tg3: Move per-int tx members to a per-int struct

This patch moves the tx_prod, tx_cons, tx_pending, tx_ring, and
tx_buffers transmit ring device members to a per-interrupt structure.
It also adds a new transmit producer mailbox member (prodmbox) and
converts the code to use it rather than a preprocessor constant.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Matt Carlson 2009-08-28 14:03:21 +00:00 committed by David S. Miller
parent 723344820a
commit f3f3f27e5b
2 changed files with 86 additions and 79 deletions

View File

@ -136,7 +136,7 @@
#define TG3_RX_JMB_MAP_SZ TG3_RX_DMA_TO_MAP_SZ(TG3_RX_JMB_DMA_SZ)
/* minimum number of free TX descriptors required to wake up TX process */
#define TG3_TX_WAKEUP_THRESH(tp) ((tp)->tx_pending / 4)
#define TG3_TX_WAKEUP_THRESH(tnapi) ((tnapi)->tx_pending / 4)
#define TG3_RAW_IP_ALIGN 2
@ -656,7 +656,7 @@ static inline unsigned int tg3_has_work(struct tg3_napi *tnapi)
work_exists = 1;
}
/* check for RX/TX work to do */
if (sblk->idx[0].tx_consumer != tp->tx_cons ||
if (sblk->idx[0].tx_consumer != tnapi->tx_cons ||
sblk->idx[0].rx_producer != tnapi->rx_rcb_ptr)
work_exists = 1;
@ -4269,11 +4269,11 @@ static void tg3_tx_recover(struct tg3 *tp)
spin_unlock(&tp->lock);
}
static inline u32 tg3_tx_avail(struct tg3 *tp)
static inline u32 tg3_tx_avail(struct tg3_napi *tnapi)
{
smp_mb();
return (tp->tx_pending -
((tp->tx_prod - tp->tx_cons) & (TG3_TX_RING_SIZE - 1)));
return tnapi->tx_pending -
((tnapi->tx_prod - tnapi->tx_cons) & (TG3_TX_RING_SIZE - 1));
}
/* Tigon3 never reports partial packet sends. So we do not
@ -4284,10 +4284,10 @@ static void tg3_tx(struct tg3_napi *tnapi)
{
struct tg3 *tp = tnapi->tp;
u32 hw_idx = tnapi->hw_status->idx[0].tx_consumer;
u32 sw_idx = tp->tx_cons;
u32 sw_idx = tnapi->tx_cons;
while (sw_idx != hw_idx) {
struct tx_ring_info *ri = &tp->tx_buffers[sw_idx];
struct tx_ring_info *ri = &tnapi->tx_buffers[sw_idx];
struct sk_buff *skb = ri->skb;
int i, tx_bug = 0;
@ -4303,7 +4303,7 @@ static void tg3_tx(struct tg3_napi *tnapi)
sw_idx = NEXT_TX(sw_idx);
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
ri = &tp->tx_buffers[sw_idx];
ri = &tnapi->tx_buffers[sw_idx];
if (unlikely(ri->skb != NULL || sw_idx == hw_idx))
tx_bug = 1;
sw_idx = NEXT_TX(sw_idx);
@ -4317,7 +4317,7 @@ static void tg3_tx(struct tg3_napi *tnapi)
}
}
tp->tx_cons = sw_idx;
tnapi->tx_cons = sw_idx;
/* Need to make the tx_cons update visible to tg3_start_xmit()
* before checking for netif_queue_stopped(). Without the
@ -4327,10 +4327,10 @@ static void tg3_tx(struct tg3_napi *tnapi)
smp_mb();
if (unlikely(netif_queue_stopped(tp->dev) &&
(tg3_tx_avail(tp) > TG3_TX_WAKEUP_THRESH(tp)))) {
(tg3_tx_avail(tnapi) > TG3_TX_WAKEUP_THRESH(tnapi)))) {
netif_tx_lock(tp->dev);
if (netif_queue_stopped(tp->dev) &&
(tg3_tx_avail(tp) > TG3_TX_WAKEUP_THRESH(tp)))
(tg3_tx_avail(tnapi) > TG3_TX_WAKEUP_THRESH(tnapi)))
netif_wake_queue(tp->dev);
netif_tx_unlock(tp->dev);
}
@ -4668,7 +4668,7 @@ static int tg3_poll_work(struct tg3_napi *tnapi, int work_done, int budget)
}
/* run TX completion thread */
if (sblk->idx[0].tx_consumer != tp->tx_cons) {
if (tnapi->hw_status->idx[0].tx_consumer != tnapi->tx_cons) {
tg3_tx(tnapi);
if (unlikely(tp->tg3_flags & TG3_FLAG_TX_RECOVERY_PENDING))
return work_done;
@ -5044,13 +5044,14 @@ static inline int tg3_40bit_overflow_test(struct tg3 *tp, dma_addr_t mapping,
#endif
}
static void tg3_set_txd(struct tg3 *, int, dma_addr_t, int, u32, u32);
static void tg3_set_txd(struct tg3_napi *, int, dma_addr_t, int, u32, u32);
/* Workaround 4GB and 40-bit hardware DMA bugs. */
static int tigon3_dma_hwbug_workaround(struct tg3 *tp, struct sk_buff *skb,
u32 last_plus_one, u32 *start,
u32 base_flags, u32 mss)
{
struct tg3_napi *tnapi = &tp->napi[0];
struct sk_buff *new_skb;
dma_addr_t new_addr = 0;
u32 entry = *start;
@ -5085,7 +5086,7 @@ static int tigon3_dma_hwbug_workaround(struct tg3 *tp, struct sk_buff *skb,
dev_kfree_skb(new_skb);
new_skb = NULL;
} else {
tg3_set_txd(tp, entry, new_addr, new_skb->len,
tg3_set_txd(tnapi, entry, new_addr, new_skb->len,
base_flags, 1 | (mss << 1));
*start = NEXT_TX(entry);
}
@ -5094,11 +5095,10 @@ static int tigon3_dma_hwbug_workaround(struct tg3 *tp, struct sk_buff *skb,
/* Now clean up the sw ring entries. */
i = 0;
while (entry != last_plus_one) {
if (i == 0) {
tp->tx_buffers[entry].skb = new_skb;
} else {
tp->tx_buffers[entry].skb = NULL;
}
if (i == 0)
tnapi->tx_buffers[entry].skb = new_skb;
else
tnapi->tx_buffers[entry].skb = NULL;
entry = NEXT_TX(entry);
i++;
}
@ -5109,11 +5109,11 @@ static int tigon3_dma_hwbug_workaround(struct tg3 *tp, struct sk_buff *skb,
return ret;
}
static void tg3_set_txd(struct tg3 *tp, int entry,
static void tg3_set_txd(struct tg3_napi *tnapi, int entry,
dma_addr_t mapping, int len, u32 flags,
u32 mss_and_is_end)
{
struct tg3_tx_buffer_desc *txd = &tp->tx_ring[entry];
struct tg3_tx_buffer_desc *txd = &tnapi->tx_ring[entry];
int is_end = (mss_and_is_end & 0x1);
u32 mss = (mss_and_is_end >> 1);
u32 vlan_tag = 0;
@ -5141,6 +5141,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
u32 len, entry, base_flags, mss;
struct skb_shared_info *sp;
dma_addr_t mapping;
struct tg3_napi *tnapi = &tp->napi[0];
len = skb_headlen(skb);
@ -5149,7 +5150,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
* interrupt. Furthermore, IRQ processing runs lockless so we have
* no IRQ context deadlocks to worry about either. Rejoice!
*/
if (unlikely(tg3_tx_avail(tp) <= (skb_shinfo(skb)->nr_frags + 1))) {
if (unlikely(tg3_tx_avail(tnapi) <= (skb_shinfo(skb)->nr_frags + 1))) {
if (!netif_queue_stopped(dev)) {
netif_stop_queue(dev);
@ -5160,7 +5161,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
return NETDEV_TX_BUSY;
}
entry = tp->tx_prod;
entry = tnapi->tx_prod;
base_flags = 0;
mss = 0;
if ((mss = skb_shinfo(skb)->gso_size) != 0) {
@ -5208,9 +5209,9 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
mapping = sp->dma_head;
tp->tx_buffers[entry].skb = skb;
tnapi->tx_buffers[entry].skb = skb;
tg3_set_txd(tp, entry, mapping, len, base_flags,
tg3_set_txd(tnapi, entry, mapping, len, base_flags,
(skb_shinfo(skb)->nr_frags == 0) | (mss << 1));
entry = NEXT_TX(entry);
@ -5225,9 +5226,9 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
len = frag->size;
mapping = sp->dma_maps[i];
tp->tx_buffers[entry].skb = NULL;
tnapi->tx_buffers[entry].skb = NULL;
tg3_set_txd(tp, entry, mapping, len,
tg3_set_txd(tnapi, entry, mapping, len,
base_flags, (i == last) | (mss << 1));
entry = NEXT_TX(entry);
@ -5235,12 +5236,12 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
}
/* Packets are ready, update Tx producer idx local and on card. */
tw32_tx_mbox((MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW), entry);
tw32_tx_mbox(tnapi->prodmbox, entry);
tp->tx_prod = entry;
if (unlikely(tg3_tx_avail(tp) <= (MAX_SKB_FRAGS + 1))) {
tnapi->tx_prod = entry;
if (unlikely(tg3_tx_avail(tnapi) <= (MAX_SKB_FRAGS + 1))) {
netif_stop_queue(dev);
if (tg3_tx_avail(tp) > TG3_TX_WAKEUP_THRESH(tp))
if (tg3_tx_avail(tnapi) > TG3_TX_WAKEUP_THRESH(tnapi))
netif_wake_queue(tp->dev);
}
@ -5258,11 +5259,12 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *, struct net_device *);
static int tg3_tso_bug(struct tg3 *tp, struct sk_buff *skb)
{
struct sk_buff *segs, *nskb;
u32 frag_cnt_est = skb_shinfo(skb)->gso_segs * 3;
/* Estimate the number of fragments in the worst case */
if (unlikely(tg3_tx_avail(tp) <= (skb_shinfo(skb)->gso_segs * 3))) {
if (unlikely(tg3_tx_avail(&tp->napi[0]) <= frag_cnt_est)) {
netif_stop_queue(tp->dev);
if (tg3_tx_avail(tp) <= (skb_shinfo(skb)->gso_segs * 3))
if (tg3_tx_avail(&tp->napi[0]) <= frag_cnt_est)
return NETDEV_TX_BUSY;
netif_wake_queue(tp->dev);
@ -5295,6 +5297,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
struct skb_shared_info *sp;
int would_hit_hwbug;
dma_addr_t mapping;
struct tg3_napi *tnapi = &tp->napi[0];
len = skb_headlen(skb);
@ -5303,7 +5306,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
* interrupt. Furthermore, IRQ processing runs lockless so we have
* no IRQ context deadlocks to worry about either. Rejoice!
*/
if (unlikely(tg3_tx_avail(tp) <= (skb_shinfo(skb)->nr_frags + 1))) {
if (unlikely(tg3_tx_avail(tnapi) <= (skb_shinfo(skb)->nr_frags + 1))) {
if (!netif_queue_stopped(dev)) {
netif_stop_queue(dev);
@ -5314,7 +5317,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
return NETDEV_TX_BUSY;
}
entry = tp->tx_prod;
entry = tnapi->tx_prod;
base_flags = 0;
if (skb->ip_summed == CHECKSUM_PARTIAL)
base_flags |= TXD_FLAG_TCPUDP_CSUM;
@ -5384,7 +5387,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
mapping = sp->dma_head;
tp->tx_buffers[entry].skb = skb;
tnapi->tx_buffers[entry].skb = skb;
would_hit_hwbug = 0;
@ -5393,7 +5396,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
else if (tg3_4g_overflow_test(mapping, len))
would_hit_hwbug = 1;
tg3_set_txd(tp, entry, mapping, len, base_flags,
tg3_set_txd(tnapi, entry, mapping, len, base_flags,
(skb_shinfo(skb)->nr_frags == 0) | (mss << 1));
entry = NEXT_TX(entry);
@ -5409,7 +5412,7 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
len = frag->size;
mapping = sp->dma_maps[i];
tp->tx_buffers[entry].skb = NULL;
tnapi->tx_buffers[entry].skb = NULL;
if (tg3_4g_overflow_test(mapping, len))
would_hit_hwbug = 1;
@ -5418,10 +5421,10 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
would_hit_hwbug = 1;
if (tp->tg3_flags2 & TG3_FLG2_HW_TSO)
tg3_set_txd(tp, entry, mapping, len,
tg3_set_txd(tnapi, entry, mapping, len,
base_flags, (i == last)|(mss << 1));
else
tg3_set_txd(tp, entry, mapping, len,
tg3_set_txd(tnapi, entry, mapping, len,
base_flags, (i == last));
entry = NEXT_TX(entry);
@ -5446,12 +5449,12 @@ static int tg3_start_xmit_dma_bug(struct sk_buff *skb, struct net_device *dev)
}
/* Packets are ready, update Tx producer idx local and on card. */
tw32_tx_mbox((MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW), entry);
tw32_tx_mbox(MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW, entry);
tp->tx_prod = entry;
if (unlikely(tg3_tx_avail(tp) <= (MAX_SKB_FRAGS + 1))) {
tnapi->tx_prod = entry;
if (unlikely(tg3_tx_avail(tnapi) <= (MAX_SKB_FRAGS + 1))) {
netif_stop_queue(dev);
if (tg3_tx_avail(tp) > TG3_TX_WAKEUP_THRESH(tp))
if (tg3_tx_avail(tnapi) > TG3_TX_WAKEUP_THRESH(tnapi))
netif_wake_queue(tp->dev);
}
@ -5522,8 +5525,8 @@ static int tg3_change_mtu(struct net_device *dev, int new_mtu)
static void tg3_rx_prodring_free(struct tg3 *tp,
struct tg3_rx_prodring_set *tpr)
{
struct ring_info *rxp;
int i;
struct ring_info *rxp;
for (i = 0; i < TG3_RX_RING_SIZE; i++) {
rxp = &tpr->rx_std_buffers[i];
@ -5710,13 +5713,14 @@ err_out:
*/
static void tg3_free_rings(struct tg3 *tp)
{
struct tg3_napi *tnapi = &tp->napi[0];
int i;
for (i = 0; i < TG3_TX_RING_SIZE; ) {
struct tx_ring_info *txp;
struct sk_buff *skb;
txp = &tp->tx_buffers[i];
txp = &tnapi->tx_buffers[i];
skb = txp->skb;
if (skb == NULL) {
@ -5751,7 +5755,7 @@ static int tg3_init_rings(struct tg3 *tp)
tg3_free_rings(tp);
/* Zero out all descriptors. */
memset(tp->tx_ring, 0, TG3_TX_RING_BYTES);
memset(tnapi->tx_ring, 0, TG3_TX_RING_BYTES);
tnapi->rx_rcb_ptr = 0;
memset(tnapi->rx_rcb, 0, TG3_RX_RCB_RING_BYTES(tp));
@ -5767,12 +5771,12 @@ static void tg3_free_consistent(struct tg3 *tp)
{
struct tg3_napi *tnapi = &tp->napi[0];
kfree(tp->tx_buffers);
tp->tx_buffers = NULL;
if (tp->tx_ring) {
kfree(tnapi->tx_buffers);
tnapi->tx_buffers = NULL;
if (tnapi->tx_ring) {
pci_free_consistent(tp->pdev, TG3_TX_RING_BYTES,
tp->tx_ring, tp->tx_desc_mapping);
tp->tx_ring = NULL;
tnapi->tx_ring, tnapi->tx_desc_mapping);
tnapi->tx_ring = NULL;
}
if (tnapi->rx_rcb) {
pci_free_consistent(tp->pdev, TG3_RX_RCB_RING_BYTES(tp),
@ -5804,14 +5808,14 @@ static int tg3_alloc_consistent(struct tg3 *tp)
if (tg3_rx_prodring_init(tp, &tp->prodring[0]))
return -ENOMEM;
tp->tx_buffers = kzalloc(sizeof(struct tx_ring_info) *
tnapi->tx_buffers = kzalloc(sizeof(struct tx_ring_info) *
TG3_TX_RING_SIZE, GFP_KERNEL);
if (!tp->tx_buffers)
if (!tnapi->tx_buffers)
goto err_out;
tp->tx_ring = pci_alloc_consistent(tp->pdev, TG3_TX_RING_BYTES,
&tp->tx_desc_mapping);
if (!tp->tx_ring)
tnapi->tx_ring = pci_alloc_consistent(tp->pdev, TG3_TX_RING_BYTES,
&tnapi->tx_desc_mapping);
if (!tnapi->tx_ring)
goto err_out;
tnapi->hw_status = pci_alloc_consistent(tp->pdev,
@ -7094,13 +7098,15 @@ static int tg3_reset_hw(struct tg3 *tp, int reset_phy)
BDINFO_FLAGS_DISABLED);
}
tp->tx_prod = 0;
tp->tx_cons = 0;
tw32_mailbox(MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW, 0);
tp->napi[0].tx_prod = 0;
tp->napi[0].tx_cons = 0;
tw32_tx_mbox(MAILBOX_SNDNIC_PROD_IDX_0 + TG3_64BIT_REG_LOW, 0);
val = MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW;
tw32_mailbox(val, 0);
tg3_set_bdinfo(tp, NIC_SRAM_SEND_RCB,
tp->tx_desc_mapping,
tp->napi[0].tx_desc_mapping,
(TG3_TX_RING_SIZE <<
BDINFO_FLAGS_MAXLEN_SHIFT),
NIC_SRAM_TX_BUFFER_DESC);
@ -9093,7 +9099,7 @@ static void tg3_get_ringparam(struct net_device *dev, struct ethtool_ringparam *
else
ering->rx_jumbo_pending = 0;
ering->tx_pending = tp->tx_pending;
ering->tx_pending = tp->napi[0].tx_pending;
}
static int tg3_set_ringparam(struct net_device *dev, struct ethtool_ringparam *ering)
@ -9123,7 +9129,7 @@ static int tg3_set_ringparam(struct net_device *dev, struct ethtool_ringparam *e
tp->rx_pending > 63)
tp->rx_pending = 63;
tp->rx_jumbo_pending = ering->rx_jumbo_pending;
tp->tx_pending = ering->tx_pending;
tp->napi[0].tx_pending = ering->tx_pending;
if (netif_running(dev)) {
tg3_halt(tp, RESET_KIND_SHUTDOWN, 1);
@ -9928,14 +9934,13 @@ static int tg3_run_loopback(struct tg3 *tp, int loopback_mode)
num_pkts = 0;
tg3_set_txd(tp, tp->tx_prod, map, tx_len, 0, 1);
tg3_set_txd(tnapi, tnapi->tx_prod, map, tx_len, 0, 1);
tp->tx_prod++;
tnapi->tx_prod++;
num_pkts++;
tw32_tx_mbox(MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW,
tp->tx_prod);
tr32_mailbox(MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW);
tw32_tx_mbox(tnapi->prodmbox, tnapi->tx_prod);
tr32_mailbox(tnapi->prodmbox);
udelay(10);
@ -9948,7 +9953,7 @@ static int tg3_run_loopback(struct tg3 *tp, int loopback_mode)
tx_idx = tnapi->hw_status->idx[0].tx_consumer;
rx_idx = rnapi->hw_status->idx[0].rx_producer;
if ((tx_idx == tp->tx_prod) &&
if ((tx_idx == tnapi->tx_prod) &&
(rx_idx == (rx_start_idx + num_pkts)))
break;
}
@ -9956,7 +9961,7 @@ static int tg3_run_loopback(struct tg3 *tp, int loopback_mode)
pci_unmap_single(tp->pdev, map, tx_len, PCI_DMA_TODEVICE);
dev_kfree_skb(skb);
if (tx_idx != tp->tx_prod)
if (tx_idx != tnapi->tx_prod)
goto out;
if (rx_idx != rx_start_idx + num_pkts)
@ -13415,11 +13420,12 @@ static int __devinit tg3_init_one(struct pci_dev *pdev,
tp->rx_pending = TG3_DEF_RX_RING_PENDING;
tp->rx_jumbo_pending = TG3_DEF_RX_JUMBO_RING_PENDING;
tp->tx_pending = TG3_DEF_TX_RING_PENDING;
tp->napi[0].tp = tp;
tp->napi[0].int_mbox = MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW;
tp->napi[0].consmbox = MAILBOX_RCVRET_CON_IDX_0 + TG3_64BIT_REG_LOW;
tp->napi[0].prodmbox = MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW;
tp->napi[0].tx_pending = TG3_DEF_TX_RING_PENDING;
netif_napi_add(dev, &tp->napi[0].napi, tg3_poll, 64);
dev->ethtool_ops = &tg3_ethtool_ops;
dev->watchdog_timeo = TG3_TX_TIMEOUT;

View File

@ -2497,13 +2497,21 @@ struct tg3_napi {
u32 last_tag;
u32 last_irq_tag;
u32 int_mbox;
u32 tx_prod;
u32 tx_cons;
u32 tx_pending;
u32 prodmbox;
u32 consmbox;
u32 rx_rcb_ptr;
struct tg3_rx_buffer_desc *rx_rcb;
struct tg3_tx_buffer_desc *tx_ring;
struct tx_ring_info *tx_buffers;
dma_addr_t status_mapping;
dma_addr_t rx_rcb_mapping;
dma_addr_t tx_desc_mapping;
};
struct tg3 {
@ -2563,13 +2571,6 @@ struct tg3 {
/* begin "tx thread" cacheline section */
void (*write32_tx_mbox) (struct tg3 *, u32,
u32);
u32 tx_prod;
u32 tx_cons;
u32 tx_pending;
struct tg3_tx_buffer_desc *tx_ring;
struct tx_ring_info *tx_buffers;
dma_addr_t tx_desc_mapping;
/* begin "rx thread" cacheline section */
struct tg3_napi napi[TG3_IRQ_MAX_VECS];