Commit 54a50941 authored by Julian Wiedmann's avatar Julian Wiedmann Committed by David S. Miller

s390/qeth: stop/wake TX queues based on their fill level

Current xmit code only stops the txq after attempting to fill an
IO buffer that hasn't been TX-completed yet. In many-connection
scenarios, this can result in frequent rejected TX attempts, requeuing
of skbs with NETDEV_TX_BUSY and extra overhead.

Now that we have a proper 1-to-1 relation between stack-side txqs and
our HW Queues, overhaul the stop/wake logic so that the xmit code
stops the txq as needed.
Given that we might map multiple skbs into a single buffer, it's crucial
to ensure that the queue always provides an _entirely_ empty IO buffer.
Otherwise large skbs (eg TSO) might not fit into the last available
buffer. So whenever qeth_do_send_packet() first utilizes an _empty_
buffer, it updates & checks the used_buffers count.

This now ensures that an skb passed to qeth_xmit() can always be mapped
into an IO buffer, so remove all of the -EBUSY roll-back handling in the
TX path. We preserve the minimal safety-checks ("Is this IO buffer
really available?"), just in case some nasty future bug ever attempts to
corrupt an in-use buffer.
Signed-off-by: default avatarJulian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent e6c15b5f
...@@ -481,6 +481,7 @@ struct qeth_out_q_stats { ...@@ -481,6 +481,7 @@ struct qeth_out_q_stats {
u64 skbs_linearized_fail; u64 skbs_linearized_fail;
u64 tso_bytes; u64 tso_bytes;
u64 packing_mode_switch; u64 packing_mode_switch;
u64 stopped;
/* rtnl_link_stats64 */ /* rtnl_link_stats64 */
u64 tx_packets; u64 tx_packets;
...@@ -511,6 +512,11 @@ struct qeth_qdio_out_q { ...@@ -511,6 +512,11 @@ struct qeth_qdio_out_q {
atomic_t set_pci_flags_count; atomic_t set_pci_flags_count;
}; };
static inline bool qeth_out_queue_is_full(struct qeth_qdio_out_q *queue)
{
return atomic_read(&queue->used_buffers) >= QDIO_MAX_BUFFERS_PER_Q;
}
struct qeth_qdio_info { struct qeth_qdio_info {
atomic_t state; atomic_t state;
/* input */ /* input */
......
...@@ -3367,7 +3367,6 @@ static void qeth_flush_buffers(struct qeth_qdio_out_q *queue, int index, ...@@ -3367,7 +3367,6 @@ static void qeth_flush_buffers(struct qeth_qdio_out_q *queue, int index,
qdio_flags = QDIO_FLAG_SYNC_OUTPUT; qdio_flags = QDIO_FLAG_SYNC_OUTPUT;
if (atomic_read(&queue->set_pci_flags_count)) if (atomic_read(&queue->set_pci_flags_count))
qdio_flags |= QDIO_FLAG_PCI_OUT; qdio_flags |= QDIO_FLAG_PCI_OUT;
atomic_add(count, &queue->used_buffers);
rc = do_QDIO(CARD_DDEV(queue->card), qdio_flags, rc = do_QDIO(CARD_DDEV(queue->card), qdio_flags,
queue->queue_no, index, count); queue->queue_no, index, count);
if (rc) { if (rc) {
...@@ -3407,7 +3406,6 @@ static void qeth_check_outbound_queue(struct qeth_qdio_out_q *queue) ...@@ -3407,7 +3406,6 @@ static void qeth_check_outbound_queue(struct qeth_qdio_out_q *queue)
* do_send_packet. So, we check if there is a * do_send_packet. So, we check if there is a
* packing buffer to be flushed here. * packing buffer to be flushed here.
*/ */
netif_stop_subqueue(queue->card->dev, queue->queue_no);
index = queue->next_buf_to_fill; index = queue->next_buf_to_fill;
q_was_packing = queue->do_pack; q_was_packing = queue->do_pack;
/* queue->do_pack may change */ /* queue->do_pack may change */
...@@ -3535,7 +3533,7 @@ static void qeth_qdio_output_handler(struct ccw_device *ccwdev, ...@@ -3535,7 +3533,7 @@ static void qeth_qdio_output_handler(struct ccw_device *ccwdev,
struct qeth_qdio_out_q *queue = card->qdio.out_qs[__queue]; struct qeth_qdio_out_q *queue = card->qdio.out_qs[__queue];
struct qeth_qdio_out_buffer *buffer; struct qeth_qdio_out_buffer *buffer;
struct net_device *dev = card->dev; struct net_device *dev = card->dev;
u16 txq; struct netdev_queue *txq;
int i; int i;
QETH_CARD_TEXT(card, 6, "qdouhdl"); QETH_CARD_TEXT(card, 6, "qdouhdl");
...@@ -3590,8 +3588,15 @@ static void qeth_qdio_output_handler(struct ccw_device *ccwdev, ...@@ -3590,8 +3588,15 @@ static void qeth_qdio_output_handler(struct ccw_device *ccwdev,
if (card->info.type != QETH_CARD_TYPE_IQD) if (card->info.type != QETH_CARD_TYPE_IQD)
qeth_check_outbound_queue(queue); qeth_check_outbound_queue(queue);
txq = IS_IQD(card) ? qeth_iqd_translate_txq(dev, __queue) : __queue; if (IS_IQD(card))
netif_wake_subqueue(dev, txq); __queue = qeth_iqd_translate_txq(dev, __queue);
txq = netdev_get_tx_queue(dev, __queue);
/* xmit may have observed the full-condition, but not yet stopped the
* txq. In which case the code below won't trigger. So before returning,
* xmit will re-check the txq's fill level and wake it up if needed.
*/
if (netif_tx_queue_stopped(txq) && !qeth_out_queue_is_full(queue))
netif_tx_wake_queue(txq);
} }
/** /**
...@@ -3845,11 +3850,13 @@ static void __qeth_fill_buffer(struct sk_buff *skb, ...@@ -3845,11 +3850,13 @@ static void __qeth_fill_buffer(struct sk_buff *skb,
* from qeth_core_header_cache. * from qeth_core_header_cache.
* @offset: when mapping the skb, start at skb->data + offset * @offset: when mapping the skb, start at skb->data + offset
* @hd_len: if > 0, build a dedicated header element of this size * @hd_len: if > 0, build a dedicated header element of this size
* flush: Prepare the buffer to be flushed, regardless of its fill level.
*/ */
static int qeth_fill_buffer(struct qeth_qdio_out_q *queue, static int qeth_fill_buffer(struct qeth_qdio_out_q *queue,
struct qeth_qdio_out_buffer *buf, struct qeth_qdio_out_buffer *buf,
struct sk_buff *skb, struct qeth_hdr *hdr, struct sk_buff *skb, struct qeth_hdr *hdr,
unsigned int offset, unsigned int hd_len) unsigned int offset, unsigned int hd_len,
bool flush)
{ {
struct qdio_buffer *buffer = buf->buffer; struct qdio_buffer *buffer = buf->buffer;
bool is_first_elem = true; bool is_first_elem = true;
...@@ -3878,8 +3885,8 @@ static int qeth_fill_buffer(struct qeth_qdio_out_q *queue, ...@@ -3878,8 +3885,8 @@ static int qeth_fill_buffer(struct qeth_qdio_out_q *queue,
QETH_TXQ_STAT_INC(queue, skbs_pack); QETH_TXQ_STAT_INC(queue, skbs_pack);
/* If the buffer still has free elements, keep using it. */ /* If the buffer still has free elements, keep using it. */
if (buf->next_element_to_fill < if (!flush && buf->next_element_to_fill <
QETH_MAX_BUFFER_ELEMENTS(queue->card)) QETH_MAX_BUFFER_ELEMENTS(queue->card))
return 0; return 0;
} }
...@@ -3896,15 +3903,31 @@ static int qeth_do_send_packet_fast(struct qeth_qdio_out_q *queue, ...@@ -3896,15 +3903,31 @@ static int qeth_do_send_packet_fast(struct qeth_qdio_out_q *queue,
{ {
int index = queue->next_buf_to_fill; int index = queue->next_buf_to_fill;
struct qeth_qdio_out_buffer *buffer = queue->bufs[index]; struct qeth_qdio_out_buffer *buffer = queue->bufs[index];
struct netdev_queue *txq;
bool stopped = false;
/* /* Just a sanity check, the wake/stop logic should ensure that we always
* check if buffer is empty to make sure that we do not 'overtake' * get a free buffer.
* ourselves and try to fill a buffer that is already primed
*/ */
if (atomic_read(&buffer->state) != QETH_QDIO_BUF_EMPTY) if (atomic_read(&buffer->state) != QETH_QDIO_BUF_EMPTY)
return -EBUSY; return -EBUSY;
qeth_fill_buffer(queue, buffer, skb, hdr, offset, hd_len);
txq = netdev_get_tx_queue(queue->card->dev, skb_get_queue_mapping(skb));
if (atomic_inc_return(&queue->used_buffers) >= QDIO_MAX_BUFFERS_PER_Q) {
/* If a TX completion happens right _here_ and misses to wake
* the txq, then our re-check below will catch the race.
*/
QETH_TXQ_STAT_INC(queue, stopped);
netif_tx_stop_queue(txq);
stopped = true;
}
qeth_fill_buffer(queue, buffer, skb, hdr, offset, hd_len, stopped);
qeth_flush_buffers(queue, index, 1); qeth_flush_buffers(queue, index, 1);
if (stopped && !qeth_out_queue_is_full(queue))
netif_tx_start_queue(txq);
return 0; return 0;
} }
...@@ -3914,6 +3937,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue, ...@@ -3914,6 +3937,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,
int elements_needed) int elements_needed)
{ {
struct qeth_qdio_out_buffer *buffer; struct qeth_qdio_out_buffer *buffer;
struct netdev_queue *txq;
bool stopped = false;
int start_index; int start_index;
int flush_count = 0; int flush_count = 0;
int do_pack = 0; int do_pack = 0;
...@@ -3925,14 +3950,17 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue, ...@@ -3925,14 +3950,17 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,
QETH_OUT_Q_LOCKED) != QETH_OUT_Q_UNLOCKED); QETH_OUT_Q_LOCKED) != QETH_OUT_Q_UNLOCKED);
start_index = queue->next_buf_to_fill; start_index = queue->next_buf_to_fill;
buffer = queue->bufs[queue->next_buf_to_fill]; buffer = queue->bufs[queue->next_buf_to_fill];
/*
* check if buffer is empty to make sure that we do not 'overtake' /* Just a sanity check, the wake/stop logic should ensure that we always
* ourselves and try to fill a buffer that is already primed * get a free buffer.
*/ */
if (atomic_read(&buffer->state) != QETH_QDIO_BUF_EMPTY) { if (atomic_read(&buffer->state) != QETH_QDIO_BUF_EMPTY) {
atomic_set(&queue->state, QETH_OUT_Q_UNLOCKED); atomic_set(&queue->state, QETH_OUT_Q_UNLOCKED);
return -EBUSY; return -EBUSY;
} }
txq = netdev_get_tx_queue(card->dev, skb_get_queue_mapping(skb));
/* check if we need to switch packing state of this queue */ /* check if we need to switch packing state of this queue */
qeth_switch_to_packing_if_needed(queue); qeth_switch_to_packing_if_needed(queue);
if (queue->do_pack) { if (queue->do_pack) {
...@@ -3947,8 +3975,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue, ...@@ -3947,8 +3975,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,
(queue->next_buf_to_fill + 1) % (queue->next_buf_to_fill + 1) %
QDIO_MAX_BUFFERS_PER_Q; QDIO_MAX_BUFFERS_PER_Q;
buffer = queue->bufs[queue->next_buf_to_fill]; buffer = queue->bufs[queue->next_buf_to_fill];
/* we did a step forward, so check buffer state
* again */ /* We stepped forward, so sanity-check again: */
if (atomic_read(&buffer->state) != if (atomic_read(&buffer->state) !=
QETH_QDIO_BUF_EMPTY) { QETH_QDIO_BUF_EMPTY) {
qeth_flush_buffers(queue, start_index, qeth_flush_buffers(queue, start_index,
...@@ -3961,8 +3989,18 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue, ...@@ -3961,8 +3989,18 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,
} }
} }
flush_count += qeth_fill_buffer(queue, buffer, skb, hdr, offset, if (buffer->next_element_to_fill == 0 &&
hd_len); atomic_inc_return(&queue->used_buffers) >= QDIO_MAX_BUFFERS_PER_Q) {
/* If a TX completion happens right _here_ and misses to wake
* the txq, then our re-check below will catch the race.
*/
QETH_TXQ_STAT_INC(queue, stopped);
netif_tx_stop_queue(txq);
stopped = true;
}
flush_count += qeth_fill_buffer(queue, buffer, skb, hdr, offset, hd_len,
stopped);
if (flush_count) if (flush_count)
qeth_flush_buffers(queue, start_index, flush_count); qeth_flush_buffers(queue, start_index, flush_count);
else if (!atomic_read(&queue->set_pci_flags_count)) else if (!atomic_read(&queue->set_pci_flags_count))
...@@ -3993,6 +4031,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue, ...@@ -3993,6 +4031,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,
if (do_pack) if (do_pack)
QETH_TXQ_STAT_ADD(queue, bufs_pack, flush_count); QETH_TXQ_STAT_ADD(queue, bufs_pack, flush_count);
if (stopped && !qeth_out_queue_is_full(queue))
netif_tx_start_queue(txq);
return rc; return rc;
} }
EXPORT_SYMBOL_GPL(qeth_do_send_packet); EXPORT_SYMBOL_GPL(qeth_do_send_packet);
...@@ -4079,9 +4119,6 @@ int qeth_xmit(struct qeth_card *card, struct sk_buff *skb, ...@@ -4079,9 +4119,6 @@ int qeth_xmit(struct qeth_card *card, struct sk_buff *skb,
} else { } else {
if (!push_len) if (!push_len)
kmem_cache_free(qeth_core_header_cache, hdr); kmem_cache_free(qeth_core_header_cache, hdr);
if (rc == -EBUSY)
/* roll back to ETH header */
skb_pull(skb, push_len);
} }
return rc; return rc;
} }
......
...@@ -38,6 +38,7 @@ static const struct qeth_stats txq_stats[] = { ...@@ -38,6 +38,7 @@ static const struct qeth_stats txq_stats[] = {
QETH_TXQ_STAT("linearized+error skbs", skbs_linearized_fail), QETH_TXQ_STAT("linearized+error skbs", skbs_linearized_fail),
QETH_TXQ_STAT("TSO bytes", tso_bytes), QETH_TXQ_STAT("TSO bytes", tso_bytes),
QETH_TXQ_STAT("Packing mode switches", packing_mode_switch), QETH_TXQ_STAT("Packing mode switches", packing_mode_switch),
QETH_TXQ_STAT("Queue stopped", stopped),
}; };
static const struct qeth_stats card_stats[] = { static const struct qeth_stats card_stats[] = {
......
...@@ -607,11 +607,8 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb, ...@@ -607,11 +607,8 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb,
int rc; int rc;
if (IS_IQD(card)) if (IS_IQD(card))
queue = card->qdio.out_qs[qeth_iqd_translate_txq(dev, txq)]; txq = qeth_iqd_translate_txq(dev, txq);
else queue = card->qdio.out_qs[txq];
queue = card->qdio.out_qs[txq];
netif_stop_subqueue(dev, txq);
if (IS_OSN(card)) if (IS_OSN(card))
rc = qeth_l2_xmit_osn(card, skb, queue); rc = qeth_l2_xmit_osn(card, skb, queue);
...@@ -622,15 +619,11 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb, ...@@ -622,15 +619,11 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb,
if (!rc) { if (!rc) {
QETH_TXQ_STAT_INC(queue, tx_packets); QETH_TXQ_STAT_INC(queue, tx_packets);
QETH_TXQ_STAT_ADD(queue, tx_bytes, tx_bytes); QETH_TXQ_STAT_ADD(queue, tx_bytes, tx_bytes);
netif_wake_subqueue(dev, txq);
return NETDEV_TX_OK; return NETDEV_TX_OK;
} else if (rc == -EBUSY) { }
return NETDEV_TX_BUSY;
} /* else fall through */
QETH_TXQ_STAT_INC(queue, tx_dropped); QETH_TXQ_STAT_INC(queue, tx_dropped);
kfree_skb(skb); kfree_skb(skb);
netif_wake_subqueue(dev, txq);
return NETDEV_TX_OK; return NETDEV_TX_OK;
} }
......
...@@ -2036,7 +2036,6 @@ static void qeth_l3_fixup_headers(struct sk_buff *skb) ...@@ -2036,7 +2036,6 @@ static void qeth_l3_fixup_headers(struct sk_buff *skb)
static int qeth_l3_xmit(struct qeth_card *card, struct sk_buff *skb, static int qeth_l3_xmit(struct qeth_card *card, struct sk_buff *skb,
struct qeth_qdio_out_q *queue, int ipv, int cast_type) struct qeth_qdio_out_q *queue, int ipv, int cast_type)
{ {
unsigned char eth_hdr[ETH_HLEN];
unsigned int hw_hdr_len; unsigned int hw_hdr_len;
int rc; int rc;
...@@ -2046,17 +2045,10 @@ static int qeth_l3_xmit(struct qeth_card *card, struct sk_buff *skb, ...@@ -2046,17 +2045,10 @@ static int qeth_l3_xmit(struct qeth_card *card, struct sk_buff *skb,
rc = skb_cow_head(skb, hw_hdr_len - ETH_HLEN); rc = skb_cow_head(skb, hw_hdr_len - ETH_HLEN);
if (rc) if (rc)
return rc; return rc;
skb_copy_from_linear_data(skb, eth_hdr, ETH_HLEN);
skb_pull(skb, ETH_HLEN); skb_pull(skb, ETH_HLEN);
qeth_l3_fixup_headers(skb); qeth_l3_fixup_headers(skb);
rc = qeth_xmit(card, skb, queue, ipv, cast_type, qeth_l3_fill_header); return qeth_xmit(card, skb, queue, ipv, cast_type, qeth_l3_fill_header);
if (rc == -EBUSY) {
/* roll back to ETH header */
skb_push(skb, ETH_HLEN);
skb_copy_to_linear_data(skb, eth_hdr, ETH_HLEN);
}
return rc;
} }
static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb, static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb,
...@@ -2091,8 +2083,6 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb, ...@@ -2091,8 +2083,6 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb,
if (cast_type == RTN_BROADCAST && !card->info.broadcast_capable) if (cast_type == RTN_BROADCAST && !card->info.broadcast_capable)
goto tx_drop; goto tx_drop;
netif_stop_subqueue(dev, txq);
if (ipv == 4 || IS_IQD(card)) if (ipv == 4 || IS_IQD(card))
rc = qeth_l3_xmit(card, skb, queue, ipv, cast_type); rc = qeth_l3_xmit(card, skb, queue, ipv, cast_type);
else else
...@@ -2102,16 +2092,12 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb, ...@@ -2102,16 +2092,12 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb,
if (!rc) { if (!rc) {
QETH_TXQ_STAT_INC(queue, tx_packets); QETH_TXQ_STAT_INC(queue, tx_packets);
QETH_TXQ_STAT_ADD(queue, tx_bytes, tx_bytes); QETH_TXQ_STAT_ADD(queue, tx_bytes, tx_bytes);
netif_wake_subqueue(dev, txq);
return NETDEV_TX_OK; return NETDEV_TX_OK;
} else if (rc == -EBUSY) { }
return NETDEV_TX_BUSY;
} /* else fall through */
tx_drop: tx_drop:
QETH_TXQ_STAT_INC(queue, tx_dropped); QETH_TXQ_STAT_INC(queue, tx_dropped);
kfree_skb(skb); kfree_skb(skb);
netif_wake_subqueue(dev, txq);
return NETDEV_TX_OK; return NETDEV_TX_OK;
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment