s2io.c

linux 内核源代码

MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);


/* Module Loadable parameters. */
S2IO_PARM_INT(tx_fifo_num, 1);
S2IO_PARM_INT(rx_ring_num, 1);


S2IO_PARM_INT(rx_ring_mode, 1);
S2IO_PARM_INT(use_continuous_tx_intrs, 1);
S2IO_PARM_INT(rmac_pause_time, 0x100);
S2IO_PARM_INT(mc_pause_threshold_q0q3, 187);
S2IO_PARM_INT(mc_pause_threshold_q4q7, 187);
S2IO_PARM_INT(shared_splits, 0);
S2IO_PARM_INT(tmac_util_period, 5);
S2IO_PARM_INT(rmac_util_period, 5);
S2IO_PARM_INT(l3l4hdr_size, 128);
/* Frequency of Rx desc syncs expressed as power of 2 */
S2IO_PARM_INT(rxsync_frequency, 3);
/* Interrupt type. Values can be 0(INTA), 2(MSI_X) */
S2IO_PARM_INT(intr_type, 2);
/* Large receive offload feature */
static unsigned int lro_enable;
module_param_named(lro, lro_enable, uint, 0);

/* Max pkts to be aggregated by LRO at one time. If not specified,
 * aggregation happens until we hit max IP pkt size(64K)
 */
S2IO_PARM_INT(lro_max_pkts, 0xFFFF);
S2IO_PARM_INT(indicate_max_pkts, 0);

S2IO_PARM_INT(napi, 1);
S2IO_PARM_INT(ufo, 0);
S2IO_PARM_INT(vlan_tag_strip, NO_STRIP_IN_PROMISC);

static unsigned int tx_fifo_len[MAX_TX_FIFOS] =
    {DEFAULT_FIFO_0_LEN, [1 ...(MAX_TX_FIFOS - 1)] = DEFAULT_FIFO_1_7_LEN};
static unsigned int rx_ring_sz[MAX_RX_RINGS] =
    {[0 ...(MAX_RX_RINGS - 1)] = SMALL_BLK_CNT};
static unsigned int rts_frm_len[MAX_RX_RINGS] =
    {[0 ...(MAX_RX_RINGS - 1)] = 0 };

module_param_array(tx_fifo_len, uint, NULL, 0);
module_param_array(rx_ring_sz, uint, NULL, 0);
module_param_array(rts_frm_len, uint, NULL, 0);

/*
 * S2IO device table.
 * This table lists all the devices that this driver supports.
 */
static struct pci_device_id s2io_tbl[] __devinitdata = {
	{PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_S2IO_WIN,
	 PCI_ANY_ID, PCI_ANY_ID},
	{PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_S2IO_UNI,
	 PCI_ANY_ID, PCI_ANY_ID},
	{PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_HERC_WIN,
         PCI_ANY_ID, PCI_ANY_ID},
        {PCI_VENDOR_ID_S2IO, PCI_DEVICE_ID_HERC_UNI,
         PCI_ANY_ID, PCI_ANY_ID},
	{0,}
};

MODULE_DEVICE_TABLE(pci, s2io_tbl);

static struct pci_error_handlers s2io_err_handler = {
	.error_detected = s2io_io_error_detected,
	.slot_reset = s2io_io_slot_reset,
	.resume = s2io_io_resume,
};

static struct pci_driver s2io_driver = {
      .name = "S2IO",
      .id_table = s2io_tbl,
      .probe = s2io_init_nic,
      .remove = __devexit_p(s2io_rem_nic),
      .err_handler = &s2io_err_handler,
};

/* A simplifier macro used both by init and free shared_mem Fns(). */
#define TXD_MEM_PAGE_CNT(len, per_each) ((len+per_each - 1) / per_each)

/**
 * init_shared_mem - Allocation and Initialization of Memory
 * @nic: Device private variable.
 * Description: The function allocates all the memory areas shared
 * between the NIC and the driver. This includes Tx descriptors,
 * Rx descriptors and the statistics block.
 */

static int init_shared_mem(struct s2io_nic *nic)
{
	u32 size;
	void *tmp_v_addr, *tmp_v_addr_next;
	dma_addr_t tmp_p_addr, tmp_p_addr_next;
	struct RxD_block *pre_rxd_blk = NULL;
	int i, j, blk_cnt;
	int lst_size, lst_per_page;
	struct net_device *dev = nic->dev;
	unsigned long tmp;
	struct buffAdd *ba;

	struct mac_info *mac_control;
	struct config_param *config;
	unsigned long long mem_allocated = 0;

	mac_control = &nic->mac_control;
	config = &nic->config;


	/* Allocation and initialization of TXDLs in FIOFs */
	size = 0;
	for (i = 0; i < config->tx_fifo_num; i++) {
		size += config->tx_cfg[i].fifo_len;
	}
	if (size > MAX_AVAILABLE_TXDS) {
		DBG_PRINT(ERR_DBG, "s2io: Requested TxDs too high, ");
		DBG_PRINT(ERR_DBG, "Requested: %d, max supported: 8192\n", size);
		return -EINVAL;
	}

	lst_size = (sizeof(struct TxD) * config->max_txds);
	lst_per_page = PAGE_SIZE / lst_size;

	for (i = 0; i < config->tx_fifo_num; i++) {
		int fifo_len = config->tx_cfg[i].fifo_len;
		int list_holder_size = fifo_len * sizeof(struct list_info_hold);
		mac_control->fifos[i].list_info = kzalloc(list_holder_size,
							  GFP_KERNEL);
		if (!mac_control->fifos[i].list_info) {
			DBG_PRINT(INFO_DBG,
				  "Malloc failed for list_info\n");
			return -ENOMEM;
		}
		mem_allocated += list_holder_size;
	}
	for (i = 0; i < config->tx_fifo_num; i++) {
		int page_num = TXD_MEM_PAGE_CNT(config->tx_cfg[i].fifo_len,
						lst_per_page);
		mac_control->fifos[i].tx_curr_put_info.offset = 0;
		mac_control->fifos[i].tx_curr_put_info.fifo_len =
		    config->tx_cfg[i].fifo_len - 1;
		mac_control->fifos[i].tx_curr_get_info.offset = 0;
		mac_control->fifos[i].tx_curr_get_info.fifo_len =
		    config->tx_cfg[i].fifo_len - 1;
		mac_control->fifos[i].fifo_no = i;
		mac_control->fifos[i].nic = nic;
		mac_control->fifos[i].max_txds = MAX_SKB_FRAGS + 2;

		for (j = 0; j < page_num; j++) {
			int k = 0;
			dma_addr_t tmp_p;
			void *tmp_v;
			tmp_v = pci_alloc_consistent(nic->pdev,
						     PAGE_SIZE, &tmp_p);
			if (!tmp_v) {
				DBG_PRINT(INFO_DBG,
					  "pci_alloc_consistent ");
				DBG_PRINT(INFO_DBG, "failed for TxDL\n");
				return -ENOMEM;
			}
			/* If we got a zero DMA address(can happen on
			 * certain platforms like PPC), reallocate.
			 * Store virtual address of page we don't want,
			 * to be freed later.
			 */
			if (!tmp_p) {
				mac_control->zerodma_virt_addr = tmp_v;
				DBG_PRINT(INIT_DBG,
				"%s: Zero DMA address for TxDL. ", dev->name);
				DBG_PRINT(INIT_DBG,
				"Virtual address %p\n", tmp_v);
				tmp_v = pci_alloc_consistent(nic->pdev,
						     PAGE_SIZE, &tmp_p);
				if (!tmp_v) {
					DBG_PRINT(INFO_DBG,
					  "pci_alloc_consistent ");
					DBG_PRINT(INFO_DBG, "failed for TxDL\n");
					return -ENOMEM;
				}
				mem_allocated += PAGE_SIZE;
			}
			while (k < lst_per_page) {
				int l = (j * lst_per_page) + k;
				if (l == config->tx_cfg[i].fifo_len)
					break;
				mac_control->fifos[i].list_info[l].list_virt_addr =
				    tmp_v + (k * lst_size);
				mac_control->fifos[i].list_info[l].list_phy_addr =
				    tmp_p + (k * lst_size);
				k++;
			}
		}
	}

	nic->ufo_in_band_v = kcalloc(size, sizeof(u64), GFP_KERNEL);
	if (!nic->ufo_in_band_v)
		return -ENOMEM;
	 mem_allocated += (size * sizeof(u64));

	/* Allocation and initialization of RXDs in Rings */
	size = 0;
	for (i = 0; i < config->rx_ring_num; i++) {
		if (config->rx_cfg[i].num_rxd %
		    (rxd_count[nic->rxd_mode] + 1)) {
			DBG_PRINT(ERR_DBG, "%s: RxD count of ", dev->name);
			DBG_PRINT(ERR_DBG, "Ring%d is not a multiple of ",
				  i);
			DBG_PRINT(ERR_DBG, "RxDs per Block");
			return FAILURE;
		}
		size += config->rx_cfg[i].num_rxd;
		mac_control->rings[i].block_count =
			config->rx_cfg[i].num_rxd /
			(rxd_count[nic->rxd_mode] + 1 );
		mac_control->rings[i].pkt_cnt = config->rx_cfg[i].num_rxd -
			mac_control->rings[i].block_count;
	}
	if (nic->rxd_mode == RXD_MODE_1)
		size = (size * (sizeof(struct RxD1)));
	else
		size = (size * (sizeof(struct RxD3)));

	for (i = 0; i < config->rx_ring_num; i++) {
		mac_control->rings[i].rx_curr_get_info.block_index = 0;
		mac_control->rings[i].rx_curr_get_info.offset = 0;
		mac_control->rings[i].rx_curr_get_info.ring_len =
		    config->rx_cfg[i].num_rxd - 1;
		mac_control->rings[i].rx_curr_put_info.block_index = 0;
		mac_control->rings[i].rx_curr_put_info.offset = 0;
		mac_control->rings[i].rx_curr_put_info.ring_len =
		    config->rx_cfg[i].num_rxd - 1;
		mac_control->rings[i].nic = nic;
		mac_control->rings[i].ring_no = i;

		blk_cnt = config->rx_cfg[i].num_rxd /
				(rxd_count[nic->rxd_mode] + 1);
		/*  Allocating all the Rx blocks */
		for (j = 0; j < blk_cnt; j++) {
			struct rx_block_info *rx_blocks;
			int l;

			rx_blocks = &mac_control->rings[i].rx_blocks[j];
			size = SIZE_OF_BLOCK; //size is always page size
			tmp_v_addr = pci_alloc_consistent(nic->pdev, size,
							  &tmp_p_addr);
			if (tmp_v_addr == NULL) {
				/*
				 * In case of failure, free_shared_mem()
				 * is called, which should free any
				 * memory that was alloced till the
				 * failure happened.
				 */
				rx_blocks->block_virt_addr = tmp_v_addr;
				return -ENOMEM;
			}
			mem_allocated += size;
			memset(tmp_v_addr, 0, size);
			rx_blocks->block_virt_addr = tmp_v_addr;
			rx_blocks->block_dma_addr = tmp_p_addr;
			rx_blocks->rxds = kmalloc(sizeof(struct rxd_info)*
						  rxd_count[nic->rxd_mode],
						  GFP_KERNEL);
			if (!rx_blocks->rxds)
				return -ENOMEM;
			mem_allocated +=
			(sizeof(struct rxd_info)* rxd_count[nic->rxd_mode]);
			for (l=0; l<rxd_count[nic->rxd_mode];l++) {
				rx_blocks->rxds[l].virt_addr =
					rx_blocks->block_virt_addr +
					(rxd_size[nic->rxd_mode] * l);
				rx_blocks->rxds[l].dma_addr =
					rx_blocks->block_dma_addr +
					(rxd_size[nic->rxd_mode] * l);
			}
		}
		/* Interlinking all Rx Blocks */
		for (j = 0; j < blk_cnt; j++) {
			tmp_v_addr =
				mac_control->rings[i].rx_blocks[j].block_virt_addr;
			tmp_v_addr_next =
				mac_control->rings[i].rx_blocks[(j + 1) %
					      blk_cnt].block_virt_addr;
			tmp_p_addr =
				mac_control->rings[i].rx_blocks[j].block_dma_addr;
			tmp_p_addr_next =
				mac_control->rings[i].rx_blocks[(j + 1) %
					      blk_cnt].block_dma_addr;

			pre_rxd_blk = (struct RxD_block *) tmp_v_addr;
			pre_rxd_blk->reserved_2_pNext_RxD_block =
			    (unsigned long) tmp_v_addr_next;
			pre_rxd_blk->pNext_RxD_Blk_physical =
			    (u64) tmp_p_addr_next;
		}
	}
	if (nic->rxd_mode == RXD_MODE_3B) {
		/*
		 * Allocation of Storages for buffer addresses in 2BUFF mode
		 * and the buffers as well.
		 */
		for (i = 0; i < config->rx_ring_num; i++) {
			blk_cnt = config->rx_cfg[i].num_rxd /
			   (rxd_count[nic->rxd_mode]+ 1);
			mac_control->rings[i].ba =
				kmalloc((sizeof(struct buffAdd *) * blk_cnt),
				     GFP_KERNEL);
			if (!mac_control->rings[i].ba)
				return -ENOMEM;
			mem_allocated +=(sizeof(struct buffAdd *) * blk_cnt);
			for (j = 0; j < blk_cnt; j++) {
				int k = 0;
				mac_control->rings[i].ba[j] =
					kmalloc((sizeof(struct buffAdd) *
						(rxd_count[nic->rxd_mode] + 1)),
						GFP_KERNEL);
				if (!mac_control->rings[i].ba[j])
					return -ENOMEM;
				mem_allocated += (sizeof(struct buffAdd) *  \
					(rxd_count[nic->rxd_mode] + 1));
				while (k != rxd_count[nic->rxd_mode]) {
					ba = &mac_control->rings[i].ba[j][k];

					ba->ba_0_org = (void *) kmalloc
					    (BUF0_LEN + ALIGN_SIZE, GFP_KERNEL);
					if (!ba->ba_0_org)
						return -ENOMEM;
					mem_allocated +=
						(BUF0_LEN + ALIGN_SIZE);
					tmp = (unsigned long)ba->ba_0_org;
					tmp += ALIGN_SIZE;
					tmp &= ~((unsigned long) ALIGN_SIZE);
					ba->ba_0 = (void *) tmp;

					ba->ba_1_org = (void *) kmalloc
					    (BUF1_LEN + ALIGN_SIZE, GFP_KERNEL);
					if (!ba->ba_1_org)
						return -ENOMEM;
					mem_allocated
						+= (BUF1_LEN + ALIGN_SIZE);
					tmp = (unsigned long) ba->ba_1_org;
					tmp += ALIGN_SIZE;
					tmp &= ~((unsigned long) ALIGN_SIZE);
					ba->ba_1 = (void *) tmp;
					k++;
				}
			}
		}
	}

	/* Allocation and initialization of Statistics block */
	size = sizeof(struct stat_block);
	mac_control->stats_mem = pci_alloc_consistent
	    (nic->pdev, size, &mac_control->stats_mem_phy);

	if (!mac_control->stats_mem) {
		/*
		 * In case of failure, free_shared_mem() is called, which
		 * should free any memory that was alloced till the
		 * failure happened.
		 */
		return -ENOMEM;
	}
	mem_allocated += size;
	mac_control->stats_mem_sz = size;

	tmp_v_addr = mac_control->stats_mem;
	mac_control->stats_info = (struct stat_block *) tmp_v_addr;
	memset(tmp_v_addr, 0, size);
	DBG_PRINT(INIT_DBG, "%s:Ring Mem PHY: 0x%llx\n", dev->name,
		  (unsigned long long) tmp_p_addr);
	mac_control->stats_info->sw_stat.mem_allocated += mem_allocated;
	return SUCCESS;
}

/**
 * free_shared_mem - Free the allocated Memory
 * @nic:  Device private variable.
 * Description: This function is to free all memory locations allocated by
 * the init_shared_mem() function and return it to the kernel.
 */

static void free_shared_mem(struct s2io_nic *nic)
{
	int i, j, blk_cnt, size;
	u32 ufo_size = 0;
	void *tmp_v_addr;
	dma_addr_t tmp_p_addr;
	struct mac_info *mac_control;
	struct config_param *config;
	int lst_size, lst_per_page;
	struct net_device *dev;
	int page_num = 0;

	if (!nic)
		return;

	dev = nic->dev;

	mac_control = &nic->mac_control;
	config = &nic->config;

	lst_size = (sizeof(struct TxD) * config->max_txds);
	lst_per_page = PAGE_SIZE / lst_size;

	for (i = 0; i < config->tx_fifo_num; i++) {
		ufo_size += config->tx_cfg[i].fifo_len;
		page_num = TXD_MEM_PAGE_CNT(config->tx_cfg[i].fifo_len,
							lst_per_page);
		for (j = 0; j < page_num; j++) {
			int mem_blks = (j * lst_per_page);
			if (!mac_control->fifos[i].list_info)
				return;
			if (!mac_control->fifos[i].list_info[mem_blks].
				 list_virt_addr)
				break;
			pci_free_consistent(nic->pdev, PAGE_SIZE,
					    mac_control->fifos[i].
					    list_info[mem_blks].
					    list_virt_addr,
					    mac_control->fifos[i].
					    list_info[mem_blks].
					    list_phy_addr);
			nic->mac_control.stats_info->sw_stat.mem_freed
						+= PAGE_SIZE;
		}
		/* If we got a zero DMA address during allocation,
		 * free the page now
		 */
		if (mac_control->zerodma_virt_addr) {
			pci_free_consistent(nic->pdev, PAGE_SIZE,
					    mac_control->zerodma_virt_addr,
					    (dma_addr_t)0);
			DBG_PRINT(INIT_DBG,
			  	"%s: Freeing TxDL with zero DMA addr. ",
				dev->name);
			DBG_PRINT(INIT_DBG, "Virtual address %p\n",
				mac_control->zerodma_virt_addr);
			nic->mac_control.stats_info->sw_stat.mem_freed
						+= PAGE_SIZE;
		}
		kfree(mac_control->fifos[i].list_info);
		nic->mac_control.stats_info->sw_stat.mem_freed +=
		(nic->config.tx_cfg[i].fifo_len *sizeof(struct list_info_hold));
	}

	size = SIZE_OF_BLOCK;