sfc: Move shared NIC code from falcon.c to new source file nic.c
authorBen Hutchings <bhutchings@solarflare.com>
Sun, 29 Nov 2009 15:14:45 +0000 (15:14 +0000)
committerDavid S. Miller <davem@davemloft.net>
Mon, 30 Nov 2009 01:23:55 +0000 (17:23 -0800)
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/sfc/Makefile
drivers/net/sfc/falcon.c
drivers/net/sfc/nic.c [new file with mode: 0644]

index 7b52fe1..223106b 100644 (file)
@@ -1,4 +1,4 @@
-sfc-y                  += efx.o falcon.o tx.o rx.o falcon_gmac.o \
+sfc-y                  += efx.o nic.o falcon.o tx.o rx.o falcon_gmac.o \
                           falcon_xmac.o selftest.o ethtool.o qt202x_phy.o \
                           mdio_10g.o tenxpress.o falcon_boards.o
 sfc-$(CONFIG_SFC_MTD)  += mtd.o
index 64b47da..48d28d8 100644 (file)
 
 /* Hardware control for SFC4000 (aka Falcon). */
 
-/**************************************************************************
- *
- * Configurable values
- *
- **************************************************************************
- */
-
-/* This is set to 16 for a good reason.  In summary, if larger than
- * 16, the descriptor cache holds more than a default socket
- * buffer's worth of packets (for UDP we can only have at most one
- * socket buffer's worth outstanding).  This combined with the fact
- * that we only get 1 TX event per descriptor cache means the NIC
- * goes idle.
- */
-#define TX_DC_ENTRIES 16
-#define TX_DC_ENTRIES_ORDER 1
-
-#define RX_DC_ENTRIES 64
-#define RX_DC_ENTRIES_ORDER 3
-
 static const unsigned int
 /* "Large" EEPROM device: Atmel AT25640 or similar
  * 8 KB, 16-bit address, 32 B write block */
@@ -63,87 +43,6 @@ default_flash_type = ((17 << SPI_DEV_TYPE_SIZE_LBN)
                      | (15 << SPI_DEV_TYPE_ERASE_SIZE_LBN)
                      | (8 << SPI_DEV_TYPE_BLOCK_SIZE_LBN));
 
-/* RX FIFO XOFF watermark
- *
- * When the amount of the RX FIFO increases used increases past this
- * watermark send XOFF. Only used if RX flow control is enabled (ethtool -A)
- * This also has an effect on RX/TX arbitration
- */
-int efx_nic_rx_xoff_thresh = -1;
-module_param_named(rx_xoff_thresh_bytes, efx_nic_rx_xoff_thresh, int, 0644);
-MODULE_PARM_DESC(rx_xoff_thresh_bytes, "RX fifo XOFF threshold");
-
-/* RX FIFO XON watermark
- *
- * When the amount of the RX FIFO used decreases below this
- * watermark send XON. Only used if TX flow control is enabled (ethtool -A)
- * This also has an effect on RX/TX arbitration
- */
-int efx_nic_rx_xon_thresh = -1;
-module_param_named(rx_xon_thresh_bytes, efx_nic_rx_xon_thresh, int, 0644);
-MODULE_PARM_DESC(rx_xon_thresh_bytes, "RX fifo XON threshold");
-
-/* If EFX_MAX_INT_ERRORS internal errors occur within
- * EFX_INT_ERROR_EXPIRE seconds, we consider the NIC broken and
- * disable it.
- */
-#define EFX_INT_ERROR_EXPIRE 3600
-#define EFX_MAX_INT_ERRORS 5
-
-/* We poll for events every FLUSH_INTERVAL ms, and check FLUSH_POLL_COUNT times
- */
-#define EFX_FLUSH_INTERVAL 10
-#define EFX_FLUSH_POLL_COUNT 100
-
-/**************************************************************************
- *
- * Falcon constants
- *
- **************************************************************************
- */
-
-/* Size and alignment of special buffers (4KB) */
-#define EFX_BUF_SIZE 4096
-
-/* Depth of RX flush request fifo */
-#define EFX_RX_FLUSH_COUNT 4
-
-/**************************************************************************
- *
- * Solarstorm hardware access
- *
- **************************************************************************/
-
-static inline void efx_write_buf_tbl(struct efx_nic *efx, efx_qword_t *value,
-                                    unsigned int index)
-{
-       efx_sram_writeq(efx, efx->membase + efx->type->buf_tbl_base,
-                       value, index);
-}
-
-/* Read the current event from the event queue */
-static inline efx_qword_t *efx_event(struct efx_channel *channel,
-                                    unsigned int index)
-{
-       return (((efx_qword_t *) (channel->eventq.addr)) + index);
-}
-
-/* See if an event is present
- *
- * We check both the high and low dword of the event for all ones.  We
- * wrote all ones when we cleared the event, and no valid event can
- * have all ones in either its high or low dwords.  This approach is
- * robust against reordering.
- *
- * Note that using a single 64-bit comparison is incorrect; even
- * though the CPU read will be atomic, the DMA write may not be.
- */
-static inline int efx_event_present(efx_qword_t *event)
-{
-       return (!(EFX_DWORD_IS_ALL_ONES(event->dword[0]) |
-                 EFX_DWORD_IS_ALL_ONES(event->dword[1])));
-}
-
 /**************************************************************************
  *
  * I2C bus - this is a bit-bashing interface using GPIO pins
@@ -174,867 +73,32 @@ static void falcon_setscl(void *data, int state)
 
 static int falcon_getsda(void *data)
 {
-       struct efx_nic *efx = (struct efx_nic *)data;
-       efx_oword_t reg;
-
-       efx_reado(efx, &reg, FR_AB_GPIO_CTL);
-       return EFX_OWORD_FIELD(reg, FRF_AB_GPIO3_IN);
-}
-
-static int falcon_getscl(void *data)
-{
-       struct efx_nic *efx = (struct efx_nic *)data;
-       efx_oword_t reg;
-
-       efx_reado(efx, &reg, FR_AB_GPIO_CTL);
-       return EFX_OWORD_FIELD(reg, FRF_AB_GPIO0_IN);
-}
-
-static struct i2c_algo_bit_data falcon_i2c_bit_operations = {
-       .setsda         = falcon_setsda,
-       .setscl         = falcon_setscl,
-       .getsda         = falcon_getsda,
-       .getscl         = falcon_getscl,
-       .udelay         = 5,
-       /* Wait up to 50 ms for slave to let us pull SCL high */
-       .timeout        = DIV_ROUND_UP(HZ, 20),
-};
-
-/**************************************************************************
- *
- * Special buffer handling
- * Special buffers are used for event queues and the TX and RX
- * descriptor rings.
- *
- *************************************************************************/
-
-/*
- * Initialise a special buffer
- *
- * This will define a buffer (previously allocated via
- * efx_alloc_special_buffer()) in the buffer table, allowing
- * it to be used for event queues, descriptor rings etc.
- */
-static void
-efx_init_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
-{
-       efx_qword_t buf_desc;
-       int index;
-       dma_addr_t dma_addr;
-       int i;
-
-       EFX_BUG_ON_PARANOID(!buffer->addr);
-
-       /* Write buffer descriptors to NIC */
-       for (i = 0; i < buffer->entries; i++) {
-               index = buffer->index + i;
-               dma_addr = buffer->dma_addr + (i * 4096);
-               EFX_LOG(efx, "mapping special buffer %d at %llx\n",
-                       index, (unsigned long long)dma_addr);
-               EFX_POPULATE_QWORD_3(buf_desc,
-                                    FRF_AZ_BUF_ADR_REGION, 0,
-                                    FRF_AZ_BUF_ADR_FBUF, dma_addr >> 12,
-                                    FRF_AZ_BUF_OWNER_ID_FBUF, 0);
-               efx_write_buf_tbl(efx, &buf_desc, index);
-       }
-}
-
-/* Unmaps a buffer and clears the buffer table entries */
-static void
-efx_fini_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
-{
-       efx_oword_t buf_tbl_upd;
-       unsigned int start = buffer->index;
-       unsigned int end = (buffer->index + buffer->entries - 1);
-
-       if (!buffer->entries)
-               return;
-
-       EFX_LOG(efx, "unmapping special buffers %d-%d\n",
-               buffer->index, buffer->index + buffer->entries - 1);
-
-       EFX_POPULATE_OWORD_4(buf_tbl_upd,
-                            FRF_AZ_BUF_UPD_CMD, 0,
-                            FRF_AZ_BUF_CLR_CMD, 1,
-                            FRF_AZ_BUF_CLR_END_ID, end,
-                            FRF_AZ_BUF_CLR_START_ID, start);
-       efx_writeo(efx, &buf_tbl_upd, FR_AZ_BUF_TBL_UPD);
-}
-
-/*
- * Allocate a new special buffer
- *
- * This allocates memory for a new buffer, clears it and allocates a
- * new buffer ID range.  It does not write into the buffer table.
- *
- * This call will allocate 4KB buffers, since 8KB buffers can't be
- * used for event queues and descriptor rings.
- */
-static int efx_alloc_special_buffer(struct efx_nic *efx,
-                                   struct efx_special_buffer *buffer,
-                                   unsigned int len)
-{
-       len = ALIGN(len, EFX_BUF_SIZE);
-
-       buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
-                                           &buffer->dma_addr);
-       if (!buffer->addr)
-               return -ENOMEM;
-       buffer->len = len;
-       buffer->entries = len / EFX_BUF_SIZE;
-       BUG_ON(buffer->dma_addr & (EFX_BUF_SIZE - 1));
-
-       /* All zeros is a potentially valid event so memset to 0xff */
-       memset(buffer->addr, 0xff, len);
-
-       /* Select new buffer ID */
-       buffer->index = efx->next_buffer_table;
-       efx->next_buffer_table += buffer->entries;
-
-       EFX_LOG(efx, "allocating special buffers %d-%d at %llx+%x "
-               "(virt %p phys %llx)\n", buffer->index,
-               buffer->index + buffer->entries - 1,
-               (u64)buffer->dma_addr, len,
-               buffer->addr, (u64)virt_to_phys(buffer->addr));
-
-       return 0;
-}
-
-static void
-efx_free_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
-{
-       if (!buffer->addr)
-               return;
-
-       EFX_LOG(efx, "deallocating special buffers %d-%d at %llx+%x "
-               "(virt %p phys %llx)\n", buffer->index,
-               buffer->index + buffer->entries - 1,
-               (u64)buffer->dma_addr, buffer->len,
-               buffer->addr, (u64)virt_to_phys(buffer->addr));
-
-       pci_free_consistent(efx->pci_dev, buffer->len, buffer->addr,
-                           buffer->dma_addr);
-       buffer->addr = NULL;
-       buffer->entries = 0;
-}
-
-/**************************************************************************
- *
- * Generic buffer handling
- * These buffers are used for interrupt status and MAC stats
- *
- **************************************************************************/
-
-int efx_nic_alloc_buffer(struct efx_nic *efx, struct efx_buffer *buffer,
-                        unsigned int len)
-{
-       buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
-                                           &buffer->dma_addr);
-       if (!buffer->addr)
-               return -ENOMEM;
-       buffer->len = len;
-       memset(buffer->addr, 0, len);
-       return 0;
-}
-
-void efx_nic_free_buffer(struct efx_nic *efx, struct efx_buffer *buffer)
-{
-       if (buffer->addr) {
-               pci_free_consistent(efx->pci_dev, buffer->len,
-                                   buffer->addr, buffer->dma_addr);
-               buffer->addr = NULL;
-       }
-}
-
-/**************************************************************************
- *
- * TX path
- *
- **************************************************************************/
-
-/* Returns a pointer to the specified transmit descriptor in the TX
- * descriptor queue belonging to the specified channel.
- */
-static inline efx_qword_t *
-efx_tx_desc(struct efx_tx_queue *tx_queue, unsigned int index)
-{
-       return (((efx_qword_t *) (tx_queue->txd.addr)) + index);
-}
-
-/* This writes to the TX_DESC_WPTR; write pointer for TX descriptor ring */
-static inline void efx_notify_tx_desc(struct efx_tx_queue *tx_queue)
-{
-       unsigned write_ptr;
-       efx_dword_t reg;
-
-       write_ptr = tx_queue->write_count & EFX_TXQ_MASK;
-       EFX_POPULATE_DWORD_1(reg, FRF_AZ_TX_DESC_WPTR_DWORD, write_ptr);
-       efx_writed_page(tx_queue->efx, &reg,
-                       FR_AZ_TX_DESC_UPD_DWORD_P0, tx_queue->queue);
-}
-
-
-/* For each entry inserted into the software descriptor ring, create a
- * descriptor in the hardware TX descriptor ring (in host memory), and
- * write a doorbell.
- */
-void efx_nic_push_buffers(struct efx_tx_queue *tx_queue)
-{
-
-       struct efx_tx_buffer *buffer;
-       efx_qword_t *txd;
-       unsigned write_ptr;
-
-       BUG_ON(tx_queue->write_count == tx_queue->insert_count);
-
-       do {
-               write_ptr = tx_queue->write_count & EFX_TXQ_MASK;
-               buffer = &tx_queue->buffer[write_ptr];
-               txd = efx_tx_desc(tx_queue, write_ptr);
-               ++tx_queue->write_count;
-
-               /* Create TX descriptor ring entry */
-               EFX_POPULATE_QWORD_4(*txd,
-                                    FSF_AZ_TX_KER_CONT, buffer->continuation,
-                                    FSF_AZ_TX_KER_BYTE_COUNT, buffer->len,
-                                    FSF_AZ_TX_KER_BUF_REGION, 0,
-                                    FSF_AZ_TX_KER_BUF_ADDR, buffer->dma_addr);
-       } while (tx_queue->write_count != tx_queue->insert_count);
-
-       wmb(); /* Ensure descriptors are written before they are fetched */
-       efx_notify_tx_desc(tx_queue);
-}
-
-/* Allocate hardware resources for a TX queue */
-int efx_nic_probe_tx(struct efx_tx_queue *tx_queue)
-{
-       struct efx_nic *efx = tx_queue->efx;
-       BUILD_BUG_ON(EFX_TXQ_SIZE < 512 || EFX_TXQ_SIZE > 4096 ||
-                    EFX_TXQ_SIZE & EFX_TXQ_MASK);
-       return efx_alloc_special_buffer(efx, &tx_queue->txd,
-                                       EFX_TXQ_SIZE * sizeof(efx_qword_t));
-}
-
-void efx_nic_init_tx(struct efx_tx_queue *tx_queue)
-{
-       efx_oword_t tx_desc_ptr;
-       struct efx_nic *efx = tx_queue->efx;
-
-       tx_queue->flushed = FLUSH_NONE;
-
-       /* Pin TX descriptor ring */
-       efx_init_special_buffer(efx, &tx_queue->txd);
-
-       /* Push TX descriptor ring to card */
-       EFX_POPULATE_OWORD_10(tx_desc_ptr,
-                             FRF_AZ_TX_DESCQ_EN, 1,
-                             FRF_AZ_TX_ISCSI_DDIG_EN, 0,
-                             FRF_AZ_TX_ISCSI_HDIG_EN, 0,
-                             FRF_AZ_TX_DESCQ_BUF_BASE_ID, tx_queue->txd.index,
-                             FRF_AZ_TX_DESCQ_EVQ_ID,
-                             tx_queue->channel->channel,
-                             FRF_AZ_TX_DESCQ_OWNER_ID, 0,
-                             FRF_AZ_TX_DESCQ_LABEL, tx_queue->queue,
-                             FRF_AZ_TX_DESCQ_SIZE,
-                             __ffs(tx_queue->txd.entries),
-                             FRF_AZ_TX_DESCQ_TYPE, 0,
-                             FRF_BZ_TX_NON_IP_DROP_DIS, 1);
-
-       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0) {
-               int csum = tx_queue->queue == EFX_TX_QUEUE_OFFLOAD_CSUM;
-               EFX_SET_OWORD_FIELD(tx_desc_ptr, FRF_BZ_TX_IP_CHKSM_DIS, !csum);
-               EFX_SET_OWORD_FIELD(tx_desc_ptr, FRF_BZ_TX_TCP_CHKSM_DIS,
-                                   !csum);
-       }
-
-       efx_writeo_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
-                        tx_queue->queue);
-
-       if (efx_nic_rev(efx) < EFX_REV_FALCON_B0) {
-               efx_oword_t reg;
-
-               /* Only 128 bits in this register */
-               BUILD_BUG_ON(EFX_TX_QUEUE_COUNT >= 128);
-
-               efx_reado(efx, &reg, FR_AA_TX_CHKSM_CFG);
-               if (tx_queue->queue == EFX_TX_QUEUE_OFFLOAD_CSUM)
-                       clear_bit_le(tx_queue->queue, (void *)&reg);
-               else
-                       set_bit_le(tx_queue->queue, (void *)&reg);
-               efx_writeo(efx, &reg, FR_AA_TX_CHKSM_CFG);
-       }
-}
-
-static void efx_flush_tx_queue(struct efx_tx_queue *tx_queue)
-{
-       struct efx_nic *efx = tx_queue->efx;
-       efx_oword_t tx_flush_descq;
-
-       tx_queue->flushed = FLUSH_PENDING;
-
-       /* Post a flush command */
-       EFX_POPULATE_OWORD_2(tx_flush_descq,
-                            FRF_AZ_TX_FLUSH_DESCQ_CMD, 1,
-                            FRF_AZ_TX_FLUSH_DESCQ, tx_queue->queue);
-       efx_writeo(efx, &tx_flush_descq, FR_AZ_TX_FLUSH_DESCQ);
-}
-
-void efx_nic_fini_tx(struct efx_tx_queue *tx_queue)
-{
-       struct efx_nic *efx = tx_queue->efx;
-       efx_oword_t tx_desc_ptr;
-
-       /* The queue should have been flushed */
-       WARN_ON(tx_queue->flushed != FLUSH_DONE);
-
-       /* Remove TX descriptor ring from card */
-       EFX_ZERO_OWORD(tx_desc_ptr);
-       efx_writeo_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
-                        tx_queue->queue);
-
-       /* Unpin TX descriptor ring */
-       efx_fini_special_buffer(efx, &tx_queue->txd);
-}
-
-/* Free buffers backing TX queue */
-void efx_nic_remove_tx(struct efx_tx_queue *tx_queue)
-{
-       efx_free_special_buffer(tx_queue->efx, &tx_queue->txd);
-}
-
-/**************************************************************************
- *
- * RX path
- *
- **************************************************************************/
-
-/* Returns a pointer to the specified descriptor in the RX descriptor queue */
-static inline efx_qword_t *
-efx_rx_desc(struct efx_rx_queue *rx_queue, unsigned int index)
-{
-       return (((efx_qword_t *) (rx_queue->rxd.addr)) + index);
-}
-
-/* This creates an entry in the RX descriptor queue */
-static inline void
-efx_build_rx_desc(struct efx_rx_queue *rx_queue, unsigned index)
-{
-       struct efx_rx_buffer *rx_buf;
-       efx_qword_t *rxd;
-
-       rxd = efx_rx_desc(rx_queue, index);
-       rx_buf = efx_rx_buffer(rx_queue, index);
-       EFX_POPULATE_QWORD_3(*rxd,
-                            FSF_AZ_RX_KER_BUF_SIZE,
-                            rx_buf->len -
-                            rx_queue->efx->type->rx_buffer_padding,
-                            FSF_AZ_RX_KER_BUF_REGION, 0,
-                            FSF_AZ_RX_KER_BUF_ADDR, rx_buf->dma_addr);
-}
-
-/* This writes to the RX_DESC_WPTR register for the specified receive
- * descriptor ring.
- */
-void efx_nic_notify_rx_desc(struct efx_rx_queue *rx_queue)
-{
-       efx_dword_t reg;
-       unsigned write_ptr;
-
-       while (rx_queue->notified_count != rx_queue->added_count) {
-               efx_build_rx_desc(rx_queue,
-                                 rx_queue->notified_count &
-                                 EFX_RXQ_MASK);
-               ++rx_queue->notified_count;
-       }
-
-       wmb();
-       write_ptr = rx_queue->added_count & EFX_RXQ_MASK;
-       EFX_POPULATE_DWORD_1(reg, FRF_AZ_RX_DESC_WPTR_DWORD, write_ptr);
-       efx_writed_page(rx_queue->efx, &reg,
-                       FR_AZ_RX_DESC_UPD_DWORD_P0, rx_queue->queue);
-}
-
-int efx_nic_probe_rx(struct efx_rx_queue *rx_queue)
-{
-       struct efx_nic *efx = rx_queue->efx;
-       BUILD_BUG_ON(EFX_RXQ_SIZE < 512 || EFX_RXQ_SIZE > 4096 ||
-                    EFX_RXQ_SIZE & EFX_RXQ_MASK);
-       return efx_alloc_special_buffer(efx, &rx_queue->rxd,
-                                       EFX_RXQ_SIZE * sizeof(efx_qword_t));
-}
-
-void efx_nic_init_rx(struct efx_rx_queue *rx_queue)
-{
-       efx_oword_t rx_desc_ptr;
-       struct efx_nic *efx = rx_queue->efx;
-       bool is_b0 = efx_nic_rev(efx) >= EFX_REV_FALCON_B0;
-       bool iscsi_digest_en = is_b0;
-
-       EFX_LOG(efx, "RX queue %d ring in special buffers %d-%d\n",
-               rx_queue->queue, rx_queue->rxd.index,
-               rx_queue->rxd.index + rx_queue->rxd.entries - 1);
-
-       rx_queue->flushed = FLUSH_NONE;
-
-       /* Pin RX descriptor ring */
-       efx_init_special_buffer(efx, &rx_queue->rxd);
-
-       /* Push RX descriptor ring to card */
-       EFX_POPULATE_OWORD_10(rx_desc_ptr,
-                             FRF_AZ_RX_ISCSI_DDIG_EN, iscsi_digest_en,
-                             FRF_AZ_RX_ISCSI_HDIG_EN, iscsi_digest_en,
-                             FRF_AZ_RX_DESCQ_BUF_BASE_ID, rx_queue->rxd.index,
-                             FRF_AZ_RX_DESCQ_EVQ_ID,
-                             rx_queue->channel->channel,
-                             FRF_AZ_RX_DESCQ_OWNER_ID, 0,
-                             FRF_AZ_RX_DESCQ_LABEL, rx_queue->queue,
-                             FRF_AZ_RX_DESCQ_SIZE,
-                             __ffs(rx_queue->rxd.entries),
-                             FRF_AZ_RX_DESCQ_TYPE, 0 /* kernel queue */ ,
-                             /* For >=B0 this is scatter so disable */
-                             FRF_AZ_RX_DESCQ_JUMBO, !is_b0,
-                             FRF_AZ_RX_DESCQ_EN, 1);
-       efx_writeo_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
-                        rx_queue->queue);
-}
-
-static void efx_flush_rx_queue(struct efx_rx_queue *rx_queue)
-{
-       struct efx_nic *efx = rx_queue->efx;
-       efx_oword_t rx_flush_descq;
-
-       rx_queue->flushed = FLUSH_PENDING;
-
-       /* Post a flush command */
-       EFX_POPULATE_OWORD_2(rx_flush_descq,
-                            FRF_AZ_RX_FLUSH_DESCQ_CMD, 1,
-                            FRF_AZ_RX_FLUSH_DESCQ, rx_queue->queue);
-       efx_writeo(efx, &rx_flush_descq, FR_AZ_RX_FLUSH_DESCQ);
-}
-
-void efx_nic_fini_rx(struct efx_rx_queue *rx_queue)
-{
-       efx_oword_t rx_desc_ptr;
-       struct efx_nic *efx = rx_queue->efx;
-
-       /* The queue should already have been flushed */
-       WARN_ON(rx_queue->flushed != FLUSH_DONE);
-
-       /* Remove RX descriptor ring from card */
-       EFX_ZERO_OWORD(rx_desc_ptr);
-       efx_writeo_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
-                        rx_queue->queue);
-
-       /* Unpin RX descriptor ring */
-       efx_fini_special_buffer(efx, &rx_queue->rxd);
-}
-
-/* Free buffers backing RX queue */
-void efx_nic_remove_rx(struct efx_rx_queue *rx_queue)
-{
-       efx_free_special_buffer(rx_queue->efx, &rx_queue->rxd);
-}
-
-/**************************************************************************
- *
- * Event queue processing
- * Event queues are processed by per-channel tasklets.
- *
- **************************************************************************/
-
-/* Update a channel's event queue's read pointer (RPTR) register
- *
- * This writes the EVQ_RPTR_REG register for the specified channel's
- * event queue.
- *
- * Note that EVQ_RPTR_REG contains the index of the "last read" event,
- * whereas channel->eventq_read_ptr contains the index of the "next to
- * read" event.
- */
-void efx_nic_eventq_read_ack(struct efx_channel *channel)
-{
-       efx_dword_t reg;
-       struct efx_nic *efx = channel->efx;
-
-       EFX_POPULATE_DWORD_1(reg, FRF_AZ_EVQ_RPTR, channel->eventq_read_ptr);
-       efx_writed_table(efx, &reg, efx->type->evq_rptr_tbl_base,
-                           channel->channel);
-}
-
-/* Use HW to insert a SW defined event */
-void efx_generate_event(struct efx_channel *channel, efx_qword_t *event)
-{
-       efx_oword_t drv_ev_reg;
-
-       BUILD_BUG_ON(FRF_AZ_DRV_EV_DATA_LBN != 0 ||
-                    FRF_AZ_DRV_EV_DATA_WIDTH != 64);
-       drv_ev_reg.u32[0] = event->u32[0];
-       drv_ev_reg.u32[1] = event->u32[1];
-       drv_ev_reg.u32[2] = 0;
-       drv_ev_reg.u32[3] = 0;
-       EFX_SET_OWORD_FIELD(drv_ev_reg, FRF_AZ_DRV_EV_QID, channel->channel);
-       efx_writeo(channel->efx, &drv_ev_reg, FR_AZ_DRV_EV);
-}
-
-/* Handle a transmit completion event
- *
- * The NIC batches TX completion events; the message we receive is of
- * the form "complete all TX events up to this index".
- */
-static void
-efx_handle_tx_event(struct efx_channel *channel, efx_qword_t *event)
-{
-       unsigned int tx_ev_desc_ptr;
-       unsigned int tx_ev_q_label;
-       struct efx_tx_queue *tx_queue;
-       struct efx_nic *efx = channel->efx;
-
-       if (likely(EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_COMP))) {
-               /* Transmit completion */
-               tx_ev_desc_ptr = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_DESC_PTR);
-               tx_ev_q_label = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_Q_LABEL);
-               tx_queue = &efx->tx_queue[tx_ev_q_label];
-               channel->irq_mod_score +=
-                       (tx_ev_desc_ptr - tx_queue->read_count) &
-                       EFX_TXQ_MASK;
-               efx_xmit_done(tx_queue, tx_ev_desc_ptr);
-       } else if (EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_WQ_FF_FULL)) {
-               /* Rewrite the FIFO write pointer */
-               tx_ev_q_label = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_Q_LABEL);
-               tx_queue = &efx->tx_queue[tx_ev_q_label];
-
-               if (efx_dev_registered(efx))
-                       netif_tx_lock(efx->net_dev);
-               efx_notify_tx_desc(tx_queue);
-               if (efx_dev_registered(efx))
-                       netif_tx_unlock(efx->net_dev);
-       } else if (EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_PKT_ERR) &&
-                  EFX_WORKAROUND_10727(efx)) {
-               efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
-       } else {
-               EFX_ERR(efx, "channel %d unexpected TX event "
-                       EFX_QWORD_FMT"\n", channel->channel,
-                       EFX_QWORD_VAL(*event));
-       }
-}
-
-/* Detect errors included in the rx_evt_pkt_ok bit. */
-static void efx_handle_rx_not_ok(struct efx_rx_queue *rx_queue,
-                                const efx_qword_t *event,
-                                bool *rx_ev_pkt_ok,
-                                bool *discard)
-{
-       struct efx_nic *efx = rx_queue->efx;
-       bool rx_ev_buf_owner_id_err, rx_ev_ip_hdr_chksum_err;
-       bool rx_ev_tcp_udp_chksum_err, rx_ev_eth_crc_err;
-       bool rx_ev_frm_trunc, rx_ev_drib_nib, rx_ev_tobe_disc;
-       bool rx_ev_other_err, rx_ev_pause_frm;
-       bool rx_ev_hdr_type, rx_ev_mcast_pkt;
-       unsigned rx_ev_pkt_type;
-
-       rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
-       rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_PKT);
-       rx_ev_tobe_disc = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_TOBE_DISC);
-       rx_ev_pkt_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_TYPE);
-       rx_ev_buf_owner_id_err = EFX_QWORD_FIELD(*event,
-                                                FSF_AZ_RX_EV_BUF_OWNER_ID_ERR);
-       rx_ev_ip_hdr_chksum_err = EFX_QWORD_FIELD(*event,
-                                                 FSF_AZ_RX_EV_IP_HDR_CHKSUM_ERR);
-       rx_ev_tcp_udp_chksum_err = EFX_QWORD_FIELD(*event,
-                                                  FSF_AZ_RX_EV_TCP_UDP_CHKSUM_ERR);
-       rx_ev_eth_crc_err = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_ETH_CRC_ERR);
-       rx_ev_frm_trunc = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_FRM_TRUNC);
-       rx_ev_drib_nib = ((efx_nic_rev(efx) >= EFX_REV_FALCON_B0) ?
-                         0 : EFX_QWORD_FIELD(*event, FSF_AA_RX_EV_DRIB_NIB));
-       rx_ev_pause_frm = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PAUSE_FRM_ERR);
-
-       /* Every error apart from tobe_disc and pause_frm */
-       rx_ev_other_err = (rx_ev_drib_nib | rx_ev_tcp_udp_chksum_err |
-                          rx_ev_buf_owner_id_err | rx_ev_eth_crc_err |
-                          rx_ev_frm_trunc | rx_ev_ip_hdr_chksum_err);
-
-       /* Count errors that are not in MAC stats.  Ignore expected
-        * checksum errors during self-test. */
-       if (rx_ev_frm_trunc)
-               ++rx_queue->channel->n_rx_frm_trunc;
-       else if (rx_ev_tobe_disc)
-               ++rx_queue->channel->n_rx_tobe_disc;
-       else if (!efx->loopback_selftest) {
-               if (rx_ev_ip_hdr_chksum_err)
-                       ++rx_queue->channel->n_rx_ip_hdr_chksum_err;
-               else if (rx_ev_tcp_udp_chksum_err)
-                       ++rx_queue->channel->n_rx_tcp_udp_chksum_err;
-       }
-
-       /* The frame must be discarded if any of these are true. */
-       *discard = (rx_ev_eth_crc_err | rx_ev_frm_trunc | rx_ev_drib_nib |
-                   rx_ev_tobe_disc | rx_ev_pause_frm);
-
-       /* TOBE_DISC is expected on unicast mismatches; don't print out an
-        * error message.  FRM_TRUNC indicates RXDP dropped the packet due
-        * to a FIFO overflow.
-        */
-#ifdef EFX_ENABLE_DEBUG
-       if (rx_ev_other_err) {
-               EFX_INFO_RL(efx, " RX queue %d unexpected RX event "
-                           EFX_QWORD_FMT "%s%s%s%s%s%s%s%s\n",
-                           rx_queue->queue, EFX_QWORD_VAL(*event),
-                           rx_ev_buf_owner_id_err ? " [OWNER_ID_ERR]" : "",
-                           rx_ev_ip_hdr_chksum_err ?
-                           " [IP_HDR_CHKSUM_ERR]" : "",
-                           rx_ev_tcp_udp_chksum_err ?
-                           " [TCP_UDP_CHKSUM_ERR]" : "",
-                           rx_ev_eth_crc_err ? " [ETH_CRC_ERR]" : "",
-                           rx_ev_frm_trunc ? " [FRM_TRUNC]" : "",
-                           rx_ev_drib_nib ? " [DRIB_NIB]" : "",
-                           rx_ev_tobe_disc ? " [TOBE_DISC]" : "",
-                           rx_ev_pause_frm ? " [PAUSE]" : "");
-       }
-#endif
-}
-
-/* Handle receive events that are not in-order. */
-static void
-efx_handle_rx_bad_index(struct efx_rx_queue *rx_queue, unsigned index)
-{
-       struct efx_nic *efx = rx_queue->efx;
-       unsigned expected, dropped;
-
-       expected = rx_queue->removed_count & EFX_RXQ_MASK;
-       dropped = (index - expected) & EFX_RXQ_MASK;
-       EFX_INFO(efx, "dropped %d events (index=%d expected=%d)\n",
-               dropped, index, expected);
-
-       efx_schedule_reset(efx, EFX_WORKAROUND_5676(efx) ?
-                          RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
-}
-
-/* Handle a packet received event
- *
- * The NIC gives a "discard" flag if it's a unicast packet with the
- * wrong destination address
- * Also "is multicast" and "matches multicast filter" flags can be used to
- * discard non-matching multicast packets.
- */
-static void
-efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
-{
-       unsigned int rx_ev_desc_ptr, rx_ev_byte_cnt;
-       unsigned int rx_ev_hdr_type, rx_ev_mcast_pkt;
-       unsigned expected_ptr;
-       bool rx_ev_pkt_ok, discard = false, checksummed;
-       struct efx_rx_queue *rx_queue;
-       struct efx_nic *efx = channel->efx;
-
-       /* Basic packet information */
-       rx_ev_byte_cnt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_BYTE_CNT);
-       rx_ev_pkt_ok = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_OK);
-       rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
-       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT));
-       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_SOP) != 1);
-       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_Q_LABEL) !=
-               channel->channel);
-
-       rx_queue = &efx->rx_queue[channel->channel];
-
-       rx_ev_desc_ptr = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_DESC_PTR);
-       expected_ptr = rx_queue->removed_count & EFX_RXQ_MASK;
-       if (unlikely(rx_ev_desc_ptr != expected_ptr))
-               efx_handle_rx_bad_index(rx_queue, rx_ev_desc_ptr);
-
-       if (likely(rx_ev_pkt_ok)) {
-               /* If packet is marked as OK and packet type is TCP/IP or
-                * UDP/IP, then we can rely on the hardware checksum.
-                */
-               checksummed =
-                       likely(efx->rx_checksum_enabled) &&
-                       (rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_TCP ||
-                        rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_UDP);
-       } else {
-               efx_handle_rx_not_ok(rx_queue, event, &rx_ev_pkt_ok, &discard);
-               checksummed = false;
-       }
-
-       /* Detect multicast packets that didn't match the filter */
-       rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_PKT);
-       if (rx_ev_mcast_pkt) {
-               unsigned int rx_ev_mcast_hash_match =
-                       EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_HASH_MATCH);
-
-               if (unlikely(!rx_ev_mcast_hash_match)) {
-                       ++channel->n_rx_mcast_mismatch;
-                       discard = true;
-               }
-       }
-
-       channel->irq_mod_score += 2;
-
-       /* Handle received packet */
-       efx_rx_packet(rx_queue, rx_ev_desc_ptr, rx_ev_byte_cnt,
-                     checksummed, discard);
-}
-
-/* Global events are basically PHY events */
-static void
-efx_handle_global_event(struct efx_channel *channel, efx_qword_t *event)
-{
-       struct efx_nic *efx = channel->efx;
-       bool handled = false;
-
-       if (EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_G_PHY0_INTR) ||
-           EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_XG_PHY0_INTR) ||
-           EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_XFP_PHY0_INTR)) {
-               /* Ignored */
-               handled = true;
-       }
-
-       if ((efx_nic_rev(efx) >= EFX_REV_FALCON_B0) &&
-           EFX_QWORD_FIELD(*event, FSF_BB_GLB_EV_XG_MGT_INTR)) {
-               efx->xmac_poll_required = true;
-               handled = true;
-       }
-
-       if (efx_nic_rev(efx) <= EFX_REV_FALCON_A1 ?
-           EFX_QWORD_FIELD(*event, FSF_AA_GLB_EV_RX_RECOVERY) :
-           EFX_QWORD_FIELD(*event, FSF_BB_GLB_EV_RX_RECOVERY)) {
-               EFX_ERR(efx, "channel %d seen global RX_RESET "
-                       "event. Resetting.\n", channel->channel);
-
-               atomic_inc(&efx->rx_reset);
-               efx_schedule_reset(efx, EFX_WORKAROUND_6555(efx) ?
-                                  RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
-               handled = true;
-       }
-
-       if (!handled)
-               EFX_ERR(efx, "channel %d unknown global event "
-                       EFX_QWORD_FMT "\n", channel->channel,
-                       EFX_QWORD_VAL(*event));
-}
-
-static void
-efx_handle_driver_event(struct efx_channel *channel, efx_qword_t *event)
-{
-       struct efx_nic *efx = channel->efx;
-       unsigned int ev_sub_code;
-       unsigned int ev_sub_data;
-
-       ev_sub_code = EFX_QWORD_FIELD(*event, FSF_AZ_DRIVER_EV_SUBCODE);
-       ev_sub_data = EFX_QWORD_FIELD(*event, FSF_AZ_DRIVER_EV_SUBDATA);
-
-       switch (ev_sub_code) {
-       case FSE_AZ_TX_DESCQ_FLS_DONE_EV:
-               EFX_TRACE(efx, "channel %d TXQ %d flushed\n",
-                         channel->channel, ev_sub_data);
-               break;
-       case FSE_AZ_RX_DESCQ_FLS_DONE_EV:
-               EFX_TRACE(efx, "channel %d RXQ %d flushed\n",
-                         channel->channel, ev_sub_data);
-               break;
-       case FSE_AZ_EVQ_INIT_DONE_EV:
-               EFX_LOG(efx, "channel %d EVQ %d initialised\n",
-                       channel->channel, ev_sub_data);
-               break;
-       case FSE_AZ_SRM_UPD_DONE_EV:
-               EFX_TRACE(efx, "channel %d SRAM update done\n",
-                         channel->channel);
-               break;
-       case FSE_AZ_WAKE_UP_EV:
-               EFX_TRACE(efx, "channel %d RXQ %d wakeup event\n",
-                         channel->channel, ev_sub_data);
-               break;
-       case FSE_AZ_TIMER_EV:
-               EFX_TRACE(efx, "channel %d RX queue %d timer expired\n",
-                         channel->channel, ev_sub_data);
-               break;
-       case FSE_AA_RX_RECOVER_EV:
-               EFX_ERR(efx, "channel %d seen DRIVER RX_RESET event. "
-                       "Resetting.\n", channel->channel);
-               atomic_inc(&efx->rx_reset);
-               efx_schedule_reset(efx,
-                                  EFX_WORKAROUND_6555(efx) ?
-                                  RESET_TYPE_RX_RECOVERY :
-                                  RESET_TYPE_DISABLE);
-               break;
-       case FSE_BZ_RX_DSC_ERROR_EV:
-               EFX_ERR(efx, "RX DMA Q %d reports descriptor fetch error."
-                       " RX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
-               efx_schedule_reset(efx, RESET_TYPE_RX_DESC_FETCH);
-               break;
-       case FSE_BZ_TX_DSC_ERROR_EV:
-               EFX_ERR(efx, "TX DMA Q %d reports descriptor fetch error."
-                       " TX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
-               efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
-               break;
-       default:
-               EFX_TRACE(efx, "channel %d unknown driver event code %d "
-                         "data %04x\n", channel->channel, ev_sub_code,
-                         ev_sub_data);
-               break;
-       }
-}
-
-int efx_nic_process_eventq(struct efx_channel *channel, int rx_quota)
-{
-       unsigned int read_ptr;
-       efx_qword_t event, *p_event;
-       int ev_code;
-       int rx_packets = 0;
-
-       read_ptr = channel->eventq_read_ptr;
-
-       do {
-               p_event = efx_event(channel, read_ptr);
-               event = *p_event;
-
-               if (!efx_event_present(&event))
-                       /* End of events */
-                       break;
-
-               EFX_TRACE(channel->efx, "channel %d event is "EFX_QWORD_FMT"\n",
-                         channel->channel, EFX_QWORD_VAL(event));
-
-               /* Clear this event by marking it all ones */
-               EFX_SET_QWORD(*p_event);
-
-               ev_code = EFX_QWORD_FIELD(event, FSF_AZ_EV_CODE);
-
-               switch (ev_code) {
-               case FSE_AZ_EV_CODE_RX_EV:
-                       efx_handle_rx_event(channel, &event);
-                       ++rx_packets;
-                       break;
-               case FSE_AZ_EV_CODE_TX_EV:
-                       efx_handle_tx_event(channel, &event);
-                       break;
-               case FSE_AZ_EV_CODE_DRV_GEN_EV:
-                       channel->eventq_magic = EFX_QWORD_FIELD(
-                               event, FSF_AZ_DRV_GEN_EV_MAGIC);
-                       EFX_LOG(channel->efx, "channel %d received generated "
-                               "event "EFX_QWORD_FMT"\n", channel->channel,
-                               EFX_QWORD_VAL(event));
-                       break;
-               case FSE_AZ_EV_CODE_GLOBAL_EV:
-                       efx_handle_global_event(channel, &event);
-                       break;
-               case FSE_AZ_EV_CODE_DRIVER_EV:
-                       efx_handle_driver_event(channel, &event);
-                       break;
-               default:
-                       EFX_ERR(channel->efx, "channel %d unknown event type %d"
-                               " (data " EFX_QWORD_FMT ")\n", channel->channel,
-                               ev_code, EFX_QWORD_VAL(event));
-               }
+       struct efx_nic *efx = (struct efx_nic *)data;
+       efx_oword_t reg;
 
-               /* Increment read pointer */
-               read_ptr = (read_ptr + 1) & EFX_EVQ_MASK;
+       efx_reado(efx, &reg, FR_AB_GPIO_CTL);
+       return EFX_OWORD_FIELD(reg, FRF_AB_GPIO3_IN);
+}
 
-       } while (rx_packets < rx_quota);
+static int falcon_getscl(void *data)
+{
+       struct efx_nic *efx = (struct efx_nic *)data;
+       efx_oword_t reg;
 
-       channel->eventq_read_ptr = read_ptr;
-       return rx_packets;
+       efx_reado(efx, &reg, FR_AB_GPIO_CTL);
+       return EFX_OWORD_FIELD(reg, FRF_AB_GPIO0_IN);
 }
 
+static struct i2c_algo_bit_data falcon_i2c_bit_operations = {
+       .setsda         = falcon_setsda,
+       .setscl         = falcon_setscl,
+       .getsda         = falcon_getsda,
+       .getscl         = falcon_getscl,
+       .udelay         = 5,
+       /* Wait up to 50 ms for slave to let us pull SCL high */
+       .timeout        = DIV_ROUND_UP(HZ, 20),
+};
+
 static void falcon_push_irq_moderation(struct efx_channel *channel)
 {
        efx_dword_t timer_cmd;
@@ -1056,135 +120,6 @@ static void falcon_push_irq_moderation(struct efx_channel *channel)
        BUILD_BUG_ON(FR_AA_TIMER_COMMAND_KER != FR_BZ_TIMER_COMMAND_P0);
        efx_writed_page_locked(efx, &timer_cmd, FR_BZ_TIMER_COMMAND_P0,
                               channel->channel);
-
-}
-
-/* Allocate buffer table entries for event queue */
-int efx_nic_probe_eventq(struct efx_channel *channel)
-{
-       struct efx_nic *efx = channel->efx;
-       BUILD_BUG_ON(EFX_EVQ_SIZE < 512 || EFX_EVQ_SIZE > 32768 ||
-                    EFX_EVQ_SIZE & EFX_EVQ_MASK);
-       return efx_alloc_special_buffer(efx, &channel->eventq,
-                                       EFX_EVQ_SIZE * sizeof(efx_qword_t));
-}
-
-void efx_nic_init_eventq(struct efx_channel *channel)
-{
-       efx_oword_t evq_ptr;
-       struct efx_nic *efx = channel->efx;
-
-       EFX_LOG(efx, "channel %d event queue in special buffers %d-%d\n",
-               channel->channel, channel->eventq.index,
-               channel->eventq.index + channel->eventq.entries - 1);
-
-       /* Pin event queue buffer */
-       efx_init_special_buffer(efx, &channel->eventq);
-
-       /* Fill event queue with all ones (i.e. empty events) */
-       memset(channel->eventq.addr, 0xff, channel->eventq.len);
-
-       /* Push event queue to card */
-       EFX_POPULATE_OWORD_3(evq_ptr,
-                            FRF_AZ_EVQ_EN, 1,
-                            FRF_AZ_EVQ_SIZE, __ffs(channel->eventq.entries),
-                            FRF_AZ_EVQ_BUF_BASE_ID, channel->eventq.index);
-       efx_writeo_table(efx, &evq_ptr, efx->type->evq_ptr_tbl_base,
-                        channel->channel);
-
-       efx->type->push_irq_moderation(channel);
-}
-
-void efx_nic_fini_eventq(struct efx_channel *channel)
-{
-       efx_oword_t eventq_ptr;
-       struct efx_nic *efx = channel->efx;
-
-       /* Remove event queue from card */
-       EFX_ZERO_OWORD(eventq_ptr);
-       efx_writeo_table(efx, &eventq_ptr, efx->type->evq_ptr_tbl_base,
-                        channel->channel);
-
-       /* Unpin event queue */
-       efx_fini_special_buffer(efx, &channel->eventq);
-}
-
-/* Free buffers backing event queue */
-void efx_nic_remove_eventq(struct efx_channel *channel)
-{
-       efx_free_special_buffer(channel->efx, &channel->eventq);
-}
-
-
-/* Generates a test event on the event queue.  A subsequent call to
- * process_eventq() should pick up the event and place the value of
- * "magic" into channel->eventq_magic;
- */
-void efx_nic_generate_test_event(struct efx_channel *channel, unsigned int magic)
-{
-       efx_qword_t test_event;
-
-       EFX_POPULATE_QWORD_2(test_event, FSF_AZ_EV_CODE,
-                            FSE_AZ_EV_CODE_DRV_GEN_EV,
-                            FSF_AZ_DRV_GEN_EV_MAGIC, magic);
-       efx_generate_event(channel, &test_event);
-}
-
-/**************************************************************************
- *
- * Flush handling
- *
- **************************************************************************/
-
-
-static void efx_poll_flush_events(struct efx_nic *efx)
-{
-       struct efx_channel *channel = &efx->channel[0];
-       struct efx_tx_queue *tx_queue;
-       struct efx_rx_queue *rx_queue;
-       unsigned int read_ptr = channel->eventq_read_ptr;
-       unsigned int end_ptr = (read_ptr - 1) & EFX_EVQ_MASK;
-
-       do {
-               efx_qword_t *event = efx_event(channel, read_ptr);
-               int ev_code, ev_sub_code, ev_queue;
-               bool ev_failed;
-
-               if (!efx_event_present(event))
-                       break;
-
-               ev_code = EFX_QWORD_FIELD(*event, FSF_AZ_EV_CODE);
-               ev_sub_code = EFX_QWORD_FIELD(*event,
-                                             FSF_AZ_DRIVER_EV_SUBCODE);
-               if (ev_code == FSE_AZ_EV_CODE_DRIVER_EV &&
-                   ev_sub_code == FSE_AZ_TX_DESCQ_FLS_DONE_EV) {
-                       ev_queue = EFX_QWORD_FIELD(*event,
-                                                  FSF_AZ_DRIVER_EV_SUBDATA);
-                       if (ev_queue < EFX_TX_QUEUE_COUNT) {
-                               tx_queue = efx->tx_queue + ev_queue;
-                               tx_queue->flushed = FLUSH_DONE;
-                       }
-               } else if (ev_code == FSE_AZ_EV_CODE_DRIVER_EV &&
-                          ev_sub_code == FSE_AZ_RX_DESCQ_FLS_DONE_EV) {
-                       ev_queue = EFX_QWORD_FIELD(
-                               *event, FSF_AZ_DRIVER_EV_RX_DESCQ_ID);
-                       ev_failed = EFX_QWORD_FIELD(
-                               *event, FSF_AZ_DRIVER_EV_RX_FLUSH_FAIL);
-                       if (ev_queue < efx->n_rx_queues) {
-                               rx_queue = efx->rx_queue + ev_queue;
-                               rx_queue->flushed =
-                                       ev_failed ? FLUSH_FAILED : FLUSH_DONE;
-                       }
-               }
-
-               /* We're about to destroy the queue anyway, so
-                * it's ok to throw away every non-flush event */
-               EFX_SET_QWORD(*event);
-
-               read_ptr = (read_ptr + 1) & EFX_EVQ_MASK;
-       } while (read_ptr != end_ptr);
-
-       channel->eventq_read_ptr = read_ptr;
 }
 
 static void falcon_deconfigure_mac_wrapper(struct efx_nic *efx);
@@ -1199,123 +134,6 @@ static void falcon_prepare_flush(struct efx_nic *efx)
        msleep(10);
 }
 
-/* Handle tx and rx flushes at the same time, since they run in
- * parallel in the hardware and there's no reason for us to
- * serialise them */
-int efx_nic_flush_queues(struct efx_nic *efx)
-{
-       struct efx_rx_queue *rx_queue;
-       struct efx_tx_queue *tx_queue;
-       int i, tx_pending, rx_pending;
-
-       /* If necessary prepare the hardware for flushing */
-       efx->type->prepare_flush(efx);
-
-       /* Flush all tx queues in parallel */
-       efx_for_each_tx_queue(tx_queue, efx)
-               efx_flush_tx_queue(tx_queue);
-
-       /* The hardware supports four concurrent rx flushes, each of which may
-        * need to be retried if there is an outstanding descriptor fetch */
-       for (i = 0; i < EFX_FLUSH_POLL_COUNT; ++i) {
-               rx_pending = tx_pending = 0;
-               efx_for_each_rx_queue(rx_queue, efx) {
-                       if (rx_queue->flushed == FLUSH_PENDING)
-                               ++rx_pending;
-               }
-               efx_for_each_rx_queue(rx_queue, efx) {
-                       if (rx_pending == EFX_RX_FLUSH_COUNT)
-                               break;
-                       if (rx_queue->flushed == FLUSH_FAILED ||
-                           rx_queue->flushed == FLUSH_NONE) {
-                               efx_flush_rx_queue(rx_queue);
-                               ++rx_pending;
-                       }
-               }
-               efx_for_each_tx_queue(tx_queue, efx) {
-                       if (tx_queue->flushed != FLUSH_DONE)
-                               ++tx_pending;
-               }
-
-               if (rx_pending == 0 && tx_pending == 0)
-                       return 0;
-
-               msleep(EFX_FLUSH_INTERVAL);
-               efx_poll_flush_events(efx);
-       }
-
-       /* Mark the queues as all flushed. We're going to return failure
-        * leading to a reset, or fake up success anyway */
-       efx_for_each_tx_queue(tx_queue, efx) {
-               if (tx_queue->flushed != FLUSH_DONE)
-                       EFX_ERR(efx, "tx queue %d flush command timed out\n",
-                               tx_queue->queue);
-               tx_queue->flushed = FLUSH_DONE;
-       }
-       efx_for_each_rx_queue(rx_queue, efx) {
-               if (rx_queue->flushed != FLUSH_DONE)
-                       EFX_ERR(efx, "rx queue %d flush command timed out\n",
-                               rx_queue->queue);
-               rx_queue->flushed = FLUSH_DONE;
-       }
-
-       if (EFX_WORKAROUND_7803(efx))
-               return 0;
-
-       return -ETIMEDOUT;
-}
-
-/**************************************************************************
- *
- * Hardware interrupts
- * The hardware interrupt handler does very little work; all the event
- * queue processing is carried out by per-channel tasklets.
- *
- **************************************************************************/
-
-/* Enable/disable/generate interrupts */
-static inline void efx_nic_interrupts(struct efx_nic *efx,
-                                     bool enabled, bool force)
-{
-       efx_oword_t int_en_reg_ker;
-
-       EFX_POPULATE_OWORD_2(int_en_reg_ker,
-                            FRF_AZ_KER_INT_KER, force,
-                            FRF_AZ_DRV_INT_EN_KER, enabled);
-       efx_writeo(efx, &int_en_reg_ker, FR_AZ_INT_EN_KER);
-}
-
-void efx_nic_enable_interrupts(struct efx_nic *efx)
-{
-       struct efx_channel *channel;
-
-       EFX_ZERO_OWORD(*((efx_oword_t *) efx->irq_status.addr));
-       wmb(); /* Ensure interrupt vector is clear before interrupts enabled */
-
-       /* Enable interrupts */
-       efx_nic_interrupts(efx, true, false);
-
-       /* Force processing of all the channels to get the EVQ RPTRs up to
-          date */
-       efx_for_each_channel(channel, efx)
-               efx_schedule_channel(channel);
-}
-
-void efx_nic_disable_interrupts(struct efx_nic *efx)
-{
-       /* Disable interrupts */
-       efx_nic_interrupts(efx, false, false);
-}
-
-/* Generate a test interrupt
- * Interrupt must already have been enabled, otherwise nasty things
- * may happen.
- */
-void efx_nic_generate_interrupt(struct efx_nic *efx)
-{
-       efx_nic_interrupts(efx, true, true);
-}
-
 /* Acknowledge a legacy interrupt from Falcon
  *
  * This acknowledges a legacy (not MSI) interrupt via INT_ACK_KER_REG.
@@ -1335,102 +153,6 @@ inline void falcon_irq_ack_a1(struct efx_nic *efx)
        efx_readd(efx, &reg, FR_AA_WORK_AROUND_BROKEN_PCI_READS);
 }
 
-/* Process a fatal interrupt
- * Disable bus mastering ASAP and schedule a reset
- */
-irqreturn_t efx_nic_fatal_interrupt(struct efx_nic *efx)
-{
-       struct falcon_nic_data *nic_data = efx->nic_data;
-       efx_oword_t *int_ker = efx->irq_status.addr;
-       efx_oword_t fatal_intr;
-       int error, mem_perr;
-
-       efx_reado(efx, &fatal_intr, FR_AZ_FATAL_INTR_KER);
-       error = EFX_OWORD_FIELD(fatal_intr, FRF_AZ_FATAL_INTR);
-
-       EFX_ERR(efx, "SYSTEM ERROR " EFX_OWORD_FMT " status "
-               EFX_OWORD_FMT ": %s\n", EFX_OWORD_VAL(*int_ker),
-               EFX_OWORD_VAL(fatal_intr),
-               error ? "disabling bus mastering" : "no recognised error");
-       if (error == 0)
-               goto out;
-
-       /* If this is a memory parity error dump which blocks are offending */
-       mem_perr = EFX_OWORD_FIELD(fatal_intr, FRF_AZ_MEM_PERR_INT_KER);
-       if (mem_perr) {
-               efx_oword_t reg;
-               efx_reado(efx, &reg, FR_AZ_MEM_STAT);
-               EFX_ERR(efx, "SYSTEM ERROR: memory parity error "
-                       EFX_OWORD_FMT "\n", EFX_OWORD_VAL(reg));
-       }
-
-       /* Disable both devices */
-       pci_clear_master(efx->pci_dev);
-       if (efx_nic_is_dual_func(efx))
-               pci_clear_master(nic_data->pci_dev2);
-       efx_nic_disable_interrupts(efx);
-
-       /* Count errors and reset or disable the NIC accordingly */
-       if (efx->int_error_count == 0 ||
-           time_after(jiffies, efx->int_error_expire)) {
-               efx->int_error_count = 0;
-               efx->int_error_expire =
-                       jiffies + EFX_INT_ERROR_EXPIRE * HZ;
-       }
-       if (++efx->int_error_count < EFX_MAX_INT_ERRORS) {
-               EFX_ERR(efx, "SYSTEM ERROR - reset scheduled\n");
-               efx_schedule_reset(efx, RESET_TYPE_INT_ERROR);
-       } else {
-               EFX_ERR(efx, "SYSTEM ERROR - max number of errors seen."
-                       "NIC will be disabled\n");
-               efx_schedule_reset(efx, RESET_TYPE_DISABLE);
-       }
-out:
-       return IRQ_HANDLED;
-}
-
-/* Handle a legacy interrupt
- * Acknowledges the interrupt and schedule event queue processing.
- */
-static irqreturn_t efx_legacy_interrupt(int irq, void *dev_id)
-{
-       struct efx_nic *efx = dev_id;
-       efx_oword_t *int_ker = efx->irq_status.addr;
-       irqreturn_t result = IRQ_NONE;
-       struct efx_channel *channel;
-       efx_dword_t reg;
-       u32 queues;
-       int syserr;
-
-       /* Read the ISR which also ACKs the interrupts */
-       efx_readd(efx, &reg, FR_BZ_INT_ISR0);
-       queues = EFX_EXTRACT_DWORD(reg, 0, 31);
-
-       /* Check to see if we have a serious error condition */
-       syserr = EFX_OWORD_FIELD(*int_ker, FSF_AZ_NET_IVEC_FATAL_INT);
-       if (unlikely(syserr))
-               return efx_nic_fatal_interrupt(efx);
-
-       /* Schedule processing of any interrupting queues */
-       efx_for_each_channel(channel, efx) {
-               if ((queues & 1) ||
-                   efx_event_present(
-                           efx_event(channel, channel->eventq_read_ptr))) {
-                       efx_schedule_channel(channel);
-                       result = IRQ_HANDLED;
-               }
-               queues >>= 1;
-       }
-
-       if (result == IRQ_HANDLED) {
-               efx->last_irq_cpu = raw_smp_processor_id();
-               EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_DWORD_FMT "\n",
-                         irq, raw_smp_processor_id(), EFX_DWORD_VAL(reg));
-       }
-
-       return result;
-}
-
 
 irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id)
 {
@@ -1477,126 +199,6 @@ irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id)
 
        return IRQ_HANDLED;
 }
-
-/* Handle an MSI interrupt
- *
- * Handle an MSI hardware interrupt.  This routine schedules event
- * queue processing.  No interrupt acknowledgement cycle is necessary.
- * Also, we never need to check that the interrupt is for us, since
- * MSI interrupts cannot be shared.
- */
-static irqreturn_t efx_msi_interrupt(int irq, void *dev_id)
-{
-       struct efx_channel *channel = dev_id;
-       struct efx_nic *efx = channel->efx;
-       efx_oword_t *int_ker = efx->irq_status.addr;
-       int syserr;
-
-       efx->last_irq_cpu = raw_smp_processor_id();
-       EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_OWORD_FMT "\n",
-                 irq, raw_smp_processor_id(), EFX_OWORD_VAL(*int_ker));
-
-       /* Check to see if we have a serious error condition */
-       syserr = EFX_OWORD_FIELD(*int_ker, FSF_AZ_NET_IVEC_FATAL_INT);
-       if (unlikely(syserr))
-               return efx_nic_fatal_interrupt(efx);
-
-       /* Schedule processing of the channel */
-       efx_schedule_channel(channel);
-
-       return IRQ_HANDLED;
-}
-
-
-/* Setup RSS indirection table.
- * This maps from the hash value of the packet to RXQ
- */
-static void efx_setup_rss_indir_table(struct efx_nic *efx)
-{
-       int i = 0;
-       unsigned long offset;
-       efx_dword_t dword;
-
-       if (efx_nic_rev(efx) < EFX_REV_FALCON_B0)
-               return;
-
-       for (offset = FR_BZ_RX_INDIRECTION_TBL;
-            offset < FR_BZ_RX_INDIRECTION_TBL + 0x800;
-            offset += 0x10) {
-               EFX_POPULATE_DWORD_1(dword, FRF_BZ_IT_QUEUE,
-                                    i % efx->n_rx_queues);
-               efx_writed(efx, &dword, offset);
-               i++;
-       }
-}
-
-/* Hook interrupt handler(s)
- * Try MSI and then legacy interrupts.
- */
-int efx_nic_init_interrupt(struct efx_nic *efx)
-{
-       struct efx_channel *channel;
-       int rc;
-
-       if (!EFX_INT_MODE_USE_MSI(efx)) {
-               irq_handler_t handler;
-               if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
-                       handler = efx_legacy_interrupt;
-               else
-                       handler = falcon_legacy_interrupt_a1;
-
-               rc = request_irq(efx->legacy_irq, handler, IRQF_SHARED,
-                                efx->name, efx);
-               if (rc) {
-                       EFX_ERR(efx, "failed to hook legacy IRQ %d\n",
-                               efx->pci_dev->irq);
-                       goto fail1;
-               }
-               return 0;
-       }
-
-       /* Hook MSI or MSI-X interrupt */
-       efx_for_each_channel(channel, efx) {
-               rc = request_irq(channel->irq, efx_msi_interrupt,
-                                IRQF_PROBE_SHARED, /* Not shared */
-                                channel->name, channel);
-               if (rc) {
-                       EFX_ERR(efx, "failed to hook IRQ %d\n", channel->irq);
-                       goto fail2;
-               }
-       }
-
-       return 0;
-
- fail2:
-       efx_for_each_channel(channel, efx)
-               free_irq(channel->irq, channel);
- fail1:
-       return rc;
-}
-
-void efx_nic_fini_interrupt(struct efx_nic *efx)
-{
-       struct efx_channel *channel;
-       efx_oword_t reg;
-
-       /* Disable MSI/MSI-X interrupts */
-       efx_for_each_channel(channel, efx) {
-               if (channel->irq)
-                       free_irq(channel->irq, channel);
-       }
-
-       /* ACK legacy interrupt */
-       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
-               efx_reado(efx, &reg, FR_BZ_INT_ISR0);
-       else
-               falcon_irq_ack_a1(efx);
-
-       /* Disable legacy interrupt */
-       if (efx->legacy_irq)
-               free_irq(efx->legacy_irq, efx);
-}
-
 /**************************************************************************
  *
  * EEPROM/flash
@@ -2440,68 +1042,6 @@ static const struct efx_nic_register_test falcon_b0_register_tests[] = {
          EFX_OWORD32(0x0003FF0F, 0x00000000, 0x00000000, 0x00000000) },
 };
 
-static bool efx_masked_compare_oword(const efx_oword_t *a, const efx_oword_t *b,
-                                    const efx_oword_t *mask)
-{
-       return ((a->u64[0] ^ b->u64[0]) & mask->u64[0]) ||
-               ((a->u64[1] ^ b->u64[1]) & mask->u64[1]);
-}
-
-int efx_nic_test_registers(struct efx_nic *efx,
-                          const struct efx_nic_register_test *regs,
-                          size_t n_regs)
-{
-       unsigned address = 0, i, j;
-       efx_oword_t mask, imask, original, reg, buf;
-
-       /* Falcon should be in loopback to isolate the XMAC from the PHY */
-       WARN_ON(!LOOPBACK_INTERNAL(efx));
-
-       for (i = 0; i < n_regs; ++i) {
-               address = regs[i].address;
-               mask = imask = regs[i].mask;
-               EFX_INVERT_OWORD(imask);
-
-               efx_reado(efx, &original, address);
-
-               /* bit sweep on and off */
-               for (j = 0; j < 128; j++) {
-                       if (!EFX_EXTRACT_OWORD32(mask, j, j))
-                               continue;
-
-                       /* Test this testable bit can be set in isolation */
-                       EFX_AND_OWORD(reg, original, mask);
-                       EFX_SET_OWORD32(reg, j, j, 1);
-
-                       efx_writeo(efx, &reg, address);
-                       efx_reado(efx, &buf, address);
-
-                       if (efx_masked_compare_oword(&reg, &buf, &mask))
-                               goto fail;
-
-                       /* Test this testable bit can be cleared in isolation */
-                       EFX_OR_OWORD(reg, original, mask);
-                       EFX_SET_OWORD32(reg, j, j, 0);
-
-                       efx_writeo(efx, &reg, address);
-                       efx_reado(efx, &buf, address);
-
-                       if (efx_masked_compare_oword(&reg, &buf, &mask))
-                               goto fail;
-               }
-
-               efx_writeo(efx, &original, address);
-       }
-
-       return 0;
-
-fail:
-       EFX_ERR(efx, "wrote "EFX_OWORD_FMT" read "EFX_OWORD_FMT
-               " at address 0x%x mask "EFX_OWORD_FMT"\n", EFX_OWORD_VAL(reg),
-               EFX_OWORD_VAL(buf), address, EFX_OWORD_VAL(mask));
-       return -EIO;
-}
-
 static int falcon_b0_test_registers(struct efx_nic *efx)
 {
        return efx_nic_test_registers(efx, falcon_b0_register_tests,
@@ -2719,7 +1259,6 @@ static int falcon_spi_device_init(struct efx_nic *efx,
        return 0;
 }
 
-
 static void falcon_remove_spi_devices(struct efx_nic *efx)
 {
        kfree(efx->spi_eeprom);
@@ -2789,14 +1328,6 @@ static int falcon_probe_nvconfig(struct efx_nic *efx)
        return rc;
 }
 
-u32 efx_nic_fpga_ver(struct efx_nic *efx)
-{
-       efx_oword_t altera_build;
-
-       efx_reado(efx, &altera_build, FR_AZ_ALTERA_BUILD);
-       return EFX_OWORD_FIELD(altera_build, FRF_AZ_ALTERA_BUILD_VER);
-}
-
 /* Probe all SPI devices on the NIC */
 static void falcon_probe_spi_devices(struct efx_nic *efx)
 {
@@ -3006,73 +1537,6 @@ static void falcon_init_rx_cfg(struct efx_nic *efx)
        efx_writeo(efx, &reg, FR_AZ_RX_CFG);
 }
 
-void efx_nic_init_common(struct efx_nic *efx)
-{
-       efx_oword_t temp;
-
-       /* Set positions of descriptor caches in SRAM. */
-       EFX_POPULATE_OWORD_1(temp, FRF_AZ_SRM_TX_DC_BASE_ADR,
-                            efx->type->tx_dc_base / 8);
-       efx_writeo(efx, &temp, FR_AZ_SRM_TX_DC_CFG);
-       EFX_POPULATE_OWORD_1(temp, FRF_AZ_SRM_RX_DC_BASE_ADR,
-                            efx->type->rx_dc_base / 8);
-       efx_writeo(efx, &temp, FR_AZ_SRM_RX_DC_CFG);
-
-       /* Set TX descriptor cache size. */
-       BUILD_BUG_ON(TX_DC_ENTRIES != (8 << TX_DC_ENTRIES_ORDER));
-       EFX_POPULATE_OWORD_1(temp, FRF_AZ_TX_DC_SIZE, TX_DC_ENTRIES_ORDER);
-       efx_writeo(efx, &temp, FR_AZ_TX_DC_CFG);
-
-       /* Set RX descriptor cache size.  Set low watermark to size-8, as
-        * this allows most efficient prefetching.
-        */
-       BUILD_BUG_ON(RX_DC_ENTRIES != (8 << RX_DC_ENTRIES_ORDER));
-       EFX_POPULATE_OWORD_1(temp, FRF_AZ_RX_DC_SIZE, RX_DC_ENTRIES_ORDER);
-       efx_writeo(efx, &temp, FR_AZ_RX_DC_CFG);
-       EFX_POPULATE_OWORD_1(temp, FRF_AZ_RX_DC_PF_LWM, RX_DC_ENTRIES - 8);
-       efx_writeo(efx, &temp, FR_AZ_RX_DC_PF_WM);
-
-       /* Program INT_KER address */
-       EFX_POPULATE_OWORD_2(temp,
-                            FRF_AZ_NORM_INT_VEC_DIS_KER,
-                            EFX_INT_MODE_USE_MSI(efx),
-                            FRF_AZ_INT_ADR_KER, efx->irq_status.dma_addr);
-       efx_writeo(efx, &temp, FR_AZ_INT_ADR_KER);
-
-       /* Enable all the genuinely fatal interrupts.  (They are still
-        * masked by the overall interrupt mask, controlled by
-        * falcon_interrupts()).
-        *
-        * Note: All other fatal interrupts are enabled
-        */
-       EFX_POPULATE_OWORD_3(temp,
-                            FRF_AZ_ILL_ADR_INT_KER_EN, 1,
-                            FRF_AZ_RBUF_OWN_INT_KER_EN, 1,
-                            FRF_AZ_TBUF_OWN_INT_KER_EN, 1);
-       EFX_INVERT_OWORD(temp);
-       efx_writeo(efx, &temp, FR_AZ_FATAL_INTR_KER);
-
-       efx_setup_rss_indir_table(efx);
-
-       /* Disable the ugly timer-based TX DMA backoff and allow TX DMA to be
-        * controlled by the RX FIFO fill level. Set arbitration to one pkt/Q.
-        */
-       efx_reado(efx, &temp, FR_AZ_TX_RESERVED);
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_RX_SPACER, 0xfe);
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_RX_SPACER_EN, 1);
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_ONE_PKT_PER_Q, 1);
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_PUSH_EN, 0);
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_DIS_NON_IP_EV, 1);
-       /* Enable SW_EV to inherit in char driver - assume harmless here */
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_SOFT_EVT_EN, 1);
-       /* Prefetch threshold 2 => fetch when descriptor cache half empty */
-       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_PREF_THRESHOLD, 2);
-       /* Squash TX of packets of 16 bytes or less */
-       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
-               EFX_SET_OWORD_FIELD(temp, FRF_BZ_TX_FLUSH_MIN_LEN_EN, 1);
-       efx_writeo(efx, &temp, FR_AZ_TX_RESERVED);
-}
-
 /* This call performs hardware-specific global initialisation, such as
  * defining the descriptor cache sizes and number of RSS channels.
  * It does not set up any buffers, descriptor rings or event queues.
diff --git a/drivers/net/sfc/nic.c b/drivers/net/sfc/nic.c
new file mode 100644 (file)
index 0000000..55dbd79
--- /dev/null
@@ -0,0 +1,1548 @@
+/****************************************************************************
+ * Driver for Solarflare Solarstorm network controllers and boards
+ * Copyright 2005-2006 Fen Systems Ltd.
+ * Copyright 2006-2008 Solarflare Communications Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ */
+
+#include <linux/bitops.h>
+#include <linux/delay.h>
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+#include "net_driver.h"
+#include "bitfield.h"
+#include "efx.h"
+#include "nic.h"
+#include "regs.h"
+#include "io.h"
+#include "workarounds.h"
+
+/**************************************************************************
+ *
+ * Configurable values
+ *
+ **************************************************************************
+ */
+
+/* This is set to 16 for a good reason.  In summary, if larger than
+ * 16, the descriptor cache holds more than a default socket
+ * buffer's worth of packets (for UDP we can only have at most one
+ * socket buffer's worth outstanding).  This combined with the fact
+ * that we only get 1 TX event per descriptor cache means the NIC
+ * goes idle.
+ */
+#define TX_DC_ENTRIES 16
+#define TX_DC_ENTRIES_ORDER 1
+
+#define RX_DC_ENTRIES 64
+#define RX_DC_ENTRIES_ORDER 3
+
+/* RX FIFO XOFF watermark
+ *
+ * When the amount of the RX FIFO increases used increases past this
+ * watermark send XOFF. Only used if RX flow control is enabled (ethtool -A)
+ * This also has an effect on RX/TX arbitration
+ */
+int efx_nic_rx_xoff_thresh = -1;
+module_param_named(rx_xoff_thresh_bytes, efx_nic_rx_xoff_thresh, int, 0644);
+MODULE_PARM_DESC(rx_xoff_thresh_bytes, "RX fifo XOFF threshold");
+
+/* RX FIFO XON watermark
+ *
+ * When the amount of the RX FIFO used decreases below this
+ * watermark send XON. Only used if TX flow control is enabled (ethtool -A)
+ * This also has an effect on RX/TX arbitration
+ */
+int efx_nic_rx_xon_thresh = -1;
+module_param_named(rx_xon_thresh_bytes, efx_nic_rx_xon_thresh, int, 0644);
+MODULE_PARM_DESC(rx_xon_thresh_bytes, "RX fifo XON threshold");
+
+/* If EFX_MAX_INT_ERRORS internal errors occur within
+ * EFX_INT_ERROR_EXPIRE seconds, we consider the NIC broken and
+ * disable it.
+ */
+#define EFX_INT_ERROR_EXPIRE 3600
+#define EFX_MAX_INT_ERRORS 5
+
+/* We poll for events every FLUSH_INTERVAL ms, and check FLUSH_POLL_COUNT times
+ */
+#define EFX_FLUSH_INTERVAL 10
+#define EFX_FLUSH_POLL_COUNT 100
+
+/* Size and alignment of special buffers (4KB) */
+#define EFX_BUF_SIZE 4096
+
+/* Depth of RX flush request fifo */
+#define EFX_RX_FLUSH_COUNT 4
+
+/**************************************************************************
+ *
+ * Solarstorm hardware access
+ *
+ **************************************************************************/
+
+static inline void efx_write_buf_tbl(struct efx_nic *efx, efx_qword_t *value,
+                                    unsigned int index)
+{
+       efx_sram_writeq(efx, efx->membase + efx->type->buf_tbl_base,
+                       value, index);
+}
+
+/* Read the current event from the event queue */
+static inline efx_qword_t *efx_event(struct efx_channel *channel,
+                                    unsigned int index)
+{
+       return (((efx_qword_t *) (channel->eventq.addr)) + index);
+}
+
+/* See if an event is present
+ *
+ * We check both the high and low dword of the event for all ones.  We
+ * wrote all ones when we cleared the event, and no valid event can
+ * have all ones in either its high or low dwords.  This approach is
+ * robust against reordering.
+ *
+ * Note that using a single 64-bit comparison is incorrect; even
+ * though the CPU read will be atomic, the DMA write may not be.
+ */
+static inline int efx_event_present(efx_qword_t *event)
+{
+       return (!(EFX_DWORD_IS_ALL_ONES(event->dword[0]) |
+                 EFX_DWORD_IS_ALL_ONES(event->dword[1])));
+}
+
+static bool efx_masked_compare_oword(const efx_oword_t *a, const efx_oword_t *b,
+                                    const efx_oword_t *mask)
+{
+       return ((a->u64[0] ^ b->u64[0]) & mask->u64[0]) ||
+               ((a->u64[1] ^ b->u64[1]) & mask->u64[1]);
+}
+
+int efx_nic_test_registers(struct efx_nic *efx,
+                          const struct efx_nic_register_test *regs,
+                          size_t n_regs)
+{
+       unsigned address = 0, i, j;
+       efx_oword_t mask, imask, original, reg, buf;
+
+       /* Falcon should be in loopback to isolate the XMAC from the PHY */
+       WARN_ON(!LOOPBACK_INTERNAL(efx));
+
+       for (i = 0; i < n_regs; ++i) {
+               address = regs[i].address;
+               mask = imask = regs[i].mask;
+               EFX_INVERT_OWORD(imask);
+
+               efx_reado(efx, &original, address);
+
+               /* bit sweep on and off */
+               for (j = 0; j < 128; j++) {
+                       if (!EFX_EXTRACT_OWORD32(mask, j, j))
+                               continue;
+
+                       /* Test this testable bit can be set in isolation */
+                       EFX_AND_OWORD(reg, original, mask);
+                       EFX_SET_OWORD32(reg, j, j, 1);
+
+                       efx_writeo(efx, &reg, address);
+                       efx_reado(efx, &buf, address);
+
+                       if (efx_masked_compare_oword(&reg, &buf, &mask))
+                               goto fail;
+
+                       /* Test this testable bit can be cleared in isolation */
+                       EFX_OR_OWORD(reg, original, mask);
+                       EFX_SET_OWORD32(reg, j, j, 0);
+
+                       efx_writeo(efx, &reg, address);
+                       efx_reado(efx, &buf, address);
+
+                       if (efx_masked_compare_oword(&reg, &buf, &mask))
+                               goto fail;
+               }
+
+               efx_writeo(efx, &original, address);
+       }
+
+       return 0;
+
+fail:
+       EFX_ERR(efx, "wrote "EFX_OWORD_FMT" read "EFX_OWORD_FMT
+               " at address 0x%x mask "EFX_OWORD_FMT"\n", EFX_OWORD_VAL(reg),
+               EFX_OWORD_VAL(buf), address, EFX_OWORD_VAL(mask));
+       return -EIO;
+}
+
+/**************************************************************************
+ *
+ * Special buffer handling
+ * Special buffers are used for event queues and the TX and RX
+ * descriptor rings.
+ *
+ *************************************************************************/
+
+/*
+ * Initialise a special buffer
+ *
+ * This will define a buffer (previously allocated via
+ * efx_alloc_special_buffer()) in the buffer table, allowing
+ * it to be used for event queues, descriptor rings etc.
+ */
+static void
+efx_init_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
+{
+       efx_qword_t buf_desc;
+       int index;
+       dma_addr_t dma_addr;
+       int i;
+
+       EFX_BUG_ON_PARANOID(!buffer->addr);
+
+       /* Write buffer descriptors to NIC */
+       for (i = 0; i < buffer->entries; i++) {
+               index = buffer->index + i;
+               dma_addr = buffer->dma_addr + (i * 4096);
+               EFX_LOG(efx, "mapping special buffer %d at %llx\n",
+                       index, (unsigned long long)dma_addr);
+               EFX_POPULATE_QWORD_3(buf_desc,
+                                    FRF_AZ_BUF_ADR_REGION, 0,
+                                    FRF_AZ_BUF_ADR_FBUF, dma_addr >> 12,
+                                    FRF_AZ_BUF_OWNER_ID_FBUF, 0);
+               efx_write_buf_tbl(efx, &buf_desc, index);
+       }
+}
+
+/* Unmaps a buffer and clears the buffer table entries */
+static void
+efx_fini_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
+{
+       efx_oword_t buf_tbl_upd;
+       unsigned int start = buffer->index;
+       unsigned int end = (buffer->index + buffer->entries - 1);
+
+       if (!buffer->entries)
+               return;
+
+       EFX_LOG(efx, "unmapping special buffers %d-%d\n",
+               buffer->index, buffer->index + buffer->entries - 1);
+
+       EFX_POPULATE_OWORD_4(buf_tbl_upd,
+                            FRF_AZ_BUF_UPD_CMD, 0,
+                            FRF_AZ_BUF_CLR_CMD, 1,
+                            FRF_AZ_BUF_CLR_END_ID, end,
+                            FRF_AZ_BUF_CLR_START_ID, start);
+       efx_writeo(efx, &buf_tbl_upd, FR_AZ_BUF_TBL_UPD);
+}
+
+/*
+ * Allocate a new special buffer
+ *
+ * This allocates memory for a new buffer, clears it and allocates a
+ * new buffer ID range.  It does not write into the buffer table.
+ *
+ * This call will allocate 4KB buffers, since 8KB buffers can't be
+ * used for event queues and descriptor rings.
+ */
+static int efx_alloc_special_buffer(struct efx_nic *efx,
+                                   struct efx_special_buffer *buffer,
+                                   unsigned int len)
+{
+       len = ALIGN(len, EFX_BUF_SIZE);
+
+       buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
+                                           &buffer->dma_addr);
+       if (!buffer->addr)
+               return -ENOMEM;
+       buffer->len = len;
+       buffer->entries = len / EFX_BUF_SIZE;
+       BUG_ON(buffer->dma_addr & (EFX_BUF_SIZE - 1));
+
+       /* All zeros is a potentially valid event so memset to 0xff */
+       memset(buffer->addr, 0xff, len);
+
+       /* Select new buffer ID */
+       buffer->index = efx->next_buffer_table;
+       efx->next_buffer_table += buffer->entries;
+
+       EFX_LOG(efx, "allocating special buffers %d-%d at %llx+%x "
+               "(virt %p phys %llx)\n", buffer->index,
+               buffer->index + buffer->entries - 1,
+               (u64)buffer->dma_addr, len,
+               buffer->addr, (u64)virt_to_phys(buffer->addr));
+
+       return 0;
+}
+
+static void
+efx_free_special_buffer(struct efx_nic *efx, struct efx_special_buffer *buffer)
+{
+       if (!buffer->addr)
+               return;
+
+       EFX_LOG(efx, "deallocating special buffers %d-%d at %llx+%x "
+               "(virt %p phys %llx)\n", buffer->index,
+               buffer->index + buffer->entries - 1,
+               (u64)buffer->dma_addr, buffer->len,
+               buffer->addr, (u64)virt_to_phys(buffer->addr));
+
+       pci_free_consistent(efx->pci_dev, buffer->len, buffer->addr,
+                           buffer->dma_addr);
+       buffer->addr = NULL;
+       buffer->entries = 0;
+}
+
+/**************************************************************************
+ *
+ * Generic buffer handling
+ * These buffers are used for interrupt status and MAC stats
+ *
+ **************************************************************************/
+
+int efx_nic_alloc_buffer(struct efx_nic *efx, struct efx_buffer *buffer,
+                        unsigned int len)
+{
+       buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
+                                           &buffer->dma_addr);
+       if (!buffer->addr)
+               return -ENOMEM;
+       buffer->len = len;
+       memset(buffer->addr, 0, len);
+       return 0;
+}
+
+void efx_nic_free_buffer(struct efx_nic *efx, struct efx_buffer *buffer)
+{
+       if (buffer->addr) {
+               pci_free_consistent(efx->pci_dev, buffer->len,
+                                   buffer->addr, buffer->dma_addr);
+               buffer->addr = NULL;
+       }
+}
+
+/**************************************************************************
+ *
+ * TX path
+ *
+ **************************************************************************/
+
+/* Returns a pointer to the specified transmit descriptor in the TX
+ * descriptor queue belonging to the specified channel.
+ */
+static inline efx_qword_t *
+efx_tx_desc(struct efx_tx_queue *tx_queue, unsigned int index)
+{
+       return (((efx_qword_t *) (tx_queue->txd.addr)) + index);
+}
+
+/* This writes to the TX_DESC_WPTR; write pointer for TX descriptor ring */
+static inline void efx_notify_tx_desc(struct efx_tx_queue *tx_queue)
+{
+       unsigned write_ptr;
+       efx_dword_t reg;
+
+       write_ptr = tx_queue->write_count & EFX_TXQ_MASK;
+       EFX_POPULATE_DWORD_1(reg, FRF_AZ_TX_DESC_WPTR_DWORD, write_ptr);
+       efx_writed_page(tx_queue->efx, &reg,
+                       FR_AZ_TX_DESC_UPD_DWORD_P0, tx_queue->queue);
+}
+
+
+/* For each entry inserted into the software descriptor ring, create a
+ * descriptor in the hardware TX descriptor ring (in host memory), and
+ * write a doorbell.
+ */
+void efx_nic_push_buffers(struct efx_tx_queue *tx_queue)
+{
+
+       struct efx_tx_buffer *buffer;
+       efx_qword_t *txd;
+       unsigned write_ptr;
+
+       BUG_ON(tx_queue->write_count == tx_queue->insert_count);
+
+       do {
+               write_ptr = tx_queue->write_count & EFX_TXQ_MASK;
+               buffer = &tx_queue->buffer[write_ptr];
+               txd = efx_tx_desc(tx_queue, write_ptr);
+               ++tx_queue->write_count;
+
+               /* Create TX descriptor ring entry */
+               EFX_POPULATE_QWORD_4(*txd,
+                                    FSF_AZ_TX_KER_CONT, buffer->continuation,
+                                    FSF_AZ_TX_KER_BYTE_COUNT, buffer->len,
+                                    FSF_AZ_TX_KER_BUF_REGION, 0,
+                                    FSF_AZ_TX_KER_BUF_ADDR, buffer->dma_addr);
+       } while (tx_queue->write_count != tx_queue->insert_count);
+
+       wmb(); /* Ensure descriptors are written before they are fetched */
+       efx_notify_tx_desc(tx_queue);
+}
+
+/* Allocate hardware resources for a TX queue */
+int efx_nic_probe_tx(struct efx_tx_queue *tx_queue)
+{
+       struct efx_nic *efx = tx_queue->efx;
+       BUILD_BUG_ON(EFX_TXQ_SIZE < 512 || EFX_TXQ_SIZE > 4096 ||
+                    EFX_TXQ_SIZE & EFX_TXQ_MASK);
+       return efx_alloc_special_buffer(efx, &tx_queue->txd,
+                                       EFX_TXQ_SIZE * sizeof(efx_qword_t));
+}
+
+void efx_nic_init_tx(struct efx_tx_queue *tx_queue)
+{
+       efx_oword_t tx_desc_ptr;
+       struct efx_nic *efx = tx_queue->efx;
+
+       tx_queue->flushed = FLUSH_NONE;
+
+       /* Pin TX descriptor ring */
+       efx_init_special_buffer(efx, &tx_queue->txd);
+
+       /* Push TX descriptor ring to card */
+       EFX_POPULATE_OWORD_10(tx_desc_ptr,
+                             FRF_AZ_TX_DESCQ_EN, 1,
+                             FRF_AZ_TX_ISCSI_DDIG_EN, 0,
+                             FRF_AZ_TX_ISCSI_HDIG_EN, 0,
+                             FRF_AZ_TX_DESCQ_BUF_BASE_ID, tx_queue->txd.index,
+                             FRF_AZ_TX_DESCQ_EVQ_ID,
+                             tx_queue->channel->channel,
+                             FRF_AZ_TX_DESCQ_OWNER_ID, 0,
+                             FRF_AZ_TX_DESCQ_LABEL, tx_queue->queue,
+                             FRF_AZ_TX_DESCQ_SIZE,
+                             __ffs(tx_queue->txd.entries),
+                             FRF_AZ_TX_DESCQ_TYPE, 0,
+                             FRF_BZ_TX_NON_IP_DROP_DIS, 1);
+
+       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0) {
+               int csum = tx_queue->queue == EFX_TX_QUEUE_OFFLOAD_CSUM;
+               EFX_SET_OWORD_FIELD(tx_desc_ptr, FRF_BZ_TX_IP_CHKSM_DIS, !csum);
+               EFX_SET_OWORD_FIELD(tx_desc_ptr, FRF_BZ_TX_TCP_CHKSM_DIS,
+                                   !csum);
+       }
+
+       efx_writeo_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
+                        tx_queue->queue);
+
+       if (efx_nic_rev(efx) < EFX_REV_FALCON_B0) {
+               efx_oword_t reg;
+
+               /* Only 128 bits in this register */
+               BUILD_BUG_ON(EFX_TX_QUEUE_COUNT >= 128);
+
+               efx_reado(efx, &reg, FR_AA_TX_CHKSM_CFG);
+               if (tx_queue->queue == EFX_TX_QUEUE_OFFLOAD_CSUM)
+                       clear_bit_le(tx_queue->queue, (void *)&reg);
+               else
+                       set_bit_le(tx_queue->queue, (void *)&reg);
+               efx_writeo(efx, &reg, FR_AA_TX_CHKSM_CFG);
+       }
+}
+
+static void efx_flush_tx_queue(struct efx_tx_queue *tx_queue)
+{
+       struct efx_nic *efx = tx_queue->efx;
+       efx_oword_t tx_flush_descq;
+
+       tx_queue->flushed = FLUSH_PENDING;
+
+       /* Post a flush command */
+       EFX_POPULATE_OWORD_2(tx_flush_descq,
+                            FRF_AZ_TX_FLUSH_DESCQ_CMD, 1,
+                            FRF_AZ_TX_FLUSH_DESCQ, tx_queue->queue);
+       efx_writeo(efx, &tx_flush_descq, FR_AZ_TX_FLUSH_DESCQ);
+}
+
+void efx_nic_fini_tx(struct efx_tx_queue *tx_queue)
+{
+       struct efx_nic *efx = tx_queue->efx;
+       efx_oword_t tx_desc_ptr;
+
+       /* The queue should have been flushed */
+       WARN_ON(tx_queue->flushed != FLUSH_DONE);
+
+       /* Remove TX descriptor ring from card */
+       EFX_ZERO_OWORD(tx_desc_ptr);
+       efx_writeo_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
+                        tx_queue->queue);
+
+       /* Unpin TX descriptor ring */
+       efx_fini_special_buffer(efx, &tx_queue->txd);
+}
+
+/* Free buffers backing TX queue */
+void efx_nic_remove_tx(struct efx_tx_queue *tx_queue)
+{
+       efx_free_special_buffer(tx_queue->efx, &tx_queue->txd);
+}
+
+/**************************************************************************
+ *
+ * RX path
+ *
+ **************************************************************************/
+
+/* Returns a pointer to the specified descriptor in the RX descriptor queue */
+static inline efx_qword_t *
+efx_rx_desc(struct efx_rx_queue *rx_queue, unsigned int index)
+{
+       return (((efx_qword_t *) (rx_queue->rxd.addr)) + index);
+}
+
+/* This creates an entry in the RX descriptor queue */
+static inline void
+efx_build_rx_desc(struct efx_rx_queue *rx_queue, unsigned index)
+{
+       struct efx_rx_buffer *rx_buf;
+       efx_qword_t *rxd;
+
+       rxd = efx_rx_desc(rx_queue, index);
+       rx_buf = efx_rx_buffer(rx_queue, index);
+       EFX_POPULATE_QWORD_3(*rxd,
+                            FSF_AZ_RX_KER_BUF_SIZE,
+                            rx_buf->len -
+                            rx_queue->efx->type->rx_buffer_padding,
+                            FSF_AZ_RX_KER_BUF_REGION, 0,
+                            FSF_AZ_RX_KER_BUF_ADDR, rx_buf->dma_addr);
+}
+
+/* This writes to the RX_DESC_WPTR register for the specified receive
+ * descriptor ring.
+ */
+void efx_nic_notify_rx_desc(struct efx_rx_queue *rx_queue)
+{
+       efx_dword_t reg;
+       unsigned write_ptr;
+
+       while (rx_queue->notified_count != rx_queue->added_count) {
+               efx_build_rx_desc(rx_queue,
+                                 rx_queue->notified_count &
+                                 EFX_RXQ_MASK);
+               ++rx_queue->notified_count;
+       }
+
+       wmb();
+       write_ptr = rx_queue->added_count & EFX_RXQ_MASK;
+       EFX_POPULATE_DWORD_1(reg, FRF_AZ_RX_DESC_WPTR_DWORD, write_ptr);
+       efx_writed_page(rx_queue->efx, &reg,
+                       FR_AZ_RX_DESC_UPD_DWORD_P0, rx_queue->queue);
+}
+
+int efx_nic_probe_rx(struct efx_rx_queue *rx_queue)
+{
+       struct efx_nic *efx = rx_queue->efx;
+       BUILD_BUG_ON(EFX_RXQ_SIZE < 512 || EFX_RXQ_SIZE > 4096 ||
+                    EFX_RXQ_SIZE & EFX_RXQ_MASK);
+       return efx_alloc_special_buffer(efx, &rx_queue->rxd,
+                                       EFX_RXQ_SIZE * sizeof(efx_qword_t));
+}
+
+void efx_nic_init_rx(struct efx_rx_queue *rx_queue)
+{
+       efx_oword_t rx_desc_ptr;
+       struct efx_nic *efx = rx_queue->efx;
+       bool is_b0 = efx_nic_rev(efx) >= EFX_REV_FALCON_B0;
+       bool iscsi_digest_en = is_b0;
+
+       EFX_LOG(efx, "RX queue %d ring in special buffers %d-%d\n",
+               rx_queue->queue, rx_queue->rxd.index,
+               rx_queue->rxd.index + rx_queue->rxd.entries - 1);
+
+       rx_queue->flushed = FLUSH_NONE;
+
+       /* Pin RX descriptor ring */
+       efx_init_special_buffer(efx, &rx_queue->rxd);
+
+       /* Push RX descriptor ring to card */
+       EFX_POPULATE_OWORD_10(rx_desc_ptr,
+                             FRF_AZ_RX_ISCSI_DDIG_EN, iscsi_digest_en,
+                             FRF_AZ_RX_ISCSI_HDIG_EN, iscsi_digest_en,
+                             FRF_AZ_RX_DESCQ_BUF_BASE_ID, rx_queue->rxd.index,
+                             FRF_AZ_RX_DESCQ_EVQ_ID,
+                             rx_queue->channel->channel,
+                             FRF_AZ_RX_DESCQ_OWNER_ID, 0,
+                             FRF_AZ_RX_DESCQ_LABEL, rx_queue->queue,
+                             FRF_AZ_RX_DESCQ_SIZE,
+                             __ffs(rx_queue->rxd.entries),
+                             FRF_AZ_RX_DESCQ_TYPE, 0 /* kernel queue */ ,
+                             /* For >=B0 this is scatter so disable */
+                             FRF_AZ_RX_DESCQ_JUMBO, !is_b0,
+                             FRF_AZ_RX_DESCQ_EN, 1);
+       efx_writeo_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
+                        rx_queue->queue);
+}
+
+static void efx_flush_rx_queue(struct efx_rx_queue *rx_queue)
+{
+       struct efx_nic *efx = rx_queue->efx;
+       efx_oword_t rx_flush_descq;
+
+       rx_queue->flushed = FLUSH_PENDING;
+
+       /* Post a flush command */
+       EFX_POPULATE_OWORD_2(rx_flush_descq,
+                            FRF_AZ_RX_FLUSH_DESCQ_CMD, 1,
+                            FRF_AZ_RX_FLUSH_DESCQ, rx_queue->queue);
+       efx_writeo(efx, &rx_flush_descq, FR_AZ_RX_FLUSH_DESCQ);
+}
+
+void efx_nic_fini_rx(struct efx_rx_queue *rx_queue)
+{
+       efx_oword_t rx_desc_ptr;
+       struct efx_nic *efx = rx_queue->efx;
+
+       /* The queue should already have been flushed */
+       WARN_ON(rx_queue->flushed != FLUSH_DONE);
+
+       /* Remove RX descriptor ring from card */
+       EFX_ZERO_OWORD(rx_desc_ptr);
+       efx_writeo_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
+                        rx_queue->queue);
+
+       /* Unpin RX descriptor ring */
+       efx_fini_special_buffer(efx, &rx_queue->rxd);
+}
+
+/* Free buffers backing RX queue */
+void efx_nic_remove_rx(struct efx_rx_queue *rx_queue)
+{
+       efx_free_special_buffer(rx_queue->efx, &rx_queue->rxd);
+}
+
+/**************************************************************************
+ *
+ * Event queue processing
+ * Event queues are processed by per-channel tasklets.
+ *
+ **************************************************************************/
+
+/* Update a channel's event queue's read pointer (RPTR) register
+ *
+ * This writes the EVQ_RPTR_REG register for the specified channel's
+ * event queue.
+ *
+ * Note that EVQ_RPTR_REG contains the index of the "last read" event,
+ * whereas channel->eventq_read_ptr contains the index of the "next to
+ * read" event.
+ */
+void efx_nic_eventq_read_ack(struct efx_channel *channel)
+{
+       efx_dword_t reg;
+       struct efx_nic *efx = channel->efx;
+
+       EFX_POPULATE_DWORD_1(reg, FRF_AZ_EVQ_RPTR, channel->eventq_read_ptr);
+       efx_writed_table(efx, &reg, efx->type->evq_rptr_tbl_base,
+                        channel->channel);
+}
+
+/* Use HW to insert a SW defined event */
+void efx_generate_event(struct efx_channel *channel, efx_qword_t *event)
+{
+       efx_oword_t drv_ev_reg;
+
+       BUILD_BUG_ON(FRF_AZ_DRV_EV_DATA_LBN != 0 ||
+                    FRF_AZ_DRV_EV_DATA_WIDTH != 64);
+       drv_ev_reg.u32[0] = event->u32[0];
+       drv_ev_reg.u32[1] = event->u32[1];
+       drv_ev_reg.u32[2] = 0;
+       drv_ev_reg.u32[3] = 0;
+       EFX_SET_OWORD_FIELD(drv_ev_reg, FRF_AZ_DRV_EV_QID, channel->channel);
+       efx_writeo(channel->efx, &drv_ev_reg, FR_AZ_DRV_EV);
+}
+
+/* Handle a transmit completion event
+ *
+ * The NIC batches TX completion events; the message we receive is of
+ * the form "complete all TX events up to this index".
+ */
+static void
+efx_handle_tx_event(struct efx_channel *channel, efx_qword_t *event)
+{
+       unsigned int tx_ev_desc_ptr;
+       unsigned int tx_ev_q_label;
+       struct efx_tx_queue *tx_queue;
+       struct efx_nic *efx = channel->efx;
+
+       if (likely(EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_COMP))) {
+               /* Transmit completion */
+               tx_ev_desc_ptr = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_DESC_PTR);
+               tx_ev_q_label = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_Q_LABEL);
+               tx_queue = &efx->tx_queue[tx_ev_q_label];
+               channel->irq_mod_score +=
+                       (tx_ev_desc_ptr - tx_queue->read_count) &
+                       EFX_TXQ_MASK;
+               efx_xmit_done(tx_queue, tx_ev_desc_ptr);
+       } else if (EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_WQ_FF_FULL)) {
+               /* Rewrite the FIFO write pointer */
+               tx_ev_q_label = EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_Q_LABEL);
+               tx_queue = &efx->tx_queue[tx_ev_q_label];
+
+               if (efx_dev_registered(efx))
+                       netif_tx_lock(efx->net_dev);
+               efx_notify_tx_desc(tx_queue);
+               if (efx_dev_registered(efx))
+                       netif_tx_unlock(efx->net_dev);
+       } else if (EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_PKT_ERR) &&
+                  EFX_WORKAROUND_10727(efx)) {
+               efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
+       } else {
+               EFX_ERR(efx, "channel %d unexpected TX event "
+                       EFX_QWORD_FMT"\n", channel->channel,
+                       EFX_QWORD_VAL(*event));
+       }
+}
+
+/* Detect errors included in the rx_evt_pkt_ok bit. */
+static void efx_handle_rx_not_ok(struct efx_rx_queue *rx_queue,
+                                const efx_qword_t *event,
+                                bool *rx_ev_pkt_ok,
+                                bool *discard)
+{
+       struct efx_nic *efx = rx_queue->efx;
+       bool rx_ev_buf_owner_id_err, rx_ev_ip_hdr_chksum_err;
+       bool rx_ev_tcp_udp_chksum_err, rx_ev_eth_crc_err;
+       bool rx_ev_frm_trunc, rx_ev_drib_nib, rx_ev_tobe_disc;
+       bool rx_ev_other_err, rx_ev_pause_frm;
+       bool rx_ev_hdr_type, rx_ev_mcast_pkt;
+       unsigned rx_ev_pkt_type;
+
+       rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
+       rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_PKT);
+       rx_ev_tobe_disc = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_TOBE_DISC);
+       rx_ev_pkt_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_TYPE);
+       rx_ev_buf_owner_id_err = EFX_QWORD_FIELD(*event,
+                                                FSF_AZ_RX_EV_BUF_OWNER_ID_ERR);
+       rx_ev_ip_hdr_chksum_err = EFX_QWORD_FIELD(*event,
+                                                 FSF_AZ_RX_EV_IP_HDR_CHKSUM_ERR);
+       rx_ev_tcp_udp_chksum_err = EFX_QWORD_FIELD(*event,
+                                                  FSF_AZ_RX_EV_TCP_UDP_CHKSUM_ERR);
+       rx_ev_eth_crc_err = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_ETH_CRC_ERR);
+       rx_ev_frm_trunc = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_FRM_TRUNC);
+       rx_ev_drib_nib = ((efx_nic_rev(efx) >= EFX_REV_FALCON_B0) ?
+                         0 : EFX_QWORD_FIELD(*event, FSF_AA_RX_EV_DRIB_NIB));
+       rx_ev_pause_frm = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PAUSE_FRM_ERR);
+
+       /* Every error apart from tobe_disc and pause_frm */
+       rx_ev_other_err = (rx_ev_drib_nib | rx_ev_tcp_udp_chksum_err |
+                          rx_ev_buf_owner_id_err | rx_ev_eth_crc_err |
+                          rx_ev_frm_trunc | rx_ev_ip_hdr_chksum_err);
+
+       /* Count errors that are not in MAC stats.  Ignore expected
+        * checksum errors during self-test. */
+       if (rx_ev_frm_trunc)
+               ++rx_queue->channel->n_rx_frm_trunc;
+       else if (rx_ev_tobe_disc)
+               ++rx_queue->channel->n_rx_tobe_disc;
+       else if (!efx->loopback_selftest) {
+               if (rx_ev_ip_hdr_chksum_err)
+                       ++rx_queue->channel->n_rx_ip_hdr_chksum_err;
+               else if (rx_ev_tcp_udp_chksum_err)
+                       ++rx_queue->channel->n_rx_tcp_udp_chksum_err;
+       }
+
+       /* The frame must be discarded if any of these are true. */
+       *discard = (rx_ev_eth_crc_err | rx_ev_frm_trunc | rx_ev_drib_nib |
+                   rx_ev_tobe_disc | rx_ev_pause_frm);
+
+       /* TOBE_DISC is expected on unicast mismatches; don't print out an
+        * error message.  FRM_TRUNC indicates RXDP dropped the packet due
+        * to a FIFO overflow.
+        */
+#ifdef EFX_ENABLE_DEBUG
+       if (rx_ev_other_err) {
+               EFX_INFO_RL(efx, " RX queue %d unexpected RX event "
+                           EFX_QWORD_FMT "%s%s%s%s%s%s%s%s\n",
+                           rx_queue->queue, EFX_QWORD_VAL(*event),
+                           rx_ev_buf_owner_id_err ? " [OWNER_ID_ERR]" : "",
+                           rx_ev_ip_hdr_chksum_err ?
+                           " [IP_HDR_CHKSUM_ERR]" : "",
+                           rx_ev_tcp_udp_chksum_err ?
+                           " [TCP_UDP_CHKSUM_ERR]" : "",
+                           rx_ev_eth_crc_err ? " [ETH_CRC_ERR]" : "",
+                           rx_ev_frm_trunc ? " [FRM_TRUNC]" : "",
+                           rx_ev_drib_nib ? " [DRIB_NIB]" : "",
+                           rx_ev_tobe_disc ? " [TOBE_DISC]" : "",
+                           rx_ev_pause_frm ? " [PAUSE]" : "");
+       }
+#endif
+}
+
+/* Handle receive events that are not in-order. */
+static void
+efx_handle_rx_bad_index(struct efx_rx_queue *rx_queue, unsigned index)
+{
+       struct efx_nic *efx = rx_queue->efx;
+       unsigned expected, dropped;
+
+       expected = rx_queue->removed_count & EFX_RXQ_MASK;
+       dropped = (index - expected) & EFX_RXQ_MASK;
+       EFX_INFO(efx, "dropped %d events (index=%d expected=%d)\n",
+               dropped, index, expected);
+
+       efx_schedule_reset(efx, EFX_WORKAROUND_5676(efx) ?
+                          RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
+}
+
+/* Handle a packet received event
+ *
+ * The NIC gives a "discard" flag if it's a unicast packet with the
+ * wrong destination address
+ * Also "is multicast" and "matches multicast filter" flags can be used to
+ * discard non-matching multicast packets.
+ */
+static void
+efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
+{
+       unsigned int rx_ev_desc_ptr, rx_ev_byte_cnt;
+       unsigned int rx_ev_hdr_type, rx_ev_mcast_pkt;
+       unsigned expected_ptr;
+       bool rx_ev_pkt_ok, discard = false, checksummed;
+       struct efx_rx_queue *rx_queue;
+       struct efx_nic *efx = channel->efx;
+
+       /* Basic packet information */
+       rx_ev_byte_cnt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_BYTE_CNT);
+       rx_ev_pkt_ok = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_OK);
+       rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
+       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT));
+       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_SOP) != 1);
+       WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_Q_LABEL) !=
+               channel->channel);
+
+       rx_queue = &efx->rx_queue[channel->channel];
+
+       rx_ev_desc_ptr = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_DESC_PTR);
+       expected_ptr = rx_queue->removed_count & EFX_RXQ_MASK;
+       if (unlikely(rx_ev_desc_ptr != expected_ptr))
+               efx_handle_rx_bad_index(rx_queue, rx_ev_desc_ptr);
+
+       if (likely(rx_ev_pkt_ok)) {
+               /* If packet is marked as OK and packet type is TCP/IP or
+                * UDP/IP, then we can rely on the hardware checksum.
+                */
+               checksummed =
+                       likely(efx->rx_checksum_enabled) &&
+                       (rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_TCP ||
+                        rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_UDP);
+       } else {
+               efx_handle_rx_not_ok(rx_queue, event, &rx_ev_pkt_ok, &discard);
+               checksummed = false;
+       }
+
+       /* Detect multicast packets that didn't match the filter */
+       rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_PKT);
+       if (rx_ev_mcast_pkt) {
+               unsigned int rx_ev_mcast_hash_match =
+                       EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_MCAST_HASH_MATCH);
+
+               if (unlikely(!rx_ev_mcast_hash_match)) {
+                       ++channel->n_rx_mcast_mismatch;
+                       discard = true;
+               }
+       }
+
+       channel->irq_mod_score += 2;
+
+       /* Handle received packet */
+       efx_rx_packet(rx_queue, rx_ev_desc_ptr, rx_ev_byte_cnt,
+                     checksummed, discard);
+}
+
+/* Global events are basically PHY events */
+static void
+efx_handle_global_event(struct efx_channel *channel, efx_qword_t *event)
+{
+       struct efx_nic *efx = channel->efx;
+       bool handled = false;
+
+       if (EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_G_PHY0_INTR) ||
+           EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_XG_PHY0_INTR) ||
+           EFX_QWORD_FIELD(*event, FSF_AB_GLB_EV_XFP_PHY0_INTR)) {
+               /* Ignored */
+               handled = true;
+       }
+
+       if ((efx_nic_rev(efx) >= EFX_REV_FALCON_B0) &&
+           EFX_QWORD_FIELD(*event, FSF_BB_GLB_EV_XG_MGT_INTR)) {
+               efx->xmac_poll_required = true;
+               handled = true;
+       }
+
+       if (efx_nic_rev(efx) <= EFX_REV_FALCON_A1 ?
+           EFX_QWORD_FIELD(*event, FSF_AA_GLB_EV_RX_RECOVERY) :
+           EFX_QWORD_FIELD(*event, FSF_BB_GLB_EV_RX_RECOVERY)) {
+               EFX_ERR(efx, "channel %d seen global RX_RESET "
+                       "event. Resetting.\n", channel->channel);
+
+               atomic_inc(&efx->rx_reset);
+               efx_schedule_reset(efx, EFX_WORKAROUND_6555(efx) ?
+                                  RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
+               handled = true;
+       }
+
+       if (!handled)
+               EFX_ERR(efx, "channel %d unknown global event "
+                       EFX_QWORD_FMT "\n", channel->channel,
+                       EFX_QWORD_VAL(*event));
+}
+
+static void
+efx_handle_driver_event(struct efx_channel *channel, efx_qword_t *event)
+{
+       struct efx_nic *efx = channel->efx;
+       unsigned int ev_sub_code;
+       unsigned int ev_sub_data;
+
+       ev_sub_code = EFX_QWORD_FIELD(*event, FSF_AZ_DRIVER_EV_SUBCODE);
+       ev_sub_data = EFX_QWORD_FIELD(*event, FSF_AZ_DRIVER_EV_SUBDATA);
+
+       switch (ev_sub_code) {
+       case FSE_AZ_TX_DESCQ_FLS_DONE_EV:
+               EFX_TRACE(efx, "channel %d TXQ %d flushed\n",
+                         channel->channel, ev_sub_data);
+               break;
+       case FSE_AZ_RX_DESCQ_FLS_DONE_EV:
+               EFX_TRACE(efx, "channel %d RXQ %d flushed\n",
+                         channel->channel, ev_sub_data);
+               break;
+       case FSE_AZ_EVQ_INIT_DONE_EV:
+               EFX_LOG(efx, "channel %d EVQ %d initialised\n",
+                       channel->channel, ev_sub_data);
+               break;
+       case FSE_AZ_SRM_UPD_DONE_EV:
+               EFX_TRACE(efx, "channel %d SRAM update done\n",
+                         channel->channel);
+               break;
+       case FSE_AZ_WAKE_UP_EV:
+               EFX_TRACE(efx, "channel %d RXQ %d wakeup event\n",
+                         channel->channel, ev_sub_data);
+               break;
+       case FSE_AZ_TIMER_EV:
+               EFX_TRACE(efx, "channel %d RX queue %d timer expired\n",
+                         channel->channel, ev_sub_data);
+               break;
+       case FSE_AA_RX_RECOVER_EV:
+               EFX_ERR(efx, "channel %d seen DRIVER RX_RESET event. "
+                       "Resetting.\n", channel->channel);
+               atomic_inc(&efx->rx_reset);
+               efx_schedule_reset(efx,
+                                  EFX_WORKAROUND_6555(efx) ?
+                                  RESET_TYPE_RX_RECOVERY :
+                                  RESET_TYPE_DISABLE);
+               break;
+       case FSE_BZ_RX_DSC_ERROR_EV:
+               EFX_ERR(efx, "RX DMA Q %d reports descriptor fetch error."
+                       " RX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
+               efx_schedule_reset(efx, RESET_TYPE_RX_DESC_FETCH);
+               break;
+       case FSE_BZ_TX_DSC_ERROR_EV:
+               EFX_ERR(efx, "TX DMA Q %d reports descriptor fetch error."
+                       " TX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
+               efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
+               break;
+       default:
+               EFX_TRACE(efx, "channel %d unknown driver event code %d "
+                         "data %04x\n", channel->channel, ev_sub_code,
+                         ev_sub_data);
+               break;
+       }
+}
+
+int efx_nic_process_eventq(struct efx_channel *channel, int rx_quota)
+{
+       unsigned int read_ptr;
+       efx_qword_t event, *p_event;
+       int ev_code;
+       int rx_packets = 0;
+
+       read_ptr = channel->eventq_read_ptr;
+
+       do {
+               p_event = efx_event(channel, read_ptr);
+               event = *p_event;
+
+               if (!efx_event_present(&event))
+                       /* End of events */
+                       break;
+
+               EFX_TRACE(channel->efx, "channel %d event is "EFX_QWORD_FMT"\n",
+                         channel->channel, EFX_QWORD_VAL(event));
+
+               /* Clear this event by marking it all ones */
+               EFX_SET_QWORD(*p_event);
+
+               ev_code = EFX_QWORD_FIELD(event, FSF_AZ_EV_CODE);
+
+               switch (ev_code) {
+               case FSE_AZ_EV_CODE_RX_EV:
+                       efx_handle_rx_event(channel, &event);
+                       ++rx_packets;
+                       break;
+               case FSE_AZ_EV_CODE_TX_EV:
+                       efx_handle_tx_event(channel, &event);
+                       break;
+               case FSE_AZ_EV_CODE_DRV_GEN_EV:
+                       channel->eventq_magic = EFX_QWORD_FIELD(
+                               event, FSF_AZ_DRV_GEN_EV_MAGIC);
+                       EFX_LOG(channel->efx, "channel %d received generated "
+                               "event "EFX_QWORD_FMT"\n", channel->channel,
+                               EFX_QWORD_VAL(event));
+                       break;
+               case FSE_AZ_EV_CODE_GLOBAL_EV:
+                       efx_handle_global_event(channel, &event);
+                       break;
+               case FSE_AZ_EV_CODE_DRIVER_EV:
+                       efx_handle_driver_event(channel, &event);
+                       break;
+               default:
+                       EFX_ERR(channel->efx, "channel %d unknown event type %d"
+                               " (data " EFX_QWORD_FMT ")\n", channel->channel,
+                               ev_code, EFX_QWORD_VAL(event));
+               }
+
+               /* Increment read pointer */
+               read_ptr = (read_ptr + 1) & EFX_EVQ_MASK;
+
+       } while (rx_packets < rx_quota);
+
+       channel->eventq_read_ptr = read_ptr;
+       return rx_packets;
+}
+
+
+/* Allocate buffer table entries for event queue */
+int efx_nic_probe_eventq(struct efx_channel *channel)
+{
+       struct efx_nic *efx = channel->efx;
+       BUILD_BUG_ON(EFX_EVQ_SIZE < 512 || EFX_EVQ_SIZE > 32768 ||
+                    EFX_EVQ_SIZE & EFX_EVQ_MASK);
+       return efx_alloc_special_buffer(efx, &channel->eventq,
+                                       EFX_EVQ_SIZE * sizeof(efx_qword_t));
+}
+
+void efx_nic_init_eventq(struct efx_channel *channel)
+{
+       efx_oword_t evq_ptr;
+       struct efx_nic *efx = channel->efx;
+
+       EFX_LOG(efx, "channel %d event queue in special buffers %d-%d\n",
+               channel->channel, channel->eventq.index,
+               channel->eventq.index + channel->eventq.entries - 1);
+
+       /* Pin event queue buffer */
+       efx_init_special_buffer(efx, &channel->eventq);
+
+       /* Fill event queue with all ones (i.e. empty events) */
+       memset(channel->eventq.addr, 0xff, channel->eventq.len);
+
+       /* Push event queue to card */
+       EFX_POPULATE_OWORD_3(evq_ptr,
+                            FRF_AZ_EVQ_EN, 1,
+                            FRF_AZ_EVQ_SIZE, __ffs(channel->eventq.entries),
+                            FRF_AZ_EVQ_BUF_BASE_ID, channel->eventq.index);
+       efx_writeo_table(efx, &evq_ptr, efx->type->evq_ptr_tbl_base,
+                        channel->channel);
+
+       efx->type->push_irq_moderation(channel);
+}
+
+void efx_nic_fini_eventq(struct efx_channel *channel)
+{
+       efx_oword_t eventq_ptr;
+       struct efx_nic *efx = channel->efx;
+
+       /* Remove event queue from card */
+       EFX_ZERO_OWORD(eventq_ptr);
+       efx_writeo_table(efx, &eventq_ptr, efx->type->evq_ptr_tbl_base,
+                        channel->channel);
+
+       /* Unpin event queue */
+       efx_fini_special_buffer(efx, &channel->eventq);
+}
+
+/* Free buffers backing event queue */
+void efx_nic_remove_eventq(struct efx_channel *channel)
+{
+       efx_free_special_buffer(channel->efx, &channel->eventq);
+}
+
+
+/* Generates a test event on the event queue.  A subsequent call to
+ * process_eventq() should pick up the event and place the value of
+ * "magic" into channel->eventq_magic;
+ */
+void efx_nic_generate_test_event(struct efx_channel *channel, unsigned int magic)
+{
+       efx_qword_t test_event;
+
+       EFX_POPULATE_QWORD_2(test_event, FSF_AZ_EV_CODE,
+                            FSE_AZ_EV_CODE_DRV_GEN_EV,
+                            FSF_AZ_DRV_GEN_EV_MAGIC, magic);
+       efx_generate_event(channel, &test_event);
+}
+
+/**************************************************************************
+ *
+ * Flush handling
+ *
+ **************************************************************************/
+
+
+static void efx_poll_flush_events(struct efx_nic *efx)
+{
+       struct efx_channel *channel = &efx->channel[0];
+       struct efx_tx_queue *tx_queue;
+       struct efx_rx_queue *rx_queue;
+       unsigned int read_ptr = channel->eventq_read_ptr;
+       unsigned int end_ptr = (read_ptr - 1) & EFX_EVQ_MASK;
+
+       do {
+               efx_qword_t *event = efx_event(channel, read_ptr);
+               int ev_code, ev_sub_code, ev_queue;
+               bool ev_failed;
+
+               if (!efx_event_present(event))
+                       break;
+
+               ev_code = EFX_QWORD_FIELD(*event, FSF_AZ_EV_CODE);
+               ev_sub_code = EFX_QWORD_FIELD(*event,
+                                             FSF_AZ_DRIVER_EV_SUBCODE);
+               if (ev_code == FSE_AZ_EV_CODE_DRIVER_EV &&
+                   ev_sub_code == FSE_AZ_TX_DESCQ_FLS_DONE_EV) {
+                       ev_queue = EFX_QWORD_FIELD(*event,
+                                                  FSF_AZ_DRIVER_EV_SUBDATA);
+                       if (ev_queue < EFX_TX_QUEUE_COUNT) {
+                               tx_queue = efx->tx_queue + ev_queue;
+                               tx_queue->flushed = FLUSH_DONE;
+                       }
+               } else if (ev_code == FSE_AZ_EV_CODE_DRIVER_EV &&
+                          ev_sub_code == FSE_AZ_RX_DESCQ_FLS_DONE_EV) {
+                       ev_queue = EFX_QWORD_FIELD(
+                               *event, FSF_AZ_DRIVER_EV_RX_DESCQ_ID);
+                       ev_failed = EFX_QWORD_FIELD(
+                               *event, FSF_AZ_DRIVER_EV_RX_FLUSH_FAIL);
+                       if (ev_queue < efx->n_rx_queues) {
+                               rx_queue = efx->rx_queue + ev_queue;
+                               rx_queue->flushed =
+                                       ev_failed ? FLUSH_FAILED : FLUSH_DONE;
+                       }
+               }
+
+               /* We're about to destroy the queue anyway, so
+                * it's ok to throw away every non-flush event */
+               EFX_SET_QWORD(*event);
+
+               read_ptr = (read_ptr + 1) & EFX_EVQ_MASK;
+       } while (read_ptr != end_ptr);
+
+       channel->eventq_read_ptr = read_ptr;
+}
+
+/* Handle tx and rx flushes at the same time, since they run in
+ * parallel in the hardware and there's no reason for us to
+ * serialise them */
+int efx_nic_flush_queues(struct efx_nic *efx)
+{
+       struct efx_rx_queue *rx_queue;
+       struct efx_tx_queue *tx_queue;
+       int i, tx_pending, rx_pending;
+
+       /* If necessary prepare the hardware for flushing */
+       efx->type->prepare_flush(efx);
+
+       /* Flush all tx queues in parallel */
+       efx_for_each_tx_queue(tx_queue, efx)
+               efx_flush_tx_queue(tx_queue);
+
+       /* The hardware supports four concurrent rx flushes, each of which may
+        * need to be retried if there is an outstanding descriptor fetch */
+       for (i = 0; i < EFX_FLUSH_POLL_COUNT; ++i) {
+               rx_pending = tx_pending = 0;
+               efx_for_each_rx_queue(rx_queue, efx) {
+                       if (rx_queue->flushed == FLUSH_PENDING)
+                               ++rx_pending;
+               }
+               efx_for_each_rx_queue(rx_queue, efx) {
+                       if (rx_pending == EFX_RX_FLUSH_COUNT)
+                               break;
+                       if (rx_queue->flushed == FLUSH_FAILED ||
+                           rx_queue->flushed == FLUSH_NONE) {
+                               efx_flush_rx_queue(rx_queue);
+                               ++rx_pending;
+                       }
+               }
+               efx_for_each_tx_queue(tx_queue, efx) {
+                       if (tx_queue->flushed != FLUSH_DONE)
+                               ++tx_pending;
+               }
+
+               if (rx_pending == 0 && tx_pending == 0)
+                       return 0;
+
+               msleep(EFX_FLUSH_INTERVAL);
+               efx_poll_flush_events(efx);
+       }
+
+       /* Mark the queues as all flushed. We're going to return failure
+        * leading to a reset, or fake up success anyway */
+       efx_for_each_tx_queue(tx_queue, efx) {
+               if (tx_queue->flushed != FLUSH_DONE)
+                       EFX_ERR(efx, "tx queue %d flush command timed out\n",
+                               tx_queue->queue);
+               tx_queue->flushed = FLUSH_DONE;
+       }
+       efx_for_each_rx_queue(rx_queue, efx) {
+               if (rx_queue->flushed != FLUSH_DONE)
+                       EFX_ERR(efx, "rx queue %d flush command timed out\n",
+                               rx_queue->queue);
+               rx_queue->flushed = FLUSH_DONE;
+       }
+
+       if (EFX_WORKAROUND_7803(efx))
+               return 0;
+
+       return -ETIMEDOUT;
+}
+
+/**************************************************************************
+ *
+ * Hardware interrupts
+ * The hardware interrupt handler does very little work; all the event
+ * queue processing is carried out by per-channel tasklets.
+ *
+ **************************************************************************/
+
+/* Enable/disable/generate interrupts */
+static inline void efx_nic_interrupts(struct efx_nic *efx,
+                                     bool enabled, bool force)
+{
+       efx_oword_t int_en_reg_ker;
+
+       EFX_POPULATE_OWORD_2(int_en_reg_ker,
+                            FRF_AZ_KER_INT_KER, force,
+                            FRF_AZ_DRV_INT_EN_KER, enabled);
+       efx_writeo(efx, &int_en_reg_ker, FR_AZ_INT_EN_KER);
+}
+
+void efx_nic_enable_interrupts(struct efx_nic *efx)
+{
+       struct efx_channel *channel;
+
+       EFX_ZERO_OWORD(*((efx_oword_t *) efx->irq_status.addr));
+       wmb(); /* Ensure interrupt vector is clear before interrupts enabled */
+
+       /* Enable interrupts */
+       efx_nic_interrupts(efx, true, false);
+
+       /* Force processing of all the channels to get the EVQ RPTRs up to
+          date */
+       efx_for_each_channel(channel, efx)
+               efx_schedule_channel(channel);
+}
+
+void efx_nic_disable_interrupts(struct efx_nic *efx)
+{
+       /* Disable interrupts */
+       efx_nic_interrupts(efx, false, false);
+}
+
+/* Generate a test interrupt
+ * Interrupt must already have been enabled, otherwise nasty things
+ * may happen.
+ */
+void efx_nic_generate_interrupt(struct efx_nic *efx)
+{
+       efx_nic_interrupts(efx, true, true);
+}
+
+/* Process a fatal interrupt
+ * Disable bus mastering ASAP and schedule a reset
+ */
+irqreturn_t efx_nic_fatal_interrupt(struct efx_nic *efx)
+{
+       struct falcon_nic_data *nic_data = efx->nic_data;
+       efx_oword_t *int_ker = efx->irq_status.addr;
+       efx_oword_t fatal_intr;
+       int error, mem_perr;
+
+       efx_reado(efx, &fatal_intr, FR_AZ_FATAL_INTR_KER);
+       error = EFX_OWORD_FIELD(fatal_intr, FRF_AZ_FATAL_INTR);
+
+       EFX_ERR(efx, "SYSTEM ERROR " EFX_OWORD_FMT " status "
+               EFX_OWORD_FMT ": %s\n", EFX_OWORD_VAL(*int_ker),
+               EFX_OWORD_VAL(fatal_intr),
+               error ? "disabling bus mastering" : "no recognised error");
+       if (error == 0)
+               goto out;
+
+       /* If this is a memory parity error dump which blocks are offending */
+       mem_perr = EFX_OWORD_FIELD(fatal_intr, FRF_AZ_MEM_PERR_INT_KER);
+       if (mem_perr) {
+               efx_oword_t reg;
+               efx_reado(efx, &reg, FR_AZ_MEM_STAT);
+               EFX_ERR(efx, "SYSTEM ERROR: memory parity error "
+                       EFX_OWORD_FMT "\n", EFX_OWORD_VAL(reg));
+       }
+
+       /* Disable both devices */
+       pci_clear_master(efx->pci_dev);
+       if (efx_nic_is_dual_func(efx))
+               pci_clear_master(nic_data->pci_dev2);
+       efx_nic_disable_interrupts(efx);
+
+       /* Count errors and reset or disable the NIC accordingly */
+       if (efx->int_error_count == 0 ||
+           time_after(jiffies, efx->int_error_expire)) {
+               efx->int_error_count = 0;
+               efx->int_error_expire =
+                       jiffies + EFX_INT_ERROR_EXPIRE * HZ;
+       }
+       if (++efx->int_error_count < EFX_MAX_INT_ERRORS) {
+               EFX_ERR(efx, "SYSTEM ERROR - reset scheduled\n");
+               efx_schedule_reset(efx, RESET_TYPE_INT_ERROR);
+       } else {
+               EFX_ERR(efx, "SYSTEM ERROR - max number of errors seen."
+                       "NIC will be disabled\n");
+               efx_schedule_reset(efx, RESET_TYPE_DISABLE);
+       }
+out:
+       return IRQ_HANDLED;
+}
+
+/* Handle a legacy interrupt
+ * Acknowledges the interrupt and schedule event queue processing.
+ */
+static irqreturn_t efx_legacy_interrupt(int irq, void *dev_id)
+{
+       struct efx_nic *efx = dev_id;
+       efx_oword_t *int_ker = efx->irq_status.addr;
+       irqreturn_t result = IRQ_NONE;
+       struct efx_channel *channel;
+       efx_dword_t reg;
+       u32 queues;
+       int syserr;
+
+       /* Read the ISR which also ACKs the interrupts */
+       efx_readd(efx, &reg, FR_BZ_INT_ISR0);
+       queues = EFX_EXTRACT_DWORD(reg, 0, 31);
+
+       /* Check to see if we have a serious error condition */
+       syserr = EFX_OWORD_FIELD(*int_ker, FSF_AZ_NET_IVEC_FATAL_INT);
+       if (unlikely(syserr))
+               return efx_nic_fatal_interrupt(efx);
+
+       /* Schedule processing of any interrupting queues */
+       efx_for_each_channel(channel, efx) {
+               if ((queues & 1) ||
+                   efx_event_present(
+                           efx_event(channel, channel->eventq_read_ptr))) {
+                               efx_schedule_channel(channel);
+                       result = IRQ_HANDLED;
+               }
+               queues >>= 1;
+       }
+
+       if (result == IRQ_HANDLED) {
+               efx->last_irq_cpu = raw_smp_processor_id();
+               EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_DWORD_FMT "\n",
+                         irq, raw_smp_processor_id(), EFX_DWORD_VAL(reg));
+       }
+
+       return result;
+}
+
+/* Handle an MSI interrupt
+ *
+ * Handle an MSI hardware interrupt.  This routine schedules event
+ * queue processing.  No interrupt acknowledgement cycle is necessary.
+ * Also, we never need to check that the interrupt is for us, since
+ * MSI interrupts cannot be shared.
+ */
+static irqreturn_t efx_msi_interrupt(int irq, void *dev_id)
+{
+       struct efx_channel *channel = dev_id;
+       struct efx_nic *efx = channel->efx;
+       efx_oword_t *int_ker = efx->irq_status.addr;
+       int syserr;
+
+       efx->last_irq_cpu = raw_smp_processor_id();
+       EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_OWORD_FMT "\n",
+                 irq, raw_smp_processor_id(), EFX_OWORD_VAL(*int_ker));
+
+       /* Check to see if we have a serious error condition */
+       syserr = EFX_OWORD_FIELD(*int_ker, FSF_AZ_NET_IVEC_FATAL_INT);
+       if (unlikely(syserr))
+               return efx_nic_fatal_interrupt(efx);
+
+       /* Schedule processing of the channel */
+       efx_schedule_channel(channel);
+
+       return IRQ_HANDLED;
+}
+
+
+/* Setup RSS indirection table.
+ * This maps from the hash value of the packet to RXQ
+ */
+static void efx_setup_rss_indir_table(struct efx_nic *efx)
+{
+       int i = 0;
+       unsigned long offset;
+       efx_dword_t dword;
+
+       if (efx_nic_rev(efx) < EFX_REV_FALCON_B0)
+               return;
+
+       for (offset = FR_BZ_RX_INDIRECTION_TBL;
+            offset < FR_BZ_RX_INDIRECTION_TBL + 0x800;
+            offset += 0x10) {
+               EFX_POPULATE_DWORD_1(dword, FRF_BZ_IT_QUEUE,
+                                    i % efx->n_rx_queues);
+               efx_writed(efx, &dword, offset);
+               i++;
+       }
+}
+
+/* Hook interrupt handler(s)
+ * Try MSI and then legacy interrupts.
+ */
+int efx_nic_init_interrupt(struct efx_nic *efx)
+{
+       struct efx_channel *channel;
+       int rc;
+
+       if (!EFX_INT_MODE_USE_MSI(efx)) {
+               irq_handler_t handler;
+               if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
+                       handler = efx_legacy_interrupt;
+               else
+                       handler = falcon_legacy_interrupt_a1;
+
+               rc = request_irq(efx->legacy_irq, handler, IRQF_SHARED,
+                                efx->name, efx);
+               if (rc) {
+                       EFX_ERR(efx, "failed to hook legacy IRQ %d\n",
+                               efx->pci_dev->irq);
+                       goto fail1;
+               }
+               return 0;
+       }
+
+       /* Hook MSI or MSI-X interrupt */
+       efx_for_each_channel(channel, efx) {
+               rc = request_irq(channel->irq, efx_msi_interrupt,
+                                IRQF_PROBE_SHARED, /* Not shared */
+                                channel->name, channel);
+               if (rc) {
+                       EFX_ERR(efx, "failed to hook IRQ %d\n", channel->irq);
+                       goto fail2;
+               }
+       }
+
+       return 0;
+
+ fail2:
+       efx_for_each_channel(channel, efx)
+               free_irq(channel->irq, channel);
+ fail1:
+       return rc;
+}
+
+void efx_nic_fini_interrupt(struct efx_nic *efx)
+{
+       struct efx_channel *channel;
+       efx_oword_t reg;
+
+       /* Disable MSI/MSI-X interrupts */
+       efx_for_each_channel(channel, efx) {
+               if (channel->irq)
+                       free_irq(channel->irq, channel);
+       }
+
+       /* ACK legacy interrupt */
+       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
+               efx_reado(efx, &reg, FR_BZ_INT_ISR0);
+       else
+               falcon_irq_ack_a1(efx);
+
+       /* Disable legacy interrupt */
+       if (efx->legacy_irq)
+               free_irq(efx->legacy_irq, efx);
+}
+
+u32 efx_nic_fpga_ver(struct efx_nic *efx)
+{
+       efx_oword_t altera_build;
+       efx_reado(efx, &altera_build, FR_AZ_ALTERA_BUILD);
+       return EFX_OWORD_FIELD(altera_build, FRF_AZ_ALTERA_BUILD_VER);
+}
+
+void efx_nic_init_common(struct efx_nic *efx)
+{
+       efx_oword_t temp;
+
+       /* Set positions of descriptor caches in SRAM. */
+       EFX_POPULATE_OWORD_1(temp, FRF_AZ_SRM_TX_DC_BASE_ADR,
+                            efx->type->tx_dc_base / 8);
+       efx_writeo(efx, &temp, FR_AZ_SRM_TX_DC_CFG);
+       EFX_POPULATE_OWORD_1(temp, FRF_AZ_SRM_RX_DC_BASE_ADR,
+                            efx->type->rx_dc_base / 8);
+       efx_writeo(efx, &temp, FR_AZ_SRM_RX_DC_CFG);
+
+       /* Set TX descriptor cache size. */
+       BUILD_BUG_ON(TX_DC_ENTRIES != (8 << TX_DC_ENTRIES_ORDER));
+       EFX_POPULATE_OWORD_1(temp, FRF_AZ_TX_DC_SIZE, TX_DC_ENTRIES_ORDER);
+       efx_writeo(efx, &temp, FR_AZ_TX_DC_CFG);
+
+       /* Set RX descriptor cache size.  Set low watermark to size-8, as
+        * this allows most efficient prefetching.
+        */
+       BUILD_BUG_ON(RX_DC_ENTRIES != (8 << RX_DC_ENTRIES_ORDER));
+       EFX_POPULATE_OWORD_1(temp, FRF_AZ_RX_DC_SIZE, RX_DC_ENTRIES_ORDER);
+       efx_writeo(efx, &temp, FR_AZ_RX_DC_CFG);
+       EFX_POPULATE_OWORD_1(temp, FRF_AZ_RX_DC_PF_LWM, RX_DC_ENTRIES - 8);
+       efx_writeo(efx, &temp, FR_AZ_RX_DC_PF_WM);
+
+       /* Program INT_KER address */
+       EFX_POPULATE_OWORD_2(temp,
+                            FRF_AZ_NORM_INT_VEC_DIS_KER,
+                            EFX_INT_MODE_USE_MSI(efx),
+                            FRF_AZ_INT_ADR_KER, efx->irq_status.dma_addr);
+       efx_writeo(efx, &temp, FR_AZ_INT_ADR_KER);
+
+       /* Enable all the genuinely fatal interrupts.  (They are still
+        * masked by the overall interrupt mask, controlled by
+        * falcon_interrupts()).
+        *
+        * Note: All other fatal interrupts are enabled
+        */
+       EFX_POPULATE_OWORD_3(temp,
+                            FRF_AZ_ILL_ADR_INT_KER_EN, 1,
+                            FRF_AZ_RBUF_OWN_INT_KER_EN, 1,
+                            FRF_AZ_TBUF_OWN_INT_KER_EN, 1);
+       EFX_INVERT_OWORD(temp);
+       efx_writeo(efx, &temp, FR_AZ_FATAL_INTR_KER);
+
+       efx_setup_rss_indir_table(efx);
+
+       /* Disable the ugly timer-based TX DMA backoff and allow TX DMA to be
+        * controlled by the RX FIFO fill level. Set arbitration to one pkt/Q.
+        */
+       efx_reado(efx, &temp, FR_AZ_TX_RESERVED);
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_RX_SPACER, 0xfe);
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_RX_SPACER_EN, 1);
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_ONE_PKT_PER_Q, 1);
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_PUSH_EN, 0);
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_DIS_NON_IP_EV, 1);
+       /* Enable SW_EV to inherit in char driver - assume harmless here */
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_SOFT_EVT_EN, 1);
+       /* Prefetch threshold 2 => fetch when descriptor cache half empty */
+       EFX_SET_OWORD_FIELD(temp, FRF_AZ_TX_PREF_THRESHOLD, 2);
+       /* Squash TX of packets of 16 bytes or less */
+       if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
+               EFX_SET_OWORD_FIELD(temp, FRF_BZ_TX_FLUSH_MIN_LEN_EN, 1);
+       efx_writeo(efx, &temp, FR_AZ_TX_RESERVED);
+}