* [dpdk-dev] [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum @ 2020-11-11 9:11 Murphy Yang 2020-12-02 6:39 ` Guo, Jia 2020-12-03 7:50 ` [dpdk-dev] [PATCH v2] " Murphy Yang 0 siblings, 2 replies; 5+ messages in thread From: Murphy Yang @ 2020-11-11 9:11 UTC (permalink / raw) To: dev; +Cc: qiming.yang, stevex.yang, beilei.xing, jia.guo, Murphy Yang 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. Fixes: 9966a00a0688 ("net/i40e: enable bad checksum flags in vector Rx") Fixes: dafadd73762e ("net/i40e: add AVX2 Rx function") Signed-off-by: Murphy Yang <murphyx.yang@intel.com> --- 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(-) diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx2.c b/drivers/net/i40e/i40e_rxtx_vec_avx2.c index 7a558fc73a..fe6ec7deef 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_avx2.c +++ b/drivers/net/i40e/i40e_rxtx_vec_avx2.c @@ -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 | diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c b/drivers/net/i40e/i40e_rxtx_vec_sse.c index 4b2b6a28fc..0bcb48e24e 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c @@ -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]); -- 2.17.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum 2020-11-11 9:11 [dpdk-dev] [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum Murphy Yang @ 2020-12-02 6:39 ` Guo, Jia 2020-12-03 7:50 ` [dpdk-dev] [PATCH v2] " Murphy Yang 1 sibling, 0 replies; 5+ messages in thread From: Guo, Jia @ 2020-12-02 6:39 UTC (permalink / raw) To: Yang, MurphyX, dev Cc: Yang, Qiming, Yang, SteveX, Xing, Beilei, Yang, MurphyX Hi, murphy > -----Original Message----- > From: Murphy Yang <murphyx.yang@intel.com> > Sent: Wednesday, November 11, 2020 5:11 PM > To: dev@dpdk.org > Cc: Yang, Qiming <qiming.yang@intel.com>; Yang, SteveX > <stevex.yang@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Guo, Jia > <jia.guo@intel.com>; Yang, MurphyX <murphyx.yang@intel.com> > Subject: [PATCH] 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. > 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? > Fixes: 9966a00a0688 ("net/i40e: enable bad checksum flags in vector Rx") > Fixes: dafadd73762e ("net/i40e: add AVX2 Rx function") > > Signed-off-by: Murphy Yang <murphyx.yang@intel.com> > --- > 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(-) > > diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx2.c > b/drivers/net/i40e/i40e_rxtx_vec_avx2.c > index 7a558fc73a..fe6ec7deef 100644 > --- a/drivers/net/i40e/i40e_rxtx_vec_avx2.c > +++ b/drivers/net/i40e/i40e_rxtx_vec_avx2.c > @@ -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); > 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 "? > const __m256i cksum_mask = _mm256_set1_epi32( > PKT_RX_IP_CKSUM_GOOD | > PKT_RX_IP_CKSUM_BAD | diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c > b/drivers/net/i40e/i40e_rxtx_vec_sse.c > index 4b2b6a28fc..0bcb48e24e 100644 > --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c > +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c > @@ -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]); > -- > 2.17.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [dpdk-dev] [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksum 2020-11-11 9:11 [dpdk-dev] [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum Murphy Yang 2020-12-02 6:39 ` Guo, Jia @ 2020-12-03 7:50 ` Murphy Yang 2020-12-04 2:18 ` Guo, Jia 1 sibling, 1 reply; 5+ messages in thread From: Murphy Yang @ 2020-12-03 7:50 UTC (permalink / raw) To: dev; +Cc: qiming.yang, stevex.yang, jia.guo, beilei.xing, Murphy Yang 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'. 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 <murphyx.yang@intel.com> --- 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(-) diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx2.c b/drivers/net/i40e/i40e_rxtx_vec_avx2.c index 7a558fc73a..fe6ec7deef 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_avx2.c +++ b/drivers/net/i40e/i40e_rxtx_vec_avx2.c @@ -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 | diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c b/drivers/net/i40e/i40e_rxtx_vec_sse.c index 4b2b6a28fc..0bcb48e24e 100644 --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c @@ -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]); -- 2.17.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksum 2020-12-03 7:50 ` [dpdk-dev] [PATCH v2] " Murphy Yang @ 2020-12-04 2:18 ` Guo, Jia 2020-12-11 2:16 ` Zhang, Qi Z 0 siblings, 1 reply; 5+ messages in thread From: Guo, Jia @ 2020-12-04 2:18 UTC (permalink / raw) To: Yang, MurphyX, dev Cc: Yang, Qiming, Yang, SteveX, Xing, Beilei, Yang, MurphyX Acked-by: Jeff Guo <jia.guo@intel.com> > -----Original Message----- > From: Murphy Yang <murphyx.yang@intel.com> > Sent: Thursday, December 3, 2020 3:51 PM > To: dev@dpdk.org > Cc: Yang, Qiming <qiming.yang@intel.com>; Yang, SteveX > <stevex.yang@intel.com>; Guo, Jia <jia.guo@intel.com>; Xing, Beilei > <beilei.xing@intel.com>; Yang, MurphyX <murphyx.yang@intel.com> > Subject: [PATCH v2] 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'. > > 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 <murphyx.yang@intel.com> > --- > 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(-) > > diff --git a/drivers/net/i40e/i40e_rxtx_vec_avx2.c > b/drivers/net/i40e/i40e_rxtx_vec_avx2.c > index 7a558fc73a..fe6ec7deef 100644 > --- a/drivers/net/i40e/i40e_rxtx_vec_avx2.c > +++ b/drivers/net/i40e/i40e_rxtx_vec_avx2.c > @@ -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 | diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c > b/drivers/net/i40e/i40e_rxtx_vec_sse.c > index 4b2b6a28fc..0bcb48e24e 100644 > --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c > +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c > @@ -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]); > -- > 2.17.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/i40e: fix incorrect checksum flag of L4 checksum 2020-12-04 2:18 ` Guo, Jia @ 2020-12-11 2:16 ` Zhang, Qi Z 0 siblings, 0 replies; 5+ messages in thread From: Zhang, Qi Z @ 2020-12-11 2:16 UTC (permalink / raw) To: Guo, Jia, Yang, MurphyX, dev Cc: Yang, Qiming, Yang, SteveX, Xing, Beilei, Yang, MurphyX > -----Original Message----- > From: dev <dev-bounces@dpdk.org> On Behalf Of Guo, Jia > Sent: Friday, December 4, 2020 10:19 AM > To: Yang, MurphyX <murphyx.yang@intel.com>; dev@dpdk.org > Cc: Yang, Qiming <qiming.yang@intel.com>; Yang, SteveX > <stevex.yang@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Yang, MurphyX > <murphyx.yang@intel.com> > Subject: Re: [dpdk-dev] [PATCH v2] net/i40e: fix incorrect checksum flag of L4 > checksum > > Acked-by: Jeff Guo <jia.guo@intel.com> > > > -----Original Message----- > > From: Murphy Yang <murphyx.yang@intel.com> > > Sent: Thursday, December 3, 2020 3:51 PM > > To: dev@dpdk.org > > Cc: Yang, Qiming <qiming.yang@intel.com>; Yang, SteveX > > <stevex.yang@intel.com>; Guo, Jia <jia.guo@intel.com>; Xing, Beilei > > <beilei.xing@intel.com>; Yang, MurphyX <murphyx.yang@intel.com> > > Subject: [PATCH v2] 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'. > > > > 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 <murphyx.yang@intel.com> Applied to dpdk-next-net-intel. Thanks Qi ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2020-12-11 2:17 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-11-11 9:11 [dpdk-dev] [PATCH] net/i40e: fix incorrect checksum flag of L4 checksum Murphy Yang 2020-12-02 6:39 ` Guo, Jia 2020-12-03 7:50 ` [dpdk-dev] [PATCH v2] " Murphy Yang 2020-12-04 2:18 ` Guo, Jia 2020-12-11 2:16 ` Zhang, Qi Z
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).