* [PATCH 0/2] support IPv4 fragment matching in transfer rules
@ 2023-06-23 16:58 Artemii Morozov
2023-06-23 16:58 ` [PATCH 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov
` (3 more replies)
0 siblings, 4 replies; 8+ messages in thread
From: Artemii Morozov @ 2023-06-23 16:58 UTC (permalink / raw)
To: dev
Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton,
Andrew Rybchenko, Ferruh Yigit
This patch seria adds support for matching fragment offsets for IPv4 packets,
but with some limitations. Exact matching is supported only for zero offset,
ranges are allowed for non-zero offsets.
Artemii Morozov (2):
common/sfc_efx/base: add MAE IP fragmentation match bits
net/sfc: support IPv4 fragment matching in transfer rules
drivers/common/sfc_efx/base/efx.h | 4 +
drivers/common/sfc_efx/base/efx_mae.c | 4 +
drivers/net/sfc/sfc_mae.c | 110 +++++++++++++++++++++++++-
drivers/net/sfc/sfc_mae.h | 4 +
4 files changed, 121 insertions(+), 1 deletion(-)
--
2.34.1
^ permalink raw reply [flat|nested] 8+ messages in thread* [PATCH 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits 2023-06-23 16:58 [PATCH 0/2] support IPv4 fragment matching in transfer rules Artemii Morozov @ 2023-06-23 16:58 ` Artemii Morozov 2023-06-23 16:58 ` [PATCH 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov ` (2 subsequent siblings) 3 siblings, 0 replies; 8+ messages in thread From: Artemii Morozov @ 2023-06-23 16:58 UTC (permalink / raw) To: dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko, Ferruh Yigit Introduce necessary infrastructure for these bits to be set and validated. Using a combination of these bits IP fragmentation can be configured. Signed-off-by: Artemii Morozov <artemii.morozov@arknetworks.am> Reviewed-by: Ivan Malov <ivan.malov@arknetworks.am> Reviewed-by: Andy Moreton <amoreton@xilinx.com> --- drivers/common/sfc_efx/base/efx.h | 4 ++++ drivers/common/sfc_efx/base/efx_mae.c | 4 ++++ 2 files changed, 8 insertions(+) diff --git a/drivers/common/sfc_efx/base/efx.h b/drivers/common/sfc_efx/base/efx.h index 77f855bfb0..efefea717f 100644 --- a/drivers/common/sfc_efx/base/efx.h +++ b/drivers/common/sfc_efx/base/efx.h @@ -4317,6 +4317,10 @@ typedef enum efx_mae_field_id_e { */ EFX_MAE_FIELD_RECIRC_ID, EFX_MAE_FIELD_CT_MARK, + + /* Single bits which can be set by efx_mae_match_spec_bit_set(). */ + EFX_MAE_FIELD_IS_IP_FRAG, + EFX_MAE_FIELD_IP_FIRST_FRAG, EFX_MAE_FIELD_NIDS } efx_mae_field_id_t; diff --git a/drivers/common/sfc_efx/base/efx_mae.c b/drivers/common/sfc_efx/base/efx_mae.c index 9ff887e04b..d36cdc71be 100644 --- a/drivers/common/sfc_efx/base/efx_mae.c +++ b/drivers/common/sfc_efx/base/efx_mae.c @@ -483,6 +483,8 @@ typedef enum efx_mae_field_cap_id_e { EFX_MAE_FIELD_ID_ENC_HAS_IVLAN = MAE_FIELD_ENC_HAS_IVLAN, EFX_MAE_FIELD_ID_RECIRC_ID = MAE_FIELD_RECIRC_ID, EFX_MAE_FIELD_ID_CT_MARK = MAE_FIELD_CT_MARK, + EFX_MAE_FIELD_ID_IS_IP_FRAG = MAE_FIELD_IS_IP_FRAG, + EFX_MAE_FIELD_ID_IP_FIRST_FRAG = MAE_FIELD_IP_FIRST_FRAG, EFX_MAE_FIELD_CAP_NIDS } efx_mae_field_cap_id_t; @@ -668,6 +670,8 @@ static const efx_mae_mv_bit_desc_t __efx_mae_action_rule_mv_bit_desc_set[] = { EFX_MAE_MV_BIT_DESC(HAS_IVLAN), EFX_MAE_MV_BIT_DESC(ENC_HAS_OVLAN), EFX_MAE_MV_BIT_DESC(ENC_HAS_IVLAN), + EFX_MAE_MV_BIT_DESC(IS_IP_FRAG), + EFX_MAE_MV_BIT_DESC(IP_FIRST_FRAG), #undef EFX_MAE_MV_BIT_DESC }; -- 2.34.1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 2/2] net/sfc: support IPv4 fragment matching in transfer rules 2023-06-23 16:58 [PATCH 0/2] support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-23 16:58 ` [PATCH 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov @ 2023-06-23 16:58 ` Artemii Morozov 2023-06-27 14:02 ` [PATCH 0/2] " Ferruh Yigit 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov 3 siblings, 0 replies; 8+ messages in thread From: Artemii Morozov @ 2023-06-23 16:58 UTC (permalink / raw) To: dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko, Ferruh Yigit This patch adds the support for IPv4 fragmented packets, but with some limitations: for non-zero fragment offset an exact match is not allowed, but ranges are allowed. Signed-off-by: Artemii Morozov <artemii.morozov@arknetworks.am> Reviewed-by: Viacheslav Galaktionov <viacheslav.galaktionov@arknetworks.am> Reviewed-by: Andy Moreton <amoreton@xilinx.com> --- drivers/net/sfc/sfc_mae.c | 110 +++++++++++++++++++++++++++++++++++++- drivers/net/sfc/sfc_mae.h | 4 ++ 2 files changed, 113 insertions(+), 1 deletion(-) diff --git a/drivers/net/sfc/sfc_mae.c b/drivers/net/sfc/sfc_mae.c index a3131c244d..d1503f39b4 100644 --- a/drivers/net/sfc/sfc_mae.c +++ b/drivers/net/sfc/sfc_mae.c @@ -1700,6 +1700,80 @@ sfc_mae_rule_process_pattern_data(struct sfc_mae_parse_ctx *ctx, if (rc != 0) goto fail; + if (pdata->l3_frag_ofst_mask != 0) { + const rte_be16_t hdr_mask = RTE_BE16(RTE_IPV4_HDR_OFFSET_MASK); + rte_be16_t value; + rte_be16_t last; + boolean_t first_frag; + boolean_t is_ip_frag; + boolean_t any_frag; + + if (pdata->l3_frag_ofst_mask & RTE_BE16(RTE_IPV4_HDR_DF_FLAG)) { + sfc_err(ctx->sa, "Don't fragment flag is not supported."); + rc = ENOTSUP; + goto fail; + } + + if ((pdata->l3_frag_ofst_mask & hdr_mask) != hdr_mask) { + sfc_err(ctx->sa, "Invalid value for fragment offset mask."); + rc = EINVAL; + goto fail; + } + + value = pdata->l3_frag_ofst_mask & pdata->l3_frag_ofst_value; + last = pdata->l3_frag_ofst_mask & pdata->l3_frag_ofst_last; + + /* + * value: last: matches: + * 0 0 Non-fragmented packet + * 1 0x1fff Non-first fragment + * 1 0x1fff+MF Any fragment + * MF 0 First fragment + */ + if (last == 0 && + (pdata->l3_frag_ofst_value & hdr_mask) != 0) { + sfc_err(ctx->sa, + "Exact matching is prohibited for non-zero offsets, but ranges are allowed."); + rc = EINVAL; + goto fail; + } + + if (value == 0 && last == 0) { + is_ip_frag = false; + any_frag = true; + } else if (value == RTE_BE16(1) && (last & hdr_mask) == hdr_mask) { + if (last & RTE_BE16(RTE_IPV4_HDR_MF_FLAG)) { + is_ip_frag = true; + any_frag = true; + } else { + is_ip_frag = true; + any_frag = false; + first_frag = false; + } + } else if (value == RTE_BE16(RTE_IPV4_HDR_MF_FLAG) && last == 0) { + is_ip_frag = true; + any_frag = false; + first_frag = true; + } else { + sfc_err(ctx->sa, "Invalid value for fragment offset."); + rc = EINVAL; + goto fail; + } + + rc = efx_mae_match_spec_bit_set(ctx->match_spec, + fremap[EFX_MAE_FIELD_IS_IP_FRAG], is_ip_frag); + if (rc != 0) + goto fail; + + if (!any_frag) { + rc = efx_mae_match_spec_bit_set(ctx->match_spec, + fremap[EFX_MAE_FIELD_IP_FIRST_FRAG], + first_frag); + if (rc != 0) + goto fail; + } + } + return 0; fail: @@ -2208,6 +2282,15 @@ static const struct sfc_mae_field_locator flocs_ipv4[] = { RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, hdr.next_proto_id), offsetof(struct rte_flow_item_ipv4, hdr.next_proto_id), }, + { + /* + * This locator is used only for building supported fields mask. + * The field is handled by sfc_mae_rule_process_pattern_data(). + */ + SFC_MAE_FIELD_HANDLING_DEFERRED, + RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, hdr.fragment_offset), + offsetof(struct rte_flow_item_ipv4, hdr.fragment_offset), + }, { EFX_MAE_FIELD_IP_TOS, RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, @@ -2230,14 +2313,30 @@ sfc_mae_rule_parse_item_ipv4(const struct rte_flow_item *item, struct sfc_mae_parse_ctx *ctx_mae = ctx->mae; struct sfc_mae_pattern_data *pdata = &ctx_mae->pattern_data; struct rte_flow_item_ipv4 supp_mask; + struct rte_flow_item item_dup; const uint8_t *spec = NULL; const uint8_t *mask = NULL; + uint8_t *last = NULL; int rc; + item_dup.spec = item->spec; + item_dup.mask = item->mask; + item_dup.last = item->last; + item_dup.type = item->type; + sfc_mae_item_build_supp_mask(flocs_ipv4, RTE_DIM(flocs_ipv4), &supp_mask, sizeof(supp_mask)); - rc = sfc_flow_parse_init(item, + /* We don't support IPv4 fragmentation in the outer frames. */ + if (ctx_mae->match_spec != ctx_mae->match_spec_action) + supp_mask.hdr.fragment_offset = 0; + + if (item != NULL && item->last != NULL) { + last = item->last; + item_dup.last = NULL; + } + + rc = sfc_flow_parse_init(&item_dup, (const void **)&spec, (const void **)&mask, (const void *)&supp_mask, &rte_flow_item_ipv4_mask, @@ -2251,12 +2350,19 @@ sfc_mae_rule_parse_item_ipv4(const struct rte_flow_item *item, if (spec != NULL) { const struct rte_flow_item_ipv4 *item_spec; const struct rte_flow_item_ipv4 *item_mask; + const struct rte_flow_item_ipv4 *item_last; item_spec = (const struct rte_flow_item_ipv4 *)spec; item_mask = (const struct rte_flow_item_ipv4 *)mask; + if (last != NULL) + item_last = (const struct rte_flow_item_ipv4 *)last; pdata->l3_next_proto_value = item_spec->hdr.next_proto_id; pdata->l3_next_proto_mask = item_mask->hdr.next_proto_id; + pdata->l3_frag_ofst_mask = item_mask->hdr.fragment_offset; + pdata->l3_frag_ofst_value = item_spec->hdr.fragment_offset; + if (last != NULL) + pdata->l3_frag_ofst_last = item_last->hdr.fragment_offset; } else { return 0; } @@ -2518,6 +2624,8 @@ static const efx_mae_field_id_t field_ids_no_remap[] = { FIELD_ID_NO_REMAP(TCP_FLAGS_BE), FIELD_ID_NO_REMAP(HAS_OVLAN), FIELD_ID_NO_REMAP(HAS_IVLAN), + FIELD_ID_NO_REMAP(IS_IP_FRAG), + FIELD_ID_NO_REMAP(IP_FIRST_FRAG), #undef FIELD_ID_NO_REMAP }; diff --git a/drivers/net/sfc/sfc_mae.h b/drivers/net/sfc/sfc_mae.h index 80585c0e93..059718e383 100644 --- a/drivers/net/sfc/sfc_mae.h +++ b/drivers/net/sfc/sfc_mae.h @@ -334,6 +334,10 @@ struct sfc_mae_pattern_data { uint8_t l3_next_proto_restriction_value; uint8_t l3_next_proto_restriction_mask; + rte_be16_t l3_frag_ofst_value; + rte_be16_t l3_frag_ofst_mask; + rte_be16_t l3_frag_ofst_last; + /* Projected state of EFX_MAE_FIELD_HAS_OVLAN match bit */ bool has_ovlan_value; bool has_ovlan_mask; -- 2.34.1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 0/2] support IPv4 fragment matching in transfer rules 2023-06-23 16:58 [PATCH 0/2] support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-23 16:58 ` [PATCH 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov 2023-06-23 16:58 ` [PATCH 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov @ 2023-06-27 14:02 ` Ferruh Yigit 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov 3 siblings, 0 replies; 8+ messages in thread From: Ferruh Yigit @ 2023-06-27 14:02 UTC (permalink / raw) To: Artemii Morozov, dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko On 6/23/2023 5:58 PM, Artemii Morozov wrote: > This patch seria adds support for matching fragment offsets for IPv4 packets, > but with some limitations. Exact matching is supported only for zero offset, > ranges are allowed for non-zero offsets. > > Artemii Morozov (2): > common/sfc_efx/base: add MAE IP fragmentation match bits > net/sfc: support IPv4 fragment matching in transfer rules > Hi Artemii, There is a build error, can you please fix it: ../drivers/net/sfc/sfc_mae.c: In function ‘sfc_mae_rule_parse_item_ipv4’: ../drivers/net/sfc/sfc_mae.c:2335:22: error: assignment discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers] 2335 | last = item->last; | ^ cc1: all warnings being treated as errors ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2 0/2] support IPv4 fragment matching in transfer rules 2023-06-23 16:58 [PATCH 0/2] support IPv4 fragment matching in transfer rules Artemii Morozov ` (2 preceding siblings ...) 2023-06-27 14:02 ` [PATCH 0/2] " Ferruh Yigit @ 2023-06-28 7:00 ` Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov ` (2 more replies) 3 siblings, 3 replies; 8+ messages in thread From: Artemii Morozov @ 2023-06-28 7:00 UTC (permalink / raw) To: dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko, Ferruh Yigit This patch seria adds support for matching fragment offsets for IPv4 packets, but with some limitations. Exact matching is supported only for zero offset, ranges are allowed for non-zero offsets. v2: * fix build issue Artemii Morozov (2): common/sfc_efx/base: add MAE IP fragmentation match bits net/sfc: support IPv4 fragment matching in transfer rules drivers/common/sfc_efx/base/efx.h | 4 + drivers/common/sfc_efx/base/efx_mae.c | 4 + drivers/net/sfc/sfc_mae.c | 110 +++++++++++++++++++++++++- drivers/net/sfc/sfc_mae.h | 4 + 4 files changed, 121 insertions(+), 1 deletion(-) -- 2.34.1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov @ 2023-06-28 7:00 ` Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-28 10:53 ` [PATCH v2 0/2] " Ferruh Yigit 2 siblings, 0 replies; 8+ messages in thread From: Artemii Morozov @ 2023-06-28 7:00 UTC (permalink / raw) To: dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko, Ferruh Yigit Introduce necessary infrastructure for these bits to be set and validated. Using a combination of these bits IP fragmentation can be configured. Signed-off-by: Artemii Morozov <artemii.morozov@arknetworks.am> Reviewed-by: Ivan Malov <ivan.malov@arknetworks.am> Reviewed-by: Andy Moreton <amoreton@xilinx.com> --- drivers/common/sfc_efx/base/efx.h | 4 ++++ drivers/common/sfc_efx/base/efx_mae.c | 4 ++++ 2 files changed, 8 insertions(+) diff --git a/drivers/common/sfc_efx/base/efx.h b/drivers/common/sfc_efx/base/efx.h index 77f855bfb0..efefea717f 100644 --- a/drivers/common/sfc_efx/base/efx.h +++ b/drivers/common/sfc_efx/base/efx.h @@ -4317,6 +4317,10 @@ typedef enum efx_mae_field_id_e { */ EFX_MAE_FIELD_RECIRC_ID, EFX_MAE_FIELD_CT_MARK, + + /* Single bits which can be set by efx_mae_match_spec_bit_set(). */ + EFX_MAE_FIELD_IS_IP_FRAG, + EFX_MAE_FIELD_IP_FIRST_FRAG, EFX_MAE_FIELD_NIDS } efx_mae_field_id_t; diff --git a/drivers/common/sfc_efx/base/efx_mae.c b/drivers/common/sfc_efx/base/efx_mae.c index 9ff887e04b..d36cdc71be 100644 --- a/drivers/common/sfc_efx/base/efx_mae.c +++ b/drivers/common/sfc_efx/base/efx_mae.c @@ -483,6 +483,8 @@ typedef enum efx_mae_field_cap_id_e { EFX_MAE_FIELD_ID_ENC_HAS_IVLAN = MAE_FIELD_ENC_HAS_IVLAN, EFX_MAE_FIELD_ID_RECIRC_ID = MAE_FIELD_RECIRC_ID, EFX_MAE_FIELD_ID_CT_MARK = MAE_FIELD_CT_MARK, + EFX_MAE_FIELD_ID_IS_IP_FRAG = MAE_FIELD_IS_IP_FRAG, + EFX_MAE_FIELD_ID_IP_FIRST_FRAG = MAE_FIELD_IP_FIRST_FRAG, EFX_MAE_FIELD_CAP_NIDS } efx_mae_field_cap_id_t; @@ -668,6 +670,8 @@ static const efx_mae_mv_bit_desc_t __efx_mae_action_rule_mv_bit_desc_set[] = { EFX_MAE_MV_BIT_DESC(HAS_IVLAN), EFX_MAE_MV_BIT_DESC(ENC_HAS_OVLAN), EFX_MAE_MV_BIT_DESC(ENC_HAS_IVLAN), + EFX_MAE_MV_BIT_DESC(IS_IP_FRAG), + EFX_MAE_MV_BIT_DESC(IP_FIRST_FRAG), #undef EFX_MAE_MV_BIT_DESC }; -- 2.34.1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2 2/2] net/sfc: support IPv4 fragment matching in transfer rules 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov @ 2023-06-28 7:00 ` Artemii Morozov 2023-06-28 10:53 ` [PATCH v2 0/2] " Ferruh Yigit 2 siblings, 0 replies; 8+ messages in thread From: Artemii Morozov @ 2023-06-28 7:00 UTC (permalink / raw) To: dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko, Ferruh Yigit This patch adds the support for IPv4 fragmented packets, but with some limitations: for non-zero fragment offset an exact match is not allowed, but ranges are allowed. Signed-off-by: Artemii Morozov <artemii.morozov@arknetworks.am> Reviewed-by: Viacheslav Galaktionov <viacheslav.galaktionov@arknetworks.am> Reviewed-by: Andy Moreton <amoreton@xilinx.com> --- drivers/net/sfc/sfc_mae.c | 110 +++++++++++++++++++++++++++++++++++++- drivers/net/sfc/sfc_mae.h | 4 ++ 2 files changed, 113 insertions(+), 1 deletion(-) diff --git a/drivers/net/sfc/sfc_mae.c b/drivers/net/sfc/sfc_mae.c index a3131c244d..63f6536243 100644 --- a/drivers/net/sfc/sfc_mae.c +++ b/drivers/net/sfc/sfc_mae.c @@ -1700,6 +1700,80 @@ sfc_mae_rule_process_pattern_data(struct sfc_mae_parse_ctx *ctx, if (rc != 0) goto fail; + if (pdata->l3_frag_ofst_mask != 0) { + const rte_be16_t hdr_mask = RTE_BE16(RTE_IPV4_HDR_OFFSET_MASK); + rte_be16_t value; + rte_be16_t last; + boolean_t first_frag; + boolean_t is_ip_frag; + boolean_t any_frag; + + if (pdata->l3_frag_ofst_mask & RTE_BE16(RTE_IPV4_HDR_DF_FLAG)) { + sfc_err(ctx->sa, "Don't fragment flag is not supported."); + rc = ENOTSUP; + goto fail; + } + + if ((pdata->l3_frag_ofst_mask & hdr_mask) != hdr_mask) { + sfc_err(ctx->sa, "Invalid value for fragment offset mask."); + rc = EINVAL; + goto fail; + } + + value = pdata->l3_frag_ofst_mask & pdata->l3_frag_ofst_value; + last = pdata->l3_frag_ofst_mask & pdata->l3_frag_ofst_last; + + /* + * value: last: matches: + * 0 0 Non-fragmented packet + * 1 0x1fff Non-first fragment + * 1 0x1fff+MF Any fragment + * MF 0 First fragment + */ + if (last == 0 && + (pdata->l3_frag_ofst_value & hdr_mask) != 0) { + sfc_err(ctx->sa, + "Exact matching is prohibited for non-zero offsets, but ranges are allowed."); + rc = EINVAL; + goto fail; + } + + if (value == 0 && last == 0) { + is_ip_frag = false; + any_frag = true; + } else if (value == RTE_BE16(1) && (last & hdr_mask) == hdr_mask) { + if (last & RTE_BE16(RTE_IPV4_HDR_MF_FLAG)) { + is_ip_frag = true; + any_frag = true; + } else { + is_ip_frag = true; + any_frag = false; + first_frag = false; + } + } else if (value == RTE_BE16(RTE_IPV4_HDR_MF_FLAG) && last == 0) { + is_ip_frag = true; + any_frag = false; + first_frag = true; + } else { + sfc_err(ctx->sa, "Invalid value for fragment offset."); + rc = EINVAL; + goto fail; + } + + rc = efx_mae_match_spec_bit_set(ctx->match_spec, + fremap[EFX_MAE_FIELD_IS_IP_FRAG], is_ip_frag); + if (rc != 0) + goto fail; + + if (!any_frag) { + rc = efx_mae_match_spec_bit_set(ctx->match_spec, + fremap[EFX_MAE_FIELD_IP_FIRST_FRAG], + first_frag); + if (rc != 0) + goto fail; + } + } + return 0; fail: @@ -2208,6 +2282,15 @@ static const struct sfc_mae_field_locator flocs_ipv4[] = { RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, hdr.next_proto_id), offsetof(struct rte_flow_item_ipv4, hdr.next_proto_id), }, + { + /* + * This locator is used only for building supported fields mask. + * The field is handled by sfc_mae_rule_process_pattern_data(). + */ + SFC_MAE_FIELD_HANDLING_DEFERRED, + RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, hdr.fragment_offset), + offsetof(struct rte_flow_item_ipv4, hdr.fragment_offset), + }, { EFX_MAE_FIELD_IP_TOS, RTE_SIZEOF_FIELD(struct rte_flow_item_ipv4, @@ -2230,14 +2313,30 @@ sfc_mae_rule_parse_item_ipv4(const struct rte_flow_item *item, struct sfc_mae_parse_ctx *ctx_mae = ctx->mae; struct sfc_mae_pattern_data *pdata = &ctx_mae->pattern_data; struct rte_flow_item_ipv4 supp_mask; + struct rte_flow_item item_dup; const uint8_t *spec = NULL; const uint8_t *mask = NULL; + const uint8_t *last = NULL; int rc; + item_dup.spec = item->spec; + item_dup.mask = item->mask; + item_dup.last = item->last; + item_dup.type = item->type; + sfc_mae_item_build_supp_mask(flocs_ipv4, RTE_DIM(flocs_ipv4), &supp_mask, sizeof(supp_mask)); - rc = sfc_flow_parse_init(item, + /* We don't support IPv4 fragmentation in the outer frames. */ + if (ctx_mae->match_spec != ctx_mae->match_spec_action) + supp_mask.hdr.fragment_offset = 0; + + if (item != NULL && item->last != NULL) { + last = item->last; + item_dup.last = NULL; + } + + rc = sfc_flow_parse_init(&item_dup, (const void **)&spec, (const void **)&mask, (const void *)&supp_mask, &rte_flow_item_ipv4_mask, @@ -2251,12 +2350,19 @@ sfc_mae_rule_parse_item_ipv4(const struct rte_flow_item *item, if (spec != NULL) { const struct rte_flow_item_ipv4 *item_spec; const struct rte_flow_item_ipv4 *item_mask; + const struct rte_flow_item_ipv4 *item_last; item_spec = (const struct rte_flow_item_ipv4 *)spec; item_mask = (const struct rte_flow_item_ipv4 *)mask; + if (last != NULL) + item_last = (const struct rte_flow_item_ipv4 *)last; pdata->l3_next_proto_value = item_spec->hdr.next_proto_id; pdata->l3_next_proto_mask = item_mask->hdr.next_proto_id; + pdata->l3_frag_ofst_mask = item_mask->hdr.fragment_offset; + pdata->l3_frag_ofst_value = item_spec->hdr.fragment_offset; + if (last != NULL) + pdata->l3_frag_ofst_last = item_last->hdr.fragment_offset; } else { return 0; } @@ -2518,6 +2624,8 @@ static const efx_mae_field_id_t field_ids_no_remap[] = { FIELD_ID_NO_REMAP(TCP_FLAGS_BE), FIELD_ID_NO_REMAP(HAS_OVLAN), FIELD_ID_NO_REMAP(HAS_IVLAN), + FIELD_ID_NO_REMAP(IS_IP_FRAG), + FIELD_ID_NO_REMAP(IP_FIRST_FRAG), #undef FIELD_ID_NO_REMAP }; diff --git a/drivers/net/sfc/sfc_mae.h b/drivers/net/sfc/sfc_mae.h index 80585c0e93..059718e383 100644 --- a/drivers/net/sfc/sfc_mae.h +++ b/drivers/net/sfc/sfc_mae.h @@ -334,6 +334,10 @@ struct sfc_mae_pattern_data { uint8_t l3_next_proto_restriction_value; uint8_t l3_next_proto_restriction_mask; + rte_be16_t l3_frag_ofst_value; + rte_be16_t l3_frag_ofst_mask; + rte_be16_t l3_frag_ofst_last; + /* Projected state of EFX_MAE_FIELD_HAS_OVLAN match bit */ bool has_ovlan_value; bool has_ovlan_mask; -- 2.34.1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 0/2] support IPv4 fragment matching in transfer rules 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov @ 2023-06-28 10:53 ` Ferruh Yigit 2 siblings, 0 replies; 8+ messages in thread From: Ferruh Yigit @ 2023-06-28 10:53 UTC (permalink / raw) To: Artemii Morozov, dev Cc: Ivan Malov, Viacheslav Galaktionov, Andy Moreton, Andrew Rybchenko On 6/28/2023 8:00 AM, Artemii Morozov wrote: > This patch seria adds support for matching fragment offsets for IPv4 packets, > but with some limitations. Exact matching is supported only for zero offset, > ranges are allowed for non-zero offsets. > > v2: > * fix build issue > > Artemii Morozov (2): > common/sfc_efx/base: add MAE IP fragmentation match bits > net/sfc: support IPv4 fragment matching in transfer rules > Base code change squashed into net/sfc one, Applied to dpdk-next-net/main, thanks. ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2023-06-28 10:54 UTC | newest] Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2023-06-23 16:58 [PATCH 0/2] support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-23 16:58 ` [PATCH 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov 2023-06-23 16:58 ` [PATCH 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-27 14:02 ` [PATCH 0/2] " Ferruh Yigit 2023-06-28 7:00 ` [PATCH v2 " Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 1/2] common/sfc_efx/base: add MAE IP fragmentation match bits Artemii Morozov 2023-06-28 7:00 ` [PATCH v2 2/2] net/sfc: support IPv4 fragment matching in transfer rules Artemii Morozov 2023-06-28 10:53 ` [PATCH v2 0/2] " Ferruh Yigit
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).