s390/qeth: fixes 2019-03-18
Subsystems (4)
cover s390/qeth: fixes 2019-03-18
Hi Dave,
please apply the following three patches to -net. The first two are fixes for minor race conditions in the probe code, while the third one gets dropwatch working (again).
Thanks,
Julian
Julian Wiedmann (3):
s390/qeth: don't erase configuration while probing s390/qeth: fix race when initializing the IP address table s390/qeth: be drop monitor friendly
drivers/s390/net/qeth_core_main.c | 5 ++++- drivers/s390/net/qeth_l2_main.c | 7 +++---- drivers/s390/net/qeth_l3_main.c | 8 ++++---- 3 files changed, 11 insertions(+), 9 deletions(-)
Signature · 1 line
2.16.4
PATCH 1/3 s390/qeth: don't erase configuration while probing
The HW trap and VNICC configuration is exposed via sysfs, and may have already been modified when qeth_l?_probe_device() attempts to initialize them. So (1) initialize the VNICC values a little earlier, and (2) don't bother about the HW trap mode, it was already initialized before.
--- drivers/s390/net/qeth_l2_main.c | 4 ++-- drivers/s390/net/qeth_l3_main.c | 1 - 2 files changed, 2 insertions(+), 3 deletions(-)
drivers/s390/net/qeth_l2_main.c +2 -2
@@ -645,6 +645,8 @@ static int qeth_l2_probe_device(struct ccwgroup_device *gdev)struct qeth_card *card =dev_get_drvdata (&gdev->dev);int rc; +qeth_l2_vnicc_set_defaults (card); +if (gdev->dev.type == &qeth_generic_devtype) { rc =qeth_l2_create_device_attributes (&gdev->dev);if (rc)
@@ -652,8 +654,6 @@ static int qeth_l2_probe_device(struct ccwgroup_device *gdev) }hash_init (card->mac_htable); - card->info.hwtrap =0 ; -qeth_l2_vnicc_set_defaults (card);return 0 ; }
drivers/s390/net/qeth_l3_main.c +0 -1
@@ -2260,7 +2260,6 @@ static int qeth_l3_probe_device(struct ccwgroup_device *gdev) }hash_init (card->ip_htable);hash_init (card->ip_mc_htable); - card->info.hwtrap =0 ;return 0 ; }
Signature · 1 line
2.16.4
PATCH 2/3 s390/qeth: fix race when initializing the IP address table
In reply to Julian WiedmannThe ucast IP table is utilized by some of the L3-specific sysfs attributes that qeth_l3_create_device_attributes() provides. So initialize the table _before_ registering the attributes.
--- drivers/s390/net/qeth_l3_main.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/s390/net/qeth_l3_main.c +3 -1
@@ -2253,12 +2253,14 @@ static int qeth_l3_probe_device(struct ccwgroup_device *gdev)struct qeth_card *card =dev_get_drvdata (&gdev->dev);int rc; +hash_init (card->ip_htable); +if (gdev->dev.type == &qeth_generic_devtype) { rc =qeth_l3_create_device_attributes (&gdev->dev);if (rc)return rc; } -hash_init (card->ip_htable); +hash_init (card->ip_mc_htable);return 0 ; }
Signature · 1 line
2.16.4
PATCH 3/3 s390/qeth: be drop monitor friendly
In reply to Julian WiedmannAs part of the TX completion path, qeth_release_skbs() frees the completed skbs with __skb_queue_purge(). This ends in kfree_skb(), reporting every completed skb as dropped.
On the other hand when dropping an skb in .ndo_start_xmit, we end up calling consume_skb()... where we should be using kfree_skb() so that drop monitors get notified.
Switch the drop/consume logic around, and also don't accumulate dropped packets in the tx_errors statistics.
--- drivers/s390/net/qeth_core_main.c | 5 ++++- drivers/s390/net/qeth_l2_main.c | 3 +-- drivers/s390/net/qeth_l3_main.c | 3 +-- 3 files changed, 6 insertions(+), 5 deletions(-)
drivers/s390/net/qeth_core_main.c +4 -1
@@ -1150,13 +1150,16 @@ static void qeth_notify_skbs(struct qeth_qdio_out_q *q,static void qeth_release_skbs (struct qeth_qdio_out_buffer *buf) { +struct sk_buff *skb; +/* release may never happen from within CQ tasklet scope */ WARN_ON_ONCE (atomic_read (&buf->state) ==QETH_QDIO_BUF_IN_CQ );if (atomic_read (&buf->state) ==QETH_QDIO_BUF_PENDING )qeth_notify_skbs (buf->q, buf,TX_NOTIFY_GENERALERROR ); -__skb_queue_purge (&buf->skb_list); +while ((skb =__skb_dequeue (&buf->skb_list)) !=NULL ) +consume_skb (skb); } static void qeth_clear_output_buffer(struct qeth_qdio_out_q *queue,
drivers/s390/net/qeth_l2_main.c +1 -2
@@ -629,8 +629,7 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb, }/* else fall through */ QETH_TXQ_STAT_INC (queue, tx_dropped); -QETH_TXQ_STAT_INC (queue, tx_errors); -dev_kfree_skb_any (skb); +kfree_skb (skb);netif_wake_queue (dev);return NETDEV_TX_OK ; }
drivers/s390/net/qeth_l3_main.c +1 -2
@@ -2096,8 +2096,7 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb, tx_drop:QETH_TXQ_STAT_INC (queue, tx_dropped); -QETH_TXQ_STAT_INC (queue, tx_errors); -dev_kfree_skb_any (skb); +kfree_skb (skb);netif_wake_queue (dev);return NETDEV_TX_OK ; }
Signature · 1 line
2.16.4
David Miller
maintainerFrom: Julian Wiedmann <[email protected]>
Date: Mon, 18 Mar 2019 16:40:53 +0100
Quoted 3 lines
please apply the following three patches to -net. The first two are fixes for minor race conditions in the probe code, while the third one gets dropwatch working (again).
Series applied, thanks Julian.