s390/qeth: updates 2021-01-28
Subsystems (4)
cover s390/qeth: updates 2021-01-28
Hi Dave & Jakub,
please apply the following patch series for qeth to netdev's net-next tree.
Nothing special, mostly fine-tuning and follow-on cleanups for earlier fixes.
Thanks,
Julian
Julian Wiedmann (5):
s390/qeth: clean up load/remove code for disciplines s390/qeth: remove qeth_get_ip_version()
s390/qeth: pass proto to qeth_l3_get_cast_type() s390/qeth: make cast type selection for af_iucv skbs robust s390/qeth: don't fake a TX completion interrupt after TX error
drivers/s390/net/qeth_core.h | 44 +++++--------- drivers/s390/net/qeth_core_main.c | 97 +++++++++++++++++-------------- drivers/s390/net/qeth_core_sys.c | 10 +--- drivers/s390/net/qeth_l2_main.c | 6 +- drivers/s390/net/qeth_l3_main.c | 90 ++++++++++++++++------------ 5 files changed, 125 insertions(+), 122 deletions(-)
Signature · 1 line
2.17.1
PATCH 1/5 s390/qeth: clean up load/remove code for disciplines
We have two usage patterns:
1. get & ->setup() a new discipline, or
2. ->remove() & put the currently loaded one.
Add corresponding helpers that hide the internals & error handling.
--- drivers/s390/net/qeth_core.h | 4 +-- drivers/s390/net/qeth_core_main.c | 45 ++++++++++++++++--------------- drivers/s390/net/qeth_core_sys.c | 10 ++----- 3 files changed, 28 insertions(+), 31 deletions(-)
drivers/s390/net/qeth_core.h +2 -2
@@ -1067,8 +1067,8 @@ extern const struct device_type qeth_generic_devtype;const char *qeth_get_cardname_short (struct qeth_card *);int qeth_resize_buffer_pool (struct qeth_card *card,unsigned int count); -int qeth_core_load_discipline (struct qeth_card *,enum qeth_discipline_id ); -void qeth_core_free_discipline (struct qeth_card *); +int qeth_setup_discipline (struct qeth_card *card,enum qeth_discipline_id disc); +void qeth_remove_discipline (struct qeth_card *card);/* exports for qeth discipline device drivers */ extern struct kmem_cache *qeth_core_header_cache;
drivers/s390/net/qeth_core_main.c +24 -21
92 lines not on this page · raw message
drivers/s390/net/qeth_core_sys.c +2 -8
@@ -384,19 +384,13 @@ static ssize_t qeth_dev_layer2_store(struct device *dev, goto out; } - card->discipline->remove (card->gdev); -qeth_core_free_discipline (card); +qeth_remove_discipline (card);free_netdev (card->dev); card->dev = ndev; } - rc =qeth_core_load_discipline (card, newdis); -if (rc) - goto out; + rc =qeth_setup_discipline (card, newdis); - rc = card->discipline->setup (card->gdev); -if (rc) -qeth_core_free_discipline (card); out:mutex_unlock (&card->discipline_mutex);return rc ? rc : count;
Signature · 1 line
2.17.1
PATCH 2/5 s390/qeth: remove qeth_get_ip_version()
In reply to Julian WiedmannReplace our home-grown helper with the more robust vlan_get_protocol(). This is pretty much a 1:1 replacement, we just need to pass around a proper ETH_P_* everyhwere and convert the old value range.
For readability also convert the protocol checks in qeth_l3_hard_start_xmit() to a switch statement.
--- drivers/s390/net/qeth_core.h | 40 ++++++++------------- drivers/s390/net/qeth_core_main.c | 18 +++++----- drivers/s390/net/qeth_l2_main.c | 6 ++-- drivers/s390/net/qeth_l3_main.c | 60 ++++++++++++++++++------------- 4 files changed, 62 insertions(+), 62 deletions(-)
drivers/s390/net/qeth_core.h +14 -26
@@ -956,24 +956,6 @@ static inline int qeth_get_elements_for_range(addr_t start, addr_t end)return PFN_UP (end) -PFN_DOWN (start); } -static inline int qeth_get_ip_version (struct sk_buff *skb) -{ -struct vlan_ethhdr *veth =vlan_eth_hdr (skb); -__be16 prot = veth->h_vlan_proto; - -if (prot ==htons (ETH_P_8021Q )) - prot = veth->h_vlan_encapsulated_proto; - -switch (prot) { -case htons (ETH_P_IPV6 ): -return 6 ; -case htons (ETH_P_IP ): -return 4 ; -default : -return 0 ; - } -} -static inline int qeth_get_ether_cast_type (struct sk_buff *skb) {u8 *addr =eth_hdr (skb)->h_dest;
@@ -984,14 +966,20 @@ static inline int qeth_get_ether_cast_type(struct sk_buff *skb)return RTN_UNICAST ; } -static inline struct dst_entry *qeth_dst_check_rcu (struct sk_buff *skb,int ipv) +static inline struct dst_entry *qeth_dst_check_rcu (struct sk_buff *skb, +__be16 proto) {struct dst_entry *dst =skb_dst (skb);struct rt6_info *rt; rt = (struct rt6_info *) dst; -if (dst) - dst =dst_check (dst, (ipv ==6 ) ?rt6_get_cookie (rt) :0 ); +if (dst) { +if (proto ==htons (ETH_P_IPV6 )) + dst =dst_check (dst,rt6_get_cookie (rt)); +else + dst =dst_check (dst,0 ); + } +return dst; }
@@ -1014,11 +1002,11 @@ static inline struct in6_addr *qeth_next_hop_v6_rcu(struct sk_buff *skb,return &ipv6_hdr (skb)->daddr; } -static inline void qeth_tx_csum (struct sk_buff *skb,u8 *flags,int ipv) +static inline void qeth_tx_csum (struct sk_buff *skb,u8 *flags,__be16 proto) { *flags |=QETH_HDR_EXT_CSUM_TRANSP_REQ ; -if ((ipv ==4 &&ip_hdr (skb)->protocol ==IPPROTO_UDP ) || - (ipv ==6 &&ipv6_hdr (skb)->nexthdr ==IPPROTO_UDP )) +if ((proto ==htons (ETH_P_IP ) &&ip_hdr (skb)->protocol ==IPPROTO_UDP ) || + (proto ==htons (ETH_P_IPV6 ) &&ipv6_hdr (skb)->nexthdr ==IPPROTO_UDP )) *flags |=QETH_HDR_EXT_UDP ; }
@@ -1145,10 +1133,10 @@ int qeth_stop(struct net_device *dev);int qeth_vm_request_mac (struct qeth_card *card);int qeth_xmit (struct qeth_card *card,struct sk_buff *skb, -struct qeth_qdio_out_q *queue,int ipv, +struct qeth_qdio_out_q *queue,__be16 proto,void (*fill_header)(struct qeth_qdio_out_q *queue,struct qeth_hdr *hdr,struct sk_buff *skb, -int ipv,unsigned int data_len)); +__be16 proto,unsigned int data_len));/* exports for OSN */ int qeth_osn_assist (struct net_device *,void *,int );
drivers/s390/net/qeth_core_main.c +10 -8
@@ -825,7 +825,8 @@ static bool qeth_next_hop_is_local_v4(struct qeth_card *card,return false;rcu_read_lock (); - next_hop =qeth_next_hop_v4_rcu (skb,qeth_dst_check_rcu (skb,4 )); + next_hop =qeth_next_hop_v4_rcu (skb, +qeth_dst_check_rcu (skb,htons (ETH_P_IP ))); key =ipv4_addr_hash (next_hop);hash_for_each_possible_rcu (card->local_addrs4, tmp, hnode, key) {
@@ -851,7 +852,8 @@ static bool qeth_next_hop_is_local_v6(struct qeth_card *card,return false;rcu_read_lock (); - next_hop =qeth_next_hop_v6_rcu (skb,qeth_dst_check_rcu (skb,6 )); + next_hop =qeth_next_hop_v6_rcu (skb, +qeth_dst_check_rcu (skb,htons (ETH_P_IPV6 ))); key =ipv6_addr_hash (next_hop);hash_for_each_possible_rcu (card->local_addrs6, tmp, hnode, key) {
@@ -3896,11 +3898,11 @@ int qeth_get_priority_queue(struct qeth_card *card, struct sk_buff *skb)switch (card->qdio.do_prio_queueing) {case QETH_PRIO_Q_ING_TOS :case QETH_PRIO_Q_ING_PREC : -switch (qeth_get_ip_version (skb)) { -case 4 : +switch (vlan_get_protocol (skb)) { +case htons (ETH_P_IP ): tos =ipv4_get_dsfield (ip_hdr (skb));break ; -case 6 : +case htons (ETH_P_IPV6 ): tos =ipv6_get_dsfield (ipv6_hdr (skb));break ;default :
@@ -4365,10 +4367,10 @@ static void qeth_fill_tso_ext(struct qeth_hdr_tso *hdr, }int qeth_xmit (struct qeth_card *card,struct sk_buff *skb, -struct qeth_qdio_out_q *queue,int ipv, +struct qeth_qdio_out_q *queue,__be16 proto,void (*fill_header)(struct qeth_qdio_out_q *queue,struct qeth_hdr *hdr,struct sk_buff *skb, -int ipv,unsigned int data_len)) +__be16 proto,unsigned int data_len)) {unsigned int proto_len, hw_hdr_len;unsigned int frame_len = skb->len;
@@ -4401,7 +4403,7 @@ int qeth_xmit(struct qeth_card *card, struct sk_buff *skb, data_offset = push_len + proto_len; }memset (hdr,0 , hw_hdr_len); -fill_header (queue, hdr, skb, ipv, frame_len); +fill_header (queue, hdr, skb, proto, frame_len);if (is_tso)qeth_fill_tso_ext ((struct qeth_hdr_tso *) hdr, frame_len - proto_len, skb, proto_len);
drivers/s390/net/qeth_l2_main.c +3 -3
@@ -157,7 +157,7 @@ static void qeth_l2_drain_rx_mode_cache(struct qeth_card *card)static void qeth_l2_fill_header (struct qeth_qdio_out_q *queue,struct qeth_hdr *hdr,struct sk_buff *skb, -int ipv,unsigned int data_len) +__be16 proto,unsigned int data_len) {int cast_type =qeth_get_ether_cast_type (skb);struct vlan_ethhdr *veth =vlan_eth_hdr (skb);
@@ -169,7 +169,7 @@ static void qeth_l2_fill_header(struct qeth_qdio_out_q *queue, }else { hdr->hdr.l2.id =QETH_HEADER_TYPE_LAYER2 ;if (skb->ip_summed ==CHECKSUM_PARTIAL ) -qeth_tx_csum (skb, &hdr->hdr.l2.flags[1 ], ipv); +qeth_tx_csum (skb, &hdr->hdr.l2.flags[1 ], proto); }/* set byte byte 3 to casting flags */
@@ -551,7 +551,7 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb,if (IS_OSN (card)) rc =qeth_l2_xmit_osn (card, skb, queue);else - rc =qeth_xmit (card, skb, queue,qeth_get_ip_version (skb), + rc =qeth_xmit (card, skb, queue,vlan_get_protocol (skb), qeth_l2_fill_header);if (!rc)
drivers/s390/net/qeth_l3_main.c +35 -25
156 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH 3/5 s390/qeth: pass proto to qeth_l3_get_cast_type()
In reply to Julian Wiedmannqeth_l3_hard_start_xmit() already determined the skb's proto. Avoid doing so a second time when it calls qeth_l3_get_cast_type().
--- drivers/s390/net/qeth_l3_main.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
drivers/s390/net/qeth_l3_main.c +6 -5
@@ -1610,9 +1610,8 @@ static int qeth_l3_get_cast_type_rcu(struct sk_buff *skb, struct dst_entry *dst, } } -static int qeth_l3_get_cast_type (struct sk_buff *skb) +static int qeth_l3_get_cast_type (struct sk_buff *skb,__be16 proto) { -__be16 proto =vlan_get_protocol (skb);struct dst_entry *dst;int cast_type;
@@ -1771,7 +1770,7 @@ static netdev_tx_t qeth_l3_hard_start_xmit(struct sk_buff *skb, }if (!(dev->flags &IFF_BROADCAST ) && -qeth_l3_get_cast_type (skb) ==RTN_BROADCAST ) +qeth_l3_get_cast_type (skb, proto) ==RTN_BROADCAST ) goto tx_drop;if (proto ==htons (ETH_P_IP ) ||IS_IQD (card))
@@ -1831,8 +1830,10 @@ static netdev_features_t qeth_l3_osa_features_check(struct sk_buff *skb,static u16 qeth_l3_iqd_select_queue (struct net_device *dev,struct sk_buff *skb,struct net_device *sb_dev) { -return qeth_iqd_select_queue (dev, skb,qeth_l3_get_cast_type (skb), - sb_dev); +__be16 proto =vlan_get_protocol (skb); + +return qeth_iqd_select_queue (dev, skb, +qeth_l3_get_cast_type (skb, proto), sb_dev); } static u16 qeth_l3_osa_select_queue(struct net_device *dev, struct sk_buff *skb,
Signature · 1 line
2.17.1
PATCH 4/5 s390/qeth: make cast type selection for af_iucv skbs robust
In reply to Julian WiedmannAs part of the TX queue selection for af_iucv skbs, qeth_l3_get_cast_type_rcu() ends up calling qeth_get_ether_cast_type(). Which is rather fragile, since such skbs don't have a proper ETH header and we rely on it being zeroed out in the right places. Add a separate case for ETH_P_AF_IUCV instead that does the right thing.
When later building the HW header for such skbs, don't hard-code the cast type but follow the same path as for other protocol types. Here the cast type should naturally come from the skb's queue mapping.
--- drivers/s390/net/qeth_l3_main.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-)
drivers/s390/net/qeth_l3_main.c +15 -12
@@ -1604,8 +1604,10 @@ static int qeth_l3_get_cast_type_rcu(struct sk_buff *skb, struct dst_entry *dst,case htons (ETH_P_IPV6 ):return ipv6_addr_is_multicast (&ipv6_hdr (skb)->daddr) ?RTN_MULTICAST :RTN_UNICAST ; +case htons (ETH_P_AF_IUCV ): +return RTN_UNICAST ;default : -/* ... and MAC address */ +/* OSA only: ... and MAC address */ return qeth_get_ether_cast_type (skb); } }
@@ -1651,14 +1653,6 @@ static void qeth_l3_fill_header(struct qeth_qdio_out_q *queue, }else { hdr->hdr.l3.id =QETH_HEADER_TYPE_LAYER3 ; -if (proto ==htons (ETH_P_AF_IUCV )) { - l3_hdr->flags =QETH_HDR_IPV6 |QETH_CAST_UNICAST ; - l3_hdr->next_hop.addr.s6_addr16[0 ] =htons (0xfe80 ); -memcpy (&l3_hdr->next_hop.addr.s6_addr32[2 ], -iucv_trans_hdr (skb)->destUserID,8 ); -return ; - } -if (skb->ip_summed ==CHECKSUM_PARTIAL ) {qeth_tx_csum (skb, &hdr->hdr.l3.ext_flags, proto);/* some HW requires combined L3+L4 csum offload: */
@@ -1687,16 +1681,25 @@ static void qeth_l3_fill_header(struct qeth_qdio_out_q *queue, cast_type =qeth_l3_get_cast_type_rcu (skb, dst, proto); l3_hdr->flags |=qeth_l3_cast_type_to_flag (cast_type); -if (proto ==htons (ETH_P_IP )) { +switch (proto) { +case htons (ETH_P_IP ): l3_hdr->next_hop.addr.s6_addr32[3 ] =qeth_next_hop_v4_rcu (skb, dst); - }else if (proto ==htons (ETH_P_IPV6 )) { +break ; +case htons (ETH_P_IPV6 ): l3_hdr->next_hop.addr = *qeth_next_hop_v6_rcu (skb, dst); hdr->hdr.l3.flags |=QETH_HDR_IPV6 ;if (!IS_IQD (card)) hdr->hdr.l3.flags |=QETH_HDR_PASSTHRU ; - }else { +break ; +case htons (ETH_P_AF_IUCV ): + l3_hdr->next_hop.addr.s6_addr16[0 ] =htons (0xfe80 ); +memcpy (&l3_hdr->next_hop.addr.s6_addr32[2 ], +iucv_trans_hdr (skb)->destUserID,8 ); + l3_hdr->flags |=QETH_HDR_IPV6 ; +break ; +default :/* OSA only: */ l3_hdr->flags |=QETH_HDR_PASSTHRU ; }
Signature · 1 line
2.17.1
PATCH 5/5 s390/qeth: don't fake a TX completion interrupt after TX error
In reply to Julian WiedmannWhen do_qdio() returns with an unexpected error, qeth_flush_buffers() kicks off a recovery action.
In such a case there's no point in starting TX completion processing, the device gets torn down anyway. So take a closer look at do_qdio()'s return value, and skip the TX completion processing accordingly.
--- drivers/s390/net/qeth_core_main.c | 34 ++++++++++++++++--------------- 1 file changed, 18 insertions(+), 16 deletions(-)
drivers/s390/net/qeth_core_main.c +18 -16
@@ -3692,24 +3692,27 @@ static void qeth_flush_buffers(struct qeth_qdio_out_q *queue, int index, rc =do_QDIO (CARD_DDEV (queue->card), qdio_flags, queue->queue_no, index, count); -/* Fake the TX completion interrupt: */ -if (IS_IQD (card)) { -unsigned int frames =READ_ONCE (queue->max_coalesced_frames); -unsigned int usecs =READ_ONCE (queue->coalesce_usecs); +switch (rc) { +case 0 : +case -ENOBUFS : +/* ignore temporary SIGA errors without busy condition */ -if (frames && queue->coalesced_frames >= frames) { -napi_schedule (&queue->napi); - queue->coalesced_frames =0 ; -QETH_TXQ_STAT_INC (queue, coal_frames); - }else if (usecs) { -qeth_tx_arm_timer (queue, usecs); +/* Fake the TX completion interrupt: */ +if (IS_IQD (card)) { +unsigned int frames =READ_ONCE (queue->max_coalesced_frames); +unsigned int usecs =READ_ONCE (queue->coalesce_usecs); + +if (frames && queue->coalesced_frames >= frames) { +napi_schedule (&queue->napi); + queue->coalesced_frames =0 ; +QETH_TXQ_STAT_INC (queue, coal_frames); + }else if (usecs) { +qeth_tx_arm_timer (queue, usecs); + } } - } -if (rc) { -/* ignore temporary SIGA errors without busy condition */ -if (rc == -ENOBUFS ) -return ; +break ; +default :QETH_CARD_TEXT (queue->card,2 ,"flushbuf" );QETH_CARD_TEXT_ (queue->card,2 ," q%d" , queue->queue_no);QETH_CARD_TEXT_ (queue->card,2 ," idx%d" , index);
@@ -3719,7 +3722,6 @@ static void qeth_flush_buffers(struct qeth_qdio_out_q *queue, int index,/* this must not happen under normal circumstances. if it * happens something is really wrong -> recover */ qeth_schedule_recovery (queue->card); -return ; } }
Signature · 1 line
2.17.1
Quoted 17 lines · Julian Wiedmann
Hi Dave & Jakub,
please apply the following patch series for qeth to netdev's net-next tree.
Nothing special, mostly fine-tuning and follow-on cleanups for earlier fixes.
Thanks,
JulianJulian Wiedmann (5):
s390/qeth: clean up load/remove code for disciplines s390/qeth: removeqeth_get_ip_version()
s390/qeth: pass proto toqeth_l3_get_cast_type()s390/qeth: make cast type selection foraf_iucvskbs robust s390/qeth: don't fake a TX completion interrupt after TX error
Quoted 6 lines
drivers/s390/net/qeth_core.h | 44 +++++--------- drivers/s390/net/qeth_core_main.c | 97 +++++++++++++++++-------------- drivers/s390/net/qeth_core_sys.c | 10 +--- drivers/s390/net/qeth_l2_main.c | 6 +- drivers/s390/net/qeth_l3_main.c | 90 ++++++++++++++++------------ 5 files changed, 125 insertions(+), 122 deletions(-)
for netdrv
This series was applied to netdev/net-next.git (refs/heads/master):
Hello:
This series was applied to netdev/net-next.git (refs/heads/master):
Quoted 10 lines · On Thu, 28 Jan 2021 12:25:46 +0100 you wrote:
Hi Dave & Jakub,
please apply the following patch series for qeth to netdev's net-next tree.
Nothing special, mostly fine-tuning and follow-on cleanups for earlier fixes.
Thanks,
Julian[...]
Here is the summary with links:
- [net-next,1/5] s390/qeth: clean up load/remove code for disciplines
https://git.kernel.org/netdev/net-next/c/ea12f1b3c828 - [net-next,2/5] s390/qeth: remove qeth_get_ip_version() https://git.kernel.org/netdev/net-next/c/17f3a8b5f5c9 - [net-next,3/5] s390/qeth: pass proto to qeth_l3_get_cast_type() https://git.kernel.org/netdev/net-next/c/c61dff3c1ef7 - [net-next,4/5] s390/qeth: make cast type selection for af_iucv skbs robust https://git.kernel.org/netdev/net-next/c/a667fee181b2 - [net-next,5/5] s390/qeth: don't fake a TX completion interrupt after TX error https://git.kernel.org/netdev/net-next/c/d6e515031517
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html