s390/qeth: updates 2019-04-25
Subsystems (4)
cover s390/qeth: updates 2019-04-25
Hi Dave,
please apply one more patch series for qeth to net-next. Nothing special, just a bunch of cleanups.
Thanks,
Julian
Arnd Bergmann (1):
s390: qeth: address type mismatch warning
Julian Wiedmann (7):
s390/qeth: remove RX seqno in skb->cb
s390/qeth: clean up stale buffer state documentation s390/qeth: use IS_* helpers for checking device type s390/qeth: don't clear Output buffers on every queue init s390/qeth: cache max number of available buffer elements s390/qeth: extract helper to determine L2 cast type s390/qeth: trust non-IP cast type in qeth_l3_fill_header()
drivers/s390/net/qeth_core.h | 74 +++++++++++------------ drivers/s390/net/qeth_core_main.c | 122 ++++++++++++++++---------------------- drivers/s390/net/qeth_core_mpc.h | 2 +- drivers/s390/net/qeth_core_sys.c | 3 +- drivers/s390/net/qeth_l2_main.c | 33 ++++------- drivers/s390/net/qeth_l3_main.c | 43 +++++--------- drivers/s390/net/qeth_l3_sys.c | 6 +- 7 files changed, 115 insertions(+), 168 deletions(-)
Signature · 1 line
2.16.4
PATCH 1/8 s390: qeth: address type mismatch warning
From: Arnd Bergmann <[email protected]>
clang produces a harmless warning for each use for the qeth_adp_supported macro:
drivers/s390/net/qeth_l2_main.c:559:31: warning: implicit conversion from enumeration type 'enum qeth_ipa_setadp_cmd' to
different enumeration type 'enum qeth_ipa_funcs' [-Wenum-conversion]
if (qeth_adp_supported(card, IPA_SETADP_SET_PROMISC_MODE))
~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~drivers/s390/net/qeth_core.h:179:41: note: expanded from macro 'qeth_adp_supported'
qeth_is_ipa_supported(&c->options.adp, f)
~~~~~~~~~~~~~~~~~~~~~ ^Add a version of this macro that uses the correct types, and remove the unused qeth_adp_enabled() macro that has the same problem.
--- drivers/s390/net/qeth_core.h | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/s390/net/qeth_core.h +7 -3
@@ -165,6 +165,12 @@ struct qeth_vnicc_info {bool rx_bcast_enabled; }; +static inline int qeth_is_adp_supported (struct qeth_ipa_info *ipa, +enum qeth_ipa_setadp_cmd func) +{ +return (ipa->supported_funcs & func); +} +static inline int qeth_is_ipa_supported (struct qeth_ipa_info *ipa,enum qeth_ipa_funcs func) {
@@ -178,9 +184,7 @@ static inline int qeth_is_ipa_enabled(struct qeth_ipa_info *ipa, }#define qeth_adp_supported (c, f) \ - qeth_is_ipa_supported(&c->options.adp, f) -#define qeth_adp_enabled (c, f) \ - qeth_is_ipa_enabled(&c->options.adp, f) + qeth_is_adp_supported(&c->options.adp, f)#define qeth_is_supported (c, f) \ qeth_is_ipa_supported(&c->options.ipa4, f)#define qeth_is_enabled (c, f) \
Signature · 1 line
2.16.4
PATCH 2/8 s390/qeth: remove RX seqno in skb->cb
In reply to Julian WiedmannIt's unclear what exact purpose this seqno may have served in the past. But it's certainly no longer used anymore, as the following napi_gro_receive() will straight away clear this part of the cb again.
--- drivers/s390/net/qeth_core.h | 1 - drivers/s390/net/qeth_l2_main.c | 2 -- 2 files changed, 3 deletions(-)
drivers/s390/net/qeth_core.h +0 -1
@@ -652,7 +652,6 @@ struct qeth_seqno {__u32 pdu_hdr;__u32 pdu_hdr_ack;__u16 ipa; -__u32 pkt_seqno; };struct qeth_reply {
drivers/s390/net/qeth_l2_main.c +0 -2
@@ -332,8 +332,6 @@ static int qeth_l2_process_inbound_buffer(struct qeth_card *card,case QETH_HEADER_TYPE_LAYER2 : skb->protocol =eth_type_trans (skb, skb->dev);qeth_rx_csum (card, skb, hdr->hdr.l2.flags[1 ]); -if (skb->protocol ==htons (ETH_P_802_2 )) - *((__u32 *)skb->cb) = ++card->seqno.pkt_seqno; len = skb->len;napi_gro_receive (&card->napi, skb);break ;
Signature · 1 line
2.16.4
PATCH 3/8 s390/qeth: clean up stale buffer state documentation
In reply to Julian WiedmannWe don't keep track of Input Buffer states, so remove the comments that make it sound like the qeth_qdio_buffer_states enum applies to Input Buffers.
--- drivers/s390/net/qeth_core.h | 41 +++++++++++++---------------------------- 1 file changed, 13 insertions(+), 28 deletions(-)
drivers/s390/net/qeth_core.h +13 -28
@@ -374,34 +374,6 @@ enum qeth_header_ids {#define QETH_HDR_EXT_CSUM_TRANSP_REQ 0x20#define QETH_HDR_EXT_UDP 0x40/*bit off for TCP*/ -enum qeth_qdio_buffer_states { -/* -* inbound: read out by driver; owned by hardware in order to be filled -* outbound: owned by driver in order to be filled -*/ -QETH_QDIO_BUF_EMPTY , -/* -* inbound: filled by hardware; owned by driver in order to be read out -* outbound: filled by driver; owned by hardware in order to be sent -*/ -QETH_QDIO_BUF_PRIMED , -/* -* inbound: not applicable -* outbound: identified to be pending in TPQ -*/ -QETH_QDIO_BUF_PENDING , -/* -* inbound: not applicable -* outbound: found in completion queue -*/ -QETH_QDIO_BUF_IN_CQ , -/* -* inbound: not applicable -* outbound: handled via transfer pending / completion queue -*/ -QETH_QDIO_BUF_HANDLED_DELAYED , -}; -enum qeth_qdio_info_states {QETH_QDIO_UNINITIALIZED ,QETH_QDIO_ALLOCATED ,
@@ -433,6 +405,19 @@ struct qeth_qdio_q {int next_buf_to_init; }; +enum qeth_qdio_out_buffer_state { +/* Owned by driver, in order to be filled. */ +QETH_QDIO_BUF_EMPTY , +/* Filled by driver; owned by hardware in order to be sent. */ +QETH_QDIO_BUF_PRIMED , +/* Identified to be pending in TPQ. */ +QETH_QDIO_BUF_PENDING , +/* Found in completion queue. */ +QETH_QDIO_BUF_IN_CQ , +/* Handled via transfer pending / completion queue. */ +QETH_QDIO_BUF_HANDLED_DELAYED , +}; +struct qeth_qdio_out_buffer {struct qdio_buffer *buffer;atomic_t state;
Signature · 1 line
2.16.4
PATCH 4/8 s390/qeth: use IS_* helpers for checking device type
In reply to Julian WiedmannWe have helper macros for all possible device types, replace all remaining open-coded accesses to the type fields.
--- drivers/s390/net/qeth_core.h | 2 +- drivers/s390/net/qeth_core_main.c | 80 +++++++++++++++------------------------ drivers/s390/net/qeth_core_mpc.h | 2 +- drivers/s390/net/qeth_core_sys.c | 3 +- drivers/s390/net/qeth_l2_main.c | 15 +++----- drivers/s390/net/qeth_l3_main.c | 24 ++++++------ drivers/s390/net/qeth_l3_sys.c | 6 +-- 7 files changed, 54 insertions(+), 78 deletions(-)
drivers/s390/net/qeth_core.h +1 -1
@@ -667,7 +667,7 @@ struct qeth_card_info {__u16 func_level;char mcl_level[QETH_MCL_LENGTH +1 ]; u8 open_when_online:1; -int guestlan; + u8 is_vm_nic:1;int mac_bits;enum qeth_card_types type;enum qeth_link_types link_type;
drivers/s390/net/qeth_core_main.c +31 -49
214 lines not on this page · raw message
drivers/s390/net/qeth_core_mpc.h +1 -1
@@ -82,7 +82,7 @@ enum qeth_card_types {#define IS_OSM (card) ((card)->info.type == QETH_CARD_TYPE_OSM)#define IS_OSN (card) ((card)->info.type == QETH_CARD_TYPE_OSN)#define IS_OSX (card) ((card)->info.type == QETH_CARD_TYPE_OSX) -#define IS_VM_NIC (card) ((card)->info.guestlan) +#define IS_VM_NIC (card) ((card)->info.is_vm_nic)#define QETH_MPC_DIFINFO_LEN_INDICATES_LINK_TYPE 0x18/* only the first two bytes are looked at in qeth_get_cardname_short */
drivers/s390/net/qeth_core_sys.c +1 -2
@@ -479,8 +479,7 @@ static ssize_t qeth_dev_isolation_store(struct device *dev,return -EINVAL ;mutex_lock (&card->conf_mutex); -if (card->info.type !=QETH_CARD_TYPE_OSD && - card->info.type !=QETH_CARD_TYPE_OSX ) { +if (!IS_OSD (card) && !IS_OSX (card)) { rc = -EOPNOTSUPP ;dev_err (&card->gdev->dev,"Adapter does not " "support QDIO data connection isolation\n" );
drivers/s390/net/qeth_l2_main.c +6 -9
@@ -336,7 +336,7 @@ static int qeth_l2_process_inbound_buffer(struct qeth_card *card,napi_gro_receive (&card->napi, skb);break ;case QETH_HEADER_TYPE_OSN : -if (card->info.type ==QETH_CARD_TYPE_OSN ) { +if (IS_OSN (card)) {skb_push (skb,sizeof (struct qeth_hdr ));skb_copy_to_linear_data (skb, hdr,sizeof (struct qeth_hdr ));
@@ -387,8 +387,7 @@ static int qeth_l2_request_initial_mac(struct qeth_card *card) }/* some devices don't support a custom MAC address: */ -if (card->info.type ==QETH_CARD_TYPE_OSM || - card->info.type ==QETH_CARD_TYPE_OSX ) +if (IS_OSM (card) ||IS_OSX (card))return (rc) ? rc : -EADDRNOTAVAIL ;eth_hw_addr_random (card->dev);
@@ -733,7 +732,7 @@ static int qeth_l2_setup_netdev(struct qeth_card *card, bool carrier_ok) card->dev->features |=NETIF_F_HW_VLAN_CTAG_FILTER ; } -if (card->info.type ==QETH_CARD_TYPE_OSD && !card->info.guestlan) { +if (IS_OSD (card) && !IS_VM_NIC (card)) { card->dev->features |=NETIF_F_SG ;/* OSA 3S and earlier has no RX/TX support */ if (qeth_is_supported (card,IPA_OUTBOUND_CHECKSUM )) {
@@ -843,8 +842,7 @@ static int qeth_l2_set_online(struct ccwgroup_device *gdev)/* softsetup */ QETH_DBF_TEXT (SETUP ,2 ,"softsetp" ); -if ((card->info.type ==QETH_CARD_TYPE_OSD ) || - (card->info.type ==QETH_CARD_TYPE_OSX )) { +if (IS_OSD (card) ||IS_OSX (card)) { rc =qeth_l2_start_ipassists (card);if (rc) goto out_remove;
@@ -1468,9 +1466,8 @@ static struct qeth_cmd_buffer *qeth_sbp_build_cmd(struct qeth_card *card,enum qeth_ipa_sbp_cmd sbp_cmd, unsigned int cmd_length) { - enum qeth_ipa_cmds ipa_cmd = (card->info.type == QETH_CARD_TYPE_IQD) ? -IPA_CMD_SETBRIDGEPORT_IQD : -IPA_CMD_SETBRIDGEPORT_OSA ; + enum qeth_ipa_cmds ipa_cmd = IS_IQD(card) ? IPA_CMD_SETBRIDGEPORT_IQD : +IPA_CMD_SETBRIDGEPORT_OSA ;struct qeth_cmd_buffer *iob;struct qeth_ipa_cmd *cmd;
drivers/s390/net/qeth_l3_main.c +11 -13
@@ -474,7 +474,7 @@ static int qeth_l3_send_setrouting(struct qeth_card *card,static int qeth_l3_correct_routing_type (struct qeth_card *card,enum qeth_routing_types *type,enum qeth_prot_versions prot) { -if (card->info.type ==QETH_CARD_TYPE_IQD ) { +if (IS_IQD (card)) {switch (*type) {case NO_ROUTER :case PRIMARY_CONNECTOR :
@@ -854,7 +854,7 @@ static int qeth_l3_softsetup_ipv6(struct qeth_card *card)QETH_CARD_TEXT (card,3 ,"softipv6" ); -if (card->info.type ==QETH_CARD_TYPE_IQD ) +if (IS_IQD (card)) goto out; rc = qeth_send_simple_setassparms(card, IPA_IPV6,
@@ -1380,8 +1380,7 @@ static int qeth_l3_process_inbound_buffer(struct qeth_card *card,switch (hdr->hdr.l3.id) {case QETH_HEADER_TYPE_LAYER3 : magic = *(__u16 *)skb->data; -if ((card->info.type ==QETH_CARD_TYPE_IQD ) && - (magic ==ETH_P_AF_IUCV )) { +if (IS_IQD (card) && magic ==ETH_P_AF_IUCV ) { len = skb->len;dev_hard_header (skb, dev,ETH_P_AF_IUCV , dev->dev_addr,"FAKELL" , len);
@@ -1460,7 +1459,7 @@ qeth_l3_handle_promisc_mode(struct qeth_card *card) (card->info.promisc_mode == SET_PROMISC_MODE_OFF))) return; -if (card->info.guestlan) {/* Guestlan trace */ +if (IS_VM_NIC (card)) {/* Guestlan trace */ if (qeth_adp_supported (card,IPA_SETADP_SET_PROMISC_MODE ))qeth_setadp_promisc_mode (card); } else if (card->options.sniffer && /* HiperSockets trace */
@@ -1557,7 +1556,7 @@ static int qeth_l3_arp_set_no_entries(struct qeth_card *card, int no_entries)* IPA_CMD_ASS_ARP_QUERY_INFO, but not IPA_CMD_ASS_ARP_SET_NO_ENTRIES; * thus we say EOPNOTSUPP for this ARP function */ -if (card->info.guestlan) +if (IS_VM_NIC (card))return -EOPNOTSUPP ;if (!qeth_is_supported (card,IPA_ARP_PROCESSING )) {return -EOPNOTSUPP ;
@@ -1789,7 +1788,7 @@ static int qeth_l3_arp_modify_entry(struct qeth_card *card,* IPA_CMD_ASS_ARP_QUERY_INFO, but not IPA_CMD_ASS_ARP_ADD_ENTRY; * thus we say EOPNOTSUPP for this ARP function */ -if (card->info.guestlan) +if (IS_VM_NIC (card))return -EOPNOTSUPP ;if (!qeth_is_supported (card,IPA_ARP_PROCESSING )) {return -EOPNOTSUPP ;
@@ -1822,7 +1821,7 @@ static int qeth_l3_arp_flush_cache(struct qeth_card *card)* IPA_CMD_ASS_ARP_QUERY_INFO, but not IPA_CMD_ASS_ARP_FLUSH_CACHE; * thus we say EOPNOTSUPP for this ARP function */ -if (card->info.guestlan || (card->info.type ==QETH_CARD_TYPE_IQD )) +if (IS_VM_NIC (card) ||IS_IQD (card))return -EOPNOTSUPP ;if (!qeth_is_supported (card,IPA_ARP_PROCESSING )) {return -EOPNOTSUPP ;
@@ -2013,7 +2012,7 @@ static void qeth_l3_fill_header(struct qeth_qdio_out_q *queue, l3_hdr->next_hop.ipv6_addr =ipv6_hdr (skb)->daddr; hdr->hdr.l3.flags |=QETH_HDR_IPV6 ; -if (card->info.type !=QETH_CARD_TYPE_IQD ) +if (!IS_IQD (card)) hdr->hdr.l3.flags |=QETH_HDR_PASSTHRU ; }rcu_read_unlock ();
@@ -2195,8 +2194,7 @@ static int qeth_l3_setup_netdev(struct qeth_card *card, bool carrier_ok)unsigned int headroom;int rc; -if (card->info.type ==QETH_CARD_TYPE_OSD || - card->info.type ==QETH_CARD_TYPE_OSX ) { +if (IS_OSD (card) ||IS_OSX (card)) {if ((card->info.link_type ==QETH_LINK_TYPE_LANE_TR ) || (card->info.link_type ==QETH_LINK_TYPE_HSTR )) {pr_info ("qeth_l3: ignoring TR device\n" );
@@ -2210,7 +2208,7 @@ static int qeth_l3_setup_netdev(struct qeth_card *card, bool carrier_ok)if (!(card->info.unique_id &UNIQUE_ID_NOT_BY_CARD )) card->dev->dev_id = card->info.unique_id &0xffff ; -if (!card->info.guestlan) { +if (!IS_VM_NIC (card)) { card->dev->features |=NETIF_F_SG ; card->dev->hw_features |=NETIF_F_TSO |NETIF_F_RXCSUM |NETIF_F_IP_CSUM ;
@@ -2234,7 +2232,7 @@ static int qeth_l3_setup_netdev(struct qeth_card *card, bool carrier_ok) headroom =sizeof (struct qeth_hdr_tso ); else headroom =sizeof (struct qeth_hdr ) +VLAN_HLEN ; - }else if (card->info.type ==QETH_CARD_TYPE_IQD ) { + }else if (IS_IQD (card)) { card->dev->flags |=IFF_NOARP ; card->dev->netdev_ops = &qeth_l3_netdev_ops; headroom =sizeof (struct qeth_hdr ) -ETH_HLEN ;
drivers/s390/net/qeth_l3_sys.c +3 -3
@@ -206,7 +206,7 @@ static ssize_t qeth_l3_dev_sniffer_store(struct device *dev,if (!card)return -EINVAL ; -if (card->info.type !=QETH_CARD_TYPE_IQD ) +if (!IS_IQD (card))return -EPERM ;if (card->options.cq ==QETH_CQ_ENABLED )return -EPERM ;
@@ -258,7 +258,7 @@ static ssize_t qeth_l3_dev_hsuid_show(struct device *dev,if (!card)return -EINVAL ; -if (card->info.type !=QETH_CARD_TYPE_IQD ) +if (!IS_IQD (card))return -EPERM ;memcpy (tmp_hsuid, card->options.hsuid,sizeof (tmp_hsuid));
@@ -276,7 +276,7 @@ static ssize_t qeth_l3_dev_hsuid_store(struct device *dev,if (!card)return -EINVAL ; -if (card->info.type !=QETH_CARD_TYPE_IQD ) +if (!IS_IQD (card))return -EPERM ;if (card->state !=CARD_STATE_DOWN )return -EPERM ;
Signature · 1 line
2.16.4
PATCH 5/8 s390/qeth: don't clear Output buffers on every queue init
In reply to Julian WiedmannOn the first initialization of a queue, its Output Buffers are in a clean state with no attached resources. On every subsequent initialization, qeth_l?_stop_card() has previously put them in a clean state via qeth_drain_output_queues(). So the call to qeth_clear_output_buffer() is redundant and can be removed.
While at it, move the initialization of the queue's card pointer into the queue allocation. It never changes afterwards.
--- drivers/s390/net/qeth_core_main.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)
drivers/s390/net/qeth_core_main.c +2 -6
@@ -2386,6 +2386,7 @@ static int qeth_alloc_qdio_queues(struct qeth_card *card) goto out_freeoutq;QETH_DBF_TEXT_ (SETUP ,2 ,"outq %i" , i);QETH_DBF_HEX (SETUP ,2 , &card->qdio.out_qs[i],sizeof (void *)); + card->qdio.out_qs[i]->card = card; card->qdio.out_qs[i]->queue_no = i;/* give outbound qeth_qdio_buffers their qdio_buffers */ for (j =0 ; j <QDIO_MAX_BUFFERS_PER_Q ; ++j) {
@@ -2697,7 +2698,7 @@ static int qeth_init_input_buffer(struct qeth_card *card,int qeth_init_qdio_queues (struct qeth_card *card) { -int i, j; +unsigned int i;int rc;QETH_DBF_TEXT (SETUP ,2 ,"initqdqs" );
@@ -2728,11 +2729,6 @@ int qeth_init_qdio_queues(struct qeth_card *card)for (i =0 ; i < card->qdio.no_out_queues; ++i) {qdio_reset_buffers (card->qdio.out_qs[i]->qdio_bufs,QDIO_MAX_BUFFERS_PER_Q ); -for (j =0 ; j <QDIO_MAX_BUFFERS_PER_Q ; ++j) { -qeth_clear_output_buffer (card->qdio.out_qs[i], - card->qdio.out_qs[i]->bufs[j]); - } - card->qdio.out_qs[i]->card = card; card->qdio.out_qs[i]->next_buf_to_fill =0 ; card->qdio.out_qs[i]->do_pack =0 ;atomic_set (&card->qdio.out_qs[i]->used_buffers,0 );
Signature · 1 line
2.16.4
PATCH 6/8 s390/qeth: cache max number of available buffer elements
In reply to Julian WiedmannThe QETH_MAX_BUFFER_ELEMENTS() macro effectively returns a constant value. To avoid some redundant pointer chasing and computations in the xmit hot path, cache this value in the queue struct.
Take this as opportunity to shrink some of the queue struct's fields to their appropriate value range, slightly reducing its total size.
--- drivers/s390/net/qeth_core.h | 10 ++++------ drivers/s390/net/qeth_core_main.c | 34 +++++++++++++++++----------------- drivers/s390/net/qeth_l2_main.c | 2 +- 3 files changed, 22 insertions(+), 24 deletions(-)
drivers/s390/net/qeth_core.h +4 -6
@@ -484,14 +484,12 @@ struct qeth_qdio_out_q {struct qeth_qdio_out_buffer *bufs[QDIO_MAX_BUFFERS_PER_Q ];struct qdio_outbuf_state *bufstates;/* convenience pointer */ struct qeth_out_q_stats stats; -int queue_no; +u8 next_buf_to_fill; +u8 max_elements; +u8 queue_no; +u8 do_pack;struct qeth_card *card;atomic_t state; -int do_pack; -/* -* index of buffer to be filled by driver; state EMPTY or PACKING -*/ -int next_buf_to_fill;/* * number of buffers that are currently filled (PRIMED) * -> these buffers are hardware-owned
drivers/s390/net/qeth_core_main.c +17 -17
@@ -1165,15 +1165,14 @@ static void qeth_clear_output_buffer(struct qeth_qdio_out_q *queue,qeth_release_skbs (buf); -for (i =0 ; i <QETH_MAX_BUFFER_ELEMENTS (queue->card); ++i) { +for (i =0 ; i < queue->max_elements; ++i) {if (buf->buffer->element[i].addr && buf->is_header[i])kmem_cache_free (qeth_core_header_cache, buf->buffer->element[i].addr); buf->is_header[i] =0 ; } -qeth_scrub_qdio_buffer (buf->buffer, -QETH_MAX_BUFFER_ELEMENTS (queue->card)); +qeth_scrub_qdio_buffer (buf->buffer, queue->max_elements); buf->next_element_to_fill =0 ;atomic_set (&buf->state,QETH_QDIO_BUF_EMPTY ); }
@@ -2727,14 +2726,15 @@ int qeth_init_qdio_queues(struct qeth_card *card)/* outbound queue */ for (i =0 ; i < card->qdio.no_out_queues; ++i) { -qdio_reset_buffers (card->qdio.out_qs[i]->qdio_bufs, -QDIO_MAX_BUFFERS_PER_Q ); - card->qdio.out_qs[i]->next_buf_to_fill =0 ; - card->qdio.out_qs[i]->do_pack =0 ; -atomic_set (&card->qdio.out_qs[i]->used_buffers,0 ); -atomic_set (&card->qdio.out_qs[i]->set_pci_flags_count,0 ); -atomic_set (&card->qdio.out_qs[i]->state, -QETH_OUT_Q_UNLOCKED ); +struct qeth_qdio_out_q *queue = card->qdio.out_qs[i]; + +qdio_reset_buffers (queue->qdio_bufs,QDIO_MAX_BUFFERS_PER_Q ); + queue->max_elements =QETH_MAX_BUFFER_ELEMENTS (card); + queue->next_buf_to_fill =0 ; + queue->do_pack =0 ; +atomic_set (&queue->used_buffers,0 ); +atomic_set (&queue->set_pci_flags_count,0 ); +atomic_set (&queue->state,QETH_OUT_Q_UNLOCKED ); }return 0 ; }
@@ -3558,7 +3558,7 @@ static void qeth_qdio_output_handler(struct ccw_device *ccwdev,/* prepare the queue slot for re-use: */ qeth_scrub_qdio_buffer (buffer->buffer, -QETH_MAX_BUFFER_ELEMENTS (card)); + queue->max_elements);if (qeth_init_qdio_out_buf (queue, bidx)) {QETH_CARD_TEXT (card,2 ,"outofbuf" );qeth_schedule_recovery (card);
@@ -3705,8 +3705,8 @@ static int qeth_add_hw_header(struct qeth_qdio_out_q *queue, unsigned int hdr_len, unsigned int proto_len, unsigned int *elements) { - const unsigned int max_elements = QETH_MAX_BUFFER_ELEMENTS(queue->card); const unsigned int contiguous = proto_len ? proto_len : 1; +const unsigned int max_elements = queue->max_elements;unsigned int __elements;addr_t start, end;bool push_ok;
@@ -3878,8 +3878,8 @@ static int qeth_fill_buffer(struct qeth_qdio_out_q *queue,QETH_TXQ_STAT_INC (queue, skbs_pack);/* If the buffer still has free elements, keep using it. */ -if (!flush && buf->next_element_to_fill < -QETH_MAX_BUFFER_ELEMENTS (queue->card)) +if (!flush && + buf->next_element_to_fill < queue->max_elements)return 0 ; }
@@ -3959,8 +3959,8 @@ int qeth_do_send_packet(struct qeth_card *card, struct qeth_qdio_out_q *queue,if (queue->do_pack) { do_pack =1 ;/* does packet fit in current buffer? */ -if ((QETH_MAX_BUFFER_ELEMENTS (card) - - buffer->next_element_to_fill) < elements_needed) { +if (buffer->next_element_to_fill + elements_needed > + queue->max_elements) {/* ... no -> set state PRIMED */ atomic_set (&buffer->state,QETH_QDIO_BUF_PRIMED ); flush_count++;
drivers/s390/net/qeth_l2_main.c +1 -1
@@ -581,7 +581,7 @@ static int qeth_l2_xmit_osn(struct qeth_card *card, struct sk_buff *skb, } elements +=qeth_count_elements (skb, hd_len); -if (elements >QETH_MAX_BUFFER_ELEMENTS (card)) { +if (elements > queue->max_elements) { rc = -E2BIG ; goto out; }
Signature · 1 line
2.16.4
PATCH 7/8 s390/qeth: extract helper to determine L2 cast type
In reply to Julian WiedmannThis de-duplicates the L2 and L3 cast-type code, and makes the L2 code a bit more robust by removing the fragile assumption that skb->data always points to the Ethernet Header. This would break in code paths where we pushed the HW header onto the skb.
--- drivers/s390/net/qeth_core.h | 10 ++++++++++ drivers/s390/net/qeth_l2_main.c | 14 +++----------- drivers/s390/net/qeth_l3_main.c | 8 +------- 3 files changed, 14 insertions(+), 18 deletions(-)
drivers/s390/net/qeth_core.h +10 -0
@@ -881,6 +881,16 @@ static inline int qeth_get_ip_version(struct sk_buff *skb) } } +static inline int qeth_get_ether_cast_type (struct sk_buff *skb) +{ +u8 *addr =eth_hdr (skb)->h_dest; + +if (is_multicast_ether_addr (addr)) +return is_broadcast_ether_addr (addr) ?RTN_BROADCAST : +RTN_MULTICAST ; +return RTN_UNICAST ; +} +static inline void qeth_rx_csum (struct qeth_card *card,struct sk_buff *skb,u8 flags) {
drivers/s390/net/qeth_l2_main.c +3 -11
@@ -161,15 +161,6 @@ static void qeth_l2_drain_rx_mode_cache(struct qeth_card *card) } } -static int qeth_l2_get_cast_type (struct sk_buff *skb) -{ -if (is_broadcast_ether_addr (skb->data)) -return RTN_BROADCAST ; -if (is_multicast_ether_addr (skb->data)) -return RTN_MULTICAST ; -return RTN_UNICAST ; -} -static void qeth_l2_fill_header (struct qeth_qdio_out_q *queue,struct qeth_hdr *hdr,struct sk_buff *skb,int ipv,int cast_type,unsigned int data_len)
@@ -611,7 +602,8 @@ static netdev_tx_t qeth_l2_hard_start_xmit(struct sk_buff *skb, rc =qeth_l2_xmit_osn (card, skb, queue); else rc =qeth_xmit (card, skb, queue,qeth_get_ip_version (skb), -qeth_l2_get_cast_type (skb), qeth_l2_fill_header); +qeth_get_ether_cast_type (skb), + qeth_l2_fill_header);if (!rc) {QETH_TXQ_STAT_INC (queue, tx_packets);
@@ -631,7 +623,7 @@ static u16 qeth_l2_select_queue(struct net_device *dev, struct sk_buff *skb,if (IS_IQD (card))return qeth_iqd_select_queue (dev, skb, -qeth_l2_get_cast_type (skb), +qeth_get_ether_cast_type (skb), sb_dev);return qeth_get_priority_queue (card, skb); }
drivers/s390/net/qeth_l3_main.c +1 -7
@@ -1918,13 +1918,7 @@ static int qeth_l3_get_cast_type(struct sk_buff *skb) RTN_MULTICAST :RTN_UNICAST ;default :/* ... and MAC address */ -if (ether_addr_equal_64bits (eth_hdr (skb)->h_dest, - skb->dev->broadcast)) -return RTN_BROADCAST ; -if (is_multicast_ether_addr (eth_hdr (skb)->h_dest)) -return RTN_MULTICAST ; -/* default to unicast */ -return RTN_UNICAST ; +return qeth_get_ether_cast_type (skb); } }
Signature · 1 line
2.16.4
PATCH 8/8 s390/qeth: trust non-IP cast type in qeth_l3_fill_header()
In reply to Julian WiedmannWhen building the L3 HW header for non-IP packets, trust the cast type that was passed as parameter. qeth_l3_get_cast_type() has most likely also used h_dest to determine the cast type, so we get the same result, and can remove that duplicated code.
In the unlikely case that we would get a _different_ cast type, then that's based off a route lookup and should be considered authoritative.
--- drivers/s390/net/qeth_l3_main.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-)
drivers/s390/net/qeth_l3_main.c +3 -8
@@ -1976,19 +1976,14 @@ static void qeth_l3_fill_header(struct qeth_qdio_out_q *queue, hdr->hdr.l3.vlan_id =ntohs (veth->h_vlan_TCI); } + l3_hdr->flags =qeth_l3_cast_type_to_flag (cast_type); +/* OSA only: */ if (!ipv) { - hdr->hdr.l3.flags =QETH_HDR_PASSTHRU ; -if (ether_addr_equal_64bits (eth_hdr (skb)->h_dest, - skb->dev->broadcast)) - hdr->hdr.l3.flags |=QETH_CAST_BROADCAST ; -else - hdr->hdr.l3.flags |= (cast_type ==RTN_MULTICAST ) ? -QETH_CAST_MULTICAST :QETH_CAST_UNICAST ; + l3_hdr->flags |=QETH_HDR_PASSTHRU ;return ; } - hdr->hdr.l3.flags =qeth_l3_cast_type_to_flag (cast_type);rcu_read_lock ();if (ipv ==4 ) {struct rtable *rt =skb_rtable (skb);
Signature · 1 line
2.16.4
David Miller
maintainerFrom: Julian Wiedmann <[email protected]>
Date: Thu, 25 Apr 2019 18:25:53 +0200
Quoted 2 lines
please apply one more patch series for qeth to net-next. Nothing special, just a bunch of cleanups.
Series applied, thanks Julian.