diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-11 08:27:49 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-11 08:27:49 +0000 |
commit | ace9429bb58fd418f0c81d4c2835699bddf6bde6 (patch) | |
tree | b2d64bc10158fdd5497876388cd68142ca374ed3 /drivers/net/ethernet/amazon/ena | |
parent | Initial commit. (diff) | |
download | linux-ace9429bb58fd418f0c81d4c2835699bddf6bde6.tar.xz linux-ace9429bb58fd418f0c81d4c2835699bddf6bde6.zip |
Adding upstream version 6.6.15.upstream/6.6.15
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'drivers/net/ethernet/amazon/ena')
-rw-r--r-- | drivers/net/ethernet/amazon/ena/Makefile | 8 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_admin_defs.h | 1198 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_com.c | 3001 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_com.h | 1025 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_common_defs.h | 21 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_eth_com.c | 646 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_eth_com.h | 261 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_eth_io_defs.h | 390 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_ethtool.c | 1066 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_netdev.c | 4832 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_netdev.h | 467 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_pci_id_tbl.h | 45 | ||||
-rw-r--r-- | drivers/net/ethernet/amazon/ena/ena_regs_defs.h | 130 |
13 files changed, 13090 insertions, 0 deletions
diff --git a/drivers/net/ethernet/amazon/ena/Makefile b/drivers/net/ethernet/amazon/ena/Makefile new file mode 100644 index 0000000000..f1f752a8f7 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/Makefile @@ -0,0 +1,8 @@ +# SPDX-License-Identifier: GPL-2.0-only +# +# Makefile for the Elastic Network Adapter (ENA) device drivers. +# + +obj-$(CONFIG_ENA_ETHERNET) += ena.o + +ena-y := ena_netdev.o ena_com.o ena_eth_com.o ena_ethtool.o diff --git a/drivers/net/ethernet/amazon/ena/ena_admin_defs.h b/drivers/net/ethernet/amazon/ena/ena_admin_defs.h new file mode 100644 index 0000000000..6de0d590be --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_admin_defs.h @@ -0,0 +1,1198 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ +#ifndef _ENA_ADMIN_H_ +#define _ENA_ADMIN_H_ + +#define ENA_ADMIN_RSS_KEY_PARTS 10 + +enum ena_admin_aq_opcode { + ENA_ADMIN_CREATE_SQ = 1, + ENA_ADMIN_DESTROY_SQ = 2, + ENA_ADMIN_CREATE_CQ = 3, + ENA_ADMIN_DESTROY_CQ = 4, + ENA_ADMIN_GET_FEATURE = 8, + ENA_ADMIN_SET_FEATURE = 9, + ENA_ADMIN_GET_STATS = 11, +}; + +enum ena_admin_aq_completion_status { + ENA_ADMIN_SUCCESS = 0, + ENA_ADMIN_RESOURCE_ALLOCATION_FAILURE = 1, + ENA_ADMIN_BAD_OPCODE = 2, + ENA_ADMIN_UNSUPPORTED_OPCODE = 3, + ENA_ADMIN_MALFORMED_REQUEST = 4, + /* Additional status is provided in ACQ entry extended_status */ + ENA_ADMIN_ILLEGAL_PARAMETER = 5, + ENA_ADMIN_UNKNOWN_ERROR = 6, + ENA_ADMIN_RESOURCE_BUSY = 7, +}; + +/* subcommands for the set/get feature admin commands */ +enum ena_admin_aq_feature_id { + ENA_ADMIN_DEVICE_ATTRIBUTES = 1, + ENA_ADMIN_MAX_QUEUES_NUM = 2, + ENA_ADMIN_HW_HINTS = 3, + ENA_ADMIN_LLQ = 4, + ENA_ADMIN_MAX_QUEUES_EXT = 7, + ENA_ADMIN_RSS_HASH_FUNCTION = 10, + ENA_ADMIN_STATELESS_OFFLOAD_CONFIG = 11, + ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG = 12, + ENA_ADMIN_MTU = 14, + ENA_ADMIN_RSS_HASH_INPUT = 18, + ENA_ADMIN_INTERRUPT_MODERATION = 20, + ENA_ADMIN_AENQ_CONFIG = 26, + ENA_ADMIN_LINK_CONFIG = 27, + ENA_ADMIN_HOST_ATTR_CONFIG = 28, + ENA_ADMIN_FEATURES_OPCODE_NUM = 32, +}; + +/* device capabilities */ +enum ena_admin_aq_caps_id { + ENA_ADMIN_ENI_STATS = 0, +}; + +enum ena_admin_placement_policy_type { + /* descriptors and headers are in host memory */ + ENA_ADMIN_PLACEMENT_POLICY_HOST = 1, + /* descriptors and headers are in device memory (a.k.a Low Latency + * Queue) + */ + ENA_ADMIN_PLACEMENT_POLICY_DEV = 3, +}; + +enum ena_admin_link_types { + ENA_ADMIN_LINK_SPEED_1G = 0x1, + ENA_ADMIN_LINK_SPEED_2_HALF_G = 0x2, + ENA_ADMIN_LINK_SPEED_5G = 0x4, + ENA_ADMIN_LINK_SPEED_10G = 0x8, + ENA_ADMIN_LINK_SPEED_25G = 0x10, + ENA_ADMIN_LINK_SPEED_40G = 0x20, + ENA_ADMIN_LINK_SPEED_50G = 0x40, + ENA_ADMIN_LINK_SPEED_100G = 0x80, + ENA_ADMIN_LINK_SPEED_200G = 0x100, + ENA_ADMIN_LINK_SPEED_400G = 0x200, +}; + +enum ena_admin_completion_policy_type { + /* completion queue entry for each sq descriptor */ + ENA_ADMIN_COMPLETION_POLICY_DESC = 0, + /* completion queue entry upon request in sq descriptor */ + ENA_ADMIN_COMPLETION_POLICY_DESC_ON_DEMAND = 1, + /* current queue head pointer is updated in OS memory upon sq + * descriptor request + */ + ENA_ADMIN_COMPLETION_POLICY_HEAD_ON_DEMAND = 2, + /* current queue head pointer is updated in OS memory for each sq + * descriptor + */ + ENA_ADMIN_COMPLETION_POLICY_HEAD = 3, +}; + +/* basic stats return ena_admin_basic_stats while extanded stats return a + * buffer (string format) with additional statistics per queue and per + * device id + */ +enum ena_admin_get_stats_type { + ENA_ADMIN_GET_STATS_TYPE_BASIC = 0, + ENA_ADMIN_GET_STATS_TYPE_EXTENDED = 1, + /* extra HW stats for specific network interface */ + ENA_ADMIN_GET_STATS_TYPE_ENI = 2, +}; + +enum ena_admin_get_stats_scope { + ENA_ADMIN_SPECIFIC_QUEUE = 0, + ENA_ADMIN_ETH_TRAFFIC = 1, +}; + +struct ena_admin_aq_common_desc { + /* 11:0 : command_id + * 15:12 : reserved12 + */ + u16 command_id; + + /* as appears in ena_admin_aq_opcode */ + u8 opcode; + + /* 0 : phase + * 1 : ctrl_data - control buffer address valid + * 2 : ctrl_data_indirect - control buffer address + * points to list of pages with addresses of control + * buffers + * 7:3 : reserved3 + */ + u8 flags; +}; + +/* used in ena_admin_aq_entry. Can point directly to control data, or to a + * page list chunk. Used also at the end of indirect mode page list chunks, + * for chaining. + */ +struct ena_admin_ctrl_buff_info { + u32 length; + + struct ena_common_mem_addr address; +}; + +struct ena_admin_sq { + u16 sq_idx; + + /* 4:0 : reserved + * 7:5 : sq_direction - 0x1 - Tx; 0x2 - Rx + */ + u8 sq_identity; + + u8 reserved1; +}; + +struct ena_admin_aq_entry { + struct ena_admin_aq_common_desc aq_common_descriptor; + + union { + u32 inline_data_w1[3]; + + struct ena_admin_ctrl_buff_info control_buffer; + } u; + + u32 inline_data_w4[12]; +}; + +struct ena_admin_acq_common_desc { + /* command identifier to associate it with the aq descriptor + * 11:0 : command_id + * 15:12 : reserved12 + */ + u16 command; + + u8 status; + + /* 0 : phase + * 7:1 : reserved1 + */ + u8 flags; + + u16 extended_status; + + /* indicates to the driver which AQ entry has been consumed by the + * device and could be reused + */ + u16 sq_head_indx; +}; + +struct ena_admin_acq_entry { + struct ena_admin_acq_common_desc acq_common_descriptor; + + u32 response_specific_data[14]; +}; + +struct ena_admin_aq_create_sq_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + /* 4:0 : reserved0_w1 + * 7:5 : sq_direction - 0x1 - Tx, 0x2 - Rx + */ + u8 sq_identity; + + u8 reserved8_w1; + + /* 3:0 : placement_policy - Describing where the SQ + * descriptor ring and the SQ packet headers reside: + * 0x1 - descriptors and headers are in OS memory, + * 0x3 - descriptors and headers in device memory + * (a.k.a Low Latency Queue) + * 6:4 : completion_policy - Describing what policy + * to use for generation completion entry (cqe) in + * the CQ associated with this SQ: 0x0 - cqe for each + * sq descriptor, 0x1 - cqe upon request in sq + * descriptor, 0x2 - current queue head pointer is + * updated in OS memory upon sq descriptor request + * 0x3 - current queue head pointer is updated in OS + * memory for each sq descriptor + * 7 : reserved15_w1 + */ + u8 sq_caps_2; + + /* 0 : is_physically_contiguous - Described if the + * queue ring memory is allocated in physical + * contiguous pages or split. + * 7:1 : reserved17_w1 + */ + u8 sq_caps_3; + + /* associated completion queue id. This CQ must be created prior to SQ + * creation + */ + u16 cq_idx; + + /* submission queue depth in entries */ + u16 sq_depth; + + /* SQ physical base address in OS memory. This field should not be + * used for Low Latency queues. Has to be page aligned. + */ + struct ena_common_mem_addr sq_ba; + + /* specifies queue head writeback location in OS memory. Valid if + * completion_policy is set to completion_policy_head_on_demand or + * completion_policy_head. Has to be cache aligned + */ + struct ena_common_mem_addr sq_head_writeback; + + u32 reserved0_w7; + + u32 reserved0_w8; +}; + +enum ena_admin_sq_direction { + ENA_ADMIN_SQ_DIRECTION_TX = 1, + ENA_ADMIN_SQ_DIRECTION_RX = 2, +}; + +struct ena_admin_acq_create_sq_resp_desc { + struct ena_admin_acq_common_desc acq_common_desc; + + u16 sq_idx; + + u16 reserved; + + /* queue doorbell address as an offset to PCIe MMIO REG BAR */ + u32 sq_doorbell_offset; + + /* low latency queue ring base address as an offset to PCIe MMIO + * LLQ_MEM BAR + */ + u32 llq_descriptors_offset; + + /* low latency queue headers' memory as an offset to PCIe MMIO + * LLQ_MEM BAR + */ + u32 llq_headers_offset; +}; + +struct ena_admin_aq_destroy_sq_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + struct ena_admin_sq sq; +}; + +struct ena_admin_acq_destroy_sq_resp_desc { + struct ena_admin_acq_common_desc acq_common_desc; +}; + +struct ena_admin_aq_create_cq_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + /* 4:0 : reserved5 + * 5 : interrupt_mode_enabled - if set, cq operates + * in interrupt mode, otherwise - polling + * 7:6 : reserved6 + */ + u8 cq_caps_1; + + /* 4:0 : cq_entry_size_words - size of CQ entry in + * 32-bit words, valid values: 4, 8. + * 7:5 : reserved7 + */ + u8 cq_caps_2; + + /* completion queue depth in # of entries. must be power of 2 */ + u16 cq_depth; + + /* msix vector assigned to this cq */ + u32 msix_vector; + + /* cq physical base address in OS memory. CQ must be physically + * contiguous + */ + struct ena_common_mem_addr cq_ba; +}; + +struct ena_admin_acq_create_cq_resp_desc { + struct ena_admin_acq_common_desc acq_common_desc; + + u16 cq_idx; + + /* actual cq depth in number of entries */ + u16 cq_actual_depth; + + u32 numa_node_register_offset; + + u32 cq_head_db_register_offset; + + u32 cq_interrupt_unmask_register_offset; +}; + +struct ena_admin_aq_destroy_cq_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + u16 cq_idx; + + u16 reserved1; +}; + +struct ena_admin_acq_destroy_cq_resp_desc { + struct ena_admin_acq_common_desc acq_common_desc; +}; + +/* ENA AQ Get Statistics command. Extended statistics are placed in control + * buffer pointed by AQ entry + */ +struct ena_admin_aq_get_stats_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + union { + /* command specific inline data */ + u32 inline_data_w1[3]; + + struct ena_admin_ctrl_buff_info control_buffer; + } u; + + /* stats type as defined in enum ena_admin_get_stats_type */ + u8 type; + + /* stats scope defined in enum ena_admin_get_stats_scope */ + u8 scope; + + u16 reserved3; + + /* queue id. used when scope is specific_queue */ + u16 queue_idx; + + /* device id, value 0xFFFF means mine. only privileged device can get + * stats of other device + */ + u16 device_id; +}; + +/* Basic Statistics Command. */ +struct ena_admin_basic_stats { + u32 tx_bytes_low; + + u32 tx_bytes_high; + + u32 tx_pkts_low; + + u32 tx_pkts_high; + + u32 rx_bytes_low; + + u32 rx_bytes_high; + + u32 rx_pkts_low; + + u32 rx_pkts_high; + + u32 rx_drops_low; + + u32 rx_drops_high; + + u32 tx_drops_low; + + u32 tx_drops_high; +}; + +/* ENI Statistics Command. */ +struct ena_admin_eni_stats { + /* The number of packets shaped due to inbound aggregate BW + * allowance being exceeded + */ + u64 bw_in_allowance_exceeded; + + /* The number of packets shaped due to outbound aggregate BW + * allowance being exceeded + */ + u64 bw_out_allowance_exceeded; + + /* The number of packets shaped due to PPS allowance being exceeded */ + u64 pps_allowance_exceeded; + + /* The number of packets shaped due to connection tracking + * allowance being exceeded and leading to failure in establishment + * of new connections + */ + u64 conntrack_allowance_exceeded; + + /* The number of packets shaped due to linklocal packet rate + * allowance being exceeded + */ + u64 linklocal_allowance_exceeded; +}; + +struct ena_admin_acq_get_stats_resp { + struct ena_admin_acq_common_desc acq_common_desc; + + union { + u64 raw[7]; + + struct ena_admin_basic_stats basic_stats; + + struct ena_admin_eni_stats eni_stats; + } u; +}; + +struct ena_admin_get_set_feature_common_desc { + /* 1:0 : select - 0x1 - current value; 0x3 - default + * value + * 7:3 : reserved3 + */ + u8 flags; + + /* as appears in ena_admin_aq_feature_id */ + u8 feature_id; + + /* The driver specifies the max feature version it supports and the + * device responds with the currently supported feature version. The + * field is zero based + */ + u8 feature_version; + + u8 reserved8; +}; + +struct ena_admin_device_attr_feature_desc { + u32 impl_id; + + u32 device_version; + + /* bitmap of ena_admin_aq_feature_id, which represents supported + * subcommands for the set/get feature admin commands. + */ + u32 supported_features; + + /* bitmap of ena_admin_aq_caps_id, which represents device + * capabilities. + */ + u32 capabilities; + + /* Indicates how many bits are used physical address access. */ + u32 phys_addr_width; + + /* Indicates how many bits are used virtual address access. */ + u32 virt_addr_width; + + /* unicast MAC address (in Network byte order) */ + u8 mac_addr[6]; + + u8 reserved7[2]; + + u32 max_mtu; +}; + +enum ena_admin_llq_header_location { + /* header is in descriptor list */ + ENA_ADMIN_INLINE_HEADER = 1, + /* header in a separate ring, implies 16B descriptor list entry */ + ENA_ADMIN_HEADER_RING = 2, +}; + +enum ena_admin_llq_ring_entry_size { + ENA_ADMIN_LIST_ENTRY_SIZE_128B = 1, + ENA_ADMIN_LIST_ENTRY_SIZE_192B = 2, + ENA_ADMIN_LIST_ENTRY_SIZE_256B = 4, +}; + +enum ena_admin_llq_num_descs_before_header { + ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_0 = 0, + ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_1 = 1, + ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_2 = 2, + ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_4 = 4, + ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_8 = 8, +}; + +/* packet descriptor list entry always starts with one or more descriptors, + * followed by a header. The rest of the descriptors are located in the + * beginning of the subsequent entry. Stride refers to how the rest of the + * descriptors are placed. This field is relevant only for inline header + * mode + */ +enum ena_admin_llq_stride_ctrl { + ENA_ADMIN_SINGLE_DESC_PER_ENTRY = 1, + ENA_ADMIN_MULTIPLE_DESCS_PER_ENTRY = 2, +}; + +enum ena_admin_accel_mode_feat { + ENA_ADMIN_DISABLE_META_CACHING = 0, + ENA_ADMIN_LIMIT_TX_BURST = 1, +}; + +struct ena_admin_accel_mode_get { + /* bit field of enum ena_admin_accel_mode_feat */ + u16 supported_flags; + + /* maximum burst size between two doorbells. The size is in bytes */ + u16 max_tx_burst_size; +}; + +struct ena_admin_accel_mode_set { + /* bit field of enum ena_admin_accel_mode_feat */ + u16 enabled_flags; + + u16 reserved; +}; + +struct ena_admin_accel_mode_req { + union { + u32 raw[2]; + + struct ena_admin_accel_mode_get get; + + struct ena_admin_accel_mode_set set; + } u; +}; + +struct ena_admin_feature_llq_desc { + u32 max_llq_num; + + u32 max_llq_depth; + + /* specify the header locations the device supports. bitfield of enum + * ena_admin_llq_header_location. + */ + u16 header_location_ctrl_supported; + + /* the header location the driver selected to use. */ + u16 header_location_ctrl_enabled; + + /* if inline header is specified - this is the size of descriptor list + * entry. If header in a separate ring is specified - this is the size + * of header ring entry. bitfield of enum ena_admin_llq_ring_entry_size. + * specify the entry sizes the device supports + */ + u16 entry_size_ctrl_supported; + + /* the entry size the driver selected to use. */ + u16 entry_size_ctrl_enabled; + + /* valid only if inline header is specified. First entry associated with + * the packet includes descriptors and header. Rest of the entries + * occupied by descriptors. This parameter defines the max number of + * descriptors precedding the header in the first entry. The field is + * bitfield of enum ena_admin_llq_num_descs_before_header and specify + * the values the device supports + */ + u16 desc_num_before_header_supported; + + /* the desire field the driver selected to use */ + u16 desc_num_before_header_enabled; + + /* valid only if inline was chosen. bitfield of enum + * ena_admin_llq_stride_ctrl + */ + u16 descriptors_stride_ctrl_supported; + + /* the stride control the driver selected to use */ + u16 descriptors_stride_ctrl_enabled; + + /* reserved */ + u32 reserved1; + + /* accelerated low latency queues requirement. driver needs to + * support those requirements in order to use accelerated llq + */ + struct ena_admin_accel_mode_req accel_mode; +}; + +struct ena_admin_queue_ext_feature_fields { + u32 max_tx_sq_num; + + u32 max_tx_cq_num; + + u32 max_rx_sq_num; + + u32 max_rx_cq_num; + + u32 max_tx_sq_depth; + + u32 max_tx_cq_depth; + + u32 max_rx_sq_depth; + + u32 max_rx_cq_depth; + + u32 max_tx_header_size; + + /* Maximum Descriptors number, including meta descriptor, allowed for a + * single Tx packet + */ + u16 max_per_packet_tx_descs; + + /* Maximum Descriptors number allowed for a single Rx packet */ + u16 max_per_packet_rx_descs; +}; + +struct ena_admin_queue_feature_desc { + u32 max_sq_num; + + u32 max_sq_depth; + + u32 max_cq_num; + + u32 max_cq_depth; + + u32 max_legacy_llq_num; + + u32 max_legacy_llq_depth; + + u32 max_header_size; + + /* Maximum Descriptors number, including meta descriptor, allowed for a + * single Tx packet + */ + u16 max_packet_tx_descs; + + /* Maximum Descriptors number allowed for a single Rx packet */ + u16 max_packet_rx_descs; +}; + +struct ena_admin_set_feature_mtu_desc { + /* exclude L2 */ + u32 mtu; +}; + +struct ena_admin_set_feature_host_attr_desc { + /* host OS info base address in OS memory. host info is 4KB of + * physically contiguous + */ + struct ena_common_mem_addr os_info_ba; + + /* host debug area base address in OS memory. debug area must be + * physically contiguous + */ + struct ena_common_mem_addr debug_ba; + + /* debug area size */ + u32 debug_area_size; +}; + +struct ena_admin_feature_intr_moder_desc { + /* interrupt delay granularity in usec */ + u16 intr_delay_resolution; + + u16 reserved; +}; + +struct ena_admin_get_feature_link_desc { + /* Link speed in Mb */ + u32 speed; + + /* bit field of enum ena_admin_link types */ + u32 supported; + + /* 0 : autoneg + * 1 : duplex - Full Duplex + * 31:2 : reserved2 + */ + u32 flags; +}; + +struct ena_admin_feature_aenq_desc { + /* bitmask for AENQ groups the device can report */ + u32 supported_groups; + + /* bitmask for AENQ groups to report */ + u32 enabled_groups; +}; + +struct ena_admin_feature_offload_desc { + /* 0 : TX_L3_csum_ipv4 + * 1 : TX_L4_ipv4_csum_part - The checksum field + * should be initialized with pseudo header checksum + * 2 : TX_L4_ipv4_csum_full + * 3 : TX_L4_ipv6_csum_part - The checksum field + * should be initialized with pseudo header checksum + * 4 : TX_L4_ipv6_csum_full + * 5 : tso_ipv4 + * 6 : tso_ipv6 + * 7 : tso_ecn + */ + u32 tx; + + /* Receive side supported stateless offload + * 0 : RX_L3_csum_ipv4 - IPv4 checksum + * 1 : RX_L4_ipv4_csum - TCP/UDP/IPv4 checksum + * 2 : RX_L4_ipv6_csum - TCP/UDP/IPv6 checksum + * 3 : RX_hash - Hash calculation + */ + u32 rx_supported; + + u32 rx_enabled; +}; + +enum ena_admin_hash_functions { + ENA_ADMIN_TOEPLITZ = 1, + ENA_ADMIN_CRC32 = 2, +}; + +struct ena_admin_feature_rss_flow_hash_control { + u32 key_parts; + + u32 reserved; + + u32 key[ENA_ADMIN_RSS_KEY_PARTS]; +}; + +struct ena_admin_feature_rss_flow_hash_function { + /* 7:0 : funcs - bitmask of ena_admin_hash_functions */ + u32 supported_func; + + /* 7:0 : selected_func - bitmask of + * ena_admin_hash_functions + */ + u32 selected_func; + + /* initial value */ + u32 init_val; +}; + +/* RSS flow hash protocols */ +enum ena_admin_flow_hash_proto { + ENA_ADMIN_RSS_TCP4 = 0, + ENA_ADMIN_RSS_UDP4 = 1, + ENA_ADMIN_RSS_TCP6 = 2, + ENA_ADMIN_RSS_UDP6 = 3, + ENA_ADMIN_RSS_IP4 = 4, + ENA_ADMIN_RSS_IP6 = 5, + ENA_ADMIN_RSS_IP4_FRAG = 6, + ENA_ADMIN_RSS_NOT_IP = 7, + /* TCPv6 with extension header */ + ENA_ADMIN_RSS_TCP6_EX = 8, + /* IPv6 with extension header */ + ENA_ADMIN_RSS_IP6_EX = 9, + ENA_ADMIN_RSS_PROTO_NUM = 16, +}; + +/* RSS flow hash fields */ +enum ena_admin_flow_hash_fields { + /* Ethernet Dest Addr */ + ENA_ADMIN_RSS_L2_DA = BIT(0), + /* Ethernet Src Addr */ + ENA_ADMIN_RSS_L2_SA = BIT(1), + /* ipv4/6 Dest Addr */ + ENA_ADMIN_RSS_L3_DA = BIT(2), + /* ipv4/6 Src Addr */ + ENA_ADMIN_RSS_L3_SA = BIT(3), + /* tcp/udp Dest Port */ + ENA_ADMIN_RSS_L4_DP = BIT(4), + /* tcp/udp Src Port */ + ENA_ADMIN_RSS_L4_SP = BIT(5), +}; + +struct ena_admin_proto_input { + /* flow hash fields (bitwise according to ena_admin_flow_hash_fields) */ + u16 fields; + + u16 reserved2; +}; + +struct ena_admin_feature_rss_hash_control { + struct ena_admin_proto_input supported_fields[ENA_ADMIN_RSS_PROTO_NUM]; + + struct ena_admin_proto_input selected_fields[ENA_ADMIN_RSS_PROTO_NUM]; + + struct ena_admin_proto_input reserved2[ENA_ADMIN_RSS_PROTO_NUM]; + + struct ena_admin_proto_input reserved3[ENA_ADMIN_RSS_PROTO_NUM]; +}; + +struct ena_admin_feature_rss_flow_hash_input { + /* supported hash input sorting + * 1 : L3_sort - support swap L3 addresses if DA is + * smaller than SA + * 2 : L4_sort - support swap L4 ports if DP smaller + * SP + */ + u16 supported_input_sort; + + /* enabled hash input sorting + * 1 : enable_L3_sort - enable swap L3 addresses if + * DA smaller than SA + * 2 : enable_L4_sort - enable swap L4 ports if DP + * smaller than SP + */ + u16 enabled_input_sort; +}; + +enum ena_admin_os_type { + ENA_ADMIN_OS_LINUX = 1, + ENA_ADMIN_OS_WIN = 2, + ENA_ADMIN_OS_DPDK = 3, + ENA_ADMIN_OS_FREEBSD = 4, + ENA_ADMIN_OS_IPXE = 5, + ENA_ADMIN_OS_ESXI = 6, + ENA_ADMIN_OS_GROUPS_NUM = 6, +}; + +struct ena_admin_host_info { + /* defined in enum ena_admin_os_type */ + u32 os_type; + + /* os distribution string format */ + u8 os_dist_str[128]; + + /* OS distribution numeric format */ + u32 os_dist; + + /* kernel version string format */ + u8 kernel_ver_str[32]; + + /* Kernel version numeric format */ + u32 kernel_ver; + + /* 7:0 : major + * 15:8 : minor + * 23:16 : sub_minor + * 31:24 : module_type + */ + u32 driver_version; + + /* features bitmap */ + u32 supported_network_features[2]; + + /* ENA spec version of driver */ + u16 ena_spec_version; + + /* ENA device's Bus, Device and Function + * 2:0 : function + * 7:3 : device + * 15:8 : bus + */ + u16 bdf; + + /* Number of CPUs */ + u16 num_cpus; + + u16 reserved; + + /* 0 : reserved + * 1 : rx_offset + * 2 : interrupt_moderation + * 3 : rx_buf_mirroring + * 4 : rss_configurable_function_key + * 5 : reserved + * 6 : rx_page_reuse + * 31:7 : reserved + */ + u32 driver_supported_features; +}; + +struct ena_admin_rss_ind_table_entry { + u16 cq_idx; + + u16 reserved; +}; + +struct ena_admin_feature_rss_ind_table { + /* min supported table size (2^min_size) */ + u16 min_size; + + /* max supported table size (2^max_size) */ + u16 max_size; + + /* table size (2^size) */ + u16 size; + + u16 reserved; + + /* index of the inline entry. 0xFFFFFFFF means invalid */ + u32 inline_index; + + /* used for updating single entry, ignored when setting the entire + * table through the control buffer. + */ + struct ena_admin_rss_ind_table_entry inline_entry; +}; + +/* When hint value is 0, driver should use it's own predefined value */ +struct ena_admin_ena_hw_hints { + /* value in ms */ + u16 mmio_read_timeout; + + /* value in ms */ + u16 driver_watchdog_timeout; + + /* Per packet tx completion timeout. value in ms */ + u16 missing_tx_completion_timeout; + + u16 missed_tx_completion_count_threshold_to_reset; + + /* value in ms */ + u16 admin_completion_tx_timeout; + + u16 netdev_wd_timeout; + + u16 max_tx_sgl_size; + + u16 max_rx_sgl_size; + + u16 reserved[8]; +}; + +struct ena_admin_get_feat_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + struct ena_admin_ctrl_buff_info control_buffer; + + struct ena_admin_get_set_feature_common_desc feat_common; + + u32 raw[11]; +}; + +struct ena_admin_queue_ext_feature_desc { + /* version */ + u8 version; + + u8 reserved1[3]; + + union { + struct ena_admin_queue_ext_feature_fields max_queue_ext; + + u32 raw[10]; + }; +}; + +struct ena_admin_get_feat_resp { + struct ena_admin_acq_common_desc acq_common_desc; + + union { + u32 raw[14]; + + struct ena_admin_device_attr_feature_desc dev_attr; + + struct ena_admin_feature_llq_desc llq; + + struct ena_admin_queue_feature_desc max_queue; + + struct ena_admin_queue_ext_feature_desc max_queue_ext; + + struct ena_admin_feature_aenq_desc aenq; + + struct ena_admin_get_feature_link_desc link; + + struct ena_admin_feature_offload_desc offload; + + struct ena_admin_feature_rss_flow_hash_function flow_hash_func; + + struct ena_admin_feature_rss_flow_hash_input flow_hash_input; + + struct ena_admin_feature_rss_ind_table ind_table; + + struct ena_admin_feature_intr_moder_desc intr_moderation; + + struct ena_admin_ena_hw_hints hw_hints; + } u; +}; + +struct ena_admin_set_feat_cmd { + struct ena_admin_aq_common_desc aq_common_descriptor; + + struct ena_admin_ctrl_buff_info control_buffer; + + struct ena_admin_get_set_feature_common_desc feat_common; + + union { + u32 raw[11]; + + /* mtu size */ + struct ena_admin_set_feature_mtu_desc mtu; + + /* host attributes */ + struct ena_admin_set_feature_host_attr_desc host_attr; + + /* AENQ configuration */ + struct ena_admin_feature_aenq_desc aenq; + + /* rss flow hash function */ + struct ena_admin_feature_rss_flow_hash_function flow_hash_func; + + /* rss flow hash input */ + struct ena_admin_feature_rss_flow_hash_input flow_hash_input; + + /* rss indirection table */ + struct ena_admin_feature_rss_ind_table ind_table; + + /* LLQ configuration */ + struct ena_admin_feature_llq_desc llq; + } u; +}; + +struct ena_admin_set_feat_resp { + struct ena_admin_acq_common_desc acq_common_desc; + + union { + u32 raw[14]; + } u; +}; + +struct ena_admin_aenq_common_desc { + u16 group; + + u16 syndrome; + + /* 0 : phase + * 7:1 : reserved - MBZ + */ + u8 flags; + + u8 reserved1[3]; + + u32 timestamp_low; + + u32 timestamp_high; +}; + +/* asynchronous event notification groups */ +enum ena_admin_aenq_group { + ENA_ADMIN_LINK_CHANGE = 0, + ENA_ADMIN_FATAL_ERROR = 1, + ENA_ADMIN_WARNING = 2, + ENA_ADMIN_NOTIFICATION = 3, + ENA_ADMIN_KEEP_ALIVE = 4, + ENA_ADMIN_AENQ_GROUPS_NUM = 5, +}; + +enum ena_admin_aenq_notification_syndrome { + ENA_ADMIN_UPDATE_HINTS = 2, +}; + +struct ena_admin_aenq_entry { + struct ena_admin_aenq_common_desc aenq_common_desc; + + /* command specific inline data */ + u32 inline_data_w4[12]; +}; + +struct ena_admin_aenq_link_change_desc { + struct ena_admin_aenq_common_desc aenq_common_desc; + + /* 0 : link_status */ + u32 flags; +}; + +struct ena_admin_aenq_keep_alive_desc { + struct ena_admin_aenq_common_desc aenq_common_desc; + + u32 rx_drops_low; + + u32 rx_drops_high; + + u32 tx_drops_low; + + u32 tx_drops_high; +}; + +struct ena_admin_ena_mmio_req_read_less_resp { + u16 req_id; + + u16 reg_off; + + /* value is valid when poll is cleared */ + u32 reg_val; +}; + +/* aq_common_desc */ +#define ENA_ADMIN_AQ_COMMON_DESC_COMMAND_ID_MASK GENMASK(11, 0) +#define ENA_ADMIN_AQ_COMMON_DESC_PHASE_MASK BIT(0) +#define ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_SHIFT 1 +#define ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_MASK BIT(1) +#define ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_SHIFT 2 +#define ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_MASK BIT(2) + +/* sq */ +#define ENA_ADMIN_SQ_SQ_DIRECTION_SHIFT 5 +#define ENA_ADMIN_SQ_SQ_DIRECTION_MASK GENMASK(7, 5) + +/* acq_common_desc */ +#define ENA_ADMIN_ACQ_COMMON_DESC_COMMAND_ID_MASK GENMASK(11, 0) +#define ENA_ADMIN_ACQ_COMMON_DESC_PHASE_MASK BIT(0) + +/* aq_create_sq_cmd */ +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_SQ_DIRECTION_SHIFT 5 +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_SQ_DIRECTION_MASK GENMASK(7, 5) +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_PLACEMENT_POLICY_MASK GENMASK(3, 0) +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_COMPLETION_POLICY_SHIFT 4 +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_COMPLETION_POLICY_MASK GENMASK(6, 4) +#define ENA_ADMIN_AQ_CREATE_SQ_CMD_IS_PHYSICALLY_CONTIGUOUS_MASK BIT(0) + +/* aq_create_cq_cmd */ +#define ENA_ADMIN_AQ_CREATE_CQ_CMD_INTERRUPT_MODE_ENABLED_SHIFT 5 +#define ENA_ADMIN_AQ_CREATE_CQ_CMD_INTERRUPT_MODE_ENABLED_MASK BIT(5) +#define ENA_ADMIN_AQ_CREATE_CQ_CMD_CQ_ENTRY_SIZE_WORDS_MASK GENMASK(4, 0) + +/* get_set_feature_common_desc */ +#define ENA_ADMIN_GET_SET_FEATURE_COMMON_DESC_SELECT_MASK GENMASK(1, 0) + +/* get_feature_link_desc */ +#define ENA_ADMIN_GET_FEATURE_LINK_DESC_AUTONEG_MASK BIT(0) +#define ENA_ADMIN_GET_FEATURE_LINK_DESC_DUPLEX_SHIFT 1 +#define ENA_ADMIN_GET_FEATURE_LINK_DESC_DUPLEX_MASK BIT(1) + +/* feature_offload_desc */ +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L3_CSUM_IPV4_MASK BIT(0) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV4_CSUM_PART_SHIFT 1 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV4_CSUM_PART_MASK BIT(1) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV4_CSUM_FULL_SHIFT 2 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV4_CSUM_FULL_MASK BIT(2) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV6_CSUM_PART_SHIFT 3 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV6_CSUM_PART_MASK BIT(3) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV6_CSUM_FULL_SHIFT 4 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV6_CSUM_FULL_MASK BIT(4) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV4_SHIFT 5 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV4_MASK BIT(5) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV6_SHIFT 6 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV6_MASK BIT(6) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_ECN_SHIFT 7 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_ECN_MASK BIT(7) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L3_CSUM_IPV4_MASK BIT(0) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV4_CSUM_SHIFT 1 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV4_CSUM_MASK BIT(1) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV6_CSUM_SHIFT 2 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV6_CSUM_MASK BIT(2) +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_HASH_SHIFT 3 +#define ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_HASH_MASK BIT(3) + +/* feature_rss_flow_hash_function */ +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_FUNCTION_FUNCS_MASK GENMASK(7, 0) +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_FUNCTION_SELECTED_FUNC_MASK GENMASK(7, 0) + +/* feature_rss_flow_hash_input */ +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L3_SORT_SHIFT 1 +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L3_SORT_MASK BIT(1) +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L4_SORT_SHIFT 2 +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L4_SORT_MASK BIT(2) +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_ENABLE_L3_SORT_SHIFT 1 +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_ENABLE_L3_SORT_MASK BIT(1) +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_ENABLE_L4_SORT_SHIFT 2 +#define ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_ENABLE_L4_SORT_MASK BIT(2) + +/* host_info */ +#define ENA_ADMIN_HOST_INFO_MAJOR_MASK GENMASK(7, 0) +#define ENA_ADMIN_HOST_INFO_MINOR_SHIFT 8 +#define ENA_ADMIN_HOST_INFO_MINOR_MASK GENMASK(15, 8) +#define ENA_ADMIN_HOST_INFO_SUB_MINOR_SHIFT 16 +#define ENA_ADMIN_HOST_INFO_SUB_MINOR_MASK GENMASK(23, 16) +#define ENA_ADMIN_HOST_INFO_MODULE_TYPE_SHIFT 24 +#define ENA_ADMIN_HOST_INFO_MODULE_TYPE_MASK GENMASK(31, 24) +#define ENA_ADMIN_HOST_INFO_FUNCTION_MASK GENMASK(2, 0) +#define ENA_ADMIN_HOST_INFO_DEVICE_SHIFT 3 +#define ENA_ADMIN_HOST_INFO_DEVICE_MASK GENMASK(7, 3) +#define ENA_ADMIN_HOST_INFO_BUS_SHIFT 8 +#define ENA_ADMIN_HOST_INFO_BUS_MASK GENMASK(15, 8) +#define ENA_ADMIN_HOST_INFO_RX_OFFSET_SHIFT 1 +#define ENA_ADMIN_HOST_INFO_RX_OFFSET_MASK BIT(1) +#define ENA_ADMIN_HOST_INFO_INTERRUPT_MODERATION_SHIFT 2 +#define ENA_ADMIN_HOST_INFO_INTERRUPT_MODERATION_MASK BIT(2) +#define ENA_ADMIN_HOST_INFO_RX_BUF_MIRRORING_SHIFT 3 +#define ENA_ADMIN_HOST_INFO_RX_BUF_MIRRORING_MASK BIT(3) +#define ENA_ADMIN_HOST_INFO_RSS_CONFIGURABLE_FUNCTION_KEY_SHIFT 4 +#define ENA_ADMIN_HOST_INFO_RSS_CONFIGURABLE_FUNCTION_KEY_MASK BIT(4) +#define ENA_ADMIN_HOST_INFO_RX_PAGE_REUSE_SHIFT 6 +#define ENA_ADMIN_HOST_INFO_RX_PAGE_REUSE_MASK BIT(6) + +/* aenq_common_desc */ +#define ENA_ADMIN_AENQ_COMMON_DESC_PHASE_MASK BIT(0) + +/* aenq_link_change_desc */ +#define ENA_ADMIN_AENQ_LINK_CHANGE_DESC_LINK_STATUS_MASK BIT(0) + +#endif /* _ENA_ADMIN_H_ */ diff --git a/drivers/net/ethernet/amazon/ena/ena_com.c b/drivers/net/ethernet/amazon/ena/ena_com.c new file mode 100644 index 0000000000..633b321d7f --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_com.c @@ -0,0 +1,3001 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#include "ena_com.h" + +/*****************************************************************************/ +/*****************************************************************************/ + +/* Timeout in micro-sec */ +#define ADMIN_CMD_TIMEOUT_US (3000000) + +#define ENA_ASYNC_QUEUE_DEPTH 16 +#define ENA_ADMIN_QUEUE_DEPTH 32 + + +#define ENA_CTRL_MAJOR 0 +#define ENA_CTRL_MINOR 0 +#define ENA_CTRL_SUB_MINOR 1 + +#define MIN_ENA_CTRL_VER \ + (((ENA_CTRL_MAJOR) << \ + (ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_SHIFT)) | \ + ((ENA_CTRL_MINOR) << \ + (ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_SHIFT)) | \ + (ENA_CTRL_SUB_MINOR)) + +#define ENA_DMA_ADDR_TO_UINT32_LOW(x) ((u32)((u64)(x))) +#define ENA_DMA_ADDR_TO_UINT32_HIGH(x) ((u32)(((u64)(x)) >> 32)) + +#define ENA_MMIO_READ_TIMEOUT 0xFFFFFFFF + +#define ENA_COM_BOUNCE_BUFFER_CNTRL_CNT 4 + +#define ENA_REGS_ADMIN_INTR_MASK 1 + +#define ENA_MAX_BACKOFF_DELAY_EXP 16U + +#define ENA_MIN_ADMIN_POLL_US 100 + +#define ENA_MAX_ADMIN_POLL_US 5000 + +/*****************************************************************************/ +/*****************************************************************************/ +/*****************************************************************************/ + +enum ena_cmd_status { + ENA_CMD_SUBMITTED, + ENA_CMD_COMPLETED, + /* Abort - canceled by the driver */ + ENA_CMD_ABORTED, +}; + +struct ena_comp_ctx { + struct completion wait_event; + struct ena_admin_acq_entry *user_cqe; + u32 comp_size; + enum ena_cmd_status status; + /* status from the device */ + u8 comp_status; + u8 cmd_opcode; + bool occupied; +}; + +struct ena_com_stats_ctx { + struct ena_admin_aq_get_stats_cmd get_cmd; + struct ena_admin_acq_get_stats_resp get_resp; +}; + +static int ena_com_mem_addr_set(struct ena_com_dev *ena_dev, + struct ena_common_mem_addr *ena_addr, + dma_addr_t addr) +{ + if ((addr & GENMASK_ULL(ena_dev->dma_addr_bits - 1, 0)) != addr) { + netdev_err(ena_dev->net_device, + "DMA address has more bits that the device supports\n"); + return -EINVAL; + } + + ena_addr->mem_addr_low = lower_32_bits(addr); + ena_addr->mem_addr_high = (u16)upper_32_bits(addr); + + return 0; +} + +static int ena_com_admin_init_sq(struct ena_com_admin_queue *admin_queue) +{ + struct ena_com_dev *ena_dev = admin_queue->ena_dev; + struct ena_com_admin_sq *sq = &admin_queue->sq; + u16 size = ADMIN_SQ_SIZE(admin_queue->q_depth); + + sq->entries = dma_alloc_coherent(admin_queue->q_dmadev, size, + &sq->dma_addr, GFP_KERNEL); + + if (!sq->entries) { + netdev_err(ena_dev->net_device, "Memory allocation failed\n"); + return -ENOMEM; + } + + sq->head = 0; + sq->tail = 0; + sq->phase = 1; + + sq->db_addr = NULL; + + return 0; +} + +static int ena_com_admin_init_cq(struct ena_com_admin_queue *admin_queue) +{ + struct ena_com_dev *ena_dev = admin_queue->ena_dev; + struct ena_com_admin_cq *cq = &admin_queue->cq; + u16 size = ADMIN_CQ_SIZE(admin_queue->q_depth); + + cq->entries = dma_alloc_coherent(admin_queue->q_dmadev, size, + &cq->dma_addr, GFP_KERNEL); + + if (!cq->entries) { + netdev_err(ena_dev->net_device, "Memory allocation failed\n"); + return -ENOMEM; + } + + cq->head = 0; + cq->phase = 1; + + return 0; +} + +static int ena_com_admin_init_aenq(struct ena_com_dev *ena_dev, + struct ena_aenq_handlers *aenq_handlers) +{ + struct ena_com_aenq *aenq = &ena_dev->aenq; + u32 addr_low, addr_high, aenq_caps; + u16 size; + + ena_dev->aenq.q_depth = ENA_ASYNC_QUEUE_DEPTH; + size = ADMIN_AENQ_SIZE(ENA_ASYNC_QUEUE_DEPTH); + aenq->entries = dma_alloc_coherent(ena_dev->dmadev, size, + &aenq->dma_addr, GFP_KERNEL); + + if (!aenq->entries) { + netdev_err(ena_dev->net_device, "Memory allocation failed\n"); + return -ENOMEM; + } + + aenq->head = aenq->q_depth; + aenq->phase = 1; + + addr_low = ENA_DMA_ADDR_TO_UINT32_LOW(aenq->dma_addr); + addr_high = ENA_DMA_ADDR_TO_UINT32_HIGH(aenq->dma_addr); + + writel(addr_low, ena_dev->reg_bar + ENA_REGS_AENQ_BASE_LO_OFF); + writel(addr_high, ena_dev->reg_bar + ENA_REGS_AENQ_BASE_HI_OFF); + + aenq_caps = 0; + aenq_caps |= ena_dev->aenq.q_depth & ENA_REGS_AENQ_CAPS_AENQ_DEPTH_MASK; + aenq_caps |= (sizeof(struct ena_admin_aenq_entry) + << ENA_REGS_AENQ_CAPS_AENQ_ENTRY_SIZE_SHIFT) & + ENA_REGS_AENQ_CAPS_AENQ_ENTRY_SIZE_MASK; + writel(aenq_caps, ena_dev->reg_bar + ENA_REGS_AENQ_CAPS_OFF); + + if (unlikely(!aenq_handlers)) { + netdev_err(ena_dev->net_device, + "AENQ handlers pointer is NULL\n"); + return -EINVAL; + } + + aenq->aenq_handlers = aenq_handlers; + + return 0; +} + +static void comp_ctxt_release(struct ena_com_admin_queue *queue, + struct ena_comp_ctx *comp_ctx) +{ + comp_ctx->occupied = false; + atomic_dec(&queue->outstanding_cmds); +} + +static struct ena_comp_ctx *get_comp_ctxt(struct ena_com_admin_queue *admin_queue, + u16 command_id, bool capture) +{ + if (unlikely(command_id >= admin_queue->q_depth)) { + netdev_err(admin_queue->ena_dev->net_device, + "Command id is larger than the queue size. cmd_id: %u queue size %d\n", + command_id, admin_queue->q_depth); + return NULL; + } + + if (unlikely(!admin_queue->comp_ctx)) { + netdev_err(admin_queue->ena_dev->net_device, + "Completion context is NULL\n"); + return NULL; + } + + if (unlikely(admin_queue->comp_ctx[command_id].occupied && capture)) { + netdev_err(admin_queue->ena_dev->net_device, + "Completion context is occupied\n"); + return NULL; + } + + if (capture) { + atomic_inc(&admin_queue->outstanding_cmds); + admin_queue->comp_ctx[command_id].occupied = true; + } + + return &admin_queue->comp_ctx[command_id]; +} + +static struct ena_comp_ctx *__ena_com_submit_admin_cmd(struct ena_com_admin_queue *admin_queue, + struct ena_admin_aq_entry *cmd, + size_t cmd_size_in_bytes, + struct ena_admin_acq_entry *comp, + size_t comp_size_in_bytes) +{ + struct ena_comp_ctx *comp_ctx; + u16 tail_masked, cmd_id; + u16 queue_size_mask; + u16 cnt; + + queue_size_mask = admin_queue->q_depth - 1; + + tail_masked = admin_queue->sq.tail & queue_size_mask; + + /* In case of queue FULL */ + cnt = (u16)atomic_read(&admin_queue->outstanding_cmds); + if (cnt >= admin_queue->q_depth) { + netdev_dbg(admin_queue->ena_dev->net_device, + "Admin queue is full.\n"); + admin_queue->stats.out_of_space++; + return ERR_PTR(-ENOSPC); + } + + cmd_id = admin_queue->curr_cmd_id; + + cmd->aq_common_descriptor.flags |= admin_queue->sq.phase & + ENA_ADMIN_AQ_COMMON_DESC_PHASE_MASK; + + cmd->aq_common_descriptor.command_id |= cmd_id & + ENA_ADMIN_AQ_COMMON_DESC_COMMAND_ID_MASK; + + comp_ctx = get_comp_ctxt(admin_queue, cmd_id, true); + if (unlikely(!comp_ctx)) + return ERR_PTR(-EINVAL); + + comp_ctx->status = ENA_CMD_SUBMITTED; + comp_ctx->comp_size = (u32)comp_size_in_bytes; + comp_ctx->user_cqe = comp; + comp_ctx->cmd_opcode = cmd->aq_common_descriptor.opcode; + + reinit_completion(&comp_ctx->wait_event); + + memcpy(&admin_queue->sq.entries[tail_masked], cmd, cmd_size_in_bytes); + + admin_queue->curr_cmd_id = (admin_queue->curr_cmd_id + 1) & + queue_size_mask; + + admin_queue->sq.tail++; + admin_queue->stats.submitted_cmd++; + + if (unlikely((admin_queue->sq.tail & queue_size_mask) == 0)) + admin_queue->sq.phase = !admin_queue->sq.phase; + + writel(admin_queue->sq.tail, admin_queue->sq.db_addr); + + return comp_ctx; +} + +static int ena_com_init_comp_ctxt(struct ena_com_admin_queue *admin_queue) +{ + struct ena_com_dev *ena_dev = admin_queue->ena_dev; + size_t size = admin_queue->q_depth * sizeof(struct ena_comp_ctx); + struct ena_comp_ctx *comp_ctx; + u16 i; + + admin_queue->comp_ctx = + devm_kzalloc(admin_queue->q_dmadev, size, GFP_KERNEL); + if (unlikely(!admin_queue->comp_ctx)) { + netdev_err(ena_dev->net_device, "Memory allocation failed\n"); + return -ENOMEM; + } + + for (i = 0; i < admin_queue->q_depth; i++) { + comp_ctx = get_comp_ctxt(admin_queue, i, false); + if (comp_ctx) + init_completion(&comp_ctx->wait_event); + } + + return 0; +} + +static struct ena_comp_ctx *ena_com_submit_admin_cmd(struct ena_com_admin_queue *admin_queue, + struct ena_admin_aq_entry *cmd, + size_t cmd_size_in_bytes, + struct ena_admin_acq_entry *comp, + size_t comp_size_in_bytes) +{ + unsigned long flags = 0; + struct ena_comp_ctx *comp_ctx; + + spin_lock_irqsave(&admin_queue->q_lock, flags); + if (unlikely(!admin_queue->running_state)) { + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + return ERR_PTR(-ENODEV); + } + comp_ctx = __ena_com_submit_admin_cmd(admin_queue, cmd, + cmd_size_in_bytes, + comp, + comp_size_in_bytes); + if (IS_ERR(comp_ctx)) + admin_queue->running_state = false; + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + + return comp_ctx; +} + +static int ena_com_init_io_sq(struct ena_com_dev *ena_dev, + struct ena_com_create_io_ctx *ctx, + struct ena_com_io_sq *io_sq) +{ + size_t size; + int dev_node = 0; + + memset(&io_sq->desc_addr, 0x0, sizeof(io_sq->desc_addr)); + + io_sq->dma_addr_bits = (u8)ena_dev->dma_addr_bits; + io_sq->desc_entry_size = + (io_sq->direction == ENA_COM_IO_QUEUE_DIRECTION_TX) ? + sizeof(struct ena_eth_io_tx_desc) : + sizeof(struct ena_eth_io_rx_desc); + + size = io_sq->desc_entry_size * io_sq->q_depth; + + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_HOST) { + dev_node = dev_to_node(ena_dev->dmadev); + set_dev_node(ena_dev->dmadev, ctx->numa_node); + io_sq->desc_addr.virt_addr = + dma_alloc_coherent(ena_dev->dmadev, size, + &io_sq->desc_addr.phys_addr, + GFP_KERNEL); + set_dev_node(ena_dev->dmadev, dev_node); + if (!io_sq->desc_addr.virt_addr) { + io_sq->desc_addr.virt_addr = + dma_alloc_coherent(ena_dev->dmadev, size, + &io_sq->desc_addr.phys_addr, + GFP_KERNEL); + } + + if (!io_sq->desc_addr.virt_addr) { + netdev_err(ena_dev->net_device, + "Memory allocation failed\n"); + return -ENOMEM; + } + } + + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + /* Allocate bounce buffers */ + io_sq->bounce_buf_ctrl.buffer_size = + ena_dev->llq_info.desc_list_entry_size; + io_sq->bounce_buf_ctrl.buffers_num = + ENA_COM_BOUNCE_BUFFER_CNTRL_CNT; + io_sq->bounce_buf_ctrl.next_to_use = 0; + + size = io_sq->bounce_buf_ctrl.buffer_size * + io_sq->bounce_buf_ctrl.buffers_num; + + dev_node = dev_to_node(ena_dev->dmadev); + set_dev_node(ena_dev->dmadev, ctx->numa_node); + io_sq->bounce_buf_ctrl.base_buffer = + devm_kzalloc(ena_dev->dmadev, size, GFP_KERNEL); + set_dev_node(ena_dev->dmadev, dev_node); + if (!io_sq->bounce_buf_ctrl.base_buffer) + io_sq->bounce_buf_ctrl.base_buffer = + devm_kzalloc(ena_dev->dmadev, size, GFP_KERNEL); + + if (!io_sq->bounce_buf_ctrl.base_buffer) { + netdev_err(ena_dev->net_device, + "Bounce buffer memory allocation failed\n"); + return -ENOMEM; + } + + memcpy(&io_sq->llq_info, &ena_dev->llq_info, + sizeof(io_sq->llq_info)); + + /* Initiate the first bounce buffer */ + io_sq->llq_buf_ctrl.curr_bounce_buf = + ena_com_get_next_bounce_buffer(&io_sq->bounce_buf_ctrl); + memset(io_sq->llq_buf_ctrl.curr_bounce_buf, + 0x0, io_sq->llq_info.desc_list_entry_size); + io_sq->llq_buf_ctrl.descs_left_in_line = + io_sq->llq_info.descs_num_before_header; + io_sq->disable_meta_caching = + io_sq->llq_info.disable_meta_caching; + + if (io_sq->llq_info.max_entries_in_tx_burst > 0) + io_sq->entries_in_tx_burst_left = + io_sq->llq_info.max_entries_in_tx_burst; + } + + io_sq->tail = 0; + io_sq->next_to_comp = 0; + io_sq->phase = 1; + + return 0; +} + +static int ena_com_init_io_cq(struct ena_com_dev *ena_dev, + struct ena_com_create_io_ctx *ctx, + struct ena_com_io_cq *io_cq) +{ + size_t size; + int prev_node = 0; + + memset(&io_cq->cdesc_addr, 0x0, sizeof(io_cq->cdesc_addr)); + + /* Use the basic completion descriptor for Rx */ + io_cq->cdesc_entry_size_in_bytes = + (io_cq->direction == ENA_COM_IO_QUEUE_DIRECTION_TX) ? + sizeof(struct ena_eth_io_tx_cdesc) : + sizeof(struct ena_eth_io_rx_cdesc_base); + + size = io_cq->cdesc_entry_size_in_bytes * io_cq->q_depth; + + prev_node = dev_to_node(ena_dev->dmadev); + set_dev_node(ena_dev->dmadev, ctx->numa_node); + io_cq->cdesc_addr.virt_addr = + dma_alloc_coherent(ena_dev->dmadev, size, + &io_cq->cdesc_addr.phys_addr, GFP_KERNEL); + set_dev_node(ena_dev->dmadev, prev_node); + if (!io_cq->cdesc_addr.virt_addr) { + io_cq->cdesc_addr.virt_addr = + dma_alloc_coherent(ena_dev->dmadev, size, + &io_cq->cdesc_addr.phys_addr, + GFP_KERNEL); + } + + if (!io_cq->cdesc_addr.virt_addr) { + netdev_err(ena_dev->net_device, "Memory allocation failed\n"); + return -ENOMEM; + } + + io_cq->phase = 1; + io_cq->head = 0; + + return 0; +} + +static void ena_com_handle_single_admin_completion(struct ena_com_admin_queue *admin_queue, + struct ena_admin_acq_entry *cqe) +{ + struct ena_comp_ctx *comp_ctx; + u16 cmd_id; + + cmd_id = cqe->acq_common_descriptor.command & + ENA_ADMIN_ACQ_COMMON_DESC_COMMAND_ID_MASK; + + comp_ctx = get_comp_ctxt(admin_queue, cmd_id, false); + if (unlikely(!comp_ctx)) { + netdev_err(admin_queue->ena_dev->net_device, + "comp_ctx is NULL. Changing the admin queue running state\n"); + admin_queue->running_state = false; + return; + } + + comp_ctx->status = ENA_CMD_COMPLETED; + comp_ctx->comp_status = cqe->acq_common_descriptor.status; + + if (comp_ctx->user_cqe) + memcpy(comp_ctx->user_cqe, (void *)cqe, comp_ctx->comp_size); + + if (!admin_queue->polling) + complete(&comp_ctx->wait_event); +} + +static void ena_com_handle_admin_completion(struct ena_com_admin_queue *admin_queue) +{ + struct ena_admin_acq_entry *cqe = NULL; + u16 comp_num = 0; + u16 head_masked; + u8 phase; + + head_masked = admin_queue->cq.head & (admin_queue->q_depth - 1); + phase = admin_queue->cq.phase; + + cqe = &admin_queue->cq.entries[head_masked]; + + /* Go over all the completions */ + while ((READ_ONCE(cqe->acq_common_descriptor.flags) & + ENA_ADMIN_ACQ_COMMON_DESC_PHASE_MASK) == phase) { + /* Do not read the rest of the completion entry before the + * phase bit was validated + */ + dma_rmb(); + ena_com_handle_single_admin_completion(admin_queue, cqe); + + head_masked++; + comp_num++; + if (unlikely(head_masked == admin_queue->q_depth)) { + head_masked = 0; + phase = !phase; + } + + cqe = &admin_queue->cq.entries[head_masked]; + } + + admin_queue->cq.head += comp_num; + admin_queue->cq.phase = phase; + admin_queue->sq.head += comp_num; + admin_queue->stats.completed_cmd += comp_num; +} + +static int ena_com_comp_status_to_errno(struct ena_com_admin_queue *admin_queue, + u8 comp_status) +{ + if (unlikely(comp_status != 0)) + netdev_err(admin_queue->ena_dev->net_device, + "Admin command failed[%u]\n", comp_status); + + switch (comp_status) { + case ENA_ADMIN_SUCCESS: + return 0; + case ENA_ADMIN_RESOURCE_ALLOCATION_FAILURE: + return -ENOMEM; + case ENA_ADMIN_UNSUPPORTED_OPCODE: + return -EOPNOTSUPP; + case ENA_ADMIN_BAD_OPCODE: + case ENA_ADMIN_MALFORMED_REQUEST: + case ENA_ADMIN_ILLEGAL_PARAMETER: + case ENA_ADMIN_UNKNOWN_ERROR: + return -EINVAL; + case ENA_ADMIN_RESOURCE_BUSY: + return -EAGAIN; + } + + return -EINVAL; +} + +static void ena_delay_exponential_backoff_us(u32 exp, u32 delay_us) +{ + exp = min_t(u32, exp, ENA_MAX_BACKOFF_DELAY_EXP); + delay_us = max_t(u32, ENA_MIN_ADMIN_POLL_US, delay_us); + delay_us = min_t(u32, delay_us * (1U << exp), ENA_MAX_ADMIN_POLL_US); + usleep_range(delay_us, 2 * delay_us); +} + +static int ena_com_wait_and_process_admin_cq_polling(struct ena_comp_ctx *comp_ctx, + struct ena_com_admin_queue *admin_queue) +{ + unsigned long flags = 0; + unsigned long timeout; + int ret; + u32 exp = 0; + + timeout = jiffies + usecs_to_jiffies(admin_queue->completion_timeout); + + while (1) { + spin_lock_irqsave(&admin_queue->q_lock, flags); + ena_com_handle_admin_completion(admin_queue); + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + + if (comp_ctx->status != ENA_CMD_SUBMITTED) + break; + + if (time_is_before_jiffies(timeout)) { + netdev_err(admin_queue->ena_dev->net_device, + "Wait for completion (polling) timeout\n"); + /* ENA didn't have any completion */ + spin_lock_irqsave(&admin_queue->q_lock, flags); + admin_queue->stats.no_completion++; + admin_queue->running_state = false; + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + + ret = -ETIME; + goto err; + } + + ena_delay_exponential_backoff_us(exp++, + admin_queue->ena_dev->ena_min_poll_delay_us); + } + + if (unlikely(comp_ctx->status == ENA_CMD_ABORTED)) { + netdev_err(admin_queue->ena_dev->net_device, + "Command was aborted\n"); + spin_lock_irqsave(&admin_queue->q_lock, flags); + admin_queue->stats.aborted_cmd++; + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + ret = -ENODEV; + goto err; + } + + WARN(comp_ctx->status != ENA_CMD_COMPLETED, "Invalid comp status %d\n", + comp_ctx->status); + + ret = ena_com_comp_status_to_errno(admin_queue, comp_ctx->comp_status); +err: + comp_ctxt_release(admin_queue, comp_ctx); + return ret; +} + +/* + * Set the LLQ configurations of the firmware + * + * The driver provides only the enabled feature values to the device, + * which in turn, checks if they are supported. + */ +static int ena_com_set_llq(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + struct ena_com_llq_info *llq_info = &ena_dev->llq_info; + int ret; + + memset(&cmd, 0x0, sizeof(cmd)); + admin_queue = &ena_dev->admin_queue; + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.feat_common.feature_id = ENA_ADMIN_LLQ; + + cmd.u.llq.header_location_ctrl_enabled = llq_info->header_location_ctrl; + cmd.u.llq.entry_size_ctrl_enabled = llq_info->desc_list_entry_size_ctrl; + cmd.u.llq.desc_num_before_header_enabled = llq_info->descs_num_before_header; + cmd.u.llq.descriptors_stride_ctrl_enabled = llq_info->desc_stride_ctrl; + + cmd.u.llq.accel_mode.u.set.enabled_flags = + BIT(ENA_ADMIN_DISABLE_META_CACHING) | + BIT(ENA_ADMIN_LIMIT_TX_BURST); + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to set LLQ configurations: %d\n", ret); + + return ret; +} + +static int ena_com_config_llq_info(struct ena_com_dev *ena_dev, + struct ena_admin_feature_llq_desc *llq_features, + struct ena_llq_configurations *llq_default_cfg) +{ + struct ena_com_llq_info *llq_info = &ena_dev->llq_info; + struct ena_admin_accel_mode_get llq_accel_mode_get; + u16 supported_feat; + int rc; + + memset(llq_info, 0, sizeof(*llq_info)); + + supported_feat = llq_features->header_location_ctrl_supported; + + if (likely(supported_feat & llq_default_cfg->llq_header_location)) { + llq_info->header_location_ctrl = + llq_default_cfg->llq_header_location; + } else { + netdev_err(ena_dev->net_device, + "Invalid header location control, supported: 0x%x\n", + supported_feat); + return -EINVAL; + } + + if (likely(llq_info->header_location_ctrl == ENA_ADMIN_INLINE_HEADER)) { + supported_feat = llq_features->descriptors_stride_ctrl_supported; + if (likely(supported_feat & llq_default_cfg->llq_stride_ctrl)) { + llq_info->desc_stride_ctrl = llq_default_cfg->llq_stride_ctrl; + } else { + if (supported_feat & ENA_ADMIN_MULTIPLE_DESCS_PER_ENTRY) { + llq_info->desc_stride_ctrl = ENA_ADMIN_MULTIPLE_DESCS_PER_ENTRY; + } else if (supported_feat & ENA_ADMIN_SINGLE_DESC_PER_ENTRY) { + llq_info->desc_stride_ctrl = ENA_ADMIN_SINGLE_DESC_PER_ENTRY; + } else { + netdev_err(ena_dev->net_device, + "Invalid desc_stride_ctrl, supported: 0x%x\n", + supported_feat); + return -EINVAL; + } + + netdev_err(ena_dev->net_device, + "Default llq stride ctrl is not supported, performing fallback, default: 0x%x, supported: 0x%x, used: 0x%x\n", + llq_default_cfg->llq_stride_ctrl, + supported_feat, llq_info->desc_stride_ctrl); + } + } else { + llq_info->desc_stride_ctrl = 0; + } + + supported_feat = llq_features->entry_size_ctrl_supported; + if (likely(supported_feat & llq_default_cfg->llq_ring_entry_size)) { + llq_info->desc_list_entry_size_ctrl = llq_default_cfg->llq_ring_entry_size; + llq_info->desc_list_entry_size = llq_default_cfg->llq_ring_entry_size_value; + } else { + if (supported_feat & ENA_ADMIN_LIST_ENTRY_SIZE_128B) { + llq_info->desc_list_entry_size_ctrl = ENA_ADMIN_LIST_ENTRY_SIZE_128B; + llq_info->desc_list_entry_size = 128; + } else if (supported_feat & ENA_ADMIN_LIST_ENTRY_SIZE_192B) { + llq_info->desc_list_entry_size_ctrl = ENA_ADMIN_LIST_ENTRY_SIZE_192B; + llq_info->desc_list_entry_size = 192; + } else if (supported_feat & ENA_ADMIN_LIST_ENTRY_SIZE_256B) { + llq_info->desc_list_entry_size_ctrl = ENA_ADMIN_LIST_ENTRY_SIZE_256B; + llq_info->desc_list_entry_size = 256; + } else { + netdev_err(ena_dev->net_device, + "Invalid entry_size_ctrl, supported: 0x%x\n", + supported_feat); + return -EINVAL; + } + + netdev_err(ena_dev->net_device, + "Default llq ring entry size is not supported, performing fallback, default: 0x%x, supported: 0x%x, used: 0x%x\n", + llq_default_cfg->llq_ring_entry_size, supported_feat, + llq_info->desc_list_entry_size); + } + if (unlikely(llq_info->desc_list_entry_size & 0x7)) { + /* The desc list entry size should be whole multiply of 8 + * This requirement comes from __iowrite64_copy() + */ + netdev_err(ena_dev->net_device, "Illegal entry size %d\n", + llq_info->desc_list_entry_size); + return -EINVAL; + } + + if (llq_info->desc_stride_ctrl == ENA_ADMIN_MULTIPLE_DESCS_PER_ENTRY) + llq_info->descs_per_entry = llq_info->desc_list_entry_size / + sizeof(struct ena_eth_io_tx_desc); + else + llq_info->descs_per_entry = 1; + + supported_feat = llq_features->desc_num_before_header_supported; + if (likely(supported_feat & llq_default_cfg->llq_num_decs_before_header)) { + llq_info->descs_num_before_header = llq_default_cfg->llq_num_decs_before_header; + } else { + if (supported_feat & ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_2) { + llq_info->descs_num_before_header = ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_2; + } else if (supported_feat & ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_1) { + llq_info->descs_num_before_header = ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_1; + } else if (supported_feat & ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_4) { + llq_info->descs_num_before_header = ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_4; + } else if (supported_feat & ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_8) { + llq_info->descs_num_before_header = ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_8; + } else { + netdev_err(ena_dev->net_device, + "Invalid descs_num_before_header, supported: 0x%x\n", + supported_feat); + return -EINVAL; + } + + netdev_err(ena_dev->net_device, + "Default llq num descs before header is not supported, performing fallback, default: 0x%x, supported: 0x%x, used: 0x%x\n", + llq_default_cfg->llq_num_decs_before_header, + supported_feat, llq_info->descs_num_before_header); + } + /* Check for accelerated queue supported */ + llq_accel_mode_get = llq_features->accel_mode.u.get; + + llq_info->disable_meta_caching = + !!(llq_accel_mode_get.supported_flags & + BIT(ENA_ADMIN_DISABLE_META_CACHING)); + + if (llq_accel_mode_get.supported_flags & BIT(ENA_ADMIN_LIMIT_TX_BURST)) + llq_info->max_entries_in_tx_burst = + llq_accel_mode_get.max_tx_burst_size / + llq_default_cfg->llq_ring_entry_size_value; + + rc = ena_com_set_llq(ena_dev); + if (rc) + netdev_err(ena_dev->net_device, + "Cannot set LLQ configuration: %d\n", rc); + + return rc; +} + +static int ena_com_wait_and_process_admin_cq_interrupts(struct ena_comp_ctx *comp_ctx, + struct ena_com_admin_queue *admin_queue) +{ + unsigned long flags = 0; + int ret; + + wait_for_completion_timeout(&comp_ctx->wait_event, + usecs_to_jiffies( + admin_queue->completion_timeout)); + + /* In case the command wasn't completed find out the root cause. + * There might be 2 kinds of errors + * 1) No completion (timeout reached) + * 2) There is completion but the device didn't get any msi-x interrupt. + */ + if (unlikely(comp_ctx->status == ENA_CMD_SUBMITTED)) { + spin_lock_irqsave(&admin_queue->q_lock, flags); + ena_com_handle_admin_completion(admin_queue); + admin_queue->stats.no_completion++; + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + + if (comp_ctx->status == ENA_CMD_COMPLETED) { + netdev_err(admin_queue->ena_dev->net_device, + "The ena device sent a completion but the driver didn't receive a MSI-X interrupt (cmd %d), autopolling mode is %s\n", + comp_ctx->cmd_opcode, + admin_queue->auto_polling ? "ON" : "OFF"); + /* Check if fallback to polling is enabled */ + if (admin_queue->auto_polling) + admin_queue->polling = true; + } else { + netdev_err(admin_queue->ena_dev->net_device, + "The ena device didn't send a completion for the admin cmd %d status %d\n", + comp_ctx->cmd_opcode, comp_ctx->status); + } + /* Check if shifted to polling mode. + * This will happen if there is a completion without an interrupt + * and autopolling mode is enabled. Continuing normal execution in such case + */ + if (!admin_queue->polling) { + admin_queue->running_state = false; + ret = -ETIME; + goto err; + } + } + + ret = ena_com_comp_status_to_errno(admin_queue, comp_ctx->comp_status); +err: + comp_ctxt_release(admin_queue, comp_ctx); + return ret; +} + +/* This method read the hardware device register through posting writes + * and waiting for response + * On timeout the function will return ENA_MMIO_READ_TIMEOUT + */ +static u32 ena_com_reg_bar_read32(struct ena_com_dev *ena_dev, u16 offset) +{ + struct ena_com_mmio_read *mmio_read = &ena_dev->mmio_read; + volatile struct ena_admin_ena_mmio_req_read_less_resp *read_resp = + mmio_read->read_resp; + u32 mmio_read_reg, ret, i; + unsigned long flags = 0; + u32 timeout = mmio_read->reg_read_to; + + might_sleep(); + + if (timeout == 0) + timeout = ENA_REG_READ_TIMEOUT; + + /* If readless is disabled, perform regular read */ + if (!mmio_read->readless_supported) + return readl(ena_dev->reg_bar + offset); + + spin_lock_irqsave(&mmio_read->lock, flags); + mmio_read->seq_num++; + + read_resp->req_id = mmio_read->seq_num + 0xDEAD; + mmio_read_reg = (offset << ENA_REGS_MMIO_REG_READ_REG_OFF_SHIFT) & + ENA_REGS_MMIO_REG_READ_REG_OFF_MASK; + mmio_read_reg |= mmio_read->seq_num & + ENA_REGS_MMIO_REG_READ_REQ_ID_MASK; + + writel(mmio_read_reg, ena_dev->reg_bar + ENA_REGS_MMIO_REG_READ_OFF); + + for (i = 0; i < timeout; i++) { + if (READ_ONCE(read_resp->req_id) == mmio_read->seq_num) + break; + + udelay(1); + } + + if (unlikely(i == timeout)) { + netdev_err(ena_dev->net_device, + "Reading reg failed for timeout. expected: req id[%u] offset[%u] actual: req id[%u] offset[%u]\n", + mmio_read->seq_num, offset, read_resp->req_id, + read_resp->reg_off); + ret = ENA_MMIO_READ_TIMEOUT; + goto err; + } + + if (read_resp->reg_off != offset) { + netdev_err(ena_dev->net_device, + "Read failure: wrong offset provided\n"); + ret = ENA_MMIO_READ_TIMEOUT; + } else { + ret = read_resp->reg_val; + } +err: + spin_unlock_irqrestore(&mmio_read->lock, flags); + + return ret; +} + +/* There are two types to wait for completion. + * Polling mode - wait until the completion is available. + * Async mode - wait on wait queue until the completion is ready + * (or the timeout expired). + * It is expected that the IRQ called ena_com_handle_admin_completion + * to mark the completions. + */ +static int ena_com_wait_and_process_admin_cq(struct ena_comp_ctx *comp_ctx, + struct ena_com_admin_queue *admin_queue) +{ + if (admin_queue->polling) + return ena_com_wait_and_process_admin_cq_polling(comp_ctx, + admin_queue); + + return ena_com_wait_and_process_admin_cq_interrupts(comp_ctx, + admin_queue); +} + +static int ena_com_destroy_io_sq(struct ena_com_dev *ena_dev, + struct ena_com_io_sq *io_sq) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_admin_aq_destroy_sq_cmd destroy_cmd; + struct ena_admin_acq_destroy_sq_resp_desc destroy_resp; + u8 direction; + int ret; + + memset(&destroy_cmd, 0x0, sizeof(destroy_cmd)); + + if (io_sq->direction == ENA_COM_IO_QUEUE_DIRECTION_TX) + direction = ENA_ADMIN_SQ_DIRECTION_TX; + else + direction = ENA_ADMIN_SQ_DIRECTION_RX; + + destroy_cmd.sq.sq_identity |= (direction << + ENA_ADMIN_SQ_SQ_DIRECTION_SHIFT) & + ENA_ADMIN_SQ_SQ_DIRECTION_MASK; + + destroy_cmd.sq.sq_idx = io_sq->idx; + destroy_cmd.aq_common_descriptor.opcode = ENA_ADMIN_DESTROY_SQ; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&destroy_cmd, + sizeof(destroy_cmd), + (struct ena_admin_acq_entry *)&destroy_resp, + sizeof(destroy_resp)); + + if (unlikely(ret && (ret != -ENODEV))) + netdev_err(ena_dev->net_device, + "Failed to destroy io sq error: %d\n", ret); + + return ret; +} + +static void ena_com_io_queue_free(struct ena_com_dev *ena_dev, + struct ena_com_io_sq *io_sq, + struct ena_com_io_cq *io_cq) +{ + size_t size; + + if (io_cq->cdesc_addr.virt_addr) { + size = io_cq->cdesc_entry_size_in_bytes * io_cq->q_depth; + + dma_free_coherent(ena_dev->dmadev, size, + io_cq->cdesc_addr.virt_addr, + io_cq->cdesc_addr.phys_addr); + + io_cq->cdesc_addr.virt_addr = NULL; + } + + if (io_sq->desc_addr.virt_addr) { + size = io_sq->desc_entry_size * io_sq->q_depth; + + dma_free_coherent(ena_dev->dmadev, size, + io_sq->desc_addr.virt_addr, + io_sq->desc_addr.phys_addr); + + io_sq->desc_addr.virt_addr = NULL; + } + + if (io_sq->bounce_buf_ctrl.base_buffer) { + devm_kfree(ena_dev->dmadev, io_sq->bounce_buf_ctrl.base_buffer); + io_sq->bounce_buf_ctrl.base_buffer = NULL; + } +} + +static int wait_for_reset_state(struct ena_com_dev *ena_dev, u32 timeout, + u16 exp_state) +{ + u32 val, exp = 0; + unsigned long timeout_stamp; + + /* Convert timeout from resolution of 100ms to us resolution. */ + timeout_stamp = jiffies + usecs_to_jiffies(100 * 1000 * timeout); + + while (1) { + val = ena_com_reg_bar_read32(ena_dev, ENA_REGS_DEV_STS_OFF); + + if (unlikely(val == ENA_MMIO_READ_TIMEOUT)) { + netdev_err(ena_dev->net_device, + "Reg read timeout occurred\n"); + return -ETIME; + } + + if ((val & ENA_REGS_DEV_STS_RESET_IN_PROGRESS_MASK) == + exp_state) + return 0; + + if (time_is_before_jiffies(timeout_stamp)) + return -ETIME; + + ena_delay_exponential_backoff_us(exp++, ena_dev->ena_min_poll_delay_us); + } +} + +static bool ena_com_check_supported_feature_id(struct ena_com_dev *ena_dev, + enum ena_admin_aq_feature_id feature_id) +{ + u32 feature_mask = 1 << feature_id; + + /* Device attributes is always supported */ + if ((feature_id != ENA_ADMIN_DEVICE_ATTRIBUTES) && + !(ena_dev->supported_features & feature_mask)) + return false; + + return true; +} + +static int ena_com_get_feature_ex(struct ena_com_dev *ena_dev, + struct ena_admin_get_feat_resp *get_resp, + enum ena_admin_aq_feature_id feature_id, + dma_addr_t control_buf_dma_addr, + u32 control_buff_size, + u8 feature_ver) +{ + struct ena_com_admin_queue *admin_queue; + struct ena_admin_get_feat_cmd get_cmd; + int ret; + + if (!ena_com_check_supported_feature_id(ena_dev, feature_id)) { + netdev_dbg(ena_dev->net_device, "Feature %d isn't supported\n", + feature_id); + return -EOPNOTSUPP; + } + + memset(&get_cmd, 0x0, sizeof(get_cmd)); + admin_queue = &ena_dev->admin_queue; + + get_cmd.aq_common_descriptor.opcode = ENA_ADMIN_GET_FEATURE; + + if (control_buff_size) + get_cmd.aq_common_descriptor.flags = + ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_MASK; + else + get_cmd.aq_common_descriptor.flags = 0; + + ret = ena_com_mem_addr_set(ena_dev, + &get_cmd.control_buffer.address, + control_buf_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + get_cmd.control_buffer.length = control_buff_size; + get_cmd.feat_common.feature_version = feature_ver; + get_cmd.feat_common.feature_id = feature_id; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *) + &get_cmd, + sizeof(get_cmd), + (struct ena_admin_acq_entry *) + get_resp, + sizeof(*get_resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to submit get_feature command %d error: %d\n", + feature_id, ret); + + return ret; +} + +static int ena_com_get_feature(struct ena_com_dev *ena_dev, + struct ena_admin_get_feat_resp *get_resp, + enum ena_admin_aq_feature_id feature_id, + u8 feature_ver) +{ + return ena_com_get_feature_ex(ena_dev, + get_resp, + feature_id, + 0, + 0, + feature_ver); +} + +int ena_com_get_current_hash_function(struct ena_com_dev *ena_dev) +{ + return ena_dev->rss.hash_func; +} + +static void ena_com_hash_key_fill_default_key(struct ena_com_dev *ena_dev) +{ + struct ena_admin_feature_rss_flow_hash_control *hash_key = + (ena_dev->rss).hash_key; + + netdev_rss_key_fill(&hash_key->key, sizeof(hash_key->key)); + /* The key buffer is stored in the device in an array of + * uint32 elements. + */ + hash_key->key_parts = ENA_ADMIN_RSS_KEY_PARTS; +} + +static int ena_com_hash_key_allocate(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + + if (!ena_com_check_supported_feature_id(ena_dev, + ENA_ADMIN_RSS_HASH_FUNCTION)) + return -EOPNOTSUPP; + + rss->hash_key = + dma_alloc_coherent(ena_dev->dmadev, sizeof(*rss->hash_key), + &rss->hash_key_dma_addr, GFP_KERNEL); + + if (unlikely(!rss->hash_key)) + return -ENOMEM; + + return 0; +} + +static void ena_com_hash_key_destroy(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + + if (rss->hash_key) + dma_free_coherent(ena_dev->dmadev, sizeof(*rss->hash_key), + rss->hash_key, rss->hash_key_dma_addr); + rss->hash_key = NULL; +} + +static int ena_com_hash_ctrl_init(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + + rss->hash_ctrl = + dma_alloc_coherent(ena_dev->dmadev, sizeof(*rss->hash_ctrl), + &rss->hash_ctrl_dma_addr, GFP_KERNEL); + + if (unlikely(!rss->hash_ctrl)) + return -ENOMEM; + + return 0; +} + +static void ena_com_hash_ctrl_destroy(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + + if (rss->hash_ctrl) + dma_free_coherent(ena_dev->dmadev, sizeof(*rss->hash_ctrl), + rss->hash_ctrl, rss->hash_ctrl_dma_addr); + rss->hash_ctrl = NULL; +} + +static int ena_com_indirect_table_allocate(struct ena_com_dev *ena_dev, + u16 log_size) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_get_feat_resp get_resp; + size_t tbl_size; + int ret; + + ret = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG, 0); + if (unlikely(ret)) + return ret; + + if ((get_resp.u.ind_table.min_size > log_size) || + (get_resp.u.ind_table.max_size < log_size)) { + netdev_err(ena_dev->net_device, + "Indirect table size doesn't fit. requested size: %d while min is:%d and max %d\n", + 1 << log_size, 1 << get_resp.u.ind_table.min_size, + 1 << get_resp.u.ind_table.max_size); + return -EINVAL; + } + + tbl_size = (1ULL << log_size) * + sizeof(struct ena_admin_rss_ind_table_entry); + + rss->rss_ind_tbl = + dma_alloc_coherent(ena_dev->dmadev, tbl_size, + &rss->rss_ind_tbl_dma_addr, GFP_KERNEL); + if (unlikely(!rss->rss_ind_tbl)) + goto mem_err1; + + tbl_size = (1ULL << log_size) * sizeof(u16); + rss->host_rss_ind_tbl = + devm_kzalloc(ena_dev->dmadev, tbl_size, GFP_KERNEL); + if (unlikely(!rss->host_rss_ind_tbl)) + goto mem_err2; + + rss->tbl_log_size = log_size; + + return 0; + +mem_err2: + tbl_size = (1ULL << log_size) * + sizeof(struct ena_admin_rss_ind_table_entry); + + dma_free_coherent(ena_dev->dmadev, tbl_size, rss->rss_ind_tbl, + rss->rss_ind_tbl_dma_addr); + rss->rss_ind_tbl = NULL; +mem_err1: + rss->tbl_log_size = 0; + return -ENOMEM; +} + +static void ena_com_indirect_table_destroy(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + size_t tbl_size = (1ULL << rss->tbl_log_size) * + sizeof(struct ena_admin_rss_ind_table_entry); + + if (rss->rss_ind_tbl) + dma_free_coherent(ena_dev->dmadev, tbl_size, rss->rss_ind_tbl, + rss->rss_ind_tbl_dma_addr); + rss->rss_ind_tbl = NULL; + + if (rss->host_rss_ind_tbl) + devm_kfree(ena_dev->dmadev, rss->host_rss_ind_tbl); + rss->host_rss_ind_tbl = NULL; +} + +static int ena_com_create_io_sq(struct ena_com_dev *ena_dev, + struct ena_com_io_sq *io_sq, u16 cq_idx) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_admin_aq_create_sq_cmd create_cmd; + struct ena_admin_acq_create_sq_resp_desc cmd_completion; + u8 direction; + int ret; + + memset(&create_cmd, 0x0, sizeof(create_cmd)); + + create_cmd.aq_common_descriptor.opcode = ENA_ADMIN_CREATE_SQ; + + if (io_sq->direction == ENA_COM_IO_QUEUE_DIRECTION_TX) + direction = ENA_ADMIN_SQ_DIRECTION_TX; + else + direction = ENA_ADMIN_SQ_DIRECTION_RX; + + create_cmd.sq_identity |= (direction << + ENA_ADMIN_AQ_CREATE_SQ_CMD_SQ_DIRECTION_SHIFT) & + ENA_ADMIN_AQ_CREATE_SQ_CMD_SQ_DIRECTION_MASK; + + create_cmd.sq_caps_2 |= io_sq->mem_queue_type & + ENA_ADMIN_AQ_CREATE_SQ_CMD_PLACEMENT_POLICY_MASK; + + create_cmd.sq_caps_2 |= (ENA_ADMIN_COMPLETION_POLICY_DESC << + ENA_ADMIN_AQ_CREATE_SQ_CMD_COMPLETION_POLICY_SHIFT) & + ENA_ADMIN_AQ_CREATE_SQ_CMD_COMPLETION_POLICY_MASK; + + create_cmd.sq_caps_3 |= + ENA_ADMIN_AQ_CREATE_SQ_CMD_IS_PHYSICALLY_CONTIGUOUS_MASK; + + create_cmd.cq_idx = cq_idx; + create_cmd.sq_depth = io_sq->q_depth; + + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_HOST) { + ret = ena_com_mem_addr_set(ena_dev, + &create_cmd.sq_ba, + io_sq->desc_addr.phys_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, + "Memory address set failed\n"); + return ret; + } + } + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&create_cmd, + sizeof(create_cmd), + (struct ena_admin_acq_entry *)&cmd_completion, + sizeof(cmd_completion)); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, + "Failed to create IO SQ. error: %d\n", ret); + return ret; + } + + io_sq->idx = cmd_completion.sq_idx; + + io_sq->db_addr = (u32 __iomem *)((uintptr_t)ena_dev->reg_bar + + (uintptr_t)cmd_completion.sq_doorbell_offset); + + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + io_sq->header_addr = (u8 __iomem *)((uintptr_t)ena_dev->mem_bar + + cmd_completion.llq_headers_offset); + + io_sq->desc_addr.pbuf_dev_addr = + (u8 __iomem *)((uintptr_t)ena_dev->mem_bar + + cmd_completion.llq_descriptors_offset); + } + + netdev_dbg(ena_dev->net_device, "Created sq[%u], depth[%u]\n", + io_sq->idx, io_sq->q_depth); + + return ret; +} + +static int ena_com_ind_tbl_convert_to_device(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_com_io_sq *io_sq; + u16 qid; + int i; + + for (i = 0; i < 1 << rss->tbl_log_size; i++) { + qid = rss->host_rss_ind_tbl[i]; + if (qid >= ENA_TOTAL_NUM_QUEUES) + return -EINVAL; + + io_sq = &ena_dev->io_sq_queues[qid]; + + if (io_sq->direction != ENA_COM_IO_QUEUE_DIRECTION_RX) + return -EINVAL; + + rss->rss_ind_tbl[i].cq_idx = io_sq->idx; + } + + return 0; +} + +static void ena_com_update_intr_delay_resolution(struct ena_com_dev *ena_dev, + u16 intr_delay_resolution) +{ + u16 prev_intr_delay_resolution = ena_dev->intr_delay_resolution; + + if (unlikely(!intr_delay_resolution)) { + netdev_err(ena_dev->net_device, + "Illegal intr_delay_resolution provided. Going to use default 1 usec resolution\n"); + intr_delay_resolution = ENA_DEFAULT_INTR_DELAY_RESOLUTION; + } + + /* update Rx */ + ena_dev->intr_moder_rx_interval = + ena_dev->intr_moder_rx_interval * + prev_intr_delay_resolution / + intr_delay_resolution; + + /* update Tx */ + ena_dev->intr_moder_tx_interval = + ena_dev->intr_moder_tx_interval * + prev_intr_delay_resolution / + intr_delay_resolution; + + ena_dev->intr_delay_resolution = intr_delay_resolution; +} + +/*****************************************************************************/ +/******************************* API ******************************/ +/*****************************************************************************/ + +int ena_com_execute_admin_command(struct ena_com_admin_queue *admin_queue, + struct ena_admin_aq_entry *cmd, + size_t cmd_size, + struct ena_admin_acq_entry *comp, + size_t comp_size) +{ + struct ena_comp_ctx *comp_ctx; + int ret; + + comp_ctx = ena_com_submit_admin_cmd(admin_queue, cmd, cmd_size, + comp, comp_size); + if (IS_ERR(comp_ctx)) { + ret = PTR_ERR(comp_ctx); + if (ret == -ENODEV) + netdev_dbg(admin_queue->ena_dev->net_device, + "Failed to submit command [%d]\n", ret); + else + netdev_err(admin_queue->ena_dev->net_device, + "Failed to submit command [%d]\n", ret); + + return ret; + } + + ret = ena_com_wait_and_process_admin_cq(comp_ctx, admin_queue); + if (unlikely(ret)) { + if (admin_queue->running_state) + netdev_err(admin_queue->ena_dev->net_device, + "Failed to process command. ret = %d\n", ret); + else + netdev_dbg(admin_queue->ena_dev->net_device, + "Failed to process command. ret = %d\n", ret); + } + return ret; +} + +int ena_com_create_io_cq(struct ena_com_dev *ena_dev, + struct ena_com_io_cq *io_cq) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_admin_aq_create_cq_cmd create_cmd; + struct ena_admin_acq_create_cq_resp_desc cmd_completion; + int ret; + + memset(&create_cmd, 0x0, sizeof(create_cmd)); + + create_cmd.aq_common_descriptor.opcode = ENA_ADMIN_CREATE_CQ; + + create_cmd.cq_caps_2 |= (io_cq->cdesc_entry_size_in_bytes / 4) & + ENA_ADMIN_AQ_CREATE_CQ_CMD_CQ_ENTRY_SIZE_WORDS_MASK; + create_cmd.cq_caps_1 |= + ENA_ADMIN_AQ_CREATE_CQ_CMD_INTERRUPT_MODE_ENABLED_MASK; + + create_cmd.msix_vector = io_cq->msix_vector; + create_cmd.cq_depth = io_cq->q_depth; + + ret = ena_com_mem_addr_set(ena_dev, + &create_cmd.cq_ba, + io_cq->cdesc_addr.phys_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&create_cmd, + sizeof(create_cmd), + (struct ena_admin_acq_entry *)&cmd_completion, + sizeof(cmd_completion)); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, + "Failed to create IO CQ. error: %d\n", ret); + return ret; + } + + io_cq->idx = cmd_completion.cq_idx; + + io_cq->unmask_reg = (u32 __iomem *)((uintptr_t)ena_dev->reg_bar + + cmd_completion.cq_interrupt_unmask_register_offset); + + if (cmd_completion.cq_head_db_register_offset) + io_cq->cq_head_db_reg = + (u32 __iomem *)((uintptr_t)ena_dev->reg_bar + + cmd_completion.cq_head_db_register_offset); + + if (cmd_completion.numa_node_register_offset) + io_cq->numa_node_cfg_reg = + (u32 __iomem *)((uintptr_t)ena_dev->reg_bar + + cmd_completion.numa_node_register_offset); + + netdev_dbg(ena_dev->net_device, "Created cq[%u], depth[%u]\n", + io_cq->idx, io_cq->q_depth); + + return ret; +} + +int ena_com_get_io_handlers(struct ena_com_dev *ena_dev, u16 qid, + struct ena_com_io_sq **io_sq, + struct ena_com_io_cq **io_cq) +{ + if (qid >= ENA_TOTAL_NUM_QUEUES) { + netdev_err(ena_dev->net_device, + "Invalid queue number %d but the max is %d\n", qid, + ENA_TOTAL_NUM_QUEUES); + return -EINVAL; + } + + *io_sq = &ena_dev->io_sq_queues[qid]; + *io_cq = &ena_dev->io_cq_queues[qid]; + + return 0; +} + +void ena_com_abort_admin_commands(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_comp_ctx *comp_ctx; + u16 i; + + if (!admin_queue->comp_ctx) + return; + + for (i = 0; i < admin_queue->q_depth; i++) { + comp_ctx = get_comp_ctxt(admin_queue, i, false); + if (unlikely(!comp_ctx)) + break; + + comp_ctx->status = ENA_CMD_ABORTED; + + complete(&comp_ctx->wait_event); + } +} + +void ena_com_wait_for_abort_completion(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + unsigned long flags = 0; + u32 exp = 0; + + spin_lock_irqsave(&admin_queue->q_lock, flags); + while (atomic_read(&admin_queue->outstanding_cmds) != 0) { + spin_unlock_irqrestore(&admin_queue->q_lock, flags); + ena_delay_exponential_backoff_us(exp++, + ena_dev->ena_min_poll_delay_us); + spin_lock_irqsave(&admin_queue->q_lock, flags); + } + spin_unlock_irqrestore(&admin_queue->q_lock, flags); +} + +int ena_com_destroy_io_cq(struct ena_com_dev *ena_dev, + struct ena_com_io_cq *io_cq) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_admin_aq_destroy_cq_cmd destroy_cmd; + struct ena_admin_acq_destroy_cq_resp_desc destroy_resp; + int ret; + + memset(&destroy_cmd, 0x0, sizeof(destroy_cmd)); + + destroy_cmd.cq_idx = io_cq->idx; + destroy_cmd.aq_common_descriptor.opcode = ENA_ADMIN_DESTROY_CQ; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&destroy_cmd, + sizeof(destroy_cmd), + (struct ena_admin_acq_entry *)&destroy_resp, + sizeof(destroy_resp)); + + if (unlikely(ret && (ret != -ENODEV))) + netdev_err(ena_dev->net_device, + "Failed to destroy IO CQ. error: %d\n", ret); + + return ret; +} + +bool ena_com_get_admin_running_state(struct ena_com_dev *ena_dev) +{ + return ena_dev->admin_queue.running_state; +} + +void ena_com_set_admin_running_state(struct ena_com_dev *ena_dev, bool state) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + unsigned long flags = 0; + + spin_lock_irqsave(&admin_queue->q_lock, flags); + ena_dev->admin_queue.running_state = state; + spin_unlock_irqrestore(&admin_queue->q_lock, flags); +} + +void ena_com_admin_aenq_enable(struct ena_com_dev *ena_dev) +{ + u16 depth = ena_dev->aenq.q_depth; + + WARN(ena_dev->aenq.head != depth, "Invalid AENQ state\n"); + + /* Init head_db to mark that all entries in the queue + * are initially available + */ + writel(depth, ena_dev->reg_bar + ENA_REGS_AENQ_HEAD_DB_OFF); +} + +int ena_com_set_aenq_config(struct ena_com_dev *ena_dev, u32 groups_flag) +{ + struct ena_com_admin_queue *admin_queue; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + struct ena_admin_get_feat_resp get_resp; + int ret; + + ret = ena_com_get_feature(ena_dev, &get_resp, ENA_ADMIN_AENQ_CONFIG, 0); + if (ret) { + dev_info(ena_dev->dmadev, "Can't get aenq configuration\n"); + return ret; + } + + if ((get_resp.u.aenq.supported_groups & groups_flag) != groups_flag) { + netdev_warn(ena_dev->net_device, + "Trying to set unsupported aenq events. supported flag: 0x%x asked flag: 0x%x\n", + get_resp.u.aenq.supported_groups, groups_flag); + return -EOPNOTSUPP; + } + + memset(&cmd, 0x0, sizeof(cmd)); + admin_queue = &ena_dev->admin_queue; + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.aq_common_descriptor.flags = 0; + cmd.feat_common.feature_id = ENA_ADMIN_AENQ_CONFIG; + cmd.u.aenq.enabled_groups = groups_flag; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to config AENQ ret: %d\n", ret); + + return ret; +} + +int ena_com_get_dma_width(struct ena_com_dev *ena_dev) +{ + u32 caps = ena_com_reg_bar_read32(ena_dev, ENA_REGS_CAPS_OFF); + u32 width; + + if (unlikely(caps == ENA_MMIO_READ_TIMEOUT)) { + netdev_err(ena_dev->net_device, "Reg read timeout occurred\n"); + return -ETIME; + } + + width = (caps & ENA_REGS_CAPS_DMA_ADDR_WIDTH_MASK) >> + ENA_REGS_CAPS_DMA_ADDR_WIDTH_SHIFT; + + netdev_dbg(ena_dev->net_device, "ENA dma width: %d\n", width); + + if ((width < 32) || width > ENA_MAX_PHYS_ADDR_SIZE_BITS) { + netdev_err(ena_dev->net_device, "DMA width illegal value: %d\n", + width); + return -EINVAL; + } + + ena_dev->dma_addr_bits = width; + + return width; +} + +int ena_com_validate_version(struct ena_com_dev *ena_dev) +{ + u32 ver; + u32 ctrl_ver; + u32 ctrl_ver_masked; + + /* Make sure the ENA version and the controller version are at least + * as the driver expects + */ + ver = ena_com_reg_bar_read32(ena_dev, ENA_REGS_VERSION_OFF); + ctrl_ver = ena_com_reg_bar_read32(ena_dev, + ENA_REGS_CONTROLLER_VERSION_OFF); + + if (unlikely((ver == ENA_MMIO_READ_TIMEOUT) || + (ctrl_ver == ENA_MMIO_READ_TIMEOUT))) { + netdev_err(ena_dev->net_device, "Reg read timeout occurred\n"); + return -ETIME; + } + + dev_info(ena_dev->dmadev, "ENA device version: %d.%d\n", + (ver & ENA_REGS_VERSION_MAJOR_VERSION_MASK) >> + ENA_REGS_VERSION_MAJOR_VERSION_SHIFT, + ver & ENA_REGS_VERSION_MINOR_VERSION_MASK); + + dev_info(ena_dev->dmadev, + "ENA controller version: %d.%d.%d implementation version %d\n", + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_MASK) >> + ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_SHIFT, + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_MASK) >> + ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_SHIFT, + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_SUBMINOR_VERSION_MASK), + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_IMPL_ID_MASK) >> + ENA_REGS_CONTROLLER_VERSION_IMPL_ID_SHIFT); + + ctrl_ver_masked = + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_MASK) | + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_MASK) | + (ctrl_ver & ENA_REGS_CONTROLLER_VERSION_SUBMINOR_VERSION_MASK); + + /* Validate the ctrl version without the implementation ID */ + if (ctrl_ver_masked < MIN_ENA_CTRL_VER) { + netdev_err(ena_dev->net_device, + "ENA ctrl version is lower than the minimal ctrl version the driver supports\n"); + return -1; + } + + return 0; +} + +static void +ena_com_free_ena_admin_queue_comp_ctx(struct ena_com_dev *ena_dev, + struct ena_com_admin_queue *admin_queue) + +{ + if (!admin_queue->comp_ctx) + return; + + devm_kfree(ena_dev->dmadev, admin_queue->comp_ctx); + + admin_queue->comp_ctx = NULL; +} + +void ena_com_admin_destroy(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_com_admin_cq *cq = &admin_queue->cq; + struct ena_com_admin_sq *sq = &admin_queue->sq; + struct ena_com_aenq *aenq = &ena_dev->aenq; + u16 size; + + ena_com_free_ena_admin_queue_comp_ctx(ena_dev, admin_queue); + + size = ADMIN_SQ_SIZE(admin_queue->q_depth); + if (sq->entries) + dma_free_coherent(ena_dev->dmadev, size, sq->entries, + sq->dma_addr); + sq->entries = NULL; + + size = ADMIN_CQ_SIZE(admin_queue->q_depth); + if (cq->entries) + dma_free_coherent(ena_dev->dmadev, size, cq->entries, + cq->dma_addr); + cq->entries = NULL; + + size = ADMIN_AENQ_SIZE(aenq->q_depth); + if (ena_dev->aenq.entries) + dma_free_coherent(ena_dev->dmadev, size, aenq->entries, + aenq->dma_addr); + aenq->entries = NULL; +} + +void ena_com_set_admin_polling_mode(struct ena_com_dev *ena_dev, bool polling) +{ + u32 mask_value = 0; + + if (polling) + mask_value = ENA_REGS_ADMIN_INTR_MASK; + + writel(mask_value, ena_dev->reg_bar + ENA_REGS_INTR_MASK_OFF); + ena_dev->admin_queue.polling = polling; +} + +void ena_com_set_admin_auto_polling_mode(struct ena_com_dev *ena_dev, + bool polling) +{ + ena_dev->admin_queue.auto_polling = polling; +} + +int ena_com_mmio_reg_read_request_init(struct ena_com_dev *ena_dev) +{ + struct ena_com_mmio_read *mmio_read = &ena_dev->mmio_read; + + spin_lock_init(&mmio_read->lock); + mmio_read->read_resp = + dma_alloc_coherent(ena_dev->dmadev, + sizeof(*mmio_read->read_resp), + &mmio_read->read_resp_dma_addr, GFP_KERNEL); + if (unlikely(!mmio_read->read_resp)) + goto err; + + ena_com_mmio_reg_read_request_write_dev_addr(ena_dev); + + mmio_read->read_resp->req_id = 0x0; + mmio_read->seq_num = 0x0; + mmio_read->readless_supported = true; + + return 0; + +err: + + return -ENOMEM; +} + +void ena_com_set_mmio_read_mode(struct ena_com_dev *ena_dev, bool readless_supported) +{ + struct ena_com_mmio_read *mmio_read = &ena_dev->mmio_read; + + mmio_read->readless_supported = readless_supported; +} + +void ena_com_mmio_reg_read_request_destroy(struct ena_com_dev *ena_dev) +{ + struct ena_com_mmio_read *mmio_read = &ena_dev->mmio_read; + + writel(0x0, ena_dev->reg_bar + ENA_REGS_MMIO_RESP_LO_OFF); + writel(0x0, ena_dev->reg_bar + ENA_REGS_MMIO_RESP_HI_OFF); + + dma_free_coherent(ena_dev->dmadev, sizeof(*mmio_read->read_resp), + mmio_read->read_resp, mmio_read->read_resp_dma_addr); + + mmio_read->read_resp = NULL; +} + +void ena_com_mmio_reg_read_request_write_dev_addr(struct ena_com_dev *ena_dev) +{ + struct ena_com_mmio_read *mmio_read = &ena_dev->mmio_read; + u32 addr_low, addr_high; + + addr_low = ENA_DMA_ADDR_TO_UINT32_LOW(mmio_read->read_resp_dma_addr); + addr_high = ENA_DMA_ADDR_TO_UINT32_HIGH(mmio_read->read_resp_dma_addr); + + writel(addr_low, ena_dev->reg_bar + ENA_REGS_MMIO_RESP_LO_OFF); + writel(addr_high, ena_dev->reg_bar + ENA_REGS_MMIO_RESP_HI_OFF); +} + +int ena_com_admin_init(struct ena_com_dev *ena_dev, + struct ena_aenq_handlers *aenq_handlers) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + u32 aq_caps, acq_caps, dev_sts, addr_low, addr_high; + int ret; + + dev_sts = ena_com_reg_bar_read32(ena_dev, ENA_REGS_DEV_STS_OFF); + + if (unlikely(dev_sts == ENA_MMIO_READ_TIMEOUT)) { + netdev_err(ena_dev->net_device, "Reg read timeout occurred\n"); + return -ETIME; + } + + if (!(dev_sts & ENA_REGS_DEV_STS_READY_MASK)) { + netdev_err(ena_dev->net_device, + "Device isn't ready, abort com init\n"); + return -ENODEV; + } + + admin_queue->q_depth = ENA_ADMIN_QUEUE_DEPTH; + + admin_queue->q_dmadev = ena_dev->dmadev; + admin_queue->polling = false; + admin_queue->curr_cmd_id = 0; + + atomic_set(&admin_queue->outstanding_cmds, 0); + + spin_lock_init(&admin_queue->q_lock); + + ret = ena_com_init_comp_ctxt(admin_queue); + if (ret) + goto error; + + ret = ena_com_admin_init_sq(admin_queue); + if (ret) + goto error; + + ret = ena_com_admin_init_cq(admin_queue); + if (ret) + goto error; + + admin_queue->sq.db_addr = (u32 __iomem *)((uintptr_t)ena_dev->reg_bar + + ENA_REGS_AQ_DB_OFF); + + addr_low = ENA_DMA_ADDR_TO_UINT32_LOW(admin_queue->sq.dma_addr); + addr_high = ENA_DMA_ADDR_TO_UINT32_HIGH(admin_queue->sq.dma_addr); + + writel(addr_low, ena_dev->reg_bar + ENA_REGS_AQ_BASE_LO_OFF); + writel(addr_high, ena_dev->reg_bar + ENA_REGS_AQ_BASE_HI_OFF); + + addr_low = ENA_DMA_ADDR_TO_UINT32_LOW(admin_queue->cq.dma_addr); + addr_high = ENA_DMA_ADDR_TO_UINT32_HIGH(admin_queue->cq.dma_addr); + + writel(addr_low, ena_dev->reg_bar + ENA_REGS_ACQ_BASE_LO_OFF); + writel(addr_high, ena_dev->reg_bar + ENA_REGS_ACQ_BASE_HI_OFF); + + aq_caps = 0; + aq_caps |= admin_queue->q_depth & ENA_REGS_AQ_CAPS_AQ_DEPTH_MASK; + aq_caps |= (sizeof(struct ena_admin_aq_entry) << + ENA_REGS_AQ_CAPS_AQ_ENTRY_SIZE_SHIFT) & + ENA_REGS_AQ_CAPS_AQ_ENTRY_SIZE_MASK; + + acq_caps = 0; + acq_caps |= admin_queue->q_depth & ENA_REGS_ACQ_CAPS_ACQ_DEPTH_MASK; + acq_caps |= (sizeof(struct ena_admin_acq_entry) << + ENA_REGS_ACQ_CAPS_ACQ_ENTRY_SIZE_SHIFT) & + ENA_REGS_ACQ_CAPS_ACQ_ENTRY_SIZE_MASK; + + writel(aq_caps, ena_dev->reg_bar + ENA_REGS_AQ_CAPS_OFF); + writel(acq_caps, ena_dev->reg_bar + ENA_REGS_ACQ_CAPS_OFF); + ret = ena_com_admin_init_aenq(ena_dev, aenq_handlers); + if (ret) + goto error; + + admin_queue->ena_dev = ena_dev; + admin_queue->running_state = true; + + return 0; +error: + ena_com_admin_destroy(ena_dev); + + return ret; +} + +int ena_com_create_io_queue(struct ena_com_dev *ena_dev, + struct ena_com_create_io_ctx *ctx) +{ + struct ena_com_io_sq *io_sq; + struct ena_com_io_cq *io_cq; + int ret; + + if (ctx->qid >= ENA_TOTAL_NUM_QUEUES) { + netdev_err(ena_dev->net_device, + "Qid (%d) is bigger than max num of queues (%d)\n", + ctx->qid, ENA_TOTAL_NUM_QUEUES); + return -EINVAL; + } + + io_sq = &ena_dev->io_sq_queues[ctx->qid]; + io_cq = &ena_dev->io_cq_queues[ctx->qid]; + + memset(io_sq, 0x0, sizeof(*io_sq)); + memset(io_cq, 0x0, sizeof(*io_cq)); + + /* Init CQ */ + io_cq->q_depth = ctx->queue_size; + io_cq->direction = ctx->direction; + io_cq->qid = ctx->qid; + + io_cq->msix_vector = ctx->msix_vector; + + io_sq->q_depth = ctx->queue_size; + io_sq->direction = ctx->direction; + io_sq->qid = ctx->qid; + + io_sq->mem_queue_type = ctx->mem_queue_type; + + if (ctx->direction == ENA_COM_IO_QUEUE_DIRECTION_TX) + /* header length is limited to 8 bits */ + io_sq->tx_max_header_size = + min_t(u32, ena_dev->tx_max_header_size, SZ_256); + + ret = ena_com_init_io_sq(ena_dev, ctx, io_sq); + if (ret) + goto error; + ret = ena_com_init_io_cq(ena_dev, ctx, io_cq); + if (ret) + goto error; + + ret = ena_com_create_io_cq(ena_dev, io_cq); + if (ret) + goto error; + + ret = ena_com_create_io_sq(ena_dev, io_sq, io_cq->idx); + if (ret) + goto destroy_io_cq; + + return 0; + +destroy_io_cq: + ena_com_destroy_io_cq(ena_dev, io_cq); +error: + ena_com_io_queue_free(ena_dev, io_sq, io_cq); + return ret; +} + +void ena_com_destroy_io_queue(struct ena_com_dev *ena_dev, u16 qid) +{ + struct ena_com_io_sq *io_sq; + struct ena_com_io_cq *io_cq; + + if (qid >= ENA_TOTAL_NUM_QUEUES) { + netdev_err(ena_dev->net_device, + "Qid (%d) is bigger than max num of queues (%d)\n", + qid, ENA_TOTAL_NUM_QUEUES); + return; + } + + io_sq = &ena_dev->io_sq_queues[qid]; + io_cq = &ena_dev->io_cq_queues[qid]; + + ena_com_destroy_io_sq(ena_dev, io_sq); + ena_com_destroy_io_cq(ena_dev, io_cq); + + ena_com_io_queue_free(ena_dev, io_sq, io_cq); +} + +int ena_com_get_link_params(struct ena_com_dev *ena_dev, + struct ena_admin_get_feat_resp *resp) +{ + return ena_com_get_feature(ena_dev, resp, ENA_ADMIN_LINK_CONFIG, 0); +} + +int ena_com_get_dev_attr_feat(struct ena_com_dev *ena_dev, + struct ena_com_dev_get_features_ctx *get_feat_ctx) +{ + struct ena_admin_get_feat_resp get_resp; + int rc; + + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_DEVICE_ATTRIBUTES, 0); + if (rc) + return rc; + + memcpy(&get_feat_ctx->dev_attr, &get_resp.u.dev_attr, + sizeof(get_resp.u.dev_attr)); + + ena_dev->supported_features = get_resp.u.dev_attr.supported_features; + ena_dev->capabilities = get_resp.u.dev_attr.capabilities; + + if (ena_dev->supported_features & BIT(ENA_ADMIN_MAX_QUEUES_EXT)) { + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_MAX_QUEUES_EXT, + ENA_FEATURE_MAX_QUEUE_EXT_VER); + if (rc) + return rc; + + if (get_resp.u.max_queue_ext.version != + ENA_FEATURE_MAX_QUEUE_EXT_VER) + return -EINVAL; + + memcpy(&get_feat_ctx->max_queue_ext, &get_resp.u.max_queue_ext, + sizeof(get_resp.u.max_queue_ext)); + ena_dev->tx_max_header_size = + get_resp.u.max_queue_ext.max_queue_ext.max_tx_header_size; + } else { + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_MAX_QUEUES_NUM, 0); + memcpy(&get_feat_ctx->max_queues, &get_resp.u.max_queue, + sizeof(get_resp.u.max_queue)); + ena_dev->tx_max_header_size = + get_resp.u.max_queue.max_header_size; + + if (rc) + return rc; + } + + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_AENQ_CONFIG, 0); + if (rc) + return rc; + + memcpy(&get_feat_ctx->aenq, &get_resp.u.aenq, + sizeof(get_resp.u.aenq)); + + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_STATELESS_OFFLOAD_CONFIG, 0); + if (rc) + return rc; + + memcpy(&get_feat_ctx->offload, &get_resp.u.offload, + sizeof(get_resp.u.offload)); + + /* Driver hints isn't mandatory admin command. So in case the + * command isn't supported set driver hints to 0 + */ + rc = ena_com_get_feature(ena_dev, &get_resp, ENA_ADMIN_HW_HINTS, 0); + + if (!rc) + memcpy(&get_feat_ctx->hw_hints, &get_resp.u.hw_hints, + sizeof(get_resp.u.hw_hints)); + else if (rc == -EOPNOTSUPP) + memset(&get_feat_ctx->hw_hints, 0x0, + sizeof(get_feat_ctx->hw_hints)); + else + return rc; + + rc = ena_com_get_feature(ena_dev, &get_resp, ENA_ADMIN_LLQ, 0); + if (!rc) + memcpy(&get_feat_ctx->llq, &get_resp.u.llq, + sizeof(get_resp.u.llq)); + else if (rc == -EOPNOTSUPP) + memset(&get_feat_ctx->llq, 0x0, sizeof(get_feat_ctx->llq)); + else + return rc; + + return 0; +} + +void ena_com_admin_q_comp_intr_handler(struct ena_com_dev *ena_dev) +{ + ena_com_handle_admin_completion(&ena_dev->admin_queue); +} + +/* ena_handle_specific_aenq_event: + * return the handler that is relevant to the specific event group + */ +static ena_aenq_handler ena_com_get_specific_aenq_cb(struct ena_com_dev *ena_dev, + u16 group) +{ + struct ena_aenq_handlers *aenq_handlers = ena_dev->aenq.aenq_handlers; + + if ((group < ENA_MAX_HANDLERS) && aenq_handlers->handlers[group]) + return aenq_handlers->handlers[group]; + + return aenq_handlers->unimplemented_handler; +} + +/* ena_aenq_intr_handler: + * handles the aenq incoming events. + * pop events from the queue and apply the specific handler + */ +void ena_com_aenq_intr_handler(struct ena_com_dev *ena_dev, void *data) +{ + struct ena_admin_aenq_entry *aenq_e; + struct ena_admin_aenq_common_desc *aenq_common; + struct ena_com_aenq *aenq = &ena_dev->aenq; + u64 timestamp; + ena_aenq_handler handler_cb; + u16 masked_head, processed = 0; + u8 phase; + + masked_head = aenq->head & (aenq->q_depth - 1); + phase = aenq->phase; + aenq_e = &aenq->entries[masked_head]; /* Get first entry */ + aenq_common = &aenq_e->aenq_common_desc; + + /* Go over all the events */ + while ((READ_ONCE(aenq_common->flags) & + ENA_ADMIN_AENQ_COMMON_DESC_PHASE_MASK) == phase) { + /* Make sure the phase bit (ownership) is as expected before + * reading the rest of the descriptor. + */ + dma_rmb(); + + timestamp = (u64)aenq_common->timestamp_low | + ((u64)aenq_common->timestamp_high << 32); + + netdev_dbg(ena_dev->net_device, + "AENQ! Group[%x] Syndrome[%x] timestamp: [%llus]\n", + aenq_common->group, aenq_common->syndrome, timestamp); + + /* Handle specific event*/ + handler_cb = ena_com_get_specific_aenq_cb(ena_dev, + aenq_common->group); + handler_cb(data, aenq_e); /* call the actual event handler*/ + + /* Get next event entry */ + masked_head++; + processed++; + + if (unlikely(masked_head == aenq->q_depth)) { + masked_head = 0; + phase = !phase; + } + aenq_e = &aenq->entries[masked_head]; + aenq_common = &aenq_e->aenq_common_desc; + } + + aenq->head += processed; + aenq->phase = phase; + + /* Don't update aenq doorbell if there weren't any processed events */ + if (!processed) + return; + + /* write the aenq doorbell after all AENQ descriptors were read */ + mb(); + writel_relaxed((u32)aenq->head, + ena_dev->reg_bar + ENA_REGS_AENQ_HEAD_DB_OFF); +} + +int ena_com_dev_reset(struct ena_com_dev *ena_dev, + enum ena_regs_reset_reason_types reset_reason) +{ + u32 stat, timeout, cap, reset_val; + int rc; + + stat = ena_com_reg_bar_read32(ena_dev, ENA_REGS_DEV_STS_OFF); + cap = ena_com_reg_bar_read32(ena_dev, ENA_REGS_CAPS_OFF); + + if (unlikely((stat == ENA_MMIO_READ_TIMEOUT) || + (cap == ENA_MMIO_READ_TIMEOUT))) { + netdev_err(ena_dev->net_device, "Reg read32 timeout occurred\n"); + return -ETIME; + } + + if ((stat & ENA_REGS_DEV_STS_READY_MASK) == 0) { + netdev_err(ena_dev->net_device, + "Device isn't ready, can't reset device\n"); + return -EINVAL; + } + + timeout = (cap & ENA_REGS_CAPS_RESET_TIMEOUT_MASK) >> + ENA_REGS_CAPS_RESET_TIMEOUT_SHIFT; + if (timeout == 0) { + netdev_err(ena_dev->net_device, "Invalid timeout value\n"); + return -EINVAL; + } + + /* start reset */ + reset_val = ENA_REGS_DEV_CTL_DEV_RESET_MASK; + reset_val |= (reset_reason << ENA_REGS_DEV_CTL_RESET_REASON_SHIFT) & + ENA_REGS_DEV_CTL_RESET_REASON_MASK; + writel(reset_val, ena_dev->reg_bar + ENA_REGS_DEV_CTL_OFF); + + /* Write again the MMIO read request address */ + ena_com_mmio_reg_read_request_write_dev_addr(ena_dev); + + rc = wait_for_reset_state(ena_dev, timeout, + ENA_REGS_DEV_STS_RESET_IN_PROGRESS_MASK); + if (rc != 0) { + netdev_err(ena_dev->net_device, + "Reset indication didn't turn on\n"); + return rc; + } + + /* reset done */ + writel(0, ena_dev->reg_bar + ENA_REGS_DEV_CTL_OFF); + rc = wait_for_reset_state(ena_dev, timeout, 0); + if (rc != 0) { + netdev_err(ena_dev->net_device, + "Reset indication didn't turn off\n"); + return rc; + } + + timeout = (cap & ENA_REGS_CAPS_ADMIN_CMD_TO_MASK) >> + ENA_REGS_CAPS_ADMIN_CMD_TO_SHIFT; + if (timeout) + /* the resolution of timeout reg is 100ms */ + ena_dev->admin_queue.completion_timeout = timeout * 100000; + else + ena_dev->admin_queue.completion_timeout = ADMIN_CMD_TIMEOUT_US; + + return 0; +} + +static int ena_get_dev_stats(struct ena_com_dev *ena_dev, + struct ena_com_stats_ctx *ctx, + enum ena_admin_get_stats_type type) +{ + struct ena_admin_aq_get_stats_cmd *get_cmd = &ctx->get_cmd; + struct ena_admin_acq_get_stats_resp *get_resp = &ctx->get_resp; + struct ena_com_admin_queue *admin_queue; + int ret; + + admin_queue = &ena_dev->admin_queue; + + get_cmd->aq_common_descriptor.opcode = ENA_ADMIN_GET_STATS; + get_cmd->aq_common_descriptor.flags = 0; + get_cmd->type = type; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)get_cmd, + sizeof(*get_cmd), + (struct ena_admin_acq_entry *)get_resp, + sizeof(*get_resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to get stats. error: %d\n", ret); + + return ret; +} + +int ena_com_get_eni_stats(struct ena_com_dev *ena_dev, + struct ena_admin_eni_stats *stats) +{ + struct ena_com_stats_ctx ctx; + int ret; + + if (!ena_com_get_cap(ena_dev, ENA_ADMIN_ENI_STATS)) { + netdev_err(ena_dev->net_device, + "Capability %d isn't supported\n", + ENA_ADMIN_ENI_STATS); + return -EOPNOTSUPP; + } + + memset(&ctx, 0x0, sizeof(ctx)); + ret = ena_get_dev_stats(ena_dev, &ctx, ENA_ADMIN_GET_STATS_TYPE_ENI); + if (likely(ret == 0)) + memcpy(stats, &ctx.get_resp.u.eni_stats, + sizeof(ctx.get_resp.u.eni_stats)); + + return ret; +} + +int ena_com_get_dev_basic_stats(struct ena_com_dev *ena_dev, + struct ena_admin_basic_stats *stats) +{ + struct ena_com_stats_ctx ctx; + int ret; + + memset(&ctx, 0x0, sizeof(ctx)); + ret = ena_get_dev_stats(ena_dev, &ctx, ENA_ADMIN_GET_STATS_TYPE_BASIC); + if (likely(ret == 0)) + memcpy(stats, &ctx.get_resp.u.basic_stats, + sizeof(ctx.get_resp.u.basic_stats)); + + return ret; +} + +int ena_com_set_dev_mtu(struct ena_com_dev *ena_dev, u32 mtu) +{ + struct ena_com_admin_queue *admin_queue; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + int ret; + + if (!ena_com_check_supported_feature_id(ena_dev, ENA_ADMIN_MTU)) { + netdev_dbg(ena_dev->net_device, "Feature %d isn't supported\n", + ENA_ADMIN_MTU); + return -EOPNOTSUPP; + } + + memset(&cmd, 0x0, sizeof(cmd)); + admin_queue = &ena_dev->admin_queue; + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.aq_common_descriptor.flags = 0; + cmd.feat_common.feature_id = ENA_ADMIN_MTU; + cmd.u.mtu.mtu = mtu; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to set mtu %d. error: %d\n", mtu, ret); + + return ret; +} + +int ena_com_get_offload_settings(struct ena_com_dev *ena_dev, + struct ena_admin_feature_offload_desc *offload) +{ + int ret; + struct ena_admin_get_feat_resp resp; + + ret = ena_com_get_feature(ena_dev, &resp, + ENA_ADMIN_STATELESS_OFFLOAD_CONFIG, 0); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, + "Failed to get offload capabilities %d\n", ret); + return ret; + } + + memcpy(offload, &resp.u.offload, sizeof(resp.u.offload)); + + return 0; +} + +int ena_com_set_hash_function(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + struct ena_admin_get_feat_resp get_resp; + int ret; + + if (!ena_com_check_supported_feature_id(ena_dev, + ENA_ADMIN_RSS_HASH_FUNCTION)) { + netdev_dbg(ena_dev->net_device, "Feature %d isn't supported\n", + ENA_ADMIN_RSS_HASH_FUNCTION); + return -EOPNOTSUPP; + } + + /* Validate hash function is supported */ + ret = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_RSS_HASH_FUNCTION, 0); + if (unlikely(ret)) + return ret; + + if (!(get_resp.u.flow_hash_func.supported_func & BIT(rss->hash_func))) { + netdev_err(ena_dev->net_device, + "Func hash %d isn't supported by device, abort\n", + rss->hash_func); + return -EOPNOTSUPP; + } + + memset(&cmd, 0x0, sizeof(cmd)); + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.aq_common_descriptor.flags = + ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_MASK; + cmd.feat_common.feature_id = ENA_ADMIN_RSS_HASH_FUNCTION; + cmd.u.flow_hash_func.init_val = rss->hash_init_val; + cmd.u.flow_hash_func.selected_func = 1 << rss->hash_func; + + ret = ena_com_mem_addr_set(ena_dev, + &cmd.control_buffer.address, + rss->hash_key_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + cmd.control_buffer.length = sizeof(*rss->hash_key); + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, + "Failed to set hash function %d. error: %d\n", + rss->hash_func, ret); + return -EINVAL; + } + + return 0; +} + +int ena_com_fill_hash_function(struct ena_com_dev *ena_dev, + enum ena_admin_hash_functions func, + const u8 *key, u16 key_len, u32 init_val) +{ + struct ena_admin_feature_rss_flow_hash_control *hash_key; + struct ena_admin_get_feat_resp get_resp; + enum ena_admin_hash_functions old_func; + struct ena_rss *rss = &ena_dev->rss; + int rc; + + hash_key = rss->hash_key; + + /* Make sure size is a mult of DWs */ + if (unlikely(key_len & 0x3)) + return -EINVAL; + + rc = ena_com_get_feature_ex(ena_dev, &get_resp, + ENA_ADMIN_RSS_HASH_FUNCTION, + rss->hash_key_dma_addr, + sizeof(*rss->hash_key), 0); + if (unlikely(rc)) + return rc; + + if (!(BIT(func) & get_resp.u.flow_hash_func.supported_func)) { + netdev_err(ena_dev->net_device, + "Flow hash function %d isn't supported\n", func); + return -EOPNOTSUPP; + } + + if ((func == ENA_ADMIN_TOEPLITZ) && key) { + if (key_len != sizeof(hash_key->key)) { + netdev_err(ena_dev->net_device, + "key len (%u) doesn't equal the supported size (%zu)\n", + key_len, sizeof(hash_key->key)); + return -EINVAL; + } + memcpy(hash_key->key, key, key_len); + hash_key->key_parts = key_len / sizeof(hash_key->key[0]); + } + + rss->hash_init_val = init_val; + old_func = rss->hash_func; + rss->hash_func = func; + rc = ena_com_set_hash_function(ena_dev); + + /* Restore the old function */ + if (unlikely(rc)) + rss->hash_func = old_func; + + return rc; +} + +int ena_com_get_hash_function(struct ena_com_dev *ena_dev, + enum ena_admin_hash_functions *func) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_get_feat_resp get_resp; + int rc; + + if (unlikely(!func)) + return -EINVAL; + + rc = ena_com_get_feature_ex(ena_dev, &get_resp, + ENA_ADMIN_RSS_HASH_FUNCTION, + rss->hash_key_dma_addr, + sizeof(*rss->hash_key), 0); + if (unlikely(rc)) + return rc; + + /* ffs() returns 1 in case the lsb is set */ + rss->hash_func = ffs(get_resp.u.flow_hash_func.selected_func); + if (rss->hash_func) + rss->hash_func--; + + *func = rss->hash_func; + + return 0; +} + +int ena_com_get_hash_key(struct ena_com_dev *ena_dev, u8 *key) +{ + struct ena_admin_feature_rss_flow_hash_control *hash_key = + ena_dev->rss.hash_key; + + if (key) + memcpy(key, hash_key->key, + (size_t)(hash_key->key_parts) * sizeof(hash_key->key[0])); + + return 0; +} + +int ena_com_get_hash_ctrl(struct ena_com_dev *ena_dev, + enum ena_admin_flow_hash_proto proto, + u16 *fields) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_get_feat_resp get_resp; + int rc; + + rc = ena_com_get_feature_ex(ena_dev, &get_resp, + ENA_ADMIN_RSS_HASH_INPUT, + rss->hash_ctrl_dma_addr, + sizeof(*rss->hash_ctrl), 0); + if (unlikely(rc)) + return rc; + + if (fields) + *fields = rss->hash_ctrl->selected_fields[proto].fields; + + return 0; +} + +int ena_com_set_hash_ctrl(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_feature_rss_hash_control *hash_ctrl = rss->hash_ctrl; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + int ret; + + if (!ena_com_check_supported_feature_id(ena_dev, + ENA_ADMIN_RSS_HASH_INPUT)) { + netdev_dbg(ena_dev->net_device, "Feature %d isn't supported\n", + ENA_ADMIN_RSS_HASH_INPUT); + return -EOPNOTSUPP; + } + + memset(&cmd, 0x0, sizeof(cmd)); + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.aq_common_descriptor.flags = + ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_MASK; + cmd.feat_common.feature_id = ENA_ADMIN_RSS_HASH_INPUT; + cmd.u.flow_hash_input.enabled_input_sort = + ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L3_SORT_MASK | + ENA_ADMIN_FEATURE_RSS_FLOW_HASH_INPUT_L4_SORT_MASK; + + ret = ena_com_mem_addr_set(ena_dev, + &cmd.control_buffer.address, + rss->hash_ctrl_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + cmd.control_buffer.length = sizeof(*hash_ctrl); + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to set hash input. error: %d\n", ret); + + return ret; +} + +int ena_com_set_default_hash_ctrl(struct ena_com_dev *ena_dev) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_feature_rss_hash_control *hash_ctrl = + rss->hash_ctrl; + u16 available_fields = 0; + int rc, i; + + /* Get the supported hash input */ + rc = ena_com_get_hash_ctrl(ena_dev, 0, NULL); + if (unlikely(rc)) + return rc; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_TCP4].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA | + ENA_ADMIN_RSS_L4_DP | ENA_ADMIN_RSS_L4_SP; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_UDP4].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA | + ENA_ADMIN_RSS_L4_DP | ENA_ADMIN_RSS_L4_SP; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_TCP6].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA | + ENA_ADMIN_RSS_L4_DP | ENA_ADMIN_RSS_L4_SP; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_UDP6].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA | + ENA_ADMIN_RSS_L4_DP | ENA_ADMIN_RSS_L4_SP; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_IP4].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_IP6].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_IP4_FRAG].fields = + ENA_ADMIN_RSS_L3_SA | ENA_ADMIN_RSS_L3_DA; + + hash_ctrl->selected_fields[ENA_ADMIN_RSS_NOT_IP].fields = + ENA_ADMIN_RSS_L2_DA | ENA_ADMIN_RSS_L2_SA; + + for (i = 0; i < ENA_ADMIN_RSS_PROTO_NUM; i++) { + available_fields = hash_ctrl->selected_fields[i].fields & + hash_ctrl->supported_fields[i].fields; + if (available_fields != hash_ctrl->selected_fields[i].fields) { + netdev_err(ena_dev->net_device, + "Hash control doesn't support all the desire configuration. proto %x supported %x selected %x\n", + i, hash_ctrl->supported_fields[i].fields, + hash_ctrl->selected_fields[i].fields); + return -EOPNOTSUPP; + } + } + + rc = ena_com_set_hash_ctrl(ena_dev); + + /* In case of failure, restore the old hash ctrl */ + if (unlikely(rc)) + ena_com_get_hash_ctrl(ena_dev, 0, NULL); + + return rc; +} + +int ena_com_fill_hash_ctrl(struct ena_com_dev *ena_dev, + enum ena_admin_flow_hash_proto proto, + u16 hash_fields) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_feature_rss_hash_control *hash_ctrl = rss->hash_ctrl; + u16 supported_fields; + int rc; + + if (proto >= ENA_ADMIN_RSS_PROTO_NUM) { + netdev_err(ena_dev->net_device, "Invalid proto num (%u)\n", + proto); + return -EINVAL; + } + + /* Get the ctrl table */ + rc = ena_com_get_hash_ctrl(ena_dev, proto, NULL); + if (unlikely(rc)) + return rc; + + /* Make sure all the fields are supported */ + supported_fields = hash_ctrl->supported_fields[proto].fields; + if ((hash_fields & supported_fields) != hash_fields) { + netdev_err(ena_dev->net_device, + "Proto %d doesn't support the required fields %x. supports only: %x\n", + proto, hash_fields, supported_fields); + } + + hash_ctrl->selected_fields[proto].fields = hash_fields; + + rc = ena_com_set_hash_ctrl(ena_dev); + + /* In case of failure, restore the old hash ctrl */ + if (unlikely(rc)) + ena_com_get_hash_ctrl(ena_dev, 0, NULL); + + return 0; +} + +int ena_com_indirect_table_fill_entry(struct ena_com_dev *ena_dev, + u16 entry_idx, u16 entry_value) +{ + struct ena_rss *rss = &ena_dev->rss; + + if (unlikely(entry_idx >= (1 << rss->tbl_log_size))) + return -EINVAL; + + if (unlikely((entry_value > ENA_TOTAL_NUM_QUEUES))) + return -EINVAL; + + rss->host_rss_ind_tbl[entry_idx] = entry_value; + + return 0; +} + +int ena_com_indirect_table_set(struct ena_com_dev *ena_dev) +{ + struct ena_com_admin_queue *admin_queue = &ena_dev->admin_queue; + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + int ret; + + if (!ena_com_check_supported_feature_id( + ena_dev, ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG)) { + netdev_dbg(ena_dev->net_device, "Feature %d isn't supported\n", + ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG); + return -EOPNOTSUPP; + } + + ret = ena_com_ind_tbl_convert_to_device(ena_dev); + if (ret) { + netdev_err(ena_dev->net_device, + "Failed to convert host indirection table to device table\n"); + return ret; + } + + memset(&cmd, 0x0, sizeof(cmd)); + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.aq_common_descriptor.flags = + ENA_ADMIN_AQ_COMMON_DESC_CTRL_DATA_INDIRECT_MASK; + cmd.feat_common.feature_id = ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG; + cmd.u.ind_table.size = rss->tbl_log_size; + cmd.u.ind_table.inline_index = 0xFFFFFFFF; + + ret = ena_com_mem_addr_set(ena_dev, + &cmd.control_buffer.address, + rss->rss_ind_tbl_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + cmd.control_buffer.length = (1ULL << rss->tbl_log_size) * + sizeof(struct ena_admin_rss_ind_table_entry); + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to set indirect table. error: %d\n", ret); + + return ret; +} + +int ena_com_indirect_table_get(struct ena_com_dev *ena_dev, u32 *ind_tbl) +{ + struct ena_rss *rss = &ena_dev->rss; + struct ena_admin_get_feat_resp get_resp; + u32 tbl_size; + int i, rc; + + tbl_size = (1ULL << rss->tbl_log_size) * + sizeof(struct ena_admin_rss_ind_table_entry); + + rc = ena_com_get_feature_ex(ena_dev, &get_resp, + ENA_ADMIN_RSS_INDIRECTION_TABLE_CONFIG, + rss->rss_ind_tbl_dma_addr, + tbl_size, 0); + if (unlikely(rc)) + return rc; + + if (!ind_tbl) + return 0; + + for (i = 0; i < (1 << rss->tbl_log_size); i++) + ind_tbl[i] = rss->host_rss_ind_tbl[i]; + + return 0; +} + +int ena_com_rss_init(struct ena_com_dev *ena_dev, u16 indr_tbl_log_size) +{ + int rc; + + memset(&ena_dev->rss, 0x0, sizeof(ena_dev->rss)); + + rc = ena_com_indirect_table_allocate(ena_dev, indr_tbl_log_size); + if (unlikely(rc)) + goto err_indr_tbl; + + /* The following function might return unsupported in case the + * device doesn't support setting the key / hash function. We can safely + * ignore this error and have indirection table support only. + */ + rc = ena_com_hash_key_allocate(ena_dev); + if (likely(!rc)) + ena_com_hash_key_fill_default_key(ena_dev); + else if (rc != -EOPNOTSUPP) + goto err_hash_key; + + rc = ena_com_hash_ctrl_init(ena_dev); + if (unlikely(rc)) + goto err_hash_ctrl; + + return 0; + +err_hash_ctrl: + ena_com_hash_key_destroy(ena_dev); +err_hash_key: + ena_com_indirect_table_destroy(ena_dev); +err_indr_tbl: + + return rc; +} + +void ena_com_rss_destroy(struct ena_com_dev *ena_dev) +{ + ena_com_indirect_table_destroy(ena_dev); + ena_com_hash_key_destroy(ena_dev); + ena_com_hash_ctrl_destroy(ena_dev); + + memset(&ena_dev->rss, 0x0, sizeof(ena_dev->rss)); +} + +int ena_com_allocate_host_info(struct ena_com_dev *ena_dev) +{ + struct ena_host_attribute *host_attr = &ena_dev->host_attr; + + host_attr->host_info = + dma_alloc_coherent(ena_dev->dmadev, SZ_4K, + &host_attr->host_info_dma_addr, GFP_KERNEL); + if (unlikely(!host_attr->host_info)) + return -ENOMEM; + + host_attr->host_info->ena_spec_version = ((ENA_COMMON_SPEC_VERSION_MAJOR << + ENA_REGS_VERSION_MAJOR_VERSION_SHIFT) | + (ENA_COMMON_SPEC_VERSION_MINOR)); + + return 0; +} + +int ena_com_allocate_debug_area(struct ena_com_dev *ena_dev, + u32 debug_area_size) +{ + struct ena_host_attribute *host_attr = &ena_dev->host_attr; + + host_attr->debug_area_virt_addr = + dma_alloc_coherent(ena_dev->dmadev, debug_area_size, + &host_attr->debug_area_dma_addr, GFP_KERNEL); + if (unlikely(!host_attr->debug_area_virt_addr)) { + host_attr->debug_area_size = 0; + return -ENOMEM; + } + + host_attr->debug_area_size = debug_area_size; + + return 0; +} + +void ena_com_delete_host_info(struct ena_com_dev *ena_dev) +{ + struct ena_host_attribute *host_attr = &ena_dev->host_attr; + + if (host_attr->host_info) { + dma_free_coherent(ena_dev->dmadev, SZ_4K, host_attr->host_info, + host_attr->host_info_dma_addr); + host_attr->host_info = NULL; + } +} + +void ena_com_delete_debug_area(struct ena_com_dev *ena_dev) +{ + struct ena_host_attribute *host_attr = &ena_dev->host_attr; + + if (host_attr->debug_area_virt_addr) { + dma_free_coherent(ena_dev->dmadev, host_attr->debug_area_size, + host_attr->debug_area_virt_addr, + host_attr->debug_area_dma_addr); + host_attr->debug_area_virt_addr = NULL; + } +} + +int ena_com_set_host_attributes(struct ena_com_dev *ena_dev) +{ + struct ena_host_attribute *host_attr = &ena_dev->host_attr; + struct ena_com_admin_queue *admin_queue; + struct ena_admin_set_feat_cmd cmd; + struct ena_admin_set_feat_resp resp; + + int ret; + + /* Host attribute config is called before ena_com_get_dev_attr_feat + * so ena_com can't check if the feature is supported. + */ + + memset(&cmd, 0x0, sizeof(cmd)); + admin_queue = &ena_dev->admin_queue; + + cmd.aq_common_descriptor.opcode = ENA_ADMIN_SET_FEATURE; + cmd.feat_common.feature_id = ENA_ADMIN_HOST_ATTR_CONFIG; + + ret = ena_com_mem_addr_set(ena_dev, + &cmd.u.host_attr.debug_ba, + host_attr->debug_area_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + ret = ena_com_mem_addr_set(ena_dev, + &cmd.u.host_attr.os_info_ba, + host_attr->host_info_dma_addr); + if (unlikely(ret)) { + netdev_err(ena_dev->net_device, "Memory address set failed\n"); + return ret; + } + + cmd.u.host_attr.debug_area_size = host_attr->debug_area_size; + + ret = ena_com_execute_admin_command(admin_queue, + (struct ena_admin_aq_entry *)&cmd, + sizeof(cmd), + (struct ena_admin_acq_entry *)&resp, + sizeof(resp)); + + if (unlikely(ret)) + netdev_err(ena_dev->net_device, + "Failed to set host attributes: %d\n", ret); + + return ret; +} + +/* Interrupt moderation */ +bool ena_com_interrupt_moderation_supported(struct ena_com_dev *ena_dev) +{ + return ena_com_check_supported_feature_id(ena_dev, + ENA_ADMIN_INTERRUPT_MODERATION); +} + +static int ena_com_update_nonadaptive_moderation_interval(struct ena_com_dev *ena_dev, + u32 coalesce_usecs, + u32 intr_delay_resolution, + u32 *intr_moder_interval) +{ + if (!intr_delay_resolution) { + netdev_err(ena_dev->net_device, + "Illegal interrupt delay granularity value\n"); + return -EFAULT; + } + + *intr_moder_interval = coalesce_usecs / intr_delay_resolution; + + return 0; +} + +int ena_com_update_nonadaptive_moderation_interval_tx(struct ena_com_dev *ena_dev, + u32 tx_coalesce_usecs) +{ + return ena_com_update_nonadaptive_moderation_interval(ena_dev, + tx_coalesce_usecs, + ena_dev->intr_delay_resolution, + &ena_dev->intr_moder_tx_interval); +} + +int ena_com_update_nonadaptive_moderation_interval_rx(struct ena_com_dev *ena_dev, + u32 rx_coalesce_usecs) +{ + return ena_com_update_nonadaptive_moderation_interval(ena_dev, + rx_coalesce_usecs, + ena_dev->intr_delay_resolution, + &ena_dev->intr_moder_rx_interval); +} + +int ena_com_init_interrupt_moderation(struct ena_com_dev *ena_dev) +{ + struct ena_admin_get_feat_resp get_resp; + u16 delay_resolution; + int rc; + + rc = ena_com_get_feature(ena_dev, &get_resp, + ENA_ADMIN_INTERRUPT_MODERATION, 0); + + if (rc) { + if (rc == -EOPNOTSUPP) { + netdev_dbg(ena_dev->net_device, + "Feature %d isn't supported\n", + ENA_ADMIN_INTERRUPT_MODERATION); + rc = 0; + } else { + netdev_err(ena_dev->net_device, + "Failed to get interrupt moderation admin cmd. rc: %d\n", + rc); + } + + /* no moderation supported, disable adaptive support */ + ena_com_disable_adaptive_moderation(ena_dev); + return rc; + } + + /* if moderation is supported by device we set adaptive moderation */ + delay_resolution = get_resp.u.intr_moderation.intr_delay_resolution; + ena_com_update_intr_delay_resolution(ena_dev, delay_resolution); + + /* Disable adaptive moderation by default - can be enabled later */ + ena_com_disable_adaptive_moderation(ena_dev); + + return 0; +} + +unsigned int ena_com_get_nonadaptive_moderation_interval_tx(struct ena_com_dev *ena_dev) +{ + return ena_dev->intr_moder_tx_interval; +} + +unsigned int ena_com_get_nonadaptive_moderation_interval_rx(struct ena_com_dev *ena_dev) +{ + return ena_dev->intr_moder_rx_interval; +} + +int ena_com_config_dev_mode(struct ena_com_dev *ena_dev, + struct ena_admin_feature_llq_desc *llq_features, + struct ena_llq_configurations *llq_default_cfg) +{ + struct ena_com_llq_info *llq_info = &ena_dev->llq_info; + int rc; + + if (!llq_features->max_llq_num) { + ena_dev->tx_mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_HOST; + return 0; + } + + rc = ena_com_config_llq_info(ena_dev, llq_features, llq_default_cfg); + if (rc) + return rc; + + ena_dev->tx_max_header_size = llq_info->desc_list_entry_size - + (llq_info->descs_num_before_header * sizeof(struct ena_eth_io_tx_desc)); + + if (unlikely(ena_dev->tx_max_header_size == 0)) { + netdev_err(ena_dev->net_device, + "The size of the LLQ entry is smaller than needed\n"); + return -EINVAL; + } + + ena_dev->tx_mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_DEV; + + return 0; +} diff --git a/drivers/net/ethernet/amazon/ena/ena_com.h b/drivers/net/ethernet/amazon/ena/ena_com.h new file mode 100644 index 0000000000..3c5081d9d2 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_com.h @@ -0,0 +1,1025 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#ifndef ENA_COM +#define ENA_COM + +#include <linux/compiler.h> +#include <linux/delay.h> +#include <linux/dma-mapping.h> +#include <linux/gfp.h> +#include <linux/io.h> +#include <linux/prefetch.h> +#include <linux/sched.h> +#include <linux/sizes.h> +#include <linux/spinlock.h> +#include <linux/types.h> +#include <linux/wait.h> +#include <linux/netdevice.h> + +#include "ena_common_defs.h" +#include "ena_admin_defs.h" +#include "ena_eth_io_defs.h" +#include "ena_regs_defs.h" + +#undef pr_fmt +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#define ENA_MAX_NUM_IO_QUEUES 128U +/* We need to queues for each IO (on for Tx and one for Rx) */ +#define ENA_TOTAL_NUM_QUEUES (2 * (ENA_MAX_NUM_IO_QUEUES)) + +#define ENA_MAX_HANDLERS 256 + +#define ENA_MAX_PHYS_ADDR_SIZE_BITS 48 + +/* Unit in usec */ +#define ENA_REG_READ_TIMEOUT 200000 + +#define ADMIN_SQ_SIZE(depth) ((depth) * sizeof(struct ena_admin_aq_entry)) +#define ADMIN_CQ_SIZE(depth) ((depth) * sizeof(struct ena_admin_acq_entry)) +#define ADMIN_AENQ_SIZE(depth) ((depth) * sizeof(struct ena_admin_aenq_entry)) + +/*****************************************************************************/ +/*****************************************************************************/ +/* ENA adaptive interrupt moderation settings */ + +#define ENA_INTR_INITIAL_TX_INTERVAL_USECS 64 +#define ENA_INTR_INITIAL_RX_INTERVAL_USECS 0 +#define ENA_DEFAULT_INTR_DELAY_RESOLUTION 1 + +#define ENA_HASH_KEY_SIZE 40 + +#define ENA_HW_HINTS_NO_TIMEOUT 0xFFFF + +#define ENA_FEATURE_MAX_QUEUE_EXT_VER 1 + +struct ena_llq_configurations { + enum ena_admin_llq_header_location llq_header_location; + enum ena_admin_llq_ring_entry_size llq_ring_entry_size; + enum ena_admin_llq_stride_ctrl llq_stride_ctrl; + enum ena_admin_llq_num_descs_before_header llq_num_decs_before_header; + u16 llq_ring_entry_size_value; +}; + +enum queue_direction { + ENA_COM_IO_QUEUE_DIRECTION_TX, + ENA_COM_IO_QUEUE_DIRECTION_RX +}; + +struct ena_com_buf { + dma_addr_t paddr; /**< Buffer physical address */ + u16 len; /**< Buffer length in bytes */ +}; + +struct ena_com_rx_buf_info { + u16 len; + u16 req_id; +}; + +struct ena_com_io_desc_addr { + u8 __iomem *pbuf_dev_addr; /* LLQ address */ + u8 *virt_addr; + dma_addr_t phys_addr; +}; + +struct ena_com_tx_meta { + u16 mss; + u16 l3_hdr_len; + u16 l3_hdr_offset; + u16 l4_hdr_len; /* In words */ +}; + +struct ena_com_llq_info { + u16 header_location_ctrl; + u16 desc_stride_ctrl; + u16 desc_list_entry_size_ctrl; + u16 desc_list_entry_size; + u16 descs_num_before_header; + u16 descs_per_entry; + u16 max_entries_in_tx_burst; + bool disable_meta_caching; +}; + +struct ena_com_io_cq { + struct ena_com_io_desc_addr cdesc_addr; + + /* Interrupt unmask register */ + u32 __iomem *unmask_reg; + + /* The completion queue head doorbell register */ + u32 __iomem *cq_head_db_reg; + + /* numa configuration register (for TPH) */ + u32 __iomem *numa_node_cfg_reg; + + /* The value to write to the above register to unmask + * the interrupt of this queue + */ + u32 msix_vector; + + enum queue_direction direction; + + /* holds the number of cdesc of the current packet */ + u16 cur_rx_pkt_cdesc_count; + /* save the first cdesc idx of the current packet */ + u16 cur_rx_pkt_cdesc_start_idx; + + u16 q_depth; + /* Caller qid */ + u16 qid; + + /* Device queue index */ + u16 idx; + u16 head; + u16 last_head_update; + u8 phase; + u8 cdesc_entry_size_in_bytes; + +} ____cacheline_aligned; + +struct ena_com_io_bounce_buffer_control { + u8 *base_buffer; + u16 next_to_use; + u16 buffer_size; + u16 buffers_num; /* Must be a power of 2 */ +}; + +/* This struct is to keep tracking the current location of the next llq entry */ +struct ena_com_llq_pkt_ctrl { + u8 *curr_bounce_buf; + u16 idx; + u16 descs_left_in_line; +}; + +struct ena_com_io_sq { + struct ena_com_io_desc_addr desc_addr; + + u32 __iomem *db_addr; + u8 __iomem *header_addr; + + enum queue_direction direction; + enum ena_admin_placement_policy_type mem_queue_type; + + bool disable_meta_caching; + + u32 msix_vector; + struct ena_com_tx_meta cached_tx_meta; + struct ena_com_llq_info llq_info; + struct ena_com_llq_pkt_ctrl llq_buf_ctrl; + struct ena_com_io_bounce_buffer_control bounce_buf_ctrl; + + u16 q_depth; + u16 qid; + + u16 idx; + u16 tail; + u16 next_to_comp; + u16 llq_last_copy_tail; + u32 tx_max_header_size; + u8 phase; + u8 desc_entry_size; + u8 dma_addr_bits; + u16 entries_in_tx_burst_left; +} ____cacheline_aligned; + +struct ena_com_admin_cq { + struct ena_admin_acq_entry *entries; + dma_addr_t dma_addr; + + u16 head; + u8 phase; +}; + +struct ena_com_admin_sq { + struct ena_admin_aq_entry *entries; + dma_addr_t dma_addr; + + u32 __iomem *db_addr; + + u16 head; + u16 tail; + u8 phase; + +}; + +struct ena_com_stats_admin { + u64 aborted_cmd; + u64 submitted_cmd; + u64 completed_cmd; + u64 out_of_space; + u64 no_completion; +}; + +struct ena_com_admin_queue { + void *q_dmadev; + struct ena_com_dev *ena_dev; + spinlock_t q_lock; /* spinlock for the admin queue */ + + struct ena_comp_ctx *comp_ctx; + u32 completion_timeout; + u16 q_depth; + struct ena_com_admin_cq cq; + struct ena_com_admin_sq sq; + + /* Indicate if the admin queue should poll for completion */ + bool polling; + + /* Define if fallback to polling mode should occur */ + bool auto_polling; + + u16 curr_cmd_id; + + /* Indicate that the ena was initialized and can + * process new admin commands + */ + bool running_state; + + /* Count the number of outstanding admin commands */ + atomic_t outstanding_cmds; + + struct ena_com_stats_admin stats; +}; + +struct ena_aenq_handlers; + +struct ena_com_aenq { + u16 head; + u8 phase; + struct ena_admin_aenq_entry *entries; + dma_addr_t dma_addr; + u16 q_depth; + struct ena_aenq_handlers *aenq_handlers; +}; + +struct ena_com_mmio_read { + struct ena_admin_ena_mmio_req_read_less_resp *read_resp; + dma_addr_t read_resp_dma_addr; + u32 reg_read_to; /* in us */ + u16 seq_num; + bool readless_supported; + /* spin lock to ensure a single outstanding read */ + spinlock_t lock; +}; + +struct ena_rss { + /* Indirect table */ + u16 *host_rss_ind_tbl; + struct ena_admin_rss_ind_table_entry *rss_ind_tbl; + dma_addr_t rss_ind_tbl_dma_addr; + u16 tbl_log_size; + + /* Hash key */ + enum ena_admin_hash_functions hash_func; + struct ena_admin_feature_rss_flow_hash_control *hash_key; + dma_addr_t hash_key_dma_addr; + u32 hash_init_val; + + /* Flow Control */ + struct ena_admin_feature_rss_hash_control *hash_ctrl; + dma_addr_t hash_ctrl_dma_addr; + +}; + +struct ena_host_attribute { + /* Debug area */ + u8 *debug_area_virt_addr; + dma_addr_t debug_area_dma_addr; + u32 debug_area_size; + + /* Host information */ + struct ena_admin_host_info *host_info; + dma_addr_t host_info_dma_addr; +}; + +/* Each ena_dev is a PCI function. */ +struct ena_com_dev { + struct ena_com_admin_queue admin_queue; + struct ena_com_aenq aenq; + struct ena_com_io_cq io_cq_queues[ENA_TOTAL_NUM_QUEUES]; + struct ena_com_io_sq io_sq_queues[ENA_TOTAL_NUM_QUEUES]; + u8 __iomem *reg_bar; + void __iomem *mem_bar; + void *dmadev; + struct net_device *net_device; + + enum ena_admin_placement_policy_type tx_mem_queue_type; + u32 tx_max_header_size; + u16 stats_func; /* Selected function for extended statistic dump */ + u16 stats_queue; /* Selected queue for extended statistic dump */ + + struct ena_com_mmio_read mmio_read; + + struct ena_rss rss; + u32 supported_features; + u32 capabilities; + u32 dma_addr_bits; + + struct ena_host_attribute host_attr; + bool adaptive_coalescing; + u16 intr_delay_resolution; + + /* interrupt moderation intervals are in usec divided by + * intr_delay_resolution, which is supplied by the device. + */ + u32 intr_moder_tx_interval; + u32 intr_moder_rx_interval; + + struct ena_intr_moder_entry *intr_moder_tbl; + + struct ena_com_llq_info llq_info; + + u32 ena_min_poll_delay_us; +}; + +struct ena_com_dev_get_features_ctx { + struct ena_admin_queue_feature_desc max_queues; + struct ena_admin_queue_ext_feature_desc max_queue_ext; + struct ena_admin_device_attr_feature_desc dev_attr; + struct ena_admin_feature_aenq_desc aenq; + struct ena_admin_feature_offload_desc offload; + struct ena_admin_ena_hw_hints hw_hints; + struct ena_admin_feature_llq_desc llq; +}; + +struct ena_com_create_io_ctx { + enum ena_admin_placement_policy_type mem_queue_type; + enum queue_direction direction; + int numa_node; + u32 msix_vector; + u16 queue_size; + u16 qid; +}; + +typedef void (*ena_aenq_handler)(void *data, + struct ena_admin_aenq_entry *aenq_e); + +/* Holds aenq handlers. Indexed by AENQ event group */ +struct ena_aenq_handlers { + ena_aenq_handler handlers[ENA_MAX_HANDLERS]; + ena_aenq_handler unimplemented_handler; +}; + +/*****************************************************************************/ +/*****************************************************************************/ + +/* ena_com_mmio_reg_read_request_init - Init the mmio reg read mechanism + * @ena_dev: ENA communication layer struct + * + * Initialize the register read mechanism. + * + * @note: This method must be the first stage in the initialization sequence. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_mmio_reg_read_request_init(struct ena_com_dev *ena_dev); + +/* ena_com_set_mmio_read_mode - Enable/disable the indirect mmio reg read mechanism + * @ena_dev: ENA communication layer struct + * @readless_supported: readless mode (enable/disable) + */ +void ena_com_set_mmio_read_mode(struct ena_com_dev *ena_dev, + bool readless_supported); + +/* ena_com_mmio_reg_read_request_write_dev_addr - Write the mmio reg read return + * value physical address. + * @ena_dev: ENA communication layer struct + */ +void ena_com_mmio_reg_read_request_write_dev_addr(struct ena_com_dev *ena_dev); + +/* ena_com_mmio_reg_read_request_destroy - Destroy the mmio reg read mechanism + * @ena_dev: ENA communication layer struct + */ +void ena_com_mmio_reg_read_request_destroy(struct ena_com_dev *ena_dev); + +/* ena_com_admin_init - Init the admin and the async queues + * @ena_dev: ENA communication layer struct + * @aenq_handlers: Those handlers to be called upon event. + * + * Initialize the admin submission and completion queues. + * Initialize the asynchronous events notification queues. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_admin_init(struct ena_com_dev *ena_dev, + struct ena_aenq_handlers *aenq_handlers); + +/* ena_com_admin_destroy - Destroy the admin and the async events queues. + * @ena_dev: ENA communication layer struct + * + * @note: Before calling this method, the caller must validate that the device + * won't send any additional admin completions/aenq. + * To achieve that, a FLR is recommended. + */ +void ena_com_admin_destroy(struct ena_com_dev *ena_dev); + +/* ena_com_dev_reset - Perform device FLR to the device. + * @ena_dev: ENA communication layer struct + * @reset_reason: Specify what is the trigger for the reset in case of an error. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_dev_reset(struct ena_com_dev *ena_dev, + enum ena_regs_reset_reason_types reset_reason); + +/* ena_com_create_io_queue - Create io queue. + * @ena_dev: ENA communication layer struct + * @ctx - create context structure + * + * Create the submission and the completion queues. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_create_io_queue(struct ena_com_dev *ena_dev, + struct ena_com_create_io_ctx *ctx); + +/* ena_com_destroy_io_queue - Destroy IO queue with the queue id - qid. + * @ena_dev: ENA communication layer struct + * @qid - the caller virtual queue id. + */ +void ena_com_destroy_io_queue(struct ena_com_dev *ena_dev, u16 qid); + +/* ena_com_get_io_handlers - Return the io queue handlers + * @ena_dev: ENA communication layer struct + * @qid - the caller virtual queue id. + * @io_sq - IO submission queue handler + * @io_cq - IO completion queue handler. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_get_io_handlers(struct ena_com_dev *ena_dev, u16 qid, + struct ena_com_io_sq **io_sq, + struct ena_com_io_cq **io_cq); + +/* ena_com_admin_aenq_enable - ENAble asynchronous event notifications + * @ena_dev: ENA communication layer struct + * + * After this method, aenq event can be received via AENQ. + */ +void ena_com_admin_aenq_enable(struct ena_com_dev *ena_dev); + +/* ena_com_set_admin_running_state - Set the state of the admin queue + * @ena_dev: ENA communication layer struct + * + * Change the state of the admin queue (enable/disable) + */ +void ena_com_set_admin_running_state(struct ena_com_dev *ena_dev, bool state); + +/* ena_com_get_admin_running_state - Get the admin queue state + * @ena_dev: ENA communication layer struct + * + * Retrieve the state of the admin queue (enable/disable) + * + * @return - current polling mode (enable/disable) + */ +bool ena_com_get_admin_running_state(struct ena_com_dev *ena_dev); + +/* ena_com_set_admin_polling_mode - Set the admin completion queue polling mode + * @ena_dev: ENA communication layer struct + * @polling: ENAble/Disable polling mode + * + * Set the admin completion mode. + */ +void ena_com_set_admin_polling_mode(struct ena_com_dev *ena_dev, bool polling); + +/* ena_com_set_admin_auto_polling_mode - Enable autoswitch to polling mode + * @ena_dev: ENA communication layer struct + * @polling: Enable/Disable polling mode + * + * Set the autopolling mode. + * If autopolling is on: + * In case of missing interrupt when data is available switch to polling. + */ +void ena_com_set_admin_auto_polling_mode(struct ena_com_dev *ena_dev, + bool polling); + +/* ena_com_admin_q_comp_intr_handler - admin queue interrupt handler + * @ena_dev: ENA communication layer struct + * + * This method goes over the admin completion queue and wakes up all the pending + * threads that wait on the commands wait event. + * + * @note: Should be called after MSI-X interrupt. + */ +void ena_com_admin_q_comp_intr_handler(struct ena_com_dev *ena_dev); + +/* ena_com_aenq_intr_handler - AENQ interrupt handler + * @ena_dev: ENA communication layer struct + * + * This method goes over the async event notification queue and calls the proper + * aenq handler. + */ +void ena_com_aenq_intr_handler(struct ena_com_dev *ena_dev, void *data); + +/* ena_com_abort_admin_commands - Abort all the outstanding admin commands. + * @ena_dev: ENA communication layer struct + * + * This method aborts all the outstanding admin commands. + * The caller should then call ena_com_wait_for_abort_completion to make sure + * all the commands were completed. + */ +void ena_com_abort_admin_commands(struct ena_com_dev *ena_dev); + +/* ena_com_wait_for_abort_completion - Wait for admin commands abort. + * @ena_dev: ENA communication layer struct + * + * This method waits until all the outstanding admin commands are completed. + */ +void ena_com_wait_for_abort_completion(struct ena_com_dev *ena_dev); + +/* ena_com_validate_version - Validate the device parameters + * @ena_dev: ENA communication layer struct + * + * This method verifies the device parameters are the same as the saved + * parameters in ena_dev. + * This method is useful after device reset, to validate the device mac address + * and the device offloads are the same as before the reset. + * + * @return - 0 on success negative value otherwise. + */ +int ena_com_validate_version(struct ena_com_dev *ena_dev); + +/* ena_com_get_link_params - Retrieve physical link parameters. + * @ena_dev: ENA communication layer struct + * @resp: Link parameters + * + * Retrieve the physical link parameters, + * like speed, auto-negotiation and full duplex support. + * + * @return - 0 on Success negative value otherwise. + */ +int ena_com_get_link_params(struct ena_com_dev *ena_dev, + struct ena_admin_get_feat_resp *resp); + +/* ena_com_get_dma_width - Retrieve physical dma address width the device + * supports. + * @ena_dev: ENA communication layer struct + * + * Retrieve the maximum physical address bits the device can handle. + * + * @return: > 0 on Success and negative value otherwise. + */ +int ena_com_get_dma_width(struct ena_com_dev *ena_dev); + +/* ena_com_set_aenq_config - Set aenq groups configurations + * @ena_dev: ENA communication layer struct + * @groups flag: bit fields flags of enum ena_admin_aenq_group. + * + * Configure which aenq event group the driver would like to receive. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_aenq_config(struct ena_com_dev *ena_dev, u32 groups_flag); + +/* ena_com_get_dev_attr_feat - Get device features + * @ena_dev: ENA communication layer struct + * @get_feat_ctx: returned context that contain the get features. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_dev_attr_feat(struct ena_com_dev *ena_dev, + struct ena_com_dev_get_features_ctx *get_feat_ctx); + +/* ena_com_get_dev_basic_stats - Get device basic statistics + * @ena_dev: ENA communication layer struct + * @stats: stats return value + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_dev_basic_stats(struct ena_com_dev *ena_dev, + struct ena_admin_basic_stats *stats); + +/* ena_com_get_eni_stats - Get extended network interface statistics + * @ena_dev: ENA communication layer struct + * @stats: stats return value + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_eni_stats(struct ena_com_dev *ena_dev, + struct ena_admin_eni_stats *stats); + +/* ena_com_set_dev_mtu - Configure the device mtu. + * @ena_dev: ENA communication layer struct + * @mtu: mtu value + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_dev_mtu(struct ena_com_dev *ena_dev, u32 mtu); + +/* ena_com_get_offload_settings - Retrieve the device offloads capabilities + * @ena_dev: ENA communication layer struct + * @offlad: offload return value + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_offload_settings(struct ena_com_dev *ena_dev, + struct ena_admin_feature_offload_desc *offload); + +/* ena_com_rss_init - Init RSS + * @ena_dev: ENA communication layer struct + * @log_size: indirection log size + * + * Allocate RSS/RFS resources. + * The caller then can configure rss using ena_com_set_hash_function, + * ena_com_set_hash_ctrl and ena_com_indirect_table_set. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_rss_init(struct ena_com_dev *ena_dev, u16 log_size); + +/* ena_com_rss_destroy - Destroy rss + * @ena_dev: ENA communication layer struct + * + * Free all the RSS/RFS resources. + */ +void ena_com_rss_destroy(struct ena_com_dev *ena_dev); + +/* ena_com_get_current_hash_function - Get RSS hash function + * @ena_dev: ENA communication layer struct + * + * Return the current hash function. + * @return: 0 or one of the ena_admin_hash_functions values. + */ +int ena_com_get_current_hash_function(struct ena_com_dev *ena_dev); + +/* ena_com_fill_hash_function - Fill RSS hash function + * @ena_dev: ENA communication layer struct + * @func: The hash function (Toeplitz or crc) + * @key: Hash key (for toeplitz hash) + * @key_len: key length (max length 10 DW) + * @init_val: initial value for the hash function + * + * Fill the ena_dev resources with the desire hash function, hash key, key_len + * and key initial value (if needed by the hash function). + * To flush the key into the device the caller should call + * ena_com_set_hash_function. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_fill_hash_function(struct ena_com_dev *ena_dev, + enum ena_admin_hash_functions func, + const u8 *key, u16 key_len, u32 init_val); + +/* ena_com_set_hash_function - Flush the hash function and it dependencies to + * the device. + * @ena_dev: ENA communication layer struct + * + * Flush the hash function and it dependencies (key, key length and + * initial value) if needed. + * + * @note: Prior to this method the caller should call ena_com_fill_hash_function + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_hash_function(struct ena_com_dev *ena_dev); + +/* ena_com_get_hash_function - Retrieve the hash function from the device. + * @ena_dev: ENA communication layer struct + * @func: hash function + * + * Retrieve the hash function from the device. + * + * @note: If the caller called ena_com_fill_hash_function but didn't flush + * it to the device, the new configuration will be lost. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_hash_function(struct ena_com_dev *ena_dev, + enum ena_admin_hash_functions *func); + +/* ena_com_get_hash_key - Retrieve the hash key + * @ena_dev: ENA communication layer struct + * @key: hash key + * + * Retrieve the hash key. + * + * @note: If the caller called ena_com_fill_hash_key but didn't flush + * it to the device, the new configuration will be lost. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_hash_key(struct ena_com_dev *ena_dev, u8 *key); +/* ena_com_fill_hash_ctrl - Fill RSS hash control + * @ena_dev: ENA communication layer struct. + * @proto: The protocol to configure. + * @hash_fields: bit mask of ena_admin_flow_hash_fields + * + * Fill the ena_dev resources with the desire hash control (the ethernet + * fields that take part of the hash) for a specific protocol. + * To flush the hash control to the device, the caller should call + * ena_com_set_hash_ctrl. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_fill_hash_ctrl(struct ena_com_dev *ena_dev, + enum ena_admin_flow_hash_proto proto, + u16 hash_fields); + +/* ena_com_set_hash_ctrl - Flush the hash control resources to the device. + * @ena_dev: ENA communication layer struct + * + * Flush the hash control (the ethernet fields that take part of the hash) + * + * @note: Prior to this method the caller should call ena_com_fill_hash_ctrl. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_hash_ctrl(struct ena_com_dev *ena_dev); + +/* ena_com_get_hash_ctrl - Retrieve the hash control from the device. + * @ena_dev: ENA communication layer struct + * @proto: The protocol to retrieve. + * @fields: bit mask of ena_admin_flow_hash_fields. + * + * Retrieve the hash control from the device. + * + * @note: If the caller called ena_com_fill_hash_ctrl but didn't flush + * it to the device, the new configuration will be lost. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_get_hash_ctrl(struct ena_com_dev *ena_dev, + enum ena_admin_flow_hash_proto proto, + u16 *fields); + +/* ena_com_set_default_hash_ctrl - Set the hash control to a default + * configuration. + * @ena_dev: ENA communication layer struct + * + * Fill the ena_dev resources with the default hash control configuration. + * To flush the hash control to the device, the caller should call + * ena_com_set_hash_ctrl. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_default_hash_ctrl(struct ena_com_dev *ena_dev); + +/* ena_com_indirect_table_fill_entry - Fill a single entry in the RSS + * indirection table + * @ena_dev: ENA communication layer struct. + * @entry_idx - indirection table entry. + * @entry_value - redirection value + * + * Fill a single entry of the RSS indirection table in the ena_dev resources. + * To flush the indirection table to the device, the called should call + * ena_com_indirect_table_set. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_indirect_table_fill_entry(struct ena_com_dev *ena_dev, + u16 entry_idx, u16 entry_value); + +/* ena_com_indirect_table_set - Flush the indirection table to the device. + * @ena_dev: ENA communication layer struct + * + * Flush the indirection hash control to the device. + * Prior to this method the caller should call ena_com_indirect_table_fill_entry + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_indirect_table_set(struct ena_com_dev *ena_dev); + +/* ena_com_indirect_table_get - Retrieve the indirection table from the device. + * @ena_dev: ENA communication layer struct + * @ind_tbl: indirection table + * + * Retrieve the RSS indirection table from the device. + * + * @note: If the caller called ena_com_indirect_table_fill_entry but didn't flush + * it to the device, the new configuration will be lost. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_indirect_table_get(struct ena_com_dev *ena_dev, u32 *ind_tbl); + +/* ena_com_allocate_host_info - Allocate host info resources. + * @ena_dev: ENA communication layer struct + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_allocate_host_info(struct ena_com_dev *ena_dev); + +/* ena_com_allocate_debug_area - Allocate debug area. + * @ena_dev: ENA communication layer struct + * @debug_area_size - debug area size. + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_allocate_debug_area(struct ena_com_dev *ena_dev, + u32 debug_area_size); + +/* ena_com_delete_debug_area - Free the debug area resources. + * @ena_dev: ENA communication layer struct + * + * Free the allocated debug area. + */ +void ena_com_delete_debug_area(struct ena_com_dev *ena_dev); + +/* ena_com_delete_host_info - Free the host info resources. + * @ena_dev: ENA communication layer struct + * + * Free the allocated host info. + */ +void ena_com_delete_host_info(struct ena_com_dev *ena_dev); + +/* ena_com_set_host_attributes - Update the device with the host + * attributes (debug area and host info) base address. + * @ena_dev: ENA communication layer struct + * + * @return: 0 on Success and negative value otherwise. + */ +int ena_com_set_host_attributes(struct ena_com_dev *ena_dev); + +/* ena_com_create_io_cq - Create io completion queue. + * @ena_dev: ENA communication layer struct + * @io_cq - io completion queue handler + + * Create IO completion queue. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_create_io_cq(struct ena_com_dev *ena_dev, + struct ena_com_io_cq *io_cq); + +/* ena_com_destroy_io_cq - Destroy io completion queue. + * @ena_dev: ENA communication layer struct + * @io_cq - io completion queue handler + + * Destroy IO completion queue. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_destroy_io_cq(struct ena_com_dev *ena_dev, + struct ena_com_io_cq *io_cq); + +/* ena_com_execute_admin_command - Execute admin command + * @admin_queue: admin queue. + * @cmd: the admin command to execute. + * @cmd_size: the command size. + * @cmd_completion: command completion return value. + * @cmd_comp_size: command completion size. + + * Submit an admin command and then wait until the device returns a + * completion. + * The completion will be copied into cmd_comp. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_execute_admin_command(struct ena_com_admin_queue *admin_queue, + struct ena_admin_aq_entry *cmd, + size_t cmd_size, + struct ena_admin_acq_entry *cmd_comp, + size_t cmd_comp_size); + +/* ena_com_init_interrupt_moderation - Init interrupt moderation + * @ena_dev: ENA communication layer struct + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_init_interrupt_moderation(struct ena_com_dev *ena_dev); + +/* ena_com_interrupt_moderation_supported - Return if interrupt moderation + * capability is supported by the device. + * + * @return - supported or not. + */ +bool ena_com_interrupt_moderation_supported(struct ena_com_dev *ena_dev); + +/* ena_com_update_nonadaptive_moderation_interval_tx - Update the + * non-adaptive interval in Tx direction. + * @ena_dev: ENA communication layer struct + * @tx_coalesce_usecs: Interval in usec. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_update_nonadaptive_moderation_interval_tx(struct ena_com_dev *ena_dev, + u32 tx_coalesce_usecs); + +/* ena_com_update_nonadaptive_moderation_interval_rx - Update the + * non-adaptive interval in Rx direction. + * @ena_dev: ENA communication layer struct + * @rx_coalesce_usecs: Interval in usec. + * + * @return - 0 on success, negative value on failure. + */ +int ena_com_update_nonadaptive_moderation_interval_rx(struct ena_com_dev *ena_dev, + u32 rx_coalesce_usecs); + +/* ena_com_get_nonadaptive_moderation_interval_tx - Retrieve the + * non-adaptive interval in Tx direction. + * @ena_dev: ENA communication layer struct + * + * @return - interval in usec + */ +unsigned int ena_com_get_nonadaptive_moderation_interval_tx(struct ena_com_dev *ena_dev); + +/* ena_com_get_nonadaptive_moderation_interval_rx - Retrieve the + * non-adaptive interval in Rx direction. + * @ena_dev: ENA communication layer struct + * + * @return - interval in usec + */ +unsigned int ena_com_get_nonadaptive_moderation_interval_rx(struct ena_com_dev *ena_dev); + +/* ena_com_config_dev_mode - Configure the placement policy of the device. + * @ena_dev: ENA communication layer struct + * @llq_features: LLQ feature descriptor, retrieve via + * ena_com_get_dev_attr_feat. + * @ena_llq_config: The default driver LLQ parameters configurations + */ +int ena_com_config_dev_mode(struct ena_com_dev *ena_dev, + struct ena_admin_feature_llq_desc *llq_features, + struct ena_llq_configurations *llq_default_config); + +/* ena_com_io_sq_to_ena_dev - Extract ena_com_dev using contained field io_sq. + * @io_sq: IO submit queue struct + * + * @return - ena_com_dev struct extracted from io_sq + */ +static inline struct ena_com_dev *ena_com_io_sq_to_ena_dev(struct ena_com_io_sq *io_sq) +{ + return container_of(io_sq, struct ena_com_dev, io_sq_queues[io_sq->qid]); +} + +/* ena_com_io_cq_to_ena_dev - Extract ena_com_dev using contained field io_cq. + * @io_sq: IO submit queue struct + * + * @return - ena_com_dev struct extracted from io_sq + */ +static inline struct ena_com_dev *ena_com_io_cq_to_ena_dev(struct ena_com_io_cq *io_cq) +{ + return container_of(io_cq, struct ena_com_dev, io_cq_queues[io_cq->qid]); +} + +static inline bool ena_com_get_adaptive_moderation_enabled(struct ena_com_dev *ena_dev) +{ + return ena_dev->adaptive_coalescing; +} + +static inline void ena_com_enable_adaptive_moderation(struct ena_com_dev *ena_dev) +{ + ena_dev->adaptive_coalescing = true; +} + +static inline void ena_com_disable_adaptive_moderation(struct ena_com_dev *ena_dev) +{ + ena_dev->adaptive_coalescing = false; +} + +/* ena_com_get_cap - query whether device supports a capability. + * @ena_dev: ENA communication layer struct + * @cap_id: enum value representing the capability + * + * @return - true if capability is supported or false otherwise + */ +static inline bool ena_com_get_cap(struct ena_com_dev *ena_dev, + enum ena_admin_aq_caps_id cap_id) +{ + return !!(ena_dev->capabilities & BIT(cap_id)); +} + +/* ena_com_update_intr_reg - Prepare interrupt register + * @intr_reg: interrupt register to update. + * @rx_delay_interval: Rx interval in usecs + * @tx_delay_interval: Tx interval in usecs + * @unmask: unmask enable/disable + * + * Prepare interrupt update register with the supplied parameters. + */ +static inline void ena_com_update_intr_reg(struct ena_eth_io_intr_reg *intr_reg, + u32 rx_delay_interval, + u32 tx_delay_interval, + bool unmask) +{ + intr_reg->intr_control = 0; + intr_reg->intr_control |= rx_delay_interval & + ENA_ETH_IO_INTR_REG_RX_INTR_DELAY_MASK; + + intr_reg->intr_control |= + (tx_delay_interval << ENA_ETH_IO_INTR_REG_TX_INTR_DELAY_SHIFT) + & ENA_ETH_IO_INTR_REG_TX_INTR_DELAY_MASK; + + if (unmask) + intr_reg->intr_control |= ENA_ETH_IO_INTR_REG_INTR_UNMASK_MASK; +} + +static inline u8 *ena_com_get_next_bounce_buffer(struct ena_com_io_bounce_buffer_control *bounce_buf_ctrl) +{ + u16 size, buffers_num; + u8 *buf; + + size = bounce_buf_ctrl->buffer_size; + buffers_num = bounce_buf_ctrl->buffers_num; + + buf = bounce_buf_ctrl->base_buffer + + (bounce_buf_ctrl->next_to_use++ & (buffers_num - 1)) * size; + + prefetchw(bounce_buf_ctrl->base_buffer + + (bounce_buf_ctrl->next_to_use & (buffers_num - 1)) * size); + + return buf; +} + +#endif /* !(ENA_COM) */ diff --git a/drivers/net/ethernet/amazon/ena/ena_common_defs.h b/drivers/net/ethernet/amazon/ena/ena_common_defs.h new file mode 100644 index 0000000000..e210c8a81f --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_common_defs.h @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ +#ifndef _ENA_COMMON_H_ +#define _ENA_COMMON_H_ + +#define ENA_COMMON_SPEC_VERSION_MAJOR 2 +#define ENA_COMMON_SPEC_VERSION_MINOR 0 + +/* ENA operates with 48-bit memory addresses. ena_mem_addr_t */ +struct ena_common_mem_addr { + u32 mem_addr_low; + + u16 mem_addr_high; + + /* MBZ */ + u16 reserved16; +}; + +#endif /* _ENA_COMMON_H_ */ diff --git a/drivers/net/ethernet/amazon/ena/ena_eth_com.c b/drivers/net/ethernet/amazon/ena/ena_eth_com.c new file mode 100644 index 0000000000..f9f886289b --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_eth_com.c @@ -0,0 +1,646 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#include "ena_eth_com.h" + +static struct ena_eth_io_rx_cdesc_base *ena_com_get_next_rx_cdesc( + struct ena_com_io_cq *io_cq) +{ + struct ena_eth_io_rx_cdesc_base *cdesc; + u16 expected_phase, head_masked; + u16 desc_phase; + + head_masked = io_cq->head & (io_cq->q_depth - 1); + expected_phase = io_cq->phase; + + cdesc = (struct ena_eth_io_rx_cdesc_base *)(io_cq->cdesc_addr.virt_addr + + (head_masked * io_cq->cdesc_entry_size_in_bytes)); + + desc_phase = (READ_ONCE(cdesc->status) & + ENA_ETH_IO_RX_CDESC_BASE_PHASE_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_PHASE_SHIFT; + + if (desc_phase != expected_phase) + return NULL; + + /* Make sure we read the rest of the descriptor after the phase bit + * has been read + */ + dma_rmb(); + + return cdesc; +} + +static void *get_sq_desc_regular_queue(struct ena_com_io_sq *io_sq) +{ + u16 tail_masked; + u32 offset; + + tail_masked = io_sq->tail & (io_sq->q_depth - 1); + + offset = tail_masked * io_sq->desc_entry_size; + + return (void *)((uintptr_t)io_sq->desc_addr.virt_addr + offset); +} + +static int ena_com_write_bounce_buffer_to_dev(struct ena_com_io_sq *io_sq, + u8 *bounce_buffer) +{ + struct ena_com_llq_info *llq_info = &io_sq->llq_info; + + u16 dst_tail_mask; + u32 dst_offset; + + dst_tail_mask = io_sq->tail & (io_sq->q_depth - 1); + dst_offset = dst_tail_mask * llq_info->desc_list_entry_size; + + if (is_llq_max_tx_burst_exists(io_sq)) { + if (unlikely(!io_sq->entries_in_tx_burst_left)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Error: trying to send more packets than tx burst allows\n"); + return -ENOSPC; + } + + io_sq->entries_in_tx_burst_left--; + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Decreasing entries_in_tx_burst_left of queue %d to %d\n", + io_sq->qid, io_sq->entries_in_tx_burst_left); + } + + /* Make sure everything was written into the bounce buffer before + * writing the bounce buffer to the device + */ + wmb(); + + /* The line is completed. Copy it to dev */ + __iowrite64_copy(io_sq->desc_addr.pbuf_dev_addr + dst_offset, + bounce_buffer, (llq_info->desc_list_entry_size) / 8); + + io_sq->tail++; + + /* Switch phase bit in case of wrap around */ + if (unlikely((io_sq->tail & (io_sq->q_depth - 1)) == 0)) + io_sq->phase ^= 1; + + return 0; +} + +static int ena_com_write_header_to_bounce(struct ena_com_io_sq *io_sq, + u8 *header_src, + u16 header_len) +{ + struct ena_com_llq_pkt_ctrl *pkt_ctrl = &io_sq->llq_buf_ctrl; + struct ena_com_llq_info *llq_info = &io_sq->llq_info; + u8 *bounce_buffer = pkt_ctrl->curr_bounce_buf; + u16 header_offset; + + if (unlikely(io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_HOST)) + return 0; + + header_offset = + llq_info->descs_num_before_header * io_sq->desc_entry_size; + + if (unlikely((header_offset + header_len) > + llq_info->desc_list_entry_size)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Trying to write header larger than llq entry can accommodate\n"); + return -EFAULT; + } + + if (unlikely(!bounce_buffer)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Bounce buffer is NULL\n"); + return -EFAULT; + } + + memcpy(bounce_buffer + header_offset, header_src, header_len); + + return 0; +} + +static void *get_sq_desc_llq(struct ena_com_io_sq *io_sq) +{ + struct ena_com_llq_pkt_ctrl *pkt_ctrl = &io_sq->llq_buf_ctrl; + u8 *bounce_buffer; + void *sq_desc; + + bounce_buffer = pkt_ctrl->curr_bounce_buf; + + if (unlikely(!bounce_buffer)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Bounce buffer is NULL\n"); + return NULL; + } + + sq_desc = bounce_buffer + pkt_ctrl->idx * io_sq->desc_entry_size; + pkt_ctrl->idx++; + pkt_ctrl->descs_left_in_line--; + + return sq_desc; +} + +static int ena_com_close_bounce_buffer(struct ena_com_io_sq *io_sq) +{ + struct ena_com_llq_pkt_ctrl *pkt_ctrl = &io_sq->llq_buf_ctrl; + struct ena_com_llq_info *llq_info = &io_sq->llq_info; + int rc; + + if (unlikely(io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_HOST)) + return 0; + + /* bounce buffer was used, so write it and get a new one */ + if (likely(pkt_ctrl->idx)) { + rc = ena_com_write_bounce_buffer_to_dev(io_sq, + pkt_ctrl->curr_bounce_buf); + if (unlikely(rc)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to write bounce buffer to device\n"); + return rc; + } + + pkt_ctrl->curr_bounce_buf = + ena_com_get_next_bounce_buffer(&io_sq->bounce_buf_ctrl); + memset(io_sq->llq_buf_ctrl.curr_bounce_buf, + 0x0, llq_info->desc_list_entry_size); + } + + pkt_ctrl->idx = 0; + pkt_ctrl->descs_left_in_line = llq_info->descs_num_before_header; + return 0; +} + +static void *get_sq_desc(struct ena_com_io_sq *io_sq) +{ + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + return get_sq_desc_llq(io_sq); + + return get_sq_desc_regular_queue(io_sq); +} + +static int ena_com_sq_update_llq_tail(struct ena_com_io_sq *io_sq) +{ + struct ena_com_llq_pkt_ctrl *pkt_ctrl = &io_sq->llq_buf_ctrl; + struct ena_com_llq_info *llq_info = &io_sq->llq_info; + int rc; + + if (!pkt_ctrl->descs_left_in_line) { + rc = ena_com_write_bounce_buffer_to_dev(io_sq, + pkt_ctrl->curr_bounce_buf); + if (unlikely(rc)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to write bounce buffer to device\n"); + return rc; + } + + pkt_ctrl->curr_bounce_buf = + ena_com_get_next_bounce_buffer(&io_sq->bounce_buf_ctrl); + memset(io_sq->llq_buf_ctrl.curr_bounce_buf, + 0x0, llq_info->desc_list_entry_size); + + pkt_ctrl->idx = 0; + if (unlikely(llq_info->desc_stride_ctrl == ENA_ADMIN_SINGLE_DESC_PER_ENTRY)) + pkt_ctrl->descs_left_in_line = 1; + else + pkt_ctrl->descs_left_in_line = + llq_info->desc_list_entry_size / io_sq->desc_entry_size; + } + + return 0; +} + +static int ena_com_sq_update_tail(struct ena_com_io_sq *io_sq) +{ + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + return ena_com_sq_update_llq_tail(io_sq); + + io_sq->tail++; + + /* Switch phase bit in case of wrap around */ + if (unlikely((io_sq->tail & (io_sq->q_depth - 1)) == 0)) + io_sq->phase ^= 1; + + return 0; +} + +static struct ena_eth_io_rx_cdesc_base * + ena_com_rx_cdesc_idx_to_ptr(struct ena_com_io_cq *io_cq, u16 idx) +{ + idx &= (io_cq->q_depth - 1); + return (struct ena_eth_io_rx_cdesc_base *) + ((uintptr_t)io_cq->cdesc_addr.virt_addr + + idx * io_cq->cdesc_entry_size_in_bytes); +} + +static u16 ena_com_cdesc_rx_pkt_get(struct ena_com_io_cq *io_cq, + u16 *first_cdesc_idx) +{ + struct ena_eth_io_rx_cdesc_base *cdesc; + u16 count = 0, head_masked; + u32 last = 0; + + do { + cdesc = ena_com_get_next_rx_cdesc(io_cq); + if (!cdesc) + break; + + ena_com_cq_inc_head(io_cq); + count++; + last = (READ_ONCE(cdesc->status) & + ENA_ETH_IO_RX_CDESC_BASE_LAST_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_LAST_SHIFT; + } while (!last); + + if (last) { + *first_cdesc_idx = io_cq->cur_rx_pkt_cdesc_start_idx; + count += io_cq->cur_rx_pkt_cdesc_count; + + head_masked = io_cq->head & (io_cq->q_depth - 1); + + io_cq->cur_rx_pkt_cdesc_count = 0; + io_cq->cur_rx_pkt_cdesc_start_idx = head_masked; + + netdev_dbg(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "ENA q_id: %d packets were completed. first desc idx %u descs# %d\n", + io_cq->qid, *first_cdesc_idx, count); + } else { + io_cq->cur_rx_pkt_cdesc_count += count; + count = 0; + } + + return count; +} + +static int ena_com_create_meta(struct ena_com_io_sq *io_sq, + struct ena_com_tx_meta *ena_meta) +{ + struct ena_eth_io_tx_meta_desc *meta_desc = NULL; + + meta_desc = get_sq_desc(io_sq); + if (unlikely(!meta_desc)) + return -EFAULT; + + memset(meta_desc, 0x0, sizeof(struct ena_eth_io_tx_meta_desc)); + + meta_desc->len_ctrl |= ENA_ETH_IO_TX_META_DESC_META_DESC_MASK; + + meta_desc->len_ctrl |= ENA_ETH_IO_TX_META_DESC_EXT_VALID_MASK; + + /* bits 0-9 of the mss */ + meta_desc->word2 |= ((u32)ena_meta->mss << + ENA_ETH_IO_TX_META_DESC_MSS_LO_SHIFT) & + ENA_ETH_IO_TX_META_DESC_MSS_LO_MASK; + /* bits 10-13 of the mss */ + meta_desc->len_ctrl |= ((ena_meta->mss >> 10) << + ENA_ETH_IO_TX_META_DESC_MSS_HI_SHIFT) & + ENA_ETH_IO_TX_META_DESC_MSS_HI_MASK; + + /* Extended meta desc */ + meta_desc->len_ctrl |= ENA_ETH_IO_TX_META_DESC_ETH_META_TYPE_MASK; + meta_desc->len_ctrl |= ((u32)io_sq->phase << + ENA_ETH_IO_TX_META_DESC_PHASE_SHIFT) & + ENA_ETH_IO_TX_META_DESC_PHASE_MASK; + + meta_desc->len_ctrl |= ENA_ETH_IO_TX_META_DESC_FIRST_MASK; + meta_desc->len_ctrl |= ENA_ETH_IO_TX_META_DESC_META_STORE_MASK; + + meta_desc->word2 |= ena_meta->l3_hdr_len & + ENA_ETH_IO_TX_META_DESC_L3_HDR_LEN_MASK; + meta_desc->word2 |= (ena_meta->l3_hdr_offset << + ENA_ETH_IO_TX_META_DESC_L3_HDR_OFF_SHIFT) & + ENA_ETH_IO_TX_META_DESC_L3_HDR_OFF_MASK; + + meta_desc->word2 |= ((u32)ena_meta->l4_hdr_len << + ENA_ETH_IO_TX_META_DESC_L4_HDR_LEN_IN_WORDS_SHIFT) & + ENA_ETH_IO_TX_META_DESC_L4_HDR_LEN_IN_WORDS_MASK; + + return ena_com_sq_update_tail(io_sq); +} + +static int ena_com_create_and_store_tx_meta_desc(struct ena_com_io_sq *io_sq, + struct ena_com_tx_ctx *ena_tx_ctx, + bool *have_meta) +{ + struct ena_com_tx_meta *ena_meta = &ena_tx_ctx->ena_meta; + + /* When disable meta caching is set, don't bother to save the meta and + * compare it to the stored version, just create the meta + */ + if (io_sq->disable_meta_caching) { + *have_meta = true; + return ena_com_create_meta(io_sq, ena_meta); + } + + if (ena_com_meta_desc_changed(io_sq, ena_tx_ctx)) { + *have_meta = true; + /* Cache the meta desc */ + memcpy(&io_sq->cached_tx_meta, ena_meta, + sizeof(struct ena_com_tx_meta)); + return ena_com_create_meta(io_sq, ena_meta); + } + + *have_meta = false; + return 0; +} + +static void ena_com_rx_set_flags(struct ena_com_io_cq *io_cq, + struct ena_com_rx_ctx *ena_rx_ctx, + struct ena_eth_io_rx_cdesc_base *cdesc) +{ + ena_rx_ctx->l3_proto = cdesc->status & + ENA_ETH_IO_RX_CDESC_BASE_L3_PROTO_IDX_MASK; + ena_rx_ctx->l4_proto = + (cdesc->status & ENA_ETH_IO_RX_CDESC_BASE_L4_PROTO_IDX_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_L4_PROTO_IDX_SHIFT; + ena_rx_ctx->l3_csum_err = + !!((cdesc->status & ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM_ERR_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM_ERR_SHIFT); + ena_rx_ctx->l4_csum_err = + !!((cdesc->status & ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_ERR_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_ERR_SHIFT); + ena_rx_ctx->l4_csum_checked = + !!((cdesc->status & ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_CHECKED_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_CHECKED_SHIFT); + ena_rx_ctx->hash = cdesc->hash; + ena_rx_ctx->frag = + (cdesc->status & ENA_ETH_IO_RX_CDESC_BASE_IPV4_FRAG_MASK) >> + ENA_ETH_IO_RX_CDESC_BASE_IPV4_FRAG_SHIFT; + + netdev_dbg(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "l3_proto %d l4_proto %d l3_csum_err %d l4_csum_err %d hash %d frag %d cdesc_status %x\n", + ena_rx_ctx->l3_proto, ena_rx_ctx->l4_proto, + ena_rx_ctx->l3_csum_err, ena_rx_ctx->l4_csum_err, + ena_rx_ctx->hash, ena_rx_ctx->frag, cdesc->status); +} + +/*****************************************************************************/ +/***************************** API **********************************/ +/*****************************************************************************/ + +int ena_com_prepare_tx(struct ena_com_io_sq *io_sq, + struct ena_com_tx_ctx *ena_tx_ctx, + int *nb_hw_desc) +{ + struct ena_eth_io_tx_desc *desc = NULL; + struct ena_com_buf *ena_bufs = ena_tx_ctx->ena_bufs; + void *buffer_to_push = ena_tx_ctx->push_header; + u16 header_len = ena_tx_ctx->header_len; + u16 num_bufs = ena_tx_ctx->num_bufs; + u16 start_tail = io_sq->tail; + int i, rc; + bool have_meta; + u64 addr_hi; + + WARN(io_sq->direction != ENA_COM_IO_QUEUE_DIRECTION_TX, "wrong Q type"); + + /* num_bufs +1 for potential meta desc */ + if (unlikely(!ena_com_sq_have_enough_space(io_sq, num_bufs + 1))) { + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Not enough space in the tx queue\n"); + return -ENOMEM; + } + + if (unlikely(header_len > io_sq->tx_max_header_size)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Header size is too large %d max header: %d\n", + header_len, io_sq->tx_max_header_size); + return -EINVAL; + } + + if (unlikely(io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV && + !buffer_to_push)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Push header wasn't provided in LLQ mode\n"); + return -EINVAL; + } + + rc = ena_com_write_header_to_bounce(io_sq, buffer_to_push, header_len); + if (unlikely(rc)) + return rc; + + rc = ena_com_create_and_store_tx_meta_desc(io_sq, ena_tx_ctx, &have_meta); + if (unlikely(rc)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to create and store tx meta desc\n"); + return rc; + } + + /* If the caller doesn't want to send packets */ + if (unlikely(!num_bufs && !header_len)) { + rc = ena_com_close_bounce_buffer(io_sq); + if (rc) + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to write buffers to LLQ\n"); + *nb_hw_desc = io_sq->tail - start_tail; + return rc; + } + + desc = get_sq_desc(io_sq); + if (unlikely(!desc)) + return -EFAULT; + memset(desc, 0x0, sizeof(struct ena_eth_io_tx_desc)); + + /* Set first desc when we don't have meta descriptor */ + if (!have_meta) + desc->len_ctrl |= ENA_ETH_IO_TX_DESC_FIRST_MASK; + + desc->buff_addr_hi_hdr_sz |= ((u32)header_len << + ENA_ETH_IO_TX_DESC_HEADER_LENGTH_SHIFT) & + ENA_ETH_IO_TX_DESC_HEADER_LENGTH_MASK; + desc->len_ctrl |= ((u32)io_sq->phase << ENA_ETH_IO_TX_DESC_PHASE_SHIFT) & + ENA_ETH_IO_TX_DESC_PHASE_MASK; + + desc->len_ctrl |= ENA_ETH_IO_TX_DESC_COMP_REQ_MASK; + + /* Bits 0-9 */ + desc->meta_ctrl |= ((u32)ena_tx_ctx->req_id << + ENA_ETH_IO_TX_DESC_REQ_ID_LO_SHIFT) & + ENA_ETH_IO_TX_DESC_REQ_ID_LO_MASK; + + desc->meta_ctrl |= (ena_tx_ctx->df << + ENA_ETH_IO_TX_DESC_DF_SHIFT) & + ENA_ETH_IO_TX_DESC_DF_MASK; + + /* Bits 10-15 */ + desc->len_ctrl |= ((ena_tx_ctx->req_id >> 10) << + ENA_ETH_IO_TX_DESC_REQ_ID_HI_SHIFT) & + ENA_ETH_IO_TX_DESC_REQ_ID_HI_MASK; + + if (ena_tx_ctx->meta_valid) { + desc->meta_ctrl |= (ena_tx_ctx->tso_enable << + ENA_ETH_IO_TX_DESC_TSO_EN_SHIFT) & + ENA_ETH_IO_TX_DESC_TSO_EN_MASK; + desc->meta_ctrl |= ena_tx_ctx->l3_proto & + ENA_ETH_IO_TX_DESC_L3_PROTO_IDX_MASK; + desc->meta_ctrl |= (ena_tx_ctx->l4_proto << + ENA_ETH_IO_TX_DESC_L4_PROTO_IDX_SHIFT) & + ENA_ETH_IO_TX_DESC_L4_PROTO_IDX_MASK; + desc->meta_ctrl |= (ena_tx_ctx->l3_csum_enable << + ENA_ETH_IO_TX_DESC_L3_CSUM_EN_SHIFT) & + ENA_ETH_IO_TX_DESC_L3_CSUM_EN_MASK; + desc->meta_ctrl |= (ena_tx_ctx->l4_csum_enable << + ENA_ETH_IO_TX_DESC_L4_CSUM_EN_SHIFT) & + ENA_ETH_IO_TX_DESC_L4_CSUM_EN_MASK; + desc->meta_ctrl |= (ena_tx_ctx->l4_csum_partial << + ENA_ETH_IO_TX_DESC_L4_CSUM_PARTIAL_SHIFT) & + ENA_ETH_IO_TX_DESC_L4_CSUM_PARTIAL_MASK; + } + + for (i = 0; i < num_bufs; i++) { + /* The first desc share the same desc as the header */ + if (likely(i != 0)) { + rc = ena_com_sq_update_tail(io_sq); + if (unlikely(rc)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to update sq tail\n"); + return rc; + } + + desc = get_sq_desc(io_sq); + if (unlikely(!desc)) + return -EFAULT; + + memset(desc, 0x0, sizeof(struct ena_eth_io_tx_desc)); + + desc->len_ctrl |= ((u32)io_sq->phase << + ENA_ETH_IO_TX_DESC_PHASE_SHIFT) & + ENA_ETH_IO_TX_DESC_PHASE_MASK; + } + + desc->len_ctrl |= ena_bufs->len & + ENA_ETH_IO_TX_DESC_LENGTH_MASK; + + addr_hi = ((ena_bufs->paddr & + GENMASK_ULL(io_sq->dma_addr_bits - 1, 32)) >> 32); + + desc->buff_addr_lo = (u32)ena_bufs->paddr; + desc->buff_addr_hi_hdr_sz |= addr_hi & + ENA_ETH_IO_TX_DESC_ADDR_HI_MASK; + ena_bufs++; + } + + /* set the last desc indicator */ + desc->len_ctrl |= ENA_ETH_IO_TX_DESC_LAST_MASK; + + rc = ena_com_sq_update_tail(io_sq); + if (unlikely(rc)) { + netdev_err(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Failed to update sq tail of the last descriptor\n"); + return rc; + } + + rc = ena_com_close_bounce_buffer(io_sq); + + *nb_hw_desc = io_sq->tail - start_tail; + return rc; +} + +int ena_com_rx_pkt(struct ena_com_io_cq *io_cq, + struct ena_com_io_sq *io_sq, + struct ena_com_rx_ctx *ena_rx_ctx) +{ + struct ena_com_rx_buf_info *ena_buf = &ena_rx_ctx->ena_bufs[0]; + struct ena_eth_io_rx_cdesc_base *cdesc = NULL; + u16 q_depth = io_cq->q_depth; + u16 cdesc_idx = 0; + u16 nb_hw_desc; + u16 i = 0; + + WARN(io_cq->direction != ENA_COM_IO_QUEUE_DIRECTION_RX, "wrong Q type"); + + nb_hw_desc = ena_com_cdesc_rx_pkt_get(io_cq, &cdesc_idx); + if (nb_hw_desc == 0) { + ena_rx_ctx->descs = nb_hw_desc; + return 0; + } + + netdev_dbg(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "Fetch rx packet: queue %d completed desc: %d\n", io_cq->qid, + nb_hw_desc); + + if (unlikely(nb_hw_desc > ena_rx_ctx->max_bufs)) { + netdev_err(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "Too many RX cdescs (%d) > MAX(%d)\n", nb_hw_desc, + ena_rx_ctx->max_bufs); + return -ENOSPC; + } + + cdesc = ena_com_rx_cdesc_idx_to_ptr(io_cq, cdesc_idx); + ena_rx_ctx->pkt_offset = cdesc->offset; + + do { + ena_buf[i].len = cdesc->length; + ena_buf[i].req_id = cdesc->req_id; + if (unlikely(ena_buf[i].req_id >= q_depth)) + return -EIO; + + if (++i >= nb_hw_desc) + break; + + cdesc = ena_com_rx_cdesc_idx_to_ptr(io_cq, cdesc_idx + i); + + } while (1); + + /* Update SQ head ptr */ + io_sq->next_to_comp += nb_hw_desc; + + netdev_dbg(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "[%s][QID#%d] Updating SQ head to: %d\n", __func__, + io_sq->qid, io_sq->next_to_comp); + + /* Get rx flags from the last pkt */ + ena_com_rx_set_flags(io_cq, ena_rx_ctx, cdesc); + + ena_rx_ctx->descs = nb_hw_desc; + + return 0; +} + +int ena_com_add_single_rx_desc(struct ena_com_io_sq *io_sq, + struct ena_com_buf *ena_buf, + u16 req_id) +{ + struct ena_eth_io_rx_desc *desc; + + WARN(io_sq->direction != ENA_COM_IO_QUEUE_DIRECTION_RX, "wrong Q type"); + + if (unlikely(!ena_com_sq_have_enough_space(io_sq, 1))) + return -ENOSPC; + + desc = get_sq_desc(io_sq); + if (unlikely(!desc)) + return -EFAULT; + + memset(desc, 0x0, sizeof(struct ena_eth_io_rx_desc)); + + desc->length = ena_buf->len; + + desc->ctrl = ENA_ETH_IO_RX_DESC_FIRST_MASK | + ENA_ETH_IO_RX_DESC_LAST_MASK | + ENA_ETH_IO_RX_DESC_COMP_REQ_MASK | + (io_sq->phase & ENA_ETH_IO_RX_DESC_PHASE_MASK); + + desc->req_id = req_id; + + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "[%s] Adding single RX desc, Queue: %u, req_id: %u\n", + __func__, io_sq->qid, req_id); + + desc->buff_addr_lo = (u32)ena_buf->paddr; + desc->buff_addr_hi = + ((ena_buf->paddr & GENMASK_ULL(io_sq->dma_addr_bits - 1, 32)) >> 32); + + return ena_com_sq_update_tail(io_sq); +} + +bool ena_com_cq_empty(struct ena_com_io_cq *io_cq) +{ + struct ena_eth_io_rx_cdesc_base *cdesc; + + cdesc = ena_com_get_next_rx_cdesc(io_cq); + if (cdesc) + return false; + else + return true; +} diff --git a/drivers/net/ethernet/amazon/ena/ena_eth_com.h b/drivers/net/ethernet/amazon/ena/ena_eth_com.h new file mode 100644 index 0000000000..372b259279 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_eth_com.h @@ -0,0 +1,261 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#ifndef ENA_ETH_COM_H_ +#define ENA_ETH_COM_H_ + +#include "ena_com.h" + +/* head update threshold in units of (queue size / ENA_COMP_HEAD_THRESH) */ +#define ENA_COMP_HEAD_THRESH 4 +/* we allow 2 DMA descriptors per LLQ entry */ +#define ENA_LLQ_ENTRY_DESC_CHUNK_SIZE (2 * sizeof(struct ena_eth_io_tx_desc)) +#define ENA_LLQ_HEADER (128UL - ENA_LLQ_ENTRY_DESC_CHUNK_SIZE) +#define ENA_LLQ_LARGE_HEADER (256UL - ENA_LLQ_ENTRY_DESC_CHUNK_SIZE) + +struct ena_com_tx_ctx { + struct ena_com_tx_meta ena_meta; + struct ena_com_buf *ena_bufs; + /* For LLQ, header buffer - pushed to the device mem space */ + void *push_header; + + enum ena_eth_io_l3_proto_index l3_proto; + enum ena_eth_io_l4_proto_index l4_proto; + u16 num_bufs; + u16 req_id; + /* For regular queue, indicate the size of the header + * For LLQ, indicate the size of the pushed buffer + */ + u16 header_len; + + u8 meta_valid; + u8 tso_enable; + u8 l3_csum_enable; + u8 l4_csum_enable; + u8 l4_csum_partial; + u8 df; /* Don't fragment */ +}; + +struct ena_com_rx_ctx { + struct ena_com_rx_buf_info *ena_bufs; + enum ena_eth_io_l3_proto_index l3_proto; + enum ena_eth_io_l4_proto_index l4_proto; + bool l3_csum_err; + bool l4_csum_err; + u8 l4_csum_checked; + /* fragmented packet */ + bool frag; + u32 hash; + u16 descs; + int max_bufs; + u8 pkt_offset; +}; + +int ena_com_prepare_tx(struct ena_com_io_sq *io_sq, + struct ena_com_tx_ctx *ena_tx_ctx, + int *nb_hw_desc); + +int ena_com_rx_pkt(struct ena_com_io_cq *io_cq, + struct ena_com_io_sq *io_sq, + struct ena_com_rx_ctx *ena_rx_ctx); + +int ena_com_add_single_rx_desc(struct ena_com_io_sq *io_sq, + struct ena_com_buf *ena_buf, + u16 req_id); + +bool ena_com_cq_empty(struct ena_com_io_cq *io_cq); + +static inline void ena_com_unmask_intr(struct ena_com_io_cq *io_cq, + struct ena_eth_io_intr_reg *intr_reg) +{ + writel(intr_reg->intr_control, io_cq->unmask_reg); +} + +static inline int ena_com_free_q_entries(struct ena_com_io_sq *io_sq) +{ + u16 tail, next_to_comp, cnt; + + next_to_comp = io_sq->next_to_comp; + tail = io_sq->tail; + cnt = tail - next_to_comp; + + return io_sq->q_depth - 1 - cnt; +} + +/* Check if the submission queue has enough space to hold required_buffers */ +static inline bool ena_com_sq_have_enough_space(struct ena_com_io_sq *io_sq, + u16 required_buffers) +{ + int temp; + + if (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_HOST) + return ena_com_free_q_entries(io_sq) >= required_buffers; + + /* This calculation doesn't need to be 100% accurate. So to reduce + * the calculation overhead just Subtract 2 lines from the free descs + * (one for the header line and one to compensate the devision + * down calculation. + */ + temp = required_buffers / io_sq->llq_info.descs_per_entry + 2; + + return ena_com_free_q_entries(io_sq) > temp; +} + +static inline bool ena_com_meta_desc_changed(struct ena_com_io_sq *io_sq, + struct ena_com_tx_ctx *ena_tx_ctx) +{ + if (!ena_tx_ctx->meta_valid) + return false; + + return !!memcmp(&io_sq->cached_tx_meta, + &ena_tx_ctx->ena_meta, + sizeof(struct ena_com_tx_meta)); +} + +static inline bool is_llq_max_tx_burst_exists(struct ena_com_io_sq *io_sq) +{ + return (io_sq->mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) && + io_sq->llq_info.max_entries_in_tx_burst > 0; +} + +static inline bool ena_com_is_doorbell_needed(struct ena_com_io_sq *io_sq, + struct ena_com_tx_ctx *ena_tx_ctx) +{ + struct ena_com_llq_info *llq_info; + int descs_after_first_entry; + int num_entries_needed = 1; + u16 num_descs; + + if (!is_llq_max_tx_burst_exists(io_sq)) + return false; + + llq_info = &io_sq->llq_info; + num_descs = ena_tx_ctx->num_bufs; + + if (llq_info->disable_meta_caching || + unlikely(ena_com_meta_desc_changed(io_sq, ena_tx_ctx))) + ++num_descs; + + if (num_descs > llq_info->descs_num_before_header) { + descs_after_first_entry = num_descs - llq_info->descs_num_before_header; + num_entries_needed += DIV_ROUND_UP(descs_after_first_entry, + llq_info->descs_per_entry); + } + + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Queue: %d num_descs: %d num_entries_needed: %d\n", + io_sq->qid, num_descs, num_entries_needed); + + return num_entries_needed > io_sq->entries_in_tx_burst_left; +} + +static inline int ena_com_write_sq_doorbell(struct ena_com_io_sq *io_sq) +{ + u16 max_entries_in_tx_burst = io_sq->llq_info.max_entries_in_tx_burst; + u16 tail = io_sq->tail; + + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Write submission queue doorbell for queue: %d tail: %d\n", + io_sq->qid, tail); + + writel(tail, io_sq->db_addr); + + if (is_llq_max_tx_burst_exists(io_sq)) { + netdev_dbg(ena_com_io_sq_to_ena_dev(io_sq)->net_device, + "Reset available entries in tx burst for queue %d to %d\n", + io_sq->qid, max_entries_in_tx_burst); + io_sq->entries_in_tx_burst_left = max_entries_in_tx_burst; + } + + return 0; +} + +static inline int ena_com_update_dev_comp_head(struct ena_com_io_cq *io_cq) +{ + u16 unreported_comp, head; + bool need_update; + + if (unlikely(io_cq->cq_head_db_reg)) { + head = io_cq->head; + unreported_comp = head - io_cq->last_head_update; + need_update = unreported_comp > (io_cq->q_depth / ENA_COMP_HEAD_THRESH); + + if (unlikely(need_update)) { + netdev_dbg(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "Write completion queue doorbell for queue %d: head: %d\n", + io_cq->qid, head); + writel(head, io_cq->cq_head_db_reg); + io_cq->last_head_update = head; + } + } + + return 0; +} + +static inline void ena_com_update_numa_node(struct ena_com_io_cq *io_cq, + u8 numa_node) +{ + struct ena_eth_io_numa_node_cfg_reg numa_cfg; + + if (!io_cq->numa_node_cfg_reg) + return; + + numa_cfg.numa_cfg = (numa_node & ENA_ETH_IO_NUMA_NODE_CFG_REG_NUMA_MASK) + | ENA_ETH_IO_NUMA_NODE_CFG_REG_ENABLED_MASK; + + writel(numa_cfg.numa_cfg, io_cq->numa_node_cfg_reg); +} + +static inline void ena_com_comp_ack(struct ena_com_io_sq *io_sq, u16 elem) +{ + io_sq->next_to_comp += elem; +} + +static inline void ena_com_cq_inc_head(struct ena_com_io_cq *io_cq) +{ + io_cq->head++; + + /* Switch phase bit in case of wrap around */ + if (unlikely((io_cq->head & (io_cq->q_depth - 1)) == 0)) + io_cq->phase ^= 1; +} + +static inline int ena_com_tx_comp_req_id_get(struct ena_com_io_cq *io_cq, + u16 *req_id) +{ + u8 expected_phase, cdesc_phase; + struct ena_eth_io_tx_cdesc *cdesc; + u16 masked_head; + + masked_head = io_cq->head & (io_cq->q_depth - 1); + expected_phase = io_cq->phase; + + cdesc = (struct ena_eth_io_tx_cdesc *) + ((uintptr_t)io_cq->cdesc_addr.virt_addr + + (masked_head * io_cq->cdesc_entry_size_in_bytes)); + + /* When the current completion descriptor phase isn't the same as the + * expected, it mean that the device still didn't update + * this completion. + */ + cdesc_phase = READ_ONCE(cdesc->flags) & ENA_ETH_IO_TX_CDESC_PHASE_MASK; + if (cdesc_phase != expected_phase) + return -EAGAIN; + + dma_rmb(); + + *req_id = READ_ONCE(cdesc->req_id); + if (unlikely(*req_id >= io_cq->q_depth)) { + netdev_err(ena_com_io_cq_to_ena_dev(io_cq)->net_device, + "Invalid req id %d\n", cdesc->req_id); + return -EINVAL; + } + + ena_com_cq_inc_head(io_cq); + + return 0; +} + +#endif /* ENA_ETH_COM_H_ */ diff --git a/drivers/net/ethernet/amazon/ena/ena_eth_io_defs.h b/drivers/net/ethernet/amazon/ena/ena_eth_io_defs.h new file mode 100644 index 0000000000..332ac0d28a --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_eth_io_defs.h @@ -0,0 +1,390 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ +#ifndef _ENA_ETH_IO_H_ +#define _ENA_ETH_IO_H_ + +enum ena_eth_io_l3_proto_index { + ENA_ETH_IO_L3_PROTO_UNKNOWN = 0, + ENA_ETH_IO_L3_PROTO_IPV4 = 8, + ENA_ETH_IO_L3_PROTO_IPV6 = 11, + ENA_ETH_IO_L3_PROTO_FCOE = 21, + ENA_ETH_IO_L3_PROTO_ROCE = 22, +}; + +enum ena_eth_io_l4_proto_index { + ENA_ETH_IO_L4_PROTO_UNKNOWN = 0, + ENA_ETH_IO_L4_PROTO_TCP = 12, + ENA_ETH_IO_L4_PROTO_UDP = 13, + ENA_ETH_IO_L4_PROTO_ROUTEABLE_ROCE = 23, +}; + +struct ena_eth_io_tx_desc { + /* 15:0 : length - Buffer length in bytes, must + * include any packet trailers that the ENA supposed + * to update like End-to-End CRC, Authentication GMAC + * etc. This length must not include the + * 'Push_Buffer' length. This length must not include + * the 4-byte added in the end for 802.3 Ethernet FCS + * 21:16 : req_id_hi - Request ID[15:10] + * 22 : reserved22 - MBZ + * 23 : meta_desc - MBZ + * 24 : phase + * 25 : reserved1 - MBZ + * 26 : first - Indicates first descriptor in + * transaction + * 27 : last - Indicates last descriptor in + * transaction + * 28 : comp_req - Indicates whether completion + * should be posted, after packet is transmitted. + * Valid only for first descriptor + * 30:29 : reserved29 - MBZ + * 31 : reserved31 - MBZ + */ + u32 len_ctrl; + + /* 3:0 : l3_proto_idx - L3 protocol. This field + * required when l3_csum_en,l3_csum or tso_en are set. + * 4 : DF - IPv4 DF, must be 0 if packet is IPv4 and + * DF flags of the IPv4 header is 0. Otherwise must + * be set to 1 + * 6:5 : reserved5 + * 7 : tso_en - Enable TSO, For TCP only. + * 12:8 : l4_proto_idx - L4 protocol. This field need + * to be set when l4_csum_en or tso_en are set. + * 13 : l3_csum_en - enable IPv4 header checksum. + * 14 : l4_csum_en - enable TCP/UDP checksum. + * 15 : ethernet_fcs_dis - when set, the controller + * will not append the 802.3 Ethernet Frame Check + * Sequence to the packet + * 16 : reserved16 + * 17 : l4_csum_partial - L4 partial checksum. when + * set to 0, the ENA calculates the L4 checksum, + * where the Destination Address required for the + * TCP/UDP pseudo-header is taken from the actual + * packet L3 header. when set to 1, the ENA doesn't + * calculate the sum of the pseudo-header, instead, + * the checksum field of the L4 is used instead. When + * TSO enabled, the checksum of the pseudo-header + * must not include the tcp length field. L4 partial + * checksum should be used for IPv6 packet that + * contains Routing Headers. + * 20:18 : reserved18 - MBZ + * 21 : reserved21 - MBZ + * 31:22 : req_id_lo - Request ID[9:0] + */ + u32 meta_ctrl; + + u32 buff_addr_lo; + + /* address high and header size + * 15:0 : addr_hi - Buffer Pointer[47:32] + * 23:16 : reserved16_w2 + * 31:24 : header_length - Header length. For Low + * Latency Queues, this fields indicates the number + * of bytes written to the headers' memory. For + * normal queues, if packet is TCP or UDP, and longer + * than max_header_size, then this field should be + * set to the sum of L4 header offset and L4 header + * size(without options), otherwise, this field + * should be set to 0. For both modes, this field + * must not exceed the max_header_size. + * max_header_size value is reported by the Max + * Queues Feature descriptor + */ + u32 buff_addr_hi_hdr_sz; +}; + +struct ena_eth_io_tx_meta_desc { + /* 9:0 : req_id_lo - Request ID[9:0] + * 11:10 : reserved10 - MBZ + * 12 : reserved12 - MBZ + * 13 : reserved13 - MBZ + * 14 : ext_valid - if set, offset fields in Word2 + * are valid Also MSS High in Word 0 and bits [31:24] + * in Word 3 + * 15 : reserved15 + * 19:16 : mss_hi + * 20 : eth_meta_type - 0: Tx Metadata Descriptor, 1: + * Extended Metadata Descriptor + * 21 : meta_store - Store extended metadata in queue + * cache + * 22 : reserved22 - MBZ + * 23 : meta_desc - MBO + * 24 : phase + * 25 : reserved25 - MBZ + * 26 : first - Indicates first descriptor in + * transaction + * 27 : last - Indicates last descriptor in + * transaction + * 28 : comp_req - Indicates whether completion + * should be posted, after packet is transmitted. + * Valid only for first descriptor + * 30:29 : reserved29 - MBZ + * 31 : reserved31 - MBZ + */ + u32 len_ctrl; + + /* 5:0 : req_id_hi + * 31:6 : reserved6 - MBZ + */ + u32 word1; + + /* 7:0 : l3_hdr_len + * 15:8 : l3_hdr_off + * 21:16 : l4_hdr_len_in_words - counts the L4 header + * length in words. there is an explicit assumption + * that L4 header appears right after L3 header and + * L4 offset is based on l3_hdr_off+l3_hdr_len + * 31:22 : mss_lo + */ + u32 word2; + + u32 reserved; +}; + +struct ena_eth_io_tx_cdesc { + /* Request ID[15:0] */ + u16 req_id; + + u8 status; + + /* flags + * 0 : phase + * 7:1 : reserved1 + */ + u8 flags; + + u16 sub_qid; + + u16 sq_head_idx; +}; + +struct ena_eth_io_rx_desc { + /* In bytes. 0 means 64KB */ + u16 length; + + /* MBZ */ + u8 reserved2; + + /* 0 : phase + * 1 : reserved1 - MBZ + * 2 : first - Indicates first descriptor in + * transaction + * 3 : last - Indicates last descriptor in transaction + * 4 : comp_req + * 5 : reserved5 - MBO + * 7:6 : reserved6 - MBZ + */ + u8 ctrl; + + u16 req_id; + + /* MBZ */ + u16 reserved6; + + u32 buff_addr_lo; + + u16 buff_addr_hi; + + /* MBZ */ + u16 reserved16_w3; +}; + +/* 4-word format Note: all ethernet parsing information are valid only when + * last=1 + */ +struct ena_eth_io_rx_cdesc_base { + /* 4:0 : l3_proto_idx + * 6:5 : src_vlan_cnt + * 7 : reserved7 - MBZ + * 12:8 : l4_proto_idx + * 13 : l3_csum_err - when set, either the L3 + * checksum error detected, or, the controller didn't + * validate the checksum. This bit is valid only when + * l3_proto_idx indicates IPv4 packet + * 14 : l4_csum_err - when set, either the L4 + * checksum error detected, or, the controller didn't + * validate the checksum. This bit is valid only when + * l4_proto_idx indicates TCP/UDP packet, and, + * ipv4_frag is not set. This bit is valid only when + * l4_csum_checked below is set. + * 15 : ipv4_frag - Indicates IPv4 fragmented packet + * 16 : l4_csum_checked - L4 checksum was verified + * (could be OK or error), when cleared the status of + * checksum is unknown + * 23:17 : reserved17 - MBZ + * 24 : phase + * 25 : l3_csum2 - second checksum engine result + * 26 : first - Indicates first descriptor in + * transaction + * 27 : last - Indicates last descriptor in + * transaction + * 29:28 : reserved28 + * 30 : buffer - 0: Metadata descriptor. 1: Buffer + * Descriptor was used + * 31 : reserved31 + */ + u32 status; + + u16 length; + + u16 req_id; + + /* 32-bit hash result */ + u32 hash; + + u16 sub_qid; + + u8 offset; + + u8 reserved; +}; + +/* 8-word format */ +struct ena_eth_io_rx_cdesc_ext { + struct ena_eth_io_rx_cdesc_base base; + + u32 buff_addr_lo; + + u16 buff_addr_hi; + + u16 reserved16; + + u32 reserved_w6; + + u32 reserved_w7; +}; + +struct ena_eth_io_intr_reg { + /* 14:0 : rx_intr_delay + * 29:15 : tx_intr_delay + * 30 : intr_unmask + * 31 : reserved + */ + u32 intr_control; +}; + +struct ena_eth_io_numa_node_cfg_reg { + /* 7:0 : numa + * 30:8 : reserved + * 31 : enabled + */ + u32 numa_cfg; +}; + +/* tx_desc */ +#define ENA_ETH_IO_TX_DESC_LENGTH_MASK GENMASK(15, 0) +#define ENA_ETH_IO_TX_DESC_REQ_ID_HI_SHIFT 16 +#define ENA_ETH_IO_TX_DESC_REQ_ID_HI_MASK GENMASK(21, 16) +#define ENA_ETH_IO_TX_DESC_META_DESC_SHIFT 23 +#define ENA_ETH_IO_TX_DESC_META_DESC_MASK BIT(23) +#define ENA_ETH_IO_TX_DESC_PHASE_SHIFT 24 +#define ENA_ETH_IO_TX_DESC_PHASE_MASK BIT(24) +#define ENA_ETH_IO_TX_DESC_FIRST_SHIFT 26 +#define ENA_ETH_IO_TX_DESC_FIRST_MASK BIT(26) +#define ENA_ETH_IO_TX_DESC_LAST_SHIFT 27 +#define ENA_ETH_IO_TX_DESC_LAST_MASK BIT(27) +#define ENA_ETH_IO_TX_DESC_COMP_REQ_SHIFT 28 +#define ENA_ETH_IO_TX_DESC_COMP_REQ_MASK BIT(28) +#define ENA_ETH_IO_TX_DESC_L3_PROTO_IDX_MASK GENMASK(3, 0) +#define ENA_ETH_IO_TX_DESC_DF_SHIFT 4 +#define ENA_ETH_IO_TX_DESC_DF_MASK BIT(4) +#define ENA_ETH_IO_TX_DESC_TSO_EN_SHIFT 7 +#define ENA_ETH_IO_TX_DESC_TSO_EN_MASK BIT(7) +#define ENA_ETH_IO_TX_DESC_L4_PROTO_IDX_SHIFT 8 +#define ENA_ETH_IO_TX_DESC_L4_PROTO_IDX_MASK GENMASK(12, 8) +#define ENA_ETH_IO_TX_DESC_L3_CSUM_EN_SHIFT 13 +#define ENA_ETH_IO_TX_DESC_L3_CSUM_EN_MASK BIT(13) +#define ENA_ETH_IO_TX_DESC_L4_CSUM_EN_SHIFT 14 +#define ENA_ETH_IO_TX_DESC_L4_CSUM_EN_MASK BIT(14) +#define ENA_ETH_IO_TX_DESC_ETHERNET_FCS_DIS_SHIFT 15 +#define ENA_ETH_IO_TX_DESC_ETHERNET_FCS_DIS_MASK BIT(15) +#define ENA_ETH_IO_TX_DESC_L4_CSUM_PARTIAL_SHIFT 17 +#define ENA_ETH_IO_TX_DESC_L4_CSUM_PARTIAL_MASK BIT(17) +#define ENA_ETH_IO_TX_DESC_REQ_ID_LO_SHIFT 22 +#define ENA_ETH_IO_TX_DESC_REQ_ID_LO_MASK GENMASK(31, 22) +#define ENA_ETH_IO_TX_DESC_ADDR_HI_MASK GENMASK(15, 0) +#define ENA_ETH_IO_TX_DESC_HEADER_LENGTH_SHIFT 24 +#define ENA_ETH_IO_TX_DESC_HEADER_LENGTH_MASK GENMASK(31, 24) + +/* tx_meta_desc */ +#define ENA_ETH_IO_TX_META_DESC_REQ_ID_LO_MASK GENMASK(9, 0) +#define ENA_ETH_IO_TX_META_DESC_EXT_VALID_SHIFT 14 +#define ENA_ETH_IO_TX_META_DESC_EXT_VALID_MASK BIT(14) +#define ENA_ETH_IO_TX_META_DESC_MSS_HI_SHIFT 16 +#define ENA_ETH_IO_TX_META_DESC_MSS_HI_MASK GENMASK(19, 16) +#define ENA_ETH_IO_TX_META_DESC_ETH_META_TYPE_SHIFT 20 +#define ENA_ETH_IO_TX_META_DESC_ETH_META_TYPE_MASK BIT(20) +#define ENA_ETH_IO_TX_META_DESC_META_STORE_SHIFT 21 +#define ENA_ETH_IO_TX_META_DESC_META_STORE_MASK BIT(21) +#define ENA_ETH_IO_TX_META_DESC_META_DESC_SHIFT 23 +#define ENA_ETH_IO_TX_META_DESC_META_DESC_MASK BIT(23) +#define ENA_ETH_IO_TX_META_DESC_PHASE_SHIFT 24 +#define ENA_ETH_IO_TX_META_DESC_PHASE_MASK BIT(24) +#define ENA_ETH_IO_TX_META_DESC_FIRST_SHIFT 26 +#define ENA_ETH_IO_TX_META_DESC_FIRST_MASK BIT(26) +#define ENA_ETH_IO_TX_META_DESC_LAST_SHIFT 27 +#define ENA_ETH_IO_TX_META_DESC_LAST_MASK BIT(27) +#define ENA_ETH_IO_TX_META_DESC_COMP_REQ_SHIFT 28 +#define ENA_ETH_IO_TX_META_DESC_COMP_REQ_MASK BIT(28) +#define ENA_ETH_IO_TX_META_DESC_REQ_ID_HI_MASK GENMASK(5, 0) +#define ENA_ETH_IO_TX_META_DESC_L3_HDR_LEN_MASK GENMASK(7, 0) +#define ENA_ETH_IO_TX_META_DESC_L3_HDR_OFF_SHIFT 8 +#define ENA_ETH_IO_TX_META_DESC_L3_HDR_OFF_MASK GENMASK(15, 8) +#define ENA_ETH_IO_TX_META_DESC_L4_HDR_LEN_IN_WORDS_SHIFT 16 +#define ENA_ETH_IO_TX_META_DESC_L4_HDR_LEN_IN_WORDS_MASK GENMASK(21, 16) +#define ENA_ETH_IO_TX_META_DESC_MSS_LO_SHIFT 22 +#define ENA_ETH_IO_TX_META_DESC_MSS_LO_MASK GENMASK(31, 22) + +/* tx_cdesc */ +#define ENA_ETH_IO_TX_CDESC_PHASE_MASK BIT(0) + +/* rx_desc */ +#define ENA_ETH_IO_RX_DESC_PHASE_MASK BIT(0) +#define ENA_ETH_IO_RX_DESC_FIRST_SHIFT 2 +#define ENA_ETH_IO_RX_DESC_FIRST_MASK BIT(2) +#define ENA_ETH_IO_RX_DESC_LAST_SHIFT 3 +#define ENA_ETH_IO_RX_DESC_LAST_MASK BIT(3) +#define ENA_ETH_IO_RX_DESC_COMP_REQ_SHIFT 4 +#define ENA_ETH_IO_RX_DESC_COMP_REQ_MASK BIT(4) + +/* rx_cdesc_base */ +#define ENA_ETH_IO_RX_CDESC_BASE_L3_PROTO_IDX_MASK GENMASK(4, 0) +#define ENA_ETH_IO_RX_CDESC_BASE_SRC_VLAN_CNT_SHIFT 5 +#define ENA_ETH_IO_RX_CDESC_BASE_SRC_VLAN_CNT_MASK GENMASK(6, 5) +#define ENA_ETH_IO_RX_CDESC_BASE_L4_PROTO_IDX_SHIFT 8 +#define ENA_ETH_IO_RX_CDESC_BASE_L4_PROTO_IDX_MASK GENMASK(12, 8) +#define ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM_ERR_SHIFT 13 +#define ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM_ERR_MASK BIT(13) +#define ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_ERR_SHIFT 14 +#define ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_ERR_MASK BIT(14) +#define ENA_ETH_IO_RX_CDESC_BASE_IPV4_FRAG_SHIFT 15 +#define ENA_ETH_IO_RX_CDESC_BASE_IPV4_FRAG_MASK BIT(15) +#define ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_CHECKED_SHIFT 16 +#define ENA_ETH_IO_RX_CDESC_BASE_L4_CSUM_CHECKED_MASK BIT(16) +#define ENA_ETH_IO_RX_CDESC_BASE_PHASE_SHIFT 24 +#define ENA_ETH_IO_RX_CDESC_BASE_PHASE_MASK BIT(24) +#define ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM2_SHIFT 25 +#define ENA_ETH_IO_RX_CDESC_BASE_L3_CSUM2_MASK BIT(25) +#define ENA_ETH_IO_RX_CDESC_BASE_FIRST_SHIFT 26 +#define ENA_ETH_IO_RX_CDESC_BASE_FIRST_MASK BIT(26) +#define ENA_ETH_IO_RX_CDESC_BASE_LAST_SHIFT 27 +#define ENA_ETH_IO_RX_CDESC_BASE_LAST_MASK BIT(27) +#define ENA_ETH_IO_RX_CDESC_BASE_BUFFER_SHIFT 30 +#define ENA_ETH_IO_RX_CDESC_BASE_BUFFER_MASK BIT(30) + +/* intr_reg */ +#define ENA_ETH_IO_INTR_REG_RX_INTR_DELAY_MASK GENMASK(14, 0) +#define ENA_ETH_IO_INTR_REG_TX_INTR_DELAY_SHIFT 15 +#define ENA_ETH_IO_INTR_REG_TX_INTR_DELAY_MASK GENMASK(29, 15) +#define ENA_ETH_IO_INTR_REG_INTR_UNMASK_SHIFT 30 +#define ENA_ETH_IO_INTR_REG_INTR_UNMASK_MASK BIT(30) + +/* numa_node_cfg_reg */ +#define ENA_ETH_IO_NUMA_NODE_CFG_REG_NUMA_MASK GENMASK(7, 0) +#define ENA_ETH_IO_NUMA_NODE_CFG_REG_ENABLED_SHIFT 31 +#define ENA_ETH_IO_NUMA_NODE_CFG_REG_ENABLED_MASK BIT(31) + +#endif /* _ENA_ETH_IO_H_ */ diff --git a/drivers/net/ethernet/amazon/ena/ena_ethtool.c b/drivers/net/ethernet/amazon/ena/ena_ethtool.c new file mode 100644 index 0000000000..d671df4b76 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_ethtool.c @@ -0,0 +1,1066 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#include <linux/ethtool.h> +#include <linux/pci.h> + +#include "ena_netdev.h" + +struct ena_stats { + char name[ETH_GSTRING_LEN]; + int stat_offset; +}; + +#define ENA_STAT_ENA_COM_ENTRY(stat) { \ + .name = #stat, \ + .stat_offset = offsetof(struct ena_com_stats_admin, stat) / sizeof(u64) \ +} + +#define ENA_STAT_ENTRY(stat, stat_type) { \ + .name = #stat, \ + .stat_offset = offsetof(struct ena_stats_##stat_type, stat) / sizeof(u64) \ +} + +#define ENA_STAT_HW_ENTRY(stat, stat_type) { \ + .name = #stat, \ + .stat_offset = offsetof(struct ena_admin_##stat_type, stat) / sizeof(u64) \ +} + +#define ENA_STAT_RX_ENTRY(stat) \ + ENA_STAT_ENTRY(stat, rx) + +#define ENA_STAT_TX_ENTRY(stat) \ + ENA_STAT_ENTRY(stat, tx) + +#define ENA_STAT_GLOBAL_ENTRY(stat) \ + ENA_STAT_ENTRY(stat, dev) + +#define ENA_STAT_ENI_ENTRY(stat) \ + ENA_STAT_HW_ENTRY(stat, eni_stats) + +static const struct ena_stats ena_stats_global_strings[] = { + ENA_STAT_GLOBAL_ENTRY(tx_timeout), + ENA_STAT_GLOBAL_ENTRY(suspend), + ENA_STAT_GLOBAL_ENTRY(resume), + ENA_STAT_GLOBAL_ENTRY(wd_expired), + ENA_STAT_GLOBAL_ENTRY(interface_up), + ENA_STAT_GLOBAL_ENTRY(interface_down), + ENA_STAT_GLOBAL_ENTRY(admin_q_pause), +}; + +static const struct ena_stats ena_stats_eni_strings[] = { + ENA_STAT_ENI_ENTRY(bw_in_allowance_exceeded), + ENA_STAT_ENI_ENTRY(bw_out_allowance_exceeded), + ENA_STAT_ENI_ENTRY(pps_allowance_exceeded), + ENA_STAT_ENI_ENTRY(conntrack_allowance_exceeded), + ENA_STAT_ENI_ENTRY(linklocal_allowance_exceeded), +}; + +static const struct ena_stats ena_stats_tx_strings[] = { + ENA_STAT_TX_ENTRY(cnt), + ENA_STAT_TX_ENTRY(bytes), + ENA_STAT_TX_ENTRY(queue_stop), + ENA_STAT_TX_ENTRY(queue_wakeup), + ENA_STAT_TX_ENTRY(dma_mapping_err), + ENA_STAT_TX_ENTRY(linearize), + ENA_STAT_TX_ENTRY(linearize_failed), + ENA_STAT_TX_ENTRY(napi_comp), + ENA_STAT_TX_ENTRY(tx_poll), + ENA_STAT_TX_ENTRY(doorbells), + ENA_STAT_TX_ENTRY(prepare_ctx_err), + ENA_STAT_TX_ENTRY(bad_req_id), + ENA_STAT_TX_ENTRY(llq_buffer_copy), + ENA_STAT_TX_ENTRY(missed_tx), + ENA_STAT_TX_ENTRY(unmask_interrupt), +}; + +static const struct ena_stats ena_stats_rx_strings[] = { + ENA_STAT_RX_ENTRY(cnt), + ENA_STAT_RX_ENTRY(bytes), + ENA_STAT_RX_ENTRY(rx_copybreak_pkt), + ENA_STAT_RX_ENTRY(csum_good), + ENA_STAT_RX_ENTRY(refil_partial), + ENA_STAT_RX_ENTRY(csum_bad), + ENA_STAT_RX_ENTRY(page_alloc_fail), + ENA_STAT_RX_ENTRY(skb_alloc_fail), + ENA_STAT_RX_ENTRY(dma_mapping_err), + ENA_STAT_RX_ENTRY(bad_desc_num), + ENA_STAT_RX_ENTRY(bad_req_id), + ENA_STAT_RX_ENTRY(empty_rx_ring), + ENA_STAT_RX_ENTRY(csum_unchecked), + ENA_STAT_RX_ENTRY(xdp_aborted), + ENA_STAT_RX_ENTRY(xdp_drop), + ENA_STAT_RX_ENTRY(xdp_pass), + ENA_STAT_RX_ENTRY(xdp_tx), + ENA_STAT_RX_ENTRY(xdp_invalid), + ENA_STAT_RX_ENTRY(xdp_redirect), +}; + +static const struct ena_stats ena_stats_ena_com_strings[] = { + ENA_STAT_ENA_COM_ENTRY(aborted_cmd), + ENA_STAT_ENA_COM_ENTRY(submitted_cmd), + ENA_STAT_ENA_COM_ENTRY(completed_cmd), + ENA_STAT_ENA_COM_ENTRY(out_of_space), + ENA_STAT_ENA_COM_ENTRY(no_completion), +}; + +#define ENA_STATS_ARRAY_GLOBAL ARRAY_SIZE(ena_stats_global_strings) +#define ENA_STATS_ARRAY_TX ARRAY_SIZE(ena_stats_tx_strings) +#define ENA_STATS_ARRAY_RX ARRAY_SIZE(ena_stats_rx_strings) +#define ENA_STATS_ARRAY_ENA_COM ARRAY_SIZE(ena_stats_ena_com_strings) +#define ENA_STATS_ARRAY_ENI(adapter) ARRAY_SIZE(ena_stats_eni_strings) + +static void ena_safe_update_stat(u64 *src, u64 *dst, + struct u64_stats_sync *syncp) +{ + unsigned int start; + + do { + start = u64_stats_fetch_begin(syncp); + *(dst) = *src; + } while (u64_stats_fetch_retry(syncp, start)); +} + +static void ena_queue_stats(struct ena_adapter *adapter, u64 **data) +{ + const struct ena_stats *ena_stats; + struct ena_ring *ring; + + u64 *ptr; + int i, j; + + for (i = 0; i < adapter->num_io_queues + adapter->xdp_num_queues; i++) { + /* Tx stats */ + ring = &adapter->tx_ring[i]; + + for (j = 0; j < ENA_STATS_ARRAY_TX; j++) { + ena_stats = &ena_stats_tx_strings[j]; + + ptr = (u64 *)&ring->tx_stats + ena_stats->stat_offset; + + ena_safe_update_stat(ptr, (*data)++, &ring->syncp); + } + /* XDP TX queues don't have a RX queue counterpart */ + if (!ENA_IS_XDP_INDEX(adapter, i)) { + /* Rx stats */ + ring = &adapter->rx_ring[i]; + + for (j = 0; j < ENA_STATS_ARRAY_RX; j++) { + ena_stats = &ena_stats_rx_strings[j]; + + ptr = (u64 *)&ring->rx_stats + + ena_stats->stat_offset; + + ena_safe_update_stat(ptr, (*data)++, &ring->syncp); + } + } + } +} + +static void ena_dev_admin_queue_stats(struct ena_adapter *adapter, u64 **data) +{ + const struct ena_stats *ena_stats; + u64 *ptr; + int i; + + for (i = 0; i < ENA_STATS_ARRAY_ENA_COM; i++) { + ena_stats = &ena_stats_ena_com_strings[i]; + + ptr = (u64 *)&adapter->ena_dev->admin_queue.stats + + ena_stats->stat_offset; + + *(*data)++ = *ptr; + } +} + +static void ena_get_stats(struct ena_adapter *adapter, + u64 *data, + bool eni_stats_needed) +{ + const struct ena_stats *ena_stats; + u64 *ptr; + int i; + + for (i = 0; i < ENA_STATS_ARRAY_GLOBAL; i++) { + ena_stats = &ena_stats_global_strings[i]; + + ptr = (u64 *)&adapter->dev_stats + ena_stats->stat_offset; + + ena_safe_update_stat(ptr, data++, &adapter->syncp); + } + + if (eni_stats_needed) { + ena_update_hw_stats(adapter); + for (i = 0; i < ENA_STATS_ARRAY_ENI(adapter); i++) { + ena_stats = &ena_stats_eni_strings[i]; + + ptr = (u64 *)&adapter->eni_stats + + ena_stats->stat_offset; + + ena_safe_update_stat(ptr, data++, &adapter->syncp); + } + } + + ena_queue_stats(adapter, &data); + ena_dev_admin_queue_stats(adapter, &data); +} + +static void ena_get_ethtool_stats(struct net_device *netdev, + struct ethtool_stats *stats, + u64 *data) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct ena_com_dev *dev = adapter->ena_dev; + + ena_get_stats(adapter, data, ena_com_get_cap(dev, ENA_ADMIN_ENI_STATS)); +} + +static int ena_get_sw_stats_count(struct ena_adapter *adapter) +{ + return adapter->num_io_queues * (ENA_STATS_ARRAY_TX + ENA_STATS_ARRAY_RX) + + adapter->xdp_num_queues * ENA_STATS_ARRAY_TX + + ENA_STATS_ARRAY_GLOBAL + ENA_STATS_ARRAY_ENA_COM; +} + +static int ena_get_hw_stats_count(struct ena_adapter *adapter) +{ + bool supported = ena_com_get_cap(adapter->ena_dev, ENA_ADMIN_ENI_STATS); + + return ENA_STATS_ARRAY_ENI(adapter) * supported; +} + +int ena_get_sset_count(struct net_device *netdev, int sset) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + switch (sset) { + case ETH_SS_STATS: + return ena_get_sw_stats_count(adapter) + + ena_get_hw_stats_count(adapter); + } + + return -EOPNOTSUPP; +} + +static void ena_queue_strings(struct ena_adapter *adapter, u8 **data) +{ + const struct ena_stats *ena_stats; + bool is_xdp; + int i, j; + + for (i = 0; i < adapter->num_io_queues + adapter->xdp_num_queues; i++) { + is_xdp = ENA_IS_XDP_INDEX(adapter, i); + /* Tx stats */ + for (j = 0; j < ENA_STATS_ARRAY_TX; j++) { + ena_stats = &ena_stats_tx_strings[j]; + + ethtool_sprintf(data, + "queue_%u_%s_%s", i, + is_xdp ? "xdp_tx" : "tx", + ena_stats->name); + } + + if (!is_xdp) { + /* RX stats, in XDP there isn't a RX queue + * counterpart + */ + for (j = 0; j < ENA_STATS_ARRAY_RX; j++) { + ena_stats = &ena_stats_rx_strings[j]; + + ethtool_sprintf(data, + "queue_%u_rx_%s", i, + ena_stats->name); + } + } + } +} + +static void ena_com_dev_strings(u8 **data) +{ + const struct ena_stats *ena_stats; + int i; + + for (i = 0; i < ENA_STATS_ARRAY_ENA_COM; i++) { + ena_stats = &ena_stats_ena_com_strings[i]; + + ethtool_sprintf(data, + "ena_admin_q_%s", ena_stats->name); + } +} + +static void ena_get_strings(struct ena_adapter *adapter, + u8 *data, + bool eni_stats_needed) +{ + const struct ena_stats *ena_stats; + int i; + + for (i = 0; i < ENA_STATS_ARRAY_GLOBAL; i++) { + ena_stats = &ena_stats_global_strings[i]; + ethtool_sprintf(&data, ena_stats->name); + } + + if (eni_stats_needed) { + for (i = 0; i < ENA_STATS_ARRAY_ENI(adapter); i++) { + ena_stats = &ena_stats_eni_strings[i]; + ethtool_sprintf(&data, ena_stats->name); + } + } + + ena_queue_strings(adapter, &data); + ena_com_dev_strings(&data); +} + +static void ena_get_ethtool_strings(struct net_device *netdev, + u32 sset, + u8 *data) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct ena_com_dev *dev = adapter->ena_dev; + + switch (sset) { + case ETH_SS_STATS: + ena_get_strings(adapter, data, ena_com_get_cap(dev, ENA_ADMIN_ENI_STATS)); + break; + } +} + +static int ena_get_link_ksettings(struct net_device *netdev, + struct ethtool_link_ksettings *link_ksettings) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct ena_com_dev *ena_dev = adapter->ena_dev; + struct ena_admin_get_feature_link_desc *link; + struct ena_admin_get_feat_resp feat_resp; + int rc; + + rc = ena_com_get_link_params(ena_dev, &feat_resp); + if (rc) + return rc; + + link = &feat_resp.u.link; + link_ksettings->base.speed = link->speed; + + if (link->flags & ENA_ADMIN_GET_FEATURE_LINK_DESC_AUTONEG_MASK) { + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, Autoneg); + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, Autoneg); + } + + link_ksettings->base.autoneg = + (link->flags & ENA_ADMIN_GET_FEATURE_LINK_DESC_AUTONEG_MASK) ? + AUTONEG_ENABLE : AUTONEG_DISABLE; + + link_ksettings->base.duplex = DUPLEX_FULL; + + return 0; +} + +static int ena_get_coalesce(struct net_device *net_dev, + struct ethtool_coalesce *coalesce, + struct kernel_ethtool_coalesce *kernel_coal, + struct netlink_ext_ack *extack) +{ + struct ena_adapter *adapter = netdev_priv(net_dev); + struct ena_com_dev *ena_dev = adapter->ena_dev; + + if (!ena_com_interrupt_moderation_supported(ena_dev)) + return -EOPNOTSUPP; + + coalesce->tx_coalesce_usecs = + ena_com_get_nonadaptive_moderation_interval_tx(ena_dev) * + ena_dev->intr_delay_resolution; + + coalesce->rx_coalesce_usecs = + ena_com_get_nonadaptive_moderation_interval_rx(ena_dev) + * ena_dev->intr_delay_resolution; + + coalesce->use_adaptive_rx_coalesce = + ena_com_get_adaptive_moderation_enabled(ena_dev); + + return 0; +} + +static void ena_update_tx_rings_nonadaptive_intr_moderation(struct ena_adapter *adapter) +{ + unsigned int val; + int i; + + val = ena_com_get_nonadaptive_moderation_interval_tx(adapter->ena_dev); + + for (i = 0; i < adapter->num_io_queues; i++) + adapter->tx_ring[i].smoothed_interval = val; +} + +static void ena_update_rx_rings_nonadaptive_intr_moderation(struct ena_adapter *adapter) +{ + unsigned int val; + int i; + + val = ena_com_get_nonadaptive_moderation_interval_rx(adapter->ena_dev); + + for (i = 0; i < adapter->num_io_queues; i++) + adapter->rx_ring[i].smoothed_interval = val; +} + +static int ena_set_coalesce(struct net_device *net_dev, + struct ethtool_coalesce *coalesce, + struct kernel_ethtool_coalesce *kernel_coal, + struct netlink_ext_ack *extack) +{ + struct ena_adapter *adapter = netdev_priv(net_dev); + struct ena_com_dev *ena_dev = adapter->ena_dev; + int rc; + + if (!ena_com_interrupt_moderation_supported(ena_dev)) + return -EOPNOTSUPP; + + rc = ena_com_update_nonadaptive_moderation_interval_tx(ena_dev, + coalesce->tx_coalesce_usecs); + if (rc) + return rc; + + ena_update_tx_rings_nonadaptive_intr_moderation(adapter); + + rc = ena_com_update_nonadaptive_moderation_interval_rx(ena_dev, + coalesce->rx_coalesce_usecs); + if (rc) + return rc; + + ena_update_rx_rings_nonadaptive_intr_moderation(adapter); + + if (coalesce->use_adaptive_rx_coalesce && + !ena_com_get_adaptive_moderation_enabled(ena_dev)) + ena_com_enable_adaptive_moderation(ena_dev); + + if (!coalesce->use_adaptive_rx_coalesce && + ena_com_get_adaptive_moderation_enabled(ena_dev)) + ena_com_disable_adaptive_moderation(ena_dev); + + return 0; +} + +static u32 ena_get_msglevel(struct net_device *netdev) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + return adapter->msg_enable; +} + +static void ena_set_msglevel(struct net_device *netdev, u32 value) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + adapter->msg_enable = value; +} + +static void ena_get_drvinfo(struct net_device *dev, + struct ethtool_drvinfo *info) +{ + struct ena_adapter *adapter = netdev_priv(dev); + + strscpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver)); + strscpy(info->bus_info, pci_name(adapter->pdev), + sizeof(info->bus_info)); +} + +static void ena_get_ringparam(struct net_device *netdev, + struct ethtool_ringparam *ring, + struct kernel_ethtool_ringparam *kernel_ring, + struct netlink_ext_ack *extack) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + ring->tx_max_pending = adapter->max_tx_ring_size; + ring->rx_max_pending = adapter->max_rx_ring_size; + if (adapter->ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + bool large_llq_supported = adapter->large_llq_header_supported; + + kernel_ring->tx_push = true; + kernel_ring->tx_push_buf_len = adapter->ena_dev->tx_max_header_size; + if (large_llq_supported) + kernel_ring->tx_push_buf_max_len = ENA_LLQ_LARGE_HEADER; + else + kernel_ring->tx_push_buf_max_len = ENA_LLQ_HEADER; + } else { + kernel_ring->tx_push = false; + kernel_ring->tx_push_buf_max_len = 0; + kernel_ring->tx_push_buf_len = 0; + } + + ring->tx_pending = adapter->tx_ring[0].ring_size; + ring->rx_pending = adapter->rx_ring[0].ring_size; +} + +static int ena_set_ringparam(struct net_device *netdev, + struct ethtool_ringparam *ring, + struct kernel_ethtool_ringparam *kernel_ring, + struct netlink_ext_ack *extack) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + u32 new_tx_size, new_rx_size, new_tx_push_buf_len; + bool changed = false; + + new_tx_size = ring->tx_pending < ENA_MIN_RING_SIZE ? + ENA_MIN_RING_SIZE : ring->tx_pending; + new_tx_size = rounddown_pow_of_two(new_tx_size); + + new_rx_size = ring->rx_pending < ENA_MIN_RING_SIZE ? + ENA_MIN_RING_SIZE : ring->rx_pending; + new_rx_size = rounddown_pow_of_two(new_rx_size); + + changed |= new_tx_size != adapter->requested_tx_ring_size || + new_rx_size != adapter->requested_rx_ring_size; + + /* This value is ignored if LLQ is not supported */ + new_tx_push_buf_len = adapter->ena_dev->tx_max_header_size; + + if ((adapter->ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) != + kernel_ring->tx_push) { + NL_SET_ERR_MSG_MOD(extack, "Push mode state cannot be modified"); + return -EINVAL; + } + + /* Validate that the push buffer is supported on the underlying device */ + if (kernel_ring->tx_push_buf_len) { + enum ena_admin_placement_policy_type placement; + + new_tx_push_buf_len = kernel_ring->tx_push_buf_len; + + placement = adapter->ena_dev->tx_mem_queue_type; + if (placement == ENA_ADMIN_PLACEMENT_POLICY_HOST) + return -EOPNOTSUPP; + + if (new_tx_push_buf_len != ENA_LLQ_HEADER && + new_tx_push_buf_len != ENA_LLQ_LARGE_HEADER) { + bool large_llq_sup = adapter->large_llq_header_supported; + char large_llq_size_str[40]; + + snprintf(large_llq_size_str, 40, ", %lu", ENA_LLQ_LARGE_HEADER); + + NL_SET_ERR_MSG_FMT_MOD(extack, + "Supported tx push buff values: [%lu%s]", + ENA_LLQ_HEADER, + large_llq_sup ? large_llq_size_str : ""); + + return -EINVAL; + } + + changed |= new_tx_push_buf_len != adapter->ena_dev->tx_max_header_size; + } + + if (!changed) + return 0; + + return ena_update_queue_params(adapter, new_tx_size, new_rx_size, + new_tx_push_buf_len); +} + +static u32 ena_flow_hash_to_flow_type(u16 hash_fields) +{ + u32 data = 0; + + if (hash_fields & ENA_ADMIN_RSS_L2_DA) + data |= RXH_L2DA; + + if (hash_fields & ENA_ADMIN_RSS_L3_DA) + data |= RXH_IP_DST; + + if (hash_fields & ENA_ADMIN_RSS_L3_SA) + data |= RXH_IP_SRC; + + if (hash_fields & ENA_ADMIN_RSS_L4_DP) + data |= RXH_L4_B_2_3; + + if (hash_fields & ENA_ADMIN_RSS_L4_SP) + data |= RXH_L4_B_0_1; + + return data; +} + +static u16 ena_flow_data_to_flow_hash(u32 hash_fields) +{ + u16 data = 0; + + if (hash_fields & RXH_L2DA) + data |= ENA_ADMIN_RSS_L2_DA; + + if (hash_fields & RXH_IP_DST) + data |= ENA_ADMIN_RSS_L3_DA; + + if (hash_fields & RXH_IP_SRC) + data |= ENA_ADMIN_RSS_L3_SA; + + if (hash_fields & RXH_L4_B_2_3) + data |= ENA_ADMIN_RSS_L4_DP; + + if (hash_fields & RXH_L4_B_0_1) + data |= ENA_ADMIN_RSS_L4_SP; + + return data; +} + +static int ena_get_rss_hash(struct ena_com_dev *ena_dev, + struct ethtool_rxnfc *cmd) +{ + enum ena_admin_flow_hash_proto proto; + u16 hash_fields; + int rc; + + cmd->data = 0; + + switch (cmd->flow_type) { + case TCP_V4_FLOW: + proto = ENA_ADMIN_RSS_TCP4; + break; + case UDP_V4_FLOW: + proto = ENA_ADMIN_RSS_UDP4; + break; + case TCP_V6_FLOW: + proto = ENA_ADMIN_RSS_TCP6; + break; + case UDP_V6_FLOW: + proto = ENA_ADMIN_RSS_UDP6; + break; + case IPV4_FLOW: + proto = ENA_ADMIN_RSS_IP4; + break; + case IPV6_FLOW: + proto = ENA_ADMIN_RSS_IP6; + break; + case ETHER_FLOW: + proto = ENA_ADMIN_RSS_NOT_IP; + break; + case AH_V4_FLOW: + case ESP_V4_FLOW: + case AH_V6_FLOW: + case ESP_V6_FLOW: + case SCTP_V4_FLOW: + case AH_ESP_V4_FLOW: + return -EOPNOTSUPP; + default: + return -EINVAL; + } + + rc = ena_com_get_hash_ctrl(ena_dev, proto, &hash_fields); + if (rc) + return rc; + + cmd->data = ena_flow_hash_to_flow_type(hash_fields); + + return 0; +} + +static int ena_set_rss_hash(struct ena_com_dev *ena_dev, + struct ethtool_rxnfc *cmd) +{ + enum ena_admin_flow_hash_proto proto; + u16 hash_fields; + + switch (cmd->flow_type) { + case TCP_V4_FLOW: + proto = ENA_ADMIN_RSS_TCP4; + break; + case UDP_V4_FLOW: + proto = ENA_ADMIN_RSS_UDP4; + break; + case TCP_V6_FLOW: + proto = ENA_ADMIN_RSS_TCP6; + break; + case UDP_V6_FLOW: + proto = ENA_ADMIN_RSS_UDP6; + break; + case IPV4_FLOW: + proto = ENA_ADMIN_RSS_IP4; + break; + case IPV6_FLOW: + proto = ENA_ADMIN_RSS_IP6; + break; + case ETHER_FLOW: + proto = ENA_ADMIN_RSS_NOT_IP; + break; + case AH_V4_FLOW: + case ESP_V4_FLOW: + case AH_V6_FLOW: + case ESP_V6_FLOW: + case SCTP_V4_FLOW: + case AH_ESP_V4_FLOW: + return -EOPNOTSUPP; + default: + return -EINVAL; + } + + hash_fields = ena_flow_data_to_flow_hash(cmd->data); + + return ena_com_fill_hash_ctrl(ena_dev, proto, hash_fields); +} + +static int ena_set_rxnfc(struct net_device *netdev, struct ethtool_rxnfc *info) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + int rc = 0; + + switch (info->cmd) { + case ETHTOOL_SRXFH: + rc = ena_set_rss_hash(adapter->ena_dev, info); + break; + case ETHTOOL_SRXCLSRLDEL: + case ETHTOOL_SRXCLSRLINS: + default: + netif_err(adapter, drv, netdev, + "Command parameter %d is not supported\n", info->cmd); + rc = -EOPNOTSUPP; + } + + return rc; +} + +static int ena_get_rxnfc(struct net_device *netdev, struct ethtool_rxnfc *info, + u32 *rules) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + int rc = 0; + + switch (info->cmd) { + case ETHTOOL_GRXRINGS: + info->data = adapter->num_io_queues; + rc = 0; + break; + case ETHTOOL_GRXFH: + rc = ena_get_rss_hash(adapter->ena_dev, info); + break; + case ETHTOOL_GRXCLSRLCNT: + case ETHTOOL_GRXCLSRULE: + case ETHTOOL_GRXCLSRLALL: + default: + netif_err(adapter, drv, netdev, + "Command parameter %d is not supported\n", info->cmd); + rc = -EOPNOTSUPP; + } + + return rc; +} + +static u32 ena_get_rxfh_indir_size(struct net_device *netdev) +{ + return ENA_RX_RSS_TABLE_SIZE; +} + +static u32 ena_get_rxfh_key_size(struct net_device *netdev) +{ + return ENA_HASH_KEY_SIZE; +} + +static int ena_indirection_table_set(struct ena_adapter *adapter, + const u32 *indir) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int i, rc; + + for (i = 0; i < ENA_RX_RSS_TABLE_SIZE; i++) { + rc = ena_com_indirect_table_fill_entry(ena_dev, + i, + ENA_IO_RXQ_IDX(indir[i])); + if (unlikely(rc)) { + netif_err(adapter, drv, adapter->netdev, + "Cannot fill indirect table (index is too large)\n"); + return rc; + } + } + + rc = ena_com_indirect_table_set(ena_dev); + if (rc) { + netif_err(adapter, drv, adapter->netdev, + "Cannot set indirect table\n"); + return rc == -EPERM ? -EOPNOTSUPP : rc; + } + return rc; +} + +static int ena_indirection_table_get(struct ena_adapter *adapter, u32 *indir) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int i, rc; + + if (!indir) + return 0; + + rc = ena_com_indirect_table_get(ena_dev, indir); + if (rc) + return rc; + + /* Our internal representation of the indices is: even indices + * for Tx and uneven indices for Rx. We need to convert the Rx + * indices to be consecutive + */ + for (i = 0; i < ENA_RX_RSS_TABLE_SIZE; i++) + indir[i] = ENA_IO_RXQ_IDX_TO_COMBINED_IDX(indir[i]); + + return rc; +} + +static int ena_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key, + u8 *hfunc) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + enum ena_admin_hash_functions ena_func; + u8 func; + int rc; + + rc = ena_indirection_table_get(adapter, indir); + if (rc) + return rc; + + /* We call this function in order to check if the device + * supports getting/setting the hash function. + */ + rc = ena_com_get_hash_function(adapter->ena_dev, &ena_func); + if (rc) { + if (rc == -EOPNOTSUPP) + rc = 0; + + return rc; + } + + rc = ena_com_get_hash_key(adapter->ena_dev, key); + if (rc) + return rc; + + switch (ena_func) { + case ENA_ADMIN_TOEPLITZ: + func = ETH_RSS_HASH_TOP; + break; + case ENA_ADMIN_CRC32: + func = ETH_RSS_HASH_CRC32; + break; + default: + netif_err(adapter, drv, netdev, + "Command parameter is not supported\n"); + return -EOPNOTSUPP; + } + + if (hfunc) + *hfunc = func; + + return 0; +} + +static int ena_set_rxfh(struct net_device *netdev, const u32 *indir, + const u8 *key, const u8 hfunc) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct ena_com_dev *ena_dev = adapter->ena_dev; + enum ena_admin_hash_functions func = 0; + int rc; + + if (indir) { + rc = ena_indirection_table_set(adapter, indir); + if (rc) + return rc; + } + + switch (hfunc) { + case ETH_RSS_HASH_NO_CHANGE: + func = ena_com_get_current_hash_function(ena_dev); + break; + case ETH_RSS_HASH_TOP: + func = ENA_ADMIN_TOEPLITZ; + break; + case ETH_RSS_HASH_CRC32: + func = ENA_ADMIN_CRC32; + break; + default: + netif_err(adapter, drv, netdev, "Unsupported hfunc %d\n", + hfunc); + return -EOPNOTSUPP; + } + + if (key || func) { + rc = ena_com_fill_hash_function(ena_dev, func, key, + ENA_HASH_KEY_SIZE, + 0xFFFFFFFF); + if (unlikely(rc)) { + netif_err(adapter, drv, netdev, "Cannot fill key\n"); + return rc == -EPERM ? -EOPNOTSUPP : rc; + } + } + + return 0; +} + +static void ena_get_channels(struct net_device *netdev, + struct ethtool_channels *channels) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + channels->max_combined = adapter->max_num_io_queues; + channels->combined_count = adapter->num_io_queues; +} + +static int ena_set_channels(struct net_device *netdev, + struct ethtool_channels *channels) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + u32 count = channels->combined_count; + /* The check for max value is already done in ethtool */ + if (count < ENA_MIN_NUM_IO_QUEUES) + return -EINVAL; + + if (!ena_xdp_legal_queue_count(adapter, count)) { + if (ena_xdp_present(adapter)) + return -EINVAL; + + xdp_clear_features_flag(netdev); + } else { + xdp_set_features_flag(netdev, + NETDEV_XDP_ACT_BASIC | + NETDEV_XDP_ACT_REDIRECT); + } + + return ena_update_queue_count(adapter, count); +} + +static int ena_get_tunable(struct net_device *netdev, + const struct ethtool_tunable *tuna, void *data) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + int ret = 0; + + switch (tuna->id) { + case ETHTOOL_RX_COPYBREAK: + *(u32 *)data = adapter->rx_copybreak; + break; + default: + ret = -EINVAL; + break; + } + + return ret; +} + +static int ena_set_tunable(struct net_device *netdev, + const struct ethtool_tunable *tuna, + const void *data) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + int ret = 0; + u32 len; + + switch (tuna->id) { + case ETHTOOL_RX_COPYBREAK: + len = *(u32 *)data; + ret = ena_set_rx_copybreak(adapter, len); + break; + default: + ret = -EINVAL; + break; + } + + return ret; +} + +static const struct ethtool_ops ena_ethtool_ops = { + .supported_coalesce_params = ETHTOOL_COALESCE_USECS | + ETHTOOL_COALESCE_USE_ADAPTIVE_RX, + .supported_ring_params = ETHTOOL_RING_USE_TX_PUSH_BUF_LEN | + ETHTOOL_RING_USE_TX_PUSH, + .get_link_ksettings = ena_get_link_ksettings, + .get_drvinfo = ena_get_drvinfo, + .get_msglevel = ena_get_msglevel, + .set_msglevel = ena_set_msglevel, + .get_link = ethtool_op_get_link, + .get_coalesce = ena_get_coalesce, + .set_coalesce = ena_set_coalesce, + .get_ringparam = ena_get_ringparam, + .set_ringparam = ena_set_ringparam, + .get_sset_count = ena_get_sset_count, + .get_strings = ena_get_ethtool_strings, + .get_ethtool_stats = ena_get_ethtool_stats, + .get_rxnfc = ena_get_rxnfc, + .set_rxnfc = ena_set_rxnfc, + .get_rxfh_indir_size = ena_get_rxfh_indir_size, + .get_rxfh_key_size = ena_get_rxfh_key_size, + .get_rxfh = ena_get_rxfh, + .set_rxfh = ena_set_rxfh, + .get_channels = ena_get_channels, + .set_channels = ena_set_channels, + .get_tunable = ena_get_tunable, + .set_tunable = ena_set_tunable, + .get_ts_info = ethtool_op_get_ts_info, +}; + +void ena_set_ethtool_ops(struct net_device *netdev) +{ + netdev->ethtool_ops = &ena_ethtool_ops; +} + +static void ena_dump_stats_ex(struct ena_adapter *adapter, u8 *buf) +{ + struct net_device *netdev = adapter->netdev; + u8 *strings_buf; + u64 *data_buf; + int strings_num; + int i, rc; + + strings_num = ena_get_sw_stats_count(adapter); + if (strings_num <= 0) { + netif_err(adapter, drv, netdev, "Can't get stats num\n"); + return; + } + + strings_buf = devm_kcalloc(&adapter->pdev->dev, + ETH_GSTRING_LEN, strings_num, + GFP_ATOMIC); + if (!strings_buf) { + netif_err(adapter, drv, netdev, + "Failed to allocate strings_buf\n"); + return; + } + + data_buf = devm_kcalloc(&adapter->pdev->dev, + strings_num, sizeof(u64), + GFP_ATOMIC); + if (!data_buf) { + netif_err(adapter, drv, netdev, + "Failed to allocate data buf\n"); + devm_kfree(&adapter->pdev->dev, strings_buf); + return; + } + + ena_get_strings(adapter, strings_buf, false); + ena_get_stats(adapter, data_buf, false); + + /* If there is a buffer, dump stats, otherwise print them to dmesg */ + if (buf) + for (i = 0; i < strings_num; i++) { + rc = snprintf(buf, ETH_GSTRING_LEN + sizeof(u64), + "%s %llu\n", + strings_buf + i * ETH_GSTRING_LEN, + data_buf[i]); + buf += rc; + } + else + for (i = 0; i < strings_num; i++) + netif_err(adapter, drv, netdev, "%s: %llu\n", + strings_buf + i * ETH_GSTRING_LEN, + data_buf[i]); + + devm_kfree(&adapter->pdev->dev, strings_buf); + devm_kfree(&adapter->pdev->dev, data_buf); +} + +void ena_dump_stats_to_buf(struct ena_adapter *adapter, u8 *buf) +{ + if (!buf) + return; + + ena_dump_stats_ex(adapter, buf); +} + +void ena_dump_stats_to_dmesg(struct ena_adapter *adapter) +{ + ena_dump_stats_ex(adapter, NULL); +} diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c new file mode 100644 index 0000000000..14e41eb577 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c @@ -0,0 +1,4832 @@ +// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#ifdef CONFIG_RFS_ACCEL +#include <linux/cpu_rmap.h> +#endif /* CONFIG_RFS_ACCEL */ +#include <linux/ethtool.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/numa.h> +#include <linux/pci.h> +#include <linux/utsname.h> +#include <linux/version.h> +#include <linux/vmalloc.h> +#include <net/ip.h> + +#include "ena_netdev.h" +#include <linux/bpf_trace.h> +#include "ena_pci_id_tbl.h" + +MODULE_AUTHOR("Amazon.com, Inc. or its affiliates"); +MODULE_DESCRIPTION(DEVICE_NAME); +MODULE_LICENSE("GPL"); + +/* Time in jiffies before concluding the transmitter is hung. */ +#define TX_TIMEOUT (5 * HZ) + +#define ENA_MAX_RINGS min_t(unsigned int, ENA_MAX_NUM_IO_QUEUES, num_possible_cpus()) + +#define DEFAULT_MSG_ENABLE (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | \ + NETIF_MSG_TX_DONE | NETIF_MSG_TX_ERR | NETIF_MSG_RX_ERR) + +static struct ena_aenq_handlers aenq_handlers; + +static struct workqueue_struct *ena_wq; + +MODULE_DEVICE_TABLE(pci, ena_pci_tbl); + +static int ena_rss_init_default(struct ena_adapter *adapter); +static void check_for_admin_com_state(struct ena_adapter *adapter); +static void ena_destroy_device(struct ena_adapter *adapter, bool graceful); +static int ena_restore_device(struct ena_adapter *adapter); + +static void ena_init_io_rings(struct ena_adapter *adapter, + int first_index, int count); +static void ena_init_napi_in_range(struct ena_adapter *adapter, int first_index, + int count); +static void ena_del_napi_in_range(struct ena_adapter *adapter, int first_index, + int count); +static int ena_setup_tx_resources(struct ena_adapter *adapter, int qid); +static int ena_setup_tx_resources_in_range(struct ena_adapter *adapter, + int first_index, + int count); +static int ena_create_io_tx_queue(struct ena_adapter *adapter, int qid); +static void ena_free_tx_resources(struct ena_adapter *adapter, int qid); +static int ena_clean_xdp_irq(struct ena_ring *xdp_ring, u32 budget); +static void ena_destroy_all_tx_queues(struct ena_adapter *adapter); +static void ena_free_all_io_tx_resources(struct ena_adapter *adapter); +static void ena_napi_disable_in_range(struct ena_adapter *adapter, + int first_index, int count); +static void ena_napi_enable_in_range(struct ena_adapter *adapter, + int first_index, int count); +static int ena_up(struct ena_adapter *adapter); +static void ena_down(struct ena_adapter *adapter); +static void ena_unmask_interrupt(struct ena_ring *tx_ring, + struct ena_ring *rx_ring); +static void ena_update_ring_numa_node(struct ena_ring *tx_ring, + struct ena_ring *rx_ring); +static void ena_unmap_tx_buff(struct ena_ring *tx_ring, + struct ena_tx_buffer *tx_info); +static int ena_create_io_tx_queues_in_range(struct ena_adapter *adapter, + int first_index, int count); +static void ena_free_all_io_tx_resources_in_range(struct ena_adapter *adapter, + int first_index, int count); + +/* Increase a stat by cnt while holding syncp seqlock on 32bit machines */ +static void ena_increase_stat(u64 *statp, u64 cnt, + struct u64_stats_sync *syncp) +{ + u64_stats_update_begin(syncp); + (*statp) += cnt; + u64_stats_update_end(syncp); +} + +static void ena_ring_tx_doorbell(struct ena_ring *tx_ring) +{ + ena_com_write_sq_doorbell(tx_ring->ena_com_io_sq); + ena_increase_stat(&tx_ring->tx_stats.doorbells, 1, &tx_ring->syncp); +} + +static void ena_tx_timeout(struct net_device *dev, unsigned int txqueue) +{ + struct ena_adapter *adapter = netdev_priv(dev); + + /* Change the state of the device to trigger reset + * Check that we are not in the middle or a trigger already + */ + + if (test_and_set_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags)) + return; + + ena_reset_device(adapter, ENA_REGS_RESET_OS_NETDEV_WD); + ena_increase_stat(&adapter->dev_stats.tx_timeout, 1, &adapter->syncp); + + netif_err(adapter, tx_err, dev, "Transmit time out\n"); +} + +static void update_rx_ring_mtu(struct ena_adapter *adapter, int mtu) +{ + int i; + + for (i = 0; i < adapter->num_io_queues; i++) + adapter->rx_ring[i].mtu = mtu; +} + +static int ena_change_mtu(struct net_device *dev, int new_mtu) +{ + struct ena_adapter *adapter = netdev_priv(dev); + int ret; + + ret = ena_com_set_dev_mtu(adapter->ena_dev, new_mtu); + if (!ret) { + netif_dbg(adapter, drv, dev, "Set MTU to %d\n", new_mtu); + update_rx_ring_mtu(adapter, new_mtu); + dev->mtu = new_mtu; + } else { + netif_err(adapter, drv, dev, "Failed to set MTU to %d\n", + new_mtu); + } + + return ret; +} + +static int ena_xmit_common(struct net_device *dev, + struct ena_ring *ring, + struct ena_tx_buffer *tx_info, + struct ena_com_tx_ctx *ena_tx_ctx, + u16 next_to_use, + u32 bytes) +{ + struct ena_adapter *adapter = netdev_priv(dev); + int rc, nb_hw_desc; + + if (unlikely(ena_com_is_doorbell_needed(ring->ena_com_io_sq, + ena_tx_ctx))) { + netif_dbg(adapter, tx_queued, dev, + "llq tx max burst size of queue %d achieved, writing doorbell to send burst\n", + ring->qid); + ena_ring_tx_doorbell(ring); + } + + /* prepare the packet's descriptors to dma engine */ + rc = ena_com_prepare_tx(ring->ena_com_io_sq, ena_tx_ctx, + &nb_hw_desc); + + /* In case there isn't enough space in the queue for the packet, + * we simply drop it. All other failure reasons of + * ena_com_prepare_tx() are fatal and therefore require a device reset. + */ + if (unlikely(rc)) { + netif_err(adapter, tx_queued, dev, + "Failed to prepare tx bufs\n"); + ena_increase_stat(&ring->tx_stats.prepare_ctx_err, 1, + &ring->syncp); + if (rc != -ENOMEM) + ena_reset_device(adapter, + ENA_REGS_RESET_DRIVER_INVALID_STATE); + return rc; + } + + u64_stats_update_begin(&ring->syncp); + ring->tx_stats.cnt++; + ring->tx_stats.bytes += bytes; + u64_stats_update_end(&ring->syncp); + + tx_info->tx_descs = nb_hw_desc; + tx_info->last_jiffies = jiffies; + tx_info->print_once = 0; + + ring->next_to_use = ENA_TX_RING_IDX_NEXT(next_to_use, + ring->ring_size); + return 0; +} + +/* This is the XDP napi callback. XDP queues use a separate napi callback + * than Rx/Tx queues. + */ +static int ena_xdp_io_poll(struct napi_struct *napi, int budget) +{ + struct ena_napi *ena_napi = container_of(napi, struct ena_napi, napi); + u32 xdp_work_done, xdp_budget; + struct ena_ring *xdp_ring; + int napi_comp_call = 0; + int ret; + + xdp_ring = ena_napi->xdp_ring; + + xdp_budget = budget; + + if (!test_bit(ENA_FLAG_DEV_UP, &xdp_ring->adapter->flags) || + test_bit(ENA_FLAG_TRIGGER_RESET, &xdp_ring->adapter->flags)) { + napi_complete_done(napi, 0); + return 0; + } + + xdp_work_done = ena_clean_xdp_irq(xdp_ring, xdp_budget); + + /* If the device is about to reset or down, avoid unmask + * the interrupt and return 0 so NAPI won't reschedule + */ + if (unlikely(!test_bit(ENA_FLAG_DEV_UP, &xdp_ring->adapter->flags))) { + napi_complete_done(napi, 0); + ret = 0; + } else if (xdp_budget > xdp_work_done) { + napi_comp_call = 1; + if (napi_complete_done(napi, xdp_work_done)) + ena_unmask_interrupt(xdp_ring, NULL); + ena_update_ring_numa_node(xdp_ring, NULL); + ret = xdp_work_done; + } else { + ret = xdp_budget; + } + + u64_stats_update_begin(&xdp_ring->syncp); + xdp_ring->tx_stats.napi_comp += napi_comp_call; + xdp_ring->tx_stats.tx_poll++; + u64_stats_update_end(&xdp_ring->syncp); + xdp_ring->tx_stats.last_napi_jiffies = jiffies; + + return ret; +} + +static int ena_xdp_tx_map_frame(struct ena_ring *xdp_ring, + struct ena_tx_buffer *tx_info, + struct xdp_frame *xdpf, + struct ena_com_tx_ctx *ena_tx_ctx) +{ + struct ena_adapter *adapter = xdp_ring->adapter; + struct ena_com_buf *ena_buf; + int push_len = 0; + dma_addr_t dma; + void *data; + u32 size; + + tx_info->xdpf = xdpf; + data = tx_info->xdpf->data; + size = tx_info->xdpf->len; + + if (xdp_ring->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + /* Designate part of the packet for LLQ */ + push_len = min_t(u32, size, xdp_ring->tx_max_header_size); + + ena_tx_ctx->push_header = data; + + size -= push_len; + data += push_len; + } + + ena_tx_ctx->header_len = push_len; + + if (size > 0) { + dma = dma_map_single(xdp_ring->dev, + data, + size, + DMA_TO_DEVICE); + if (unlikely(dma_mapping_error(xdp_ring->dev, dma))) + goto error_report_dma_error; + + tx_info->map_linear_data = 0; + + ena_buf = tx_info->bufs; + ena_buf->paddr = dma; + ena_buf->len = size; + + ena_tx_ctx->ena_bufs = ena_buf; + ena_tx_ctx->num_bufs = tx_info->num_of_bufs = 1; + } + + return 0; + +error_report_dma_error: + ena_increase_stat(&xdp_ring->tx_stats.dma_mapping_err, 1, + &xdp_ring->syncp); + netif_warn(adapter, tx_queued, adapter->netdev, "Failed to map xdp buff\n"); + + return -EINVAL; +} + +static int ena_xdp_xmit_frame(struct ena_ring *xdp_ring, + struct net_device *dev, + struct xdp_frame *xdpf, + int flags) +{ + struct ena_com_tx_ctx ena_tx_ctx = {}; + struct ena_tx_buffer *tx_info; + u16 next_to_use, req_id; + int rc; + + next_to_use = xdp_ring->next_to_use; + req_id = xdp_ring->free_ids[next_to_use]; + tx_info = &xdp_ring->tx_buffer_info[req_id]; + tx_info->num_of_bufs = 0; + + rc = ena_xdp_tx_map_frame(xdp_ring, tx_info, xdpf, &ena_tx_ctx); + if (unlikely(rc)) + return rc; + + ena_tx_ctx.req_id = req_id; + + rc = ena_xmit_common(dev, + xdp_ring, + tx_info, + &ena_tx_ctx, + next_to_use, + xdpf->len); + if (rc) + goto error_unmap_dma; + + /* trigger the dma engine. ena_ring_tx_doorbell() + * calls a memory barrier inside it. + */ + if (flags & XDP_XMIT_FLUSH) + ena_ring_tx_doorbell(xdp_ring); + + return rc; + +error_unmap_dma: + ena_unmap_tx_buff(xdp_ring, tx_info); + tx_info->xdpf = NULL; + return rc; +} + +static int ena_xdp_xmit(struct net_device *dev, int n, + struct xdp_frame **frames, u32 flags) +{ + struct ena_adapter *adapter = netdev_priv(dev); + struct ena_ring *xdp_ring; + int qid, i, nxmit = 0; + + if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK)) + return -EINVAL; + + if (!test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + return -ENETDOWN; + + /* We assume that all rings have the same XDP program */ + if (!READ_ONCE(adapter->rx_ring->xdp_bpf_prog)) + return -ENXIO; + + qid = smp_processor_id() % adapter->xdp_num_queues; + qid += adapter->xdp_first_ring; + xdp_ring = &adapter->tx_ring[qid]; + + /* Other CPU ids might try to send thorugh this queue */ + spin_lock(&xdp_ring->xdp_tx_lock); + + for (i = 0; i < n; i++) { + if (ena_xdp_xmit_frame(xdp_ring, dev, frames[i], 0)) + break; + nxmit++; + } + + /* Ring doorbell to make device aware of the packets */ + if (flags & XDP_XMIT_FLUSH) + ena_ring_tx_doorbell(xdp_ring); + + spin_unlock(&xdp_ring->xdp_tx_lock); + + /* Return number of packets sent */ + return nxmit; +} + +static int ena_xdp_execute(struct ena_ring *rx_ring, struct xdp_buff *xdp) +{ + u32 verdict = ENA_XDP_PASS; + struct bpf_prog *xdp_prog; + struct ena_ring *xdp_ring; + struct xdp_frame *xdpf; + u64 *xdp_stat; + + xdp_prog = READ_ONCE(rx_ring->xdp_bpf_prog); + + if (!xdp_prog) + goto out; + + verdict = bpf_prog_run_xdp(xdp_prog, xdp); + + switch (verdict) { + case XDP_TX: + xdpf = xdp_convert_buff_to_frame(xdp); + if (unlikely(!xdpf)) { + trace_xdp_exception(rx_ring->netdev, xdp_prog, verdict); + xdp_stat = &rx_ring->rx_stats.xdp_aborted; + verdict = ENA_XDP_DROP; + break; + } + + /* Find xmit queue */ + xdp_ring = rx_ring->xdp_ring; + + /* The XDP queues are shared between XDP_TX and XDP_REDIRECT */ + spin_lock(&xdp_ring->xdp_tx_lock); + + if (ena_xdp_xmit_frame(xdp_ring, rx_ring->netdev, xdpf, + XDP_XMIT_FLUSH)) + xdp_return_frame(xdpf); + + spin_unlock(&xdp_ring->xdp_tx_lock); + xdp_stat = &rx_ring->rx_stats.xdp_tx; + verdict = ENA_XDP_TX; + break; + case XDP_REDIRECT: + if (likely(!xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog))) { + xdp_stat = &rx_ring->rx_stats.xdp_redirect; + verdict = ENA_XDP_REDIRECT; + break; + } + trace_xdp_exception(rx_ring->netdev, xdp_prog, verdict); + xdp_stat = &rx_ring->rx_stats.xdp_aborted; + verdict = ENA_XDP_DROP; + break; + case XDP_ABORTED: + trace_xdp_exception(rx_ring->netdev, xdp_prog, verdict); + xdp_stat = &rx_ring->rx_stats.xdp_aborted; + verdict = ENA_XDP_DROP; + break; + case XDP_DROP: + xdp_stat = &rx_ring->rx_stats.xdp_drop; + verdict = ENA_XDP_DROP; + break; + case XDP_PASS: + xdp_stat = &rx_ring->rx_stats.xdp_pass; + verdict = ENA_XDP_PASS; + break; + default: + bpf_warn_invalid_xdp_action(rx_ring->netdev, xdp_prog, verdict); + xdp_stat = &rx_ring->rx_stats.xdp_invalid; + verdict = ENA_XDP_DROP; + } + + ena_increase_stat(xdp_stat, 1, &rx_ring->syncp); +out: + return verdict; +} + +static void ena_init_all_xdp_queues(struct ena_adapter *adapter) +{ + adapter->xdp_first_ring = adapter->num_io_queues; + adapter->xdp_num_queues = adapter->num_io_queues; + + ena_init_io_rings(adapter, + adapter->xdp_first_ring, + adapter->xdp_num_queues); +} + +static int ena_setup_and_create_all_xdp_queues(struct ena_adapter *adapter) +{ + u32 xdp_first_ring = adapter->xdp_first_ring; + u32 xdp_num_queues = adapter->xdp_num_queues; + int rc = 0; + + rc = ena_setup_tx_resources_in_range(adapter, xdp_first_ring, xdp_num_queues); + if (rc) + goto setup_err; + + rc = ena_create_io_tx_queues_in_range(adapter, xdp_first_ring, xdp_num_queues); + if (rc) + goto create_err; + + return 0; + +create_err: + ena_free_all_io_tx_resources_in_range(adapter, xdp_first_ring, xdp_num_queues); +setup_err: + return rc; +} + +/* Provides a way for both kernel and bpf-prog to know + * more about the RX-queue a given XDP frame arrived on. + */ +static int ena_xdp_register_rxq_info(struct ena_ring *rx_ring) +{ + int rc; + + rc = xdp_rxq_info_reg(&rx_ring->xdp_rxq, rx_ring->netdev, rx_ring->qid, 0); + + if (rc) { + netif_err(rx_ring->adapter, ifup, rx_ring->netdev, + "Failed to register xdp rx queue info. RX queue num %d rc: %d\n", + rx_ring->qid, rc); + goto err; + } + + rc = xdp_rxq_info_reg_mem_model(&rx_ring->xdp_rxq, MEM_TYPE_PAGE_SHARED, + NULL); + + if (rc) { + netif_err(rx_ring->adapter, ifup, rx_ring->netdev, + "Failed to register xdp rx queue info memory model. RX queue num %d rc: %d\n", + rx_ring->qid, rc); + xdp_rxq_info_unreg(&rx_ring->xdp_rxq); + } + +err: + return rc; +} + +static void ena_xdp_unregister_rxq_info(struct ena_ring *rx_ring) +{ + xdp_rxq_info_unreg_mem_model(&rx_ring->xdp_rxq); + xdp_rxq_info_unreg(&rx_ring->xdp_rxq); +} + +static void ena_xdp_exchange_program_rx_in_range(struct ena_adapter *adapter, + struct bpf_prog *prog, + int first, int count) +{ + struct bpf_prog *old_bpf_prog; + struct ena_ring *rx_ring; + int i = 0; + + for (i = first; i < count; i++) { + rx_ring = &adapter->rx_ring[i]; + old_bpf_prog = xchg(&rx_ring->xdp_bpf_prog, prog); + + if (!old_bpf_prog && prog) { + ena_xdp_register_rxq_info(rx_ring); + rx_ring->rx_headroom = XDP_PACKET_HEADROOM; + } else if (old_bpf_prog && !prog) { + ena_xdp_unregister_rxq_info(rx_ring); + rx_ring->rx_headroom = NET_SKB_PAD; + } + } +} + +static void ena_xdp_exchange_program(struct ena_adapter *adapter, + struct bpf_prog *prog) +{ + struct bpf_prog *old_bpf_prog = xchg(&adapter->xdp_bpf_prog, prog); + + ena_xdp_exchange_program_rx_in_range(adapter, + prog, + 0, + adapter->num_io_queues); + + if (old_bpf_prog) + bpf_prog_put(old_bpf_prog); +} + +static int ena_destroy_and_free_all_xdp_queues(struct ena_adapter *adapter) +{ + bool was_up; + int rc; + + was_up = test_bit(ENA_FLAG_DEV_UP, &adapter->flags); + + if (was_up) + ena_down(adapter); + + adapter->xdp_first_ring = 0; + adapter->xdp_num_queues = 0; + ena_xdp_exchange_program(adapter, NULL); + if (was_up) { + rc = ena_up(adapter); + if (rc) + return rc; + } + return 0; +} + +static int ena_xdp_set(struct net_device *netdev, struct netdev_bpf *bpf) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct bpf_prog *prog = bpf->prog; + struct bpf_prog *old_bpf_prog; + int rc, prev_mtu; + bool is_up; + + is_up = test_bit(ENA_FLAG_DEV_UP, &adapter->flags); + rc = ena_xdp_allowed(adapter); + if (rc == ENA_XDP_ALLOWED) { + old_bpf_prog = adapter->xdp_bpf_prog; + if (prog) { + if (!is_up) { + ena_init_all_xdp_queues(adapter); + } else if (!old_bpf_prog) { + ena_down(adapter); + ena_init_all_xdp_queues(adapter); + } + ena_xdp_exchange_program(adapter, prog); + + if (is_up && !old_bpf_prog) { + rc = ena_up(adapter); + if (rc) + return rc; + } + xdp_features_set_redirect_target(netdev, false); + } else if (old_bpf_prog) { + xdp_features_clear_redirect_target(netdev); + rc = ena_destroy_and_free_all_xdp_queues(adapter); + if (rc) + return rc; + } + + prev_mtu = netdev->max_mtu; + netdev->max_mtu = prog ? ENA_XDP_MAX_MTU : adapter->max_mtu; + + if (!old_bpf_prog) + netif_info(adapter, drv, adapter->netdev, + "XDP program is set, changing the max_mtu from %d to %d", + prev_mtu, netdev->max_mtu); + + } else if (rc == ENA_XDP_CURRENT_MTU_TOO_LARGE) { + netif_err(adapter, drv, adapter->netdev, + "Failed to set xdp program, the current MTU (%d) is larger than the maximum allowed MTU (%lu) while xdp is on", + netdev->mtu, ENA_XDP_MAX_MTU); + NL_SET_ERR_MSG_MOD(bpf->extack, + "Failed to set xdp program, the current MTU is larger than the maximum allowed MTU. Check the dmesg for more info"); + return -EINVAL; + } else if (rc == ENA_XDP_NO_ENOUGH_QUEUES) { + netif_err(adapter, drv, adapter->netdev, + "Failed to set xdp program, the Rx/Tx channel count should be at most half of the maximum allowed channel count. The current queue count (%d), the maximal queue count (%d)\n", + adapter->num_io_queues, adapter->max_num_io_queues); + NL_SET_ERR_MSG_MOD(bpf->extack, + "Failed to set xdp program, there is no enough space for allocating XDP queues, Check the dmesg for more info"); + return -EINVAL; + } + + return 0; +} + +/* This is the main xdp callback, it's used by the kernel to set/unset the xdp + * program as well as to query the current xdp program id. + */ +static int ena_xdp(struct net_device *netdev, struct netdev_bpf *bpf) +{ + switch (bpf->command) { + case XDP_SETUP_PROG: + return ena_xdp_set(netdev, bpf); + default: + return -EINVAL; + } + return 0; +} + +static int ena_init_rx_cpu_rmap(struct ena_adapter *adapter) +{ +#ifdef CONFIG_RFS_ACCEL + u32 i; + int rc; + + adapter->netdev->rx_cpu_rmap = alloc_irq_cpu_rmap(adapter->num_io_queues); + if (!adapter->netdev->rx_cpu_rmap) + return -ENOMEM; + for (i = 0; i < adapter->num_io_queues; i++) { + int irq_idx = ENA_IO_IRQ_IDX(i); + + rc = irq_cpu_rmap_add(adapter->netdev->rx_cpu_rmap, + pci_irq_vector(adapter->pdev, irq_idx)); + if (rc) { + free_irq_cpu_rmap(adapter->netdev->rx_cpu_rmap); + adapter->netdev->rx_cpu_rmap = NULL; + return rc; + } + } +#endif /* CONFIG_RFS_ACCEL */ + return 0; +} + +static void ena_init_io_rings_common(struct ena_adapter *adapter, + struct ena_ring *ring, u16 qid) +{ + ring->qid = qid; + ring->pdev = adapter->pdev; + ring->dev = &adapter->pdev->dev; + ring->netdev = adapter->netdev; + ring->napi = &adapter->ena_napi[qid].napi; + ring->adapter = adapter; + ring->ena_dev = adapter->ena_dev; + ring->per_napi_packets = 0; + ring->cpu = 0; + ring->numa_node = 0; + ring->no_interrupt_event_cnt = 0; + u64_stats_init(&ring->syncp); +} + +static void ena_init_io_rings(struct ena_adapter *adapter, + int first_index, int count) +{ + struct ena_com_dev *ena_dev; + struct ena_ring *txr, *rxr; + int i; + + ena_dev = adapter->ena_dev; + + for (i = first_index; i < first_index + count; i++) { + txr = &adapter->tx_ring[i]; + rxr = &adapter->rx_ring[i]; + + /* TX common ring state */ + ena_init_io_rings_common(adapter, txr, i); + + /* TX specific ring state */ + txr->ring_size = adapter->requested_tx_ring_size; + txr->tx_max_header_size = ena_dev->tx_max_header_size; + txr->tx_mem_queue_type = ena_dev->tx_mem_queue_type; + txr->sgl_size = adapter->max_tx_sgl_size; + txr->smoothed_interval = + ena_com_get_nonadaptive_moderation_interval_tx(ena_dev); + txr->disable_meta_caching = adapter->disable_meta_caching; + spin_lock_init(&txr->xdp_tx_lock); + + /* Don't init RX queues for xdp queues */ + if (!ENA_IS_XDP_INDEX(adapter, i)) { + /* RX common ring state */ + ena_init_io_rings_common(adapter, rxr, i); + + /* RX specific ring state */ + rxr->ring_size = adapter->requested_rx_ring_size; + rxr->rx_copybreak = adapter->rx_copybreak; + rxr->sgl_size = adapter->max_rx_sgl_size; + rxr->smoothed_interval = + ena_com_get_nonadaptive_moderation_interval_rx(ena_dev); + rxr->empty_rx_queue = 0; + rxr->rx_headroom = NET_SKB_PAD; + adapter->ena_napi[i].dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE; + rxr->xdp_ring = &adapter->tx_ring[i + adapter->num_io_queues]; + } + } +} + +/* ena_setup_tx_resources - allocate I/O Tx resources (Descriptors) + * @adapter: network interface device structure + * @qid: queue index + * + * Return 0 on success, negative on failure + */ +static int ena_setup_tx_resources(struct ena_adapter *adapter, int qid) +{ + struct ena_ring *tx_ring = &adapter->tx_ring[qid]; + struct ena_irq *ena_irq = &adapter->irq_tbl[ENA_IO_IRQ_IDX(qid)]; + int size, i, node; + + if (tx_ring->tx_buffer_info) { + netif_err(adapter, ifup, + adapter->netdev, "tx_buffer_info info is not NULL"); + return -EEXIST; + } + + size = sizeof(struct ena_tx_buffer) * tx_ring->ring_size; + node = cpu_to_node(ena_irq->cpu); + + tx_ring->tx_buffer_info = vzalloc_node(size, node); + if (!tx_ring->tx_buffer_info) { + tx_ring->tx_buffer_info = vzalloc(size); + if (!tx_ring->tx_buffer_info) + goto err_tx_buffer_info; + } + + size = sizeof(u16) * tx_ring->ring_size; + tx_ring->free_ids = vzalloc_node(size, node); + if (!tx_ring->free_ids) { + tx_ring->free_ids = vzalloc(size); + if (!tx_ring->free_ids) + goto err_tx_free_ids; + } + + size = tx_ring->tx_max_header_size; + tx_ring->push_buf_intermediate_buf = vzalloc_node(size, node); + if (!tx_ring->push_buf_intermediate_buf) { + tx_ring->push_buf_intermediate_buf = vzalloc(size); + if (!tx_ring->push_buf_intermediate_buf) + goto err_push_buf_intermediate_buf; + } + + /* Req id ring for TX out of order completions */ + for (i = 0; i < tx_ring->ring_size; i++) + tx_ring->free_ids[i] = i; + + /* Reset tx statistics */ + memset(&tx_ring->tx_stats, 0x0, sizeof(tx_ring->tx_stats)); + + tx_ring->next_to_use = 0; + tx_ring->next_to_clean = 0; + tx_ring->cpu = ena_irq->cpu; + tx_ring->numa_node = node; + return 0; + +err_push_buf_intermediate_buf: + vfree(tx_ring->free_ids); + tx_ring->free_ids = NULL; +err_tx_free_ids: + vfree(tx_ring->tx_buffer_info); + tx_ring->tx_buffer_info = NULL; +err_tx_buffer_info: + return -ENOMEM; +} + +/* ena_free_tx_resources - Free I/O Tx Resources per Queue + * @adapter: network interface device structure + * @qid: queue index + * + * Free all transmit software resources + */ +static void ena_free_tx_resources(struct ena_adapter *adapter, int qid) +{ + struct ena_ring *tx_ring = &adapter->tx_ring[qid]; + + vfree(tx_ring->tx_buffer_info); + tx_ring->tx_buffer_info = NULL; + + vfree(tx_ring->free_ids); + tx_ring->free_ids = NULL; + + vfree(tx_ring->push_buf_intermediate_buf); + tx_ring->push_buf_intermediate_buf = NULL; +} + +static int ena_setup_tx_resources_in_range(struct ena_adapter *adapter, + int first_index, + int count) +{ + int i, rc = 0; + + for (i = first_index; i < first_index + count; i++) { + rc = ena_setup_tx_resources(adapter, i); + if (rc) + goto err_setup_tx; + } + + return 0; + +err_setup_tx: + + netif_err(adapter, ifup, adapter->netdev, + "Tx queue %d: allocation failed\n", i); + + /* rewind the index freeing the rings as we go */ + while (first_index < i--) + ena_free_tx_resources(adapter, i); + return rc; +} + +static void ena_free_all_io_tx_resources_in_range(struct ena_adapter *adapter, + int first_index, int count) +{ + int i; + + for (i = first_index; i < first_index + count; i++) + ena_free_tx_resources(adapter, i); +} + +/* ena_free_all_io_tx_resources - Free I/O Tx Resources for All Queues + * @adapter: board private structure + * + * Free all transmit software resources + */ +static void ena_free_all_io_tx_resources(struct ena_adapter *adapter) +{ + ena_free_all_io_tx_resources_in_range(adapter, + 0, + adapter->xdp_num_queues + + adapter->num_io_queues); +} + +/* ena_setup_rx_resources - allocate I/O Rx resources (Descriptors) + * @adapter: network interface device structure + * @qid: queue index + * + * Returns 0 on success, negative on failure + */ +static int ena_setup_rx_resources(struct ena_adapter *adapter, + u32 qid) +{ + struct ena_ring *rx_ring = &adapter->rx_ring[qid]; + struct ena_irq *ena_irq = &adapter->irq_tbl[ENA_IO_IRQ_IDX(qid)]; + int size, node, i; + + if (rx_ring->rx_buffer_info) { + netif_err(adapter, ifup, adapter->netdev, + "rx_buffer_info is not NULL"); + return -EEXIST; + } + + /* alloc extra element so in rx path + * we can always prefetch rx_info + 1 + */ + size = sizeof(struct ena_rx_buffer) * (rx_ring->ring_size + 1); + node = cpu_to_node(ena_irq->cpu); + + rx_ring->rx_buffer_info = vzalloc_node(size, node); + if (!rx_ring->rx_buffer_info) { + rx_ring->rx_buffer_info = vzalloc(size); + if (!rx_ring->rx_buffer_info) + return -ENOMEM; + } + + size = sizeof(u16) * rx_ring->ring_size; + rx_ring->free_ids = vzalloc_node(size, node); + if (!rx_ring->free_ids) { + rx_ring->free_ids = vzalloc(size); + if (!rx_ring->free_ids) { + vfree(rx_ring->rx_buffer_info); + rx_ring->rx_buffer_info = NULL; + return -ENOMEM; + } + } + + /* Req id ring for receiving RX pkts out of order */ + for (i = 0; i < rx_ring->ring_size; i++) + rx_ring->free_ids[i] = i; + + /* Reset rx statistics */ + memset(&rx_ring->rx_stats, 0x0, sizeof(rx_ring->rx_stats)); + + rx_ring->next_to_clean = 0; + rx_ring->next_to_use = 0; + rx_ring->cpu = ena_irq->cpu; + rx_ring->numa_node = node; + + return 0; +} + +/* ena_free_rx_resources - Free I/O Rx Resources + * @adapter: network interface device structure + * @qid: queue index + * + * Free all receive software resources + */ +static void ena_free_rx_resources(struct ena_adapter *adapter, + u32 qid) +{ + struct ena_ring *rx_ring = &adapter->rx_ring[qid]; + + vfree(rx_ring->rx_buffer_info); + rx_ring->rx_buffer_info = NULL; + + vfree(rx_ring->free_ids); + rx_ring->free_ids = NULL; +} + +/* ena_setup_all_rx_resources - allocate I/O Rx queues resources for all queues + * @adapter: board private structure + * + * Return 0 on success, negative on failure + */ +static int ena_setup_all_rx_resources(struct ena_adapter *adapter) +{ + int i, rc = 0; + + for (i = 0; i < adapter->num_io_queues; i++) { + rc = ena_setup_rx_resources(adapter, i); + if (rc) + goto err_setup_rx; + } + + return 0; + +err_setup_rx: + + netif_err(adapter, ifup, adapter->netdev, + "Rx queue %d: allocation failed\n", i); + + /* rewind the index freeing the rings as we go */ + while (i--) + ena_free_rx_resources(adapter, i); + return rc; +} + +/* ena_free_all_io_rx_resources - Free I/O Rx Resources for All Queues + * @adapter: board private structure + * + * Free all receive software resources + */ +static void ena_free_all_io_rx_resources(struct ena_adapter *adapter) +{ + int i; + + for (i = 0; i < adapter->num_io_queues; i++) + ena_free_rx_resources(adapter, i); +} + +static struct page *ena_alloc_map_page(struct ena_ring *rx_ring, + dma_addr_t *dma) +{ + struct page *page; + + /* This would allocate the page on the same NUMA node the executing code + * is running on. + */ + page = dev_alloc_page(); + if (!page) { + ena_increase_stat(&rx_ring->rx_stats.page_alloc_fail, 1, + &rx_ring->syncp); + return ERR_PTR(-ENOSPC); + } + + /* To enable NIC-side port-mirroring, AKA SPAN port, + * we make the buffer readable from the nic as well + */ + *dma = dma_map_page(rx_ring->dev, page, 0, ENA_PAGE_SIZE, + DMA_BIDIRECTIONAL); + if (unlikely(dma_mapping_error(rx_ring->dev, *dma))) { + ena_increase_stat(&rx_ring->rx_stats.dma_mapping_err, 1, + &rx_ring->syncp); + __free_page(page); + return ERR_PTR(-EIO); + } + + return page; +} + +static int ena_alloc_rx_buffer(struct ena_ring *rx_ring, + struct ena_rx_buffer *rx_info) +{ + int headroom = rx_ring->rx_headroom; + struct ena_com_buf *ena_buf; + struct page *page; + dma_addr_t dma; + int tailroom; + + /* restore page offset value in case it has been changed by device */ + rx_info->buf_offset = headroom; + + /* if previous allocated page is not used */ + if (unlikely(rx_info->page)) + return 0; + + /* We handle DMA here */ + page = ena_alloc_map_page(rx_ring, &dma); + if (unlikely(IS_ERR(page))) + return PTR_ERR(page); + + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "Allocate page %p, rx_info %p\n", page, rx_info); + + tailroom = SKB_DATA_ALIGN(sizeof(struct skb_shared_info)); + + rx_info->page = page; + rx_info->dma_addr = dma; + rx_info->page_offset = 0; + ena_buf = &rx_info->ena_buf; + ena_buf->paddr = dma + headroom; + ena_buf->len = ENA_PAGE_SIZE - headroom - tailroom; + + return 0; +} + +static void ena_unmap_rx_buff_attrs(struct ena_ring *rx_ring, + struct ena_rx_buffer *rx_info, + unsigned long attrs) +{ + dma_unmap_page_attrs(rx_ring->dev, rx_info->dma_addr, ENA_PAGE_SIZE, + DMA_BIDIRECTIONAL, attrs); +} + +static void ena_free_rx_page(struct ena_ring *rx_ring, + struct ena_rx_buffer *rx_info) +{ + struct page *page = rx_info->page; + + if (unlikely(!page)) { + netif_warn(rx_ring->adapter, rx_err, rx_ring->netdev, + "Trying to free unallocated buffer\n"); + return; + } + + ena_unmap_rx_buff_attrs(rx_ring, rx_info, 0); + + __free_page(page); + rx_info->page = NULL; +} + +static int ena_refill_rx_bufs(struct ena_ring *rx_ring, u32 num) +{ + u16 next_to_use, req_id; + u32 i; + int rc; + + next_to_use = rx_ring->next_to_use; + + for (i = 0; i < num; i++) { + struct ena_rx_buffer *rx_info; + + req_id = rx_ring->free_ids[next_to_use]; + + rx_info = &rx_ring->rx_buffer_info[req_id]; + + rc = ena_alloc_rx_buffer(rx_ring, rx_info); + if (unlikely(rc < 0)) { + netif_warn(rx_ring->adapter, rx_err, rx_ring->netdev, + "Failed to allocate buffer for rx queue %d\n", + rx_ring->qid); + break; + } + rc = ena_com_add_single_rx_desc(rx_ring->ena_com_io_sq, + &rx_info->ena_buf, + req_id); + if (unlikely(rc)) { + netif_warn(rx_ring->adapter, rx_status, rx_ring->netdev, + "Failed to add buffer for rx queue %d\n", + rx_ring->qid); + break; + } + next_to_use = ENA_RX_RING_IDX_NEXT(next_to_use, + rx_ring->ring_size); + } + + if (unlikely(i < num)) { + ena_increase_stat(&rx_ring->rx_stats.refil_partial, 1, + &rx_ring->syncp); + netif_warn(rx_ring->adapter, rx_err, rx_ring->netdev, + "Refilled rx qid %d with only %d buffers (from %d)\n", + rx_ring->qid, i, num); + } + + /* ena_com_write_sq_doorbell issues a wmb() */ + if (likely(i)) + ena_com_write_sq_doorbell(rx_ring->ena_com_io_sq); + + rx_ring->next_to_use = next_to_use; + + return i; +} + +static void ena_free_rx_bufs(struct ena_adapter *adapter, + u32 qid) +{ + struct ena_ring *rx_ring = &adapter->rx_ring[qid]; + u32 i; + + for (i = 0; i < rx_ring->ring_size; i++) { + struct ena_rx_buffer *rx_info = &rx_ring->rx_buffer_info[i]; + + if (rx_info->page) + ena_free_rx_page(rx_ring, rx_info); + } +} + +/* ena_refill_all_rx_bufs - allocate all queues Rx buffers + * @adapter: board private structure + */ +static void ena_refill_all_rx_bufs(struct ena_adapter *adapter) +{ + struct ena_ring *rx_ring; + int i, rc, bufs_num; + + for (i = 0; i < adapter->num_io_queues; i++) { + rx_ring = &adapter->rx_ring[i]; + bufs_num = rx_ring->ring_size - 1; + rc = ena_refill_rx_bufs(rx_ring, bufs_num); + + if (unlikely(rc != bufs_num)) + netif_warn(rx_ring->adapter, rx_status, rx_ring->netdev, + "Refilling Queue %d failed. allocated %d buffers from: %d\n", + i, rc, bufs_num); + } +} + +static void ena_free_all_rx_bufs(struct ena_adapter *adapter) +{ + int i; + + for (i = 0; i < adapter->num_io_queues; i++) + ena_free_rx_bufs(adapter, i); +} + +static void ena_unmap_tx_buff(struct ena_ring *tx_ring, + struct ena_tx_buffer *tx_info) +{ + struct ena_com_buf *ena_buf; + u32 cnt; + int i; + + ena_buf = tx_info->bufs; + cnt = tx_info->num_of_bufs; + + if (unlikely(!cnt)) + return; + + if (tx_info->map_linear_data) { + dma_unmap_single(tx_ring->dev, + dma_unmap_addr(ena_buf, paddr), + dma_unmap_len(ena_buf, len), + DMA_TO_DEVICE); + ena_buf++; + cnt--; + } + + /* unmap remaining mapped pages */ + for (i = 0; i < cnt; i++) { + dma_unmap_page(tx_ring->dev, dma_unmap_addr(ena_buf, paddr), + dma_unmap_len(ena_buf, len), DMA_TO_DEVICE); + ena_buf++; + } +} + +/* ena_free_tx_bufs - Free Tx Buffers per Queue + * @tx_ring: TX ring for which buffers be freed + */ +static void ena_free_tx_bufs(struct ena_ring *tx_ring) +{ + bool print_once = true; + u32 i; + + for (i = 0; i < tx_ring->ring_size; i++) { + struct ena_tx_buffer *tx_info = &tx_ring->tx_buffer_info[i]; + + if (!tx_info->skb) + continue; + + if (print_once) { + netif_notice(tx_ring->adapter, ifdown, tx_ring->netdev, + "Free uncompleted tx skb qid %d idx 0x%x\n", + tx_ring->qid, i); + print_once = false; + } else { + netif_dbg(tx_ring->adapter, ifdown, tx_ring->netdev, + "Free uncompleted tx skb qid %d idx 0x%x\n", + tx_ring->qid, i); + } + + ena_unmap_tx_buff(tx_ring, tx_info); + + dev_kfree_skb_any(tx_info->skb); + } + netdev_tx_reset_queue(netdev_get_tx_queue(tx_ring->netdev, + tx_ring->qid)); +} + +static void ena_free_all_tx_bufs(struct ena_adapter *adapter) +{ + struct ena_ring *tx_ring; + int i; + + for (i = 0; i < adapter->num_io_queues + adapter->xdp_num_queues; i++) { + tx_ring = &adapter->tx_ring[i]; + ena_free_tx_bufs(tx_ring); + } +} + +static void ena_destroy_all_tx_queues(struct ena_adapter *adapter) +{ + u16 ena_qid; + int i; + + for (i = 0; i < adapter->num_io_queues + adapter->xdp_num_queues; i++) { + ena_qid = ENA_IO_TXQ_IDX(i); + ena_com_destroy_io_queue(adapter->ena_dev, ena_qid); + } +} + +static void ena_destroy_all_rx_queues(struct ena_adapter *adapter) +{ + u16 ena_qid; + int i; + + for (i = 0; i < adapter->num_io_queues; i++) { + ena_qid = ENA_IO_RXQ_IDX(i); + cancel_work_sync(&adapter->ena_napi[i].dim.work); + ena_com_destroy_io_queue(adapter->ena_dev, ena_qid); + } +} + +static void ena_destroy_all_io_queues(struct ena_adapter *adapter) +{ + ena_destroy_all_tx_queues(adapter); + ena_destroy_all_rx_queues(adapter); +} + +static int handle_invalid_req_id(struct ena_ring *ring, u16 req_id, + struct ena_tx_buffer *tx_info, bool is_xdp) +{ + if (tx_info) + netif_err(ring->adapter, + tx_done, + ring->netdev, + "tx_info doesn't have valid %s. qid %u req_id %u", + is_xdp ? "xdp frame" : "skb", ring->qid, req_id); + else + netif_err(ring->adapter, + tx_done, + ring->netdev, + "Invalid req_id %u in qid %u\n", + req_id, ring->qid); + + ena_increase_stat(&ring->tx_stats.bad_req_id, 1, &ring->syncp); + ena_reset_device(ring->adapter, ENA_REGS_RESET_INV_TX_REQ_ID); + + return -EFAULT; +} + +static int validate_tx_req_id(struct ena_ring *tx_ring, u16 req_id) +{ + struct ena_tx_buffer *tx_info; + + tx_info = &tx_ring->tx_buffer_info[req_id]; + if (likely(tx_info->skb)) + return 0; + + return handle_invalid_req_id(tx_ring, req_id, tx_info, false); +} + +static int validate_xdp_req_id(struct ena_ring *xdp_ring, u16 req_id) +{ + struct ena_tx_buffer *tx_info; + + tx_info = &xdp_ring->tx_buffer_info[req_id]; + if (likely(tx_info->xdpf)) + return 0; + + return handle_invalid_req_id(xdp_ring, req_id, tx_info, true); +} + +static int ena_clean_tx_irq(struct ena_ring *tx_ring, u32 budget) +{ + struct netdev_queue *txq; + bool above_thresh; + u32 tx_bytes = 0; + u32 total_done = 0; + u16 next_to_clean; + u16 req_id; + int tx_pkts = 0; + int rc; + + next_to_clean = tx_ring->next_to_clean; + txq = netdev_get_tx_queue(tx_ring->netdev, tx_ring->qid); + + while (tx_pkts < budget) { + struct ena_tx_buffer *tx_info; + struct sk_buff *skb; + + rc = ena_com_tx_comp_req_id_get(tx_ring->ena_com_io_cq, + &req_id); + if (rc) { + if (unlikely(rc == -EINVAL)) + handle_invalid_req_id(tx_ring, req_id, NULL, + false); + break; + } + + /* validate that the request id points to a valid skb */ + rc = validate_tx_req_id(tx_ring, req_id); + if (rc) + break; + + tx_info = &tx_ring->tx_buffer_info[req_id]; + skb = tx_info->skb; + + /* prefetch skb_end_pointer() to speedup skb_shinfo(skb) */ + prefetch(&skb->end); + + tx_info->skb = NULL; + tx_info->last_jiffies = 0; + + ena_unmap_tx_buff(tx_ring, tx_info); + + netif_dbg(tx_ring->adapter, tx_done, tx_ring->netdev, + "tx_poll: q %d skb %p completed\n", tx_ring->qid, + skb); + + tx_bytes += skb->len; + dev_kfree_skb(skb); + tx_pkts++; + total_done += tx_info->tx_descs; + + tx_ring->free_ids[next_to_clean] = req_id; + next_to_clean = ENA_TX_RING_IDX_NEXT(next_to_clean, + tx_ring->ring_size); + } + + tx_ring->next_to_clean = next_to_clean; + ena_com_comp_ack(tx_ring->ena_com_io_sq, total_done); + ena_com_update_dev_comp_head(tx_ring->ena_com_io_cq); + + netdev_tx_completed_queue(txq, tx_pkts, tx_bytes); + + netif_dbg(tx_ring->adapter, tx_done, tx_ring->netdev, + "tx_poll: q %d done. total pkts: %d\n", + tx_ring->qid, tx_pkts); + + /* need to make the rings circular update visible to + * ena_start_xmit() before checking for netif_queue_stopped(). + */ + smp_mb(); + + above_thresh = ena_com_sq_have_enough_space(tx_ring->ena_com_io_sq, + ENA_TX_WAKEUP_THRESH); + if (unlikely(netif_tx_queue_stopped(txq) && above_thresh)) { + __netif_tx_lock(txq, smp_processor_id()); + above_thresh = + ena_com_sq_have_enough_space(tx_ring->ena_com_io_sq, + ENA_TX_WAKEUP_THRESH); + if (netif_tx_queue_stopped(txq) && above_thresh && + test_bit(ENA_FLAG_DEV_UP, &tx_ring->adapter->flags)) { + netif_tx_wake_queue(txq); + ena_increase_stat(&tx_ring->tx_stats.queue_wakeup, 1, + &tx_ring->syncp); + } + __netif_tx_unlock(txq); + } + + return tx_pkts; +} + +static struct sk_buff *ena_alloc_skb(struct ena_ring *rx_ring, void *first_frag, u16 len) +{ + struct sk_buff *skb; + + if (!first_frag) + skb = napi_alloc_skb(rx_ring->napi, len); + else + skb = napi_build_skb(first_frag, len); + + if (unlikely(!skb)) { + ena_increase_stat(&rx_ring->rx_stats.skb_alloc_fail, 1, + &rx_ring->syncp); + + netif_dbg(rx_ring->adapter, rx_err, rx_ring->netdev, + "Failed to allocate skb. first_frag %s\n", + first_frag ? "provided" : "not provided"); + } + + return skb; +} + +static bool ena_try_rx_buf_page_reuse(struct ena_rx_buffer *rx_info, u16 buf_len, + u16 len, int pkt_offset) +{ + struct ena_com_buf *ena_buf = &rx_info->ena_buf; + + /* More than ENA_MIN_RX_BUF_SIZE left in the reused buffer + * for data + headroom + tailroom. + */ + if (SKB_DATA_ALIGN(len + pkt_offset) + ENA_MIN_RX_BUF_SIZE <= ena_buf->len) { + page_ref_inc(rx_info->page); + rx_info->page_offset += buf_len; + ena_buf->paddr += buf_len; + ena_buf->len -= buf_len; + return true; + } + + return false; +} + +static struct sk_buff *ena_rx_skb(struct ena_ring *rx_ring, + struct ena_com_rx_buf_info *ena_bufs, + u32 descs, + u16 *next_to_clean) +{ + int tailroom = SKB_DATA_ALIGN(sizeof(struct skb_shared_info)); + bool is_xdp_loaded = ena_xdp_present_ring(rx_ring); + struct ena_rx_buffer *rx_info; + struct ena_adapter *adapter; + int page_offset, pkt_offset; + dma_addr_t pre_reuse_paddr; + u16 len, req_id, buf = 0; + bool reuse_rx_buf_page; + struct sk_buff *skb; + void *buf_addr; + int buf_offset; + u16 buf_len; + + len = ena_bufs[buf].len; + req_id = ena_bufs[buf].req_id; + + rx_info = &rx_ring->rx_buffer_info[req_id]; + + if (unlikely(!rx_info->page)) { + adapter = rx_ring->adapter; + netif_err(adapter, rx_err, rx_ring->netdev, + "Page is NULL. qid %u req_id %u\n", rx_ring->qid, req_id); + ena_increase_stat(&rx_ring->rx_stats.bad_req_id, 1, &rx_ring->syncp); + ena_reset_device(adapter, ENA_REGS_RESET_INV_RX_REQ_ID); + return NULL; + } + + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "rx_info %p page %p\n", + rx_info, rx_info->page); + + buf_offset = rx_info->buf_offset; + pkt_offset = buf_offset - rx_ring->rx_headroom; + page_offset = rx_info->page_offset; + buf_addr = page_address(rx_info->page) + page_offset; + + if (len <= rx_ring->rx_copybreak) { + skb = ena_alloc_skb(rx_ring, NULL, len); + if (unlikely(!skb)) + return NULL; + + skb_copy_to_linear_data(skb, buf_addr + buf_offset, len); + dma_sync_single_for_device(rx_ring->dev, + dma_unmap_addr(&rx_info->ena_buf, paddr) + pkt_offset, + len, + DMA_FROM_DEVICE); + + skb_put(skb, len); + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "RX allocated small packet. len %d.\n", skb->len); + skb->protocol = eth_type_trans(skb, rx_ring->netdev); + rx_ring->free_ids[*next_to_clean] = req_id; + *next_to_clean = ENA_RX_RING_IDX_ADD(*next_to_clean, descs, + rx_ring->ring_size); + return skb; + } + + buf_len = SKB_DATA_ALIGN(len + buf_offset + tailroom); + + /* If XDP isn't loaded try to reuse part of the RX buffer */ + reuse_rx_buf_page = !is_xdp_loaded && + ena_try_rx_buf_page_reuse(rx_info, buf_len, len, pkt_offset); + + if (!reuse_rx_buf_page) + ena_unmap_rx_buff_attrs(rx_ring, rx_info, DMA_ATTR_SKIP_CPU_SYNC); + + skb = ena_alloc_skb(rx_ring, buf_addr, buf_len); + if (unlikely(!skb)) + return NULL; + + /* Populate skb's linear part */ + skb_reserve(skb, buf_offset); + skb_put(skb, len); + skb->protocol = eth_type_trans(skb, rx_ring->netdev); + + do { + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "RX skb updated. len %d. data_len %d\n", + skb->len, skb->data_len); + + if (!reuse_rx_buf_page) + rx_info->page = NULL; + + rx_ring->free_ids[*next_to_clean] = req_id; + *next_to_clean = + ENA_RX_RING_IDX_NEXT(*next_to_clean, + rx_ring->ring_size); + if (likely(--descs == 0)) + break; + + buf++; + len = ena_bufs[buf].len; + req_id = ena_bufs[buf].req_id; + + rx_info = &rx_ring->rx_buffer_info[req_id]; + + /* rx_info->buf_offset includes rx_ring->rx_headroom */ + buf_offset = rx_info->buf_offset; + pkt_offset = buf_offset - rx_ring->rx_headroom; + buf_len = SKB_DATA_ALIGN(len + buf_offset + tailroom); + page_offset = rx_info->page_offset; + + pre_reuse_paddr = dma_unmap_addr(&rx_info->ena_buf, paddr); + + reuse_rx_buf_page = !is_xdp_loaded && + ena_try_rx_buf_page_reuse(rx_info, buf_len, len, pkt_offset); + + dma_sync_single_for_cpu(rx_ring->dev, + pre_reuse_paddr + pkt_offset, + len, + DMA_FROM_DEVICE); + + if (!reuse_rx_buf_page) + ena_unmap_rx_buff_attrs(rx_ring, rx_info, + DMA_ATTR_SKIP_CPU_SYNC); + + skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_info->page, + page_offset + buf_offset, len, buf_len); + + } while (1); + + return skb; +} + +/* ena_rx_checksum - indicate in skb if hw indicated a good cksum + * @adapter: structure containing adapter specific data + * @ena_rx_ctx: received packet context/metadata + * @skb: skb currently being received and modified + */ +static void ena_rx_checksum(struct ena_ring *rx_ring, + struct ena_com_rx_ctx *ena_rx_ctx, + struct sk_buff *skb) +{ + /* Rx csum disabled */ + if (unlikely(!(rx_ring->netdev->features & NETIF_F_RXCSUM))) { + skb->ip_summed = CHECKSUM_NONE; + return; + } + + /* For fragmented packets the checksum isn't valid */ + if (ena_rx_ctx->frag) { + skb->ip_summed = CHECKSUM_NONE; + return; + } + + /* if IP and error */ + if (unlikely((ena_rx_ctx->l3_proto == ENA_ETH_IO_L3_PROTO_IPV4) && + (ena_rx_ctx->l3_csum_err))) { + /* ipv4 checksum error */ + skb->ip_summed = CHECKSUM_NONE; + ena_increase_stat(&rx_ring->rx_stats.csum_bad, 1, + &rx_ring->syncp); + netif_dbg(rx_ring->adapter, rx_err, rx_ring->netdev, + "RX IPv4 header checksum error\n"); + return; + } + + /* if TCP/UDP */ + if (likely((ena_rx_ctx->l4_proto == ENA_ETH_IO_L4_PROTO_TCP) || + (ena_rx_ctx->l4_proto == ENA_ETH_IO_L4_PROTO_UDP))) { + if (unlikely(ena_rx_ctx->l4_csum_err)) { + /* TCP/UDP checksum error */ + ena_increase_stat(&rx_ring->rx_stats.csum_bad, 1, + &rx_ring->syncp); + netif_dbg(rx_ring->adapter, rx_err, rx_ring->netdev, + "RX L4 checksum error\n"); + skb->ip_summed = CHECKSUM_NONE; + return; + } + + if (likely(ena_rx_ctx->l4_csum_checked)) { + skb->ip_summed = CHECKSUM_UNNECESSARY; + ena_increase_stat(&rx_ring->rx_stats.csum_good, 1, + &rx_ring->syncp); + } else { + ena_increase_stat(&rx_ring->rx_stats.csum_unchecked, 1, + &rx_ring->syncp); + skb->ip_summed = CHECKSUM_NONE; + } + } else { + skb->ip_summed = CHECKSUM_NONE; + return; + } + +} + +static void ena_set_rx_hash(struct ena_ring *rx_ring, + struct ena_com_rx_ctx *ena_rx_ctx, + struct sk_buff *skb) +{ + enum pkt_hash_types hash_type; + + if (likely(rx_ring->netdev->features & NETIF_F_RXHASH)) { + if (likely((ena_rx_ctx->l4_proto == ENA_ETH_IO_L4_PROTO_TCP) || + (ena_rx_ctx->l4_proto == ENA_ETH_IO_L4_PROTO_UDP))) + + hash_type = PKT_HASH_TYPE_L4; + else + hash_type = PKT_HASH_TYPE_NONE; + + /* Override hash type if the packet is fragmented */ + if (ena_rx_ctx->frag) + hash_type = PKT_HASH_TYPE_NONE; + + skb_set_hash(skb, ena_rx_ctx->hash, hash_type); + } +} + +static int ena_xdp_handle_buff(struct ena_ring *rx_ring, struct xdp_buff *xdp, u16 num_descs) +{ + struct ena_rx_buffer *rx_info; + int ret; + + /* XDP multi-buffer packets not supported */ + if (unlikely(num_descs > 1)) { + netdev_err_once(rx_ring->adapter->netdev, + "xdp: dropped unsupported multi-buffer packets\n"); + ena_increase_stat(&rx_ring->rx_stats.xdp_drop, 1, &rx_ring->syncp); + return ENA_XDP_DROP; + } + + rx_info = &rx_ring->rx_buffer_info[rx_ring->ena_bufs[0].req_id]; + xdp_prepare_buff(xdp, page_address(rx_info->page), + rx_info->buf_offset, + rx_ring->ena_bufs[0].len, false); + + ret = ena_xdp_execute(rx_ring, xdp); + + /* The xdp program might expand the headers */ + if (ret == ENA_XDP_PASS) { + rx_info->buf_offset = xdp->data - xdp->data_hard_start; + rx_ring->ena_bufs[0].len = xdp->data_end - xdp->data; + } + + return ret; +} +/* ena_clean_rx_irq - Cleanup RX irq + * @rx_ring: RX ring to clean + * @napi: napi handler + * @budget: how many packets driver is allowed to clean + * + * Returns the number of cleaned buffers. + */ +static int ena_clean_rx_irq(struct ena_ring *rx_ring, struct napi_struct *napi, + u32 budget) +{ + u16 next_to_clean = rx_ring->next_to_clean; + struct ena_com_rx_ctx ena_rx_ctx; + struct ena_rx_buffer *rx_info; + struct ena_adapter *adapter; + u32 res_budget, work_done; + int rx_copybreak_pkt = 0; + int refill_threshold; + struct sk_buff *skb; + int refill_required; + struct xdp_buff xdp; + int xdp_flags = 0; + int total_len = 0; + int xdp_verdict; + u8 pkt_offset; + int rc = 0; + int i; + + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "%s qid %d\n", __func__, rx_ring->qid); + res_budget = budget; + xdp_init_buff(&xdp, ENA_PAGE_SIZE, &rx_ring->xdp_rxq); + + do { + xdp_verdict = ENA_XDP_PASS; + skb = NULL; + ena_rx_ctx.ena_bufs = rx_ring->ena_bufs; + ena_rx_ctx.max_bufs = rx_ring->sgl_size; + ena_rx_ctx.descs = 0; + ena_rx_ctx.pkt_offset = 0; + rc = ena_com_rx_pkt(rx_ring->ena_com_io_cq, + rx_ring->ena_com_io_sq, + &ena_rx_ctx); + if (unlikely(rc)) + goto error; + + if (unlikely(ena_rx_ctx.descs == 0)) + break; + + /* First descriptor might have an offset set by the device */ + rx_info = &rx_ring->rx_buffer_info[rx_ring->ena_bufs[0].req_id]; + pkt_offset = ena_rx_ctx.pkt_offset; + rx_info->buf_offset += pkt_offset; + + netif_dbg(rx_ring->adapter, rx_status, rx_ring->netdev, + "rx_poll: q %d got packet from ena. descs #: %d l3 proto %d l4 proto %d hash: %x\n", + rx_ring->qid, ena_rx_ctx.descs, ena_rx_ctx.l3_proto, + ena_rx_ctx.l4_proto, ena_rx_ctx.hash); + + dma_sync_single_for_cpu(rx_ring->dev, + dma_unmap_addr(&rx_info->ena_buf, paddr) + pkt_offset, + rx_ring->ena_bufs[0].len, + DMA_FROM_DEVICE); + + if (ena_xdp_present_ring(rx_ring)) + xdp_verdict = ena_xdp_handle_buff(rx_ring, &xdp, ena_rx_ctx.descs); + + /* allocate skb and fill it */ + if (xdp_verdict == ENA_XDP_PASS) + skb = ena_rx_skb(rx_ring, + rx_ring->ena_bufs, + ena_rx_ctx.descs, + &next_to_clean); + + if (unlikely(!skb)) { + for (i = 0; i < ena_rx_ctx.descs; i++) { + int req_id = rx_ring->ena_bufs[i].req_id; + + rx_ring->free_ids[next_to_clean] = req_id; + next_to_clean = + ENA_RX_RING_IDX_NEXT(next_to_clean, + rx_ring->ring_size); + + /* Packets was passed for transmission, unmap it + * from RX side. + */ + if (xdp_verdict & ENA_XDP_FORWARDED) { + ena_unmap_rx_buff_attrs(rx_ring, + &rx_ring->rx_buffer_info[req_id], + DMA_ATTR_SKIP_CPU_SYNC); + rx_ring->rx_buffer_info[req_id].page = NULL; + } + } + if (xdp_verdict != ENA_XDP_PASS) { + xdp_flags |= xdp_verdict; + total_len += ena_rx_ctx.ena_bufs[0].len; + res_budget--; + continue; + } + break; + } + + ena_rx_checksum(rx_ring, &ena_rx_ctx, skb); + + ena_set_rx_hash(rx_ring, &ena_rx_ctx, skb); + + skb_record_rx_queue(skb, rx_ring->qid); + + if (rx_ring->ena_bufs[0].len <= rx_ring->rx_copybreak) + rx_copybreak_pkt++; + + total_len += skb->len; + + napi_gro_receive(napi, skb); + + res_budget--; + } while (likely(res_budget)); + + work_done = budget - res_budget; + rx_ring->per_napi_packets += work_done; + u64_stats_update_begin(&rx_ring->syncp); + rx_ring->rx_stats.bytes += total_len; + rx_ring->rx_stats.cnt += work_done; + rx_ring->rx_stats.rx_copybreak_pkt += rx_copybreak_pkt; + u64_stats_update_end(&rx_ring->syncp); + + rx_ring->next_to_clean = next_to_clean; + + refill_required = ena_com_free_q_entries(rx_ring->ena_com_io_sq); + refill_threshold = + min_t(int, rx_ring->ring_size / ENA_RX_REFILL_THRESH_DIVIDER, + ENA_RX_REFILL_THRESH_PACKET); + + /* Optimization, try to batch new rx buffers */ + if (refill_required > refill_threshold) { + ena_com_update_dev_comp_head(rx_ring->ena_com_io_cq); + ena_refill_rx_bufs(rx_ring, refill_required); + } + + if (xdp_flags & ENA_XDP_REDIRECT) + xdp_do_flush_map(); + + return work_done; + +error: + if (xdp_flags & ENA_XDP_REDIRECT) + xdp_do_flush(); + + adapter = netdev_priv(rx_ring->netdev); + + if (rc == -ENOSPC) { + ena_increase_stat(&rx_ring->rx_stats.bad_desc_num, 1, + &rx_ring->syncp); + ena_reset_device(adapter, ENA_REGS_RESET_TOO_MANY_RX_DESCS); + } else { + ena_increase_stat(&rx_ring->rx_stats.bad_req_id, 1, + &rx_ring->syncp); + ena_reset_device(adapter, ENA_REGS_RESET_INV_RX_REQ_ID); + } + return 0; +} + +static void ena_dim_work(struct work_struct *w) +{ + struct dim *dim = container_of(w, struct dim, work); + struct dim_cq_moder cur_moder = + net_dim_get_rx_moderation(dim->mode, dim->profile_ix); + struct ena_napi *ena_napi = container_of(dim, struct ena_napi, dim); + + ena_napi->rx_ring->smoothed_interval = cur_moder.usec; + dim->state = DIM_START_MEASURE; +} + +static void ena_adjust_adaptive_rx_intr_moderation(struct ena_napi *ena_napi) +{ + struct dim_sample dim_sample; + struct ena_ring *rx_ring = ena_napi->rx_ring; + + if (!rx_ring->per_napi_packets) + return; + + rx_ring->non_empty_napi_events++; + + dim_update_sample(rx_ring->non_empty_napi_events, + rx_ring->rx_stats.cnt, + rx_ring->rx_stats.bytes, + &dim_sample); + + net_dim(&ena_napi->dim, dim_sample); + + rx_ring->per_napi_packets = 0; +} + +static void ena_unmask_interrupt(struct ena_ring *tx_ring, + struct ena_ring *rx_ring) +{ + u32 rx_interval = tx_ring->smoothed_interval; + struct ena_eth_io_intr_reg intr_reg; + + /* Rx ring can be NULL when for XDP tx queues which don't have an + * accompanying rx_ring pair. + */ + if (rx_ring) + rx_interval = ena_com_get_adaptive_moderation_enabled(rx_ring->ena_dev) ? + rx_ring->smoothed_interval : + ena_com_get_nonadaptive_moderation_interval_rx(rx_ring->ena_dev); + + /* Update intr register: rx intr delay, + * tx intr delay and interrupt unmask + */ + ena_com_update_intr_reg(&intr_reg, + rx_interval, + tx_ring->smoothed_interval, + true); + + ena_increase_stat(&tx_ring->tx_stats.unmask_interrupt, 1, + &tx_ring->syncp); + + /* It is a shared MSI-X. + * Tx and Rx CQ have pointer to it. + * So we use one of them to reach the intr reg + * The Tx ring is used because the rx_ring is NULL for XDP queues + */ + ena_com_unmask_intr(tx_ring->ena_com_io_cq, &intr_reg); +} + +static void ena_update_ring_numa_node(struct ena_ring *tx_ring, + struct ena_ring *rx_ring) +{ + int cpu = get_cpu(); + int numa_node; + + /* Check only one ring since the 2 rings are running on the same cpu */ + if (likely(tx_ring->cpu == cpu)) + goto out; + + tx_ring->cpu = cpu; + if (rx_ring) + rx_ring->cpu = cpu; + + numa_node = cpu_to_node(cpu); + + if (likely(tx_ring->numa_node == numa_node)) + goto out; + + put_cpu(); + + if (numa_node != NUMA_NO_NODE) { + ena_com_update_numa_node(tx_ring->ena_com_io_cq, numa_node); + tx_ring->numa_node = numa_node; + if (rx_ring) { + rx_ring->numa_node = numa_node; + ena_com_update_numa_node(rx_ring->ena_com_io_cq, + numa_node); + } + } + + return; +out: + put_cpu(); +} + +static int ena_clean_xdp_irq(struct ena_ring *xdp_ring, u32 budget) +{ + u32 total_done = 0; + u16 next_to_clean; + int tx_pkts = 0; + u16 req_id; + int rc; + + if (unlikely(!xdp_ring)) + return 0; + next_to_clean = xdp_ring->next_to_clean; + + while (tx_pkts < budget) { + struct ena_tx_buffer *tx_info; + struct xdp_frame *xdpf; + + rc = ena_com_tx_comp_req_id_get(xdp_ring->ena_com_io_cq, + &req_id); + if (rc) { + if (unlikely(rc == -EINVAL)) + handle_invalid_req_id(xdp_ring, req_id, NULL, + true); + break; + } + + /* validate that the request id points to a valid xdp_frame */ + rc = validate_xdp_req_id(xdp_ring, req_id); + if (rc) + break; + + tx_info = &xdp_ring->tx_buffer_info[req_id]; + xdpf = tx_info->xdpf; + + tx_info->xdpf = NULL; + tx_info->last_jiffies = 0; + ena_unmap_tx_buff(xdp_ring, tx_info); + + netif_dbg(xdp_ring->adapter, tx_done, xdp_ring->netdev, + "tx_poll: q %d skb %p completed\n", xdp_ring->qid, + xdpf); + + tx_pkts++; + total_done += tx_info->tx_descs; + + xdp_return_frame(xdpf); + xdp_ring->free_ids[next_to_clean] = req_id; + next_to_clean = ENA_TX_RING_IDX_NEXT(next_to_clean, + xdp_ring->ring_size); + } + + xdp_ring->next_to_clean = next_to_clean; + ena_com_comp_ack(xdp_ring->ena_com_io_sq, total_done); + ena_com_update_dev_comp_head(xdp_ring->ena_com_io_cq); + + netif_dbg(xdp_ring->adapter, tx_done, xdp_ring->netdev, + "tx_poll: q %d done. total pkts: %d\n", + xdp_ring->qid, tx_pkts); + + return tx_pkts; +} + +static int ena_io_poll(struct napi_struct *napi, int budget) +{ + struct ena_napi *ena_napi = container_of(napi, struct ena_napi, napi); + struct ena_ring *tx_ring, *rx_ring; + int tx_work_done; + int rx_work_done = 0; + int tx_budget; + int napi_comp_call = 0; + int ret; + + tx_ring = ena_napi->tx_ring; + rx_ring = ena_napi->rx_ring; + + tx_budget = tx_ring->ring_size / ENA_TX_POLL_BUDGET_DIVIDER; + + if (!test_bit(ENA_FLAG_DEV_UP, &tx_ring->adapter->flags) || + test_bit(ENA_FLAG_TRIGGER_RESET, &tx_ring->adapter->flags)) { + napi_complete_done(napi, 0); + return 0; + } + + tx_work_done = ena_clean_tx_irq(tx_ring, tx_budget); + /* On netpoll the budget is zero and the handler should only clean the + * tx completions. + */ + if (likely(budget)) + rx_work_done = ena_clean_rx_irq(rx_ring, napi, budget); + + /* If the device is about to reset or down, avoid unmask + * the interrupt and return 0 so NAPI won't reschedule + */ + if (unlikely(!test_bit(ENA_FLAG_DEV_UP, &tx_ring->adapter->flags) || + test_bit(ENA_FLAG_TRIGGER_RESET, &tx_ring->adapter->flags))) { + napi_complete_done(napi, 0); + ret = 0; + + } else if ((budget > rx_work_done) && (tx_budget > tx_work_done)) { + napi_comp_call = 1; + + /* Update numa and unmask the interrupt only when schedule + * from the interrupt context (vs from sk_busy_loop) + */ + if (napi_complete_done(napi, rx_work_done) && + READ_ONCE(ena_napi->interrupts_masked)) { + smp_rmb(); /* make sure interrupts_masked is read */ + WRITE_ONCE(ena_napi->interrupts_masked, false); + /* We apply adaptive moderation on Rx path only. + * Tx uses static interrupt moderation. + */ + if (ena_com_get_adaptive_moderation_enabled(rx_ring->ena_dev)) + ena_adjust_adaptive_rx_intr_moderation(ena_napi); + + ena_update_ring_numa_node(tx_ring, rx_ring); + ena_unmask_interrupt(tx_ring, rx_ring); + } + + ret = rx_work_done; + } else { + ret = budget; + } + + u64_stats_update_begin(&tx_ring->syncp); + tx_ring->tx_stats.napi_comp += napi_comp_call; + tx_ring->tx_stats.tx_poll++; + u64_stats_update_end(&tx_ring->syncp); + + tx_ring->tx_stats.last_napi_jiffies = jiffies; + + return ret; +} + +static irqreturn_t ena_intr_msix_mgmnt(int irq, void *data) +{ + struct ena_adapter *adapter = (struct ena_adapter *)data; + + ena_com_admin_q_comp_intr_handler(adapter->ena_dev); + + /* Don't call the aenq handler before probe is done */ + if (likely(test_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags))) + ena_com_aenq_intr_handler(adapter->ena_dev, data); + + return IRQ_HANDLED; +} + +/* ena_intr_msix_io - MSI-X Interrupt Handler for Tx/Rx + * @irq: interrupt number + * @data: pointer to a network interface private napi device structure + */ +static irqreturn_t ena_intr_msix_io(int irq, void *data) +{ + struct ena_napi *ena_napi = data; + + /* Used to check HW health */ + WRITE_ONCE(ena_napi->first_interrupt, true); + + WRITE_ONCE(ena_napi->interrupts_masked, true); + smp_wmb(); /* write interrupts_masked before calling napi */ + + napi_schedule_irqoff(&ena_napi->napi); + + return IRQ_HANDLED; +} + +/* Reserve a single MSI-X vector for management (admin + aenq). + * plus reserve one vector for each potential io queue. + * the number of potential io queues is the minimum of what the device + * supports and the number of vCPUs. + */ +static int ena_enable_msix(struct ena_adapter *adapter) +{ + int msix_vecs, irq_cnt; + + if (test_bit(ENA_FLAG_MSIX_ENABLED, &adapter->flags)) { + netif_err(adapter, probe, adapter->netdev, + "Error, MSI-X is already enabled\n"); + return -EPERM; + } + + /* Reserved the max msix vectors we might need */ + msix_vecs = ENA_MAX_MSIX_VEC(adapter->max_num_io_queues); + netif_dbg(adapter, probe, adapter->netdev, + "Trying to enable MSI-X, vectors %d\n", msix_vecs); + + irq_cnt = pci_alloc_irq_vectors(adapter->pdev, ENA_MIN_MSIX_VEC, + msix_vecs, PCI_IRQ_MSIX); + + if (irq_cnt < 0) { + netif_err(adapter, probe, adapter->netdev, + "Failed to enable MSI-X. irq_cnt %d\n", irq_cnt); + return -ENOSPC; + } + + if (irq_cnt != msix_vecs) { + netif_notice(adapter, probe, adapter->netdev, + "Enable only %d MSI-X (out of %d), reduce the number of queues\n", + irq_cnt, msix_vecs); + adapter->num_io_queues = irq_cnt - ENA_ADMIN_MSIX_VEC; + } + + if (ena_init_rx_cpu_rmap(adapter)) + netif_warn(adapter, probe, adapter->netdev, + "Failed to map IRQs to CPUs\n"); + + adapter->msix_vecs = irq_cnt; + set_bit(ENA_FLAG_MSIX_ENABLED, &adapter->flags); + + return 0; +} + +static void ena_setup_mgmnt_intr(struct ena_adapter *adapter) +{ + u32 cpu; + + snprintf(adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].name, + ENA_IRQNAME_SIZE, "ena-mgmnt@pci:%s", + pci_name(adapter->pdev)); + adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].handler = + ena_intr_msix_mgmnt; + adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].data = adapter; + adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].vector = + pci_irq_vector(adapter->pdev, ENA_MGMNT_IRQ_IDX); + cpu = cpumask_first(cpu_online_mask); + adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].cpu = cpu; + cpumask_set_cpu(cpu, + &adapter->irq_tbl[ENA_MGMNT_IRQ_IDX].affinity_hint_mask); +} + +static void ena_setup_io_intr(struct ena_adapter *adapter) +{ + struct net_device *netdev; + int irq_idx, i, cpu; + int io_queue_count; + + netdev = adapter->netdev; + io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + + for (i = 0; i < io_queue_count; i++) { + irq_idx = ENA_IO_IRQ_IDX(i); + cpu = i % num_online_cpus(); + + snprintf(adapter->irq_tbl[irq_idx].name, ENA_IRQNAME_SIZE, + "%s-Tx-Rx-%d", netdev->name, i); + adapter->irq_tbl[irq_idx].handler = ena_intr_msix_io; + adapter->irq_tbl[irq_idx].data = &adapter->ena_napi[i]; + adapter->irq_tbl[irq_idx].vector = + pci_irq_vector(adapter->pdev, irq_idx); + adapter->irq_tbl[irq_idx].cpu = cpu; + + cpumask_set_cpu(cpu, + &adapter->irq_tbl[irq_idx].affinity_hint_mask); + } +} + +static int ena_request_mgmnt_irq(struct ena_adapter *adapter) +{ + unsigned long flags = 0; + struct ena_irq *irq; + int rc; + + irq = &adapter->irq_tbl[ENA_MGMNT_IRQ_IDX]; + rc = request_irq(irq->vector, irq->handler, flags, irq->name, + irq->data); + if (rc) { + netif_err(adapter, probe, adapter->netdev, + "Failed to request admin irq\n"); + return rc; + } + + netif_dbg(adapter, probe, adapter->netdev, + "Set affinity hint of mgmnt irq.to 0x%lx (irq vector: %d)\n", + irq->affinity_hint_mask.bits[0], irq->vector); + + irq_set_affinity_hint(irq->vector, &irq->affinity_hint_mask); + + return rc; +} + +static int ena_request_io_irq(struct ena_adapter *adapter) +{ + u32 io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + unsigned long flags = 0; + struct ena_irq *irq; + int rc = 0, i, k; + + if (!test_bit(ENA_FLAG_MSIX_ENABLED, &adapter->flags)) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to request I/O IRQ: MSI-X is not enabled\n"); + return -EINVAL; + } + + for (i = ENA_IO_IRQ_FIRST_IDX; i < ENA_MAX_MSIX_VEC(io_queue_count); i++) { + irq = &adapter->irq_tbl[i]; + rc = request_irq(irq->vector, irq->handler, flags, irq->name, + irq->data); + if (rc) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to request I/O IRQ. index %d rc %d\n", + i, rc); + goto err; + } + + netif_dbg(adapter, ifup, adapter->netdev, + "Set affinity hint of irq. index %d to 0x%lx (irq vector: %d)\n", + i, irq->affinity_hint_mask.bits[0], irq->vector); + + irq_set_affinity_hint(irq->vector, &irq->affinity_hint_mask); + } + + return rc; + +err: + for (k = ENA_IO_IRQ_FIRST_IDX; k < i; k++) { + irq = &adapter->irq_tbl[k]; + free_irq(irq->vector, irq->data); + } + + return rc; +} + +static void ena_free_mgmnt_irq(struct ena_adapter *adapter) +{ + struct ena_irq *irq; + + irq = &adapter->irq_tbl[ENA_MGMNT_IRQ_IDX]; + synchronize_irq(irq->vector); + irq_set_affinity_hint(irq->vector, NULL); + free_irq(irq->vector, irq->data); +} + +static void ena_free_io_irq(struct ena_adapter *adapter) +{ + u32 io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + struct ena_irq *irq; + int i; + +#ifdef CONFIG_RFS_ACCEL + if (adapter->msix_vecs >= 1) { + free_irq_cpu_rmap(adapter->netdev->rx_cpu_rmap); + adapter->netdev->rx_cpu_rmap = NULL; + } +#endif /* CONFIG_RFS_ACCEL */ + + for (i = ENA_IO_IRQ_FIRST_IDX; i < ENA_MAX_MSIX_VEC(io_queue_count); i++) { + irq = &adapter->irq_tbl[i]; + irq_set_affinity_hint(irq->vector, NULL); + free_irq(irq->vector, irq->data); + } +} + +static void ena_disable_msix(struct ena_adapter *adapter) +{ + if (test_and_clear_bit(ENA_FLAG_MSIX_ENABLED, &adapter->flags)) + pci_free_irq_vectors(adapter->pdev); +} + +static void ena_disable_io_intr_sync(struct ena_adapter *adapter) +{ + u32 io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + int i; + + if (!netif_running(adapter->netdev)) + return; + + for (i = ENA_IO_IRQ_FIRST_IDX; i < ENA_MAX_MSIX_VEC(io_queue_count); i++) + synchronize_irq(adapter->irq_tbl[i].vector); +} + +static void ena_del_napi_in_range(struct ena_adapter *adapter, + int first_index, + int count) +{ + int i; + + for (i = first_index; i < first_index + count; i++) { + netif_napi_del(&adapter->ena_napi[i].napi); + + WARN_ON(!ENA_IS_XDP_INDEX(adapter, i) && + adapter->ena_napi[i].xdp_ring); + } +} + +static void ena_init_napi_in_range(struct ena_adapter *adapter, + int first_index, int count) +{ + int i; + + for (i = first_index; i < first_index + count; i++) { + struct ena_napi *napi = &adapter->ena_napi[i]; + + netif_napi_add(adapter->netdev, &napi->napi, + ENA_IS_XDP_INDEX(adapter, i) ? ena_xdp_io_poll : ena_io_poll); + + if (!ENA_IS_XDP_INDEX(adapter, i)) { + napi->rx_ring = &adapter->rx_ring[i]; + napi->tx_ring = &adapter->tx_ring[i]; + } else { + napi->xdp_ring = &adapter->tx_ring[i]; + } + napi->qid = i; + } +} + +static void ena_napi_disable_in_range(struct ena_adapter *adapter, + int first_index, + int count) +{ + int i; + + for (i = first_index; i < first_index + count; i++) + napi_disable(&adapter->ena_napi[i].napi); +} + +static void ena_napi_enable_in_range(struct ena_adapter *adapter, + int first_index, + int count) +{ + int i; + + for (i = first_index; i < first_index + count; i++) + napi_enable(&adapter->ena_napi[i].napi); +} + +/* Configure the Rx forwarding */ +static int ena_rss_configure(struct ena_adapter *adapter) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int rc; + + /* In case the RSS table wasn't initialized by probe */ + if (!ena_dev->rss.tbl_log_size) { + rc = ena_rss_init_default(adapter); + if (rc && (rc != -EOPNOTSUPP)) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to init RSS rc: %d\n", rc); + return rc; + } + } + + /* Set indirect table */ + rc = ena_com_indirect_table_set(ena_dev); + if (unlikely(rc && rc != -EOPNOTSUPP)) + return rc; + + /* Configure hash function (if supported) */ + rc = ena_com_set_hash_function(ena_dev); + if (unlikely(rc && (rc != -EOPNOTSUPP))) + return rc; + + /* Configure hash inputs (if supported) */ + rc = ena_com_set_hash_ctrl(ena_dev); + if (unlikely(rc && (rc != -EOPNOTSUPP))) + return rc; + + return 0; +} + +static int ena_up_complete(struct ena_adapter *adapter) +{ + int rc; + + rc = ena_rss_configure(adapter); + if (rc) + return rc; + + ena_change_mtu(adapter->netdev, adapter->netdev->mtu); + + ena_refill_all_rx_bufs(adapter); + + /* enable transmits */ + netif_tx_start_all_queues(adapter->netdev); + + ena_napi_enable_in_range(adapter, + 0, + adapter->xdp_num_queues + adapter->num_io_queues); + + return 0; +} + +static int ena_create_io_tx_queue(struct ena_adapter *adapter, int qid) +{ + struct ena_com_create_io_ctx ctx; + struct ena_com_dev *ena_dev; + struct ena_ring *tx_ring; + u32 msix_vector; + u16 ena_qid; + int rc; + + ena_dev = adapter->ena_dev; + + tx_ring = &adapter->tx_ring[qid]; + msix_vector = ENA_IO_IRQ_IDX(qid); + ena_qid = ENA_IO_TXQ_IDX(qid); + + memset(&ctx, 0x0, sizeof(ctx)); + + ctx.direction = ENA_COM_IO_QUEUE_DIRECTION_TX; + ctx.qid = ena_qid; + ctx.mem_queue_type = ena_dev->tx_mem_queue_type; + ctx.msix_vector = msix_vector; + ctx.queue_size = tx_ring->ring_size; + ctx.numa_node = tx_ring->numa_node; + + rc = ena_com_create_io_queue(ena_dev, &ctx); + if (rc) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to create I/O TX queue num %d rc: %d\n", + qid, rc); + return rc; + } + + rc = ena_com_get_io_handlers(ena_dev, ena_qid, + &tx_ring->ena_com_io_sq, + &tx_ring->ena_com_io_cq); + if (rc) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to get TX queue handlers. TX queue num %d rc: %d\n", + qid, rc); + ena_com_destroy_io_queue(ena_dev, ena_qid); + return rc; + } + + ena_com_update_numa_node(tx_ring->ena_com_io_cq, ctx.numa_node); + return rc; +} + +static int ena_create_io_tx_queues_in_range(struct ena_adapter *adapter, + int first_index, int count) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int rc, i; + + for (i = first_index; i < first_index + count; i++) { + rc = ena_create_io_tx_queue(adapter, i); + if (rc) + goto create_err; + } + + return 0; + +create_err: + while (i-- > first_index) + ena_com_destroy_io_queue(ena_dev, ENA_IO_TXQ_IDX(i)); + + return rc; +} + +static int ena_create_io_rx_queue(struct ena_adapter *adapter, int qid) +{ + struct ena_com_dev *ena_dev; + struct ena_com_create_io_ctx ctx; + struct ena_ring *rx_ring; + u32 msix_vector; + u16 ena_qid; + int rc; + + ena_dev = adapter->ena_dev; + + rx_ring = &adapter->rx_ring[qid]; + msix_vector = ENA_IO_IRQ_IDX(qid); + ena_qid = ENA_IO_RXQ_IDX(qid); + + memset(&ctx, 0x0, sizeof(ctx)); + + ctx.qid = ena_qid; + ctx.direction = ENA_COM_IO_QUEUE_DIRECTION_RX; + ctx.mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_HOST; + ctx.msix_vector = msix_vector; + ctx.queue_size = rx_ring->ring_size; + ctx.numa_node = rx_ring->numa_node; + + rc = ena_com_create_io_queue(ena_dev, &ctx); + if (rc) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to create I/O RX queue num %d rc: %d\n", + qid, rc); + return rc; + } + + rc = ena_com_get_io_handlers(ena_dev, ena_qid, + &rx_ring->ena_com_io_sq, + &rx_ring->ena_com_io_cq); + if (rc) { + netif_err(adapter, ifup, adapter->netdev, + "Failed to get RX queue handlers. RX queue num %d rc: %d\n", + qid, rc); + goto err; + } + + ena_com_update_numa_node(rx_ring->ena_com_io_cq, ctx.numa_node); + + return rc; +err: + ena_com_destroy_io_queue(ena_dev, ena_qid); + return rc; +} + +static int ena_create_all_io_rx_queues(struct ena_adapter *adapter) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int rc, i; + + for (i = 0; i < adapter->num_io_queues; i++) { + rc = ena_create_io_rx_queue(adapter, i); + if (rc) + goto create_err; + INIT_WORK(&adapter->ena_napi[i].dim.work, ena_dim_work); + } + + return 0; + +create_err: + while (i--) { + cancel_work_sync(&adapter->ena_napi[i].dim.work); + ena_com_destroy_io_queue(ena_dev, ENA_IO_RXQ_IDX(i)); + } + + return rc; +} + +static void set_io_rings_size(struct ena_adapter *adapter, + int new_tx_size, + int new_rx_size) +{ + int i; + + for (i = 0; i < adapter->num_io_queues; i++) { + adapter->tx_ring[i].ring_size = new_tx_size; + adapter->rx_ring[i].ring_size = new_rx_size; + } +} + +/* This function allows queue allocation to backoff when the system is + * low on memory. If there is not enough memory to allocate io queues + * the driver will try to allocate smaller queues. + * + * The backoff algorithm is as follows: + * 1. Try to allocate TX and RX and if successful. + * 1.1. return success + * + * 2. Divide by 2 the size of the larger of RX and TX queues (or both if their size is the same). + * + * 3. If TX or RX is smaller than 256 + * 3.1. return failure. + * 4. else + * 4.1. go back to 1. + */ +static int create_queues_with_size_backoff(struct ena_adapter *adapter) +{ + int rc, cur_rx_ring_size, cur_tx_ring_size; + int new_rx_ring_size, new_tx_ring_size; + + /* current queue sizes might be set to smaller than the requested + * ones due to past queue allocation failures. + */ + set_io_rings_size(adapter, adapter->requested_tx_ring_size, + adapter->requested_rx_ring_size); + + while (1) { + if (ena_xdp_present(adapter)) { + rc = ena_setup_and_create_all_xdp_queues(adapter); + + if (rc) + goto err_setup_tx; + } + rc = ena_setup_tx_resources_in_range(adapter, + 0, + adapter->num_io_queues); + if (rc) + goto err_setup_tx; + + rc = ena_create_io_tx_queues_in_range(adapter, + 0, + adapter->num_io_queues); + if (rc) + goto err_create_tx_queues; + + rc = ena_setup_all_rx_resources(adapter); + if (rc) + goto err_setup_rx; + + rc = ena_create_all_io_rx_queues(adapter); + if (rc) + goto err_create_rx_queues; + + return 0; + +err_create_rx_queues: + ena_free_all_io_rx_resources(adapter); +err_setup_rx: + ena_destroy_all_tx_queues(adapter); +err_create_tx_queues: + ena_free_all_io_tx_resources(adapter); +err_setup_tx: + if (rc != -ENOMEM) { + netif_err(adapter, ifup, adapter->netdev, + "Queue creation failed with error code %d\n", + rc); + return rc; + } + + cur_tx_ring_size = adapter->tx_ring[0].ring_size; + cur_rx_ring_size = adapter->rx_ring[0].ring_size; + + netif_err(adapter, ifup, adapter->netdev, + "Not enough memory to create queues with sizes TX=%d, RX=%d\n", + cur_tx_ring_size, cur_rx_ring_size); + + new_tx_ring_size = cur_tx_ring_size; + new_rx_ring_size = cur_rx_ring_size; + + /* Decrease the size of the larger queue, or + * decrease both if they are the same size. + */ + if (cur_rx_ring_size <= cur_tx_ring_size) + new_tx_ring_size = cur_tx_ring_size / 2; + if (cur_rx_ring_size >= cur_tx_ring_size) + new_rx_ring_size = cur_rx_ring_size / 2; + + if (new_tx_ring_size < ENA_MIN_RING_SIZE || + new_rx_ring_size < ENA_MIN_RING_SIZE) { + netif_err(adapter, ifup, adapter->netdev, + "Queue creation failed with the smallest possible queue size of %d for both queues. Not retrying with smaller queues\n", + ENA_MIN_RING_SIZE); + return rc; + } + + netif_err(adapter, ifup, adapter->netdev, + "Retrying queue creation with sizes TX=%d, RX=%d\n", + new_tx_ring_size, + new_rx_ring_size); + + set_io_rings_size(adapter, new_tx_ring_size, + new_rx_ring_size); + } +} + +static int ena_up(struct ena_adapter *adapter) +{ + int io_queue_count, rc, i; + + netif_dbg(adapter, ifup, adapter->netdev, "%s\n", __func__); + + io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + ena_setup_io_intr(adapter); + + /* napi poll functions should be initialized before running + * request_irq(), to handle a rare condition where there is a pending + * interrupt, causing the ISR to fire immediately while the poll + * function wasn't set yet, causing a null dereference + */ + ena_init_napi_in_range(adapter, 0, io_queue_count); + + rc = ena_request_io_irq(adapter); + if (rc) + goto err_req_irq; + + rc = create_queues_with_size_backoff(adapter); + if (rc) + goto err_create_queues_with_backoff; + + rc = ena_up_complete(adapter); + if (rc) + goto err_up; + + if (test_bit(ENA_FLAG_LINK_UP, &adapter->flags)) + netif_carrier_on(adapter->netdev); + + ena_increase_stat(&adapter->dev_stats.interface_up, 1, + &adapter->syncp); + + set_bit(ENA_FLAG_DEV_UP, &adapter->flags); + + /* Enable completion queues interrupt */ + for (i = 0; i < adapter->num_io_queues; i++) + ena_unmask_interrupt(&adapter->tx_ring[i], + &adapter->rx_ring[i]); + + /* schedule napi in case we had pending packets + * from the last time we disable napi + */ + for (i = 0; i < io_queue_count; i++) + napi_schedule(&adapter->ena_napi[i].napi); + + return rc; + +err_up: + ena_destroy_all_tx_queues(adapter); + ena_free_all_io_tx_resources(adapter); + ena_destroy_all_rx_queues(adapter); + ena_free_all_io_rx_resources(adapter); +err_create_queues_with_backoff: + ena_free_io_irq(adapter); +err_req_irq: + ena_del_napi_in_range(adapter, 0, io_queue_count); + + return rc; +} + +static void ena_down(struct ena_adapter *adapter) +{ + int io_queue_count = adapter->num_io_queues + adapter->xdp_num_queues; + + netif_info(adapter, ifdown, adapter->netdev, "%s\n", __func__); + + clear_bit(ENA_FLAG_DEV_UP, &adapter->flags); + + ena_increase_stat(&adapter->dev_stats.interface_down, 1, + &adapter->syncp); + + netif_carrier_off(adapter->netdev); + netif_tx_disable(adapter->netdev); + + /* After this point the napi handler won't enable the tx queue */ + ena_napi_disable_in_range(adapter, 0, io_queue_count); + + /* After destroy the queue there won't be any new interrupts */ + + if (test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags)) { + int rc; + + rc = ena_com_dev_reset(adapter->ena_dev, adapter->reset_reason); + if (rc) + netif_err(adapter, ifdown, adapter->netdev, + "Device reset failed\n"); + /* stop submitting admin commands on a device that was reset */ + ena_com_set_admin_running_state(adapter->ena_dev, false); + } + + ena_destroy_all_io_queues(adapter); + + ena_disable_io_intr_sync(adapter); + ena_free_io_irq(adapter); + ena_del_napi_in_range(adapter, 0, io_queue_count); + + ena_free_all_tx_bufs(adapter); + ena_free_all_rx_bufs(adapter); + ena_free_all_io_tx_resources(adapter); + ena_free_all_io_rx_resources(adapter); +} + +/* ena_open - Called when a network interface is made active + * @netdev: network interface device structure + * + * Returns 0 on success, negative value on failure + * + * The open entry point is called when a network interface is made + * active by the system (IFF_UP). At this point all resources needed + * for transmit and receive operations are allocated, the interrupt + * handler is registered with the OS, the watchdog timer is started, + * and the stack is notified that the interface is ready. + */ +static int ena_open(struct net_device *netdev) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + int rc; + + /* Notify the stack of the actual queue counts. */ + rc = netif_set_real_num_tx_queues(netdev, adapter->num_io_queues); + if (rc) { + netif_err(adapter, ifup, netdev, "Can't set num tx queues\n"); + return rc; + } + + rc = netif_set_real_num_rx_queues(netdev, adapter->num_io_queues); + if (rc) { + netif_err(adapter, ifup, netdev, "Can't set num rx queues\n"); + return rc; + } + + rc = ena_up(adapter); + if (rc) + return rc; + + return rc; +} + +/* ena_close - Disables a network interface + * @netdev: network interface device structure + * + * Returns 0, this is not allowed to fail + * + * The close entry point is called when an interface is de-activated + * by the OS. The hardware is still under the drivers control, but + * needs to be disabled. A global MAC reset is issued to stop the + * hardware, and all transmit and receive resources are freed. + */ +static int ena_close(struct net_device *netdev) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + + netif_dbg(adapter, ifdown, netdev, "%s\n", __func__); + + if (!test_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags)) + return 0; + + if (test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + ena_down(adapter); + + /* Check for device status and issue reset if needed*/ + check_for_admin_com_state(adapter); + if (unlikely(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags))) { + netif_err(adapter, ifdown, adapter->netdev, + "Destroy failure, restarting device\n"); + ena_dump_stats_to_dmesg(adapter); + /* rtnl lock already obtained in dev_ioctl() layer */ + ena_destroy_device(adapter, false); + ena_restore_device(adapter); + } + + return 0; +} + +int ena_update_queue_params(struct ena_adapter *adapter, + u32 new_tx_size, + u32 new_rx_size, + u32 new_llq_header_len) +{ + bool dev_was_up, large_llq_changed = false; + int rc = 0; + + dev_was_up = test_bit(ENA_FLAG_DEV_UP, &adapter->flags); + ena_close(adapter->netdev); + adapter->requested_tx_ring_size = new_tx_size; + adapter->requested_rx_ring_size = new_rx_size; + ena_init_io_rings(adapter, + 0, + adapter->xdp_num_queues + + adapter->num_io_queues); + + large_llq_changed = adapter->ena_dev->tx_mem_queue_type == + ENA_ADMIN_PLACEMENT_POLICY_DEV; + large_llq_changed &= + new_llq_header_len != adapter->ena_dev->tx_max_header_size; + + /* a check that the configuration is valid is done by caller */ + if (large_llq_changed) { + adapter->large_llq_header_enabled = !adapter->large_llq_header_enabled; + + ena_destroy_device(adapter, false); + rc = ena_restore_device(adapter); + } + + return dev_was_up && !rc ? ena_up(adapter) : rc; +} + +int ena_set_rx_copybreak(struct ena_adapter *adapter, u32 rx_copybreak) +{ + struct ena_ring *rx_ring; + int i; + + if (rx_copybreak > min_t(u16, adapter->netdev->mtu, ENA_PAGE_SIZE)) + return -EINVAL; + + adapter->rx_copybreak = rx_copybreak; + + for (i = 0; i < adapter->num_io_queues; i++) { + rx_ring = &adapter->rx_ring[i]; + rx_ring->rx_copybreak = rx_copybreak; + } + + return 0; +} + +int ena_update_queue_count(struct ena_adapter *adapter, u32 new_channel_count) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + int prev_channel_count; + bool dev_was_up; + + dev_was_up = test_bit(ENA_FLAG_DEV_UP, &adapter->flags); + ena_close(adapter->netdev); + prev_channel_count = adapter->num_io_queues; + adapter->num_io_queues = new_channel_count; + if (ena_xdp_present(adapter) && + ena_xdp_allowed(adapter) == ENA_XDP_ALLOWED) { + adapter->xdp_first_ring = new_channel_count; + adapter->xdp_num_queues = new_channel_count; + if (prev_channel_count > new_channel_count) + ena_xdp_exchange_program_rx_in_range(adapter, + NULL, + new_channel_count, + prev_channel_count); + else + ena_xdp_exchange_program_rx_in_range(adapter, + adapter->xdp_bpf_prog, + prev_channel_count, + new_channel_count); + } + + /* We need to destroy the rss table so that the indirection + * table will be reinitialized by ena_up() + */ + ena_com_rss_destroy(ena_dev); + ena_init_io_rings(adapter, + 0, + adapter->xdp_num_queues + + adapter->num_io_queues); + return dev_was_up ? ena_open(adapter->netdev) : 0; +} + +static void ena_tx_csum(struct ena_com_tx_ctx *ena_tx_ctx, + struct sk_buff *skb, + bool disable_meta_caching) +{ + u32 mss = skb_shinfo(skb)->gso_size; + struct ena_com_tx_meta *ena_meta = &ena_tx_ctx->ena_meta; + u8 l4_protocol = 0; + + if ((skb->ip_summed == CHECKSUM_PARTIAL) || mss) { + ena_tx_ctx->l4_csum_enable = 1; + if (mss) { + ena_tx_ctx->tso_enable = 1; + ena_meta->l4_hdr_len = tcp_hdr(skb)->doff; + ena_tx_ctx->l4_csum_partial = 0; + } else { + ena_tx_ctx->tso_enable = 0; + ena_meta->l4_hdr_len = 0; + ena_tx_ctx->l4_csum_partial = 1; + } + + switch (ip_hdr(skb)->version) { + case IPVERSION: + ena_tx_ctx->l3_proto = ENA_ETH_IO_L3_PROTO_IPV4; + if (ip_hdr(skb)->frag_off & htons(IP_DF)) + ena_tx_ctx->df = 1; + if (mss) + ena_tx_ctx->l3_csum_enable = 1; + l4_protocol = ip_hdr(skb)->protocol; + break; + case 6: + ena_tx_ctx->l3_proto = ENA_ETH_IO_L3_PROTO_IPV6; + l4_protocol = ipv6_hdr(skb)->nexthdr; + break; + default: + break; + } + + if (l4_protocol == IPPROTO_TCP) + ena_tx_ctx->l4_proto = ENA_ETH_IO_L4_PROTO_TCP; + else + ena_tx_ctx->l4_proto = ENA_ETH_IO_L4_PROTO_UDP; + + ena_meta->mss = mss; + ena_meta->l3_hdr_len = skb_network_header_len(skb); + ena_meta->l3_hdr_offset = skb_network_offset(skb); + ena_tx_ctx->meta_valid = 1; + } else if (disable_meta_caching) { + memset(ena_meta, 0, sizeof(*ena_meta)); + ena_tx_ctx->meta_valid = 1; + } else { + ena_tx_ctx->meta_valid = 0; + } +} + +static int ena_check_and_linearize_skb(struct ena_ring *tx_ring, + struct sk_buff *skb) +{ + int num_frags, header_len, rc; + + num_frags = skb_shinfo(skb)->nr_frags; + header_len = skb_headlen(skb); + + if (num_frags < tx_ring->sgl_size) + return 0; + + if ((num_frags == tx_ring->sgl_size) && + (header_len < tx_ring->tx_max_header_size)) + return 0; + + ena_increase_stat(&tx_ring->tx_stats.linearize, 1, &tx_ring->syncp); + + rc = skb_linearize(skb); + if (unlikely(rc)) { + ena_increase_stat(&tx_ring->tx_stats.linearize_failed, 1, + &tx_ring->syncp); + } + + return rc; +} + +static int ena_tx_map_skb(struct ena_ring *tx_ring, + struct ena_tx_buffer *tx_info, + struct sk_buff *skb, + void **push_hdr, + u16 *header_len) +{ + struct ena_adapter *adapter = tx_ring->adapter; + struct ena_com_buf *ena_buf; + dma_addr_t dma; + u32 skb_head_len, frag_len, last_frag; + u16 push_len = 0; + u16 delta = 0; + int i = 0; + + skb_head_len = skb_headlen(skb); + tx_info->skb = skb; + ena_buf = tx_info->bufs; + + if (tx_ring->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + /* When the device is LLQ mode, the driver will copy + * the header into the device memory space. + * the ena_com layer assume the header is in a linear + * memory space. + * This assumption might be wrong since part of the header + * can be in the fragmented buffers. + * Use skb_header_pointer to make sure the header is in a + * linear memory space. + */ + + push_len = min_t(u32, skb->len, tx_ring->tx_max_header_size); + *push_hdr = skb_header_pointer(skb, 0, push_len, + tx_ring->push_buf_intermediate_buf); + *header_len = push_len; + if (unlikely(skb->data != *push_hdr)) { + ena_increase_stat(&tx_ring->tx_stats.llq_buffer_copy, 1, + &tx_ring->syncp); + + delta = push_len - skb_head_len; + } + } else { + *push_hdr = NULL; + *header_len = min_t(u32, skb_head_len, + tx_ring->tx_max_header_size); + } + + netif_dbg(adapter, tx_queued, adapter->netdev, + "skb: %p header_buf->vaddr: %p push_len: %d\n", skb, + *push_hdr, push_len); + + if (skb_head_len > push_len) { + dma = dma_map_single(tx_ring->dev, skb->data + push_len, + skb_head_len - push_len, DMA_TO_DEVICE); + if (unlikely(dma_mapping_error(tx_ring->dev, dma))) + goto error_report_dma_error; + + ena_buf->paddr = dma; + ena_buf->len = skb_head_len - push_len; + + ena_buf++; + tx_info->num_of_bufs++; + tx_info->map_linear_data = 1; + } else { + tx_info->map_linear_data = 0; + } + + last_frag = skb_shinfo(skb)->nr_frags; + + for (i = 0; i < last_frag; i++) { + const skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; + + frag_len = skb_frag_size(frag); + + if (unlikely(delta >= frag_len)) { + delta -= frag_len; + continue; + } + + dma = skb_frag_dma_map(tx_ring->dev, frag, delta, + frag_len - delta, DMA_TO_DEVICE); + if (unlikely(dma_mapping_error(tx_ring->dev, dma))) + goto error_report_dma_error; + + ena_buf->paddr = dma; + ena_buf->len = frag_len - delta; + ena_buf++; + tx_info->num_of_bufs++; + delta = 0; + } + + return 0; + +error_report_dma_error: + ena_increase_stat(&tx_ring->tx_stats.dma_mapping_err, 1, + &tx_ring->syncp); + netif_warn(adapter, tx_queued, adapter->netdev, "Failed to map skb\n"); + + tx_info->skb = NULL; + + tx_info->num_of_bufs += i; + ena_unmap_tx_buff(tx_ring, tx_info); + + return -EINVAL; +} + +/* Called with netif_tx_lock. */ +static netdev_tx_t ena_start_xmit(struct sk_buff *skb, struct net_device *dev) +{ + struct ena_adapter *adapter = netdev_priv(dev); + struct ena_tx_buffer *tx_info; + struct ena_com_tx_ctx ena_tx_ctx; + struct ena_ring *tx_ring; + struct netdev_queue *txq; + void *push_hdr; + u16 next_to_use, req_id, header_len; + int qid, rc; + + netif_dbg(adapter, tx_queued, dev, "%s skb %p\n", __func__, skb); + /* Determine which tx ring we will be placed on */ + qid = skb_get_queue_mapping(skb); + tx_ring = &adapter->tx_ring[qid]; + txq = netdev_get_tx_queue(dev, qid); + + rc = ena_check_and_linearize_skb(tx_ring, skb); + if (unlikely(rc)) + goto error_drop_packet; + + skb_tx_timestamp(skb); + + next_to_use = tx_ring->next_to_use; + req_id = tx_ring->free_ids[next_to_use]; + tx_info = &tx_ring->tx_buffer_info[req_id]; + tx_info->num_of_bufs = 0; + + WARN(tx_info->skb, "SKB isn't NULL req_id %d\n", req_id); + + rc = ena_tx_map_skb(tx_ring, tx_info, skb, &push_hdr, &header_len); + if (unlikely(rc)) + goto error_drop_packet; + + memset(&ena_tx_ctx, 0x0, sizeof(struct ena_com_tx_ctx)); + ena_tx_ctx.ena_bufs = tx_info->bufs; + ena_tx_ctx.push_header = push_hdr; + ena_tx_ctx.num_bufs = tx_info->num_of_bufs; + ena_tx_ctx.req_id = req_id; + ena_tx_ctx.header_len = header_len; + + /* set flags and meta data */ + ena_tx_csum(&ena_tx_ctx, skb, tx_ring->disable_meta_caching); + + rc = ena_xmit_common(dev, + tx_ring, + tx_info, + &ena_tx_ctx, + next_to_use, + skb->len); + if (rc) + goto error_unmap_dma; + + netdev_tx_sent_queue(txq, skb->len); + + /* stop the queue when no more space available, the packet can have up + * to sgl_size + 2. one for the meta descriptor and one for header + * (if the header is larger than tx_max_header_size). + */ + if (unlikely(!ena_com_sq_have_enough_space(tx_ring->ena_com_io_sq, + tx_ring->sgl_size + 2))) { + netif_dbg(adapter, tx_queued, dev, "%s stop queue %d\n", + __func__, qid); + + netif_tx_stop_queue(txq); + ena_increase_stat(&tx_ring->tx_stats.queue_stop, 1, + &tx_ring->syncp); + + /* There is a rare condition where this function decide to + * stop the queue but meanwhile clean_tx_irq updates + * next_to_completion and terminates. + * The queue will remain stopped forever. + * To solve this issue add a mb() to make sure that + * netif_tx_stop_queue() write is vissible before checking if + * there is additional space in the queue. + */ + smp_mb(); + + if (ena_com_sq_have_enough_space(tx_ring->ena_com_io_sq, + ENA_TX_WAKEUP_THRESH)) { + netif_tx_wake_queue(txq); + ena_increase_stat(&tx_ring->tx_stats.queue_wakeup, 1, + &tx_ring->syncp); + } + } + + if (netif_xmit_stopped(txq) || !netdev_xmit_more()) + /* trigger the dma engine. ena_ring_tx_doorbell() + * calls a memory barrier inside it. + */ + ena_ring_tx_doorbell(tx_ring); + + return NETDEV_TX_OK; + +error_unmap_dma: + ena_unmap_tx_buff(tx_ring, tx_info); + tx_info->skb = NULL; + +error_drop_packet: + dev_kfree_skb(skb); + return NETDEV_TX_OK; +} + +static u16 ena_select_queue(struct net_device *dev, struct sk_buff *skb, + struct net_device *sb_dev) +{ + u16 qid; + /* we suspect that this is good for in--kernel network services that + * want to loop incoming skb rx to tx in normal user generated traffic, + * most probably we will not get to this + */ + if (skb_rx_queue_recorded(skb)) + qid = skb_get_rx_queue(skb); + else + qid = netdev_pick_tx(dev, skb, NULL); + + return qid; +} + +static void ena_config_host_info(struct ena_com_dev *ena_dev, struct pci_dev *pdev) +{ + struct device *dev = &pdev->dev; + struct ena_admin_host_info *host_info; + int rc; + + /* Allocate only the host info */ + rc = ena_com_allocate_host_info(ena_dev); + if (rc) { + dev_err(dev, "Cannot allocate host info\n"); + return; + } + + host_info = ena_dev->host_attr.host_info; + + host_info->bdf = pci_dev_id(pdev); + host_info->os_type = ENA_ADMIN_OS_LINUX; + host_info->kernel_ver = LINUX_VERSION_CODE; + strscpy(host_info->kernel_ver_str, utsname()->version, + sizeof(host_info->kernel_ver_str) - 1); + host_info->os_dist = 0; + strncpy(host_info->os_dist_str, utsname()->release, + sizeof(host_info->os_dist_str) - 1); + host_info->driver_version = + (DRV_MODULE_GEN_MAJOR) | + (DRV_MODULE_GEN_MINOR << ENA_ADMIN_HOST_INFO_MINOR_SHIFT) | + (DRV_MODULE_GEN_SUBMINOR << ENA_ADMIN_HOST_INFO_SUB_MINOR_SHIFT) | + ("K"[0] << ENA_ADMIN_HOST_INFO_MODULE_TYPE_SHIFT); + host_info->num_cpus = num_online_cpus(); + + host_info->driver_supported_features = + ENA_ADMIN_HOST_INFO_RX_OFFSET_MASK | + ENA_ADMIN_HOST_INFO_INTERRUPT_MODERATION_MASK | + ENA_ADMIN_HOST_INFO_RX_BUF_MIRRORING_MASK | + ENA_ADMIN_HOST_INFO_RSS_CONFIGURABLE_FUNCTION_KEY_MASK | + ENA_ADMIN_HOST_INFO_RX_PAGE_REUSE_MASK; + + rc = ena_com_set_host_attributes(ena_dev); + if (rc) { + if (rc == -EOPNOTSUPP) + dev_warn(dev, "Cannot set host attributes\n"); + else + dev_err(dev, "Cannot set host attributes\n"); + + goto err; + } + + return; + +err: + ena_com_delete_host_info(ena_dev); +} + +static void ena_config_debug_area(struct ena_adapter *adapter) +{ + u32 debug_area_size; + int rc, ss_count; + + ss_count = ena_get_sset_count(adapter->netdev, ETH_SS_STATS); + if (ss_count <= 0) { + netif_err(adapter, drv, adapter->netdev, + "SS count is negative\n"); + return; + } + + /* allocate 32 bytes for each string and 64bit for the value */ + debug_area_size = ss_count * ETH_GSTRING_LEN + sizeof(u64) * ss_count; + + rc = ena_com_allocate_debug_area(adapter->ena_dev, debug_area_size); + if (rc) { + netif_err(adapter, drv, adapter->netdev, + "Cannot allocate debug area\n"); + return; + } + + rc = ena_com_set_host_attributes(adapter->ena_dev); + if (rc) { + if (rc == -EOPNOTSUPP) + netif_warn(adapter, drv, adapter->netdev, + "Cannot set host attributes\n"); + else + netif_err(adapter, drv, adapter->netdev, + "Cannot set host attributes\n"); + goto err; + } + + return; +err: + ena_com_delete_debug_area(adapter->ena_dev); +} + +int ena_update_hw_stats(struct ena_adapter *adapter) +{ + int rc; + + rc = ena_com_get_eni_stats(adapter->ena_dev, &adapter->eni_stats); + if (rc) { + netdev_err(adapter->netdev, "Failed to get ENI stats\n"); + return rc; + } + + return 0; +} + +static void ena_get_stats64(struct net_device *netdev, + struct rtnl_link_stats64 *stats) +{ + struct ena_adapter *adapter = netdev_priv(netdev); + struct ena_ring *rx_ring, *tx_ring; + unsigned int start; + u64 rx_drops; + u64 tx_drops; + int i; + + if (!test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + return; + + for (i = 0; i < adapter->num_io_queues; i++) { + u64 bytes, packets; + + tx_ring = &adapter->tx_ring[i]; + + do { + start = u64_stats_fetch_begin(&tx_ring->syncp); + packets = tx_ring->tx_stats.cnt; + bytes = tx_ring->tx_stats.bytes; + } while (u64_stats_fetch_retry(&tx_ring->syncp, start)); + + stats->tx_packets += packets; + stats->tx_bytes += bytes; + + rx_ring = &adapter->rx_ring[i]; + + do { + start = u64_stats_fetch_begin(&rx_ring->syncp); + packets = rx_ring->rx_stats.cnt; + bytes = rx_ring->rx_stats.bytes; + } while (u64_stats_fetch_retry(&rx_ring->syncp, start)); + + stats->rx_packets += packets; + stats->rx_bytes += bytes; + } + + do { + start = u64_stats_fetch_begin(&adapter->syncp); + rx_drops = adapter->dev_stats.rx_drops; + tx_drops = adapter->dev_stats.tx_drops; + } while (u64_stats_fetch_retry(&adapter->syncp, start)); + + stats->rx_dropped = rx_drops; + stats->tx_dropped = tx_drops; + + stats->multicast = 0; + stats->collisions = 0; + + stats->rx_length_errors = 0; + stats->rx_crc_errors = 0; + stats->rx_frame_errors = 0; + stats->rx_fifo_errors = 0; + stats->rx_missed_errors = 0; + stats->tx_window_errors = 0; + + stats->rx_errors = 0; + stats->tx_errors = 0; +} + +static const struct net_device_ops ena_netdev_ops = { + .ndo_open = ena_open, + .ndo_stop = ena_close, + .ndo_start_xmit = ena_start_xmit, + .ndo_select_queue = ena_select_queue, + .ndo_get_stats64 = ena_get_stats64, + .ndo_tx_timeout = ena_tx_timeout, + .ndo_change_mtu = ena_change_mtu, + .ndo_set_mac_address = NULL, + .ndo_validate_addr = eth_validate_addr, + .ndo_bpf = ena_xdp, + .ndo_xdp_xmit = ena_xdp_xmit, +}; + +static void ena_calc_io_queue_size(struct ena_adapter *adapter, + struct ena_com_dev_get_features_ctx *get_feat_ctx) +{ + struct ena_admin_feature_llq_desc *llq = &get_feat_ctx->llq; + struct ena_com_dev *ena_dev = adapter->ena_dev; + u32 tx_queue_size = ENA_DEFAULT_RING_SIZE; + u32 rx_queue_size = ENA_DEFAULT_RING_SIZE; + u32 max_tx_queue_size; + u32 max_rx_queue_size; + + /* If this function is called after driver load, the ring sizes have already + * been configured. Take it into account when recalculating ring size. + */ + if (adapter->tx_ring->ring_size) + tx_queue_size = adapter->tx_ring->ring_size; + + if (adapter->rx_ring->ring_size) + rx_queue_size = adapter->rx_ring->ring_size; + + if (ena_dev->supported_features & BIT(ENA_ADMIN_MAX_QUEUES_EXT)) { + struct ena_admin_queue_ext_feature_fields *max_queue_ext = + &get_feat_ctx->max_queue_ext.max_queue_ext; + max_rx_queue_size = min_t(u32, max_queue_ext->max_rx_cq_depth, + max_queue_ext->max_rx_sq_depth); + max_tx_queue_size = max_queue_ext->max_tx_cq_depth; + + if (ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + max_tx_queue_size = min_t(u32, max_tx_queue_size, + llq->max_llq_depth); + else + max_tx_queue_size = min_t(u32, max_tx_queue_size, + max_queue_ext->max_tx_sq_depth); + + adapter->max_tx_sgl_size = min_t(u16, ENA_PKT_MAX_BUFS, + max_queue_ext->max_per_packet_tx_descs); + adapter->max_rx_sgl_size = min_t(u16, ENA_PKT_MAX_BUFS, + max_queue_ext->max_per_packet_rx_descs); + } else { + struct ena_admin_queue_feature_desc *max_queues = + &get_feat_ctx->max_queues; + max_rx_queue_size = min_t(u32, max_queues->max_cq_depth, + max_queues->max_sq_depth); + max_tx_queue_size = max_queues->max_cq_depth; + + if (ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + max_tx_queue_size = min_t(u32, max_tx_queue_size, + llq->max_llq_depth); + else + max_tx_queue_size = min_t(u32, max_tx_queue_size, + max_queues->max_sq_depth); + + adapter->max_tx_sgl_size = min_t(u16, ENA_PKT_MAX_BUFS, + max_queues->max_packet_tx_descs); + adapter->max_rx_sgl_size = min_t(u16, ENA_PKT_MAX_BUFS, + max_queues->max_packet_rx_descs); + } + + max_tx_queue_size = rounddown_pow_of_two(max_tx_queue_size); + max_rx_queue_size = rounddown_pow_of_two(max_rx_queue_size); + + /* When forcing large headers, we multiply the entry size by 2, and therefore divide + * the queue size by 2, leaving the amount of memory used by the queues unchanged. + */ + if (adapter->large_llq_header_enabled) { + if ((llq->entry_size_ctrl_supported & ENA_ADMIN_LIST_ENTRY_SIZE_256B) && + ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) { + max_tx_queue_size /= 2; + dev_info(&adapter->pdev->dev, + "Forcing large headers and decreasing maximum TX queue size to %d\n", + max_tx_queue_size); + } else { + dev_err(&adapter->pdev->dev, + "Forcing large headers failed: LLQ is disabled or device does not support large headers\n"); + + adapter->large_llq_header_enabled = false; + } + } + + tx_queue_size = clamp_val(tx_queue_size, ENA_MIN_RING_SIZE, + max_tx_queue_size); + rx_queue_size = clamp_val(rx_queue_size, ENA_MIN_RING_SIZE, + max_rx_queue_size); + + tx_queue_size = rounddown_pow_of_two(tx_queue_size); + rx_queue_size = rounddown_pow_of_two(rx_queue_size); + + adapter->max_tx_ring_size = max_tx_queue_size; + adapter->max_rx_ring_size = max_rx_queue_size; + adapter->requested_tx_ring_size = tx_queue_size; + adapter->requested_rx_ring_size = rx_queue_size; +} + +static int ena_device_validate_params(struct ena_adapter *adapter, + struct ena_com_dev_get_features_ctx *get_feat_ctx) +{ + struct net_device *netdev = adapter->netdev; + int rc; + + rc = ether_addr_equal(get_feat_ctx->dev_attr.mac_addr, + adapter->mac_addr); + if (!rc) { + netif_err(adapter, drv, netdev, + "Error, mac address are different\n"); + return -EINVAL; + } + + if (get_feat_ctx->dev_attr.max_mtu < netdev->mtu) { + netif_err(adapter, drv, netdev, + "Error, device max mtu is smaller than netdev MTU\n"); + return -EINVAL; + } + + return 0; +} + +static void set_default_llq_configurations(struct ena_adapter *adapter, + struct ena_llq_configurations *llq_config, + struct ena_admin_feature_llq_desc *llq) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + + llq_config->llq_header_location = ENA_ADMIN_INLINE_HEADER; + llq_config->llq_stride_ctrl = ENA_ADMIN_MULTIPLE_DESCS_PER_ENTRY; + llq_config->llq_num_decs_before_header = ENA_ADMIN_LLQ_NUM_DESCS_BEFORE_HEADER_2; + + adapter->large_llq_header_supported = + !!(ena_dev->supported_features & BIT(ENA_ADMIN_LLQ)); + adapter->large_llq_header_supported &= + !!(llq->entry_size_ctrl_supported & + ENA_ADMIN_LIST_ENTRY_SIZE_256B); + + if ((llq->entry_size_ctrl_supported & ENA_ADMIN_LIST_ENTRY_SIZE_256B) && + adapter->large_llq_header_enabled) { + llq_config->llq_ring_entry_size = ENA_ADMIN_LIST_ENTRY_SIZE_256B; + llq_config->llq_ring_entry_size_value = 256; + } else { + llq_config->llq_ring_entry_size = ENA_ADMIN_LIST_ENTRY_SIZE_128B; + llq_config->llq_ring_entry_size_value = 128; + } +} + +static int ena_set_queues_placement_policy(struct pci_dev *pdev, + struct ena_com_dev *ena_dev, + struct ena_admin_feature_llq_desc *llq, + struct ena_llq_configurations *llq_default_configurations) +{ + int rc; + u32 llq_feature_mask; + + llq_feature_mask = 1 << ENA_ADMIN_LLQ; + if (!(ena_dev->supported_features & llq_feature_mask)) { + dev_warn(&pdev->dev, + "LLQ is not supported Fallback to host mode policy.\n"); + ena_dev->tx_mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_HOST; + return 0; + } + + if (!ena_dev->mem_bar) { + netdev_err(ena_dev->net_device, + "LLQ is advertised as supported but device doesn't expose mem bar\n"); + ena_dev->tx_mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_HOST; + return 0; + } + + rc = ena_com_config_dev_mode(ena_dev, llq, llq_default_configurations); + if (unlikely(rc)) { + dev_err(&pdev->dev, + "Failed to configure the device mode. Fallback to host mode policy.\n"); + ena_dev->tx_mem_queue_type = ENA_ADMIN_PLACEMENT_POLICY_HOST; + } + + return 0; +} + +static int ena_map_llq_mem_bar(struct pci_dev *pdev, struct ena_com_dev *ena_dev, + int bars) +{ + bool has_mem_bar = !!(bars & BIT(ENA_MEM_BAR)); + + if (!has_mem_bar) + return 0; + + ena_dev->mem_bar = devm_ioremap_wc(&pdev->dev, + pci_resource_start(pdev, ENA_MEM_BAR), + pci_resource_len(pdev, ENA_MEM_BAR)); + + if (!ena_dev->mem_bar) + return -EFAULT; + + return 0; +} + +static int ena_device_init(struct ena_adapter *adapter, struct pci_dev *pdev, + struct ena_com_dev_get_features_ctx *get_feat_ctx, + bool *wd_state) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + struct ena_llq_configurations llq_config; + struct device *dev = &pdev->dev; + bool readless_supported; + u32 aenq_groups; + int dma_width; + int rc; + + rc = ena_com_mmio_reg_read_request_init(ena_dev); + if (rc) { + dev_err(dev, "Failed to init mmio read less\n"); + return rc; + } + + /* The PCIe configuration space revision id indicate if mmio reg + * read is disabled + */ + readless_supported = !(pdev->revision & ENA_MMIO_DISABLE_REG_READ); + ena_com_set_mmio_read_mode(ena_dev, readless_supported); + + rc = ena_com_dev_reset(ena_dev, ENA_REGS_RESET_NORMAL); + if (rc) { + dev_err(dev, "Can not reset device\n"); + goto err_mmio_read_less; + } + + rc = ena_com_validate_version(ena_dev); + if (rc) { + dev_err(dev, "Device version is too low\n"); + goto err_mmio_read_less; + } + + dma_width = ena_com_get_dma_width(ena_dev); + if (dma_width < 0) { + dev_err(dev, "Invalid dma width value %d", dma_width); + rc = dma_width; + goto err_mmio_read_less; + } + + rc = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(dma_width)); + if (rc) { + dev_err(dev, "dma_set_mask_and_coherent failed %d\n", rc); + goto err_mmio_read_less; + } + + /* ENA admin level init */ + rc = ena_com_admin_init(ena_dev, &aenq_handlers); + if (rc) { + dev_err(dev, + "Can not initialize ena admin queue with device\n"); + goto err_mmio_read_less; + } + + /* To enable the msix interrupts the driver needs to know the number + * of queues. So the driver uses polling mode to retrieve this + * information + */ + ena_com_set_admin_polling_mode(ena_dev, true); + + ena_config_host_info(ena_dev, pdev); + + /* Get Device Attributes*/ + rc = ena_com_get_dev_attr_feat(ena_dev, get_feat_ctx); + if (rc) { + dev_err(dev, "Cannot get attribute for ena device rc=%d\n", rc); + goto err_admin_init; + } + + /* Try to turn all the available aenq groups */ + aenq_groups = BIT(ENA_ADMIN_LINK_CHANGE) | + BIT(ENA_ADMIN_FATAL_ERROR) | + BIT(ENA_ADMIN_WARNING) | + BIT(ENA_ADMIN_NOTIFICATION) | + BIT(ENA_ADMIN_KEEP_ALIVE); + + aenq_groups &= get_feat_ctx->aenq.supported_groups; + + rc = ena_com_set_aenq_config(ena_dev, aenq_groups); + if (rc) { + dev_err(dev, "Cannot configure aenq groups rc= %d\n", rc); + goto err_admin_init; + } + + *wd_state = !!(aenq_groups & BIT(ENA_ADMIN_KEEP_ALIVE)); + + set_default_llq_configurations(adapter, &llq_config, &get_feat_ctx->llq); + + rc = ena_set_queues_placement_policy(pdev, ena_dev, &get_feat_ctx->llq, + &llq_config); + if (rc) { + dev_err(dev, "ENA device init failed\n"); + goto err_admin_init; + } + + ena_calc_io_queue_size(adapter, get_feat_ctx); + + return 0; + +err_admin_init: + ena_com_delete_host_info(ena_dev); + ena_com_admin_destroy(ena_dev); +err_mmio_read_less: + ena_com_mmio_reg_read_request_destroy(ena_dev); + + return rc; +} + +static int ena_enable_msix_and_set_admin_interrupts(struct ena_adapter *adapter) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + struct device *dev = &adapter->pdev->dev; + int rc; + + rc = ena_enable_msix(adapter); + if (rc) { + dev_err(dev, "Can not reserve msix vectors\n"); + return rc; + } + + ena_setup_mgmnt_intr(adapter); + + rc = ena_request_mgmnt_irq(adapter); + if (rc) { + dev_err(dev, "Can not setup management interrupts\n"); + goto err_disable_msix; + } + + ena_com_set_admin_polling_mode(ena_dev, false); + + ena_com_admin_aenq_enable(ena_dev); + + return 0; + +err_disable_msix: + ena_disable_msix(adapter); + + return rc; +} + +static void ena_destroy_device(struct ena_adapter *adapter, bool graceful) +{ + struct net_device *netdev = adapter->netdev; + struct ena_com_dev *ena_dev = adapter->ena_dev; + bool dev_up; + + if (!test_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags)) + return; + + netif_carrier_off(netdev); + + del_timer_sync(&adapter->timer_service); + + dev_up = test_bit(ENA_FLAG_DEV_UP, &adapter->flags); + adapter->dev_up_before_reset = dev_up; + if (!graceful) + ena_com_set_admin_running_state(ena_dev, false); + + if (test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + ena_down(adapter); + + /* Stop the device from sending AENQ events (in case reset flag is set + * and device is up, ena_down() already reset the device. + */ + if (!(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags) && dev_up)) + ena_com_dev_reset(adapter->ena_dev, adapter->reset_reason); + + ena_free_mgmnt_irq(adapter); + + ena_disable_msix(adapter); + + ena_com_abort_admin_commands(ena_dev); + + ena_com_wait_for_abort_completion(ena_dev); + + ena_com_admin_destroy(ena_dev); + + ena_com_mmio_reg_read_request_destroy(ena_dev); + + /* return reset reason to default value */ + adapter->reset_reason = ENA_REGS_RESET_NORMAL; + + clear_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags); + clear_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags); +} + +static int ena_restore_device(struct ena_adapter *adapter) +{ + struct ena_com_dev_get_features_ctx get_feat_ctx; + struct ena_com_dev *ena_dev = adapter->ena_dev; + struct pci_dev *pdev = adapter->pdev; + struct ena_ring *txr; + int rc, count, i; + bool wd_state; + + set_bit(ENA_FLAG_ONGOING_RESET, &adapter->flags); + rc = ena_device_init(adapter, adapter->pdev, &get_feat_ctx, &wd_state); + if (rc) { + dev_err(&pdev->dev, "Can not initialize device\n"); + goto err; + } + adapter->wd_state = wd_state; + + count = adapter->xdp_num_queues + adapter->num_io_queues; + for (i = 0 ; i < count; i++) { + txr = &adapter->tx_ring[i]; + txr->tx_mem_queue_type = ena_dev->tx_mem_queue_type; + txr->tx_max_header_size = ena_dev->tx_max_header_size; + } + + rc = ena_device_validate_params(adapter, &get_feat_ctx); + if (rc) { + dev_err(&pdev->dev, "Validation of device parameters failed\n"); + goto err_device_destroy; + } + + rc = ena_enable_msix_and_set_admin_interrupts(adapter); + if (rc) { + dev_err(&pdev->dev, "Enable MSI-X failed\n"); + goto err_device_destroy; + } + /* If the interface was up before the reset bring it up */ + if (adapter->dev_up_before_reset) { + rc = ena_up(adapter); + if (rc) { + dev_err(&pdev->dev, "Failed to create I/O queues\n"); + goto err_disable_msix; + } + } + + set_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags); + + clear_bit(ENA_FLAG_ONGOING_RESET, &adapter->flags); + if (test_bit(ENA_FLAG_LINK_UP, &adapter->flags)) + netif_carrier_on(adapter->netdev); + + mod_timer(&adapter->timer_service, round_jiffies(jiffies + HZ)); + adapter->last_keep_alive_jiffies = jiffies; + + return rc; +err_disable_msix: + ena_free_mgmnt_irq(adapter); + ena_disable_msix(adapter); +err_device_destroy: + ena_com_abort_admin_commands(ena_dev); + ena_com_wait_for_abort_completion(ena_dev); + ena_com_admin_destroy(ena_dev); + ena_com_dev_reset(ena_dev, ENA_REGS_RESET_DRIVER_INVALID_STATE); + ena_com_mmio_reg_read_request_destroy(ena_dev); +err: + clear_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags); + clear_bit(ENA_FLAG_ONGOING_RESET, &adapter->flags); + dev_err(&pdev->dev, + "Reset attempt failed. Can not reset the device\n"); + + return rc; +} + +static void ena_fw_reset_device(struct work_struct *work) +{ + struct ena_adapter *adapter = + container_of(work, struct ena_adapter, reset_task); + + rtnl_lock(); + + if (likely(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags))) { + ena_destroy_device(adapter, false); + ena_restore_device(adapter); + + dev_err(&adapter->pdev->dev, "Device reset completed successfully\n"); + } + + rtnl_unlock(); +} + +static int check_for_rx_interrupt_queue(struct ena_adapter *adapter, + struct ena_ring *rx_ring) +{ + struct ena_napi *ena_napi = container_of(rx_ring->napi, struct ena_napi, napi); + + if (likely(READ_ONCE(ena_napi->first_interrupt))) + return 0; + + if (ena_com_cq_empty(rx_ring->ena_com_io_cq)) + return 0; + + rx_ring->no_interrupt_event_cnt++; + + if (rx_ring->no_interrupt_event_cnt == ENA_MAX_NO_INTERRUPT_ITERATIONS) { + netif_err(adapter, rx_err, adapter->netdev, + "Potential MSIX issue on Rx side Queue = %d. Reset the device\n", + rx_ring->qid); + + ena_reset_device(adapter, ENA_REGS_RESET_MISS_INTERRUPT); + return -EIO; + } + + return 0; +} + +static int check_missing_comp_in_tx_queue(struct ena_adapter *adapter, + struct ena_ring *tx_ring) +{ + struct ena_napi *ena_napi = container_of(tx_ring->napi, struct ena_napi, napi); + unsigned int time_since_last_napi; + unsigned int missing_tx_comp_to; + bool is_tx_comp_time_expired; + struct ena_tx_buffer *tx_buf; + unsigned long last_jiffies; + u32 missed_tx = 0; + int i, rc = 0; + + for (i = 0; i < tx_ring->ring_size; i++) { + tx_buf = &tx_ring->tx_buffer_info[i]; + last_jiffies = tx_buf->last_jiffies; + + if (last_jiffies == 0) + /* no pending Tx at this location */ + continue; + + is_tx_comp_time_expired = time_is_before_jiffies(last_jiffies + + 2 * adapter->missing_tx_completion_to); + + if (unlikely(!READ_ONCE(ena_napi->first_interrupt) && is_tx_comp_time_expired)) { + /* If after graceful period interrupt is still not + * received, we schedule a reset + */ + netif_err(adapter, tx_err, adapter->netdev, + "Potential MSIX issue on Tx side Queue = %d. Reset the device\n", + tx_ring->qid); + ena_reset_device(adapter, ENA_REGS_RESET_MISS_INTERRUPT); + return -EIO; + } + + is_tx_comp_time_expired = time_is_before_jiffies(last_jiffies + + adapter->missing_tx_completion_to); + + if (unlikely(is_tx_comp_time_expired)) { + if (!tx_buf->print_once) { + time_since_last_napi = jiffies_to_usecs(jiffies - tx_ring->tx_stats.last_napi_jiffies); + missing_tx_comp_to = jiffies_to_msecs(adapter->missing_tx_completion_to); + netif_notice(adapter, tx_err, adapter->netdev, + "Found a Tx that wasn't completed on time, qid %d, index %d. %u usecs have passed since last napi execution. Missing Tx timeout value %u msecs\n", + tx_ring->qid, i, time_since_last_napi, missing_tx_comp_to); + } + + tx_buf->print_once = 1; + missed_tx++; + } + } + + if (unlikely(missed_tx > adapter->missing_tx_completion_threshold)) { + netif_err(adapter, tx_err, adapter->netdev, + "The number of lost tx completions is above the threshold (%d > %d). Reset the device\n", + missed_tx, + adapter->missing_tx_completion_threshold); + ena_reset_device(adapter, ENA_REGS_RESET_MISS_TX_CMPL); + rc = -EIO; + } + + ena_increase_stat(&tx_ring->tx_stats.missed_tx, missed_tx, + &tx_ring->syncp); + + return rc; +} + +static void check_for_missing_completions(struct ena_adapter *adapter) +{ + struct ena_ring *tx_ring; + struct ena_ring *rx_ring; + int i, budget, rc; + int io_queue_count; + + io_queue_count = adapter->xdp_num_queues + adapter->num_io_queues; + /* Make sure the driver doesn't turn the device in other process */ + smp_rmb(); + + if (!test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + return; + + if (test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags)) + return; + + if (adapter->missing_tx_completion_to == ENA_HW_HINTS_NO_TIMEOUT) + return; + + budget = ENA_MONITORED_TX_QUEUES; + + for (i = adapter->last_monitored_tx_qid; i < io_queue_count; i++) { + tx_ring = &adapter->tx_ring[i]; + rx_ring = &adapter->rx_ring[i]; + + rc = check_missing_comp_in_tx_queue(adapter, tx_ring); + if (unlikely(rc)) + return; + + rc = !ENA_IS_XDP_INDEX(adapter, i) ? + check_for_rx_interrupt_queue(adapter, rx_ring) : 0; + if (unlikely(rc)) + return; + + budget--; + if (!budget) + break; + } + + adapter->last_monitored_tx_qid = i % io_queue_count; +} + +/* trigger napi schedule after 2 consecutive detections */ +#define EMPTY_RX_REFILL 2 +/* For the rare case where the device runs out of Rx descriptors and the + * napi handler failed to refill new Rx descriptors (due to a lack of memory + * for example). + * This case will lead to a deadlock: + * The device won't send interrupts since all the new Rx packets will be dropped + * The napi handler won't allocate new Rx descriptors so the device will be + * able to send new packets. + * + * This scenario can happen when the kernel's vm.min_free_kbytes is too small. + * It is recommended to have at least 512MB, with a minimum of 128MB for + * constrained environment). + * + * When such a situation is detected - Reschedule napi + */ +static void check_for_empty_rx_ring(struct ena_adapter *adapter) +{ + struct ena_ring *rx_ring; + int i, refill_required; + + if (!test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) + return; + + if (test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags)) + return; + + for (i = 0; i < adapter->num_io_queues; i++) { + rx_ring = &adapter->rx_ring[i]; + + refill_required = ena_com_free_q_entries(rx_ring->ena_com_io_sq); + if (unlikely(refill_required == (rx_ring->ring_size - 1))) { + rx_ring->empty_rx_queue++; + + if (rx_ring->empty_rx_queue >= EMPTY_RX_REFILL) { + ena_increase_stat(&rx_ring->rx_stats.empty_rx_ring, 1, + &rx_ring->syncp); + + netif_err(adapter, drv, adapter->netdev, + "Trigger refill for ring %d\n", i); + + napi_schedule(rx_ring->napi); + rx_ring->empty_rx_queue = 0; + } + } else { + rx_ring->empty_rx_queue = 0; + } + } +} + +/* Check for keep alive expiration */ +static void check_for_missing_keep_alive(struct ena_adapter *adapter) +{ + unsigned long keep_alive_expired; + + if (!adapter->wd_state) + return; + + if (adapter->keep_alive_timeout == ENA_HW_HINTS_NO_TIMEOUT) + return; + + keep_alive_expired = adapter->last_keep_alive_jiffies + + adapter->keep_alive_timeout; + if (unlikely(time_is_before_jiffies(keep_alive_expired))) { + netif_err(adapter, drv, adapter->netdev, + "Keep alive watchdog timeout.\n"); + ena_increase_stat(&adapter->dev_stats.wd_expired, 1, + &adapter->syncp); + ena_reset_device(adapter, ENA_REGS_RESET_KEEP_ALIVE_TO); + } +} + +static void check_for_admin_com_state(struct ena_adapter *adapter) +{ + if (unlikely(!ena_com_get_admin_running_state(adapter->ena_dev))) { + netif_err(adapter, drv, adapter->netdev, + "ENA admin queue is not in running state!\n"); + ena_increase_stat(&adapter->dev_stats.admin_q_pause, 1, + &adapter->syncp); + ena_reset_device(adapter, ENA_REGS_RESET_ADMIN_TO); + } +} + +static void ena_update_hints(struct ena_adapter *adapter, + struct ena_admin_ena_hw_hints *hints) +{ + struct net_device *netdev = adapter->netdev; + + if (hints->admin_completion_tx_timeout) + adapter->ena_dev->admin_queue.completion_timeout = + hints->admin_completion_tx_timeout * 1000; + + if (hints->mmio_read_timeout) + /* convert to usec */ + adapter->ena_dev->mmio_read.reg_read_to = + hints->mmio_read_timeout * 1000; + + if (hints->missed_tx_completion_count_threshold_to_reset) + adapter->missing_tx_completion_threshold = + hints->missed_tx_completion_count_threshold_to_reset; + + if (hints->missing_tx_completion_timeout) { + if (hints->missing_tx_completion_timeout == ENA_HW_HINTS_NO_TIMEOUT) + adapter->missing_tx_completion_to = ENA_HW_HINTS_NO_TIMEOUT; + else + adapter->missing_tx_completion_to = + msecs_to_jiffies(hints->missing_tx_completion_timeout); + } + + if (hints->netdev_wd_timeout) + netdev->watchdog_timeo = msecs_to_jiffies(hints->netdev_wd_timeout); + + if (hints->driver_watchdog_timeout) { + if (hints->driver_watchdog_timeout == ENA_HW_HINTS_NO_TIMEOUT) + adapter->keep_alive_timeout = ENA_HW_HINTS_NO_TIMEOUT; + else + adapter->keep_alive_timeout = + msecs_to_jiffies(hints->driver_watchdog_timeout); + } +} + +static void ena_update_host_info(struct ena_admin_host_info *host_info, + struct net_device *netdev) +{ + host_info->supported_network_features[0] = + netdev->features & GENMASK_ULL(31, 0); + host_info->supported_network_features[1] = + (netdev->features & GENMASK_ULL(63, 32)) >> 32; +} + +static void ena_timer_service(struct timer_list *t) +{ + struct ena_adapter *adapter = from_timer(adapter, t, timer_service); + u8 *debug_area = adapter->ena_dev->host_attr.debug_area_virt_addr; + struct ena_admin_host_info *host_info = + adapter->ena_dev->host_attr.host_info; + + check_for_missing_keep_alive(adapter); + + check_for_admin_com_state(adapter); + + check_for_missing_completions(adapter); + + check_for_empty_rx_ring(adapter); + + if (debug_area) + ena_dump_stats_to_buf(adapter, debug_area); + + if (host_info) + ena_update_host_info(host_info, adapter->netdev); + + if (unlikely(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags))) { + netif_err(adapter, drv, adapter->netdev, + "Trigger reset is on\n"); + ena_dump_stats_to_dmesg(adapter); + queue_work(ena_wq, &adapter->reset_task); + return; + } + + /* Reset the timer */ + mod_timer(&adapter->timer_service, round_jiffies(jiffies + HZ)); +} + +static u32 ena_calc_max_io_queue_num(struct pci_dev *pdev, + struct ena_com_dev *ena_dev, + struct ena_com_dev_get_features_ctx *get_feat_ctx) +{ + u32 io_tx_sq_num, io_tx_cq_num, io_rx_num, max_num_io_queues; + + if (ena_dev->supported_features & BIT(ENA_ADMIN_MAX_QUEUES_EXT)) { + struct ena_admin_queue_ext_feature_fields *max_queue_ext = + &get_feat_ctx->max_queue_ext.max_queue_ext; + io_rx_num = min_t(u32, max_queue_ext->max_rx_sq_num, + max_queue_ext->max_rx_cq_num); + + io_tx_sq_num = max_queue_ext->max_tx_sq_num; + io_tx_cq_num = max_queue_ext->max_tx_cq_num; + } else { + struct ena_admin_queue_feature_desc *max_queues = + &get_feat_ctx->max_queues; + io_tx_sq_num = max_queues->max_sq_num; + io_tx_cq_num = max_queues->max_cq_num; + io_rx_num = min_t(u32, io_tx_sq_num, io_tx_cq_num); + } + + /* In case of LLQ use the llq fields for the tx SQ/CQ */ + if (ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + io_tx_sq_num = get_feat_ctx->llq.max_llq_num; + + max_num_io_queues = min_t(u32, num_online_cpus(), ENA_MAX_NUM_IO_QUEUES); + max_num_io_queues = min_t(u32, max_num_io_queues, io_rx_num); + max_num_io_queues = min_t(u32, max_num_io_queues, io_tx_sq_num); + max_num_io_queues = min_t(u32, max_num_io_queues, io_tx_cq_num); + /* 1 IRQ for mgmnt and 1 IRQs for each IO direction */ + max_num_io_queues = min_t(u32, max_num_io_queues, pci_msix_vec_count(pdev) - 1); + + return max_num_io_queues; +} + +static void ena_set_dev_offloads(struct ena_com_dev_get_features_ctx *feat, + struct net_device *netdev) +{ + netdev_features_t dev_features = 0; + + /* Set offload features */ + if (feat->offload.tx & + ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV4_CSUM_PART_MASK) + dev_features |= NETIF_F_IP_CSUM; + + if (feat->offload.tx & + ENA_ADMIN_FEATURE_OFFLOAD_DESC_TX_L4_IPV6_CSUM_PART_MASK) + dev_features |= NETIF_F_IPV6_CSUM; + + if (feat->offload.tx & ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV4_MASK) + dev_features |= NETIF_F_TSO; + + if (feat->offload.tx & ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_IPV6_MASK) + dev_features |= NETIF_F_TSO6; + + if (feat->offload.tx & ENA_ADMIN_FEATURE_OFFLOAD_DESC_TSO_ECN_MASK) + dev_features |= NETIF_F_TSO_ECN; + + if (feat->offload.rx_supported & + ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV4_CSUM_MASK) + dev_features |= NETIF_F_RXCSUM; + + if (feat->offload.rx_supported & + ENA_ADMIN_FEATURE_OFFLOAD_DESC_RX_L4_IPV6_CSUM_MASK) + dev_features |= NETIF_F_RXCSUM; + + netdev->features = + dev_features | + NETIF_F_SG | + NETIF_F_RXHASH | + NETIF_F_HIGHDMA; + + netdev->hw_features |= netdev->features; + netdev->vlan_features |= netdev->features; +} + +static void ena_set_conf_feat_params(struct ena_adapter *adapter, + struct ena_com_dev_get_features_ctx *feat) +{ + struct net_device *netdev = adapter->netdev; + + /* Copy mac address */ + if (!is_valid_ether_addr(feat->dev_attr.mac_addr)) { + eth_hw_addr_random(netdev); + ether_addr_copy(adapter->mac_addr, netdev->dev_addr); + } else { + ether_addr_copy(adapter->mac_addr, feat->dev_attr.mac_addr); + eth_hw_addr_set(netdev, adapter->mac_addr); + } + + /* Set offload features */ + ena_set_dev_offloads(feat, netdev); + + adapter->max_mtu = feat->dev_attr.max_mtu; + netdev->max_mtu = adapter->max_mtu; + netdev->min_mtu = ENA_MIN_MTU; +} + +static int ena_rss_init_default(struct ena_adapter *adapter) +{ + struct ena_com_dev *ena_dev = adapter->ena_dev; + struct device *dev = &adapter->pdev->dev; + int rc, i; + u32 val; + + rc = ena_com_rss_init(ena_dev, ENA_RX_RSS_TABLE_LOG_SIZE); + if (unlikely(rc)) { + dev_err(dev, "Cannot init indirect table\n"); + goto err_rss_init; + } + + for (i = 0; i < ENA_RX_RSS_TABLE_SIZE; i++) { + val = ethtool_rxfh_indir_default(i, adapter->num_io_queues); + rc = ena_com_indirect_table_fill_entry(ena_dev, i, + ENA_IO_RXQ_IDX(val)); + if (unlikely(rc)) { + dev_err(dev, "Cannot fill indirect table\n"); + goto err_fill_indir; + } + } + + rc = ena_com_fill_hash_function(ena_dev, ENA_ADMIN_TOEPLITZ, NULL, + ENA_HASH_KEY_SIZE, 0xFFFFFFFF); + if (unlikely(rc && (rc != -EOPNOTSUPP))) { + dev_err(dev, "Cannot fill hash function\n"); + goto err_fill_indir; + } + + rc = ena_com_set_default_hash_ctrl(ena_dev); + if (unlikely(rc && (rc != -EOPNOTSUPP))) { + dev_err(dev, "Cannot fill hash control\n"); + goto err_fill_indir; + } + + return 0; + +err_fill_indir: + ena_com_rss_destroy(ena_dev); +err_rss_init: + + return rc; +} + +static void ena_release_bars(struct ena_com_dev *ena_dev, struct pci_dev *pdev) +{ + int release_bars = pci_select_bars(pdev, IORESOURCE_MEM) & ENA_BAR_MASK; + + pci_release_selected_regions(pdev, release_bars); +} + +/* ena_probe - Device Initialization Routine + * @pdev: PCI device information struct + * @ent: entry in ena_pci_tbl + * + * Returns 0 on success, negative on failure + * + * ena_probe initializes an adapter identified by a pci_dev structure. + * The OS initialization, configuring of the adapter private structure, + * and a hardware reset occur. + */ +static int ena_probe(struct pci_dev *pdev, const struct pci_device_id *ent) +{ + struct ena_com_dev_get_features_ctx get_feat_ctx; + struct ena_com_dev *ena_dev = NULL; + struct ena_adapter *adapter; + struct net_device *netdev; + static int adapters_found; + u32 max_num_io_queues; + bool wd_state; + int bars, rc; + + dev_dbg(&pdev->dev, "%s\n", __func__); + + rc = pci_enable_device_mem(pdev); + if (rc) { + dev_err(&pdev->dev, "pci_enable_device_mem() failed!\n"); + return rc; + } + + rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(ENA_MAX_PHYS_ADDR_SIZE_BITS)); + if (rc) { + dev_err(&pdev->dev, "dma_set_mask_and_coherent failed %d\n", rc); + goto err_disable_device; + } + + pci_set_master(pdev); + + ena_dev = vzalloc(sizeof(*ena_dev)); + if (!ena_dev) { + rc = -ENOMEM; + goto err_disable_device; + } + + bars = pci_select_bars(pdev, IORESOURCE_MEM) & ENA_BAR_MASK; + rc = pci_request_selected_regions(pdev, bars, DRV_MODULE_NAME); + if (rc) { + dev_err(&pdev->dev, "pci_request_selected_regions failed %d\n", + rc); + goto err_free_ena_dev; + } + + ena_dev->reg_bar = devm_ioremap(&pdev->dev, + pci_resource_start(pdev, ENA_REG_BAR), + pci_resource_len(pdev, ENA_REG_BAR)); + if (!ena_dev->reg_bar) { + dev_err(&pdev->dev, "Failed to remap regs bar\n"); + rc = -EFAULT; + goto err_free_region; + } + + ena_dev->ena_min_poll_delay_us = ENA_ADMIN_POLL_DELAY_US; + + ena_dev->dmadev = &pdev->dev; + + netdev = alloc_etherdev_mq(sizeof(struct ena_adapter), ENA_MAX_RINGS); + if (!netdev) { + dev_err(&pdev->dev, "alloc_etherdev_mq failed\n"); + rc = -ENOMEM; + goto err_free_region; + } + + SET_NETDEV_DEV(netdev, &pdev->dev); + adapter = netdev_priv(netdev); + adapter->ena_dev = ena_dev; + adapter->netdev = netdev; + adapter->pdev = pdev; + adapter->msg_enable = DEFAULT_MSG_ENABLE; + + ena_dev->net_device = netdev; + + pci_set_drvdata(pdev, adapter); + + rc = ena_map_llq_mem_bar(pdev, ena_dev, bars); + if (rc) { + dev_err(&pdev->dev, "ENA LLQ bar mapping failed\n"); + goto err_netdev_destroy; + } + + rc = ena_device_init(adapter, pdev, &get_feat_ctx, &wd_state); + if (rc) { + dev_err(&pdev->dev, "ENA device init failed\n"); + if (rc == -ETIME) + rc = -EPROBE_DEFER; + goto err_netdev_destroy; + } + + /* Initial TX and RX interrupt delay. Assumes 1 usec granularity. + * Updated during device initialization with the real granularity + */ + ena_dev->intr_moder_tx_interval = ENA_INTR_INITIAL_TX_INTERVAL_USECS; + ena_dev->intr_moder_rx_interval = ENA_INTR_INITIAL_RX_INTERVAL_USECS; + ena_dev->intr_delay_resolution = ENA_DEFAULT_INTR_DELAY_RESOLUTION; + max_num_io_queues = ena_calc_max_io_queue_num(pdev, ena_dev, &get_feat_ctx); + if (unlikely(!max_num_io_queues)) { + rc = -EFAULT; + goto err_device_destroy; + } + + ena_set_conf_feat_params(adapter, &get_feat_ctx); + + adapter->reset_reason = ENA_REGS_RESET_NORMAL; + + adapter->num_io_queues = max_num_io_queues; + adapter->max_num_io_queues = max_num_io_queues; + adapter->last_monitored_tx_qid = 0; + + adapter->xdp_first_ring = 0; + adapter->xdp_num_queues = 0; + + adapter->rx_copybreak = ENA_DEFAULT_RX_COPYBREAK; + if (ena_dev->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) + adapter->disable_meta_caching = + !!(get_feat_ctx.llq.accel_mode.u.get.supported_flags & + BIT(ENA_ADMIN_DISABLE_META_CACHING)); + + adapter->wd_state = wd_state; + + snprintf(adapter->name, ENA_NAME_MAX_LEN, "ena_%d", adapters_found); + + rc = ena_com_init_interrupt_moderation(adapter->ena_dev); + if (rc) { + dev_err(&pdev->dev, + "Failed to query interrupt moderation feature\n"); + goto err_device_destroy; + } + + ena_init_io_rings(adapter, + 0, + adapter->xdp_num_queues + + adapter->num_io_queues); + + netdev->netdev_ops = &ena_netdev_ops; + netdev->watchdog_timeo = TX_TIMEOUT; + ena_set_ethtool_ops(netdev); + + netdev->priv_flags |= IFF_UNICAST_FLT; + + u64_stats_init(&adapter->syncp); + + rc = ena_enable_msix_and_set_admin_interrupts(adapter); + if (rc) { + dev_err(&pdev->dev, + "Failed to enable and set the admin interrupts\n"); + goto err_worker_destroy; + } + rc = ena_rss_init_default(adapter); + if (rc && (rc != -EOPNOTSUPP)) { + dev_err(&pdev->dev, "Cannot init RSS rc: %d\n", rc); + goto err_free_msix; + } + + ena_config_debug_area(adapter); + + if (ena_xdp_legal_queue_count(adapter, adapter->num_io_queues)) + netdev->xdp_features = NETDEV_XDP_ACT_BASIC | + NETDEV_XDP_ACT_REDIRECT; + + memcpy(adapter->netdev->perm_addr, adapter->mac_addr, netdev->addr_len); + + netif_carrier_off(netdev); + + rc = register_netdev(netdev); + if (rc) { + dev_err(&pdev->dev, "Cannot register net device\n"); + goto err_rss; + } + + INIT_WORK(&adapter->reset_task, ena_fw_reset_device); + + adapter->last_keep_alive_jiffies = jiffies; + adapter->keep_alive_timeout = ENA_DEVICE_KALIVE_TIMEOUT; + adapter->missing_tx_completion_to = TX_TIMEOUT; + adapter->missing_tx_completion_threshold = MAX_NUM_OF_TIMEOUTED_PACKETS; + + ena_update_hints(adapter, &get_feat_ctx.hw_hints); + + timer_setup(&adapter->timer_service, ena_timer_service, 0); + mod_timer(&adapter->timer_service, round_jiffies(jiffies + HZ)); + + dev_info(&pdev->dev, + "%s found at mem %lx, mac addr %pM\n", + DEVICE_NAME, (long)pci_resource_start(pdev, 0), + netdev->dev_addr); + + set_bit(ENA_FLAG_DEVICE_RUNNING, &adapter->flags); + + adapters_found++; + + return 0; + +err_rss: + ena_com_delete_debug_area(ena_dev); + ena_com_rss_destroy(ena_dev); +err_free_msix: + ena_com_dev_reset(ena_dev, ENA_REGS_RESET_INIT_ERR); + /* stop submitting admin commands on a device that was reset */ + ena_com_set_admin_running_state(ena_dev, false); + ena_free_mgmnt_irq(adapter); + ena_disable_msix(adapter); +err_worker_destroy: + del_timer(&adapter->timer_service); +err_device_destroy: + ena_com_delete_host_info(ena_dev); + ena_com_admin_destroy(ena_dev); +err_netdev_destroy: + free_netdev(netdev); +err_free_region: + ena_release_bars(ena_dev, pdev); +err_free_ena_dev: + vfree(ena_dev); +err_disable_device: + pci_disable_device(pdev); + return rc; +} + +/*****************************************************************************/ + +/* __ena_shutoff - Helper used in both PCI remove/shutdown routines + * @pdev: PCI device information struct + * @shutdown: Is it a shutdown operation? If false, means it is a removal + * + * __ena_shutoff is a helper routine that does the real work on shutdown and + * removal paths; the difference between those paths is with regards to whether + * dettach or unregister the netdevice. + */ +static void __ena_shutoff(struct pci_dev *pdev, bool shutdown) +{ + struct ena_adapter *adapter = pci_get_drvdata(pdev); + struct ena_com_dev *ena_dev; + struct net_device *netdev; + + ena_dev = adapter->ena_dev; + netdev = adapter->netdev; + +#ifdef CONFIG_RFS_ACCEL + if ((adapter->msix_vecs >= 1) && (netdev->rx_cpu_rmap)) { + free_irq_cpu_rmap(netdev->rx_cpu_rmap); + netdev->rx_cpu_rmap = NULL; + } +#endif /* CONFIG_RFS_ACCEL */ + + /* Make sure timer and reset routine won't be called after + * freeing device resources. + */ + del_timer_sync(&adapter->timer_service); + cancel_work_sync(&adapter->reset_task); + + rtnl_lock(); /* lock released inside the below if-else block */ + adapter->reset_reason = ENA_REGS_RESET_SHUTDOWN; + ena_destroy_device(adapter, true); + + if (shutdown) { + netif_device_detach(netdev); + dev_close(netdev); + rtnl_unlock(); + } else { + rtnl_unlock(); + unregister_netdev(netdev); + free_netdev(netdev); + } + + ena_com_rss_destroy(ena_dev); + + ena_com_delete_debug_area(ena_dev); + + ena_com_delete_host_info(ena_dev); + + ena_release_bars(ena_dev, pdev); + + pci_disable_device(pdev); + + vfree(ena_dev); +} + +/* ena_remove - Device Removal Routine + * @pdev: PCI device information struct + * + * ena_remove is called by the PCI subsystem to alert the driver + * that it should release a PCI device. + */ + +static void ena_remove(struct pci_dev *pdev) +{ + __ena_shutoff(pdev, false); +} + +/* ena_shutdown - Device Shutdown Routine + * @pdev: PCI device information struct + * + * ena_shutdown is called by the PCI subsystem to alert the driver that + * a shutdown/reboot (or kexec) is happening and device must be disabled. + */ + +static void ena_shutdown(struct pci_dev *pdev) +{ + __ena_shutoff(pdev, true); +} + +/* ena_suspend - PM suspend callback + * @dev_d: Device information struct + */ +static int __maybe_unused ena_suspend(struct device *dev_d) +{ + struct pci_dev *pdev = to_pci_dev(dev_d); + struct ena_adapter *adapter = pci_get_drvdata(pdev); + + ena_increase_stat(&adapter->dev_stats.suspend, 1, &adapter->syncp); + + rtnl_lock(); + if (unlikely(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags))) { + dev_err(&pdev->dev, + "Ignoring device reset request as the device is being suspended\n"); + clear_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags); + } + ena_destroy_device(adapter, true); + rtnl_unlock(); + return 0; +} + +/* ena_resume - PM resume callback + * @dev_d: Device information struct + */ +static int __maybe_unused ena_resume(struct device *dev_d) +{ + struct ena_adapter *adapter = dev_get_drvdata(dev_d); + int rc; + + ena_increase_stat(&adapter->dev_stats.resume, 1, &adapter->syncp); + + rtnl_lock(); + rc = ena_restore_device(adapter); + rtnl_unlock(); + return rc; +} + +static SIMPLE_DEV_PM_OPS(ena_pm_ops, ena_suspend, ena_resume); + +static struct pci_driver ena_pci_driver = { + .name = DRV_MODULE_NAME, + .id_table = ena_pci_tbl, + .probe = ena_probe, + .remove = ena_remove, + .shutdown = ena_shutdown, + .driver.pm = &ena_pm_ops, + .sriov_configure = pci_sriov_configure_simple, +}; + +static int __init ena_init(void) +{ + int ret; + + ena_wq = create_singlethread_workqueue(DRV_MODULE_NAME); + if (!ena_wq) { + pr_err("Failed to create workqueue\n"); + return -ENOMEM; + } + + ret = pci_register_driver(&ena_pci_driver); + if (ret) + destroy_workqueue(ena_wq); + + return ret; +} + +static void __exit ena_cleanup(void) +{ + pci_unregister_driver(&ena_pci_driver); + + if (ena_wq) { + destroy_workqueue(ena_wq); + ena_wq = NULL; + } +} + +/****************************************************************************** + ******************************** AENQ Handlers ******************************* + *****************************************************************************/ +/* ena_update_on_link_change: + * Notify the network interface about the change in link status + */ +static void ena_update_on_link_change(void *adapter_data, + struct ena_admin_aenq_entry *aenq_e) +{ + struct ena_adapter *adapter = (struct ena_adapter *)adapter_data; + struct ena_admin_aenq_link_change_desc *aenq_desc = + (struct ena_admin_aenq_link_change_desc *)aenq_e; + int status = aenq_desc->flags & + ENA_ADMIN_AENQ_LINK_CHANGE_DESC_LINK_STATUS_MASK; + + if (status) { + netif_dbg(adapter, ifup, adapter->netdev, "%s\n", __func__); + set_bit(ENA_FLAG_LINK_UP, &adapter->flags); + if (!test_bit(ENA_FLAG_ONGOING_RESET, &adapter->flags)) + netif_carrier_on(adapter->netdev); + } else { + clear_bit(ENA_FLAG_LINK_UP, &adapter->flags); + netif_carrier_off(adapter->netdev); + } +} + +static void ena_keep_alive_wd(void *adapter_data, + struct ena_admin_aenq_entry *aenq_e) +{ + struct ena_adapter *adapter = (struct ena_adapter *)adapter_data; + struct ena_admin_aenq_keep_alive_desc *desc; + u64 rx_drops; + u64 tx_drops; + + desc = (struct ena_admin_aenq_keep_alive_desc *)aenq_e; + adapter->last_keep_alive_jiffies = jiffies; + + rx_drops = ((u64)desc->rx_drops_high << 32) | desc->rx_drops_low; + tx_drops = ((u64)desc->tx_drops_high << 32) | desc->tx_drops_low; + + u64_stats_update_begin(&adapter->syncp); + /* These stats are accumulated by the device, so the counters indicate + * all drops since last reset. + */ + adapter->dev_stats.rx_drops = rx_drops; + adapter->dev_stats.tx_drops = tx_drops; + u64_stats_update_end(&adapter->syncp); +} + +static void ena_notification(void *adapter_data, + struct ena_admin_aenq_entry *aenq_e) +{ + struct ena_adapter *adapter = (struct ena_adapter *)adapter_data; + struct ena_admin_ena_hw_hints *hints; + + WARN(aenq_e->aenq_common_desc.group != ENA_ADMIN_NOTIFICATION, + "Invalid group(%x) expected %x\n", + aenq_e->aenq_common_desc.group, + ENA_ADMIN_NOTIFICATION); + + switch (aenq_e->aenq_common_desc.syndrome) { + case ENA_ADMIN_UPDATE_HINTS: + hints = (struct ena_admin_ena_hw_hints *) + (&aenq_e->inline_data_w4); + ena_update_hints(adapter, hints); + break; + default: + netif_err(adapter, drv, adapter->netdev, + "Invalid aenq notification link state %d\n", + aenq_e->aenq_common_desc.syndrome); + } +} + +/* This handler will called for unknown event group or unimplemented handlers*/ +static void unimplemented_aenq_handler(void *data, + struct ena_admin_aenq_entry *aenq_e) +{ + struct ena_adapter *adapter = (struct ena_adapter *)data; + + netif_err(adapter, drv, adapter->netdev, + "Unknown event was received or event with unimplemented handler\n"); +} + +static struct ena_aenq_handlers aenq_handlers = { + .handlers = { + [ENA_ADMIN_LINK_CHANGE] = ena_update_on_link_change, + [ENA_ADMIN_NOTIFICATION] = ena_notification, + [ENA_ADMIN_KEEP_ALIVE] = ena_keep_alive_wd, + }, + .unimplemented_handler = unimplemented_aenq_handler +}; + +module_init(ena_init); +module_exit(ena_cleanup); diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.h b/drivers/net/ethernet/amazon/ena/ena_netdev.h new file mode 100644 index 0000000000..33c923e126 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.h @@ -0,0 +1,467 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#ifndef ENA_H +#define ENA_H + +#include <linux/bitops.h> +#include <linux/dim.h> +#include <linux/etherdevice.h> +#include <linux/if_vlan.h> +#include <linux/inetdevice.h> +#include <linux/interrupt.h> +#include <linux/netdevice.h> +#include <linux/skbuff.h> +#include <net/xdp.h> +#include <uapi/linux/bpf.h> + +#include "ena_com.h" +#include "ena_eth_com.h" + +#define DRV_MODULE_GEN_MAJOR 2 +#define DRV_MODULE_GEN_MINOR 1 +#define DRV_MODULE_GEN_SUBMINOR 0 + +#define DRV_MODULE_NAME "ena" + +#define DEVICE_NAME "Elastic Network Adapter (ENA)" + +/* 1 for AENQ + ADMIN */ +#define ENA_ADMIN_MSIX_VEC 1 +#define ENA_MAX_MSIX_VEC(io_queues) (ENA_ADMIN_MSIX_VEC + (io_queues)) + +/* The ENA buffer length fields is 16 bit long. So when PAGE_SIZE == 64kB the + * driver passes 0. + * Since the max packet size the ENA handles is ~9kB limit the buffer length to + * 16kB. + */ +#if PAGE_SIZE > SZ_16K +#define ENA_PAGE_SIZE (_AC(SZ_16K, UL)) +#else +#define ENA_PAGE_SIZE PAGE_SIZE +#endif + +#define ENA_MIN_MSIX_VEC 2 + +#define ENA_REG_BAR 0 +#define ENA_MEM_BAR 2 +#define ENA_BAR_MASK (BIT(ENA_REG_BAR) | BIT(ENA_MEM_BAR)) + +#define ENA_DEFAULT_RING_SIZE (1024) +#define ENA_MIN_RING_SIZE (256) + +#define ENA_MIN_RX_BUF_SIZE (2048) + +#define ENA_MIN_NUM_IO_QUEUES (1) + +#define ENA_TX_WAKEUP_THRESH (MAX_SKB_FRAGS + 2) +#define ENA_DEFAULT_RX_COPYBREAK (256 - NET_IP_ALIGN) + +#define ENA_MIN_MTU 128 + +#define ENA_NAME_MAX_LEN 20 +#define ENA_IRQNAME_SIZE 40 + +#define ENA_PKT_MAX_BUFS 19 + +#define ENA_RX_RSS_TABLE_LOG_SIZE 7 +#define ENA_RX_RSS_TABLE_SIZE (1 << ENA_RX_RSS_TABLE_LOG_SIZE) + +/* The number of tx packet completions that will be handled each NAPI poll + * cycle is ring_size / ENA_TX_POLL_BUDGET_DIVIDER. + */ +#define ENA_TX_POLL_BUDGET_DIVIDER 4 + +/* Refill Rx queue when number of required descriptors is above + * QUEUE_SIZE / ENA_RX_REFILL_THRESH_DIVIDER or ENA_RX_REFILL_THRESH_PACKET + */ +#define ENA_RX_REFILL_THRESH_DIVIDER 8 +#define ENA_RX_REFILL_THRESH_PACKET 256 + +/* Number of queues to check for missing queues per timer service */ +#define ENA_MONITORED_TX_QUEUES 4 +/* Max timeout packets before device reset */ +#define MAX_NUM_OF_TIMEOUTED_PACKETS 128 + +#define ENA_TX_RING_IDX_NEXT(idx, ring_size) (((idx) + 1) & ((ring_size) - 1)) + +#define ENA_RX_RING_IDX_NEXT(idx, ring_size) (((idx) + 1) & ((ring_size) - 1)) +#define ENA_RX_RING_IDX_ADD(idx, n, ring_size) \ + (((idx) + (n)) & ((ring_size) - 1)) + +#define ENA_IO_TXQ_IDX(q) (2 * (q)) +#define ENA_IO_RXQ_IDX(q) (2 * (q) + 1) +#define ENA_IO_TXQ_IDX_TO_COMBINED_IDX(q) ((q) / 2) +#define ENA_IO_RXQ_IDX_TO_COMBINED_IDX(q) (((q) - 1) / 2) + +#define ENA_MGMNT_IRQ_IDX 0 +#define ENA_IO_IRQ_FIRST_IDX 1 +#define ENA_IO_IRQ_IDX(q) (ENA_IO_IRQ_FIRST_IDX + (q)) + +#define ENA_ADMIN_POLL_DELAY_US 100 + +/* ENA device should send keep alive msg every 1 sec. + * We wait for 6 sec just to be on the safe side. + */ +#define ENA_DEVICE_KALIVE_TIMEOUT (6 * HZ) +#define ENA_MAX_NO_INTERRUPT_ITERATIONS 3 + +#define ENA_MMIO_DISABLE_REG_READ BIT(0) + +/* The max MTU size is configured to be the ethernet frame size without + * the overhead of the ethernet header, which can have a VLAN header, and + * a frame check sequence (FCS). + * The buffer size we share with the device is defined to be ENA_PAGE_SIZE + */ + +#define ENA_XDP_MAX_MTU (ENA_PAGE_SIZE - ETH_HLEN - ETH_FCS_LEN - \ + VLAN_HLEN - XDP_PACKET_HEADROOM - \ + SKB_DATA_ALIGN(sizeof(struct skb_shared_info))) + +#define ENA_IS_XDP_INDEX(adapter, index) (((index) >= (adapter)->xdp_first_ring) && \ + ((index) < (adapter)->xdp_first_ring + (adapter)->xdp_num_queues)) + +struct ena_irq { + irq_handler_t handler; + void *data; + int cpu; + u32 vector; + cpumask_t affinity_hint_mask; + char name[ENA_IRQNAME_SIZE]; +}; + +struct ena_napi { + u8 first_interrupt ____cacheline_aligned; + u8 interrupts_masked; + struct napi_struct napi; + struct ena_ring *tx_ring; + struct ena_ring *rx_ring; + struct ena_ring *xdp_ring; + u32 qid; + struct dim dim; +}; + +struct ena_tx_buffer { + struct sk_buff *skb; + /* num of ena desc for this specific skb + * (includes data desc and metadata desc) + */ + u32 tx_descs; + /* num of buffers used by this skb */ + u32 num_of_bufs; + + /* XDP buffer structure which is used for sending packets in + * the xdp queues + */ + struct xdp_frame *xdpf; + + /* Indicate if bufs[0] map the linear data of the skb. */ + u8 map_linear_data; + + /* Used for detect missing tx packets to limit the number of prints */ + u32 print_once; + /* Save the last jiffies to detect missing tx packets + * + * sets to non zero value on ena_start_xmit and set to zero on + * napi and timer_Service_routine. + * + * while this value is not protected by lock, + * a given packet is not expected to be handled by ena_start_xmit + * and by napi/timer_service at the same time. + */ + unsigned long last_jiffies; + struct ena_com_buf bufs[ENA_PKT_MAX_BUFS]; +} ____cacheline_aligned; + +struct ena_rx_buffer { + struct sk_buff *skb; + struct page *page; + dma_addr_t dma_addr; + u32 page_offset; + u32 buf_offset; + struct ena_com_buf ena_buf; +} ____cacheline_aligned; + +struct ena_stats_tx { + u64 cnt; + u64 bytes; + u64 queue_stop; + u64 prepare_ctx_err; + u64 queue_wakeup; + u64 dma_mapping_err; + u64 linearize; + u64 linearize_failed; + u64 napi_comp; + u64 tx_poll; + u64 doorbells; + u64 bad_req_id; + u64 llq_buffer_copy; + u64 missed_tx; + u64 unmask_interrupt; + u64 last_napi_jiffies; +}; + +struct ena_stats_rx { + u64 cnt; + u64 bytes; + u64 rx_copybreak_pkt; + u64 csum_good; + u64 refil_partial; + u64 csum_bad; + u64 page_alloc_fail; + u64 skb_alloc_fail; + u64 dma_mapping_err; + u64 bad_desc_num; + u64 bad_req_id; + u64 empty_rx_ring; + u64 csum_unchecked; + u64 xdp_aborted; + u64 xdp_drop; + u64 xdp_pass; + u64 xdp_tx; + u64 xdp_invalid; + u64 xdp_redirect; +}; + +struct ena_ring { + /* Holds the empty requests for TX/RX + * out of order completions + */ + u16 *free_ids; + + union { + struct ena_tx_buffer *tx_buffer_info; + struct ena_rx_buffer *rx_buffer_info; + }; + + /* cache ptr to avoid using the adapter */ + struct device *dev; + struct pci_dev *pdev; + struct napi_struct *napi; + struct net_device *netdev; + struct ena_com_dev *ena_dev; + struct ena_adapter *adapter; + struct ena_com_io_cq *ena_com_io_cq; + struct ena_com_io_sq *ena_com_io_sq; + struct bpf_prog *xdp_bpf_prog; + struct xdp_rxq_info xdp_rxq; + spinlock_t xdp_tx_lock; /* synchronize XDP TX/Redirect traffic */ + /* Used for rx queues only to point to the xdp tx ring, to + * which traffic should be redirected from this rx ring. + */ + struct ena_ring *xdp_ring; + + u16 next_to_use; + u16 next_to_clean; + u16 rx_copybreak; + u16 rx_headroom; + u16 qid; + u16 mtu; + u16 sgl_size; + + /* The maximum header length the device can handle */ + u8 tx_max_header_size; + + bool disable_meta_caching; + u16 no_interrupt_event_cnt; + + /* cpu and NUMA for TPH */ + int cpu; + int numa_node; + + /* number of tx/rx_buffer_info's entries */ + int ring_size; + + enum ena_admin_placement_policy_type tx_mem_queue_type; + + struct ena_com_rx_buf_info ena_bufs[ENA_PKT_MAX_BUFS]; + u32 smoothed_interval; + u32 per_napi_packets; + u16 non_empty_napi_events; + struct u64_stats_sync syncp; + union { + struct ena_stats_tx tx_stats; + struct ena_stats_rx rx_stats; + }; + + u8 *push_buf_intermediate_buf; + int empty_rx_queue; +} ____cacheline_aligned; + +struct ena_stats_dev { + u64 tx_timeout; + u64 suspend; + u64 resume; + u64 wd_expired; + u64 interface_up; + u64 interface_down; + u64 admin_q_pause; + u64 rx_drops; + u64 tx_drops; +}; + +enum ena_flags_t { + ENA_FLAG_DEVICE_RUNNING, + ENA_FLAG_DEV_UP, + ENA_FLAG_LINK_UP, + ENA_FLAG_MSIX_ENABLED, + ENA_FLAG_TRIGGER_RESET, + ENA_FLAG_ONGOING_RESET +}; + +/* adapter specific private data structure */ +struct ena_adapter { + struct ena_com_dev *ena_dev; + /* OS defined structs */ + struct net_device *netdev; + struct pci_dev *pdev; + + /* rx packets that shorter that this len will be copied to the skb + * header + */ + u32 rx_copybreak; + u32 max_mtu; + + u32 num_io_queues; + u32 max_num_io_queues; + + int msix_vecs; + + u32 missing_tx_completion_threshold; + + u32 requested_tx_ring_size; + u32 requested_rx_ring_size; + + u32 max_tx_ring_size; + u32 max_rx_ring_size; + + u32 msg_enable; + + /* large_llq_header_enabled is used for two purposes: + * 1. Indicates that large LLQ has been requested. + * 2. Indicates whether large LLQ is set or not after device + * initialization / configuration. + */ + bool large_llq_header_enabled; + bool large_llq_header_supported; + + u16 max_tx_sgl_size; + u16 max_rx_sgl_size; + + u8 mac_addr[ETH_ALEN]; + + unsigned long keep_alive_timeout; + unsigned long missing_tx_completion_to; + + char name[ENA_NAME_MAX_LEN]; + + unsigned long flags; + /* TX */ + struct ena_ring tx_ring[ENA_MAX_NUM_IO_QUEUES] + ____cacheline_aligned_in_smp; + + /* RX */ + struct ena_ring rx_ring[ENA_MAX_NUM_IO_QUEUES] + ____cacheline_aligned_in_smp; + + struct ena_napi ena_napi[ENA_MAX_NUM_IO_QUEUES]; + + struct ena_irq irq_tbl[ENA_MAX_MSIX_VEC(ENA_MAX_NUM_IO_QUEUES)]; + + /* timer service */ + struct work_struct reset_task; + struct timer_list timer_service; + + bool wd_state; + bool dev_up_before_reset; + bool disable_meta_caching; + unsigned long last_keep_alive_jiffies; + + struct u64_stats_sync syncp; + struct ena_stats_dev dev_stats; + struct ena_admin_eni_stats eni_stats; + + /* last queue index that was checked for uncompleted tx packets */ + u32 last_monitored_tx_qid; + + enum ena_regs_reset_reason_types reset_reason; + + struct bpf_prog *xdp_bpf_prog; + u32 xdp_first_ring; + u32 xdp_num_queues; +}; + +void ena_set_ethtool_ops(struct net_device *netdev); + +void ena_dump_stats_to_dmesg(struct ena_adapter *adapter); + +void ena_dump_stats_to_buf(struct ena_adapter *adapter, u8 *buf); + +int ena_update_hw_stats(struct ena_adapter *adapter); + +int ena_update_queue_params(struct ena_adapter *adapter, + u32 new_tx_size, + u32 new_rx_size, + u32 new_llq_header_len); + +int ena_update_queue_count(struct ena_adapter *adapter, u32 new_channel_count); + +int ena_set_rx_copybreak(struct ena_adapter *adapter, u32 rx_copybreak); + +int ena_get_sset_count(struct net_device *netdev, int sset); + +static inline void ena_reset_device(struct ena_adapter *adapter, + enum ena_regs_reset_reason_types reset_reason) +{ + adapter->reset_reason = reset_reason; + /* Make sure reset reason is set before triggering the reset */ + smp_mb__before_atomic(); + set_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags); +} + +enum ena_xdp_errors_t { + ENA_XDP_ALLOWED = 0, + ENA_XDP_CURRENT_MTU_TOO_LARGE, + ENA_XDP_NO_ENOUGH_QUEUES, +}; + +enum ENA_XDP_ACTIONS { + ENA_XDP_PASS = 0, + ENA_XDP_TX = BIT(0), + ENA_XDP_REDIRECT = BIT(1), + ENA_XDP_DROP = BIT(2) +}; + +#define ENA_XDP_FORWARDED (ENA_XDP_TX | ENA_XDP_REDIRECT) + +static inline bool ena_xdp_present(struct ena_adapter *adapter) +{ + return !!adapter->xdp_bpf_prog; +} + +static inline bool ena_xdp_present_ring(struct ena_ring *ring) +{ + return !!ring->xdp_bpf_prog; +} + +static inline bool ena_xdp_legal_queue_count(struct ena_adapter *adapter, + u32 queues) +{ + return 2 * queues <= adapter->max_num_io_queues; +} + +static inline enum ena_xdp_errors_t ena_xdp_allowed(struct ena_adapter *adapter) +{ + enum ena_xdp_errors_t rc = ENA_XDP_ALLOWED; + + if (adapter->netdev->mtu > ENA_XDP_MAX_MTU) + rc = ENA_XDP_CURRENT_MTU_TOO_LARGE; + else if (!ena_xdp_legal_queue_count(adapter, adapter->num_io_queues)) + rc = ENA_XDP_NO_ENOUGH_QUEUES; + + return rc; +} + +#endif /* !(ENA_H) */ diff --git a/drivers/net/ethernet/amazon/ena/ena_pci_id_tbl.h b/drivers/net/ethernet/amazon/ena/ena_pci_id_tbl.h new file mode 100644 index 0000000000..3ecdf29160 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_pci_id_tbl.h @@ -0,0 +1,45 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ + +#ifndef ENA_PCI_ID_TBL_H_ +#define ENA_PCI_ID_TBL_H_ + +#ifndef PCI_VENDOR_ID_AMAZON +#define PCI_VENDOR_ID_AMAZON 0x1d0f +#endif + +#ifndef PCI_DEV_ID_ENA_PF +#define PCI_DEV_ID_ENA_PF 0x0ec2 +#endif + +#ifndef PCI_DEV_ID_ENA_LLQ_PF +#define PCI_DEV_ID_ENA_LLQ_PF 0x1ec2 +#endif + +#ifndef PCI_DEV_ID_ENA_VF +#define PCI_DEV_ID_ENA_VF 0xec20 +#endif + +#ifndef PCI_DEV_ID_ENA_LLQ_VF +#define PCI_DEV_ID_ENA_LLQ_VF 0xec21 +#endif + +#ifndef PCI_DEV_ID_ENA_RESRV0 +#define PCI_DEV_ID_ENA_RESRV0 0x0051 +#endif + +#define ENA_PCI_ID_TABLE_ENTRY(devid) \ + {PCI_DEVICE(PCI_VENDOR_ID_AMAZON, devid)}, + +static const struct pci_device_id ena_pci_tbl[] = { + ENA_PCI_ID_TABLE_ENTRY(PCI_DEV_ID_ENA_RESRV0) + ENA_PCI_ID_TABLE_ENTRY(PCI_DEV_ID_ENA_PF) + ENA_PCI_ID_TABLE_ENTRY(PCI_DEV_ID_ENA_LLQ_PF) + ENA_PCI_ID_TABLE_ENTRY(PCI_DEV_ID_ENA_VF) + ENA_PCI_ID_TABLE_ENTRY(PCI_DEV_ID_ENA_LLQ_VF) + { } +}; + +#endif /* ENA_PCI_ID_TBL_H_ */ diff --git a/drivers/net/ethernet/amazon/ena/ena_regs_defs.h b/drivers/net/ethernet/amazon/ena/ena_regs_defs.h new file mode 100644 index 0000000000..1e007a41a5 --- /dev/null +++ b/drivers/net/ethernet/amazon/ena/ena_regs_defs.h @@ -0,0 +1,130 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */ +/* + * Copyright 2015-2020 Amazon.com, Inc. or its affiliates. All rights reserved. + */ +#ifndef _ENA_REGS_H_ +#define _ENA_REGS_H_ + +enum ena_regs_reset_reason_types { + ENA_REGS_RESET_NORMAL = 0, + ENA_REGS_RESET_KEEP_ALIVE_TO = 1, + ENA_REGS_RESET_ADMIN_TO = 2, + ENA_REGS_RESET_MISS_TX_CMPL = 3, + ENA_REGS_RESET_INV_RX_REQ_ID = 4, + ENA_REGS_RESET_INV_TX_REQ_ID = 5, + ENA_REGS_RESET_TOO_MANY_RX_DESCS = 6, + ENA_REGS_RESET_INIT_ERR = 7, + ENA_REGS_RESET_DRIVER_INVALID_STATE = 8, + ENA_REGS_RESET_OS_TRIGGER = 9, + ENA_REGS_RESET_OS_NETDEV_WD = 10, + ENA_REGS_RESET_SHUTDOWN = 11, + ENA_REGS_RESET_USER_TRIGGER = 12, + ENA_REGS_RESET_GENERIC = 13, + ENA_REGS_RESET_MISS_INTERRUPT = 14, +}; + +/* ena_registers offsets */ + +/* 0 base */ +#define ENA_REGS_VERSION_OFF 0x0 +#define ENA_REGS_CONTROLLER_VERSION_OFF 0x4 +#define ENA_REGS_CAPS_OFF 0x8 +#define ENA_REGS_CAPS_EXT_OFF 0xc +#define ENA_REGS_AQ_BASE_LO_OFF 0x10 +#define ENA_REGS_AQ_BASE_HI_OFF 0x14 +#define ENA_REGS_AQ_CAPS_OFF 0x18 +#define ENA_REGS_ACQ_BASE_LO_OFF 0x20 +#define ENA_REGS_ACQ_BASE_HI_OFF 0x24 +#define ENA_REGS_ACQ_CAPS_OFF 0x28 +#define ENA_REGS_AQ_DB_OFF 0x2c +#define ENA_REGS_ACQ_TAIL_OFF 0x30 +#define ENA_REGS_AENQ_CAPS_OFF 0x34 +#define ENA_REGS_AENQ_BASE_LO_OFF 0x38 +#define ENA_REGS_AENQ_BASE_HI_OFF 0x3c +#define ENA_REGS_AENQ_HEAD_DB_OFF 0x40 +#define ENA_REGS_AENQ_TAIL_OFF 0x44 +#define ENA_REGS_INTR_MASK_OFF 0x4c +#define ENA_REGS_DEV_CTL_OFF 0x54 +#define ENA_REGS_DEV_STS_OFF 0x58 +#define ENA_REGS_MMIO_REG_READ_OFF 0x5c +#define ENA_REGS_MMIO_RESP_LO_OFF 0x60 +#define ENA_REGS_MMIO_RESP_HI_OFF 0x64 +#define ENA_REGS_RSS_IND_ENTRY_UPDATE_OFF 0x68 + +/* version register */ +#define ENA_REGS_VERSION_MINOR_VERSION_MASK 0xff +#define ENA_REGS_VERSION_MAJOR_VERSION_SHIFT 8 +#define ENA_REGS_VERSION_MAJOR_VERSION_MASK 0xff00 + +/* controller_version register */ +#define ENA_REGS_CONTROLLER_VERSION_SUBMINOR_VERSION_MASK 0xff +#define ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_SHIFT 8 +#define ENA_REGS_CONTROLLER_VERSION_MINOR_VERSION_MASK 0xff00 +#define ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_SHIFT 16 +#define ENA_REGS_CONTROLLER_VERSION_MAJOR_VERSION_MASK 0xff0000 +#define ENA_REGS_CONTROLLER_VERSION_IMPL_ID_SHIFT 24 +#define ENA_REGS_CONTROLLER_VERSION_IMPL_ID_MASK 0xff000000 + +/* caps register */ +#define ENA_REGS_CAPS_CONTIGUOUS_QUEUE_REQUIRED_MASK 0x1 +#define ENA_REGS_CAPS_RESET_TIMEOUT_SHIFT 1 +#define ENA_REGS_CAPS_RESET_TIMEOUT_MASK 0x3e +#define ENA_REGS_CAPS_DMA_ADDR_WIDTH_SHIFT 8 +#define ENA_REGS_CAPS_DMA_ADDR_WIDTH_MASK 0xff00 +#define ENA_REGS_CAPS_ADMIN_CMD_TO_SHIFT 16 +#define ENA_REGS_CAPS_ADMIN_CMD_TO_MASK 0xf0000 + +/* aq_caps register */ +#define ENA_REGS_AQ_CAPS_AQ_DEPTH_MASK 0xffff +#define ENA_REGS_AQ_CAPS_AQ_ENTRY_SIZE_SHIFT 16 +#define ENA_REGS_AQ_CAPS_AQ_ENTRY_SIZE_MASK 0xffff0000 + +/* acq_caps register */ +#define ENA_REGS_ACQ_CAPS_ACQ_DEPTH_MASK 0xffff +#define ENA_REGS_ACQ_CAPS_ACQ_ENTRY_SIZE_SHIFT 16 +#define ENA_REGS_ACQ_CAPS_ACQ_ENTRY_SIZE_MASK 0xffff0000 + +/* aenq_caps register */ +#define ENA_REGS_AENQ_CAPS_AENQ_DEPTH_MASK 0xffff +#define ENA_REGS_AENQ_CAPS_AENQ_ENTRY_SIZE_SHIFT 16 +#define ENA_REGS_AENQ_CAPS_AENQ_ENTRY_SIZE_MASK 0xffff0000 + +/* dev_ctl register */ +#define ENA_REGS_DEV_CTL_DEV_RESET_MASK 0x1 +#define ENA_REGS_DEV_CTL_AQ_RESTART_SHIFT 1 +#define ENA_REGS_DEV_CTL_AQ_RESTART_MASK 0x2 +#define ENA_REGS_DEV_CTL_QUIESCENT_SHIFT 2 +#define ENA_REGS_DEV_CTL_QUIESCENT_MASK 0x4 +#define ENA_REGS_DEV_CTL_IO_RESUME_SHIFT 3 +#define ENA_REGS_DEV_CTL_IO_RESUME_MASK 0x8 +#define ENA_REGS_DEV_CTL_RESET_REASON_SHIFT 28 +#define ENA_REGS_DEV_CTL_RESET_REASON_MASK 0xf0000000 + +/* dev_sts register */ +#define ENA_REGS_DEV_STS_READY_MASK 0x1 +#define ENA_REGS_DEV_STS_AQ_RESTART_IN_PROGRESS_SHIFT 1 +#define ENA_REGS_DEV_STS_AQ_RESTART_IN_PROGRESS_MASK 0x2 +#define ENA_REGS_DEV_STS_AQ_RESTART_FINISHED_SHIFT 2 +#define ENA_REGS_DEV_STS_AQ_RESTART_FINISHED_MASK 0x4 +#define ENA_REGS_DEV_STS_RESET_IN_PROGRESS_SHIFT 3 +#define ENA_REGS_DEV_STS_RESET_IN_PROGRESS_MASK 0x8 +#define ENA_REGS_DEV_STS_RESET_FINISHED_SHIFT 4 +#define ENA_REGS_DEV_STS_RESET_FINISHED_MASK 0x10 +#define ENA_REGS_DEV_STS_FATAL_ERROR_SHIFT 5 +#define ENA_REGS_DEV_STS_FATAL_ERROR_MASK 0x20 +#define ENA_REGS_DEV_STS_QUIESCENT_STATE_IN_PROGRESS_SHIFT 6 +#define ENA_REGS_DEV_STS_QUIESCENT_STATE_IN_PROGRESS_MASK 0x40 +#define ENA_REGS_DEV_STS_QUIESCENT_STATE_ACHIEVED_SHIFT 7 +#define ENA_REGS_DEV_STS_QUIESCENT_STATE_ACHIEVED_MASK 0x80 + +/* mmio_reg_read register */ +#define ENA_REGS_MMIO_REG_READ_REQ_ID_MASK 0xffff +#define ENA_REGS_MMIO_REG_READ_REG_OFF_SHIFT 16 +#define ENA_REGS_MMIO_REG_READ_REG_OFF_MASK 0xffff0000 + +/* rss_ind_entry_update register */ +#define ENA_REGS_RSS_IND_ENTRY_UPDATE_INDEX_MASK 0xffff +#define ENA_REGS_RSS_IND_ENTRY_UPDATE_CQ_IDX_SHIFT 16 +#define ENA_REGS_RSS_IND_ENTRY_UPDATE_CQ_IDX_MASK 0xffff0000 + +#endif /* _ENA_REGS_H_ */ |