/*      $NetBSD: if_enavar.h,v 1.10 2024/02/09 22:08:35 andvar Exp $    */

/*-
* BSD LICENSE
*
* Copyright (c) 2015-2017 Amazon.com, Inc. or its affiliates.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* $FreeBSD: head/sys/dev/ena/ena.h 333450 2018-05-10 09:06:21Z mw $
*
*/

#ifndef ENA_H
#define ENA_H

#include <sys/types.h>
#include <sys/atomic.h>
#include <sys/pcq.h>

#include "external/bsd/ena-com/ena_com.h"
#include "external/bsd/ena-com/ena_eth_com.h"

#define DRV_MODULE_VER_MAJOR    0
#define DRV_MODULE_VER_MINOR    8
#define DRV_MODULE_VER_SUBMINOR 1

#define DRV_MODULE_NAME         "ena"

#ifndef DRV_MODULE_VERSION
#define DRV_MODULE_VERSION                              \
       ___STRING(DRV_MODULE_VER_MAJOR) "."             \
       ___STRING(DRV_MODULE_VER_MINOR) "."             \
       ___STRING(DRV_MODULE_VER_SUBMINOR)
#endif
#define DEVICE_NAME     "Elastic Network Adapter (ENA)"
#define DEVICE_DESC     "ENA adapter"

/* Calculate DMA mask - width for ena cannot exceed 48, so it is safe */
#define ENA_DMA_BIT_MASK(x)             ((1ULL << (x)) - 1ULL)

/* 1 for AENQ + ADMIN */
#define ENA_ADMIN_MSIX_VEC              1
#define ENA_MAX_MSIX_VEC(io_queues)     (ENA_ADMIN_MSIX_VEC + (io_queues))

#define ENA_REG_BAR                     PCI_BAR(0)
#define ENA_MEM_BAR                     PCI_BAR(2)

#define ENA_BUS_DMA_SEGS                32

#define ENA_DEFAULT_RING_SIZE           1024

#define ENA_RX_REFILL_THRESH_DIVIDER    8

#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)

#define ENA_HASH_KEY_SIZE               40

#define ENA_MAX_FRAME_LEN               10000
#define ENA_MIN_FRAME_LEN               60

#define ENA_TX_CLEANUP_THRESHOLD        128

#define DB_THRESHOLD    64

#define TX_COMMIT       32
/*
* TX budget for cleaning. It should be half of the RX budget to reduce amount
*  of TCP retransmissions.
*/
#define TX_BUDGET       128
/* RX cleanup budget. -1 stands for infinity. */
#define RX_BUDGET       256
/*
* How many times we can repeat cleanup in the io irq handling routine if the
* RX or TX budget was depleted.
*/
#define CLEAN_BUDGET    8

#define RX_IRQ_INTERVAL 20
#define TX_IRQ_INTERVAL 50

#define ENA_MIN_MTU             128

#define ENA_TSO_MAXSIZE         65536

#define ENA_MMIO_DISABLE_REG_READ       BIT(0)

#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_IO_TXQ_IDX(q)               (2 * (q))
#define ENA_IO_RXQ_IDX(q)               (2 * (q) + 1)

#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))

/*
* ENA device should send keep alive msg every 1 sec.
* We wait for 6 sec just to be on the safe side.
*/
#define DEFAULT_KEEP_ALIVE_TO           (SBT_1S * 6)

/* Time in jiffies before concluding the transmitter is hung. */
#define DEFAULT_TX_CMP_TO               (SBT_1S * 5)

/* Number of queues to check for missing queues per timer tick */
#define DEFAULT_TX_MONITORED_QUEUES     (4)

/* Max number of timeouted packets before device reset */
#define DEFAULT_TX_CMP_THRESHOLD        (128)

/*
* Supported PCI vendor and devices IDs
*/
#define PCI_VENDOR_ID_AMAZON    0x1d0f

#define PCI_DEV_ID_ENA_PF       0x0ec2
#define PCI_DEV_ID_ENA_LLQ_PF   0x1ec2
#define PCI_DEV_ID_ENA_VF       0xec20
#define PCI_DEV_ID_ENA_LLQ_VF   0xec21

/*
* Flags indicating current ENA driver state
*/
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,
       ENA_FLAG_DEV_UP_BEFORE_RESET,
       ENA_FLAG_RSS_ACTIVE,
       ENA_FLAGS_NUMBER = ENA_FLAG_RSS_ACTIVE
};

#define ENA_FLAG_BITMASK(bit)   (~(uint32_t)__BIT(bit))
#define ENA_FLAG_ZERO(adapter)  (adapter)->flags = 0;
#define ENA_FLAG_ISSET(bit, adapter)    ((adapter)->flags & __BIT(bit))
#define ENA_FLAG_SET_ATOMIC(bit, adapter)       \
       atomic_or_32(&(adapter)->flags, __BIT(bit))
#define ENA_FLAG_CLEAR_ATOMIC(bit, adapter)     \
       atomic_and_32(&(adapter)->flags, ENA_FLAG_BITMASK(bit))

typedef __int64_t sbintime_t;

struct msix_entry {
       int entry;
       int vector;
};

typedef struct _ena_vendor_info_t {
       unsigned int vendor_id;
       unsigned int device_id;
       unsigned int index;
} ena_vendor_info_t;

struct ena_que {
       struct ena_adapter *adapter;
       struct ena_ring *tx_ring;
       struct ena_ring *rx_ring;
       uint32_t id;
       int cpu;
};

struct ena_tx_buffer {
       struct mbuf *mbuf;
       /* # of ena desc for this specific mbuf
        * (includes data desc and metadata desc) */
       unsigned int tx_descs;
       /* # of buffers used by this mbuf */
       unsigned int num_of_bufs;
       bus_dmamap_t map;

       /* Used to detect missing tx packets */
       struct bintime timestamp;
       bool print_once;

       struct ena_com_buf bufs[ENA_PKT_MAX_BUFS];
} __aligned(CACHE_LINE_SIZE);

struct ena_rx_buffer {
       struct mbuf *mbuf;
       bus_dmamap_t map;
       struct ena_com_buf ena_buf;
} __aligned(CACHE_LINE_SIZE);

struct ena_stats_tx {
       char name[16];
       struct evcnt cnt;
       struct evcnt bytes;
       struct evcnt prepare_ctx_err;
       struct evcnt dma_mapping_err;
       struct evcnt doorbells;
       struct evcnt missing_tx_comp;
       struct evcnt bad_req_id;
       struct evcnt collapse;
       struct evcnt collapse_err;
       struct evcnt pcq_drops;
};

struct ena_stats_rx {
       char name[16];
       struct evcnt cnt;
       struct evcnt bytes;
       struct evcnt refil_partial;
       struct evcnt bad_csum;
       struct evcnt mbuf_alloc_fail;
       struct evcnt dma_mapping_err;
       struct evcnt bad_desc_num;
       struct evcnt bad_req_id;
       struct evcnt empty_rx_ring;
};

/*
* Locking notes:
* + For TX, a field in ena_ring is protected by ring_mtx (a spin mutex).
*   - protect them only when I/F is up.
*   - when I/F is down or attaching, detaching, no need to protect them.
* + For RX, a field "stopping" is protected by ring_mtx (a spin mutex).
*   - other fields in ena_ring are not protected.
* + a fields in ena_adapter is protected by global_mtx (a adaptive mutex).
*
* + a field marked "stable" is unlocked.
* + a field marked "atomic" is unlocked,
*   but must use atomic ops to read/write.
*
* Lock order:
* + global_mtx -> ring_mtx
*/
struct ena_ring {
       /* Holds the empty requests for TX/RX out of order completions */
       union {
               uint16_t *free_tx_ids;
               uint16_t *free_rx_ids;
       };
       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;

       uint16_t qid;

       /* Determines if device will use LLQ or normal mode for TX */
       enum ena_admin_placement_policy_type tx_mem_queue_type;
       /* The maximum length the driver can push to the device (For LLQ) */
       uint8_t tx_max_header_size;

       struct ena_com_rx_buf_info ena_bufs[ENA_PKT_MAX_BUFS];

       /*
        * Fields used for Adaptive Interrupt Modulation - to be implemented in
        * the future releases
        */
       uint32_t  smoothed_interval;
       enum ena_intr_moder_level moder_tbl_idx;

       struct ena_que *que;
#ifdef LRO
       struct lro_ctrl lro;
#endif

       uint16_t next_to_use;
       uint16_t next_to_clean;

       union {
               struct ena_tx_buffer *tx_buffer_info; /* context of tx packet */
               struct ena_rx_buffer *rx_buffer_info; /* context of rx packet */
       };
       int ring_size; /* number of tx/rx_buffer_info's entries */

       pcq_t *br; /* only for TX */

       kmutex_t ring_mtx;
       char mtx_name[16];

       union {
               struct {
                       struct work enqueue_task;
                       struct workqueue *enqueue_tq;
               };
               struct {
                       struct work cleanup_task;
                       struct workqueue *cleanup_tq;
               };
       };
       u_int task_pending; /* atomic */
       bool stopping;

       union {
               struct ena_stats_tx tx_stats;
               struct ena_stats_rx rx_stats;
       };

       int empty_rx_queue;
} __aligned(CACHE_LINE_SIZE);

struct ena_stats_dev {
       char name[16];
       struct evcnt wd_expired;
       struct evcnt interface_up;
       struct evcnt interface_down;
       struct evcnt admin_q_pause;
};

struct ena_hw_stats {
       char name[16];
       struct evcnt rx_packets;
       struct evcnt tx_packets;

       struct evcnt rx_bytes;
       struct evcnt tx_bytes;

       struct evcnt rx_drops;
};

/* Board specific private data structure */
struct ena_adapter {
       struct ena_com_dev *ena_dev;

       /* OS defined structs */
       device_t pdev;
       struct ethercom sc_ec;
       struct ifnet *ifp;              /* set to point to sc_ec */
       struct ifmedia  media;

       /* OS resources */
       kmutex_t global_mtx;

       void *sc_ihs[ENA_MAX_MSIX_VEC(ENA_MAX_NUM_IO_QUEUES)];
       pci_intr_handle_t *sc_intrs;
       int sc_nintrs;
       struct pci_attach_args sc_pa;

       /* Registers */
       bus_space_handle_t sc_bhandle;
       bus_space_tag_t sc_btag;
       bus_addr_t sc_memaddr;
       bus_size_t sc_mapsize;

       /* DMA tag used throughout the driver adapter for Tx and Rx */
       bus_dma_tag_t sc_dmat;
       int dma_width;

       uint32_t max_mtu;

       uint16_t max_tx_sgl_size;
       uint16_t max_rx_sgl_size;

       uint32_t tx_offload_cap;

       /* Tx fast path data */
       int num_queues;

       unsigned int tx_ring_size;
       unsigned int rx_ring_size;

       /* RSS*/
       uint8_t rss_ind_tbl[ENA_RX_RSS_TABLE_SIZE];
       bool rss_support;
       int initialized;

       uint8_t mac_addr[ETHER_ADDR_LEN];
       /* mdio and phy*/

       uint32_t flags; /* atomic */

       /* Queue will represent one TX and one RX ring */
       struct ena_que que[ENA_MAX_NUM_IO_QUEUES]
           __aligned(CACHE_LINE_SIZE); /* stable */

       /* TX */
       struct ena_ring tx_ring[ENA_MAX_NUM_IO_QUEUES]
           __aligned(CACHE_LINE_SIZE);

       /* RX */
       struct ena_ring rx_ring[ENA_MAX_NUM_IO_QUEUES]
           __aligned(CACHE_LINE_SIZE);

       /* Timer service */
       struct callout timer_service;
       sbintime_t keep_alive_timestamp;
       uint32_t next_monitored_tx_qid;
       struct work reset_task;
       struct workqueue *reset_tq;
       int wd_active;
       sbintime_t keep_alive_timeout;
       sbintime_t missing_tx_timeout;
       uint32_t missing_tx_max_queues;
       uint32_t missing_tx_threshold;

       /* Statistics */
       struct ena_stats_dev dev_stats;
       struct ena_hw_stats hw_stats;

       enum ena_regs_reset_reason_types reset_reason;
};

#define ENA_RING_MTX_LOCK(_ring)        mutex_enter(&(_ring)->ring_mtx)
#define ENA_RING_MTX_TRYLOCK(_ring)     mutex_tryenter(&(_ring)->ring_mtx)
#define ENA_RING_MTX_UNLOCK(_ring)      mutex_exit(&(_ring)->ring_mtx)
#define ENA_RING_MTX_OWNED(_ring)       mutex_owned(&(_ring)->ring_mtx)

#define ENA_CORE_MTX_LOCK(_adapter)             mutex_enter(&(_adapter)->global_mtx)
#define ENA_CORE_MTX_TRYLOCK(_adapter)  mutex_tryenter(&(_adapter)->global_mtx)
#define ENA_CORE_MTX_UNLOCK(_adapter)   mutex_exit(&(_adapter)->global_mtx)
#define ENA_CORE_MTX_OWNED(_adapter)    mutex_owned(&(_adapter)->global_mtx)

static inline int ena_mbuf_count(struct mbuf *mbuf)
{
       int count = 1;

       while ((mbuf = mbuf->m_next) != NULL)
               ++count;

       return count;
}

/* provide FreeBSD-compatible macros */
#define if_getcapenable(ifp)            (ifp)->if_capenable
#define if_setcapenable(ifp, s)         SET((ifp)->if_capenable, s)
#define if_getcapabilities(ifp)         (ifp)->if_capabilities
#define if_setcapabilities(ifp, s)      SET((ifp)->if_capabilities, s)
#define if_setcapabilitiesbit(ifp, s, c) do {   \
               CLR((ifp)->if_capabilities, c); \
               SET((ifp)->if_capabilities, s); \
       } while (0)
#define if_getsoftc(ifp)                (ifp)->if_softc
#define if_setmtu(ifp, new_mtu)         (ifp)->if_mtu = (new_mtu)
#define if_getdrvflags(ifp)             (ifp)->if_flags
#define if_setdrvflagbits(ifp, s, c)    do {    \
               CLR((ifp)->if_flags, c);        \
               SET((ifp)->if_flags, s);        \
       } while (0)
#define if_setflags(ifp, s)             SET((ifp)->if_flags, s)
#define if_sethwassistbits(ifp, s, c)   do {            \
               CLR((ifp)->if_csum_flags_rx, c);        \
               SET((ifp)->if_csum_flags_rx, s);        \
       } while (0)
#define if_clearhwassist(ifp)           (ifp)->if_csum_flags_rx = 0
#define if_setbaudrate(ifp, r)          (ifp)->if_baudrate = (r)
#define if_setdev(ifp, dev)             do { } while (0)
#define if_setsoftc(ifp, softc)         (ifp)->if_softc = (softc)
#define if_setinitfn(ifp, initfn)       (ifp)->if_init = (initfn)
#define if_settransmitfn(ifp, txfn)     (ifp)->if_transmit = (txfn)
#define if_setioctlfn(ifp, ioctlfn)     (ifp)->if_ioctl = (ioctlfn)
#define if_setsendqlen(ifp, sqlen)      \
       IFQ_SET_MAXLEN(&(ifp)->if_snd, uimax(sqlen, IFQ_MAXLEN))
#define if_setsendqready(ifp)           IFQ_SET_READY(&(ifp)->if_snd)
#define if_setifheaderlen(ifp, len)     (ifp)->if_hdrlen = (len)

#define SBT_1S  ((sbintime_t)1 << 32)
#define bintime_clear(a)        ((a)->sec = (a)->frac = 0)
#define bintime_isset(a)        ((a)->sec || (a)->frac)

static __inline sbintime_t
bttosbt(const struct bintime _bt)
{
       return (((sbintime_t)_bt.sec << 32) + (_bt.frac >> 32));
}

static __inline sbintime_t
getsbinuptime(void)
{
       struct bintime _bt;

       getbinuptime(&_bt);
       return (bttosbt(_bt));
}

/* Intentionally non-atomic, it's just unnecessary overhead */
#define counter_u64_add(x, cnt)                 (x).ev_count += (cnt)
#define counter_u64_zero(x)                     (x).ev_count = 0
#define counter_u64_free(x)                     evcnt_detach(&(x))

#define counter_u64_add_protected(x, cnt)       (x).ev_count += (cnt)
#define counter_enter()                         do {} while (0)
#define counter_exit()                          do {} while (0)

/* Misc other constants */
#define mp_ncpus                        ncpu
#define osreldate                       __NetBSD_Version__

/*
* XXX XXX XXX just to make compile, must provide replacement XXX XXX XXX
* Other than that, TODO:
* - decide whether to import <sys/buf_ring.h>
* - recheck the M_CSUM/IPCAP mapping
* - recheck workqueue use - FreeBSD taskqueues might have different semantics
*/
#define buf_ring_alloc(a, b, c, d)      (void *)&a
#define drbr_free(ifp, b)               do { } while (0)
#define drbr_flush(ifp, b)              IFQ_PURGE(&(ifp)->if_snd)
#define drbr_advance(ifp, b)                                    \
       ({                                                      \
               struct mbuf *__m;                               \
               IFQ_DEQUEUE(&(ifp)->if_snd, __m);               \
               __m;                                            \
       })
#define drbr_putback(ifp, b, m)         do { } while (0)
#define drbr_empty(ifp, b)              IFQ_IS_EMPTY(&(ifp)->if_snd)
#define drbr_peek(ifp, b)                                       \
       ({                                                      \
               struct mbuf *__m;                               \
               IFQ_POLL(&(ifp)->if_snd, __m);                  \
               __m;                                            \
       })
#define drbr_enqueue(ifp, b, m)                                 \
       ({                                                      \
               int __err;                                      \
               IFQ_ENQUEUE(&(ifp)->if_snd, m, __err);          \
               __err;                                          \
       })
#define m_getjcl(a, b, c, d)            NULL
#define MJUM16BYTES                     MCLBYTES
#define m_append(m, len, cp)            ena_m_append(m, len, cp)
#define m_collapse(m, how, maxfrags)    m_defrag(m, how)        /* XXX */
/* XXX XXX XXX */

static inline int
ena_m_append(struct mbuf *m0, int len, const void *cpv)
{
       struct mbuf *m, *n;
       int remainder, space;
       const char *cp = cpv;

       KASSERT(len != M_COPYALL);
       for (m = m0; m->m_next != NULL; m = m->m_next)
               continue;
       remainder = len;
       space = M_TRAILINGSPACE(m);
       if (space > 0) {
               /*
                * Copy into available space.
                */
               if (space > remainder)
                       space = remainder;
               memmove(mtod(m, char *) + m->m_len, cp, space);
               m->m_len += space;
               cp = cp + space, remainder -= space;
       }
       while (remainder > 0) {
               /*
                * Allocate a new mbuf; could check space
                * and allocate a cluster instead.
                */
               n = m_get(M_DONTWAIT, m->m_type);
               if (n == NULL)
                       break;
               n->m_len = uimin(MLEN, remainder);
               memmove(mtod(n, void *), cp, n->m_len);
               cp += n->m_len, remainder -= n->m_len;
               m->m_next = n;
               m = n;
       }
       if (m0->m_flags & M_PKTHDR)
               m0->m_pkthdr.len += len - remainder;
       return (remainder == 0);
}
#endif /* !(ENA_H) */