X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fice%2Fbase%2Fice_flex_pipe.c;h=999ad6be3c5547b353348da01e9481e0a113e18c;hb=58f58cfe709cf167f9684e814cdbc8ab7535a798;hp=5253780793d257765b7f927f179d0759c2202d09;hpb=f0c38010536931377da5f9e99ecbc916e2cc78a0;p=dpdk.git diff --git a/drivers/net/ice/base/ice_flex_pipe.c b/drivers/net/ice/base/ice_flex_pipe.c index 5253780793..999ad6be3c 100644 --- a/drivers/net/ice/base/ice_flex_pipe.c +++ b/drivers/net/ice/base/ice_flex_pipe.c @@ -1,5 +1,5 @@ /* SPDX-License-Identifier: BSD-3-Clause - * Copyright(c) 2001-2019 + * Copyright(c) 2001-2020 Intel Corporation */ #include "ice_common.h" @@ -7,19 +7,12 @@ #include "ice_protocol_type.h" #include "ice_flow.h" +/* To support tunneling entries by PF, the package will append the PF number to + * the label; for example TNL_VXLAN_PF0, TNL_VXLAN_PF1, TNL_VXLAN_PF2, etc. + */ static const struct ice_tunnel_type_scan tnls[] = { - { TNL_VXLAN, "TNL_VXLAN" }, - { TNL_GTPC, "TNL_GTPC" }, - { TNL_GTPC_TEID, "TNL_GTPC_TEID" }, - { TNL_GTPU, "TNL_GTPC" }, - { TNL_GTPU_TEID, "TNL_GTPU_TEID" }, - { TNL_VXLAN_GPE, "TNL_VXLAN_GPE" }, - { TNL_GENEVE, "TNL_GENEVE" }, - { TNL_NAT, "TNL_NAT" }, - { TNL_ROCE_V2, "TNL_ROCE_V2" }, - { TNL_MPLSO_UDP, "TNL_MPLSO_UDP" }, - { TNL_UDP2_END, "TNL_UDP2_END" }, - { TNL_UPD_END, "TNL_UPD_END" }, + { TNL_VXLAN, "TNL_VXLAN_PF" }, + { TNL_GENEVE, "TNL_GENEVE_PF" }, { TNL_LAST, "" } }; @@ -138,10 +131,11 @@ static struct ice_buf_table *ice_find_buf_table(struct ice_seg *ice_seg) { struct ice_nvm_table *nvms; - nvms = (struct ice_nvm_table *)(ice_seg->device_table + - LE32_TO_CPU(ice_seg->device_table_count)); + nvms = (struct ice_nvm_table *) + (ice_seg->device_table + + LE32_TO_CPU(ice_seg->device_table_count)); - return (struct ice_buf_table *) + return (_FORCE_ struct ice_buf_table *) (nvms->vers + LE32_TO_CPU(nvms->table_count)); } @@ -405,7 +399,7 @@ ice_find_boost_entry(struct ice_seg *ice_seg, u16 addr, * Handles enumeration of individual label entries. */ static void * -ice_label_enum_handler(u32 __always_unused sect_type, void *section, u32 index, +ice_label_enum_handler(u32 __ALWAYS_UNUSED sect_type, void *section, u32 index, u32 *offset) { struct ice_label_section *labels; @@ -468,7 +462,7 @@ ice_enum_labels(struct ice_seg *ice_seg, u32 type, struct ice_pkg_enum *state, * since the first call to ice_enum_labels requires a pointer to an actual * ice_seg structure. */ -void ice_init_pkg_hints(struct ice_hw *hw, struct ice_seg *ice_seg) +static void ice_init_pkg_hints(struct ice_hw *hw, struct ice_seg *ice_seg) { struct ice_pkg_enum state; char *label_name; @@ -476,6 +470,7 @@ void ice_init_pkg_hints(struct ice_hw *hw, struct ice_seg *ice_seg) int i; ice_memset(&hw->tnl, 0, sizeof(hw->tnl), ICE_NONDMA_MEM); + ice_memset(&state, 0, sizeof(state), ICE_NONDMA_MEM); if (!ice_seg) return; @@ -485,8 +480,17 @@ void ice_init_pkg_hints(struct ice_hw *hw, struct ice_seg *ice_seg) while (label_name && hw->tnl.count < ICE_TUNNEL_MAX_ENTRIES) { for (i = 0; tnls[i].type != TNL_LAST; i++) { - if (!strncmp(label_name, tnls[i].label_prefix, - strlen(tnls[i].label_prefix))) { + size_t len = strlen(tnls[i].label_prefix); + + /* Look for matching label start, before continuing */ + if (strncmp(label_name, tnls[i].label_prefix, len)) + continue; + + /* Make sure this label matches our PF. Note that the PF + * character ('0' - '7') will be located where our + * prefix string's null terminator is located. + */ + if ((label_name[len] - '0') == hw->pf_id) { hw->tnl.tbl[hw->tnl.count].type = tnls[i].type; hw->tnl.tbl[hw->tnl.count].valid = false; hw->tnl.tbl[hw->tnl.count].in_use = false; @@ -606,7 +610,7 @@ ice_gen_key_word(u8 val, u8 valid, u8 dont_care, u8 nvr_mtch, u8 *key, static bool ice_bits_max_set(const u8 *mask, u16 size, u16 max) { u16 count = 0; - u16 i, j; + u16 i; /* check each byte */ for (i = 0; i < size; i++) { @@ -622,11 +626,9 @@ static bool ice_bits_max_set(const u8 *mask, u16 size, u16 max) return false; /* count the bits in this byte, checking threshold */ - for (j = 0; j < BITS_PER_BYTE; j++) { - count += (mask[i] & (0x1 << j)) ? 1 : 0; - if (count > max) - return false; - } + count += ice_hweight8(mask[i]); + if (count > max) + return false; } return true; @@ -638,7 +640,7 @@ static bool ice_bits_max_set(const u8 *mask, u16 size, u16 max) * @size: the size of the complete key in bytes (must be even) * @val: array of 8-bit values that makes up the value portion of the key * @upd: array of 8-bit masks that determine what key portion to update - * @dc: array of 8-bit masks that make up the dont' care mask + * @dc: array of 8-bit masks that make up the don't care mask * @nm: array of 8-bit masks that make up the never match mask * @off: the offset of the first byte in the key to update * @len: the number of bytes in the key update @@ -646,7 +648,7 @@ static bool ice_bits_max_set(const u8 *mask, u16 size, u16 max) * This function generates a key from a value, a don't care mask and a never * match mask. * upd, dc, and nm are optional parameters, and can be NULL: - * upd == NULL --> udp mask is all 1's (update all bits) + * upd == NULL --> upd mask is all 1's (update all bits) * dc == NULL --> dc mask is all 0's (no don't care bits) * nm == NULL --> nm mask is all 0's (no never match bits) */ @@ -704,7 +706,7 @@ ice_acquire_global_cfg_lock(struct ice_hw *hw, { enum ice_status status; - ice_debug(hw, ICE_DBG_TRACE, "ice_acquire_global_cfg_lock"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); status = ice_acquire_res(hw, ICE_GLOBAL_CFG_LOCK_RES_ID, access, ICE_GLOBAL_CFG_LOCK_TIMEOUT); @@ -734,10 +736,10 @@ static void ice_release_global_cfg_lock(struct ice_hw *hw) * * This function will request ownership of the change lock. */ -static enum ice_status +enum ice_status ice_acquire_change_lock(struct ice_hw *hw, enum ice_aq_res_access_type access) { - ice_debug(hw, ICE_DBG_TRACE, "ice_acquire_change_lock"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); return ice_acquire_res(hw, ICE_CHANGE_LOCK_RES_ID, access, ICE_CHANGE_LOCK_TIMEOUT); @@ -749,9 +751,9 @@ ice_acquire_change_lock(struct ice_hw *hw, enum ice_aq_res_access_type access) * * This function will release the change lock using the proper Admin Command. */ -static void ice_release_change_lock(struct ice_hw *hw) +void ice_release_change_lock(struct ice_hw *hw) { - ice_debug(hw, ICE_DBG_TRACE, "ice_release_change_lock"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); ice_release_res(hw, ICE_CHANGE_LOCK_RES_ID); } @@ -777,7 +779,7 @@ ice_aq_download_pkg(struct ice_hw *hw, struct ice_buf_hdr *pkg_buf, struct ice_aq_desc desc; enum ice_status status; - ice_debug(hw, ICE_DBG_TRACE, "ice_aq_download_pkg"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); if (error_offset) *error_offset = 0; @@ -806,28 +808,6 @@ ice_aq_download_pkg(struct ice_hw *hw, struct ice_buf_hdr *pkg_buf, return status; } -/** - * ice_aq_upload_section - * @hw: pointer to the hardware structure - * @pkg_buf: the package buffer which will receive the section - * @buf_size: the size of the package buffer - * @cd: pointer to command details structure or NULL - * - * Upload Section (0x0C41) - */ -enum ice_status -ice_aq_upload_section(struct ice_hw *hw, struct ice_buf_hdr *pkg_buf, - u16 buf_size, struct ice_sq_cd *cd) -{ - struct ice_aq_desc desc; - - ice_debug(hw, ICE_DBG_TRACE, "ice_aq_upload_section"); - ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_upload_section); - desc.flags |= CPU_TO_LE16(ICE_AQ_FLAG_RD); - - return ice_aq_send_cmd(hw, &desc, pkg_buf, buf_size, cd); -} - /** * ice_aq_update_pkg * @hw: pointer to the hardware structure @@ -849,7 +829,7 @@ ice_aq_update_pkg(struct ice_hw *hw, struct ice_buf_hdr *pkg_buf, u16 buf_size, struct ice_aq_desc desc; enum ice_status status; - ice_debug(hw, ICE_DBG_TRACE, "ice_aq_update_pkg"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); if (error_offset) *error_offset = 0; @@ -888,16 +868,17 @@ ice_aq_update_pkg(struct ice_hw *hw, struct ice_buf_hdr *pkg_buf, u16 buf_size, * success it returns a pointer to the segment header, otherwise it will * return NULL. */ -struct ice_generic_seg_hdr * +static struct ice_generic_seg_hdr * ice_find_seg_in_pkg(struct ice_hw *hw, u32 seg_type, struct ice_pkg_hdr *pkg_hdr) { u32 i; ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); - ice_debug(hw, ICE_DBG_PKG, "Package version: %d.%d.%d.%d\n", - pkg_hdr->format_ver.major, pkg_hdr->format_ver.minor, - pkg_hdr->format_ver.update, pkg_hdr->format_ver.draft); + ice_debug(hw, ICE_DBG_PKG, "Package format version: %d.%d.%d.%d\n", + pkg_hdr->pkg_format_ver.major, pkg_hdr->pkg_format_ver.minor, + pkg_hdr->pkg_format_ver.update, + pkg_hdr->pkg_format_ver.draft); /* Search all package segments for the requested segment type */ for (i = 0; i < LE32_TO_CPU(pkg_hdr->seg_count); i++) { @@ -932,9 +913,8 @@ ice_update_pkg(struct ice_hw *hw, struct ice_buf *bufs, u32 count) return status; for (i = 0; i < count; i++) { - bool last = ((i + 1) == count); - struct ice_buf_hdr *bh = (struct ice_buf_hdr *)(bufs + i); + bool last = ((i + 1) == count); status = ice_aq_update_pkg(hw, bh, LE16_TO_CPU(bh->data_end), last, &offset, &info, NULL); @@ -980,9 +960,19 @@ ice_dwnld_cfg_bufs(struct ice_hw *hw, struct ice_buf *bufs, u32 count) if (LE32_TO_CPU(bh->section_entry[0].type) & ICE_METADATA_BUF) return ICE_SUCCESS; + /* reset pkg_dwnld_status in case this function is called in the + * reset/rebuild flow + */ + hw->pkg_dwnld_status = ICE_AQ_RC_OK; + status = ice_acquire_global_cfg_lock(hw, ICE_RES_WRITE); - if (status) + if (status) { + if (status == ICE_ERR_AQ_NO_WORK) + hw->pkg_dwnld_status = ICE_AQ_RC_EEXIST; + else + hw->pkg_dwnld_status = hw->adminq.sq_last_status; return status; + } for (i = 0; i < count; i++) { bool last = ((i + 1) == count); @@ -1003,9 +993,11 @@ ice_dwnld_cfg_bufs(struct ice_hw *hw, struct ice_buf *bufs, u32 count) bh = (struct ice_buf_hdr *)(bufs + i); - status = ice_aq_download_pkg(hw, bh, LE16_TO_CPU(bh->data_end), - last, &offset, &info, NULL); + status = ice_aq_download_pkg(hw, bh, ICE_PKG_BUF_SIZE, last, + &offset, &info, NULL); + /* Save AQ status from download package */ + hw->pkg_dwnld_status = hw->adminq.sq_last_status; if (status) { ice_debug(hw, ICE_DBG_PKG, "Pkg download failed: err %d off %d inf %d\n", @@ -1038,7 +1030,7 @@ ice_aq_get_pkg_info_list(struct ice_hw *hw, { struct ice_aq_desc desc; - ice_debug(hw, ICE_DBG_TRACE, "ice_aq_get_pkg_info_list"); + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_get_pkg_info_list); return ice_aq_send_cmd(hw, &desc, pkg_info, buf_size, cd); @@ -1051,18 +1043,21 @@ ice_aq_get_pkg_info_list(struct ice_hw *hw, * * Handles the download of a complete package. */ -enum ice_status ice_download_pkg(struct ice_hw *hw, struct ice_seg *ice_seg) +static enum ice_status +ice_download_pkg(struct ice_hw *hw, struct ice_seg *ice_seg) { struct ice_buf_table *ice_buf_tbl; ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); - ice_debug(hw, ICE_DBG_PKG, "Segment version: %d.%d.%d.%d\n", - ice_seg->hdr.seg_ver.major, ice_seg->hdr.seg_ver.minor, - ice_seg->hdr.seg_ver.update, ice_seg->hdr.seg_ver.draft); + ice_debug(hw, ICE_DBG_PKG, "Segment format version: %d.%d.%d.%d\n", + ice_seg->hdr.seg_format_ver.major, + ice_seg->hdr.seg_format_ver.minor, + ice_seg->hdr.seg_format_ver.update, + ice_seg->hdr.seg_format_ver.draft); ice_debug(hw, ICE_DBG_PKG, "Seg: type 0x%X, size %d, name %s\n", LE32_TO_CPU(ice_seg->hdr.seg_type), - LE32_TO_CPU(ice_seg->hdr.seg_size), ice_seg->hdr.seg_name); + LE32_TO_CPU(ice_seg->hdr.seg_size), ice_seg->hdr.seg_id); ice_buf_tbl = ice_find_buf_table(ice_seg); @@ -1080,15 +1075,11 @@ enum ice_status ice_download_pkg(struct ice_hw *hw, struct ice_seg *ice_seg) * * Saves off the package details into the HW structure. */ -enum ice_status +static enum ice_status ice_init_pkg_info(struct ice_hw *hw, struct ice_pkg_hdr *pkg_hdr) { - struct ice_aqc_get_pkg_info_resp *pkg_info; struct ice_global_metadata_seg *meta_seg; struct ice_generic_seg_hdr *seg_hdr; - enum ice_status status; - u16 size; - u32 i; ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); if (!pkg_hdr) @@ -1113,23 +1104,41 @@ ice_init_pkg_info(struct ice_hw *hw, struct ice_pkg_hdr *pkg_hdr) seg_hdr = ice_find_seg_in_pkg(hw, SEGMENT_TYPE_ICE, pkg_hdr); if (seg_hdr) { - hw->ice_pkg_ver = seg_hdr->seg_ver; - ice_memcpy(hw->ice_pkg_name, seg_hdr->seg_name, + hw->ice_pkg_ver = seg_hdr->seg_format_ver; + ice_memcpy(hw->ice_pkg_name, seg_hdr->seg_id, sizeof(hw->ice_pkg_name), ICE_NONDMA_TO_NONDMA); - ice_debug(hw, ICE_DBG_PKG, "Ice Pkg: %d.%d.%d.%d, %s\n", - seg_hdr->seg_ver.major, seg_hdr->seg_ver.minor, - seg_hdr->seg_ver.update, seg_hdr->seg_ver.draft, - seg_hdr->seg_name); + ice_debug(hw, ICE_DBG_PKG, "Ice Seg: %d.%d.%d.%d, %s\n", + seg_hdr->seg_format_ver.major, + seg_hdr->seg_format_ver.minor, + seg_hdr->seg_format_ver.update, + seg_hdr->seg_format_ver.draft, + seg_hdr->seg_id); } else { ice_debug(hw, ICE_DBG_INIT, "Did not find ice segment in driver package\n"); return ICE_ERR_CFG; } -#define ICE_PKG_CNT 4 - size = sizeof(*pkg_info) + (sizeof(pkg_info->pkg_info[0]) * - (ICE_PKG_CNT - 1)); + return ICE_SUCCESS; +} + +/** + * ice_get_pkg_info + * @hw: pointer to the hardware structure + * + * Store details of the package currently loaded in HW into the HW structure. + */ +static enum ice_status ice_get_pkg_info(struct ice_hw *hw) +{ + struct ice_aqc_get_pkg_info_resp *pkg_info; + enum ice_status status; + u16 size; + u32 i; + + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); + + size = ice_struct_size(pkg_info, pkg_info, ICE_PKG_CNT - 1); pkg_info = (struct ice_aqc_get_pkg_info_resp *)ice_malloc(hw, size); if (!pkg_info) return ICE_ERR_NO_MEMORY; @@ -1146,10 +1155,13 @@ ice_init_pkg_info(struct ice_hw *hw, struct ice_pkg_hdr *pkg_hdr) if (pkg_info->pkg_info[i].is_active) { flags[place++] = 'A'; hw->active_pkg_ver = pkg_info->pkg_info[i].ver; + hw->active_track_id = + LE32_TO_CPU(pkg_info->pkg_info[i].track_id); ice_memcpy(hw->active_pkg_name, pkg_info->pkg_info[i].name, - sizeof(hw->active_pkg_name), + sizeof(pkg_info->pkg_info[i].name), ICE_NONDMA_TO_NONDMA); + hw->active_pkg_in_nvm = pkg_info->pkg_info[i].is_in_nvm; } if (pkg_info->pkg_info[i].is_active_at_boot) flags[place++] = 'B'; @@ -1172,41 +1184,6 @@ init_pkg_free_alloc: return status; } -/** - * ice_find_label_value - * @ice_seg: pointer to the ice segment (non-NULL) - * @name: name of the label to search for - * @type: the section type that will contain the label - * @value: pointer to a value that will return the label's value if found - * - * Finds a label's value given the label name and the section type to search. - * The ice_seg parameter must not be NULL since the first call to - * ice_enum_labels requires a pointer to an actual ice_seg structure. - */ -enum ice_status -ice_find_label_value(struct ice_seg *ice_seg, char const *name, u32 type, - u16 *value) -{ - struct ice_pkg_enum state; - char *label_name; - u16 val; - - if (!ice_seg) - return ICE_ERR_PARAM; - - do { - label_name = ice_enum_labels(ice_seg, type, &state, &val); - if (label_name && !strcmp(label_name, name)) { - *value = val; - return ICE_SUCCESS; - } - - ice_seg = NULL; - } while (label_name); - - return ICE_ERR_CFG; -} - /** * ice_verify_pkg - verify package * @pkg: pointer to the package buffer @@ -1223,10 +1200,10 @@ static enum ice_status ice_verify_pkg(struct ice_pkg_hdr *pkg, u32 len) if (len < sizeof(*pkg)) return ICE_ERR_BUF_TOO_SHORT; - if (pkg->format_ver.major != ICE_PKG_FMT_VER_MAJ || - pkg->format_ver.minor != ICE_PKG_FMT_VER_MNR || - pkg->format_ver.update != ICE_PKG_FMT_VER_UPD || - pkg->format_ver.draft != ICE_PKG_FMT_VER_DFT) + if (pkg->pkg_format_ver.major != ICE_PKG_FMT_VER_MAJ || + pkg->pkg_format_ver.minor != ICE_PKG_FMT_VER_MNR || + pkg->pkg_format_ver.update != ICE_PKG_FMT_VER_UPD || + pkg->pkg_format_ver.draft != ICE_PKG_FMT_VER_DFT) return ICE_ERR_CFG; /* pkg must have at least one segment */ @@ -1235,7 +1212,7 @@ static enum ice_status ice_verify_pkg(struct ice_pkg_hdr *pkg, u32 len) return ICE_ERR_CFG; /* make sure segment array fits in package length */ - if (len < sizeof(*pkg) + ((seg_count - 1) * sizeof(pkg->seg_offset))) + if (len < ice_struct_size(pkg, seg_offset, seg_count - 1)) return ICE_ERR_BUF_TOO_SHORT; /* all segments must fit within length */ @@ -1283,12 +1260,96 @@ static void ice_init_pkg_regs(struct ice_hw *hw) #define ICE_SW_BLK_INP_MASK_L 0xFFFFFFFF #define ICE_SW_BLK_INP_MASK_H 0x0000FFFF #define ICE_SW_BLK_IDX 0 + if (hw->dcf_enabled) + return; /* setup Switch block input mask, which is 48-bits in two parts */ wr32(hw, GL_PREEXT_L2_PMASK0(ICE_SW_BLK_IDX), ICE_SW_BLK_INP_MASK_L); wr32(hw, GL_PREEXT_L2_PMASK1(ICE_SW_BLK_IDX), ICE_SW_BLK_INP_MASK_H); } +/** + * ice_chk_pkg_version - check package version for compatibility with driver + * @pkg_ver: pointer to a version structure to check + * + * Check to make sure that the package about to be downloaded is compatible with + * the driver. To be compatible, the major and minor components of the package + * version must match our ICE_PKG_SUPP_VER_MAJ and ICE_PKG_SUPP_VER_MNR + * definitions. + */ +static enum ice_status ice_chk_pkg_version(struct ice_pkg_ver *pkg_ver) +{ + if (pkg_ver->major != ICE_PKG_SUPP_VER_MAJ || + pkg_ver->minor != ICE_PKG_SUPP_VER_MNR) + return ICE_ERR_NOT_SUPPORTED; + + return ICE_SUCCESS; +} + +/** + * ice_chk_pkg_compat + * @hw: pointer to the hardware structure + * @ospkg: pointer to the package hdr + * @seg: pointer to the package segment hdr + * + * This function checks the package version compatibility with driver and NVM + */ +static enum ice_status +ice_chk_pkg_compat(struct ice_hw *hw, struct ice_pkg_hdr *ospkg, + struct ice_seg **seg) +{ + struct ice_aqc_get_pkg_info_resp *pkg; + enum ice_status status; + u16 size; + u32 i; + + ice_debug(hw, ICE_DBG_TRACE, "%s\n", __func__); + + /* Check package version compatibility */ + status = ice_chk_pkg_version(&hw->pkg_ver); + if (status) { + ice_debug(hw, ICE_DBG_INIT, "Package version check failed.\n"); + return status; + } + + /* find ICE segment in given package */ + *seg = (struct ice_seg *)ice_find_seg_in_pkg(hw, SEGMENT_TYPE_ICE, + ospkg); + if (!*seg) { + ice_debug(hw, ICE_DBG_INIT, "no ice segment in package.\n"); + return ICE_ERR_CFG; + } + + /* Check if FW is compatible with the OS package */ + size = ice_struct_size(pkg, pkg_info, ICE_PKG_CNT - 1); + pkg = (struct ice_aqc_get_pkg_info_resp *)ice_malloc(hw, size); + if (!pkg) + return ICE_ERR_NO_MEMORY; + + status = ice_aq_get_pkg_info_list(hw, pkg, size, NULL); + if (status) + goto fw_ddp_compat_free_alloc; + + for (i = 0; i < LE32_TO_CPU(pkg->count); i++) { + /* loop till we find the NVM package */ + if (!pkg->pkg_info[i].is_in_nvm) + continue; + if ((*seg)->hdr.seg_format_ver.major != + pkg->pkg_info[i].ver.major || + (*seg)->hdr.seg_format_ver.minor > + pkg->pkg_info[i].ver.minor) { + status = ICE_ERR_FW_DDP_MISMATCH; + ice_debug(hw, ICE_DBG_INIT, + "OS package is not compatible with NVM.\n"); + } + /* done processing NVM package so break */ + break; + } +fw_ddp_compat_free_alloc: + ice_free(hw, pkg); + return status; +} + /** * ice_init_pkg - initialize/download package * @hw: pointer to the hardware structure @@ -1336,12 +1397,12 @@ enum ice_status ice_init_pkg(struct ice_hw *hw, u8 *buf, u32 len) if (status) return status; - /* find segment in given package */ - seg = (struct ice_seg *)ice_find_seg_in_pkg(hw, SEGMENT_TYPE_ICE, pkg); - if (!seg) { - ice_debug(hw, ICE_DBG_INIT, "no ice segment in package.\n"); - return ICE_ERR_CFG; - } + /* before downloading the package, check package version for + * compatibility with driver + */ + status = ice_chk_pkg_compat(hw, pkg, &seg); + if (status) + return status; /* initialize package hints and then download package */ ice_init_pkg_hints(hw, seg); @@ -1352,12 +1413,23 @@ enum ice_status ice_init_pkg(struct ice_hw *hw, u8 *buf, u32 len) status = ICE_SUCCESS; } + /* Get information on the package currently loaded in HW, then make sure + * the driver is compatible with this version. + */ + if (!status) { + status = ice_get_pkg_info(hw); + if (!status) + status = ice_chk_pkg_version(&hw->active_pkg_ver); + } + if (!status) { hw->seg = seg; - /* on successful package download, update other required - * registers to support the package + /* on successful package download update other required + * registers to support the package and fill HW tables + * with package content. */ ice_init_pkg_regs(hw); + ice_fill_blk_tbls(hw); } else { ice_debug(hw, ICE_DBG_INIT, "package load failed, %d\n", status); @@ -1419,7 +1491,7 @@ enum ice_status ice_copy_and_init_pkg(struct ice_hw *hw, const u8 *buf, u32 len) * Allocates a package buffer and returns a pointer to the buffer header. * Note: all package contents must be in Little Endian form. */ -struct ice_buf_build *ice_pkg_buf_alloc(struct ice_hw *hw) +static struct ice_buf_build *ice_pkg_buf_alloc(struct ice_hw *hw) { struct ice_buf_build *bld; struct ice_buf_hdr *buf; @@ -1429,8 +1501,8 @@ struct ice_buf_build *ice_pkg_buf_alloc(struct ice_hw *hw) return NULL; buf = (struct ice_buf_hdr *)bld; - buf->data_end = CPU_TO_LE16(sizeof(*buf) - - sizeof(buf->section_entry[0])); + buf->data_end = CPU_TO_LE16(offsetof(struct ice_buf_hdr, + section_entry)); return bld; } @@ -1466,11 +1538,86 @@ ice_sw_fv_handler(u32 sect_type, void *section, u32 index, u32 *offset) return fv_section->fv + index; } +/** + * ice_get_sw_prof_type - determine switch profile type + * @hw: pointer to the HW structure + * @fv: pointer to the switch field vector + */ +static enum ice_prof_type +ice_get_sw_prof_type(struct ice_hw *hw, struct ice_fv *fv) +{ + u16 i; + + for (i = 0; i < hw->blk[ICE_BLK_SW].es.fvw; i++) { + /* UDP tunnel will have UDP_OF protocol ID and VNI offset */ + if (fv->ew[i].prot_id == (u8)ICE_PROT_UDP_OF && + fv->ew[i].off == ICE_VNI_OFFSET) + return ICE_PROF_TUN_UDP; + + /* GRE tunnel will have GRE protocol */ + if (fv->ew[i].prot_id == (u8)ICE_PROT_GRE_OF) + return ICE_PROF_TUN_GRE; + + /* PPPOE tunnel will have PPPOE protocol */ + if (fv->ew[i].prot_id == (u8)ICE_PROT_PPPOE) + return ICE_PROF_TUN_PPPOE; + } + + return ICE_PROF_NON_TUN; +} + +/** + * ice_get_sw_fv_bitmap - Get switch field vector bitmap based on profile type + * @hw: pointer to hardware structure + * @req_profs: type of profiles requested + * @bm: pointer to memory for returning the bitmap of field vectors + */ +void +ice_get_sw_fv_bitmap(struct ice_hw *hw, enum ice_prof_type req_profs, + ice_bitmap_t *bm) +{ + struct ice_pkg_enum state; + struct ice_seg *ice_seg; + struct ice_fv *fv; + + ice_memset(&state, 0, sizeof(state), ICE_NONDMA_MEM); + + if (req_profs == ICE_PROF_ALL) { + u16 i; + + for (i = 0; i < ICE_MAX_NUM_PROFILES; i++) + ice_set_bit(i, bm); + return; + } + + ice_zero_bitmap(bm, ICE_MAX_NUM_PROFILES); + + ice_seg = hw->seg; + do { + enum ice_prof_type prof_type; + u32 offset; + + fv = (struct ice_fv *) + ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW, + &offset, ice_sw_fv_handler); + ice_seg = NULL; + + if (fv) { + /* Determine field vector type */ + prof_type = ice_get_sw_prof_type(hw, fv); + + if (req_profs & prof_type) + ice_set_bit((u16)offset, bm); + } + } while (fv); +} + /** * ice_get_sw_fv_list * @hw: pointer to the HW structure * @prot_ids: field vector to search for with a given protocol ID * @ids_cnt: lookup/protocol count + * @bm: bitmap of field vectors to consider * @fv_list: Head of a list * * Finds all the field vector entries from switch block that contain @@ -1481,8 +1628,8 @@ ice_sw_fv_handler(u32 sect_type, void *section, u32 index, u32 *offset) * allocated for every list entry. */ enum ice_status -ice_get_sw_fv_list(struct ice_hw *hw, u16 *prot_ids, u8 ids_cnt, - struct LIST_HEAD_TYPE *fv_list) +ice_get_sw_fv_list(struct ice_hw *hw, u8 *prot_ids, u16 ids_cnt, + ice_bitmap_t *bm, struct LIST_HEAD_TYPE *fv_list) { struct ice_sw_fv_list_entry *fvl; struct ice_sw_fv_list_entry *tmp; @@ -1491,18 +1638,29 @@ ice_get_sw_fv_list(struct ice_hw *hw, u16 *prot_ids, u8 ids_cnt, struct ice_fv *fv; u32 offset; + ice_memset(&state, 0, sizeof(state), ICE_NONDMA_MEM); + if (!ids_cnt || !hw->seg) return ICE_ERR_PARAM; ice_seg = hw->seg; do { - u8 i; + u16 i; fv = (struct ice_fv *) ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW, &offset, ice_sw_fv_handler); + if (!fv) + break; + ice_seg = NULL; + + /* If field vector is not in the bitmap list, then skip this + * profile. + */ + if (!ice_is_bit_set(bm, (u16)offset)) + continue; - for (i = 0; i < ids_cnt && fv; i++) { + for (i = 0; i < ids_cnt; i++) { int j; /* This code assumes that if a switch field vector line @@ -1526,7 +1684,6 @@ ice_get_sw_fv_list(struct ice_hw *hw, u16 *prot_ids, u8 ids_cnt, break; } } - ice_seg = NULL; } while (fv); if (LIST_EMPTY(fv_list)) return ICE_ERR_CFG; @@ -1543,40 +1700,57 @@ err: } /** - * ice_pkg_buf_alloc_single_section - * @hw: pointer to the HW structure - * @type: the section type value - * @size: the size of the section to reserve (in bytes) - * @section: returns pointer to the section - * - * Allocates a package buffer with a single section. - * Note: all package contents must be in Little Endian form. + * ice_init_prof_result_bm - Initialize the profile result index bitmap + * @hw: pointer to hardware structure */ -static struct ice_buf_build * -ice_pkg_buf_alloc_single_section(struct ice_hw *hw, u32 type, u16 size, - void **section) +void ice_init_prof_result_bm(struct ice_hw *hw) { - struct ice_buf_build *buf; + struct ice_pkg_enum state; + struct ice_seg *ice_seg; + struct ice_fv *fv; - if (!section) - return NULL; + ice_memset(&state, 0, sizeof(state), ICE_NONDMA_MEM); - buf = ice_pkg_buf_alloc(hw); - if (!buf) - return NULL; + if (!hw->seg) + return; + + ice_seg = hw->seg; + do { + u32 off; + u16 i; - if (ice_pkg_buf_reserve_section(buf, 1)) - goto ice_pkg_buf_alloc_single_section_err; + fv = (struct ice_fv *) + ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW, + &off, ice_sw_fv_handler); + ice_seg = NULL; + if (!fv) + break; - *section = ice_pkg_buf_alloc_section(buf, type, size); - if (!*section) - goto ice_pkg_buf_alloc_single_section_err; + ice_zero_bitmap(hw->switch_info->prof_res_bm[off], + ICE_MAX_FV_WORDS); - return buf; + /* Determine empty field vector indices, these can be + * used for recipe results. Skip index 0, since it is + * always used for Switch ID. + */ + for (i = 1; i < ICE_MAX_FV_WORDS; i++) + if (fv->ew[i].prot_id == ICE_PROT_INVALID && + fv->ew[i].off == ICE_FV_OFFSET_INVAL) + ice_set_bit(i, + hw->switch_info->prof_res_bm[off]); + } while (fv); +} -ice_pkg_buf_alloc_single_section_err: - ice_pkg_buf_free(hw, buf); - return NULL; +/** + * ice_pkg_buf_free + * @hw: pointer to the HW structure + * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) + * + * Frees a package buffer + */ +static void ice_pkg_buf_free(struct ice_hw *hw, struct ice_buf_build *bld) +{ + ice_free(hw, bld); } /** @@ -1592,7 +1766,7 @@ ice_pkg_buf_alloc_single_section_err: * result in some wasted space in the buffer. * Note: all package contents must be in Little Endian form. */ -enum ice_status +static enum ice_status ice_pkg_buf_reserve_section(struct ice_buf_build *bld, u16 count) { struct ice_buf_hdr *buf; @@ -1620,48 +1794,6 @@ ice_pkg_buf_reserve_section(struct ice_buf_build *bld, u16 count) return ICE_SUCCESS; } -/** - * ice_pkg_buf_unreserve_section - * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) - * @count: the number of sections to unreserve - * - * Unreserves one or more section table entries in a package buffer, releasing - * space that can be used for section data. This routine can be called - * multiple times as long as they are made before calling - * ice_pkg_buf_alloc_section(). Once ice_pkg_buf_alloc_section() - * is called once, the number of sections that can be allocated will not be able - * to be increased; not using all reserved sections is fine, but this will - * result in some wasted space in the buffer. - * Note: all package contents must be in Little Endian form. - */ -enum ice_status -ice_pkg_buf_unreserve_section(struct ice_buf_build *bld, u16 count) -{ - struct ice_buf_hdr *buf; - u16 section_count; - u16 data_end; - - if (!bld) - return ICE_ERR_PARAM; - - buf = (struct ice_buf_hdr *)&bld->buf; - - /* already an active section, can't decrease table size */ - section_count = LE16_TO_CPU(buf->section_count); - if (section_count > 0) - return ICE_ERR_CFG; - - if (count > bld->reserved_section_table_entries) - return ICE_ERR_CFG; - bld->reserved_section_table_entries -= count; - - data_end = LE16_TO_CPU(buf->data_end) - - (count * sizeof(buf->section_entry[0])); - buf->data_end = CPU_TO_LE16(data_end); - - return ICE_SUCCESS; -} - /** * ice_pkg_buf_alloc_section * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) @@ -1674,7 +1806,7 @@ ice_pkg_buf_unreserve_section(struct ice_buf_build *bld, u16 count) * section contents. * Note: all package contents must be in Little Endian form. */ -void * +static void * ice_pkg_buf_alloc_section(struct ice_buf_build *bld, u32 type, u16 size) { struct ice_buf_hdr *buf; @@ -1716,25 +1848,7 @@ ice_pkg_buf_alloc_section(struct ice_buf_build *bld, u32 type, u16 size) } /** - * ice_pkg_buf_get_free_space - * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) - * - * Returns the number of free bytes remaining in the buffer. - * Note: all package contents must be in Little Endian form. - */ -u16 ice_pkg_buf_get_free_space(struct ice_buf_build *bld) -{ - struct ice_buf_hdr *buf; - - if (!bld) - return 0; - - buf = (struct ice_buf_hdr *)&bld->buf; - return ICE_MAX_S_DATA_END - LE16_TO_CPU(buf->data_end); -} - -/** - * ice_pkg_buf_get_active_sections + * ice_pkg_buf_get_active_sections * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) * * Returns the number of active sections. Before using the package buffer @@ -1743,7 +1857,7 @@ u16 ice_pkg_buf_get_free_space(struct ice_buf_build *bld) * not be used. * Note: all package contents must be in Little Endian form. */ -u16 ice_pkg_buf_get_active_sections(struct ice_buf_build *bld) +static u16 ice_pkg_buf_get_active_sections(struct ice_buf_build *bld) { struct ice_buf_hdr *buf; @@ -1755,12 +1869,12 @@ u16 ice_pkg_buf_get_active_sections(struct ice_buf_build *bld) } /** - * ice_pkg_buf_header + * ice_pkg_buf * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) * * Return a pointer to the buffer's header */ -struct ice_buf *ice_pkg_buf(struct ice_buf_build *bld) +static struct ice_buf *ice_pkg_buf(struct ice_buf_build *bld) { if (!bld) return NULL; @@ -1769,135 +1883,384 @@ struct ice_buf *ice_pkg_buf(struct ice_buf_build *bld) } /** - * ice_pkg_buf_free + * ice_tunnel_port_in_use_hlpr - helper function to determine tunnel usage * @hw: pointer to the HW structure - * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc()) + * @port: port to search for + * @index: optionally returns index * - * Frees a package buffer + * Returns whether a port is already in use as a tunnel, and optionally its + * index */ -void ice_pkg_buf_free(struct ice_hw *hw, struct ice_buf_build *bld) +static bool ice_tunnel_port_in_use_hlpr(struct ice_hw *hw, u16 port, u16 *index) { - ice_free(hw, bld); + u16 i; + + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].in_use && hw->tnl.tbl[i].port == port) { + if (index) + *index = i; + return true; + } + + return false; } -/* PTG Management */ +/** + * ice_tunnel_port_in_use + * @hw: pointer to the HW structure + * @port: port to search for + * @index: optionally returns index + * + * Returns whether a port is already in use as a tunnel, and optionally its + * index + */ +bool ice_tunnel_port_in_use(struct ice_hw *hw, u16 port, u16 *index) +{ + bool res; + + ice_acquire_lock(&hw->tnl_lock); + res = ice_tunnel_port_in_use_hlpr(hw, port, index); + ice_release_lock(&hw->tnl_lock); + + return res; +} /** - * ice_ptg_update_xlt1 - Updates packet type groups in HW via XLT1 table - * @hw: pointer to the hardware structure - * @blk: HW block + * ice_tunnel_get_type + * @hw: pointer to the HW structure + * @port: port to search for + * @type: returns tunnel index + * + * For a given port number, will return the type of tunnel. + */ +bool +ice_tunnel_get_type(struct ice_hw *hw, u16 port, enum ice_tunnel_type *type) +{ + bool res = false; + u16 i; + + ice_acquire_lock(&hw->tnl_lock); + + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].in_use && hw->tnl.tbl[i].port == port) { + *type = hw->tnl.tbl[i].type; + res = true; + break; + } + + ice_release_lock(&hw->tnl_lock); + + return res; +} + +/** + * ice_find_free_tunnel_entry + * @hw: pointer to the HW structure + * @type: tunnel type + * @index: optionally returns index + * + * Returns whether there is a free tunnel entry, and optionally its index + */ +static bool +ice_find_free_tunnel_entry(struct ice_hw *hw, enum ice_tunnel_type type, + u16 *index) +{ + u16 i; + + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].valid && !hw->tnl.tbl[i].in_use && + hw->tnl.tbl[i].type == type) { + if (index) + *index = i; + return true; + } + + return false; +} + +/** + * ice_get_open_tunnel_port - retrieve an open tunnel port + * @hw: pointer to the HW structure + * @type: tunnel type (TNL_ALL will return any open port) + * @port: returns open port + */ +bool +ice_get_open_tunnel_port(struct ice_hw *hw, enum ice_tunnel_type type, + u16 *port) +{ + bool res = false; + u16 i; + + ice_acquire_lock(&hw->tnl_lock); + + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].valid && hw->tnl.tbl[i].in_use && + (type == TNL_ALL || hw->tnl.tbl[i].type == type)) { + *port = hw->tnl.tbl[i].port; + res = true; + break; + } + + ice_release_lock(&hw->tnl_lock); + + return res; +} + +/** + * ice_create_tunnel + * @hw: pointer to the HW structure + * @type: type of tunnel + * @port: port of tunnel to create * - * This function will update the XLT1 hardware table to reflect the new - * packet type group configuration. + * Create a tunnel by updating the parse graph in the parser. We do that by + * creating a package buffer with the tunnel info and issuing an update package + * command. */ -enum ice_status ice_ptg_update_xlt1(struct ice_hw *hw, enum ice_block blk) +enum ice_status +ice_create_tunnel(struct ice_hw *hw, enum ice_tunnel_type type, u16 port) { - struct ice_xlt1_section *sect; + struct ice_boost_tcam_section *sect_rx, *sect_tx; + enum ice_status status = ICE_ERR_MAX_LIMIT; struct ice_buf_build *bld; - enum ice_status status; u16 index; - bld = ice_pkg_buf_alloc_single_section(hw, ice_sect_id(blk, ICE_XLT1), - ICE_XLT1_SIZE(ICE_XLT1_CNT), - (void **)§); - if (!bld) - return ICE_ERR_NO_MEMORY; + ice_acquire_lock(&hw->tnl_lock); + + if (ice_tunnel_port_in_use_hlpr(hw, port, &index)) { + hw->tnl.tbl[index].ref++; + status = ICE_SUCCESS; + goto ice_create_tunnel_end; + } + + if (!ice_find_free_tunnel_entry(hw, type, &index)) { + status = ICE_ERR_OUT_OF_RANGE; + goto ice_create_tunnel_end; + } + + bld = ice_pkg_buf_alloc(hw); + if (!bld) { + status = ICE_ERR_NO_MEMORY; + goto ice_create_tunnel_end; + } + + /* allocate 2 sections, one for Rx parser, one for Tx parser */ + if (ice_pkg_buf_reserve_section(bld, 2)) + goto ice_create_tunnel_err; + + sect_rx = (struct ice_boost_tcam_section *) + ice_pkg_buf_alloc_section(bld, ICE_SID_RXPARSER_BOOST_TCAM, + sizeof(*sect_rx)); + if (!sect_rx) + goto ice_create_tunnel_err; + sect_rx->count = CPU_TO_LE16(1); + + sect_tx = (struct ice_boost_tcam_section *) + ice_pkg_buf_alloc_section(bld, ICE_SID_TXPARSER_BOOST_TCAM, + sizeof(*sect_tx)); + if (!sect_tx) + goto ice_create_tunnel_err; + sect_tx->count = CPU_TO_LE16(1); + + /* copy original boost entry to update package buffer */ + ice_memcpy(sect_rx->tcam, hw->tnl.tbl[index].boost_entry, + sizeof(*sect_rx->tcam), ICE_NONDMA_TO_NONDMA); + + /* over-write the never-match dest port key bits with the encoded port + * bits + */ + ice_set_key((u8 *)§_rx->tcam[0].key, sizeof(sect_rx->tcam[0].key), + (u8 *)&port, NULL, NULL, NULL, + (u16)offsetof(struct ice_boost_key_value, hv_dst_port_key), + sizeof(sect_rx->tcam[0].key.key.hv_dst_port_key)); - sect->count = CPU_TO_LE16(ICE_XLT1_CNT); - sect->offset = CPU_TO_LE16(0); - for (index = 0; index < ICE_XLT1_CNT; index++) - sect->value[index] = hw->blk[blk].xlt1.ptypes[index].ptg; + /* exact copy of entry to Tx section entry */ + ice_memcpy(sect_tx->tcam, sect_rx->tcam, sizeof(*sect_tx->tcam), + ICE_NONDMA_TO_NONDMA); status = ice_update_pkg(hw, ice_pkg_buf(bld), 1); + if (!status) { + hw->tnl.tbl[index].port = port; + hw->tnl.tbl[index].in_use = true; + hw->tnl.tbl[index].ref = 1; + } +ice_create_tunnel_err: ice_pkg_buf_free(hw, bld); +ice_create_tunnel_end: + ice_release_lock(&hw->tnl_lock); + return status; } /** - * ice_ptg_find_ptype - Search for packet type group using packet type (ptype) - * @hw: pointer to the hardware structure - * @blk: HW block - * @ptype: the ptype to search for - * @ptg: pointer to variable that receives the PTG + * ice_destroy_tunnel + * @hw: pointer to the HW structure + * @port: port of tunnel to destroy (ignored if the all parameter is true) + * @all: flag that states to destroy all tunnels * - * This function will search the PTGs for a particular ptype, returning the - * PTG ID that contains it through the ptg parameter, with the value of - * ICE_DEFAULT_PTG (0) meaning it is part the default PTG. + * Destroys a tunnel or all tunnels by creating an update package buffer + * targeting the specific updates requested and then performing an update + * package. */ -enum ice_status -ice_ptg_find_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 *ptg) +enum ice_status ice_destroy_tunnel(struct ice_hw *hw, u16 port, bool all) { - if (ptype >= ICE_XLT1_CNT || !ptg) - return ICE_ERR_PARAM; + struct ice_boost_tcam_section *sect_rx, *sect_tx; + enum ice_status status = ICE_ERR_MAX_LIMIT; + struct ice_buf_build *bld; + u16 count = 0; + u16 index; + u16 size; + u16 i; - *ptg = hw->blk[blk].xlt1.ptypes[ptype].ptg; - return ICE_SUCCESS; + ice_acquire_lock(&hw->tnl_lock); + + if (!all && ice_tunnel_port_in_use_hlpr(hw, port, &index)) + if (hw->tnl.tbl[index].ref > 1) { + hw->tnl.tbl[index].ref--; + status = ICE_SUCCESS; + goto ice_destroy_tunnel_end; + } + + /* determine count */ + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].valid && hw->tnl.tbl[i].in_use && + (all || hw->tnl.tbl[i].port == port)) + count++; + + if (!count) { + status = ICE_ERR_PARAM; + goto ice_destroy_tunnel_end; + } + + /* size of section - there is at least one entry */ + size = ice_struct_size(sect_rx, tcam, count - 1); + + bld = ice_pkg_buf_alloc(hw); + if (!bld) { + status = ICE_ERR_NO_MEMORY; + goto ice_destroy_tunnel_end; + } + + /* allocate 2 sections, one for Rx parser, one for Tx parser */ + if (ice_pkg_buf_reserve_section(bld, 2)) + goto ice_destroy_tunnel_err; + + sect_rx = (struct ice_boost_tcam_section *) + ice_pkg_buf_alloc_section(bld, ICE_SID_RXPARSER_BOOST_TCAM, + size); + if (!sect_rx) + goto ice_destroy_tunnel_err; + sect_rx->count = CPU_TO_LE16(1); + + sect_tx = (struct ice_boost_tcam_section *) + ice_pkg_buf_alloc_section(bld, ICE_SID_TXPARSER_BOOST_TCAM, + size); + if (!sect_tx) + goto ice_destroy_tunnel_err; + sect_tx->count = CPU_TO_LE16(1); + + /* copy original boost entry to update package buffer, one copy to Rx + * section, another copy to the Tx section + */ + for (i = 0; i < hw->tnl.count && i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].valid && hw->tnl.tbl[i].in_use && + (all || hw->tnl.tbl[i].port == port)) { + ice_memcpy(sect_rx->tcam + i, + hw->tnl.tbl[i].boost_entry, + sizeof(*sect_rx->tcam), + ICE_NONDMA_TO_NONDMA); + ice_memcpy(sect_tx->tcam + i, + hw->tnl.tbl[i].boost_entry, + sizeof(*sect_tx->tcam), + ICE_NONDMA_TO_NONDMA); + hw->tnl.tbl[i].marked = true; + } + + status = ice_update_pkg(hw, ice_pkg_buf(bld), 1); + if (!status) + for (i = 0; i < hw->tnl.count && + i < ICE_TUNNEL_MAX_ENTRIES; i++) + if (hw->tnl.tbl[i].marked) { + hw->tnl.tbl[i].ref = 0; + hw->tnl.tbl[i].port = 0; + hw->tnl.tbl[i].in_use = false; + hw->tnl.tbl[i].marked = false; + } + +ice_destroy_tunnel_err: + ice_pkg_buf_free(hw, bld); + +ice_destroy_tunnel_end: + ice_release_lock(&hw->tnl_lock); + + return status; } /** - * ice_ptg_alloc_val - Allocates a new packet type group ID by value + * ice_find_prot_off - find prot ID and offset pair, based on prof and FV index * @hw: pointer to the hardware structure - * @blk: HW block - * @ptg: the ptg to allocate - * - * This function allocates a given packet type group ID specified by the ptg - * parameter. + * @blk: hardware block + * @prof: profile ID + * @fv_idx: field vector word index + * @prot: variable to receive the protocol ID + * @off: variable to receive the protocol offset */ -static -void ice_ptg_alloc_val(struct ice_hw *hw, enum ice_block blk, u8 ptg) +enum ice_status +ice_find_prot_off(struct ice_hw *hw, enum ice_block blk, u8 prof, u16 fv_idx, + u8 *prot, u16 *off) { - hw->blk[blk].xlt1.ptg_tbl[ptg].in_use = true; + struct ice_fv_word *fv_ext; + + if (prof >= hw->blk[blk].es.count) + return ICE_ERR_PARAM; + + if (fv_idx >= hw->blk[blk].es.fvw) + return ICE_ERR_PARAM; + + fv_ext = hw->blk[blk].es.t + (prof * hw->blk[blk].es.fvw); + + *prot = fv_ext[fv_idx].prot_id; + *off = fv_ext[fv_idx].off; + + return ICE_SUCCESS; } +/* PTG Management */ + /** - * ice_ptg_alloc - Find a free entry and allocates a new packet type group ID + * ice_ptg_find_ptype - Search for packet type group using packet type (ptype) * @hw: pointer to the hardware structure * @blk: HW block + * @ptype: the ptype to search for + * @ptg: pointer to variable that receives the PTG * - * This function allocates and returns a new packet type group ID. Note - * that 0 is the default packet type group, so successfully created PTGs will - * have a non-zero ID value; which means a 0 return value indicates an error. + * This function will search the PTGs for a particular ptype, returning the + * PTG ID that contains it through the PTG parameter, with the value of + * ICE_DEFAULT_PTG (0) meaning it is part the default PTG. */ -u8 ice_ptg_alloc(struct ice_hw *hw, enum ice_block blk) +static enum ice_status +ice_ptg_find_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 *ptg) { - u16 i; - - /* Skip the default PTG of 0 */ - for (i = 1; i < ICE_MAX_PTGS; i++) - if (!hw->blk[blk].xlt1.ptg_tbl[i].in_use) { - /* found a free PTG ID */ - ice_ptg_alloc_val(hw, blk, i); - return (u8)i; - } + if (ptype >= ICE_XLT1_CNT || !ptg) + return ICE_ERR_PARAM; - return 0; + *ptg = hw->blk[blk].xlt1.ptypes[ptype].ptg; + return ICE_SUCCESS; } /** - * ice_ptg_free - Frees a packet type group + * ice_ptg_alloc_val - Allocates a new packet type group ID by value * @hw: pointer to the hardware structure * @blk: HW block - * @ptg: the ptg ID to free + * @ptg: the PTG to allocate * - * This function frees a packet type group, and returns all the current ptypes - * within it to the default PTG. + * This function allocates a given packet type group ID specified by the PTG + * parameter. */ -void ice_ptg_free(struct ice_hw *hw, enum ice_block blk, u8 ptg) +static void ice_ptg_alloc_val(struct ice_hw *hw, enum ice_block blk, u8 ptg) { - struct ice_ptg_ptype *p, *temp; - - hw->blk[blk].xlt1.ptg_tbl[ptg].in_use = false; - p = hw->blk[blk].xlt1.ptg_tbl[ptg].first_ptype; - while (p) { - p->ptg = ICE_DEFAULT_PTG; - temp = p->next_ptype; - p->next_ptype = NULL; - p = temp; - } - - hw->blk[blk].xlt1.ptg_tbl[ptg].first_ptype = NULL; + hw->blk[blk].xlt1.ptg_tbl[ptg].in_use = true; } /** @@ -1905,9 +2268,9 @@ void ice_ptg_free(struct ice_hw *hw, enum ice_block blk, u8 ptg) * @hw: pointer to the hardware structure * @blk: HW block * @ptype: the ptype to remove - * @ptg: the ptg to remove the ptype from + * @ptg: the PTG to remove the ptype from * - * This function will remove the ptype from the specific ptg, and move it to + * This function will remove the ptype from the specific PTG, and move it to * the default PTG (ICE_DEFAULT_PTG). */ static enum ice_status @@ -1950,14 +2313,14 @@ ice_ptg_remove_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 ptg) * @hw: pointer to the hardware structure * @blk: HW block * @ptype: the ptype to add or move - * @ptg: the ptg to add or move the ptype to + * @ptg: the PTG to add or move the ptype to * * This function will either add or move a ptype to a particular PTG depending * on if the ptype is already part of another group. Note that using a * a destination PTG ID of ICE_DEFAULT_PTG (0) will move the ptype to the * default PTG. */ -enum ice_status +static enum ice_status ice_ptg_add_mv_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 ptg) { enum ice_status status; @@ -2003,7 +2366,7 @@ struct ice_blk_size_details { u16 xlt2; /* # XLT2 entries */ u16 prof_tcam; /* # profile ID TCAM entries */ u16 prof_id; /* # profile IDs */ - u8 prof_cdid_bits; /* # cdid one-hot bits used in key */ + u8 prof_cdid_bits; /* # CDID one-hot bits used in key */ u16 prof_redir; /* # profile redirection entries */ u16 es; /* # extraction sequence entries */ u16 fvw; /* # field vector words */ @@ -2065,12 +2428,10 @@ ice_match_prop_lst(struct LIST_HEAD_TYPE *list1, struct LIST_HEAD_TYPE *list2) u16 count = 0; /* compare counts */ - LIST_FOR_EACH_ENTRY(tmp1, list1, ice_vsig_prof, list) { + LIST_FOR_EACH_ENTRY(tmp1, list1, ice_vsig_prof, list) count++; - } - LIST_FOR_EACH_ENTRY(tmp2, list2, ice_vsig_prof, list) { + LIST_FOR_EACH_ENTRY(tmp2, list2, ice_vsig_prof, list) chk_count++; - } if (!count || count != chk_count) return false; @@ -2093,71 +2454,6 @@ ice_match_prop_lst(struct LIST_HEAD_TYPE *list1, struct LIST_HEAD_TYPE *list2) /* VSIG Management */ -/** - * ice_vsig_update_xlt2_sect - update one section of XLT2 table - * @hw: pointer to the hardware structure - * @blk: HW block - * @vsi: HW VSI number to program - * @vsig: vsig for the VSI - * - * This function will update the XLT2 hardware table with the input VSI - * group configuration. - */ -static enum ice_status -ice_vsig_update_xlt2_sect(struct ice_hw *hw, enum ice_block blk, u16 vsi, - u16 vsig) -{ - struct ice_xlt2_section *sect; - struct ice_buf_build *bld; - enum ice_status status; - - bld = ice_pkg_buf_alloc_single_section(hw, ice_sect_id(blk, ICE_XLT2), - sizeof(struct ice_xlt2_section), - (void **)§); - if (!bld) - return ICE_ERR_NO_MEMORY; - - sect->count = CPU_TO_LE16(1); - sect->offset = CPU_TO_LE16(vsi); - sect->value[0] = CPU_TO_LE16(vsig); - - status = ice_update_pkg(hw, ice_pkg_buf(bld), 1); - - ice_pkg_buf_free(hw, bld); - - return status; -} - -/** - * ice_vsig_update_xlt2 - update XLT2 table with VSIG configuration - * @hw: pointer to the hardware structure - * @blk: HW block - * - * This function will update the XLT2 hardware table with the input VSI - * group configuration of used vsis. - */ -enum ice_status ice_vsig_update_xlt2(struct ice_hw *hw, enum ice_block blk) -{ - u16 vsi; - - for (vsi = 0; vsi < ICE_MAX_VSI; vsi++) { - /* update only vsis that have been changed */ - if (hw->blk[blk].xlt2.vsis[vsi].changed) { - enum ice_status status; - u16 vsig; - - vsig = hw->blk[blk].xlt2.vsis[vsi].vsig; - status = ice_vsig_update_xlt2_sect(hw, blk, vsi, vsig); - if (status) - return status; - - hw->blk[blk].xlt2.vsis[vsi].changed = 0; - } - } - - return ICE_SUCCESS; -} - /** * ice_vsig_find_vsi - find a VSIG that contains a specified VSI * @hw: pointer to the hardware structure @@ -2187,9 +2483,9 @@ ice_vsig_find_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 *vsig) * ice_vsig_alloc_val - allocate a new VSIG by value * @hw: pointer to the hardware structure * @blk: HW block - * @vsig: the vsig to allocate + * @vsig: the VSIG to allocate * - * This function will allocate a given VSIG specified by the vsig parameter. + * This function will allocate a given VSIG specified by the VSIG parameter. */ static u16 ice_vsig_alloc_val(struct ice_hw *hw, enum ice_block blk, u16 vsig) { @@ -2237,20 +2533,19 @@ static u16 ice_vsig_alloc(struct ice_hw *hw, enum ice_block blk) * for, the list must match exactly, including the order in which the * characteristics are listed. */ -enum ice_status +static enum ice_status ice_find_dup_props_vsig(struct ice_hw *hw, enum ice_block blk, struct LIST_HEAD_TYPE *chs, u16 *vsig) { struct ice_xlt2 *xlt2 = &hw->blk[blk].xlt2; u16 i; - for (i = 0; i < xlt2->count; i++) { + for (i = 0; i < xlt2->count; i++) if (xlt2->vsig_tbl[i].in_use && ice_match_prop_lst(chs, &xlt2->vsig_tbl[i].prop_lst)) { *vsig = ICE_VSIG_VALUE(i, hw->pf_id); return ICE_SUCCESS; } - } return ICE_ERR_DOES_NOT_EXIST; } @@ -2264,7 +2559,7 @@ ice_find_dup_props_vsig(struct ice_hw *hw, enum ice_block blk, * The function will remove all VSIs associated with the input VSIG and move * them to the DEFAULT_VSIG and mark the VSIG available. */ -enum ice_status +static enum ice_status ice_vsig_free(struct ice_hw *hw, enum ice_block blk, u16 vsig) { struct ice_vsig_prof *dtmp, *del; @@ -2281,21 +2576,23 @@ ice_vsig_free(struct ice_hw *hw, enum ice_block blk, u16 vsig) hw->blk[blk].xlt2.vsig_tbl[idx].in_use = false; vsi_cur = hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi; - if (!vsi_cur) - return ICE_ERR_CFG; - - /* remove all vsis associated with this VSIG XLT2 entry */ - do { - struct ice_vsig_vsi *tmp = vsi_cur->next_vsi; - - vsi_cur->vsig = ICE_DEFAULT_VSIG; - vsi_cur->changed = 1; - vsi_cur->next_vsi = NULL; - vsi_cur = tmp; - } while (vsi_cur); - - /* NULL terminate head of vsi list */ - hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi = NULL; + /* If the VSIG has at least 1 VSI then iterate through the + * list and remove the VSIs before deleting the group. + */ + if (vsi_cur) { + /* remove all vsis associated with this VSIG XLT2 entry */ + do { + struct ice_vsig_vsi *tmp = vsi_cur->next_vsi; + + vsi_cur->vsig = ICE_DEFAULT_VSIG; + vsi_cur->changed = 1; + vsi_cur->next_vsi = NULL; + vsi_cur = tmp; + } while (vsi_cur); + + /* NULL terminate head of VSI list */ + hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi = NULL; + } /* free characteristic list */ LIST_FOR_EACH_ENTRY_SAFE(del, dtmp, @@ -2305,38 +2602,99 @@ ice_vsig_free(struct ice_hw *hw, enum ice_block blk, u16 vsig) ice_free(hw, del); } + /* if VSIG characteristic list was cleared for reset + * re-initialize the list head + */ + INIT_LIST_HEAD(&hw->blk[blk].xlt2.vsig_tbl[idx].prop_lst); + return ICE_SUCCESS; } /** - * ice_vsig_add_mv_vsi - add or move a VSI to a VSI group + * ice_vsig_remove_vsi - remove VSI from VSIG * @hw: pointer to the hardware structure * @blk: HW block - * @vsi: VSI to move - * @vsig: destination VSI group + * @vsi: VSI to remove + * @vsig: VSI group to remove from * - * This function will move or add the input VSI to the target VSIG. - * The function will find the original VSIG the VSI belongs to and - * move the entry to the DEFAULT_VSIG, update the original VSIG and - * then move entry to the new VSIG. + * The function will remove the input VSI from its VSI group and move it + * to the DEFAULT_VSIG. */ -enum ice_status -ice_vsig_add_mv_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) +static enum ice_status +ice_vsig_remove_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) { - struct ice_vsig_vsi *tmp; - enum ice_status status; - u16 orig_vsig, idx; + struct ice_vsig_vsi **vsi_head, *vsi_cur, *vsi_tgt; + u16 idx; idx = vsig & ICE_VSIG_IDX_M; if (vsi >= ICE_MAX_VSI || idx >= ICE_MAX_VSIGS) return ICE_ERR_PARAM; - /* if VSIG not in use and VSIG is not default type this VSIG - * doesn't exist. - */ - if (!hw->blk[blk].xlt2.vsig_tbl[idx].in_use && - vsig != ICE_DEFAULT_VSIG) + if (!hw->blk[blk].xlt2.vsig_tbl[idx].in_use) + return ICE_ERR_DOES_NOT_EXIST; + + /* entry already in default VSIG, don't have to remove */ + if (idx == ICE_DEFAULT_VSIG) + return ICE_SUCCESS; + + vsi_head = &hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi; + if (!(*vsi_head)) + return ICE_ERR_CFG; + + vsi_tgt = &hw->blk[blk].xlt2.vsis[vsi]; + vsi_cur = (*vsi_head); + + /* iterate the VSI list, skip over the entry to be removed */ + while (vsi_cur) { + if (vsi_tgt == vsi_cur) { + (*vsi_head) = vsi_cur->next_vsi; + break; + } + vsi_head = &vsi_cur->next_vsi; + vsi_cur = vsi_cur->next_vsi; + } + + /* verify if VSI was removed from group list */ + if (!vsi_cur) + return ICE_ERR_DOES_NOT_EXIST; + + vsi_cur->vsig = ICE_DEFAULT_VSIG; + vsi_cur->changed = 1; + vsi_cur->next_vsi = NULL; + + return ICE_SUCCESS; +} + +/** + * ice_vsig_add_mv_vsi - add or move a VSI to a VSI group + * @hw: pointer to the hardware structure + * @blk: HW block + * @vsi: VSI to move + * @vsig: destination VSI group + * + * This function will move or add the input VSI to the target VSIG. + * The function will find the original VSIG the VSI belongs to and + * move the entry to the DEFAULT_VSIG, update the original VSIG and + * then move entry to the new VSIG. + */ +static enum ice_status +ice_vsig_add_mv_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) +{ + struct ice_vsig_vsi *tmp; + enum ice_status status; + u16 orig_vsig, idx; + + idx = vsig & ICE_VSIG_IDX_M; + + if (vsi >= ICE_MAX_VSI || idx >= ICE_MAX_VSIGS) + return ICE_ERR_PARAM; + + /* if VSIG not in use and VSIG is not default type this VSIG + * doesn't exist. + */ + if (!hw->blk[blk].xlt2.vsig_tbl[idx].in_use && + vsig != ICE_DEFAULT_VSIG) return ICE_ERR_DOES_NOT_EXIST; status = ice_vsig_find_vsi(hw, blk, vsi, &orig_vsig); @@ -2372,79 +2730,98 @@ ice_vsig_add_mv_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) } /** - * ice_vsig_remove_vsi - remove VSI from VSIG + * ice_prof_has_mask_idx - determine if profile index masking is identical * @hw: pointer to the hardware structure * @blk: HW block - * @vsi: VSI to remove - * @vsig: VSI group to remove from - * - * The function will remove the input VSI from its VSI group and move it - * to the DEFAULT_VSIG. + * @prof: profile to check + * @idx: profile index to check + * @mask: mask to match */ -enum ice_status -ice_vsig_remove_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) +static bool +ice_prof_has_mask_idx(struct ice_hw *hw, enum ice_block blk, u8 prof, u16 idx, + u16 mask) { - struct ice_vsig_vsi **vsi_head, *vsi_cur, *vsi_tgt; - u16 idx; - - idx = vsig & ICE_VSIG_IDX_M; - - if (vsi >= ICE_MAX_VSI || idx >= ICE_MAX_VSIGS) - return ICE_ERR_PARAM; - - if (!hw->blk[blk].xlt2.vsig_tbl[idx].in_use) - return ICE_ERR_DOES_NOT_EXIST; - - /* entry already in default VSIG, don't have to remove */ - if (idx == ICE_DEFAULT_VSIG) - return ICE_SUCCESS; - - vsi_head = &hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi; - if (!(*vsi_head)) - return ICE_ERR_CFG; + bool expect_no_mask = false; + bool found = false; + bool match = false; + u16 i; - vsi_tgt = &hw->blk[blk].xlt2.vsis[vsi]; - vsi_cur = (*vsi_head); + /* If mask is 0x0000 or 0xffff, then there is no masking */ + if (mask == 0 || mask == 0xffff) + expect_no_mask = true; + + /* Scan the enabled masks on this profile, for the specified idx */ + for (i = hw->blk[blk].masks.first; i < hw->blk[blk].masks.first + + hw->blk[blk].masks.count; i++) + if (hw->blk[blk].es.mask_ena[prof] & BIT(i)) + if (hw->blk[blk].masks.masks[i].in_use && + hw->blk[blk].masks.masks[i].idx == idx) { + found = true; + if (hw->blk[blk].masks.masks[i].mask == mask) + match = true; + break; + } - /* iterate the VSI list, skip over the entry to be removed */ - while (vsi_cur) { - if (vsi_tgt == vsi_cur) { - (*vsi_head) = vsi_cur->next_vsi; - break; - } - vsi_head = &vsi_cur->next_vsi; - vsi_cur = vsi_cur->next_vsi; + if (expect_no_mask) { + if (found) + return false; + } else { + if (!match) + return false; } - /* verify if VSI was removed from group list */ - if (!vsi_cur) - return ICE_ERR_DOES_NOT_EXIST; + return true; +} - vsi_cur->vsig = ICE_DEFAULT_VSIG; - vsi_cur->changed = 1; - vsi_cur->next_vsi = NULL; +/** + * ice_prof_has_mask - determine if profile masking is identical + * @hw: pointer to the hardware structure + * @blk: HW block + * @prof: profile to check + * @masks: masks to match + */ +static bool +ice_prof_has_mask(struct ice_hw *hw, enum ice_block blk, u8 prof, u16 *masks) +{ + u16 i; - return ICE_SUCCESS; + /* es->mask_ena[prof] will have the mask */ + for (i = 0; i < hw->blk[blk].es.fvw; i++) + if (!ice_prof_has_mask_idx(hw, blk, prof, i, masks[i])) + return false; + + return true; } /** - * ice_find_prof_id - find profile ID for a given field vector + * ice_find_prof_id_with_mask - find profile ID for a given field vector * @hw: pointer to the hardware structure * @blk: HW block * @fv: field vector to search for + * @masks: masks for fv * @prof_id: receives the profile ID */ static enum ice_status -ice_find_prof_id(struct ice_hw *hw, enum ice_block blk, - struct ice_fv_word *fv, u8 *prof_id) +ice_find_prof_id_with_mask(struct ice_hw *hw, enum ice_block blk, + struct ice_fv_word *fv, u16 *masks, u8 *prof_id) { struct ice_es *es = &hw->blk[blk].es; - u16 off, i; + u8 i; - for (i = 0; i < es->count; i++) { - off = i * es->fvw; + /* For FD and RSS, we don't want to re-use an existed profile with the + * same field vector and mask. This will cause rule interference. + */ + if (blk == ICE_BLK_FD || blk == ICE_BLK_RSS) + return ICE_ERR_DOES_NOT_EXIST; + + for (i = 0; i < (u8)es->count; i++) { + u16 off = i * es->fvw; - if (memcmp(&es->t[off], fv, es->fvw * 2)) + if (memcmp(&es->t[off], fv, es->fvw * sizeof(*fv))) + continue; + + /* check if masks settings are the same for this profile */ + if (masks && !ice_prof_has_mask(hw, blk, i, masks)) continue; *prof_id = i; @@ -2512,217 +2889,26 @@ static bool ice_tcam_ent_rsrc_type(enum ice_block blk, u16 *rsrc_type) return true; } -/** - * ice_workaround_get_res_blk - determine the block from a resource type - * @type: type of resource - * @blk: pointer to a enum that will receive the block type - * @tcam: pointer to variable that will be set to true for a TCAM resource type - */ -static enum -ice_status ice_workaround_get_res_blk(u16 type, enum ice_block *blk, bool *tcam) -{ - /* just need to support TCAM entries and Profile IDs for now */ - *tcam = false; - - switch (type) { - case ICE_AQC_RES_TYPE_SWITCH_PROF_BLDR_TCAM: - *blk = ICE_BLK_SW; - *tcam = true; - break; - case ICE_AQC_RES_TYPE_ACL_PROF_BLDR_TCAM: - *blk = ICE_BLK_ACL; - *tcam = true; - break; - case ICE_AQC_RES_TYPE_FD_PROF_BLDR_TCAM: - *blk = ICE_BLK_FD; - *tcam = true; - break; - case ICE_AQC_RES_TYPE_HASH_PROF_BLDR_TCAM: - *blk = ICE_BLK_RSS; - *tcam = true; - break; - case ICE_AQC_RES_TYPE_QHASH_PROF_BLDR_TCAM: - *blk = ICE_BLK_PE; - *tcam = true; - break; - case ICE_AQC_RES_TYPE_SWITCH_PROF_BLDR_PROFID: - *blk = ICE_BLK_SW; - break; - case ICE_AQC_RES_TYPE_ACL_PROF_BLDR_PROFID: - *blk = ICE_BLK_ACL; - break; - case ICE_AQC_RES_TYPE_FD_PROF_BLDR_PROFID: - *blk = ICE_BLK_FD; - break; - case ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID: - *blk = ICE_BLK_RSS; - break; - case ICE_AQC_RES_TYPE_QHASH_PROF_BLDR_PROFID: - *blk = ICE_BLK_PE; - break; - default: - return ICE_ERR_PARAM; - } - - return ICE_SUCCESS; -} - -/** - * ice_alloc_res_workaround - * @hw: pointer to the hw struct - * @type: type of resource - * @num: number of resources to allocate - * @res: pointer to array that will receive the resources - */ -static enum ice_status -ice_alloc_res_workaround(struct ice_hw *hw, u16 type, u16 num, u16 *res) -{ - enum ice_block blk; - u16 count = 0; - bool tcam; - u16 first; - u16 last; - u16 max; - u16 i; - -/* Number of PFs we support with this workaround */ -#define ICE_WA_PF_COUNT 4 -#define ICE_WA_1ST_TCAM 4 -#define ICE_WA_1ST_FV 4 - - /* Only allow our supported PFs */ - if (hw->pf_id >= ICE_WA_PF_COUNT) - return ICE_ERR_AQ_ERROR; - - if (ice_workaround_get_res_blk(type, &blk, &tcam)) - return ICE_ERR_AQ_ERROR; - - if (tcam) { - /* range of entries based on PF */ - max = hw->blk[blk].prof.count / ICE_WA_PF_COUNT; - first = max * hw->pf_id; - last = first + max; - - /* Profile IDs - start at non-zero index for PROF ID TCAM table - * The first few entries are for bypass, default and errors - * (only relevant for PF 0) - */ - first += hw->pf_id ? 0 : ICE_WA_1ST_TCAM; - - for (i = first; i < last && count < num; i++) { - if (!hw->blk[blk].prof.resource_used_hack[i]) { - res[count++] = i; - hw->blk[blk].prof.resource_used_hack[i] = true; - } - } - - /* handle failure case */ - if (count < num) { - for (i = 0; i < count; i++) { - hw->blk[blk].prof.resource_used_hack[res[i]] = - false; - res[i] = 0; - } - - return ICE_ERR_AQ_ERROR; - } - } else { - /* range of entries based on PF */ - max = hw->blk[blk].es.count / ICE_WA_PF_COUNT; - first = max * hw->pf_id; - last = first + max; - - /* FV index - start at non-zero index for Field vector table - * The first few entries are for bypass, default and errors - * (only relevant for PF 0) - */ - first += hw->pf_id ? 0 : ICE_WA_1ST_FV; - - for (i = first; i < last && count < num; i++) { - if (!hw->blk[blk].es.resource_used_hack[i]) { - res[count++] = i; - hw->blk[blk].es.resource_used_hack[i] = true; - } - } - - /* handle failure case */ - if (count < num) { - for (i = 0; i < count; i++) { - hw->blk[blk].es.resource_used_hack[res[i]] = - false; - res[i] = 0; - } - - return ICE_ERR_AQ_ERROR; - } - } - - return ICE_SUCCESS; -} - -/** - * ice_free_res_workaround - * @hw: pointer to the hw struct - * @type: type of resource to free - * @num: number of resources - * @res: array of resource ids to free - */ -static enum ice_status -ice_free_res_workaround(struct ice_hw *hw, u16 type, u16 num, u16 *res) -{ - enum ice_block blk; - bool tcam = false; - u16 i; - - if (ice_workaround_get_res_blk(type, &blk, &tcam)) - return ICE_ERR_AQ_ERROR; - - if (tcam) { - /* TCAM entries */ - for (i = 0; i < num; i++) { - if (res[i] < hw->blk[blk].prof.count) { - u16 idx = res[i]; - - ice_free_hw_res(hw, type, 1, &idx); - hw->blk[blk].prof.resource_used_hack[res[i]] = - false; - } - } - - } else { - /* Profile IDs */ - for (i = 0; i < num; i++) { - if (res[i] < hw->blk[blk].es.count) { - u16 idx = res[i]; - - ice_free_hw_res(hw, type, 1, &idx); - hw->blk[blk].es.resource_used_hack[res[i]] = - false; - } - } - } - - return ICE_SUCCESS; -} - /** * ice_alloc_tcam_ent - allocate hardware TCAM entry * @hw: pointer to the HW struct * @blk: the block to allocate the TCAM for + * @btm: true to allocate from bottom of table, false to allocate from top * @tcam_idx: pointer to variable to receive the TCAM entry * * This function allocates a new entry in a Profile ID TCAM for a specific * block. */ static enum ice_status -ice_alloc_tcam_ent(struct ice_hw *hw, enum ice_block blk, u16 *tcam_idx) +ice_alloc_tcam_ent(struct ice_hw *hw, enum ice_block blk, bool btm, + u16 *tcam_idx) { u16 res_type; if (!ice_tcam_ent_rsrc_type(blk, &res_type)) return ICE_ERR_PARAM; - return ice_alloc_res_workaround(hw, res_type, 1, tcam_idx); + return ice_alloc_hw_res(hw, res_type, 1, btm, tcam_idx); } /** @@ -2741,7 +2927,7 @@ ice_free_tcam_ent(struct ice_hw *hw, enum ice_block blk, u16 tcam_idx) if (!ice_tcam_ent_rsrc_type(blk, &res_type)) return ICE_ERR_PARAM; - return ice_free_res_workaround(hw, res_type, 1, &tcam_idx); + return ice_free_hw_res(hw, res_type, 1, &tcam_idx); } /** @@ -2763,7 +2949,7 @@ ice_alloc_prof_id(struct ice_hw *hw, enum ice_block blk, u8 *prof_id) if (!ice_prof_id_rsrc_type(blk, &res_type)) return ICE_ERR_PARAM; - status = ice_alloc_res_workaround(hw, res_type, 1, &get_prof); + status = ice_alloc_hw_res(hw, res_type, 1, false, &get_prof); if (!status) *prof_id = (u8)get_prof; @@ -2787,15 +2973,7 @@ ice_free_prof_id(struct ice_hw *hw, enum ice_block blk, u8 prof_id) if (!ice_prof_id_rsrc_type(blk, &res_type)) return ICE_ERR_PARAM; - return ice_free_res_workaround(hw, res_type, 1, &tmp_prof_id); - /* The following code is a WORKAROUND until DCR 076 is available. - * DCR 076 - Update to Profile ID TCAM Resource Allocation - * - * Once the DCR 076 changes are available in FW, this code can be - * restored. Original code: - * - * return ice_free_res(hw, res_type, 1, &tmp_prof_id); - */ + return ice_free_hw_res(hw, res_type, 1, &tmp_prof_id); } /** @@ -2816,22 +2994,330 @@ ice_prof_inc_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) } /** - * ice_prof_dec_ref - decrement reference count for profile + * ice_write_prof_mask_reg - write profile mask register * @hw: pointer to the HW struct - * @blk: the block from which to free the profile ID - * @prof_id: the profile ID for which to decrement the reference count + * @blk: hardware block + * @mask_idx: mask index + * @idx: index of the FV which will use the mask + * @mask: the 16-bit mask + */ +static void +ice_write_prof_mask_reg(struct ice_hw *hw, enum ice_block blk, u16 mask_idx, + u16 idx, u16 mask) +{ + u32 offset; + u32 val; + + switch (blk) { + case ICE_BLK_RSS: + offset = GLQF_HMASK(mask_idx); + val = (idx << GLQF_HMASK_MSK_INDEX_S) & + GLQF_HMASK_MSK_INDEX_M; + val |= (mask << GLQF_HMASK_MASK_S) & GLQF_HMASK_MASK_M; + break; + case ICE_BLK_FD: + offset = GLQF_FDMASK(mask_idx); + val = (idx << GLQF_FDMASK_MSK_INDEX_S) & + GLQF_FDMASK_MSK_INDEX_M; + val |= (mask << GLQF_FDMASK_MASK_S) & + GLQF_FDMASK_MASK_M; + break; + default: + ice_debug(hw, ICE_DBG_PKG, "No profile masks for block %d\n", + blk); + return; + } + + wr32(hw, offset, val); + ice_debug(hw, ICE_DBG_PKG, "write mask, blk %d (%d): %x = %x\n", + blk, idx, offset, val); +} + +/** + * ice_write_prof_mask_enable_res - write profile mask enable register + * @hw: pointer to the HW struct + * @blk: hardware block + * @prof_id: profile ID + * @enable_mask: enable mask + */ +static void +ice_write_prof_mask_enable_res(struct ice_hw *hw, enum ice_block blk, + u16 prof_id, u32 enable_mask) +{ + u32 offset; + + switch (blk) { + case ICE_BLK_RSS: + offset = GLQF_HMASK_SEL(prof_id); + break; + case ICE_BLK_FD: + offset = GLQF_FDMASK_SEL(prof_id); + break; + default: + ice_debug(hw, ICE_DBG_PKG, "No profile masks for block %d\n", + blk); + return; + } + + wr32(hw, offset, enable_mask); + ice_debug(hw, ICE_DBG_PKG, "write mask enable, blk %d (%d): %x = %x\n", + blk, prof_id, offset, enable_mask); +} + +/** + * ice_init_prof_masks - initial prof masks + * @hw: pointer to the HW struct + * @blk: hardware block + */ +static void ice_init_prof_masks(struct ice_hw *hw, enum ice_block blk) +{ + u16 per_pf; + u16 i; + + ice_init_lock(&hw->blk[blk].masks.lock); + + per_pf = ICE_PROF_MASK_COUNT / hw->dev_caps.num_funcs; + + hw->blk[blk].masks.count = per_pf; + hw->blk[blk].masks.first = hw->pf_id * per_pf; + + ice_memset(hw->blk[blk].masks.masks, 0, + sizeof(hw->blk[blk].masks.masks), ICE_NONDMA_MEM); + + for (i = hw->blk[blk].masks.first; + i < hw->blk[blk].masks.first + hw->blk[blk].masks.count; i++) + ice_write_prof_mask_reg(hw, blk, i, 0, 0); +} + +/** + * ice_init_all_prof_masks - initial all prof masks + * @hw: pointer to the HW struct + */ +void ice_init_all_prof_masks(struct ice_hw *hw) +{ + ice_init_prof_masks(hw, ICE_BLK_RSS); + ice_init_prof_masks(hw, ICE_BLK_FD); +} + +/** + * ice_alloc_prof_mask - allocate profile mask + * @hw: pointer to the HW struct + * @blk: hardware block + * @idx: index of FV which will use the mask + * @mask: the 16-bit mask + * @mask_idx: variable to receive the mask index */ static enum ice_status -ice_prof_dec_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) +ice_alloc_prof_mask(struct ice_hw *hw, enum ice_block blk, u16 idx, u16 mask, + u16 *mask_idx) { - if (prof_id > hw->blk[blk].es.count) + bool found_unused = false, found_copy = false; + enum ice_status status = ICE_ERR_MAX_LIMIT; + u16 unused_idx = 0, copy_idx = 0; + u16 i; + + if (blk != ICE_BLK_RSS && blk != ICE_BLK_FD) return ICE_ERR_PARAM; - if (hw->blk[blk].es.ref_count[prof_id] > 0) { - if (!--hw->blk[blk].es.ref_count[prof_id]) - return ice_free_prof_id(hw, blk, prof_id); + ice_acquire_lock(&hw->blk[blk].masks.lock); + + for (i = hw->blk[blk].masks.first; + i < hw->blk[blk].masks.first + hw->blk[blk].masks.count; i++) + if (hw->blk[blk].masks.masks[i].in_use) { + /* if mask is in use and it exactly duplicates the + * desired mask and index, then in can be reused + */ + if (hw->blk[blk].masks.masks[i].mask == mask && + hw->blk[blk].masks.masks[i].idx == idx) { + found_copy = true; + copy_idx = i; + break; + } + } else { + /* save off unused index, but keep searching in case + * there is an exact match later on + */ + if (!found_unused) { + found_unused = true; + unused_idx = i; + } + } + + if (found_copy) + i = copy_idx; + else if (found_unused) + i = unused_idx; + else + goto err_ice_alloc_prof_mask; + + /* update mask for a new entry */ + if (found_unused) { + hw->blk[blk].masks.masks[i].in_use = true; + hw->blk[blk].masks.masks[i].mask = mask; + hw->blk[blk].masks.masks[i].idx = idx; + hw->blk[blk].masks.masks[i].ref = 0; + ice_write_prof_mask_reg(hw, blk, i, idx, mask); + } + + hw->blk[blk].masks.masks[i].ref++; + *mask_idx = i; + status = ICE_SUCCESS; + +err_ice_alloc_prof_mask: + ice_release_lock(&hw->blk[blk].masks.lock); + + return status; +} + +/** + * ice_free_prof_mask - free profile mask + * @hw: pointer to the HW struct + * @blk: hardware block + * @mask_idx: index of mask + */ +static enum ice_status +ice_free_prof_mask(struct ice_hw *hw, enum ice_block blk, u16 mask_idx) +{ + if (blk != ICE_BLK_RSS && blk != ICE_BLK_FD) + return ICE_ERR_PARAM; + + if (!(mask_idx >= hw->blk[blk].masks.first && + mask_idx < hw->blk[blk].masks.first + hw->blk[blk].masks.count)) + return ICE_ERR_DOES_NOT_EXIST; + + ice_acquire_lock(&hw->blk[blk].masks.lock); + + if (!hw->blk[blk].masks.masks[mask_idx].in_use) + goto exit_ice_free_prof_mask; + + if (hw->blk[blk].masks.masks[mask_idx].ref > 1) { + hw->blk[blk].masks.masks[mask_idx].ref--; + goto exit_ice_free_prof_mask; } + /* remove mask */ + hw->blk[blk].masks.masks[mask_idx].in_use = false; + hw->blk[blk].masks.masks[mask_idx].mask = 0; + hw->blk[blk].masks.masks[mask_idx].idx = 0; + + /* update mask as unused entry */ + ice_debug(hw, ICE_DBG_PKG, "Free mask, blk %d, mask %d\n", blk, + mask_idx); + ice_write_prof_mask_reg(hw, blk, mask_idx, 0, 0); + +exit_ice_free_prof_mask: + ice_release_lock(&hw->blk[blk].masks.lock); + + return ICE_SUCCESS; +} + +/** + * ice_free_prof_masks - free all profile masks for a profile + * @hw: pointer to the HW struct + * @blk: hardware block + * @prof_id: profile ID + */ +static enum ice_status +ice_free_prof_masks(struct ice_hw *hw, enum ice_block blk, u16 prof_id) +{ + u32 mask_bm; + u16 i; + + if (blk != ICE_BLK_RSS && blk != ICE_BLK_FD) + return ICE_ERR_PARAM; + + mask_bm = hw->blk[blk].es.mask_ena[prof_id]; + for (i = 0; i < BITS_PER_BYTE * sizeof(mask_bm); i++) + if (mask_bm & BIT(i)) + ice_free_prof_mask(hw, blk, i); + + return ICE_SUCCESS; +} + +/** + * ice_shutdown_prof_masks - releases lock for masking + * @hw: pointer to the HW struct + * @blk: hardware block + * + * This should be called before unloading the driver + */ +static void ice_shutdown_prof_masks(struct ice_hw *hw, enum ice_block blk) +{ + u16 i; + + ice_acquire_lock(&hw->blk[blk].masks.lock); + + for (i = hw->blk[blk].masks.first; + i < hw->blk[blk].masks.first + hw->blk[blk].masks.count; i++) { + ice_write_prof_mask_reg(hw, blk, i, 0, 0); + + hw->blk[blk].masks.masks[i].in_use = false; + hw->blk[blk].masks.masks[i].idx = 0; + hw->blk[blk].masks.masks[i].mask = 0; + } + + ice_release_lock(&hw->blk[blk].masks.lock); + ice_destroy_lock(&hw->blk[blk].masks.lock); +} + +/** + * ice_shutdown_all_prof_masks - releases all locks for masking + * @hw: pointer to the HW struct + * + * This should be called before unloading the driver + */ +void ice_shutdown_all_prof_masks(struct ice_hw *hw) +{ + ice_shutdown_prof_masks(hw, ICE_BLK_RSS); + ice_shutdown_prof_masks(hw, ICE_BLK_FD); +} + +/** + * ice_update_prof_masking - set registers according to masking + * @hw: pointer to the HW struct + * @blk: hardware block + * @prof_id: profile ID + * @masks: masks + */ +static enum ice_status +ice_update_prof_masking(struct ice_hw *hw, enum ice_block blk, u16 prof_id, + u16 *masks) +{ + bool err = false; + u32 ena_mask = 0; + u16 idx; + u16 i; + + /* Only support FD and RSS masking, otherwise nothing to be done */ + if (blk != ICE_BLK_RSS && blk != ICE_BLK_FD) + return ICE_SUCCESS; + + for (i = 0; i < hw->blk[blk].es.fvw; i++) + if (masks[i] && masks[i] != 0xFFFF) { + if (!ice_alloc_prof_mask(hw, blk, i, masks[i], &idx)) { + ena_mask |= BIT(idx); + } else { + /* not enough bitmaps */ + err = true; + break; + } + } + + if (err) { + /* free any bitmaps we have allocated */ + for (i = 0; i < BITS_PER_BYTE * sizeof(ena_mask); i++) + if (ena_mask & BIT(i)) + ice_free_prof_mask(hw, blk, i); + + return ICE_ERR_OUT_OF_RANGE; + } + + /* enable the masks for this profile */ + ice_write_prof_mask_enable_res(hw, blk, prof_id, ena_mask); + + /* store enabled masks with profile so that they can be freed later */ + hw->blk[blk].es.mask_ena[prof_id] = ena_mask; + return ICE_SUCCESS; } @@ -2840,7 +3326,7 @@ ice_prof_dec_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) * @hw: pointer to the HW struct * @blk: the block in which to write the extraction sequence * @prof_id: the profile ID to write - * @fv: pointer to the extraction sequence to write + * @fv: pointer to the extraction sequence to write - NULL to clear extraction */ static void ice_write_es(struct ice_hw *hw, enum ice_block blk, u8 prof_id, @@ -2849,8 +3335,37 @@ ice_write_es(struct ice_hw *hw, enum ice_block blk, u8 prof_id, u16 off; off = prof_id * hw->blk[blk].es.fvw; - ice_memcpy(&hw->blk[blk].es.t[off], fv, hw->blk[blk].es.fvw * 2, - ICE_NONDMA_TO_NONDMA); + if (!fv) { + ice_memset(&hw->blk[blk].es.t[off], 0, hw->blk[blk].es.fvw * + sizeof(*fv), ICE_NONDMA_MEM); + hw->blk[blk].es.written[prof_id] = false; + } else { + ice_memcpy(&hw->blk[blk].es.t[off], fv, hw->blk[blk].es.fvw * + sizeof(*fv), ICE_NONDMA_TO_NONDMA); + } +} + +/** + * ice_prof_dec_ref - decrement reference count for profile + * @hw: pointer to the HW struct + * @blk: the block from which to free the profile ID + * @prof_id: the profile ID for which to decrement the reference count + */ +static enum ice_status +ice_prof_dec_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) +{ + if (prof_id > hw->blk[blk].es.count) + return ICE_ERR_PARAM; + + if (hw->blk[blk].es.ref_count[prof_id] > 0) { + if (!--hw->blk[blk].es.ref_count[prof_id]) { + ice_write_es(hw, blk, prof_id, NULL); + ice_free_prof_masks(hw, blk, prof_id); + return ice_free_prof_id(hw, blk, prof_id); + } + } + + return ICE_SUCCESS; } /* Block / table section IDs */ @@ -2896,6 +3411,64 @@ static const u32 ice_blk_sids[ICE_BLK_COUNT][ICE_SID_OFF_COUNT] = { } }; +/** + * ice_init_sw_xlt1_db - init software XLT1 database from HW tables + * @hw: pointer to the hardware structure + * @blk: the HW block to initialize + */ +static void ice_init_sw_xlt1_db(struct ice_hw *hw, enum ice_block blk) +{ + u16 pt; + + for (pt = 0; pt < hw->blk[blk].xlt1.count; pt++) { + u8 ptg; + + ptg = hw->blk[blk].xlt1.t[pt]; + if (ptg != ICE_DEFAULT_PTG) { + ice_ptg_alloc_val(hw, blk, ptg); + ice_ptg_add_mv_ptype(hw, blk, pt, ptg); + } + } +} + +/** + * ice_init_sw_xlt2_db - init software XLT2 database from HW tables + * @hw: pointer to the hardware structure + * @blk: the HW block to initialize + */ +static void ice_init_sw_xlt2_db(struct ice_hw *hw, enum ice_block blk) +{ + u16 vsi; + + for (vsi = 0; vsi < hw->blk[blk].xlt2.count; vsi++) { + u16 vsig; + + vsig = hw->blk[blk].xlt2.t[vsi]; + if (vsig) { + ice_vsig_alloc_val(hw, blk, vsig); + ice_vsig_add_mv_vsi(hw, blk, vsi, vsig); + /* no changes at this time, since this has been + * initialized from the original package + */ + hw->blk[blk].xlt2.vsis[vsi].changed = 0; + } + } +} + +/** + * ice_init_sw_db - init software database from HW tables + * @hw: pointer to the hardware structure + */ +static void ice_init_sw_db(struct ice_hw *hw) +{ + u16 i; + + for (i = 0; i < ICE_BLK_COUNT; i++) { + ice_init_sw_xlt1_db(hw, (enum ice_block)i); + ice_init_sw_xlt2_db(hw, (enum ice_block)i); + } +} + /** * ice_fill_tbl - Reads content of a single table type into database * @hw: pointer to the hardware structure @@ -2921,7 +3494,7 @@ static void ice_fill_tbl(struct ice_hw *hw, enum ice_block block_id, u32 sid) void *sect; /* if the HW segment pointer is null then the first iteration of - * ice_pkg_enum_section() will fail. In this case the Hw tables will + * ice_pkg_enum_section() will fail. In this case the HW tables will * not be filled and return success. */ if (!hw->seg) { @@ -2954,7 +3527,7 @@ static void ice_fill_tbl(struct ice_hw *hw, enum ice_block block_id, u32 sid) case ICE_SID_XLT2_ACL: case ICE_SID_XLT2_PE: xlt2 = (struct ice_xlt2_section *)sect; - src = (u8 *)xlt2->value; + src = (_FORCE_ u8 *)xlt2->value; sect_len = LE16_TO_CPU(xlt2->count) * sizeof(*hw->blk[block_id].xlt2.t); dst = (u8 *)hw->blk[block_id].xlt2.t; @@ -2994,12 +3567,12 @@ static void ice_fill_tbl(struct ice_hw *hw, enum ice_block block_id, u32 sid) case ICE_SID_FLD_VEC_PE: es = (struct ice_sw_fv_section *)sect; src = (u8 *)es->fv; - sect_len = LE16_TO_CPU(es->count) * - hw->blk[block_id].es.fvw * + sect_len = (u32)(LE16_TO_CPU(es->count) * + hw->blk[block_id].es.fvw) * sizeof(*hw->blk[block_id].es.t); dst = (u8 *)hw->blk[block_id].es.t; - dst_len = hw->blk[block_id].es.count * - hw->blk[block_id].es.fvw * + dst_len = (u32)(hw->blk[block_id].es.count * + hw->blk[block_id].es.fvw) * sizeof(*hw->blk[block_id].es.t); break; default: @@ -3013,7 +3586,7 @@ static void ice_fill_tbl(struct ice_hw *hw, enum ice_block block_id, u32 sid) return; /* if the sum of section size and offset exceed destination size - * then we are out of bounds of the Hw table size for that PF. + * then we are out of bounds of the HW table size for that PF. * Changing section length to fill the remaining table space * of that PF. */ @@ -3027,76 +3600,82 @@ static void ice_fill_tbl(struct ice_hw *hw, enum ice_block block_id, u32 sid) } /** - * ice_fill_blk_tbls - Read package content for tables of a block + * ice_fill_blk_tbls - Read package context for tables * @hw: pointer to the hardware structure - * @block_id: The block ID which contains the tables to be copied * * Reads the current package contents and populates the driver - * database with the data it contains to allow for advanced driver - * features. - */ -static void ice_fill_blk_tbls(struct ice_hw *hw, enum ice_block block_id) -{ - ice_fill_tbl(hw, block_id, hw->blk[block_id].xlt1.sid); - ice_fill_tbl(hw, block_id, hw->blk[block_id].xlt2.sid); - ice_fill_tbl(hw, block_id, hw->blk[block_id].prof.sid); - ice_fill_tbl(hw, block_id, hw->blk[block_id].prof_redir.sid); - ice_fill_tbl(hw, block_id, hw->blk[block_id].es.sid); -} - -/** - * ice_free_flow_profs - free flow profile entries - * @hw: pointer to the hardware structure + * database with the data iteratively for all advanced feature + * blocks. Assume that the HW tables have been allocated. */ -static void ice_free_flow_profs(struct ice_hw *hw) +void ice_fill_blk_tbls(struct ice_hw *hw) { u8 i; for (i = 0; i < ICE_BLK_COUNT; i++) { - struct ice_flow_prof *p, *tmp; - - if (!&hw->fl_profs[i]) - continue; - - /* This call is being made as part of resource deallocation - * during unload. Lock acquire and release will not be - * necessary here. - */ - LIST_FOR_EACH_ENTRY_SAFE(p, tmp, &hw->fl_profs[i], - ice_flow_prof, l_entry) { - struct ice_flow_entry *e, *t; - - LIST_FOR_EACH_ENTRY_SAFE(e, t, &p->entries, - ice_flow_entry, l_entry) - ice_flow_rem_entry(hw, ICE_FLOW_ENTRY_HNDL(e)); + enum ice_block blk_id = (enum ice_block)i; - LIST_DEL(&p->l_entry); - if (p->acts) - ice_free(hw, p->acts); - ice_free(hw, p); - } - - ice_destroy_lock(&hw->fl_profs_locks[i]); + ice_fill_tbl(hw, blk_id, hw->blk[blk_id].xlt1.sid); + ice_fill_tbl(hw, blk_id, hw->blk[blk_id].xlt2.sid); + ice_fill_tbl(hw, blk_id, hw->blk[blk_id].prof.sid); + ice_fill_tbl(hw, blk_id, hw->blk[blk_id].prof_redir.sid); + ice_fill_tbl(hw, blk_id, hw->blk[blk_id].es.sid); } + + ice_init_sw_db(hw); } /** - * ice_free_prof_map - frees the profile map + * ice_free_prof_map - free profile map * @hw: pointer to the hardware structure - * @blk: the HW block which contains the profile map to be freed + * @blk_idx: HW block index */ -static void ice_free_prof_map(struct ice_hw *hw, enum ice_block blk) +static void ice_free_prof_map(struct ice_hw *hw, u8 blk_idx) { + struct ice_es *es = &hw->blk[blk_idx].es; struct ice_prof_map *del, *tmp; - if (LIST_EMPTY(&hw->blk[blk].es.prof_map)) - return; - - LIST_FOR_EACH_ENTRY_SAFE(del, tmp, &hw->blk[blk].es.prof_map, + ice_acquire_lock(&es->prof_map_lock); + LIST_FOR_EACH_ENTRY_SAFE(del, tmp, &es->prof_map, ice_prof_map, list) { LIST_DEL(&del->list); ice_free(hw, del); } + INIT_LIST_HEAD(&es->prof_map); + ice_release_lock(&es->prof_map_lock); +} + +/** + * ice_free_flow_profs - free flow profile entries + * @hw: pointer to the hardware structure + * @blk_idx: HW block index + */ +static void ice_free_flow_profs(struct ice_hw *hw, u8 blk_idx) +{ + struct ice_flow_prof *p, *tmp; + + ice_acquire_lock(&hw->fl_profs_locks[blk_idx]); + LIST_FOR_EACH_ENTRY_SAFE(p, tmp, &hw->fl_profs[blk_idx], + ice_flow_prof, l_entry) { + struct ice_flow_entry *e, *t; + + LIST_FOR_EACH_ENTRY_SAFE(e, t, &p->entries, + ice_flow_entry, l_entry) + ice_flow_rem_entry(hw, (enum ice_block)blk_idx, + ICE_FLOW_ENTRY_HNDL(e)); + + LIST_DEL(&p->l_entry); + if (p->acts) + ice_free(hw, p->acts); + + ice_destroy_lock(&p->entries_lock); + ice_free(hw, p); + } + ice_release_lock(&hw->fl_profs_locks[blk_idx]); + + /* if driver is in reset and tables are being cleared + * re-initialize the flow profile list heads + */ + INIT_LIST_HEAD(&hw->fl_profs[blk_idx]); } /** @@ -3122,10 +3701,20 @@ static void ice_free_vsig_tbl(struct ice_hw *hw, enum ice_block blk) */ void ice_free_hw_tbls(struct ice_hw *hw) { + struct ice_rss_cfg *r, *rt; u8 i; for (i = 0; i < ICE_BLK_COUNT; i++) { - ice_free_prof_map(hw, (enum ice_block)i); + if (hw->blk[i].is_list_init) { + struct ice_es *es = &hw->blk[i].es; + + ice_free_prof_map(hw, i); + ice_destroy_lock(&es->prof_map_lock); + ice_free_flow_profs(hw, i); + ice_destroy_lock(&hw->fl_profs_locks[i]); + + hw->blk[i].is_list_init = false; + } ice_free_vsig_tbl(hw, (enum ice_block)i); ice_free(hw, hw->blk[i].xlt1.ptypes); ice_free(hw, hw->blk[i].xlt1.ptg_tbl); @@ -3137,88 +3726,84 @@ void ice_free_hw_tbls(struct ice_hw *hw) ice_free(hw, hw->blk[i].prof_redir.t); ice_free(hw, hw->blk[i].es.t); ice_free(hw, hw->blk[i].es.ref_count); - - ice_free(hw, hw->blk[i].es.resource_used_hack); - ice_free(hw, hw->blk[i].prof.resource_used_hack); + ice_free(hw, hw->blk[i].es.written); + ice_free(hw, hw->blk[i].es.mask_ena); } + LIST_FOR_EACH_ENTRY_SAFE(r, rt, &hw->rss_list_head, + ice_rss_cfg, l_entry) { + LIST_DEL(&r->l_entry); + ice_free(hw, r); + } + ice_destroy_lock(&hw->rss_locks); + if (!hw->dcf_enabled) + ice_shutdown_all_prof_masks(hw); ice_memset(hw->blk, 0, sizeof(hw->blk), ICE_NONDMA_MEM); - - ice_free_flow_profs(hw); } /** * ice_init_flow_profs - init flow profile locks and list heads * @hw: pointer to the hardware structure + * @blk_idx: HW block index */ -static void ice_init_flow_profs(struct ice_hw *hw) -{ - u8 i; - - for (i = 0; i < ICE_BLK_COUNT; i++) { - ice_init_lock(&hw->fl_profs_locks[i]); - INIT_LIST_HEAD(&hw->fl_profs[i]); - } -} - -/** - * ice_init_sw_xlt1_db - init software XLT1 database from HW tables - * @hw: pointer to the hardware structure - * @blk: the HW block to initialize - */ -static -void ice_init_sw_xlt1_db(struct ice_hw *hw, enum ice_block blk) +static void ice_init_flow_profs(struct ice_hw *hw, u8 blk_idx) { - u16 pt; - - for (pt = 0; pt < hw->blk[blk].xlt1.count; pt++) { - u8 ptg; - - ptg = hw->blk[blk].xlt1.t[pt]; - if (ptg != ICE_DEFAULT_PTG) { - ice_ptg_alloc_val(hw, blk, ptg); - ice_ptg_add_mv_ptype(hw, blk, pt, ptg); - } - } + ice_init_lock(&hw->fl_profs_locks[blk_idx]); + INIT_LIST_HEAD(&hw->fl_profs[blk_idx]); } /** - * ice_init_sw_xlt2_db - init software XLT2 database from HW tables + * ice_clear_hw_tbls - clear HW tables and flow profiles * @hw: pointer to the hardware structure - * @blk: the HW block to initialize */ -static -void ice_init_sw_xlt2_db(struct ice_hw *hw, enum ice_block blk) +void ice_clear_hw_tbls(struct ice_hw *hw) { - u16 vsi; + u8 i; - for (vsi = 0; vsi < hw->blk[blk].xlt2.count; vsi++) { - u16 vsig; + for (i = 0; i < ICE_BLK_COUNT; i++) { + struct ice_prof_redir *prof_redir = &hw->blk[i].prof_redir; + struct ice_prof_tcam *prof = &hw->blk[i].prof; + struct ice_xlt1 *xlt1 = &hw->blk[i].xlt1; + struct ice_xlt2 *xlt2 = &hw->blk[i].xlt2; + struct ice_es *es = &hw->blk[i].es; - vsig = hw->blk[blk].xlt2.t[vsi]; - if (vsig) { - ice_vsig_alloc_val(hw, blk, vsig); - ice_vsig_add_mv_vsi(hw, blk, vsi, vsig); - /* no changes at this time, since this has been - * initialized from the original package - */ - hw->blk[blk].xlt2.vsis[vsi].changed = 0; + if (hw->blk[i].is_list_init) { + ice_free_prof_map(hw, i); + ice_free_flow_profs(hw, i); } - } -} -/** - * ice_init_sw_db - init software database from HW tables - * @hw: pointer to the hardware structure - */ -static -void ice_init_sw_db(struct ice_hw *hw) -{ - u16 i; + ice_free_vsig_tbl(hw, (enum ice_block)i); - for (i = 0; i < ICE_BLK_COUNT; i++) { - ice_init_sw_xlt1_db(hw, (enum ice_block)i); - ice_init_sw_xlt2_db(hw, (enum ice_block)i); + ice_memset(xlt1->ptypes, 0, xlt1->count * sizeof(*xlt1->ptypes), + ICE_NONDMA_MEM); + ice_memset(xlt1->ptg_tbl, 0, + ICE_MAX_PTGS * sizeof(*xlt1->ptg_tbl), + ICE_NONDMA_MEM); + ice_memset(xlt1->t, 0, xlt1->count * sizeof(*xlt1->t), + ICE_NONDMA_MEM); + + ice_memset(xlt2->vsis, 0, xlt2->count * sizeof(*xlt2->vsis), + ICE_NONDMA_MEM); + ice_memset(xlt2->vsig_tbl, 0, + xlt2->count * sizeof(*xlt2->vsig_tbl), + ICE_NONDMA_MEM); + ice_memset(xlt2->t, 0, xlt2->count * sizeof(*xlt2->t), + ICE_NONDMA_MEM); + + ice_memset(prof->t, 0, prof->count * sizeof(*prof->t), + ICE_NONDMA_MEM); + ice_memset(prof_redir->t, 0, + prof_redir->count * sizeof(*prof_redir->t), + ICE_NONDMA_MEM); + + ice_memset(es->t, 0, es->count * sizeof(*es->t) * es->fvw, + ICE_NONDMA_MEM); + ice_memset(es->ref_count, 0, es->count * sizeof(*es->ref_count), + ICE_NONDMA_MEM); + ice_memset(es->written, 0, es->count * sizeof(*es->written), + ICE_NONDMA_MEM); + ice_memset(es->mask_ena, 0, es->count * sizeof(*es->mask_ena), + ICE_NONDMA_MEM); } } @@ -3230,14 +3815,25 @@ enum ice_status ice_init_hw_tbls(struct ice_hw *hw) { u8 i; - ice_init_flow_profs(hw); - + ice_init_lock(&hw->rss_locks); + INIT_LIST_HEAD(&hw->rss_list_head); + if (!hw->dcf_enabled) + ice_init_all_prof_masks(hw); for (i = 0; i < ICE_BLK_COUNT; i++) { struct ice_prof_redir *prof_redir = &hw->blk[i].prof_redir; struct ice_prof_tcam *prof = &hw->blk[i].prof; struct ice_xlt1 *xlt1 = &hw->blk[i].xlt1; struct ice_xlt2 *xlt2 = &hw->blk[i].xlt2; struct ice_es *es = &hw->blk[i].es; + u16 j; + + if (hw->blk[i].is_list_init) + continue; + + ice_init_flow_profs(hw, i); + ice_init_lock(&es->prof_map_lock); + INIT_LIST_HEAD(&es->prof_map); + hw->blk[i].is_list_init = true; hw->blk[i].overwrite = blk_sizes[i].overwrite; es->reverse = blk_sizes[i].reverse; @@ -3275,6 +3871,9 @@ enum ice_status ice_init_hw_tbls(struct ice_hw *hw) if (!xlt2->vsig_tbl) goto err; + for (j = 0; j < xlt2->count; j++) + INIT_LIST_HEAD(&xlt2->vsig_tbl[j].prop_lst); + xlt2->t = (u16 *)ice_calloc(hw, xlt2->count, sizeof(*xlt2->t)); if (!xlt2->t) goto err; @@ -3301,37 +3900,21 @@ enum ice_status ice_init_hw_tbls(struct ice_hw *hw) es->count = blk_sizes[i].es; es->fvw = blk_sizes[i].fvw; es->t = (struct ice_fv_word *) - ice_calloc(hw, es->count * es->fvw, sizeof(*es->t)); - + ice_calloc(hw, (u32)(es->count * es->fvw), + sizeof(*es->t)); if (!es->t) goto err; es->ref_count = (u16 *) ice_calloc(hw, es->count, sizeof(*es->ref_count)); + es->written = (u8 *) + ice_calloc(hw, es->count, sizeof(*es->written)); + es->mask_ena = (u32 *) + ice_calloc(hw, es->count, sizeof(*es->mask_ena)); if (!es->ref_count) goto err; - - es->resource_used_hack = (u8 *) - ice_calloc(hw, hw->blk[i].es.count, sizeof(u8)); - - if (!es->resource_used_hack) - goto err; - - prof->resource_used_hack = (u8 *)ice_calloc(hw, prof->count, - sizeof(u8)); - - if (!prof->resource_used_hack) - goto err; - - INIT_LIST_HEAD(&es->prof_map); - - /* Now that tables are allocated, read in package data */ - ice_fill_blk_tbls(hw, (enum ice_block)i); } - - ice_init_sw_db(hw); - return ICE_SUCCESS; err: @@ -3345,7 +3928,7 @@ err: * @blk: the block in which to write profile ID to * @ptg: packet type group (PTG) portion of key * @vsig: VSIG portion of key - * @cdid: cdid portion of key + * @cdid: CDID portion of key * @flags: flag portion of key * @vl_msk: valid mask * @dc_msk: don't care mask @@ -3388,7 +3971,7 @@ ice_prof_gen_key(struct ice_hw *hw, enum ice_block blk, u8 ptg, u16 vsig, default: ice_debug(hw, ICE_DBG_PKG, "Error in profile config\n"); break; - }; + } return ice_set_key(key, ICE_TCAM_KEY_SZ, (u8 *)&inkey, vl_msk, dc_msk, nm_msk, 0, ICE_TCAM_KEY_SZ / 2); @@ -3402,7 +3985,7 @@ ice_prof_gen_key(struct ice_hw *hw, enum ice_block blk, u8 ptg, u16 vsig, * @prof_id: profile ID * @ptg: packet type group (PTG) portion of key * @vsig: VSIG portion of key - * @cdid: cdid portion of key + * @cdid: CDID portion of key * @flags: flag portion of key * @vl_msk: valid mask * @dc_msk: don't care mask @@ -3440,6 +4023,7 @@ ice_vsig_get_ref(struct ice_hw *hw, enum ice_block blk, u16 vsig, u16 *refs) { u16 idx = vsig & ICE_VSIG_IDX_M; struct ice_vsig_vsi *ptr; + *refs = 0; if (!hw->blk[blk].xlt2.vsig_tbl[idx].in_use) @@ -3454,43 +4038,6 @@ ice_vsig_get_ref(struct ice_hw *hw, enum ice_block blk, u16 vsig, u16 *refs) return ICE_SUCCESS; } -/** - * ice_get_ptg - get or allocate a ptg for a ptype - * @hw: pointer to the hardware structure - * @blk: HW block - * @ptype: the ptype to retrieve the PTG for - * @ptg: receives the PTG of the ptype - * @add: receive boolean indicating whether PTG was added or not - */ -static enum ice_status -ice_get_ptg(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 *ptg, - bool *add) -{ - enum ice_status status; - - *ptg = ICE_DEFAULT_PTG; - *add = false; - - status = ice_ptg_find_ptype(hw, blk, ptype, ptg); - if (status) - return status; - - if (*ptg == ICE_DEFAULT_PTG) { - /* need to allocate a PTG, and add ptype to it */ - *ptg = ice_ptg_alloc(hw, blk); - if (*ptg == ICE_DEFAULT_PTG) - return ICE_ERR_HW_TABLE; - - status = ice_ptg_add_mv_ptype(hw, blk, ptype, *ptg); - if (status) - return ICE_ERR_HW_TABLE; - - *add = true; - } - - return ICE_SUCCESS; -}; - /** * ice_has_prof_vsig - check to see if VSIG has a specific profile * @hw: pointer to the hardware structure @@ -3505,10 +4052,9 @@ ice_has_prof_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl) struct ice_vsig_prof *ent; LIST_FOR_EACH_ENTRY(ent, &hw->blk[blk].xlt2.vsig_tbl[idx].prop_lst, - ice_vsig_prof, list) { + ice_vsig_prof, list) if (ent->profile_cookie == hdl) return true; - } ice_debug(hw, ICE_DBG_INIT, "Characteristic list for VSI group %d not found.\n", @@ -3530,7 +4076,7 @@ ice_prof_bld_es(struct ice_hw *hw, enum ice_block blk, u16 vec_size = hw->blk[blk].es.fvw * sizeof(struct ice_fv_word); struct ice_chs_chg *tmp; - LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) { + LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) if (tmp->type == ICE_PTG_ES_ADD && tmp->add_prof) { u16 off = tmp->prof_id * hw->blk[blk].es.fvw; struct ice_pkg_es *p; @@ -3551,7 +4097,6 @@ ice_prof_bld_es(struct ice_hw *hw, enum ice_block blk, ice_memcpy(p->es, &hw->blk[blk].es.t[off], vec_size, ICE_NONDMA_TO_NONDMA); } - } return ICE_SUCCESS; } @@ -3569,7 +4114,7 @@ ice_prof_bld_tcam(struct ice_hw *hw, enum ice_block blk, { struct ice_chs_chg *tmp; - LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) { + LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) if (tmp->type == ICE_TCAM_ADD && tmp->add_tcam_idx) { struct ice_prof_id_section *p; u32 id; @@ -3590,7 +4135,6 @@ ice_prof_bld_tcam(struct ice_hw *hw, enum ice_block blk, sizeof(hw->blk[blk].prof.t->key), ICE_NONDMA_TO_NONDMA); } - } return ICE_SUCCESS; } @@ -3607,7 +4151,7 @@ ice_prof_bld_xlt1(enum ice_block blk, struct ice_buf_build *bld, { struct ice_chs_chg *tmp; - LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) { + LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) if (tmp->type == ICE_PTG_ES_ADD && tmp->add_ptg) { struct ice_xlt1_section *p; u32 id; @@ -3623,7 +4167,6 @@ ice_prof_bld_xlt1(enum ice_block blk, struct ice_buf_build *bld, p->offset = CPU_TO_LE16(tmp->ptype); p->value[0] = tmp->ptg; } - } return ICE_SUCCESS; } @@ -3641,19 +4184,13 @@ ice_prof_bld_xlt2(enum ice_block blk, struct ice_buf_build *bld, struct ice_chs_chg *tmp; LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) { - bool found = false; - - if (tmp->type == ICE_VSIG_ADD) - found = true; - else if (tmp->type == ICE_VSI_MOVE) - found = true; - else if (tmp->type == ICE_VSIG_REM) - found = true; - - if (found) { - struct ice_xlt2_section *p; - u32 id; + struct ice_xlt2_section *p; + u32 id; + switch (tmp->type) { + case ICE_VSIG_ADD: + case ICE_VSI_MOVE: + case ICE_VSIG_REM: id = ice_sect_id(blk, ICE_XLT2); p = (struct ice_xlt2_section *) ice_pkg_buf_alloc_section(bld, id, sizeof(*p)); @@ -3664,6 +4201,9 @@ ice_prof_bld_xlt2(enum ice_block blk, struct ice_buf_build *bld, p->count = CPU_TO_LE16(1); p->offset = CPU_TO_LE16(tmp->vsi); p->value[0] = CPU_TO_LE16(tmp->vsig); + break; + default: + break; } } @@ -3683,12 +4223,12 @@ ice_upd_prof_hw(struct ice_hw *hw, enum ice_block blk, struct ice_buf_build *b; struct ice_chs_chg *tmp; enum ice_status status; - u16 pkg_sects = 0; - u16 sects = 0; + u16 pkg_sects; u16 xlt1 = 0; u16 xlt2 = 0; u16 tcam = 0; u16 es = 0; + u16 sects; /* count number of sections we need */ LIST_FOR_EACH_ENTRY(tmp, chgs, ice_chs_chg, list_entry) { @@ -3763,20 +4303,283 @@ ice_upd_prof_hw(struct ice_hw *hw, enum ice_block blk, /* update package */ status = ice_update_pkg(hw, ice_pkg_buf(b), 1); if (status == ICE_ERR_AQ_ERROR) - ice_debug(hw, ICE_DBG_INIT, "Unable to update HW profile."); + ice_debug(hw, ICE_DBG_INIT, "Unable to update HW profile\n"); error_tmp: ice_pkg_buf_free(hw, b); return status; } +/** + * ice_update_fd_mask - set Flow Director Field Vector mask for a profile + * @hw: pointer to the HW struct + * @prof_id: profile ID + * @mask_sel: mask select + * + * This function enable any of the masks selected by the mask select parameter + * for the profile specified. + */ +static void ice_update_fd_mask(struct ice_hw *hw, u16 prof_id, u32 mask_sel) +{ + wr32(hw, GLQF_FDMASK_SEL(prof_id), mask_sel); + + ice_debug(hw, ICE_DBG_INIT, "fd mask(%d): %x = %x\n", prof_id, + GLQF_FDMASK_SEL(prof_id), mask_sel); +} + +struct ice_fd_src_dst_pair { + u8 prot_id; + u8 count; + u16 off; +}; + +static const struct ice_fd_src_dst_pair ice_fd_pairs[] = { + /* These are defined in pairs */ + { ICE_PROT_IPV4_OF_OR_S, 2, 12 }, + { ICE_PROT_IPV4_OF_OR_S, 2, 16 }, + + { ICE_PROT_IPV4_IL, 2, 12 }, + { ICE_PROT_IPV4_IL, 2, 16 }, + + { ICE_PROT_IPV6_OF_OR_S, 8, 8 }, + { ICE_PROT_IPV6_OF_OR_S, 8, 24 }, + + { ICE_PROT_IPV6_IL, 8, 8 }, + { ICE_PROT_IPV6_IL, 8, 24 }, + + { ICE_PROT_TCP_IL, 1, 0 }, + { ICE_PROT_TCP_IL, 1, 2 }, + + { ICE_PROT_UDP_OF, 1, 0 }, + { ICE_PROT_UDP_OF, 1, 2 }, + + { ICE_PROT_UDP_IL_OR_S, 1, 0 }, + { ICE_PROT_UDP_IL_OR_S, 1, 2 }, + + { ICE_PROT_SCTP_IL, 1, 0 }, + { ICE_PROT_SCTP_IL, 1, 2 } +}; + +#define ICE_FD_SRC_DST_PAIR_COUNT ARRAY_SIZE(ice_fd_pairs) + +/** + * ice_update_fd_swap - set register appropriately for a FD FV extraction + * @hw: pointer to the HW struct + * @prof_id: profile ID + * @es: extraction sequence (length of array is determined by the block) + */ +static enum ice_status +ice_update_fd_swap(struct ice_hw *hw, u16 prof_id, struct ice_fv_word *es) +{ + ice_declare_bitmap(pair_list, ICE_FD_SRC_DST_PAIR_COUNT); + u8 pair_start[ICE_FD_SRC_DST_PAIR_COUNT] = { 0 }; +#define ICE_FD_FV_NOT_FOUND (-2) + s8 first_free = ICE_FD_FV_NOT_FOUND; + u8 used[ICE_MAX_FV_WORDS] = { 0 }; + s8 orig_free, si; + u32 mask_sel = 0; + u8 i, j, k; + + ice_zero_bitmap(pair_list, ICE_FD_SRC_DST_PAIR_COUNT); + + /* This code assumes that the Flow Director field vectors are assigned + * from the end of the FV indexes working towards the zero index, that + * only complete fields will be included and will be consecutive, and + * that there are no gaps between valid indexes. + */ + + /* Determine swap fields present */ + for (i = 0; i < hw->blk[ICE_BLK_FD].es.fvw; i++) { + /* Find the first free entry, assuming right to left population. + * This is where we can start adding additional pairs if needed. + */ + if (first_free == ICE_FD_FV_NOT_FOUND && es[i].prot_id != + ICE_PROT_INVALID) + first_free = i - 1; + + for (j = 0; j < ICE_FD_SRC_DST_PAIR_COUNT; j++) + if (es[i].prot_id == ice_fd_pairs[j].prot_id && + es[i].off == ice_fd_pairs[j].off) { + ice_set_bit(j, pair_list); + pair_start[j] = i; + } + } + + orig_free = first_free; + + /* determine missing swap fields that need to be added */ + for (i = 0; i < ICE_FD_SRC_DST_PAIR_COUNT; i += 2) { + u8 bit1 = ice_is_bit_set(pair_list, i + 1); + u8 bit0 = ice_is_bit_set(pair_list, i); + + if (bit0 ^ bit1) { + u8 index; + + /* add the appropriate 'paired' entry */ + if (!bit0) + index = i; + else + index = i + 1; + + /* check for room */ + if (first_free + 1 < (s8)ice_fd_pairs[index].count) + return ICE_ERR_MAX_LIMIT; + + /* place in extraction sequence */ + for (k = 0; k < ice_fd_pairs[index].count; k++) { + es[first_free - k].prot_id = + ice_fd_pairs[index].prot_id; + es[first_free - k].off = + ice_fd_pairs[index].off + (k * 2); + + if (k > first_free) + return ICE_ERR_OUT_OF_RANGE; + + /* keep track of non-relevant fields */ + mask_sel |= BIT(first_free - k); + } + + pair_start[index] = first_free; + first_free -= ice_fd_pairs[index].count; + } + } + + /* fill in the swap array */ + si = hw->blk[ICE_BLK_FD].es.fvw - 1; + while (si >= 0) { + u8 indexes_used = 1; + + /* assume flat at this index */ +#define ICE_SWAP_VALID 0x80 + used[si] = si | ICE_SWAP_VALID; + + if (orig_free == ICE_FD_FV_NOT_FOUND || si <= orig_free) { + si -= indexes_used; + continue; + } + + /* check for a swap location */ + for (j = 0; j < ICE_FD_SRC_DST_PAIR_COUNT; j++) + if (es[si].prot_id == ice_fd_pairs[j].prot_id && + es[si].off == ice_fd_pairs[j].off) { + u8 idx; + + /* determine the appropriate matching field */ + idx = j + ((j % 2) ? -1 : 1); + + indexes_used = ice_fd_pairs[idx].count; + for (k = 0; k < indexes_used; k++) { + used[si - k] = (pair_start[idx] - k) | + ICE_SWAP_VALID; + } + + break; + } + + si -= indexes_used; + } + + /* for each set of 4 swap and 4 inset indexes, write the appropriate + * register + */ + for (j = 0; j < hw->blk[ICE_BLK_FD].es.fvw / 4; j++) { + u32 raw_swap = 0; + u32 raw_in = 0; + + for (k = 0; k < 4; k++) { + u8 idx; + + idx = (j * 4) + k; + if (used[idx] && !(mask_sel & BIT(idx))) { + raw_swap |= used[idx] << (k * BITS_PER_BYTE); +#define ICE_INSET_DFLT 0x9f + raw_in |= ICE_INSET_DFLT << (k * BITS_PER_BYTE); + } + } + + /* write the appropriate swap register set */ + wr32(hw, GLQF_FDSWAP(prof_id, j), raw_swap); + + ice_debug(hw, ICE_DBG_INIT, "swap wr(%d, %d): %x = %08x\n", + prof_id, j, GLQF_FDSWAP(prof_id, j), raw_swap); + + /* write the appropriate inset register set */ + wr32(hw, GLQF_FDINSET(prof_id, j), raw_in); + + ice_debug(hw, ICE_DBG_INIT, "inset wr(%d, %d): %x = %08x\n", + prof_id, j, GLQF_FDINSET(prof_id, j), raw_in); + } + + /* initially clear the mask select for this profile */ + ice_update_fd_mask(hw, prof_id, 0); + + return ICE_SUCCESS; +} + +/* The entries here needs to match the order of enum ice_ptype_attrib */ +static const struct ice_ptype_attrib_info ice_ptype_attributes[] = { + { ICE_GTP_PDU_EH, ICE_GTP_PDU_FLAG_MASK }, + { ICE_GTP_SESSION, ICE_GTP_FLAGS_MASK }, + { ICE_GTP_DOWNLINK, ICE_GTP_FLAGS_MASK }, + { ICE_GTP_UPLINK, ICE_GTP_FLAGS_MASK }, +}; + +/** + * ice_get_ptype_attrib_info - get ptype attribute information + * @type: attribute type + * @info: pointer to variable to the attribute information + */ +static void +ice_get_ptype_attrib_info(enum ice_ptype_attrib_type type, + struct ice_ptype_attrib_info *info) +{ + *info = ice_ptype_attributes[type]; +} + +/** + * ice_add_prof_attrib - add any PTG with attributes to profile + * @prof: pointer to the profile to which PTG entries will be added + * @ptg: PTG to be added + * @ptype: PTYPE that needs to be looked up + * @attr: array of attributes that will be considered + * @attr_cnt: number of elements in the attribute array + */ +static enum ice_status +ice_add_prof_attrib(struct ice_prof_map *prof, u8 ptg, u16 ptype, + const struct ice_ptype_attributes *attr, u16 attr_cnt) +{ + bool found = false; + u16 i; + + for (i = 0; i < attr_cnt; i++) { + if (attr[i].ptype == ptype) { + found = true; + + prof->ptg[prof->ptg_cnt] = ptg; + ice_get_ptype_attrib_info(attr[i].attrib, + &prof->attr[prof->ptg_cnt]); + + if (++prof->ptg_cnt >= ICE_MAX_PTG_PER_PROFILE) + return ICE_ERR_MAX_LIMIT; + } + } + + if (!found) + return ICE_ERR_DOES_NOT_EXIST; + + return ICE_SUCCESS; +} + /** * ice_add_prof - add profile * @hw: pointer to the HW struct * @blk: hardware block * @id: profile tracking ID * @ptypes: array of bitmaps indicating ptypes (ICE_FLOW_PTYPE_MAX bits) + * @attr: array of attributes + * @attr_cnt: number of elements in attrib array * @es: extraction sequence (length of array is determined by the block) + * @masks: mask for extraction sequence * * This function registers a profile, which matches a set of PTYPES with a * particular extraction sequence. While the hardware profile is allocated @@ -3785,19 +4588,40 @@ error_tmp: */ enum ice_status ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], - struct ice_fv_word *es) + const struct ice_ptype_attributes *attr, u16 attr_cnt, + struct ice_fv_word *es, u16 *masks) { u32 bytes = DIVIDE_AND_ROUND_UP(ICE_FLOW_PTYPE_MAX, BITS_PER_BYTE); + ice_declare_bitmap(ptgs_used, ICE_XLT1_CNT); struct ice_prof_map *prof; enum ice_status status; - u32 byte = 0; + u8 byte = 0; u8 prof_id; + ice_zero_bitmap(ptgs_used, ICE_XLT1_CNT); + + ice_acquire_lock(&hw->blk[blk].es.prof_map_lock); + /* search for existing profile */ - status = ice_find_prof_id(hw, blk, es, &prof_id); + status = ice_find_prof_id_with_mask(hw, blk, es, masks, &prof_id); if (status) { /* allocate profile ID */ status = ice_alloc_prof_id(hw, blk, &prof_id); + if (status) + goto err_ice_add_prof; + if (blk == ICE_BLK_FD) { + /* For Flow Director block, the extraction sequence may + * need to be altered in the case where there are paired + * fields that have no match. This is necessary because + * for Flow Director, src and dest fields need to paired + * for filter programming and these values are swapped + * during Tx. + */ + status = ice_update_fd_swap(hw, prof_id, es); + if (status) + goto err_ice_add_prof; + } + status = ice_update_prof_masking(hw, blk, prof_id, masks); if (status) goto err_ice_add_prof; @@ -3805,6 +4629,8 @@ ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], ice_write_es(hw, blk, prof_id, es); } + ice_prof_inc_ref(hw, blk, prof_id); + /* add profile info */ prof = (struct ice_prof_map *)ice_malloc(hw, sizeof(*prof)); @@ -3813,12 +4639,12 @@ ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], prof->profile_cookie = id; prof->prof_id = prof_id; - prof->ptype_count = 0; + prof->ptg_cnt = 0; prof->context = 0; /* build list of ptgs */ - while (bytes && prof->ptype_count < ICE_MAX_PTYPE_PER_PROFILE) { - u32 bit; + while (bytes && prof->ptg_cnt < ICE_MAX_PTG_PER_PROFILE) { + u8 bit; if (!ptypes[byte]) { bytes--; @@ -3827,21 +4653,47 @@ ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], } /* Examine 8 bits per byte */ for (bit = 0; bit < 8; bit++) { - if (ptypes[byte] & 1 << bit) { + if (ptypes[byte] & BIT(bit)) { u16 ptype; + u8 ptg; u8 m; - ptype = byte * 8 + bit; - if (ptype < ICE_FLOW_PTYPE_MAX) { - prof->ptype[prof->ptype_count] = ptype; + ptype = byte * BITS_PER_BYTE + bit; + + /* The package should place all ptypes in a + * non-zero PTG, so the following call should + * never fail. + */ + if (ice_ptg_find_ptype(hw, blk, ptype, &ptg)) + continue; + + /* If PTG is already added, skip and continue */ + if (ice_is_bit_set(ptgs_used, ptg)) + continue; - if (++prof->ptype_count >= - ICE_MAX_PTYPE_PER_PROFILE) + ice_set_bit(ptg, ptgs_used); + /* Check to see there are any attributes for + * this ptype, and add them if found. + */ + status = ice_add_prof_attrib(prof, ptg, ptype, + attr, attr_cnt); + if (status == ICE_ERR_MAX_LIMIT) + break; + if (status) { + /* This is simple a ptype/PTG with no + * attribute + */ + prof->ptg[prof->ptg_cnt] = ptg; + prof->attr[prof->ptg_cnt].flags = 0; + prof->attr[prof->ptg_cnt].mask = 0; + + if (++prof->ptg_cnt >= + ICE_MAX_PTG_PER_PROFILE) break; } /* nothing left in byte, then exit */ - m = ~((1 << (bit + 1)) - 1); + m = ~(u8)((1 << (bit + 1)) - 1); if (!(ptypes[byte] & m)) break; } @@ -3850,11 +4702,12 @@ ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], bytes--; byte++; } - LIST_ADD(&prof->list, &hw->blk[blk].es.prof_map); - return ICE_SUCCESS; + LIST_ADD(&prof->list, &hw->blk[blk].es.prof_map); + status = ICE_SUCCESS; err_ice_add_prof: + ice_release_lock(&hw->blk[blk].es.prof_map_lock); return status; } @@ -3865,6 +4718,7 @@ err_ice_add_prof: * @id: profile tracking ID * * This will search for a profile tracking ID which was previously added. + * The profile map lock should be held before calling this function. */ struct ice_prof_map * ice_search_prof_id(struct ice_hw *hw, enum ice_block blk, u64 id) @@ -3872,51 +4726,11 @@ ice_search_prof_id(struct ice_hw *hw, enum ice_block blk, u64 id) struct ice_prof_map *entry = NULL; struct ice_prof_map *map; - LIST_FOR_EACH_ENTRY(map, &hw->blk[blk].es.prof_map, ice_prof_map, - list) { + LIST_FOR_EACH_ENTRY(map, &hw->blk[blk].es.prof_map, ice_prof_map, list) if (map->profile_cookie == id) { entry = map; break; } - } - - return entry; -} - -/** - * ice_set_prof_context - Set context for a given profile - * @hw: pointer to the HW struct - * @blk: hardware block - * @id: profile tracking ID - * @cntxt: context - */ -struct ice_prof_map * -ice_set_prof_context(struct ice_hw *hw, enum ice_block blk, u64 id, u64 cntxt) -{ - struct ice_prof_map *entry; - - entry = ice_search_prof_id(hw, blk, id); - if (entry) - entry->context = cntxt; - - return entry; -} - -/** - * ice_get_prof_context - Get context for a given profile - * @hw: pointer to the HW struct - * @blk: hardware block - * @id: profile tracking ID - * @cntxt: pointer to variable to receive the context - */ -struct ice_prof_map * -ice_get_prof_context(struct ice_hw *hw, enum ice_block blk, u64 id, u64 *cntxt) -{ - struct ice_prof_map *entry; - - entry = ice_search_prof_id(hw, blk, id); - if (entry) - *cntxt = entry->context; return entry; } @@ -3934,9 +4748,8 @@ ice_vsig_prof_id_count(struct ice_hw *hw, enum ice_block blk, u16 vsig) struct ice_vsig_prof *p; LIST_FOR_EACH_ENTRY(p, &hw->blk[blk].xlt2.vsig_tbl[idx].prop_lst, - ice_vsig_prof, list) { + ice_vsig_prof, list) count++; - } return count; } @@ -3981,16 +4794,14 @@ ice_rem_prof_id(struct ice_hw *hw, enum ice_block blk, enum ice_status status; u16 i; - for (i = 0; i < prof->tcam_count; i++) { - prof->tcam[i].in_use = false; - status = ice_rel_tcam_idx(hw, blk, prof->tcam[i].tcam_idx); - if (!status) - status = ice_prof_dec_ref(hw, blk, - prof->tcam[i].prof_id); - - if (status) - return ICE_ERR_HW_TABLE; - } + for (i = 0; i < prof->tcam_count; i++) + if (prof->tcam[i].in_use) { + prof->tcam[i].in_use = false; + status = ice_rel_tcam_idx(hw, blk, + prof->tcam[i].tcam_idx); + if (status) + return ICE_ERR_HW_TABLE; + } return ICE_SUCCESS; } @@ -4025,36 +4836,29 @@ ice_rem_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, /* Move all VSIS associated with this VSIG to the default VSIG */ vsi_cur = hw->blk[blk].xlt2.vsig_tbl[idx].first_vsi; - if (!vsi_cur) - return ICE_ERR_CFG; - - do { - struct ice_vsig_vsi *tmp = vsi_cur->next_vsi; - struct ice_chs_chg *p; - - p = (struct ice_chs_chg *)ice_malloc(hw, sizeof(*p)); - if (!p) - goto err_ice_rem_vsig; - - p->type = ICE_VSIG_REM; - p->orig_vsig = vsig; - p->vsig = ICE_DEFAULT_VSIG; - p->vsi = vsi_cur - hw->blk[blk].xlt2.vsis; + /* If the VSIG has at least 1 VSI then iterate through the list + * and remove the VSIs before deleting the group. + */ + if (vsi_cur) + do { + struct ice_vsig_vsi *tmp = vsi_cur->next_vsi; + struct ice_chs_chg *p; - LIST_ADD(&p->list_entry, chg); + p = (struct ice_chs_chg *)ice_malloc(hw, sizeof(*p)); + if (!p) + return ICE_ERR_NO_MEMORY; - status = ice_vsig_free(hw, blk, vsig); - if (status) - return status; + p->type = ICE_VSIG_REM; + p->orig_vsig = vsig; + p->vsig = ICE_DEFAULT_VSIG; + p->vsi = vsi_cur - hw->blk[blk].xlt2.vsis; - vsi_cur = tmp; - } while (vsi_cur); + LIST_ADD(&p->list_entry, chg); - return ICE_SUCCESS; + vsi_cur = tmp; + } while (vsi_cur); -err_ice_rem_vsig: - /* the caller will free up the change list */ - return ICE_ERR_NO_MEMORY; + return ice_vsig_free(hw, blk, vsig); } /** @@ -4075,7 +4879,7 @@ ice_rem_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, LIST_FOR_EACH_ENTRY_SAFE(p, t, &hw->blk[blk].xlt2.vsig_tbl[idx].prop_lst, - ice_vsig_prof, list) { + ice_vsig_prof, list) if (p->profile_cookie == hdl) { if (ice_vsig_prof_id_count(hw, blk, vsig) == 1) /* this is the last profile, remove the VSIG */ @@ -4088,7 +4892,6 @@ ice_rem_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, } return status; } - } return ICE_ERR_DOES_NOT_EXIST; } @@ -4109,7 +4912,7 @@ ice_rem_flow_all(struct ice_hw *hw, enum ice_block blk, u64 id) INIT_LIST_HEAD(&chg); - for (i = 1; i < ICE_MAX_VSIGS; i++) { + for (i = 1; i < ICE_MAX_VSIGS; i++) if (hw->blk[blk].xlt2.vsig_tbl[i].in_use) { if (ice_has_prof_vsig(hw, blk, i, id)) { status = ice_rem_prof_id_vsig(hw, blk, i, id, @@ -4118,7 +4921,6 @@ ice_rem_flow_all(struct ice_hw *hw, enum ice_block blk, u64 id) goto err_ice_rem_flow_all; } } - } status = ice_upd_prof_hw(hw, blk, &chg); @@ -4143,80 +4945,84 @@ err_ice_rem_flow_all: */ enum ice_status ice_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 id) { - enum ice_status status; struct ice_prof_map *pmap; + enum ice_status status; - pmap = ice_search_prof_id(hw, blk, id); - if (!pmap) - return ICE_ERR_DOES_NOT_EXIST; - - status = ice_free_prof_id(hw, blk, pmap->prof_id); + ice_acquire_lock(&hw->blk[blk].es.prof_map_lock); - if (status) - return status; + pmap = ice_search_prof_id(hw, blk, id); + if (!pmap) { + status = ICE_ERR_DOES_NOT_EXIST; + goto err_ice_rem_prof; + } /* remove all flows with this profile */ status = ice_rem_flow_all(hw, blk, pmap->profile_cookie); if (status) - return status; + goto err_ice_rem_prof; + + /* dereference profile, and possibly remove */ + ice_prof_dec_ref(hw, blk, pmap->prof_id); + LIST_DEL(&pmap->list); ice_free(hw, pmap); - return ICE_SUCCESS; +err_ice_rem_prof: + ice_release_lock(&hw->blk[blk].es.prof_map_lock); + return status; } /** - * ice_get_prof_ptgs - get ptgs for profile + * ice_get_prof - get profile * @hw: pointer to the HW struct * @blk: hardware block * @hdl: profile handle * @chg: change list */ static enum ice_status -ice_get_prof_ptgs(struct ice_hw *hw, enum ice_block blk, u64 hdl, - struct LIST_HEAD_TYPE *chg) +ice_get_prof(struct ice_hw *hw, enum ice_block blk, u64 hdl, + struct LIST_HEAD_TYPE *chg) { + enum ice_status status = ICE_SUCCESS; struct ice_prof_map *map; struct ice_chs_chg *p; u16 i; + ice_acquire_lock(&hw->blk[blk].es.prof_map_lock); /* Get the details on the profile specified by the handle ID */ map = ice_search_prof_id(hw, blk, hdl); - if (!map) - return ICE_ERR_DOES_NOT_EXIST; - - for (i = 0; i < map->ptype_count; i++) { - enum ice_status status; - bool add; - u8 ptg; - - status = ice_get_ptg(hw, blk, map->ptype[i], &ptg, &add); - if (status) - goto err_ice_get_prof_ptgs; + if (!map) { + status = ICE_ERR_DOES_NOT_EXIST; + goto err_ice_get_prof; + } - if (add || !hw->blk[blk].es.ref_count[map->prof_id]) { - /* add PTG to change list */ + for (i = 0; i < map->ptg_cnt; i++) + if (!hw->blk[blk].es.written[map->prof_id]) { + /* add ES to change list */ p = (struct ice_chs_chg *)ice_malloc(hw, sizeof(*p)); - if (!p) - goto err_ice_get_prof_ptgs; + if (!p) { + status = ICE_ERR_NO_MEMORY; + goto err_ice_get_prof; + } p->type = ICE_PTG_ES_ADD; - p->ptype = map->ptype[i]; - p->ptg = ptg; - p->add_ptg = add; + p->ptype = 0; + p->ptg = map->ptg[i]; + p->attr = map->attr[i]; + p->add_ptg = 0; - p->add_prof = !hw->blk[blk].es.ref_count[map->prof_id]; + p->add_prof = 1; p->prof_id = map->prof_id; + hw->blk[blk].es.written[map->prof_id] = true; + LIST_ADD(&p->list_entry, chg); } - } - - return ICE_SUCCESS; -err_ice_get_prof_ptgs: +err_ice_get_prof: + ice_release_lock(&hw->blk[blk].es.prof_map_lock); /* let caller clean up the change list */ - return ICE_ERR_NO_MEMORY; + return status; } /** @@ -4240,13 +5046,12 @@ ice_get_profs_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, struct ice_vsig_prof *p; /* copy to the input list */ - p = (struct ice_vsig_prof *)ice_malloc(hw, sizeof(*p)); + p = (struct ice_vsig_prof *)ice_memdup(hw, ent1, sizeof(*p), + ICE_NONDMA_TO_NONDMA); if (!p) goto err_ice_get_profs_vsig; - ice_memcpy(p, ent1, sizeof(*p), ICE_NONDMA_TO_NONDMA); - - LIST_ADD(&p->list, lst); + LIST_ADD_TAIL(&p->list, lst); } return ICE_SUCCESS; @@ -4271,36 +5076,40 @@ static enum ice_status ice_add_prof_to_lst(struct ice_hw *hw, enum ice_block blk, struct LIST_HEAD_TYPE *lst, u64 hdl) { - struct ice_vsig_prof *p; + enum ice_status status = ICE_SUCCESS; struct ice_prof_map *map; + struct ice_vsig_prof *p; u16 i; + ice_acquire_lock(&hw->blk[blk].es.prof_map_lock); map = ice_search_prof_id(hw, blk, hdl); - if (!map) - return ICE_ERR_DOES_NOT_EXIST; + if (!map) { + status = ICE_ERR_DOES_NOT_EXIST; + goto err_ice_add_prof_to_lst; + } p = (struct ice_vsig_prof *)ice_malloc(hw, sizeof(*p)); - if (!p) - return ICE_ERR_NO_MEMORY; + if (!p) { + status = ICE_ERR_NO_MEMORY; + goto err_ice_add_prof_to_lst; + } p->profile_cookie = map->profile_cookie; p->prof_id = map->prof_id; - p->tcam_count = map->ptype_count; - - for (i = 0; i < map->ptype_count; i++) { - u8 ptg; + p->tcam_count = map->ptg_cnt; + for (i = 0; i < map->ptg_cnt; i++) { p->tcam[i].prof_id = map->prof_id; p->tcam[i].tcam_idx = ICE_INVALID_TCAM; - - ice_ptg_find_ptype(hw, blk, map->ptype[i], &ptg); - - p->tcam[i].ptg = ptg; + p->tcam[i].ptg = map->ptg[i]; + p->tcam[i].attr = map->attr[i]; } LIST_ADD(&p->list, lst); - return ICE_SUCCESS; +err_ice_add_prof_to_lst: + ice_release_lock(&hw->blk[blk].es.prof_map_lock); + return status; } /** @@ -4326,6 +5135,7 @@ ice_move_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig, status = ice_vsig_find_vsi(hw, blk, vsi, &orig_vsig); if (!status) status = ice_vsig_add_mv_vsi(hw, blk, vsi, vsig); + if (status) { ice_free(hw, p); return status; @@ -4341,12 +5151,44 @@ ice_move_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig, return ICE_SUCCESS; } +/** + * ice_set_tcam_flags - set TCAM flag don't care mask + * @mask: mask for flags + * @dc_mask: pointer to the don't care mask + */ +static void ice_set_tcam_flags(u16 mask, u8 dc_mask[ICE_TCAM_KEY_VAL_SZ]) +{ + u16 *flag_word; + + /* flags are lowest u16 */ + flag_word = (u16 *)dc_mask; + *flag_word = ~mask; +} + +/** + * ice_rem_chg_tcam_ent - remove a specific TCAM entry from change list + * @hw: pointer to the HW struct + * @idx: the index of the TCAM entry to remove + * @chg: the list of change structures to search + */ +static void +ice_rem_chg_tcam_ent(struct ice_hw *hw, u16 idx, struct LIST_HEAD_TYPE *chg) +{ + struct ice_chs_chg *pos, *tmp; + + LIST_FOR_EACH_ENTRY_SAFE(tmp, pos, chg, ice_chs_chg, list_entry) + if (tmp->type == ICE_TCAM_ADD && tmp->tcam_idx == idx) { + LIST_DEL(&tmp->list_entry); + ice_free(hw, tmp); + } +} + /** * ice_prof_tcam_ena_dis - add enable or disable TCAM change * @hw: pointer to the HW struct * @blk: hardware block * @enable: true to enable, false to disable - * @vsig: the vsig of the TCAM entry + * @vsig: the VSIG of the TCAM entry * @tcam: pointer the TCAM info structure of the TCAM to disable * @chg: the change list * @@ -4360,29 +5202,49 @@ ice_prof_tcam_ena_dis(struct ice_hw *hw, enum ice_block blk, bool enable, enum ice_status status; struct ice_chs_chg *p; - /* Default: enable means change the low flag bit to don't care */ - u8 dc_msk[ICE_TCAM_KEY_VAL_SZ] = { 0x01, 0x00, 0x00, 0x00, 0x00 }; + u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF }; + u8 dc_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0x00, 0x00, 0x00 }; u8 nm_msk[ICE_TCAM_KEY_VAL_SZ] = { 0x00, 0x00, 0x00, 0x00, 0x00 }; - u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0x01, 0x00, 0x00, 0x00, 0x00 }; - /* If disabled, change the low flag bit to never match */ + /* if disabling, free the TCAM */ if (!enable) { - dc_msk[0] = 0x00; - nm_msk[0] = 0x01; + status = ice_rel_tcam_idx(hw, blk, tcam->tcam_idx); + + /* if we have already created a change for this TCAM entry, then + * we need to remove that entry, in order to prevent writing to + * a TCAM entry we no longer will have ownership of. + */ + ice_rem_chg_tcam_ent(hw, tcam->tcam_idx, chg); + tcam->tcam_idx = 0; + tcam->in_use = 0; + return status; } + /* for re-enabling, reallocate a TCAM */ + /* for entries with empty attribute masks, allocate entry from + * the bottom of the tcam table; otherwise, allocate from the + * top of the table in order to give it higher priority + */ + status = ice_alloc_tcam_ent(hw, blk, tcam->attr.mask == 0, + &tcam->tcam_idx); + if (status) + return status; + /* add TCAM to change list */ p = (struct ice_chs_chg *)ice_malloc(hw, sizeof(*p)); if (!p) return ICE_ERR_NO_MEMORY; + /* set don't care masks for TCAM flags */ + ice_set_tcam_flags(tcam->attr.mask, dc_msk); + status = ice_tcam_write_entry(hw, blk, tcam->tcam_idx, tcam->prof_id, - tcam->ptg, vsig, 0, 0, vl_msk, dc_msk, - nm_msk); + tcam->ptg, vsig, 0, tcam->attr.flags, + vl_msk, dc_msk, nm_msk); if (status) goto err_ice_prof_tcam_ena_dis; - tcam->in_use = enable; + tcam->in_use = 1; p->type = ICE_TCAM_ADD; p->add_tcam_idx = true; @@ -4401,6 +5263,32 @@ err_ice_prof_tcam_ena_dis: return status; } +/** + * ice_ptg_attr_in_use - determine if PTG and attribute pair is in use + * @ptg_attr: pointer to the PTG and attribute pair to check + * @ptgs_used: bitmap that denotes which PTGs are in use + * @attr_used: array of PTG and attributes pairs already used + * @attr_cnt: count of entries in the attr_used array + */ +static bool +ice_ptg_attr_in_use(struct ice_tcam_inf *ptg_attr, ice_bitmap_t *ptgs_used, + struct ice_tcam_inf *attr_used[], u16 attr_cnt) +{ + u16 i; + + if (!ice_is_bit_set(ptgs_used, ptg_attr->ptg)) + return false; + + /* the PTG is used, so now look for correct attributes */ + for (i = 0; i < attr_cnt; i++) + if (attr_used[i]->ptg == ptg_attr->ptg && + attr_used[i]->attr.flags == ptg_attr->attr.flags && + attr_used[i]->attr.mask == ptg_attr->attr.mask) + return true; + + return false; +} + /** * ice_adj_prof_priorities - adjust profile based on priorities * @hw: pointer to the HW struct @@ -4413,11 +5301,19 @@ ice_adj_prof_priorities(struct ice_hw *hw, enum ice_block blk, u16 vsig, struct LIST_HEAD_TYPE *chg) { ice_declare_bitmap(ptgs_used, ICE_XLT1_CNT); + struct ice_tcam_inf **attr_used; + enum ice_status status = ICE_SUCCESS; struct ice_vsig_prof *t; - enum ice_status status; + u16 attr_used_cnt = 0; u16 idx; - ice_memset(ptgs_used, 0, sizeof(ptgs_used), ICE_NONDMA_MEM); +#define ICE_MAX_PTG_ATTRS 1024 + attr_used = (struct ice_tcam_inf **)ice_calloc(hw, ICE_MAX_PTG_ATTRS, + sizeof(*attr_used)); + if (!attr_used) + return ICE_ERR_NO_MEMORY; + + ice_zero_bitmap(ptgs_used, ICE_XLT1_CNT); idx = vsig & ICE_VSIG_IDX_M; /* Priority is based on the order in which the profiles are added. The @@ -4434,11 +5330,15 @@ ice_adj_prof_priorities(struct ice_hw *hw, enum ice_block blk, u16 vsig, u16 i; for (i = 0; i < t->tcam_count; i++) { + bool used; + /* Scan the priorities from newest to oldest. * Make sure that the newest profiles take priority. */ - if (ice_is_bit_set(ptgs_used, t->tcam[i].ptg) && - t->tcam[i].in_use) { + used = ice_ptg_attr_in_use(&t->tcam[i], ptgs_used, + attr_used, attr_used_cnt); + + if (used && t->tcam[i].in_use) { /* need to mark this PTG as never match, as it * was already in use and therefore duplicate * (and lower priority) @@ -4448,9 +5348,8 @@ ice_adj_prof_priorities(struct ice_hw *hw, enum ice_block blk, u16 vsig, &t->tcam[i], chg); if (status) - return status; - } else if (!ice_is_bit_set(ptgs_used, t->tcam[i].ptg) && - !t->tcam[i].in_use) { + goto err_ice_adj_prof_priorities; + } else if (!used && !t->tcam[i].in_use) { /* need to enable this PTG, as it in not in use * and not enabled (highest priority) */ @@ -4459,15 +5358,22 @@ ice_adj_prof_priorities(struct ice_hw *hw, enum ice_block blk, u16 vsig, &t->tcam[i], chg); if (status) - return status; + goto err_ice_adj_prof_priorities; } /* keep track of used ptgs */ ice_set_bit(t->tcam[i].ptg, ptgs_used); + if (attr_used_cnt < ICE_MAX_PTG_ATTRS) + attr_used[attr_used_cnt++] = &t->tcam[i]; + else + ice_debug(hw, ICE_DBG_INIT, + "Warn: ICE_MAX_PTG_ATTRS exceeded\n"); } } - return ICE_SUCCESS; +err_ice_adj_prof_priorities: + ice_free(hw, attr_used); + return status; } /** @@ -4476,25 +5382,22 @@ ice_adj_prof_priorities(struct ice_hw *hw, enum ice_block blk, u16 vsig, * @blk: hardware block * @vsig: the VSIG to which this profile is to be added * @hdl: the profile handle indicating the profile to add + * @rev: true to add entries to the end of the list * @chg: the change list */ static enum ice_status ice_add_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, - struct LIST_HEAD_TYPE *chg) + bool rev, struct LIST_HEAD_TYPE *chg) { /* Masks that ignore flags */ u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF }; u8 dc_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0x00, 0x00, 0x00 }; u8 nm_msk[ICE_TCAM_KEY_VAL_SZ] = { 0x00, 0x00, 0x00, 0x00, 0x00 }; + enum ice_status status = ICE_SUCCESS; struct ice_prof_map *map; struct ice_vsig_prof *t; struct ice_chs_chg *p; - u16 i; - - /* Get the details on the profile specified by the handle ID */ - map = ice_search_prof_id(hw, blk, hdl); - if (!map) - return ICE_ERR_DOES_NOT_EXIST; + u16 vsig_idx, i; /* Error, if this VSIG already has this profile */ if (ice_has_prof_vsig(hw, blk, vsig, hdl)) @@ -4503,41 +5406,47 @@ ice_add_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, /* new VSIG profile structure */ t = (struct ice_vsig_prof *)ice_malloc(hw, sizeof(*t)); if (!t) + return ICE_ERR_NO_MEMORY; + + ice_acquire_lock(&hw->blk[blk].es.prof_map_lock); + /* Get the details on the profile specified by the handle ID */ + map = ice_search_prof_id(hw, blk, hdl); + if (!map) { + status = ICE_ERR_DOES_NOT_EXIST; goto err_ice_add_prof_id_vsig; + } t->profile_cookie = map->profile_cookie; t->prof_id = map->prof_id; - t->tcam_count = map->ptype_count; + t->tcam_count = map->ptg_cnt; /* create TCAM entries */ - for (i = 0; i < map->ptype_count; i++) { - enum ice_status status; + for (i = 0; i < map->ptg_cnt; i++) { u16 tcam_idx; - bool add; - u8 ptg; - - /* If properly sequenced, we should never have to allocate new - * PTGs - */ - status = ice_get_ptg(hw, blk, map->ptype[i], &ptg, &add); - if (status) - goto err_ice_add_prof_id_vsig; /* add TCAM to change list */ p = (struct ice_chs_chg *)ice_malloc(hw, sizeof(*p)); - if (!p) + if (!p) { + status = ICE_ERR_NO_MEMORY; goto err_ice_add_prof_id_vsig; + } /* allocate the TCAM entry index */ - status = ice_alloc_tcam_ent(hw, blk, &tcam_idx); + /* for entries with empty attribute masks, allocate entry from + * the bottom of the tcam table; otherwise, allocate from the + * top of the table in order to give it higher priority + */ + status = ice_alloc_tcam_ent(hw, blk, map->attr[i].mask == 0, + &tcam_idx); if (status) { ice_free(hw, p); goto err_ice_add_prof_id_vsig; } - t->tcam[i].ptg = ptg; + t->tcam[i].ptg = map->ptg[i]; t->tcam[i].prof_id = map->prof_id; t->tcam[i].tcam_idx = tcam_idx; + t->tcam[i].attr = map->attr[i]; t->tcam[i].in_use = true; p->type = ICE_TCAM_ADD; @@ -4547,33 +5456,41 @@ ice_add_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, p->vsig = vsig; p->tcam_idx = t->tcam[i].tcam_idx; + /* set don't care masks for TCAM flags */ + ice_set_tcam_flags(t->tcam[i].attr.mask, dc_msk); + /* write the TCAM entry */ status = ice_tcam_write_entry(hw, blk, t->tcam[i].tcam_idx, t->tcam[i].prof_id, - t->tcam[i].ptg, vsig, 0, 0, - vl_msk, dc_msk, nm_msk); - if (status) + t->tcam[i].ptg, vsig, 0, + t->tcam[i].attr.flags, vl_msk, + dc_msk, nm_msk); + if (status) { + ice_free(hw, p); goto err_ice_add_prof_id_vsig; - - /* this increments the reference count of how many TCAM entries - * are using this HW profile ID - */ - status = ice_prof_inc_ref(hw, blk, t->tcam[i].prof_id); + } /* log change */ LIST_ADD(&p->list_entry, chg); } /* add profile to VSIG */ - LIST_ADD(&t->list, - &hw->blk[blk].xlt2.vsig_tbl[(vsig & ICE_VSIG_IDX_M)].prop_lst); + vsig_idx = vsig & ICE_VSIG_IDX_M; + if (rev) + LIST_ADD_TAIL(&t->list, + &hw->blk[blk].xlt2.vsig_tbl[vsig_idx].prop_lst); + else + LIST_ADD(&t->list, + &hw->blk[blk].xlt2.vsig_tbl[vsig_idx].prop_lst); - return ICE_SUCCESS; + ice_release_lock(&hw->blk[blk].es.prof_map_lock); + return status; err_ice_add_prof_id_vsig: + ice_release_lock(&hw->blk[blk].es.prof_map_lock); /* let caller clean up the change list */ ice_free(hw, t); - return ICE_ERR_NO_MEMORY; + return status; } /** @@ -4606,7 +5523,7 @@ ice_create_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl, if (status) goto err_ice_create_prof_id_vsig; - status = ice_add_prof_id_vsig(hw, blk, new_vsig, hdl, chg); + status = ice_add_prof_id_vsig(hw, blk, new_vsig, hdl, false, chg); if (status) goto err_ice_create_prof_id_vsig; @@ -4626,16 +5543,18 @@ err_ice_create_prof_id_vsig: } /** - * ice_create_vsig_from_list - create a new VSIG with a list of profiles + * ice_create_vsig_from_lst - create a new VSIG with a list of profiles * @hw: pointer to the HW struct * @blk: hardware block * @vsi: the initial VSI that will be in VSIG * @lst: the list of profile that will be added to the VSIG + * @new_vsig: return of new VSIG * @chg: the change list */ static enum ice_status ice_create_vsig_from_lst(struct ice_hw *hw, enum ice_block blk, u16 vsi, - struct LIST_HEAD_TYPE *lst, struct LIST_HEAD_TYPE *chg) + struct LIST_HEAD_TYPE *lst, u16 *new_vsig, + struct LIST_HEAD_TYPE *chg) { struct ice_vsig_prof *t; enum ice_status status; @@ -4650,12 +5569,15 @@ ice_create_vsig_from_lst(struct ice_hw *hw, enum ice_block blk, u16 vsi, return status; LIST_FOR_EACH_ENTRY(t, lst, ice_vsig_prof, list) { + /* Reverse the order here since we are copying the list */ status = ice_add_prof_id_vsig(hw, blk, vsig, t->profile_cookie, - chg); + true, chg); if (status) return status; } + *new_vsig = vsig; + return ICE_SUCCESS; } @@ -4690,6 +5612,47 @@ ice_find_prof_vsig(struct ice_hw *hw, enum ice_block blk, u64 hdl, u16 *vsig) return status == ICE_SUCCESS; } +/** + * ice_add_vsi_flow - add VSI flow + * @hw: pointer to the HW struct + * @blk: hardware block + * @vsi: input VSI + * @vsig: target VSIG to include the input VSI + * + * Calling this function will add the VSI to a given VSIG and + * update the HW tables accordingly. This call can be used to + * add multiple VSIs to a VSIG if we know beforehand that those + * VSIs have the same characteristics of the VSIG. This will + * save time in generating a new VSIG and TCAMs till a match is + * found and subsequent rollback when a matching VSIG is found. + */ +enum ice_status +ice_add_vsi_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) +{ + struct ice_chs_chg *tmp, *del; + struct LIST_HEAD_TYPE chg; + enum ice_status status; + + /* if target VSIG is default the move is invalid */ + if ((vsig & ICE_VSIG_IDX_M) == ICE_DEFAULT_VSIG) + return ICE_ERR_PARAM; + + INIT_LIST_HEAD(&chg); + + /* move VSI to the VSIG that matches */ + status = ice_move_vsi(hw, blk, vsi, vsig, &chg); + /* update hardware if success */ + if (!status) + status = ice_upd_prof_hw(hw, blk, &chg); + + LIST_FOR_EACH_ENTRY_SAFE(del, tmp, &chg, ice_chs_chg, list_entry) { + LIST_DEL(&del->list_entry); + ice_free(hw, del); + } + + return status; +} + /** * ice_add_prof_id_flow - add profile flow * @hw: pointer to the HW struct @@ -4707,16 +5670,15 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) struct ice_vsig_prof *tmp1, *del1; struct LIST_HEAD_TYPE union_lst; struct ice_chs_chg *tmp, *del; - struct LIST_HEAD_TYPE chrs; struct LIST_HEAD_TYPE chg; enum ice_status status; - u16 vsig, or_vsig = 0; + u16 vsig; INIT_LIST_HEAD(&union_lst); - INIT_LIST_HEAD(&chrs); INIT_LIST_HEAD(&chg); - status = ice_get_prof_ptgs(hw, blk, hdl, &chg); + /* Get profile */ + status = ice_get_prof(hw, blk, hdl, &chg); if (status) return status; @@ -4724,9 +5686,10 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) status = ice_vsig_find_vsi(hw, blk, vsi, &vsig); if (!status && vsig) { bool only_vsi; + u16 or_vsig; u16 ref; - /* found in vsig */ + /* found in VSIG */ or_vsig = vsig; /* make sure that there is no overlap/conflict between the new @@ -4758,14 +5721,13 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) /* search for an existing VSIG with an exact charc match */ status = ice_find_dup_props_vsig(hw, blk, &union_lst, &vsig); if (!status) { - /* found an exact match */ - /* move vsi to the VSIG that matches */ + /* move VSI to the VSIG that matches */ status = ice_move_vsi(hw, blk, vsi, vsig, &chg); if (status) goto err_ice_add_prof_id_flow; - /* remove original VSIG if we just moved the only VSI - * from it + /* VSI has been moved out of or_vsig. If the or_vsig had + * only that VSI it is now empty and can be removed. */ if (only_vsi) { status = ice_rem_vsig(hw, blk, or_vsig, &chg); @@ -4778,7 +5740,8 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) * not sharing entries and we can simply add the new * profile to the VSIG. */ - status = ice_add_prof_id_vsig(hw, blk, vsig, hdl, &chg); + status = ice_add_prof_id_vsig(hw, blk, vsig, hdl, false, + &chg); if (status) goto err_ice_add_prof_id_flow; @@ -4789,7 +5752,8 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) } else { /* No match, so we need a new VSIG */ status = ice_create_vsig_from_lst(hw, blk, vsi, - &union_lst, &chg); + &union_lst, &vsig, + &chg); if (status) goto err_ice_add_prof_id_flow; @@ -4832,42 +5796,9 @@ err_ice_add_prof_id_flow: ice_free(hw, del1); } - LIST_FOR_EACH_ENTRY_SAFE(del1, tmp1, &chrs, ice_vsig_prof, list) { - LIST_DEL(&del1->list); - ice_free(hw, del1); - } - return status; } -/** - * ice_add_flow - add flow - * @hw: pointer to the HW struct - * @blk: hardware block - * @vsi: array of VSIs to enable with the profile specified by ID - * @count: number of elements in the VSI array - * @id: profile tracking ID - * - * Calling this function will update the hardware tables to enable the - * profile indicated by the ID parameter for the VSIs specified in the VSI - * array. Once successfully called, the flow will be enabled. - */ -enum ice_status -ice_add_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi[], u8 count, - u64 id) -{ - enum ice_status status; - u16 i; - - for (i = 0; i < count; i++) { - status = ice_add_prof_id_flow(hw, blk, vsi[i], id); - if (status) - return status; - } - - return ICE_SUCCESS; -} - /** * ice_rem_prof_from_list - remove a profile from list * @hw: pointer to the HW struct @@ -4879,13 +5810,12 @@ ice_rem_prof_from_list(struct ice_hw *hw, struct LIST_HEAD_TYPE *lst, u64 hdl) { struct ice_vsig_prof *ent, *tmp; - LIST_FOR_EACH_ENTRY_SAFE(ent, tmp, lst, ice_vsig_prof, list) { + LIST_FOR_EACH_ENTRY_SAFE(ent, tmp, lst, ice_vsig_prof, list) if (ent->profile_cookie == hdl) { LIST_DEL(&ent->list); ice_free(hw, ent); return ICE_SUCCESS; } - } return ICE_ERR_DOES_NOT_EXIST; } @@ -4989,7 +5919,8 @@ ice_rem_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) * new VSIG and TCAM entries */ status = ice_create_vsig_from_lst(hw, blk, vsi, - ©, &chg); + ©, &vsig, + &chg); if (status) goto err_ice_rem_prof_id_flow; @@ -5021,31 +5952,3 @@ err_ice_rem_prof_id_flow: return status; } - -/** - * ice_rem_flow - remove flow - * @hw: pointer to the HW struct - * @blk: hardware block - * @vsi: array of VSIs from which to remove the profile specified by ID - * @count: number of elements in the VSI array - * @id: profile tracking ID - * - * The function will remove flows from the specified VSIs that were enabled - * using ice_add_flow. The ID value will indicated which profile will be - * removed. Once successfully called, the flow will be disabled. - */ -enum ice_status -ice_rem_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi[], u8 count, - u64 id) -{ - enum ice_status status; - u16 i; - - for (i = 0; i < count; i++) { - status = ice_rem_prof_id_flow(hw, blk, vsi[i], id); - if (status) - return status; - } - - return ICE_SUCCESS; -}