flags' for sse and 'l3_l4_flags_shuf' for avx2 to ensure that the 'ol_flags' can match correct flags.
---
v2:
- update the comments.
drivers/net/i40e/i40e_rxtx_vec_avx2.c | 40 ++++++++++++++++----------- drivers/net/i40e/i40e_rxtx_vec_sse.c | 20 ++++++++------ 2 files changed, 35 insertions(+), 25 deletions(-)
drivers/net/i40e/i40e_rxtx_vec_avx2.c +24 -16
@@ -342,24 +342,32 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts,*/ const __m256i l3_l4_flags_shuf =_mm256_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 ,/* second 128-bits */ 0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 ); + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 ); const __m256i cksum_mask = _mm256_set1_epi32( PKT_RX_IP_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD |
drivers/net/i40e/i40e_rxtx_vec_sse.c +11 -9
@@ -254,16 +254,18 @@ desc_to_olflags_v(struct i40e_rx_queue *rxq, volatile union i40e_rx_desc *rxdp,const __m128i l3_l4e_flags =_mm_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD | -PKT_RX_L4_CKSUM_BAD ) >>1 , - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_EIP_CKSUM_BAD ) >>1 , - (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_BAD ) >>1 , -PKT_RX_IP_CKSUM_BAD >>1 , - (PKT_RX_IP_CKSUM_GOOD |PKT_RX_L4_CKSUM_GOOD ) >>1 ); + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_GOOD | +PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_GOOD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_BAD ) >>1 , + (PKT_RX_L4_CKSUM_GOOD |PKT_RX_IP_CKSUM_GOOD ) >>1 );/* Unpack "status" from quadword 1, bits 0:32 */ vlan0 =_mm_unpackhi_epi32 (descs[0 ], descs[1 ]);
Signature · 1 line
2.17.1
Quoted 166 lines
-----Original Message-----
From: Murphy Yang <[email protected]>
Sent: Thursday, December 3, 2020 3:51 PM
To: [email protected]
Cc: Yang, Qiming <[email protected]>; Yang, SteveX <[email protected]>; Guo, Jia <[email protected]>; Xing, Beilei <[email protected]>; Yang, MurphyX <[email protected]>
Subject: [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksumWhen send tunneled packet that inner L4 checksum value is correct, the
test_pmdoutput log shows 'ol_flags' value is 'PKT_RX_L4_CKSUM_UNKNOWN', but expected value is 'PKT_RX_L4_CKSUM_GOOD'.If the inner l4 checksum is correctly, mark the '
PKT_RX_L4_CKSUM_GOOD' flag to 'l3_l4e_flags' for sse and 'l3_l4_flags_shuf' for avx2 to ensure that the 'ol_flags' can match correct flags.Fixes 9966a00a0688 ("net/i40e: enable bad checksum flags in vector Rx")Fixes dafadd73762e ("net/i40e: add AVX2 Rx function")Signed-off-by Murphy Yang <[email protected]>---
v2:
- update the comments.drivers/net/i40e/i40e_rxtx_vec_avx2.c | 40 ++++++++++++++++----------- drivers/net/i40e/i40e_rxtx_vec_sse.c | 20 ++++++++------ 2 files changed, 35 insertions(+), 25 deletions(-)drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-8@@ -342,24 +342,32 @@ _recv_raw_pkts_vec_avx2(struct i40e_rx_queue*rxq, struct rte_mbuf **rx_pkts, */ const __m256i l3_l4_flags_shuf =_mm256_set_epi8 (0 ,0 ,0 ,0 ,0 ,0 ,0 ,0 ,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |PKT_RX_L4_CKSUM_BAD |PKT_RX_IP_CKSUM_BAD ) >>1 , - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD) >> 1, - (PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD) >> 1, - PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1/* second 128-bits */ 0, 0, 0, 0, 0, 0, 0, 0, - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD|PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-2- PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_avx2.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_avx2.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_avx2.c +0
-0const __m256i cksum_mask = _mm256_set1_epi32( PKT_RX_IP_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD| diff --gita/drivers/net/i40e/i40e_rxtx_vec_sse.cb/drivers/net/i40e/i40e_rxtx_vec_sse.c
index4b2b6a28fc..0bcb48e24e100644drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1@@ -254,16 +254,18 @@ desc_to_olflags_v(struct i40e_rx_queue *rxq,volatile union i40e_rx_desc *rxdp, const __m128i l3_l4e_flags = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,/* shift right 1 bit to make sure it not exceed 255 */ - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-2- PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_EIP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1- (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-2- PKT_RX_IP_CKSUM_BAD >> 1, - (PKT_RX_IP_CKSUM_GOOD |
PKT_RX_L4_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_BAD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_EIP_CKSUM_BAD |
PKT_RX_L4_CKSUM_GOOD|drivers/net/i40e/i40e_rxtx_vec_sse.c +2
-0+ PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_BAD |
PKT_RX_IP_CKSUM_GOOD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_BAD) >> 1,drivers/net/i40e/i40e_rxtx_vec_sse.c +1
-0+ (PKT_RX_L4_CKSUM_GOOD |
PKT_RX_IP_CKSUM_GOOD) >> 1);drivers/net/i40e/i40e_rxtx_vec_sse.c +0
-1/* Unpack "status" from quadword 1, bits 0:32 */ vlan0 =_mm_unpackhi_epi32 (descs[0 ], descs[1 ]); --2.17.1
Zhang, Qi Z
AppliedQuoted 12 lines
-----Original Message-----
From: dev <[email protected]> On Behalf Of Guo, Jia
Sent: Friday, December 4, 2020 10:19 AM
To: Yang, MurphyX <[email protected]>; [email protected]
Cc: Yang, Qiming <[email protected]>; Yang, SteveX <[email protected]>; Xing, Beilei <[email protected]>; Yang, MurphyX <[email protected]>
Subject: Re: [dpdk-dev] [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksum
Quoted 26 lines
Acked-by Jeff Guo <[email protected]>Quoted 24 lines
-----Original Message-----
From: Murphy Yang <[email protected]>
Sent: Thursday, December 3, 2020 3:51 PM
To: [email protected]
Cc: Yang, Qiming <[email protected]>; Yang, SteveX <[email protected]>; Guo, Jia <[email protected]>; Xing, Beilei <[email protected]>; Yang, MurphyX <[email protected]>
Subject: [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksumWhen send tunneled packet that inner L4 checksum value is correct, the
test_pmdoutput log shows 'ol_flags' value is 'PKT_RX_L4_CKSUM_UNKNOWN', but expected value is 'PKT_RX_L4_CKSUM_GOOD'.If the inner l4 checksum is correctly, mark the '
PKT_RX_L4_CKSUM_GOOD' flag to 'l3_l4e_flags' for sse and 'l3_l4_flags_shuf' for avx2 to ensure that the 'ol_flags' can match correct flags.Fixes 9966a00a0688 ("net/i40e: enable bad checksum flags in vectorRx")
Fixes dafadd73762e ("net/i40e: add AVX2 Rx function")Signed-off-by Murphy Yang <[email protected]>
Applied to dpdk-next-net-intel.
Thanks
Qi