via-velocity.c

Linux Kernel 2.6.9 for OMAP1710

/**
 *	velocity_free_rings	-	free PCI ring pointers
 *	@vptr: Velocity to free from
 *
 *	Clean up the PCI ring buffers allocated to this velocity.
 */

static void velocity_free_rings(struct velocity_info *vptr)
{
	int size;

	size = vptr->options.numrx * sizeof(struct rx_desc) + 
	       vptr->options.numtx * sizeof(struct tx_desc) * vptr->num_txq;

	pci_free_consistent(vptr->pdev, size, vptr->rd_ring, vptr->rd_pool_dma);

	size = vptr->options.numtx * PKT_BUF_SZ * vptr->num_txq;

	pci_free_consistent(vptr->pdev, size, vptr->tx_bufs, vptr->tx_bufs_dma);
}

static inline void velocity_give_many_rx_descs(struct velocity_info *vptr)
{
	struct mac_regs *regs = vptr->mac_regs;
	int avail, dirty, unusable;

	/*
	 * RD number must be equal to 4X per hardware spec
	 * (programming guide rev 1.20, p.13)
	 */
	if (vptr->rd_filled < 4)
		return;

	wmb();

	unusable = vptr->rd_filled & 0x0003;
	dirty = vptr->rd_dirty - unusable;
	for (avail = vptr->rd_filled & 0xfffc; avail; avail--) {
		dirty = (dirty > 0) ? dirty - 1 : vptr->options.numrx - 1;
		vptr->rd_ring[dirty].rdesc0.owner = OWNED_BY_NIC;
	}

	writew(vptr->rd_filled & 0xfffc, &regs->RBRDU);
	vptr->rd_filled = unusable;
}

static int velocity_rx_refill(struct velocity_info *vptr)
{
	int dirty = vptr->rd_dirty, done = 0, ret = 0;

	do {
		struct rx_desc *rd = vptr->rd_ring + dirty;

		/* Fine for an all zero Rx desc at init time as well */
		if (rd->rdesc0.owner == OWNED_BY_NIC)
			break;

		if (!vptr->rd_info[dirty].skb) {
			ret = velocity_alloc_rx_buf(vptr, dirty);
			if (ret < 0)
				break;
		}
		done++;
		dirty = (dirty < vptr->options.numrx - 1) ? dirty + 1 : 0;	
	} while (dirty != vptr->rd_curr);

	if (done) {
		vptr->rd_dirty = dirty;
		vptr->rd_filled += done;
		velocity_give_many_rx_descs(vptr);
	}

	return ret;
}

/**
 *	velocity_init_rd_ring	-	set up receive ring
 *	@vptr: velocity to configure
 *
 *	Allocate and set up the receive buffers for each ring slot and
 *	assign them to the network adapter.
 */

static int velocity_init_rd_ring(struct velocity_info *vptr)
{
	int ret = -ENOMEM;
	unsigned int rsize = sizeof(struct velocity_rd_info) * 
					vptr->options.numrx;

	vptr->rd_info = kmalloc(rsize, GFP_KERNEL);
	if(vptr->rd_info == NULL)
		goto out;
	memset(vptr->rd_info, 0, rsize);

	vptr->rd_filled = vptr->rd_dirty = vptr->rd_curr = 0;

	ret = velocity_rx_refill(vptr);
	if (ret < 0) {
		VELOCITY_PRT(MSG_LEVEL_ERR, KERN_ERR
			"%s: failed to allocate RX buffer.\n", vptr->dev->name);
		velocity_free_rd_ring(vptr);
	}
out:
	return ret;
}

/**
 *	velocity_free_rd_ring	-	free receive ring
 *	@vptr: velocity to clean up
 *
 *	Free the receive buffers for each ring slot and any
 *	attached socket buffers that need to go away.
 */

static void velocity_free_rd_ring(struct velocity_info *vptr)
{
	int i;

	if (vptr->rd_info == NULL)
		return;

	for (i = 0; i < vptr->options.numrx; i++) {
		struct velocity_rd_info *rd_info = &(vptr->rd_info[i]);

		if (!rd_info->skb)
			continue;
		pci_unmap_single(vptr->pdev, rd_info->skb_dma, vptr->rx_buf_sz,
				 PCI_DMA_FROMDEVICE);
		rd_info->skb_dma = (dma_addr_t) NULL;

		dev_kfree_skb(rd_info->skb);
		rd_info->skb = NULL;
	}

	kfree(vptr->rd_info);
	vptr->rd_info = NULL;
}

/**
 *	velocity_init_td_ring	-	set up transmit ring
 *	@vptr:	velocity
 *
 *	Set up the transmit ring and chain the ring pointers together.
 *	Returns zero on success or a negative posix errno code for
 *	failure.
 */
 
static int velocity_init_td_ring(struct velocity_info *vptr)
{
	int i, j;
	dma_addr_t curr;
	struct tx_desc *td;
	struct velocity_td_info *td_info;
	unsigned int tsize = sizeof(struct velocity_td_info) * 
					vptr->options.numtx;

	/* Init the TD ring entries */
	for (j = 0; j < vptr->num_txq; j++) {
		curr = vptr->td_pool_dma[j];

		vptr->td_infos[j] = kmalloc(tsize, GFP_KERNEL);
		if(vptr->td_infos[j] == NULL)
		{
			while(--j >= 0)
				kfree(vptr->td_infos[j]);
			return -ENOMEM;
		}
		memset(vptr->td_infos[j], 0, tsize);

		for (i = 0; i < vptr->options.numtx; i++, curr += sizeof(struct tx_desc)) {
			td = &(vptr->td_rings[j][i]);
			td_info = &(vptr->td_infos[j][i]);
			td_info->buf = vptr->tx_bufs +
				(j * vptr->options.numtx + i) * PKT_BUF_SZ;
			td_info->buf_dma = vptr->tx_bufs_dma +
				(j * vptr->options.numtx + i) * PKT_BUF_SZ;
		}
		vptr->td_tail[j] = vptr->td_curr[j] = vptr->td_used[j] = 0;
	}
	return 0;
}

/*
 *	FIXME: could we merge this with velocity_free_tx_buf ?
 */

static void velocity_free_td_ring_entry(struct velocity_info *vptr,
							 int q, int n)
{
	struct velocity_td_info * td_info = &(vptr->td_infos[q][n]);
	int i;
	
	if (td_info == NULL)
		return;
		
	if (td_info->skb) {
		for (i = 0; i < td_info->nskb_dma; i++)
		{
			if (td_info->skb_dma[i]) {
				pci_unmap_single(vptr->pdev, td_info->skb_dma[i], 
					td_info->skb->len, PCI_DMA_TODEVICE);
				td_info->skb_dma[i] = (dma_addr_t) NULL;
			}
		}
		dev_kfree_skb(td_info->skb);
		td_info->skb = NULL;
	}
}

/**
 *	velocity_free_td_ring	-	free td ring
 *	@vptr: velocity
 *
 *	Free up the transmit ring for this particular velocity adapter.
 *	We free the ring contents but not the ring itself.
 */
 
static void velocity_free_td_ring(struct velocity_info *vptr)
{
	int i, j;

	for (j = 0; j < vptr->num_txq; j++) {
		if (vptr->td_infos[j] == NULL)
			continue;
		for (i = 0; i < vptr->options.numtx; i++) {
			velocity_free_td_ring_entry(vptr, j, i);

		}
		if (vptr->td_infos[j]) {
			kfree(vptr->td_infos[j]);
			vptr->td_infos[j] = NULL;
		}
	}
}

/**
 *	velocity_rx_srv		-	service RX interrupt
 *	@vptr: velocity
 *	@status: adapter status (unused)
 *
 *	Walk the receive ring of the velocity adapter and remove
 *	any received packets from the receive queue. Hand the ring
 *	slots back to the adapter for reuse.
 */
 
static int velocity_rx_srv(struct velocity_info *vptr, int status)
{
	struct net_device_stats *stats = &vptr->stats;
	int rd_curr = vptr->rd_curr;
	int works = 0;

	do {
		struct rx_desc *rd = vptr->rd_ring + rd_curr;

		if (!vptr->rd_info[rd_curr].skb)
			break;

		if (rd->rdesc0.owner == OWNED_BY_NIC)
			break;

		rmb();

		/*
		 *	Don't drop CE or RL error frame although RXOK is off
		 */
		if ((rd->rdesc0.RSR & RSR_RXOK) || (!(rd->rdesc0.RSR & RSR_RXOK) && (rd->rdesc0.RSR & (RSR_CE | RSR_RL)))) {
			if (velocity_receive_frame(vptr, rd_curr) < 0)
				stats->rx_dropped++;
		} else {
			if (rd->rdesc0.RSR & RSR_CRC)
				stats->rx_crc_errors++;
			if (rd->rdesc0.RSR & RSR_FAE)
				stats->rx_frame_errors++;

			stats->rx_dropped++;
		}

		rd->inten = 1;

		vptr->dev->last_rx = jiffies;

		rd_curr++;
		if (rd_curr >= vptr->options.numrx)
			rd_curr = 0;
	} while (++works <= 15);

	vptr->rd_curr = rd_curr;

	if (works > 0 && velocity_rx_refill(vptr) < 0) {
		VELOCITY_PRT(MSG_LEVEL_ERR, KERN_ERR
			"%s: rx buf allocation failure\n", vptr->dev->name);
	}

	VAR_USED(stats);
	return works;
}

/**
 *	velocity_rx_csum	-	checksum process
 *	@rd: receive packet descriptor
 *	@skb: network layer packet buffer
 *
 *	Process the status bits for the received packet and determine
 *	if the checksum was computed and verified by the hardware
 */
 
static inline void velocity_rx_csum(struct rx_desc *rd, struct sk_buff *skb)
{
	skb->ip_summed = CHECKSUM_NONE;

	if (rd->rdesc1.CSM & CSM_IPKT) {
		if (rd->rdesc1.CSM & CSM_IPOK) {
			if ((rd->rdesc1.CSM & CSM_TCPKT) || 
					(rd->rdesc1.CSM & CSM_UDPKT)) {
				if (!(rd->rdesc1.CSM & CSM_TUPOK)) {
					return;
				}
			}
			skb->ip_summed = CHECKSUM_UNNECESSARY;
		}
	}
}

/**
 *	velocity_rx_copy	-	in place Rx copy for small packets
 *	@rx_skb: network layer packet buffer candidate
 *	@pkt_size: received data size
 *	@rd: receive packet descriptor
 *	@dev: network device
 *
 *	Replace the current skb that is scheduled for Rx processing by a
 *	shorter, immediatly allocated skb, if the received packet is small
 *	enough. This function returns a negative value if the received
 *	packet is too big or if memory is exhausted.
 */
static inline int velocity_rx_copy(struct sk_buff **rx_skb, int pkt_size,
				   struct velocity_info *vptr)
{
	int ret = -1;

	if (pkt_size < rx_copybreak) {
		struct sk_buff *new_skb;

		new_skb = dev_alloc_skb(pkt_size + 2);
		if (new_skb) {
			new_skb->dev = vptr->dev;
			new_skb->ip_summed = rx_skb[0]->ip_summed;

			if (vptr->flags & VELOCITY_FLAGS_IP_ALIGN)
				skb_reserve(new_skb, 2);

			memcpy(new_skb->data, rx_skb[0]->tail, pkt_size);
			*rx_skb = new_skb;
			ret = 0;
		}
		
	}
	return ret;
}

/**
 *	velocity_iph_realign	-	IP header alignment
 *	@vptr: velocity we are handling
 *	@skb: network layer packet buffer
 *	@pkt_size: received data size
 *
 *	Align IP header on a 2 bytes boundary. This behavior can be
 *	configured by the user.
 */
static inline void velocity_iph_realign(struct velocity_info *vptr,
					struct sk_buff *skb, int pkt_size)
{
	/* FIXME - memmove ? */
	if (vptr->flags & VELOCITY_FLAGS_IP_ALIGN) {
		int i;

		for (i = pkt_size; i >= 0; i--)
			*(skb->data + i + 2) = *(skb->data + i);
		skb_reserve(skb, 2);
	}
}

/**
 *	velocity_receive_frame	-	received packet processor
 *	@vptr: velocity we are handling
 *	@idx: ring index
 *	
 *	A packet has arrived. We process the packet and if appropriate
 *	pass the frame up the network stack
 */
 
static int velocity_receive_frame(struct velocity_info *vptr, int idx)
{
	void (*pci_action)(struct pci_dev *, dma_addr_t, size_t, int);
	struct net_device_stats *stats = &vptr->stats;
	struct velocity_rd_info *rd_info = &(vptr->rd_info[idx]);
	struct rx_desc *rd = &(vptr->rd_ring[idx]);
	int pkt_len = rd->rdesc0.len;
	struct sk_buff *skb;

	if (rd->rdesc0.RSR & (RSR_STP | RSR_EDP)) {
		VELOCITY_PRT(MSG_LEVEL_VERBOSE, KERN_ERR " %s : the received frame span multple RDs.\n", vptr->dev->name);
		stats->rx_length_errors++;
		return -EINVAL;
	}

	if (rd->rdesc0.RSR & RSR_MAR)
		vptr->stats.multicast++;

	skb = rd_info->skb;
	skb->dev = vptr->dev;

	pci_dma_sync_single_for_cpu(vptr->pdev, rd_info->skb_dma,
				    vptr->rx_buf_sz, PCI_DMA_FROMDEVICE);

	/*
	 *	Drop frame not meeting IEEE 802.3
	 */
	 
	if (vptr->flags & VELOCITY_FLAGS_VAL_PKT_LEN) {
		if (rd->rdesc0.RSR & RSR_RL) {
			stats->rx_length_errors++;
			return -EINVAL;
		}
	}

	pci_action = pci_dma_sync_single_for_device;

	velocity_rx_csum(rd, skb);

	if (velocity_rx_copy(&skb, pkt_len, vptr) < 0) {
		velocity_iph_realign(vptr, skb, pkt_len);
		pci_action = pci_unmap_single;
		rd_info->skb = NULL;
	}

	pci_action(vptr->pdev, rd_info->skb_dma, vptr->rx_buf_sz,
		   PCI_DMA_FROMDEVICE);

	skb_put(skb, pkt_len - 4);
	skb->protocol = eth_type_trans(skb, skb->dev);	

	stats->rx_bytes += pkt_len;
	netif_rx(skb);

	return 0;
}

/**
 *	velocity_alloc_rx_buf	-	allocate aligned receive buffer
 *	@vptr: velocity
 *	@idx: ring index
 *
 *	Allocate a new full sized buffer for the reception of a frame and
 *	map it into PCI space for the hardware to use. The hardware
 *	requires *64* byte alignment of the buffer which makes life
 *	less fun than would be ideal.
 */
 
static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
{
	struct rx_desc *rd = &(vptr->rd_ring[idx]);
	struct velocity_rd_info *rd_info = &(vptr->rd_info[idx]);

	rd_info->skb = dev_alloc_skb(vptr->rx_buf_sz + 64);
	if (rd_info->skb == NULL)
		return -ENOMEM;

	/*
	 *	Do the gymnastics to get the buffer head for data at
	 *	64byte alignment.
	 */
	skb_reserve(rd_info->skb, (unsigned long) rd_info->skb->tail & 63);
	rd_info->skb->dev = vptr->dev;
	rd_info->skb_dma = pci_map_single(vptr->pdev, rd_info->skb->tail, vptr->rx_buf_sz, PCI_DMA_FROMDEVICE);
	
	/*
	 *	Fill in the descriptor to match
 	 */	
 	 
	*((u32 *) & (rd->rdesc0)) = 0;
	rd->len = cpu_to_le32(vptr->rx_buf_sz);
	rd->inten = 1;
	rd->pa_low = cpu_to_le32(rd_info->skb_dma);
	rd->pa_high = 0;
	return 0;
}

/**
 *	tx_srv		-	transmit interrupt service
 *	@vptr; Velocity
 *	@status:
 *