From: Julian Wiedmann <jwi@linux.ibm.com>
Subject: s390/qeth: add statistics for consumed buffer elements
Patch-mainline: v4.19-rc1
Git-commit: d2a274b25be7218f8400037868a756640e8a4b0d
References: FATE#326350, LTC#169511, bsc#1113509
Summary: qeth: performance improvements
Description: This adds recent functional and performance improvements for the
qeth network driver.
Primarily this brings Scatter-Gather support for HiperSockets,
reduced CPU consumption in the L3 IPv4 transmit path for OSA,
improved Promiscuous Mode performance due to IFF_UNICAST_FLT,
support for Scatter-Gather on z/VM virtual NICs, and
support for delayed GRO flushing.
For sanity & stability reasons, this effectively constitutes a
backport of the qeth device driver from 4.19 mainline.
Upstream-Description:
s390/qeth: add statistics for consumed buffer elements
Nowadays an skb fragment typically spans over multiple pages. So replace
the obsolete, SG-only 'fragments' counter with one that tracks the
consumed buffer elements. This is what actually matters for performance.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Acked-by: Petr Tesarik <ptesarik@suse.com>
---
drivers/s390/net/qeth_core.h | 2 +-
drivers/s390/net/qeth_core_main.c | 4 ++--
drivers/s390/net/qeth_l2_main.c | 13 +++++++------
drivers/s390/net/qeth_l3_main.c | 28 ++++++++++++++--------------
4 files changed, 24 insertions(+), 23 deletions(-)
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -102,6 +102,7 @@ struct qeth_dbf_info {
struct qeth_perf_stats {
unsigned int bufs_rec;
unsigned int bufs_sent;
+ unsigned int buf_elements_sent;
unsigned int skbs_sent_pack;
unsigned int bufs_sent_pack;
@@ -135,7 +136,6 @@ struct qeth_perf_stats {
unsigned int large_send_bytes;
unsigned int large_send_cnt;
unsigned int sg_skbs_sent;
- unsigned int sg_frags_sent;
/* initial values when measuring starts */
unsigned long initial_rx_packets;
unsigned long initial_tx_packets;
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -5968,7 +5968,7 @@ static struct {
{"tx skbs packing"},
{"tx buffers packing"},
{"tx sg skbs"},
- {"tx sg frags"},
+ {"tx buffer elements"},
/* 10 */{"rx sg skbs"},
{"rx sg frags"},
{"rx sg page allocs"},
@@ -6027,7 +6027,7 @@ void qeth_core_get_ethtool_stats(struct
data[6] = card->perf_stats.skbs_sent_pack;
data[7] = card->perf_stats.bufs_sent_pack;
data[8] = card->perf_stats.sg_skbs_sent;
- data[9] = card->perf_stats.sg_frags_sent;
+ data[9] = card->perf_stats.buf_elements_sent;
data[10] = card->perf_stats.sg_skbs_rx;
data[11] = card->perf_stats.sg_frags_rx;
data[12] = card->perf_stats.sg_alloc_page_rx;
--- a/drivers/s390/net/qeth_l2_main.c
+++ b/drivers/s390/net/qeth_l2_main.c
@@ -672,10 +672,11 @@ static int qeth_l2_xmit_osa(struct qeth_
int ipv)
{
int push_len = sizeof(struct qeth_hdr);
- unsigned int elements, nr_frags;
unsigned int hdr_elements = 0;
struct qeth_hdr *hdr = NULL;
unsigned int hd_len = 0;
+ unsigned int elements;
+ bool is_sg;
int rc;
/* fix hardware limitation: as long as we do not have sbal
@@ -693,7 +694,6 @@ static int qeth_l2_xmit_osa(struct qeth_
if (rc)
return rc;
}
- nr_frags = skb_shinfo(skb)->nr_frags;
rc = skb_cow_head(skb, push_len);
if (rc)
@@ -720,15 +720,16 @@ static int qeth_l2_xmit_osa(struct qeth_
}
elements += hdr_elements;
+ is_sg = skb_is_nonlinear(skb);
/* TODO: remove the skb_orphan() once TX completion is fast enough */
skb_orphan(skb);
rc = qeth_do_send_packet(card, queue, skb, hdr, 0, hd_len, elements);
out:
if (!rc) {
- if (card->options.performance_stats && nr_frags) {
- card->perf_stats.sg_skbs_sent++;
- /* nr_frags + skb->data */
- card->perf_stats.sg_frags_sent += nr_frags + 1;
+ if (card->options.performance_stats) {
+ card->perf_stats.buf_elements_sent += elements;
+ if (is_sg)
+ card->perf_stats.sg_skbs_sent++;
}
} else {
if (hd_len)
--- a/drivers/s390/net/qeth_l3_main.c
+++ b/drivers/s390/net/qeth_l3_main.c
@@ -2172,12 +2172,13 @@ static int qeth_l3_xmit_offload(struct q
int cast_type)
{
const unsigned int hw_hdr_len = sizeof(struct qeth_hdr);
- unsigned int frame_len, nr_frags;
unsigned char eth_hdr[ETH_HLEN];
unsigned int hdr_elements = 0;
struct qeth_hdr *hdr = NULL;
int elements, push_len, rc;
unsigned int hd_len = 0;
+ unsigned int frame_len;
+ bool is_sg;
/* compress skb to fit into one IO buffer: */
if (!qeth_get_elements_no(card, skb, 0, 0)) {
@@ -2200,7 +2201,6 @@ static int qeth_l3_xmit_offload(struct q
skb_copy_from_linear_data(skb, eth_hdr, ETH_HLEN);
skb_pull(skb, ETH_HLEN);
frame_len = skb->len;
- nr_frags = skb_shinfo(skb)->nr_frags;
push_len = qeth_push_hdr(skb, &hdr, hw_hdr_len);
if (push_len < 0)
@@ -2223,6 +2223,7 @@ static int qeth_l3_xmit_offload(struct q
else
qeth_l3_fill_header(card, hdr, skb, ipv, cast_type, frame_len);
+ is_sg = skb_is_nonlinear(skb);
if (IS_IQD(card)) {
rc = qeth_do_send_packet_fast(queue, skb, hdr, 0, hd_len);
} else {
@@ -2233,10 +2234,10 @@ static int qeth_l3_xmit_offload(struct q
}
out:
if (!rc) {
- if (card->options.performance_stats && nr_frags) {
- card->perf_stats.sg_skbs_sent++;
- /* nr_frags + skb->data */
- card->perf_stats.sg_frags_sent += nr_frags + 1;
+ if (card->options.performance_stats) {
+ card->perf_stats.buf_elements_sent += elements;
+ if (is_sg)
+ card->perf_stats.sg_skbs_sent++;
}
} else {
if (!push_len)
@@ -2254,14 +2255,14 @@ out:
static int qeth_l3_xmit(struct qeth_card *card, struct sk_buff *skb,
struct qeth_qdio_out_q *queue, int ipv, int cast_type)
{
- unsigned int hd_len, nr_frags;
int elements, len, rc;
__be16 *tag;
struct qeth_hdr *hdr = NULL;
int hdr_elements = 0;
struct sk_buff *new_skb = NULL;
int tx_bytes = skb->len;
- bool use_tso;
+ unsigned int hd_len;
+ bool use_tso, is_sg;
/* Ignore segment size from skb_is_gso(), 1 page is always used. */
use_tso = skb_is_gso(skb) &&
@@ -2303,7 +2304,6 @@ static int qeth_l3_xmit(struct qeth_card
if (rc)
goto out;
}
- nr_frags = skb_shinfo(new_skb)->nr_frags;
if (use_tso) {
hdr = skb_push(new_skb, sizeof(struct qeth_hdr_tso));
@@ -2340,6 +2340,8 @@ static int qeth_l3_xmit(struct qeth_card
rc = -EINVAL;
goto out;
}
+
+ is_sg = skb_is_nonlinear(new_skb);
rc = qeth_do_send_packet(card, queue, new_skb, hdr, hd_len, hd_len,
elements);
out:
@@ -2347,15 +2349,13 @@ out:
if (new_skb != skb)
dev_kfree_skb_any(skb);
if (card->options.performance_stats) {
+ card->perf_stats.buf_elements_sent += elements;
+ if (is_sg)
+ card->perf_stats.sg_skbs_sent++;
if (use_tso) {
card->perf_stats.large_send_bytes += tx_bytes;
card->perf_stats.large_send_cnt++;
}
- if (nr_frags) {
- card->perf_stats.sg_skbs_sent++;
- /* nr_frags + skb->data */
- card->perf_stats.sg_frags_sent += nr_frags + 1;
- }
}
} else {
if (new_skb != skb)