From: "Liu, Mingxia" <mingxia.liu@intel.com>
To: "Qiao, Wenjing" <wenjing.qiao@intel.com>,
"Zhang, Yuying" <yuying.zhang@intel.com>,
"dev@dpdk.org" <dev@dpdk.org>,
"Zhang, Qi Z" <qi.z.zhang@intel.com>,
"Wu, Jingjing" <jingjing.wu@intel.com>,
"Xing, Beilei" <beilei.xing@intel.com>
Subject: RE: [PATCH v3 2/9] net/cpfl: add flow json parser
Date: Fri, 8 Sep 2023 06:26:52 +0000 [thread overview]
Message-ID: <PH0PR11MB5877A249B62C545CF39BB2DFECEDA@PH0PR11MB5877.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20230906093407.3635038-3-wenjing.qiao@intel.com>
> -----Original Message-----
> From: Qiao, Wenjing <wenjing.qiao@intel.com>
> Sent: Wednesday, September 6, 2023 5:34 PM
> To: Zhang, Yuying <yuying.zhang@intel.com>; dev@dpdk.org; Zhang, Qi Z
> <qi.z.zhang@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>
> Cc: Liu, Mingxia <mingxia.liu@intel.com>; Qiao, Wenjing
> <wenjing.qiao@intel.com>
> Subject: [PATCH v3 2/9] net/cpfl: add flow json parser
>
> A JSON file will be used to direct DPDK CPF PMD to
> parse rte_flow tokens into low level hardware resources
> defined in a DDP package file.
>
> Signed-off-by: Wenjing Qiao <wenjing.qiao@intel.com>
> ---
> drivers/net/cpfl/cpfl_ethdev.h | 70 +
> drivers/net/cpfl/cpfl_flow_parser.c | 1910 +++++++++++++++++++++++++++
> drivers/net/cpfl/cpfl_flow_parser.h | 236 ++++
> drivers/net/cpfl/meson.build | 3 +
> 4 files changed, 2219 insertions(+)
> create mode 100644 drivers/net/cpfl/cpfl_flow_parser.c
> create mode 100644 drivers/net/cpfl/cpfl_flow_parser.h
>
> +static int
> +cpfl_flow_js_pattern_key_attr(json_object *cjson_pr_key_attr, struct
> cpfl_flow_js_pr *js_pr)
> +{
> + int i, len;
> + struct cpfl_flow_js_pr_key_attr *attr;
> +
> + len = json_object_array_length(cjson_pr_key_attr);
> + js_pr->key.attributes = rte_malloc(NULL, sizeof(struct
> cpfl_flow_js_pr_key_attr), 0);
> + if (!js_pr->key.attributes) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + js_pr->key.attr_size = len;
> + attr = js_pr->key.attributes;
> + for (i = 0; i < len; i++) {
> + json_object *object;
> + const char *name;
> + uint16_t value = 0;
> + int ret;
> +
> + object = json_object_array_get_idx(cjson_pr_key_attr, i);
> + name = cpfl_json_object_to_string(object, "Name");
> + if (!name) {
> + rte_free(js_pr->key.attributes);
> + PMD_DRV_LOG(ERR, "Can not parse string 'Name'.");
> + return -EINVAL;
[Liu, Mingxia] Better to use goto statement as other similar function do?
> + }
> + ret = cpfl_json_object_to_uint16(object, "Value", &value);
> + if (ret < 0) {
> + rte_free(js_pr->key.attributes);
> + PMD_DRV_LOG(ERR, "Can not parse 'value'.");
> + return -EINVAL;
> + }
> + if (strcmp(name, "ingress") == 0) {
> + attr->ingress = value;
> + } else if (strcmp(name, "egress") == 0) {
> + attr->egress = value;
> + } else {
> + /* TODO: more... */
> + rte_free(js_pr->key.attributes);
> + PMD_DRV_LOG(ERR, "Not support attr name: %s.",
> name);
> + return -EINVAL;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_key_proto_field(json_object *cjson_field,
> + struct cpfl_flow_js_pr_key_proto *js_field)
> +{
> + int len, i;
> +
> + if (!cjson_field)
> + return 0;
> + len = json_object_array_length(cjson_field);
> + js_field->fields_size = len;
> + if (len == 0)
> + return 0;
> + js_field->fields =
> + rte_malloc(NULL, sizeof(struct cpfl_flow_js_pr_key_proto_field) * len,
> 0);
> + if (!js_field->fields) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + for (i = 0; i < len; i++) {
> + json_object *object;
> + const char *name, *mask;
> +
> + object = json_object_array_get_idx(cjson_field, i);
> + name = cpfl_json_object_to_string(object, "name");
> + if (!name) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'name'.");
> + goto err;
> + }
> + if (strlen(name) > CPFL_FLOW_JSON_STR_SIZE_MAX) {
> + PMD_DRV_LOG(ERR, "The 'name' is too long.");
> + goto err;
> + }
> + memcpy(js_field->fields[i].name, name, strlen(name));
> +
> + if (js_field->type == RTE_FLOW_ITEM_TYPE_ETH ||
> + js_field->type == RTE_FLOW_ITEM_TYPE_IPV4) {
> + mask = cpfl_json_object_to_string(object, "mask");
> + if (!mask) {
> + PMD_DRV_LOG(ERR, "Can not parse string
> 'mask'.");
> + goto err;
> + }
> + memcpy(js_field->fields[i].mask, mask, strlen(mask));
[Liu, Mingxia] Need to check the length and validation of mask?
> + } else {
> + uint32_t mask_32b;
> + int ret;
> +
> + ret = cpfl_json_object_to_uint32(object, "mask",
> &mask_32b);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse uint32
> 'mask'.");
> + goto err;
> + }
> + js_field->fields[i].mask_32b = mask_32b;
> + }
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(js_field->fields);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_key_proto(json_object *cjson_pr_key_proto, struct
> cpfl_flow_js_pr *js_pr)
> +{
> + int len, i, ret;
> +
> + len = json_object_array_length(cjson_pr_key_proto);
> + js_pr->key.proto_size = len;
> + js_pr->key.protocols = rte_malloc(NULL, sizeof(struct
> cpfl_flow_js_pr_key_proto) * len, 0);
> + if (!js_pr->key.protocols) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> +
> + for (i = 0; i < len; i++) {
> + json_object *object, *cjson_pr_key_proto_fields;
> + const char *type;
> + enum rte_flow_item_type item_type;
> +
> + object = json_object_array_get_idx(cjson_pr_key_proto, i);
> + /* pr->key->proto->type */
> + type = cpfl_json_object_to_string(object, "type");
> + if (!type) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'type'.");
> + goto err;
> + }
> + item_type = cpfl_get_item_type_by_str(type);
> + if (item_type == RTE_FLOW_ITEM_TYPE_VOID)
> + goto err;
> + js_pr->key.protocols[i].type = item_type;
> + /* pr->key->proto->fields */
> + cjson_pr_key_proto_fields = json_object_object_get(object,
> "fields");
> + ret =
> cpfl_flow_js_pattern_key_proto_field(cjson_pr_key_proto_fields,
> + &js_pr-
> >key.protocols[i]);
> + if (ret < 0)
> + goto err;
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(js_pr->key.protocols);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_act_fv_proto(json_object *cjson_value, struct
> cpfl_flow_js_fv *js_fv)
> +{
> + uint16_t layer = 0, offset = 0, mask = 0;
> + const char *header;
> + enum rte_flow_item_type type;
> + int ret;
> +
> + ret = cpfl_json_object_to_uint16(cjson_value, "layer", &layer);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'value'.");
> + return -EINVAL;
> + }
> +
> + header = cpfl_json_object_to_string(cjson_value, "header");
> + if (!header) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'header'.");
> + return -EINVAL;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_value, "offset", &offset);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'offset'.");
> + return -EINVAL;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_value, "mask", &mask);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'mask'.");
> + return -EINVAL;
> + }
> + js_fv->proto.layer = layer;
> + js_fv->proto.offset = offset;
> + js_fv->proto.mask = mask;
> + type = cpfl_get_item_type_by_str(header);
> + if (type == RTE_FLOW_ITEM_TYPE_VOID)
> + return -EINVAL;
> + js_fv->proto.header = type;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_act_fv_metadata(json_object *cjson_value, struct
> cpfl_flow_js_fv *js_fv)
> +{
> + int ret;
> +
> + ret = cpfl_json_object_to_uint16(cjson_value, "type", &js_fv-
> >meta.type);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'size'.");
> + return ret;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_value, "offset", &js_fv-
> >meta.offset);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'size'.");
> + return ret;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_value, "mask", &js_fv-
> >meta.mask);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'size'.");
> + return ret;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_act_fv(json_object *cjson_fv, struct
> cpfl_flow_js_pr_action *js_act)
> +{
> + int len, i;
> +
> + len = json_object_array_length(cjson_fv);
> + js_act->sem.fv = rte_malloc(NULL, sizeof(struct cpfl_flow_js_fv) * len,
> 0);
> + if (!js_act->sem.fv) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + js_act->sem.fv_size = len;
> + for (i = 0; i < len; i++) {
> + struct cpfl_flow_js_fv *js_fv;
> + json_object *object, *cjson_value;
> + uint16_t offset = 0;
> + const char *type;
> + int ret;
> +
> + object = json_object_array_get_idx(cjson_fv, i);
> + js_fv = &js_act->sem.fv[i];
> + ret = cpfl_json_object_to_uint16(object, "offset", &offset);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'offset'.");
> + goto err;
> + }
> + js_fv->offset = offset;
> + type = cpfl_json_object_to_string(object, "type");
> + if (!type) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'type'.");
> + goto err;
> + }
> + cjson_value = json_object_object_get(object, "value");
> + if (strcmp(type, "immediate") == 0) {
> + js_fv->type = CPFL_FV_TYPE_IMMEDIATE;
> + js_fv->immediate = json_object_get_int(cjson_value);
> + } else if (strcmp(type, "metadata") == 0) {
> + js_fv->type = CPFL_FV_TYPE_METADATA;
> + cpfl_flow_js_pattern_act_fv_metadata(cjson_value,
> js_fv);
> + } else if (strcmp(type, "protocol") == 0) {
> + js_fv->type = CPFL_FV_TYPE_PROTOCOL;
> + cpfl_flow_js_pattern_act_fv_proto(cjson_value, js_fv);
> + } else {
> + PMD_DRV_LOG(ERR, "Not support this type: %s.",
> type);
> + goto err;
> + }
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(js_act->sem.fv);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_per_act(json_object *cjson_per_act, struct
> cpfl_flow_js_pr_action *js_act)
> +{
> + const char *type;
> + int ret;
> +
> + /* pr->actions->type */
> + type = cpfl_json_object_to_string(cjson_per_act, "type");
> + if (!type) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'type'.");
> + return -EINVAL;
> + }
> + /* pr->actions->data */
> + if (strcmp(type, "sem") == 0) {
> + json_object *cjson_fv, *cjson_pr_action_sem;
> +
> + js_act->type = CPFL_JS_PR_ACTION_TYPE_SEM;
> + cjson_pr_action_sem = json_object_object_get(cjson_per_act,
> "data");
> + ret = cpfl_json_object_to_uint16(cjson_pr_action_sem,
> "profile",
> + &js_act->sem.prof);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'profile'.");
> + return -EINVAL;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_pr_action_sem,
> "subprofile",
> + &js_act->sem.subprof);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'subprofile'.");
> + return -EINVAL;
> + }
> + ret = cpfl_json_object_to_uint16(cjson_pr_action_sem,
> "keysize",
> + &js_act->sem.keysize);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'keysize'.");
> + return -EINVAL;
> + }
> + cjson_fv = json_object_object_get(cjson_pr_action_sem,
> "fieldvectors");
> + ret = cpfl_flow_js_pattern_act_fv(cjson_fv, js_act);
> + if (ret < 0)
> + return ret;
> + } else {
> + PMD_DRV_LOG(ERR, "Not support this type: %s.", type);
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_act(json_object *cjson_pr_act, struct cpfl_flow_js_pr
> *js_pr)
> +{
> + int i, len, ret;
> +
> + len = json_object_array_length(cjson_pr_act);
> + js_pr->actions = rte_malloc(NULL, sizeof(struct cpfl_flow_js_pr_action)
> * len, 0);
> + if (!js_pr->actions) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + js_pr->actions_size = len;
> + for (i = 0; i < len; i++) {
> + struct cpfl_flow_js_pr_action *js_act;
> + json_object *object;
> +
> + object = json_object_array_get_idx(cjson_pr_act, i);
> + js_act = &js_pr->actions[i];
> + ret = cpfl_flow_js_pattern_per_act(object, js_act);
> + if (ret < 0) {
> + rte_free(js_pr->actions);
> + PMD_DRV_LOG(ERR, "Can not parse pattern action.");
> + return -EINVAL;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_pattern_rule(json_object *json_root, struct cpfl_flow_js_parser
> *parser)
> +{
> + json_object *cjson_pr;
> + int i, len;
> +
> + /* Pattern Rules */
> + cjson_pr = json_object_object_get(json_root, "patterns");
> + if (!cjson_pr) {
> + PMD_DRV_LOG(ERR, "The patterns is mandatory.");
> + return -EINVAL;
> + }
> +
> + len = json_object_array_length(cjson_pr);
> + parser->patterns = rte_malloc(NULL, sizeof(struct cpfl_flow_js_pr) * len,
> 0);
> + if (!parser->patterns) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + parser->pr_size = len;
> + for (i = 0; i < len; i++) {
> + json_object *object, *cjson_pr_actions, *cjson_pr_key,
> *cjson_pr_key_proto,
> + *cjson_pr_key_attr;
> + int ret;
> +
> + object = json_object_array_get_idx(cjson_pr, i);
> + /* pr->key */
> + cjson_pr_key = json_object_object_get(object, "key");
> + /* pr->key->protocols */
> + cjson_pr_key_proto = json_object_object_get(cjson_pr_key,
> "protocols");
> + ret = cpfl_flow_js_pattern_key_proto(cjson_pr_key_proto,
> &parser->patterns[i]);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse key->protocols.");
> + goto err;
> + }
> + /* pr->key->attributes */
> + cjson_pr_key_attr = json_object_object_get(cjson_pr_key,
> "attributes");
> + ret = cpfl_flow_js_pattern_key_attr(cjson_pr_key_attr, &parser-
> >patterns[i]);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse key->attributes.");
> + goto err;
> + }
> + /* pr->actions */
> + cjson_pr_actions = json_object_object_get(object, "actions");
> + ret = cpfl_flow_js_pattern_act(cjson_pr_actions, &parser-
> >patterns[i]);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse pattern action.");
> + goto err;
> + }
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(parser->patterns);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_mr_key(json_object *cjson_mr_key, struct cpfl_flow_js_mr_key
> *js_mr_key)
> +{
> + int len, i;
> +
> + len = json_object_array_length(cjson_mr_key);
> + js_mr_key->actions = rte_malloc(NULL, sizeof(struct
> cpfl_flow_js_mr_key_action) * len, 0);
> + if (!js_mr_key->actions) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + js_mr_key->actions_size = len;
> + for (i = 0; i < len; i++) {
> + json_object *object, *cjson_mr_key_data;
> + const char *type;
> + enum rte_flow_action_type act_type;
> +
> + object = json_object_array_get_idx(cjson_mr_key, i);
> + /* mr->key->actions->type */
> + type = cpfl_json_object_to_string(object, "type");
> + if (!type) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'type'.");
> + goto err;
> + }
> + act_type = cpfl_get_action_type_by_str(type);
> + if (act_type == RTE_FLOW_ACTION_TYPE_VOID)
> + goto err;
> + js_mr_key->actions[i].type = act_type;
> + /* mr->key->actions->data */
> + cjson_mr_key_data = json_object_object_get(object, "data");
> + if (js_mr_key->actions[i].type ==
> RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP) {
> + json_object *cjson_mr_key_proto;
> + int proto_size, j;
> + struct cpfl_flow_js_mr_key_action_vxlan_encap
> *encap;
> +
> + cjson_mr_key_proto =
> json_object_object_get(cjson_mr_key_data, "protocols");
> + encap = &js_mr_key->actions[i].encap;
> + if (!cjson_mr_key_proto) {
> + encap->proto_size = 0;
> + continue;
> + }
> + proto_size =
> json_object_array_length(cjson_mr_key_proto);
> + encap->proto_size = proto_size;
> + for (j = 0; j < proto_size; j++) {
> + const char *s;
> + json_object *subobject;
> + enum rte_flow_item_type proto_type;
> +
> + subobject =
> json_object_array_get_idx(cjson_mr_key_proto, j);
> + s = json_object_get_string(subobject);
> + proto_type = cpfl_get_item_type_by_str(s);
> + if (proto_type ==
> RTE_FLOW_ITEM_TYPE_VOID) {
> + PMD_DRV_LOG(ERR, "parse
> VXLAN_ENCAP failed.");
> + goto err;
> + }
> + encap->protocols[j] = proto_type;
> + }
> + } else if (js_mr_key->actions[i].type !=
> RTE_FLOW_ACTION_TYPE_VXLAN_DECAP) {
> + PMD_DRV_LOG(ERR, "not support this type: %d.",
> js_mr_key->actions[i].type);
> + goto err;
> + }
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(js_mr_key->actions);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_mr_layout(json_object *cjson_layout, struct
> cpfl_flow_js_mr_action_mod *js_mod)
> +{
> + int len, i;
> +
> + len = json_object_array_length(cjson_layout);
> + js_mod->layout_size = len;
> + if (len == 0)
> + return 0;
> + js_mod->layout = rte_malloc(NULL, sizeof(struct
> cpfl_flow_js_mr_layout) * len, 0);
> + if (!js_mod->layout) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> +
> + for (i = 0; i < len; i++) {
> + json_object *object;
> + int index = 0, size = 0, offset = 0, ret;
> + const char *hint;
> +
> + object = json_object_array_get_idx(cjson_layout, i);
> + ret = cpfl_json_object_to_int(object, "index", &index);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'index'.");
> + goto err;
> + }
> + js_mod->layout[i].index = index;
> + ret = cpfl_json_object_to_int(object, "size", &size);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'size'.");
> + goto err;
> + }
> + js_mod->layout[i].size = size;
> + ret = cpfl_json_object_to_int(object, "offset", &offset);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'offset'.");
> + goto err;
> + }
> + js_mod->layout[i].offset = offset;
> + hint = cpfl_json_object_to_string(object, "hint");
> + if (!hint) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'hint'.");
> + goto err;
> + }
> + memcpy(js_mod->layout[i].hint, hint, strlen(hint));
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(js_mod->layout);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_flow_js_mr_action(json_object *cjson_mr_act, struct
> cpfl_flow_js_mr_action *js_mr_act)
> +{
> + json_object *cjson_mr_action_data;
> + const char *type;
> +
> + /* mr->action->type */
> + type = cpfl_json_object_to_string(cjson_mr_act, "type");
> + if (!type) {
> + PMD_DRV_LOG(ERR, "Can not parse string 'type'.");
> + return -EINVAL;
> + }
> + /* mr->action->data */
> + cjson_mr_action_data = json_object_object_get(cjson_mr_act, "data");
> + if (strcmp(type, "mod") == 0) {
> + json_object *layout;
> + uint16_t profile = 0;
> + int ret;
> +
> + js_mr_act->type = CPFL_JS_MR_ACTION_TYPE_MOD;
> + ret = cpfl_json_object_to_uint16(cjson_mr_action_data,
> "profile", &profile);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse 'profile'.");
> + return -EINVAL;
> + }
> + js_mr_act->mod.prof = profile;
> + layout = json_object_object_get(cjson_mr_action_data,
> "layout");
> + ret = cpfl_flow_js_mr_layout(layout, &js_mr_act->mod);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "Can not parse layout.");
> + return ret;
> + }
> + } else {
> + PMD_DRV_LOG(ERR, "not support this type: %s.", type);
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_flow_js_mod_rule(json_object *json_root, struct cpfl_flow_js_parser
> *parser)
> +{
> + json_object *cjson_mr;
> + int i, len;
> +
> + cjson_mr = json_object_object_get(json_root, "modifications");
> + if (!cjson_mr) {
> + PMD_DRV_LOG(INFO, "The modifications is optional.");
> + return 0;
> + }
> + len = json_object_array_length(cjson_mr);
> + parser->mr_size = len;
> + if (len == 0)
> + return 0;
> + parser->modifications = rte_malloc(NULL, sizeof(struct cpfl_flow_js_mr)
> * len, 0);
> + if (!parser->modifications) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory.");
> + return -ENOMEM;
> + }
> + for (i = 0; i < len; i++) {
> + int ret;
> + json_object *object, *cjson_mr_key, *cjson_mr_action,
> *cjson_mr_key_action;
> +
> + object = json_object_array_get_idx(cjson_mr, i);
> + /* mr->key */
> + cjson_mr_key = json_object_object_get(object, "key");
> + /* mr->key->actions */
> + cjson_mr_key_action = json_object_object_get(cjson_mr_key,
> "actions");
> + ret = cpfl_flow_js_mr_key(cjson_mr_key_action, &parser-
> >modifications[i].key);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "parse mr_key failed.");
> + goto err;
> + }
> + /* mr->action */
> + cjson_mr_action = json_object_object_get(object, "action");
> + ret = cpfl_flow_js_mr_action(cjson_mr_action, &parser-
> >modifications[i].action);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "parse mr_action failed.");
> + goto err;
> + }
> + }
> +
> + return 0;
> +
> +err:
> + rte_free(parser->modifications);
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_parser_init(json_object *json_root, struct cpfl_flow_js_parser *parser)
> +{
> + int ret = 0;
> +
> + ret = cpfl_flow_js_pattern_rule(json_root, parser);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "parse pattern_rule failed.");
> + return ret;
> + }
> + ret = cpfl_flow_js_mod_rule(json_root, parser);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "parse mod_rule failed.");
> + return ret;
> + }
> +
> + return 0;
> +}
> +
> +int
> +cpfl_parser_create(struct cpfl_flow_js_parser **flow_parser, const char
> *filename)
> +{
> + struct cpfl_flow_js_parser *parser;
> + json_object *root;
> + int ret;
> +
> + parser = rte_zmalloc("flow_parser", sizeof(struct cpfl_flow_js_parser),
> 0);
> + if (!parser) {
> + PMD_DRV_LOG(ERR, "Not enough memory to create flow
> parser.");
> + return -ENOMEM;
> + }
> + root = json_object_from_file(filename);
> + if (!root) {
> + PMD_DRV_LOG(ERR, "Can not load JSON file: %s.", filename);
> + rte_free(parser);
> + return -EINVAL;
[Liu, Mingxia] better to use goto free_parser; ?
> + }
> + ret = cpfl_parser_init(root, parser);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "parser init failed.");
> + goto free_parser;
[Liu, Mingxia] If failed, need to json_object_put(root) ?
> + }
> + *flow_parser = parser;
> + ret = json_object_put(root);
> + if (ret != 1) {
> + PMD_DRV_LOG(ERR, "Free json_object failed.");
> + return -EINVAL;
> + }
> +
> + return 0;
> +free_parser:
> + rte_free(parser);
> + return -EINVAL;
> +}
> +
> +static void
> +cpfl_parser_free_pr_action(struct cpfl_flow_js_pr_action *pr_act)
> +{
> + if (pr_act->type == CPFL_JS_PR_ACTION_TYPE_SEM)
> + rte_free(pr_act->sem.fv);
> +}
> +
> +int
> +cpfl_parser_destroy(struct cpfl_flow_js_parser *parser)
> +{
> + int i, j;
> +
> + if (!parser)
> + return 0;
> +
> + for (i = 0; i < parser->pr_size; i++) {
> + struct cpfl_flow_js_pr *pattern = &parser->patterns[i];
> +
> + if (!pattern)
> + return -EINVAL;
> + for (j = 0; j < pattern->key.proto_size; j++)
> + rte_free(pattern->key.protocols[j].fields);
> + rte_free(pattern->key.protocols);
> + rte_free(pattern->key.attributes);
> +
> + for (j = 0; j < pattern->actions_size; j++) {
> + struct cpfl_flow_js_pr_action *pr_act;
> +
> + pr_act = &pattern->actions[j];
> + cpfl_parser_free_pr_action(pr_act);
> + }
> + rte_free(pattern->actions);
> + }
> + rte_free(parser->patterns);
> + for (i = 0; i < parser->mr_size; i++) {
> + struct cpfl_flow_js_mr *mr = &parser->modifications[i];
> +
> + if (!mr)
> + return -EINVAL;
> + rte_free(mr->key.actions);
> + rte_free(mr->action.mod.layout);
> + }
> + rte_free(parser->modifications);
> + rte_free(parser);
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_get_items_length(const struct rte_flow_item *items)
> +{
> + int length = 0;
> + const struct rte_flow_item *item = items;
> +
> + while ((item + length++)->type != RTE_FLOW_ITEM_TYPE_END)
> + continue;
> + return length;
> +}
> +
> +static int
> +cpfl_get_actions_length(const struct rte_flow_action *actions)
> +{
> + int length = 0;
> + const struct rte_flow_action *action = actions;
> +
> + while ((action + length++)->type != RTE_FLOW_ACTION_TYPE_END)
> + continue;
> + return length;
> +}
> +
> +static int
> +cpfl_parse_fv_protocol(struct cpfl_flow_js_fv *js_fv, const struct rte_flow_item
> *items,
> + uint16_t offset, uint8_t *fv)
> +{
> + uint16_t v_layer, v_offset, v_mask;
> + enum rte_flow_item_type v_header;
> + int j, layer, length;
> + uint16_t temp_fv;
> +
> + length = cpfl_get_items_length(items);
> + v_layer = js_fv->proto.layer;
> + v_header = js_fv->proto.header;
> + v_offset = js_fv->proto.offset;
> + v_mask = js_fv->proto.mask;
> + layer = 0;
> + for (j = 0; j < length - 1; j++) {
> + if (items[j].type == v_header) {
> + if (layer == v_layer) {
> + /* copy out 16 bits from offset */
> + const uint8_t *pointer;
> +
> + pointer = &(((const uint8_t
> *)(items[j].spec))[v_offset]);
> + temp_fv = ntohs((*((const uint16_t *)pointer))
> & v_mask);
> + fv[2 * offset] = (uint8_t)((temp_fv & 0xff00) >>
> 8);
[Liu, Mingxia] ((temp_fv & 0xff00) >> 8) can be simplified to temp_fv >> 8, there are other places that has the same issue.
> + fv[2 * offset + 1] = (uint8_t)(temp_fv &
> 0x00ff);
> + break;
> + }
> + layer++;
> + } /* TODO: more type... */
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_parse_fieldvectors(struct cpfl_itf *itf, struct cpfl_flow_js_fv *js_fvs, int
> size,
> + uint8_t *fv, const struct rte_flow_item *items)
> +{
> + int i, ret;
> +
> + for (i = 0; i < size; i++) {
> + uint16_t offset, temp_fv, value_int;
> + enum cpfl_flow_js_fv_type type;
> + struct cpfl_flow_js_fv *js_fv;
> +
> + js_fv = &js_fvs[i];
> + offset = js_fv->offset;
> + type = js_fv->type;
> + if (type == CPFL_FV_TYPE_IMMEDIATE) {
> + value_int = js_fv->immediate;
> + temp_fv = (value_int << 8) & 0xff00;
> + fv[2 * offset] = (uint8_t)((temp_fv & 0xff00) >> 8);
> + fv[2 * offset + 1] = (uint8_t)(temp_fv & 0x00ff);
> + } else if (type == CPFL_FV_TYPE_METADATA) {
> + uint16_t type, v_offset, mask;
> +
> + type = js_fv->meta.type;
> + v_offset = js_fv->meta.offset;
> + mask = js_fv->meta.mask;
> + temp_fv = cpfl_metadata_read16(&itf->adapter->meta,
> type, v_offset) & mask;
> + fv[2 * offset] = (uint8_t)(temp_fv & 0x00ff);
> + fv[2 * offset + 1] = (uint8_t)((temp_fv & 0xff00) >> 8);
> + } else if (type == CPFL_FV_TYPE_PROTOCOL) {
> + ret = cpfl_parse_fv_protocol(js_fv, items, offset, fv);
> + if (ret)
> + return ret;
> + } else {
> + PMD_DRV_LOG(DEBUG, "not support this type: %d.",
> type);
> + return -EINVAL;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_parse_pr_actions(struct cpfl_itf *itf,
> + struct cpfl_flow_js_pr_action *actions,
> + int size,
> + const struct rte_flow_item *items,
> + const struct rte_flow_attr *attr,
> + struct cpfl_flow_pr_action *pr_action)
> +{
> + int i, ret;
> +
> + for (i = 0; i < size; i++) {
> + struct cpfl_flow_js_pr_action *pr_act;
> + enum cpfl_flow_pr_action_type type;
> +
> + pr_act = &actions[i];
> + /* pr->actions->type */
> + type = pr_act->type;
> + /* pr->actions->data */
> + if (attr->group % 10 == 1 && type ==
> CPFL_JS_PR_ACTION_TYPE_SEM) {
> + struct cpfl_flow_js_pr_action_sem *sem = &pr_act-
> >sem;
> +
> + pr_action->type = CPFL_JS_PR_ACTION_TYPE_SEM;
> + pr_action->sem.prof = sem->prof;
> + pr_action->sem.subprof = sem->subprof;
> + pr_action->sem.keysize = sem->keysize;
> + memset(pr_action->sem.cpfl_flow_pr_fv, 0,
> + sizeof(pr_action->sem.cpfl_flow_pr_fv));
> + ret = cpfl_parse_fieldvectors(itf, sem->fv, sem->fv_size,
> + pr_action-
> >sem.cpfl_flow_pr_fv, items);
> + return ret;
> + } else if (attr->group > 4 || attr->group == 0) {
> + return -EPERM;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_str2mac(const char *mask, uint8_t *addr_bytes)
> +{
> + int i, size, j;
> + uint8_t n;
> +
> + size = strlen(mask);
> + n = 0;
> + j = 0;
> + for (i = 0; i < size; i++) {
> + char ch = mask[i];
> +
> + if (ch == ':') {
> + if (j >= RTE_ETHER_ADDR_LEN)
> + return -EINVAL;
> + addr_bytes[j++] = n;
> + n = 0;
> + } else if (ch >= 'a' && ch <= 'f') {
> + n = n * 16 + ch - 'a' + 10;
> + } else if (ch >= 'A' && ch <= 'F') {
> + n = n * 16 + ch - 'A' + 10;
> + } else if (ch >= '0' && ch <= '9') {
> + n = n * 16 + ch - '0';
> + } else {
> + return -EINVAL;
> + }
> + }
> + if (j < RTE_ETHER_ADDR_LEN)
> + addr_bytes[j++] = n;
> +
> + if (j != RTE_ETHER_ADDR_LEN)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_eth_mask(const char *mask, const uint8_t
> addr_bytes[RTE_ETHER_ADDR_LEN])
> +{
> + int i, ret;
> + uint8_t mask_bytes[RTE_ETHER_ADDR_LEN] = { 0 };
> +
> + ret = cpfl_str2mac(mask, mask_bytes);
> + if (ret < 0) {
> + PMD_DRV_LOG(ERR, "translate mac address from string to
> uint8_t[] failed.");
> + return -EINVAL;
> + }
> + /* validate eth mask addr if match */
> + for (i = 0; i < RTE_ETHER_ADDR_LEN; i++) {
> + if (mask_bytes[i] != addr_bytes[i])
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_ipv4_mask(const char *mask, rte_be32_t addr)
> +{
> + uint32_t out_addr;
> +
> + /* 0: success; -EINVAL: invalid; -ENOTSUP: fail */
> + int ret = inet_pton(AF_INET, mask, &out_addr);
> +
> + if (ret < 0)
> + return -EINVAL;
> + /* validate ipv4 mask addr if match */
> + if (out_addr != addr)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_eth(struct cpfl_flow_js_pr_key_proto *proto, const struct
> rte_flow_item_eth *eth_mask)
> +{
> + int field_size, j;
> + int flag_dst_addr, flag_src_addr, flag_ether_type;
> + struct cpfl_flow_js_pr_key_proto_field *field;
> +
> + if (!proto)
> + return 0;
> + field_size = proto->fields_size;
> + if (field_size != 0 && !eth_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && eth_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && !eth_mask)
> + return 0;
> +
> + flag_dst_addr = false;
> + flag_src_addr = false;
> + flag_ether_type = false;
> + for (j = 0; j < field_size; j++) {
> + const char *name, *s_mask;
> +
> + field = &proto->fields[j];
> + /* match: rte_flow_item_eth.dst, more see Field Mapping
> + */
> + name = field->name;
> + /* match: rte_flow_item->mask */
> + if (strcmp(name, "src_addr") == 0) {
> + s_mask = field->mask;
> + if (cpfl_check_eth_mask(s_mask, eth_mask-
> >src.addr_bytes) < 0)
> + return -EINVAL;
> + flag_src_addr = true;
> + } else if (strcmp(name, "dst_addr") == 0) {
> + s_mask = field->mask;
> + if (cpfl_check_eth_mask(s_mask, eth_mask-
> >dst.addr_bytes) < 0)
> + return -EINVAL;
> + flag_dst_addr = true;
> + } else if (strcmp(name, "ether_type") == 0) {
> + uint16_t mask = (uint16_t)field->mask_32b;
> +
> + if (mask != eth_mask->type)
> + return -EINVAL;
> + flag_ether_type = true;
> + } else {
> + /* TODO: more type... */
> + PMD_DRV_LOG(ERR, "not support this name.");
> + return -EINVAL;
> + }
> + }
> + if (!flag_src_addr) {
> + if (strcmp((const char *)eth_mask->src.addr_bytes,
> "\x00\x00\x00\x00\x00\x00") != 0)
> + return -EINVAL;
> + }
> + if (!flag_dst_addr) {
> + if (strcmp((const char *)eth_mask->dst.addr_bytes,
> "\x00\x00\x00\x00\x00\x00") != 0)
> + return -EINVAL;
> + }
> + if (!flag_ether_type) {
> + if (eth_mask->hdr.ether_type != (rte_be16_t)0)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_ipv4(struct cpfl_flow_js_pr_key_proto *proto, const struct
> rte_flow_item_ipv4 *ipv4_mask)
> +{
> + int field_size, j;
> + int flag_next_proto_id, flag_src_addr, flag_dst_addr;
> + struct cpfl_flow_js_pr_key_proto_field *field;
> +
> + if (!proto)
> + return 0;
> +
> + field_size = proto->fields_size;
> + if (field_size != 0 && !ipv4_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && ipv4_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && !ipv4_mask)
> + return 0;
> +
> + flag_dst_addr = false;
> + flag_src_addr = false;
> + flag_next_proto_id = false;
> + for (j = 0; j < field_size; j++) {
> + const char *name;
> +
> + field = &proto->fields[j];
> + name = field->name;
> + if (strcmp(name, "src_addr") == 0) {
> + /* match: rte_flow_item->mask */
> + const char *mask;
> +
> + mask = field->mask;
> + if (cpfl_check_ipv4_mask(mask, ipv4_mask-
> >hdr.src_addr) < 0)
> + return -EINVAL;
> + flag_src_addr = true;
> + } else if (strcmp(name, "dst_addr") == 0) {
> + const char *mask;
> +
> + mask = field->mask;
> + if (cpfl_check_ipv4_mask(mask, ipv4_mask-
> >hdr.dst_addr) < 0)
> + return -EINVAL;
> + flag_dst_addr = true;
> + } else if (strcmp(name, "next_proto_id") == 0) {
> + uint8_t mask;
> +
> + mask = (uint8_t)field->mask_32b;
> + if (mask != ipv4_mask->hdr.next_proto_id)
> + return -EINVAL;
> + flag_next_proto_id = true;
> + } else {
> + PMD_DRV_LOG(ERR, "not support this name.");
> + return -EINVAL;
> + }
> + }
> + if (!flag_src_addr) {
> + if (ipv4_mask->hdr.src_addr != (rte_be32_t)0)
> + return -EINVAL;
> + }
> + if (!flag_dst_addr) {
> + if (ipv4_mask->hdr.dst_addr != (rte_be32_t)0)
> + return -EINVAL;
> + }
> + if (!flag_next_proto_id) {
> + if (ipv4_mask->hdr.next_proto_id != (uint8_t)0)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_tcp(struct cpfl_flow_js_pr_key_proto *proto, const struct
> rte_flow_item_tcp *tcp_mask)
> +{
> + int field_size, j;
> + int flag_src_port, flag_dst_port;
> + struct cpfl_flow_js_pr_key_proto_field *field;
> +
> + if (!proto)
> + return 0;
> +
> + field_size = proto->fields_size;
> + if (field_size != 0 && !tcp_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && tcp_mask)
> + return -EINVAL;
> +
> + if (field_size == 0 && !tcp_mask)
> + return 0;
> +
> + flag_src_port = false;
> + flag_dst_port = false;
> + for (j = 0; j < field_size; j++) {
> + const char *name;
> + uint16_t mask;
> +
> + field = &proto->fields[j];
> + /* match: rte_flow_item_eth.dst */
> + name = field->name;
> + /* match: rte_flow_item->mask */
> + mask = (uint16_t)field->mask_32b;
> + if (strcmp(name, "src_port") == 0) {
> + if (tcp_mask->hdr.src_port != mask)
> + return -EINVAL;
> + flag_src_port = true;
> + } else if (strcmp(name, "dst_port") == 0) {
> + if (tcp_mask->hdr.dst_port != mask)
> + return -EINVAL;
> + flag_dst_port = true;
> + } else {
> + PMD_DRV_LOG(ERR, "not support this name.");
> + return -EINVAL;
> + }
> + }
> + if (!flag_src_port) {
> + if (tcp_mask->hdr.src_port != (rte_be16_t)0)
> + return -EINVAL;
> + }
> + if (!flag_dst_port) {
> + if (tcp_mask->hdr.dst_port != (rte_be16_t)0)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_udp(struct cpfl_flow_js_pr_key_proto *proto, const struct
> rte_flow_item_udp *udp_mask)
> +{
> + int field_size, j;
> + bool flag_src_port, flag_dst_port;
> + struct cpfl_flow_js_pr_key_proto_field *field;
> +
> + if (!proto)
> + return 0;
> + field_size = proto->fields_size;
> + if (field_size != 0 && !udp_mask)
> + return -EINVAL;
> + if (field_size == 0 && udp_mask)
> + return -EINVAL;
> + if (field_size == 0 && !udp_mask)
> + return 0;
> + flag_src_port = false;
> + flag_dst_port = false;
> + for (j = 0; j < field_size; j++) {
> + const char *name;
> + uint16_t mask;
> +
> + field = &proto->fields[j];
> + /* match: rte_flow_item_eth.dst */
> + name = field->name; /* match: rte_flow_item->mask */
> + mask = (uint16_t)field->mask_32b;
> + if (strcmp(name, "src_port") == 0) {
> + if (udp_mask->hdr.src_port != mask)
> + return -EINVAL;
> + flag_src_port = true;
> + } else if (strcmp(name, "dst_port") == 0) {
> + if (udp_mask->hdr.dst_port != mask)
> + return -EINVAL;
> + flag_dst_port = true;
> + } else {
> + PMD_DRV_LOG(ERR, "not support this name.");
> + return -EINVAL;
> + }
> + }
> + if (!flag_src_port) {
> + if (udp_mask->hdr.src_port != (rte_be16_t)0)
> + return -EINVAL;
> + }
> + if (!flag_dst_port) {
> + if (udp_mask->hdr.dst_port != (rte_be16_t)0)
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_vxlan(struct cpfl_flow_js_pr_key_proto *proto,
> + const struct rte_flow_item_vxlan *vxlan_mask)
> +{
> + int field_size, j;
> + struct cpfl_flow_js_pr_key_proto_field *field;
> +
> + if (!proto)
> + return 0;
> + field_size = proto->fields_size;
> + if (field_size != 0 && !vxlan_mask)
> + return -EINVAL;
> + if (field_size == 0 && vxlan_mask)
> + return -EINVAL;
> + if (field_size == 0 && !vxlan_mask)
> + return 0;
> + for (j = 0; j < field_size; j++) {
> + const char *name;
> + int64_t mask;
> +
> + field = &proto->fields[j];
> + name = field->name;
> + /* match: rte_flow_item->mask */
> + mask = (int64_t)field->mask_32b;
> + if (strcmp(name, "vx_vni") == 0) {
> + if ((int64_t)RTE_BE32(vxlan_mask->hdr.vx_vni) !=
> mask)
> + return -EINVAL;
> + } else {
> + PMD_DRV_LOG(ERR, "not support this name.");
> + return -EINVAL;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_icmp(struct cpfl_flow_js_pr_key_proto *proto, const struct
> rte_flow_item_icmp *icmp_mask)
> +{
> + int field_size;
> +
> + if (!proto)
> + return 0;
> + field_size = proto->fields_size;
> + if (field_size != 0 && !icmp_mask)
> + return -EINVAL;
> + if (field_size == 0 && icmp_mask)
> + return -EINVAL;
> + if (field_size == 0 && !icmp_mask)
> + return 0;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_pattern_key_proto(struct cpfl_flow_js_pr_key_proto *protocols,
> + int proto_size,
> + const struct rte_flow_item *items)
> +{
> + int i, length;
> + int j = 0;
> +
> + length = cpfl_get_items_length(items);
> + if (proto_size > length - 1)
> + return -EINVAL;
> + for (i = 0; i < proto_size; i++) {
> + struct cpfl_flow_js_pr_key_proto *key_proto;
> + enum rte_flow_item_type type;
> +
> + key_proto = &protocols[i];
> + /* pr->key->proto->type */
> + type = key_proto->type;
> + /* pr->key->proto->fields */
> + switch (type) {
> + case RTE_FLOW_ITEM_TYPE_ETH:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_ETH) {
> + const struct rte_flow_item_eth *eth_mask;
> + int ret;
> +
> + eth_mask = (const struct rte_flow_item_eth
> *)items[i].mask;
> + ret = cpfl_check_eth(key_proto, eth_mask);
> + if (ret < 0)
> + return ret;
> + } else {
> + return -EINVAL;
> + }
> + break;
> + case RTE_FLOW_ITEM_TYPE_IPV4:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_IPV4) {
> + const struct rte_flow_item_ipv4 *ipv4_mask;
> + int ret;
> +
> + ipv4_mask = (const struct rte_flow_item_ipv4
> *)items[i].mask;
> + ret = cpfl_check_ipv4(key_proto, ipv4_mask);
> + if (ret < 0)
> + return ret;
> + } else {
> + return -EINVAL;
> + }
> + break;
> + case RTE_FLOW_ITEM_TYPE_TCP:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_TCP) {
> + const struct rte_flow_item_tcp *tcp_mask;
> + int ret;
> +
> + tcp_mask = (const struct rte_flow_item_tcp
> *)items[i].mask;
> + ret = cpfl_check_tcp(key_proto, tcp_mask);
> + if (ret < 0)
> + return ret;
> + } else {
> + return -EINVAL;
> + }
> + break;
> + case RTE_FLOW_ITEM_TYPE_UDP:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_UDP) {
> + const struct rte_flow_item_udp *udp_mask;
> + int ret;
> +
> + udp_mask = (const struct rte_flow_item_udp
> *)items[i].mask;
> + ret = cpfl_check_udp(key_proto, udp_mask);
> + if (ret < 0)
> + return ret;
> + } else {
> + return -EINVAL;
> + }
> + break;
> + case RTE_FLOW_ITEM_TYPE_VXLAN:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_VXLAN) {
> + const struct rte_flow_item_vxlan *vxlan_mask;
> + int ret;
> +
> + vxlan_mask = (const struct
> rte_flow_item_vxlan *)items[i].mask;
> + ret = cpfl_check_vxlan(key_proto, vxlan_mask);
> + if (ret < 0)
> + return ret;
> + } else {
> + return -EINVAL;
> + }
> + break;
> + case RTE_FLOW_ITEM_TYPE_ICMP:
> + if (items[j++].type == RTE_FLOW_ITEM_TYPE_ICMP) {
> + const struct rte_flow_item_icmp *icmp_mask;
> + int ret;
> +
> + icmp_mask = (const struct rte_flow_item_icmp
> *)items[i].mask;
> + ret = cpfl_check_icmp(key_proto, icmp_mask);
> + if (ret < 0)
> + return ret;
> +
> + } else {
> + return -EINVAL;
> + }
> + break;
> + default:
> + PMD_DRV_LOG(ERR, "Not support this type: %d.",
> type);
> + return -EPERM;
> + }
> + }
> + if (items[j].type != RTE_FLOW_ITEM_TYPE_END)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_pattern_key_attr(struct cpfl_flow_js_pr_key_attr *key_attr,
> + const struct rte_flow_attr *attr)
> +{
> + /* match: struct rte_flow_attr(ingress,egress) */
> + if (key_attr->ingress != attr->ingress) {
> + PMD_DRV_LOG(DEBUG, "ingress not match.");
> + return -EINVAL;
> + }
> + if (key_attr->egress != attr->egress) {
> + PMD_DRV_LOG(DEBUG, "egress not match.");
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_check_pattern_key(struct cpfl_flow_js_pr *pattern,
> + const struct rte_flow_item *items,
> + const struct rte_flow_attr *attr)
> +{
> + int ret;
> +
> + /* pr->key */
> + /* pr->key->protocols */
> + ret = cpfl_check_pattern_key_proto(pattern->key.protocols,
> + pattern->key.proto_size, items);
> + if (ret < 0)
> + return -EINVAL;
> + /* pr->key->attributes */
> + ret = cpfl_check_pattern_key_attr(pattern->key.attributes, attr);
> + if (ret < 0)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> +/* output: struct cpfl_flow_pr_action* pr_action */
> +static int
> +cpfl_parse_pattern_rules(struct cpfl_itf *itf,
> + struct cpfl_flow_js_parser *parser,
> + const struct rte_flow_item *items,
> + const struct rte_flow_attr *attr,
> + struct cpfl_flow_pr_action *pr_action)
> +{
> + int i, size;
> + struct cpfl_flow_js_pr *pattern;
> +
> + size = parser->pr_size;
> + for (i = 0; i < size; i++) {
> + int ret;
> +
> + pattern = &parser->patterns[i];
> + ret = cpfl_check_pattern_key(pattern, items, attr);
> + if (ret < 0)
> + continue;
> + /* pr->actions */
> + ret = cpfl_parse_pr_actions(itf, pattern->actions, pattern-
> >actions_size,
> + items, attr, pr_action);
> + return ret;
> + }
> +
> + return -EINVAL;
> +}
> +
> +int
> +cpfl_flow_parse_items(struct cpfl_itf *itf,
> + struct cpfl_flow_js_parser *parser,
> + const struct rte_flow_item *items,
> + const struct rte_flow_attr *attr,
> + struct cpfl_flow_pr_action *pr_action)
> +{
> + int ret;
> +
> + /* Pattern Rules */
> + ret = cpfl_parse_pattern_rules(itf, parser, items, attr, pr_action);
> + return ret;
> +}
> +
> +/* modifications rules */
> +static int
> +cpfl_check_actions_vxlan_encap(struct cpfl_flow_mr_key_action_vxlan_encap
> *encap,
> + const struct rte_flow_action *action)
> +{
> + const struct rte_flow_action_vxlan_encap *action_vxlan_encap;
> + struct rte_flow_item *definition;
> + int def_length, i, proto_size;
> +
> + action_vxlan_encap = (const struct rte_flow_action_vxlan_encap
> *)action->conf;
> + definition = action_vxlan_encap->definition;
> + def_length = cpfl_get_items_length(definition);
> + proto_size = encap->proto_size;
> + if (proto_size != def_length - 1) {
> + PMD_DRV_LOG(DEBUG, "protocols not match.");
> + return -EINVAL;
> + }
> + for (i = 0; i < proto_size; i++) {
> + enum rte_flow_item_type proto;
> +
> + proto = encap->protocols[i];
> + if (proto == RTE_FLOW_ITEM_TYPE_VLAN) {
> + if (definition[i].type != RTE_FLOW_ITEM_TYPE_VOID) {
> + PMD_DRV_LOG(DEBUG, "protocols not
> match.");
> + return -EINVAL;
> + }
> + } else if (proto != definition[i].type) {
> + PMD_DRV_LOG(DEBUG, "protocols not match.");
> + return -EINVAL;
> + }
> + }
> +
> + return 0;
> +}
> +
> +/* output: struct cpfl_flow_mr_key_action *mr_key_action */
> +/* check and parse */
> +static int
> +cpfl_parse_mr_key_action(struct cpfl_flow_js_mr_key_action *key_acts, int
> size,
> + const struct rte_flow_action *actions,
> + struct cpfl_flow_mr_key_action *mr_key_action)
> +{
> + int actions_length, i;
> + int j = 0;
> + int ret;
> +
> + actions_length = cpfl_get_actions_length(actions);
> + if (size > actions_length - 1)
> + return -EINVAL;
> + for (i = 0; i < size; i++) {
> + enum rte_flow_action_type type;
> + struct cpfl_flow_js_mr_key_action *key_act;
> +
> + key_act = &key_acts[i];
> + /* mr->key->actions->type */
> + type = key_act->type;
> + /* mr->key->actions->data */
> + /* match: <type> action matches
> RTE_FLOW_ACTION_TYPE_<type> */
> + if (type == RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP) {
> + int proto_size, k;
> + struct cpfl_flow_mr_key_action_vxlan_encap *encap;
> +
> + while (j < actions_length &&
> + actions[j].type !=
> RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP) {
> + j++;
> + }
> + if (j >= actions_length)
> + return -EINVAL;
> + mr_key_action[i].type =
> RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP;
> + mr_key_action[i].encap.action = &actions[j];
> + encap = &mr_key_action[i].encap;
> +
> + proto_size = key_act->encap.proto_size;
> + encap->proto_size = proto_size;
> + for (k = 0; k < proto_size; k++) {
> + enum rte_flow_item_type proto;
> +
> + proto = key_act->encap.protocols[k];
> + encap->protocols[k] = proto;
> + }
> + ret = cpfl_check_actions_vxlan_encap(encap,
> &actions[j]);
> + if (ret < 0)
> + return -EINVAL;
> +
> + j++;
> + } else if (type == RTE_FLOW_ACTION_TYPE_VXLAN_DECAP) {
> + while (j < actions_length &&
> + actions[j].type !=
> RTE_FLOW_ACTION_TYPE_VXLAN_DECAP) {
> + j++;
> + }
> + if (j >= actions_length)
> + return -EINVAL;
> +
> + mr_key_action[i].type =
> RTE_FLOW_ACTION_TYPE_VXLAN_DECAP;
> + j++;
> + } else {
> + PMD_DRV_LOG(ERR, "Not support this type: %d.",
> type);
> + return -EPERM;
> + }
> + }
> +
> + return 0;
> +}
> +
> +/* output: uint8_t *buffer, uint16_t *byte_len */
> +static int
> +cpfl_parse_layout(struct cpfl_flow_js_mr_layout *layouts, int layout_size,
> + struct cpfl_flow_mr_key_action *mr_key_action,
> + uint8_t *buffer, uint16_t *byte_len)
> +{
> + int i;
> + int start = 0;
> +
> + for (i = 0; i < layout_size; i++) {
> + int index, size, offset;
> + const char *hint;
> + const uint8_t *addr;
> + struct cpfl_flow_mr_key_action *temp;
> + struct cpfl_flow_js_mr_layout *layout;
> +
> + layout = &layouts[i];
> + /* index links to the element of the actions array. */
> + index = layout->index;
> + size = layout->size;
> + offset = layout->offset;
> + if (index == -1) {
> + hint = "dummpy";
> + start += size;
> + continue;
> + }
> + hint = layout->hint;
> + addr = NULL;
> + temp = mr_key_action + index;
> + if (temp->type == RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP) {
> + const struct rte_flow_action_vxlan_encap
> *action_vxlan_encap;
> + struct rte_flow_item *definition;
> + int def_length, k;
> +
> + action_vxlan_encap =
> + (const struct rte_flow_action_vxlan_encap *)temp-
> >encap.action->conf;
> + definition = action_vxlan_encap->definition;
> + def_length = cpfl_get_items_length(definition);
> + for (k = 0; k < def_length - 1; k++) {
> + if ((strcmp(hint, "eth") == 0 &&
> + definition[k].type ==
> RTE_FLOW_ITEM_TYPE_ETH) ||
> + (strcmp(hint, "ipv4") == 0 &&
> + definition[k].type ==
> RTE_FLOW_ITEM_TYPE_IPV4) ||
> + (strcmp(hint, "udp") == 0 &&
> + definition[k].type ==
> RTE_FLOW_ITEM_TYPE_UDP) ||
> + (strcmp(hint, "tcp") == 0 &&
> + definition[k].type ==
> RTE_FLOW_ITEM_TYPE_TCP) ||
> + (strcmp(hint, "vxlan") == 0 &&
> + definition[k].type ==
> RTE_FLOW_ITEM_TYPE_VXLAN)) {
> + addr = (const uint8_t
> *)(definition[k].spec);
> + if (start > 255) {
> + *byte_len = 0;
> + PMD_DRV_LOG(ERR, "byte
> length is too long%s",
> + hint);
> + return -EINVAL;
> + }
> + memcpy(buffer + start, addr + offset,
> size);
> + break;
> + } /* TODO: more hint... */
> + }
> + if (k == def_length - 1) {
> + *byte_len = 0;
> + PMD_DRV_LOG(ERR, "can not find
> corresponding hint: %s", hint);
> + return -EINVAL;
> + }
> + } else {
> + *byte_len = 0;
> + PMD_DRV_LOG(ERR, "Not support this type: %d.",
> temp->type);
> + return -EINVAL;
> + }
> + /* else TODO: more type... */
> +
> + start += size;
> + }
> + *byte_len = start;
> +
> + return 0;
> +}
> +
> +static int
> +cpfl_parse_mr_action(struct cpfl_flow_js_mr_action *action,
> + struct cpfl_flow_mr_key_action *mr_key_action,
> + struct cpfl_flow_mr_action *mr_action)
> +{
> + enum cpfl_flow_mr_action_type type;
> +
> + /* mr->action->type */
> + type = action->type;
> + /* mr->action->data */
> + if (type == CPFL_JS_MR_ACTION_TYPE_MOD) {
> + struct cpfl_flow_js_mr_layout *layout;
> +
> + mr_action->type = CPFL_JS_MR_ACTION_TYPE_MOD;
> + mr_action->mod.byte_len = 0;
> + mr_action->mod.prof = action->mod.prof;
> + layout = action->mod.layout;
> + if (layout) {
> + int ret;
> +
> + memset(mr_action->mod.data, 0, sizeof(mr_action-
> >mod.data));
> + ret = cpfl_parse_layout(layout, action-
> >mod.layout_size, mr_key_action,
> + mr_action->mod.data,
> &mr_action->mod.byte_len);
> + if (ret < 0)
> + return -EINVAL;
> + }
> + return 0;
> + }
> + PMD_DRV_LOG(ERR, "Not support this type: %d.", type);
> +
> + return -EINVAL;
> +}
> +
> +static int
> +cpfl_check_mod_key(struct cpfl_flow_js_mr *mr, const struct rte_flow_action
> *actions,
> + struct cpfl_flow_mr_key_action *mr_key_action)
> +{
> + int key_action_size;
> +
> + /* mr->key->actions */
> + key_action_size = mr->key.actions_size;
> + return cpfl_parse_mr_key_action(mr->key.actions, key_action_size,
> actions, mr_key_action);
> +}
> +
> +/* output: struct cpfl_flow_mr_action *mr_action */
> +static int
> +cpfl_parse_mod_rules(struct cpfl_flow_js_parser *parser, const struct
> rte_flow_action *actions,
> + struct cpfl_flow_mr_action *mr_action)
> +{
> + int i, size;
> + struct cpfl_flow_mr_key_action
> mr_key_action[CPFL_MOD_KEY_NUM_MAX] = {0};
> +
> + size = parser->mr_size;
> +
> + for (i = 0; i < size; i++) {
> + int ret;
> + struct cpfl_flow_js_mr *mr;
> +
> + mr = &parser->modifications[i];
> + ret = cpfl_check_mod_key(mr, actions, mr_key_action);
> + if (ret < 0)
> + continue;
> + /* mr->action */
> + ret = cpfl_parse_mr_action(&mr->action, mr_key_action,
> mr_action);
> + if (!ret)
> + return 0;
> + }
> +
> + return -EINVAL;
> +}
> +
> +int
> +cpfl_flow_parse_actions(struct cpfl_flow_js_parser *parser, const struct
> rte_flow_action *actions,
> + struct cpfl_flow_mr_action *mr_action)
> +{
> + /* modifications rules */
> + if (!parser->modifications) {
> + PMD_DRV_LOG(INFO, "The modifications is optional.");
> + return 0;
> + }
> +
> + return cpfl_parse_mod_rules(parser, actions, mr_action);
> +}
> +
> +void
> +cpfl_metadata_write16(struct cpfl_metadata *meta, int type, int offset,
> uint16_t data)
> +{
> + rte_memcpy(&meta->chunks[type].data[offset],
> + &data,
> + sizeof(uint16_t));
> +}
> +
> +void
> +cpfl_metadata_write32(struct cpfl_metadata *meta, int type, int offset,
> uint32_t data)
> +{
> + rte_memcpy(&meta->chunks[type].data[offset],
> + &data,
> + sizeof(uint32_t));
> +}
> +
> +uint16_t
> +cpfl_metadata_read16(struct cpfl_metadata *meta, int type, int offset)
> +{
> + return *((uint16_t *)(&meta->chunks[type].data[offset]));
> +}
> +
> +bool
> +cpfl_metadata_write_port_id(struct cpfl_itf *itf)
> +{
> + uint32_t dev_id;
> + const int type = 0;
> + const int offset = 5;
> +
> + dev_id = cpfl_get_port_id(itf);
> + if (dev_id == CPFL_INVALID_HW_ID) {
> + PMD_DRV_LOG(ERR, "fail to get hw ID\n");
> + return false;
> + }
> + dev_id = dev_id << 3;
> + cpfl_metadata_write16(&itf->adapter->meta, type, offset, dev_id);
[Liu, Mingxia] better to change the type of 'dev_id' to uint16_t?
> +
> + return true;
> +}
> +
> +bool
> +cpfl_metadata_write_targetvsi(struct cpfl_itf *itf)
> +{
> + uint32_t dev_id;
> + const int type = 6;
> + const int offset = 2;
> +
> + dev_id = cpfl_get_vsi_id(itf);
> + if (dev_id == CPFL_INVALID_HW_ID) {
> + PMD_DRV_LOG(ERR, "fail to get hw ID");
> + return false;
> + }
> + dev_id = dev_id << 1;
> + cpfl_metadata_write16(&itf->adapter->meta, type, offset, dev_id);
> +
[Liu, Mingxia] better to change the type of 'dev_id' to uint16_t?
> + return true;
> +}
> +
> +bool
> +cpfl_metadata_write_sourcevsi(struct cpfl_itf *itf)
> +{
> + uint32_t dev_id;
> + const int type = 6;
> + const int offset = 0;
> +
> + dev_id = cpfl_get_vsi_id(itf);
> + if (dev_id == CPFL_INVALID_HW_ID) {
> + PMD_DRV_LOG(ERR, "fail to get hw ID");
> + return false;
> + }
> + cpfl_metadata_write16(&itf->adapter->meta, type, offset, dev_id);
> +
[Liu, Mingxia] better to change the type of 'dev_id' to uint16_t?
> + return true;
> +}
> +
> +void
> +cpfl_metadata_init(struct cpfl_metadata *meta)
> +{
> + int i;
> +
> + for (i = 0; i < CPFL_META_LENGTH; i++)
> + meta->chunks[i].type = i;
> +}
> +
> +bool cpfl_metadata_write_vsi(struct cpfl_itf *itf)
> +{
> + uint32_t dev_id;
> + const int type = 0;
> + const int offset = 24;
> +
> + dev_id = cpfl_get_vsi_id(itf);
> + if (dev_id == CPFL_INVALID_HW_ID) {
> + PMD_DRV_LOG(ERR, "fail to get hw ID");
> + return false;
> + }
> + cpfl_metadata_write16(&itf->adapter->meta, type, offset, dev_id);
> +
[Liu, Mingxia] better to change the type of 'dev_id' to uint16_t?
> + return true;
> +}
next prev parent reply other threads:[~2023-09-08 6:27 UTC|newest]
Thread overview: 128+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-12 7:55 [PATCH v1 0/5] add rte flow support for cpfl Yuying Zhang
2023-08-12 7:55 ` [PATCH v1 1/5] net/cpfl: setup rte flow skeleton Yuying Zhang
2023-08-25 3:55 ` Xing, Beilei
2023-08-12 7:55 ` [PATCH v1 2/5] common/idpf/base: refine idpf ctlq message structure Yuying Zhang
2023-08-25 5:55 ` Xing, Beilei
2023-08-12 7:55 ` [PATCH v1 3/5] net/cpfl: add cpfl control queue message handle Yuying Zhang
2023-08-25 6:23 ` Xing, Beilei
2023-08-12 7:55 ` [PATCH v1 4/5] net/cpfl: add fxp rule module Yuying Zhang
2023-08-25 7:35 ` Xing, Beilei
2023-08-25 8:42 ` Xing, Beilei
2023-08-12 7:55 ` [PATCH v1 5/5] net/cpfl: add fxp flow engine Yuying Zhang
2023-08-25 9:15 ` Xing, Beilei
2023-09-01 11:31 ` [PATCH v2 0/8] add rte flow support for cpfl Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 1/8] net/cpfl: parse flow parser file in devargs Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 2/8] net/cpfl: add flow json parser Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 3/8] net/cpfl: add FXP low level implementation Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 4/8] net/cpfl: setup ctrl path Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 5/8] net/cpfl: set up rte flow skeleton Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 6/8] net/cpfl: add fxp rule module Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 7/8] net/cpfl: add fxp flow engine Yuying Zhang
2023-09-01 11:31 ` [PATCH v2 8/8] net/cpfl: add flow support for representor Yuying Zhang
2023-09-06 9:33 ` [PATCH v3 0/9] add rte flow support for cpfl Wenjing Qiao
2023-08-15 16:50 ` [PATCH v4 " Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 1/9] net/cpfl: add json parser for rte flow pattern rules Zhang, Yuying
2023-09-15 15:11 ` Stephen Hemminger
2023-08-15 16:50 ` [PATCH v4 2/9] net/cpfl: add mod rule parser support for rte flow Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 3/9] net/cpfl: set up rte flow skeleton Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 4/9] net/cpfl: add FXP low level implementation Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 5/9] net/cpfl: add fxp rule module Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 6/9] net/cpfl: add fxp flow engine Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 7/9] net/cpfl: add flow support for representor Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 8/9] app/test-pmd: refine encap content Zhang, Yuying
2023-08-15 16:50 ` [PATCH v4 9/9] net/cpfl: fix incorrect status calculation Zhang, Yuying
2023-09-06 9:33 ` [PATCH v3 1/9] net/cpfl: parse flow parser file in devargs Wenjing Qiao
2023-09-11 0:48 ` Wu, Jingjing
2023-09-06 9:34 ` [PATCH v3 2/9] net/cpfl: add flow json parser Wenjing Qiao
2023-09-08 6:26 ` Liu, Mingxia [this message]
2023-09-11 6:24 ` Wu, Jingjing
2023-09-06 9:34 ` [PATCH v3 3/9] net/cpfl: add FXP low level implementation Wenjing Qiao
2023-09-06 9:34 ` [PATCH v3 4/9] net/cpfl: setup ctrl path Wenjing Qiao
2023-09-11 6:30 ` Liu, Mingxia
2023-09-11 6:36 ` Wu, Jingjing
2023-09-06 9:34 ` [PATCH v3 5/9] net/cpfl: set up rte flow skeleton Wenjing Qiao
2023-09-06 9:34 ` [PATCH v3 6/9] net/cpfl: add fxp rule module Wenjing Qiao
2023-09-12 7:40 ` FW: " Liu, Mingxia
2023-09-06 9:34 ` [PATCH v3 7/9] net/cpfl: add fxp flow engine Wenjing Qiao
2023-09-06 9:34 ` [PATCH v3 8/9] net/cpfl: add flow support for representor Wenjing Qiao
2023-09-06 9:34 ` [PATCH v3 9/9] app/test-pmd: refine encap content Wenjing Qiao
2023-09-15 10:00 ` [PATCH v5 0/9] add rte flow support for cpfl Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 0/8] " Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 1/8] net/cpfl: add json parser for rte flow pattern rules Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 2/8] net/cpfl: add mod rule parser support for rte flow Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 3/8] net/cpfl: set up rte flow skeleton Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 4/8] net/cpfl: set up control path Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 5/8] net/cpfl: add FXP low level implementation Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 6/8] net/cpfl: add fxp rule module Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 7/8] net/cpfl: add fxp flow engine Zhang, Yuying
2023-08-22 1:02 ` [PATCH v6 8/8] net/cpfl: add flow support for representor Zhang, Yuying
2023-09-26 18:16 ` [PATCH v7 0/8] add rte flow support for cpfl yuying.zhang
2023-09-26 18:16 ` [PATCH v7 1/8] net/cpfl: add json parser for rte flow pattern rules yuying.zhang
2023-09-26 19:03 ` Stephen Hemminger
2023-09-27 1:21 ` Zhang, Qi Z
2023-09-26 18:16 ` [PATCH v7 2/8] net/cpfl: build action mapping rules from JSON yuying.zhang
2023-09-26 18:16 ` [PATCH v7 3/8] net/cpfl: set up rte flow skeleton yuying.zhang
2023-09-26 18:16 ` [PATCH v7 4/8] net/cpfl: set up control path yuying.zhang
2023-09-26 18:17 ` [PATCH v7 5/8] net/cpfl: add FXP low level implementation yuying.zhang
2023-09-26 18:17 ` [PATCH v7 6/8] net/cpfl: add fxp rule module yuying.zhang
2023-09-28 3:29 ` Zhang, Qi Z
2023-09-26 18:17 ` [PATCH v7 7/8] net/cpfl: add fxp flow engine yuying.zhang
2023-09-26 18:17 ` [PATCH v7 8/8] net/cpfl: add flow support for representor yuying.zhang
2023-09-27 12:54 ` [PATCH v8 0/9] add rte flow support for cpfl yuying.zhang
2023-09-27 12:54 ` [PATCH v8 1/9] net/cpfl: add json parser for rte flow pattern rules yuying.zhang
2023-09-27 12:54 ` [PATCH v8 2/9] net/cpfl: build action mapping rules from JSON yuying.zhang
2023-09-27 12:54 ` [PATCH v8 3/9] net/cpfl: set up rte flow skeleton yuying.zhang
2023-09-27 12:54 ` [PATCH v8 4/9] net/cpfl: set up control path yuying.zhang
2023-09-27 12:54 ` [PATCH v8 5/9] net/cpfl: add FXP low level implementation yuying.zhang
2023-09-27 12:54 ` [PATCH v8 6/9] net/cpfl: add fxp rule module yuying.zhang
2023-09-27 12:54 ` [PATCH v8 7/9] net/cpfl: add fxp flow engine yuying.zhang
2023-09-27 12:54 ` [PATCH v8 8/9] net/cpfl: add flow support for representor yuying.zhang
2023-09-27 12:54 ` [PATCH v8 9/9] net/cpfl: add support of to represented port action yuying.zhang
2023-09-28 3:37 ` [PATCH v8 0/9] add rte flow support for cpfl Zhang, Qi Z
2023-09-28 8:44 ` [PATCH v9 " yuying.zhang
2023-09-08 16:05 ` [PATCH v10 " Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 1/9] net/cpfl: parse flow offloading hint from JSON Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 2/9] net/cpfl: build action mapping rules " Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 3/9] net/cpfl: set up flow offloading skeleton Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 4/9] net/cpfl: set up control path Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 5/9] net/cpfl: add FXP low level implementation Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 6/9] net/cpfl: implement FXP rule creation and destroying Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 7/9] net/cpfl: adapt FXP to flow engine Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 8/9] net/cpfl: support flow ops on representor Zhang, Yuying
2023-09-08 16:05 ` [PATCH v10 9/9] net/cpfl: support represented port action Zhang, Yuying
2023-09-28 8:44 ` [PATCH v9 1/9] net/cpfl: add json parser for rte flow pattern rules yuying.zhang
2023-09-28 8:44 ` [PATCH v9 2/9] net/cpfl: build action mapping rules from JSON yuying.zhang
2023-09-28 8:44 ` [PATCH v9 3/9] net/cpfl: set up rte flow skeleton yuying.zhang
2023-10-15 13:01 ` Thomas Monjalon
2023-10-16 3:07 ` Zhang, Qi Z
2023-09-28 8:44 ` [PATCH v9 4/9] net/cpfl: set up control path yuying.zhang
2023-09-28 8:44 ` [PATCH v9 5/9] net/cpfl: add FXP low level implementation yuying.zhang
2023-09-28 8:44 ` [PATCH v9 6/9] net/cpfl: add fxp rule module yuying.zhang
2023-09-28 8:44 ` [PATCH v9 7/9] net/cpfl: add fxp flow engine yuying.zhang
2023-09-28 8:44 ` [PATCH v9 8/9] net/cpfl: add flow support for representor yuying.zhang
2023-09-28 8:44 ` [PATCH v9 9/9] net/cpfl: add support of to represented port action yuying.zhang
2023-09-28 12:45 ` [PATCH v9 0/9] add rte flow support for cpfl Zhang, Qi Z
2023-09-28 16:04 ` Stephen Hemminger
2023-10-09 4:00 ` [PATCH v10 " Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 1/9] net/cpfl: parse flow offloading hint from JSON Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 2/9] net/cpfl: build action mapping rules " Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 3/9] net/cpfl: set up flow offloading skeleton Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 4/9] net/cpfl: set up control path Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 5/9] net/cpfl: add FXP low level implementation Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 6/9] net/cpfl: implement FXP rule creation and destroying Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 7/9] net/cpfl: adapt FXP to flow engine Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 8/9] net/cpfl: support flow ops on representor Zhang, Yuying
2023-10-09 4:00 ` [PATCH v10 9/9] net/cpfl: support represented port action Zhang, Yuying
2023-10-10 1:31 ` [PATCH v10 0/9] add rte flow support for cpfl Zhang, Qi Z
2023-10-15 11:21 ` [PATCH v9 " Thomas Monjalon
2023-09-15 10:00 ` [PATCH v5 1/9] net/cpfl: add json parser for rte flow pattern rules Zhang, Yuying
2023-09-15 11:14 ` Zhang, Qi Z
2023-09-15 10:00 ` [PATCH v5 2/9] net/cpfl: add mod rule parser support for rte flow Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 3/9] net/cpfl: set up rte flow skeleton Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 4/9] net/cpfl: add FXP low level implementation Zhang, Yuying
2023-09-15 11:19 ` Zhang, Qi Z
2023-09-15 10:00 ` [PATCH v5 5/9] net/cpfl: add fxp rule module Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 6/9] net/cpfl: add fxp flow engine Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 7/9] net/cpfl: add flow support for representor Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 8/9] app/test-pmd: refine encap content Zhang, Yuying
2023-09-15 10:00 ` [PATCH v5 9/9] net/cpfl: fix incorrect status calculation Zhang, Yuying
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=PH0PR11MB5877A249B62C545CF39BB2DFECEDA@PH0PR11MB5877.namprd11.prod.outlook.com \
--to=mingxia.liu@intel.com \
--cc=beilei.xing@intel.com \
--cc=dev@dpdk.org \
--cc=jingjing.wu@intel.com \
--cc=qi.z.zhang@intel.com \
--cc=wenjing.qiao@intel.com \
--cc=yuying.zhang@intel.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).